Neurophysiology in Neurosurgery

CONTRIBUTORS

Numbers in parentheses indicate the pages on which the author’s contribution(s) begin.

RICK ABBOT (219), Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York RON L. ALTERMAN (405), Department of Neurosurgery, Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York VAHE E. AMASSIAN (3), Departments of Physiology and Pharmacology, and Neurology, State University of New York, Health Science Center at Brooklyn, New York ALBINO BRICOLO (267), Section of Neurosurgery, Department of Neurological Sciences and Vision, Verona University, Verona, Italy VEDRAN DELETIS (25, 197, 319), Division of Intraoperative Neurophysiology, Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York FRED J. EPSTEIN (55, 319), Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York LEO HAPPEL (169), Louisiana State University Medical Center, New Orleans, Louisiana GEORGE I. JALLO (55), Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York DAVID KLINE (169), Louisiana State University Medical Center, New Orleans, Louisiana

xi 1

xii

Contributors

KARL F. KOTHBAUER (73), Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York MATEVZˇ J. KRZˇAN (153), Department of Neurology, Children’s Hospital, University Medical Center, Ljubljana, Slovenia PATRICK MERTENS (93), Department of Neurosurgery, Hopital Neurologique Pierre Wertheimer, University of Lyon, Lyon, France AAGE R. MØLLER (291), Callier Center for Communication Disorders, University of Texas at Dallas, Dallas, Texas NOBUHITO MOROTA (319), Department of Neurosurgery, National Children’s Medical Center, National Center for Child Health and Development, Tokyo, Japan GEORG NEULOH (339), Department of Neurosurgery, University of Bonn, Germany YASUNARI NIIMI (119), Center for Endovascular Surgery, Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York FRANCESCO SALA (119, 267), Section of Neurosurgery, Department of Neurological Sciences and Vision, Verona University, Verona, Italy JOHANNES SCHRAMM (339), Department of Neurosurgery, University of Bonn, Germany JAY L. SHILS (405), Division of Intraoperative Neurophysiology and Department of Neurosurgery, Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York MARC SINDOU (93), Department of Neurosurgery, Hopital Neurologique Pierre Wertheimer, University of Lyon, Lyon, France TOD B. SLOAN (451), Department of Anesthesiology, University of Texas Health Science Center, San Antonio, Texas MICHELE TAGLIATI (405), Department of Neurology, Beth Israel Medical Center, New York RICHARD J. TOLEIKIS (231), Department of Anesthesiology, Rush-PresbyterianSt. Luke’s Medical Center, Rush University, Chicago, Illinois DAVID B. VODUSˇEK (197), University Institute of Clinical Neurophysiology, University Medical Centre, Ljubljana, Slovenia

ACKNOWLEDGMENTS

Many thanks to David Hershberger and Marco Campos for their editorial help as well as to Dr. Andrea Szelényi, Dr. Adauri Bueno de Camargo and Dr. Klaus Novak for their careful review of the manuscript and for their insightful suggestions. Special thanks to Linda and Carlos Schejola for their generous support which has given us the opportunity to build equipment for simultaneous recording of neurophysiological and intraoperative neurosurgical data.

xv

PREFACE

Prior to the advent of current intraoperative neurophysiological monitoring methodologies, it was difficult for neurosurgeons to determine the extent to which a tumor, such as a low-grade malignancy spinal cord tumor, should be removed. Ten years ago it was not uncommon for many patients, even those in whom only partial tumor removal had been performed, to experience postoperative deficits like paraplegia or even quadriplegia. Today, intraoperative neurophysiological (ION) methodologies for monitoring the motor system can provide surgeons with real-time data that has been shown to consistently correspond with post-operative motor status. Surgeons now have at their disposal a reliable criteria for determining the extent of tumor removal and for preventing serious intraoperative neurological injury to the motor system. This ability to ascertain the motor pathways’ functional integrity has become one of ION’s most important achievements, fulfilling its primary goals of prevention and documentation of intraoperatively-induced neurological injury. As new ION methodologies develop, our ability to monitor and identify (map) different nervous system structures has also significantly evolved. Today, the wealth of data being drawn directly from exposed nervous tissue is rapidly creating a second goal for ION: expanded exploration of the nervous system’s physiology. Thanks to the introduction of these new methodological approaches, the pursuit of this goal promises to contribute new knowledge about parts of the nervous system which were previously inaccessible. Since modern ION methodologies are finally providing reliable data concerning the integrity of the motor and other parts of the nervous system, it has increasingly become an important tool in several surgical disciplines. ION is now an interdisciplinary field incorporating knowledge and experience from neurosurgery, neurology, orthopedic surgery, neurophysiology, anesthesiology, interventional neuroradiology, and biomedical engineering.

xiii

xiv

Preface

Furthermore, new experiences in functional and stereotactic neurosurgery for the precise targeting of deep-brain structures and chronically implantable deep-brain electrodes have refined the treatment of movement disorders. Precise physiological targeting now allows the neurosurgeon to access deep brain areas where minor errors can make the difference between success and failure. Achievements in ION have been recognized by the neuroscience community (Clinical Examinations in Neurology, Mayo Clinic and Mayo Foundation, Mosby Year Book, 1991; Neurosurgery: The Scientific Basis of Clinical Practice, Blackwell Science Publ., 1998; Neuroprotective Agents, Annals of New York Academy of Science, Vol. 939, 2001). Indeed, thanks to technological advancements and the rapid growth of methodologies over the last ten years, we believe ION is traversing a fruitful and important period that will continue to strengthen and establish its relevance in the medical community. We are pleased to present 17 chapters dealing with current developments in ION, most of them written by peer-recognized experts from around the world. We invite the reader to make extensive use of the accompanying CD at the end of the book, which provides video and audio/visual examples of some of the methodologies presented in the book. We believe that it is a powerful educational and instructional tool. We hope that this book will continue to stimulate interest in the integration of ION into neurosurgery, and consequently, continue to decrease the number of patients experiencing intraoperatively induced injury to the nervous system. New York, NY Summer, 2002

Vedran Deletis Jay L. Shils

Cover image: Mrs. San-San Chiang, Senior Neurophysiological Technician, checking electrodes attached to a patient’s head which has been fixed in a frame.

CHAPTER

1

Animal and Human Motor System Neurophysiology Related to Intraoperative Monitoring VAHE E. AMASSIAN Departments of Physiology and Pharmacology, and Neurology, State University of New York, Health Science Center at Brooklyn, New York

1 Introduction 2 Corticospinal Responses 2.1 Configuration of CT Waves 2.2 Eliciting D Waves 2.3 Eliciting I Waves 3 Muscle Responses References

ABSTRACT This chapter uses data from animal models and human subjects to describe some physiological principles underlying intraoperative spinal cord monitoring of the motor pathways. In the first type of monitoring, conducted impulses in the corticospinal tract (CT) are recorded following transcranial electrical stimulation (TES) or transcranial magnetic stimulation (TMS). Single pulses elicit direct (D), i.e., unrelayed CT discharges, which are followed, if the anesthesia is light, by multiple indirect (I) waves that are transsynaptically generated in motor cortex. Corticocortical afferent inputs from parietal areas generate the first I wave, and subsequent I waves result from excitation by a vertically oriented interneuron chain in the motor cortex, which functions like a “clock” in quantizing time in periods of 1.3–2.0 ms. The D wave increases in amplitude monotonically with TES or TMS intensity; over portions of the relation between stimulus intensity and D response, the curve is very approximately linear, so that a percent decrease during an operation approximately reflects a block in conduction proportionately in the number of conducting CT fibers. By contrast, muscle activation by the CT volleys involves a highly nonlinear transfer function from CT to motoneuron, which may render this measure oversensitive

Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

3

4

Vahe E. Amassian to CT damage. Anesthesia reduces not only tonic facilitatory bombardment of the motoneuron, but also I wave components, thereby diminishing temporal facilitation of the motoneuron. Cooling the motor cortex reduces, especially later, I waves and leads to loss of the muscle response. Such diminished excitation of the motoneuron can be counteracted by using a high-frequency train of TES, which at the appropriate interstimulus interval can lead to excitatory postsynaptic potential (EPSP) summation and motoneuron discharge.

1 INTRODUCTION Monitoring motor responses for stimulating the human cerebral cortex has clearly been used for much more than clinical mapping studies of the exposed brain at operation [1]. The discovery that transcranial electrical stimulation (TES) of the human motor cortex could cause muscle activation [2] and its replacement in (awake) humans by the less painful transcranial magnetic stimulation (TMS) greatly expanded the opportunities for motor pathway monitoring. The current uses include the following: 1. Intraoperative monitoring was the focus of this and other presentations during the symposium entitled “Intraoperative Neurophysiological Monitoring in Neurosurgery” (Second International Symposium in New York City, November 20–21, 2000). Such recordings immediately affect surgical procedures and also aid in predicting outcome [3]. 2. Preoperative mapping of motor representation, e.g., prior to cortical removal or implanting of a grid for cortical stimulation. The combination of surface visualization of the cerebral cortex with appropriate MRI software and mapping with focal TMS appears especially advantageous [4]. 3. Diagnosis and pathophysiological research, e.g., of lesions of the motor system such as those caused by multiple sclerosis, amyotrophic lateral sclerosis, strokes, myoclonus, etc., have yielded unique data on central motor delays [5]. However, rarely has the clinical diagnosis depended on the electrophysiological monitoring. Rather, such monitoring has provided a quantitative functional measure of the effect of the lesion, which may assist in correlating the imaged extent of the lesion and the results of neurological examination. Such a comparison potentially aids in explaining the underlying pathophysiology of motor system disease and the changes in brain function, i.e., “reorganization” following lesions [4, 6, 7]. In intraoperative motor system monitoring, the physiological and pathophysiological mechanisms concern two types of response to stimulation of the cerebral cortex. The first, recorded from the corticospinal tract (CT), is relatively insensitive to anesthetic conditions and is “approximately” linearly related to stimulus intensity, but it is of low amplitude, and the recordings often

Animal and Human Motor System Neurophysiology

5

include troublesome stimulus artifacts requiring special attention. In the second type of recording, the electromyography (EMG) of transcranially activated muscles is of greater amplitude than the CT response and therefore is easily recorded by widely used techniques, but it is highly sensitive to anesthetic level and requires particular attention to the parameters of cortical stimulation, to which it is nonlinearly related, and a high-frequency short train of stimuli is usually needed.

2 CORTICOSPINAL RESPONSES

2.1 CONFIGURATION OF CT WAVES 2.1.1 Conducted Impulses Soon after population-conducted impulses were recorded in cats and monkeys, it was apparent that recordings from the surface of the medullar pyramid or the dorsum of the spinal cord differed markedly from those among the CT fibers [8]. With single electrical pulse stimulation of the motor cortex, recordings from the pial surface reveal a brief positive–larger negative deflection of the wave, whose latency is clearly accounted for by conduction time in fast CT fibers; i.e., the wave is direct (D) without intervening synapses (Fig. 1.1, left).

FIGURE 1.1 Conducted and blocked D and I responses in squirrel monkey and conducted responses in humans. Left, indicated number of summed CT responses to motor cortical stimulation. Monkey anesthetized with pentobarbital. Middle and right, epidural recordings from humans at indicated cervical level; TMS stimulation was oriented to induce either a lateromedial (L-M) or postero-anterior (P-A) electric field. At left, CT response increases in amplitude many-fold when conducted response is blocked. On the right, interrupted lines indicate the “correction” factor applied to I waves relative to D waves recorded in humans. Absolute amplitude of all waves would increase. Left records reprinted from [10]; middle and right records reprinted from [11].

6

Vahe E. Amassian

Following the D wave are less well defined and usually much smaller waves, which upon appropriate analysis (see Section 2.3) were proved to be indirect (I), being CT fiber discharges following transsynaptic activation in motor cortex. Depending on the parameters of cortical stimulation, such conducted D waves, with or without I discharges, constitute the great majority of epidural recordings in humans. The explanation for the D wave configuration readily derives from classical volume conductor theory. When the focal recording is referenced against a “distant” electrode, the approach of an impulse in each CT fiber is signaled by an initial positivity; the isoelectric transition signals the arrival of the impulse, and the subsequent negativity results from fiber activation under the focal electrode. Finally, propagation of the impulse beyond the focal electrode with recovery toward the resting potential yields a low-amplitude, long-duration positivity. Whereas in animal preparations the final positivity may be barely distinguishable, in human D recordings it is usually more prominent than the initial positivity, even when differentiation of the preceding negativity is avoided by using an adequate band pass at the low-frequency end [9]. Thus, given the triphasic potential associated with single CT fiber recordings, the amplitude of the population D wave reflects a number of factors: (1) The first of these is the synchronicity of the potentials in each fiber; slowing of conduction velocity in part of the responding population would lead to additional reduction in peak amplitude through phase cancellation. A reduced D wave amplitude accompanied by a broadening of the negative component might evidence diminished synchronization. (2) The second factor is the number of activated, fast CT fibers. (3) The action potential amplitude in individual fibers may be changed by physiological phenomena such as refractoriness or supernormality [9] or by damage.

2.1.2 Blocked Impulses When a semi-microelectrode is inserted into the CT (or pyramid), the configuration of the D wave is markedly altered (Fig. 1.1, left). The polarity of the D wave becomes substantially monophasic with a duration approximating the previous (+ − +) wave [10]. The blocked region acts only as a source for the approaching sink, or more accurately, the injury potential is briefly reduced. Furthermore, the amplitude of the D wave substantially increases (nearly 5 × in Fig. 1.1). However, a much greater increase in amplitude (× 15.5) of the I wave is recorded; i.e., a threefold increase in the I:D ratio can occur when a conducted CT response is blocked. In Fig. 1.1, right, this “correction” factor is applied to conducted human CT responses [11]. Such a relative increase in the amplitude of a blocked response is usually attributed to lessened phase cancellation of slightly asynchronous triphasic potentials when these are converted by

Animal and Human Motor System Neurophysiology

7

FIGURE 1.2 Relationship of electrical stimulus intensity to PT D wave amplitude. Subcortical white matter of cat stimulated with 0.19 ms pulses at indicated current strengths via bipolar electrodes. Reprinted from [13].

block to monophasic potentials; although this may be an important factor, it may not be the only factor [12]. The recording of a monophasic D wave provides a basis for constructing relationships between stimulus intensity and D wave response [13]. Electrical stimulation of the subcortical white matter at increasing intensity yields a monotonic relationship to the D wave amplitude (Fig. 1.2). Theoretically, a group of fast CT fibers, with diameters (and therefore thresholds) normally distributed around a mean value, would, exposed to the same electrical stimulus, yield a sigmoid relation (i.e., the integral of the normal distribution function) between stimulus intensity and population D response. Experimentally, the D wave response only approximates a sigmoid function. However, the CT in cats, monkeys, and humans contains a wide range of fiber diameters, and an electrical stimulus in the white matter cannot be expected to affect fibers uniformly; although the largest CT fibers that were closest to the stimulating electrodes would first be excited, at increasing intensities nearby higher-threshold fibers and distant low-threshold fibers (subject to an inverse square reduction in intensity) would both be stimulated. Further complicating the stimulusresponse relationship is that the lowest threshold CT fibers first excited would be expected to contribute the largest action potentials to the population response. Thus the amplitude of the population D wave only approximately measures the number of synchronously active fibers (cf. the EMG response). With magnetic stimulation (Fig. 1.3), the large stimulus artifact precludes eliciting a maximum D wave, but at submaximal stimulus intensities, the D wave clearly increases monotonically [10] in the “quasilinear” manner of the intermediate range of intensities in Fig. 1.2 [13]. Given that insertion of a recording electrode into the human CT would not be justifiable in routine monitoring, the question arises as to whether there are any serious disadvantages in being restricted to recording conducted responses.

8

Vahe E. Amassian

FIGURE 1.3 Effect of changing TMS stimulus intensity on CT D wave responses. Monkey maintained under pentobarbital anesthesia. Round coil (9.4 cm outer diameter) in lateral-sagittal position so one set of windings over CT projections. One hundred percent output equivalent to 2.2 T. Five responses superimposed on each trace. Reprinted from [14].

Clearly, the normal conducted D wave amplitude remains a valuable measure of the number of conducting fibers, limited only by the factor of pathological desynchronization and reduced action potential amplitude in the individual fibers. However, the major reduction in I:D wave amplitude ratio renders conducted CT activity a technically unacceptable measure of thresholds of D and I activation with focal anodal versus cathodal stimulation or electrical versus magnetic stimulation. The conversion toward a monophasic potential with fiber injury permits a functional identification of the site of experimental traumatic injury to the spinal cord [15], which can be compared with the rostral limit of the contusion (Fig. 1.4). A similar change toward positivity was caused by a spinal cord tumor [16]. It would be valuable to record the potential changes in the epi- or subdural D wave when recorded at several levels (i.e., above, at, or just below a traumatized level of spinal cord), to compare the “electrophysiological” level with the MRI and neurological levels. Such measurements would also serve to validate applying to the human the I/D amplitude correction for blocked CT impulses deduced from the monkey (Fig. 1.1, right).

2.2 ELICITING D WAVES Transient electrical stimulation of motor cortex in animals and human readily elicits D activation, but there are important differences in the efficacies and sites of activation related to the intensity, the polarity (+ or −) of the focal stimulus, and the species. In cats, the CT neuron may be excited at three sites: at the initial segment (IS) region and adjoining membrane, at its recurrent axon collaterals (with increased latency through slowed propagation), and in white matter.

Animal and Human Motor System Neurophysiology

9

FIGURE 1.4 Effect of weight drop on spinal cord and on feline CT response recorded epidurally at indicated distances above the contused area in photograph. Cat anesthetized with pentobarbital. Weight drop was 47 g × 7.2 cm, producing a 96% block in conduction. Effect of impact occurs within 1 s, precluding vascular factors or edema in the loss of conduction. Two hundred responses summed on each trace. Reprinted from [15].

Hern et al. [17] described in monkeys that focal anodal stimulation could elicit D excitation alone in conducted impulse recordings; however, later recordings of blocked CT fibers revealed that when the stimuli were a little above threshold (e.g., 50%), both D and I waves were elicited by either polarity of stimulation [10]. Nevertheless, in the monkey, focal anodal and cathodal stimulation generates D waves at different sites; with cathodal stimulation, at a given intensity D waves are highly variable, while they are almost constant in amplitude with anodal stimulation (Fig. 1.5). Presumably, with cathodal stimulation, D activation can occur close to the spike trigger zone, e.g., at the IS and adjoining membrane, while with anodal stimulation, it occurs at a white matter site electrotonically isolated from the synapses. Intracellular recording in the cat supports this explanation, threshold bipolar stimulation generating a D impulse when CT and other motor cortical neurons are relatively depolarized (Fig. 1.6); when less depolarized, transsynaptic excitation was necessary [18]. The site within white matter where focal anodal stimulation elicits D activation at the lowest threshold is now believed from modeling studies to be at bends in the CT fibers [19, 20]. Facilitation occurs when just threshold intensities of anodal TES and TMS are combined only at brief intervals (e.g., 30 µs; [21]). The site of electrotonic summation of the two stimuli is inferred to be at

10

Vahe E. Amassian

FIGURE 1.5 Effect of changing surface anodal and cathodal stimulus intensities on CT responses. Squirrel monkey lightly anesthetized with pentobarbital, three superimposed responses on each trace. Reprinted from [10].

FIGURE 1.6 Relationship of site of D activation to membrane potential of PT neuron. At right, intracellular recording of PT neuron, which follows 1:1 at 500 Hz antidromic stimulation. (A) and (B) show latency variation with threshold stimulation at motor cortical sites A and B. (C) column is an intensity series. (D) shows clear onset of EPSP. Middle, above shows latency of first discharge plotted as a function of membrane potential; below, latency distributions with stimulation at sites A and B. At left, arrows in diagram show excitation at IS-axon hillock region, at recurrent collateral (initial discharge in middle, bottom histogram) and at bend in fiber in white matter. Reprinted from [18].

Animal and Human Motor System Neurophysiology

11

bends of CT fibers in the white matter, where the directions of the orthogonal electric fields would both lead to outward current from the node. What implication do these findings have in the human? The great majority of conducted D wave recordings will depend on TES, rather than stimulating the surgically exposed pial surface of motor cortex. Necessarily, the electric field is developed over a wide area of cortex and subcortex because the closer the interpolar distance, the greater the proportion of current shunted through the superficial tissues. Given the high resistance of the cranium, there seems no advantage in using electrodes with small surface areas. Our standard technique uses a focal scalp electrode (dimension: 2 × 2 cm), over motor cortex, with a large semicircular reference electrode (area: 25 cm2) at an interpolar distance of approximately 6 cm [22]. Muscle responses during voluntary contraction are obtained at a higher threshold with focal cathodal than with anodal TES, but the increased current strengths required are greater than expected from the differences in D threshold in monkeys or in humans [23]. Monitoring stimulating current, e.g., by an inductive probe on a stimulating lead (P6016, Tektronix, Inc., Beverton, OR), is valuable in monitoring stimulus intensity. Electrode impedances are readily assessed by substituting for the subject an appropriate precision resistor (e.g., 1 KΩ); the electrode impedances are then calculated from the ratio in currents flowing at the given voltage output of the stimulator. Initially, 25 µm-thick stainless steel electrodes were used, but these proved unsuitable because if bent, they no longer conformed to the scalp; aluminum foil has proved a satisfactory substitute because it is easily flattened and does not distort brief electrical stimulating pulses (e.g., 0.2 ms in duration). Practical considerations may require specific stimulating electrode montages appropriate for the requirements of a particular group of patients (cf. the use of Corkscrew electrodes, Nicolet, Madison, WI [9]). The sites of stimulation C3 vs. C4 are of particular interest because of the likelihood of D excitation at both sites at the stimulus intensities used; this excitation increases the size of the CT D wave through summation. A possible problem could arise if the combined D wave amplitude fell by x% due to unilateral pressure on one lateral column and actually resulted from twofold damage to that column. A suggested montage to identify such a problem would be to connect the (−) output of the stimulator to the posteriorly sited large reference electrode and the (+) output to a two-pole, three-way switch permitting a connection to (1) both focal anodes; (2) the right anode; and (3) the left anode. Alternatively, if a focal anodal pulse were delivered 1 ms after the first pulse at the same site, the CT fibers would be in the refractory period [9]. Therefore, focal anodic stimuli given at C3 and C4 1 ms apart should generate independent D waves. The initial animal experiments on the CT were all conducted under varying levels of barbiturate or chloralose anesthesia, without any description of anesthetic effects on the D wave amplitude. However, it is evident that an increased

12

Vahe E. Amassian

anesthetic level could reduce D wave amplitude of cathodally elicited responses if the level of a net steady-state facilitatory bombardment of CT neurons were depressed. Unfortunately, there is no evidence in humans that focal cathodal (or anodal) stimuli preferentially excite CT neurons in the grey matter where transsynaptic conditioning effects could occur. On the contrary, the earliest D wave recordings indicated increasingly subcortical sites of excitation with increasing focal anodic stimulus intensities [24–26]. Therefore, the reported depression of CT responses to anodal TES excitation at higher anesthetic levels would reflect either a direct effect on CT fibers [25] or an indirect effect, e.g., through anesthetic-induced redistribution of fluids and therefore of stimulating currents within the brain [12]. Significantly, the anesthetic mediated effect on D wave amplitude under 2% isofluoride caused a reduction to 46% with TES, but only to 80% with stimulation of the exposed motor cortex. Furthermore, an increased level of anesthesia increases the latency of the CT response to TES, thereby supporting the hypothesis that the site of TES stimulation shifts with anesthetic level. The above account has focused on the use of TES rather than TMS in eliciting D waves, although it has long been known that the appropriate lateral orientation of the coil can elicit muscle responses with latencies matching those with focal anodic TES [22, 27]. Proof of preferential direct CT excitation with the appropriate orientation of the coil was secured in the monkey [14] and in humans [11]. Although there appears to be little advantage of TMS over TES during surgery, the use of TMS facilitates preoperative tests on the same patient prior to anesthesia.

2.3 ELICITING I WAVES The monumental Golgi studies of connectivity in cerebral cortex by Lorente de Nó [28] led to the hypothesis that repetitive I waves result from excitatory bombardment of CT neurons by interneuron chains [8]. Subsequently, other mechanisms have been proposed, which will be briefly reviewed later in this section. Many powerful synaptic inputs impinge either directly or indirectly on large CT neurons, including (1) extrinsic inputs from (a) inter-areal corticocortical afferents, (b) callosal afferents, and (c) specific thalamocortical afferents; and (2) intrinsic inputs from local motor cortical neurons. A difficulty in analyzing the source of transsynaptic excitation of CT neurons when electrical (or magnetic) stimuli are applied over the motor cortex is not knowing whether the CT neurons respond to one or other of the extrinsic inputs or to an intrinsic input. In the following account, it is presumed that if I waves are elicited by distant cortical stimulation, they are most likely generated by corticocortical inputs; if loss or production of particular I waves occurs selectively with procedures applied

13

Animal and Human Motor System Neurophysiology

locally to motor cortex, such as cooling or microelectrode stimulation, then intrinsic inputs may elicit such I waves. 2.3.1 Extrinsic Inputs Among powerful extrinsic inputs, some can readily be eliminated as necessary for generating I waves. Thus excitation of specific thalamocortical afferents derived from Nucleus ventralis lateralis (VL) and anterior (VA), which are mono- and disynaptically excitatory to large CT neurons, and those from Nucleus ventralis posterior (VP), which are polysynaptically excitatory [ 291, was eliminated in cats by massive radio frequency lesions of these thalamic nuclei. After allowing lo-27 days for degeneration of the afferent fibers, I activation of pyramidal tract (PT) fibers was still obtainable with pericruciate stimulation of motor cortex [lo]. It would clearly be of interest to stimulate motor cortex with TMS in patients who previously had unilateral Nucleus VL-VA lesions for Parkinsonism to determine if the threshold for motor responses was unelevated. Among corticocortical inputs, those in the human from the contralateral hemisphere clearly can have excitatory effects, as when visual information is projected onto the nondominant hemisphere and subsequently transferred transcallosally for vocalization. However, the predominant transcallosal effect of single-pulse TMS is inhibitory [30]. By contrast, large multiple I waves were readily elicited in monkey CT by electrical stimulation of parietal or premotor cortex (Fig. 1.7 [8,31]). In humans, facilitatory interaction between combined near-threshold parietal and premotor cortical TMS is elicited at time intervals excluding any direct summation of the electric fields [32]. Therefore, the TMS and most likely TES over motor cortex most likely elicit multiple I wave responses through generating an incoming volley in corticocortical afferent fibers. BEFORE

AFTER

MOTOR

ABLATION

CORTICAL

ABLATION

FIGURE 1.7 Corticospinal responses in monkey to electrical stimulation of cortical loci before (above) and after (below) precentral ablation. The inset shows only D response when white matter exposed by ablation is stimulated. Reprinted from [31].

14

Vahe E. Amassian

Significantly, when the lateral windings of a circular coil centered on the vertex overlie motor cortex, muscle activation occurs a minimum of 1–2 ms later than with focal anodic stimulation; the later activation by TMS is manifest also in single motor unit recordings; i.e., it does not reflect activation of slower motoneurons but their activation by I discharge [33]. With TMS applied through a figure 8 coil, optimal motor response occurs when the midpoint of the junction region of a figure 8 coil is close to the interaural line, i.e., over motor cortex and approximately at 90° to the central sulcus [34, 35]. The failure of single-pulse TMS, applied parasagittally over motor cortex, to excite directly the largest neurons in motor cortex, the Betz cells, which lie in a variety of orientations in the banks and crown of the folds of the precentral gyrus, makes it very unlikely that smaller neurons are directly excited. Modeling experiments with a peripheral nerve in a skull volume conductor [19] implies that excitation would preferentially occur at the bend in the corticocortical fibers toward grey matter. Excitation occurs at a bend optimally by the induced electric field at its peak and not at its spatial derivative, as with a linear nerve [20]. As indicated previously, optimal TMS parasagittal excitation occurs when the center of the junction region, i.e., the peak electric field, is over motor cortex. Furthermore, optimal TMS excitation occurs when the direction of the induced electric field is posterior-anterior (P-A) [27, 33], i.e., appropriate for current exiting the bend of parietal corticocortical fibers. Responses from corticocortical fibers from more anterior portions of the frontal lobe, such as Area 6, are optimally elicited when the induced field is A-P directed, i.e., exiting the bend of the posteriorly directed afferent fibers. It must be emphasized that the difficulty of TMS in directly exciting neuropil in grey matter applies only to single stimuli. For example, when a nearthreshold TMS pulse applied in the P-A orientation over motor cortex is briefly (e.g., 1–2 ms) followed by a weaker second pulse, (e.g., 70% of the intensity of the first), the first dorsal interosseous (FDI) response is markedly facilitated [36, 37]. Because the corticocortical fibers excited by the larger first TMS pulse would be refractory to the even weaker second pulse, other neural elements whose threshold was reduced by the transsynaptic effect of the first pulse most likely mediated the facilitation. A likely site where the electric field induced by the second TMS pulse and EPSPs from corticocortical action of the first could interact is the initial segment and adjoining membrane of the motor cortical neurons. A major question that arises is whether the multiple I waves result from the transsynaptic conductance changes induced by a single volley in thalamocortical [38] or corticocortical fibers, or whether they reflect excitation by intrinsic motor cortical neurons or a combination of both monosynaptic (corticocortical) and polysynaptic (local interneuronal) excitations. Clearly, invoking monosynaptic activation of CT neurons by a synaptic input requires recording the minimum EPSP delays and attaining subsequent delays for firing level in PT neurons

16

Vahe E. Amassian

unlikely that a prolonged conductance change alone generates the I waves. Furthermore, in humans, two near-threshold TMS pulses presented at intervals that are harmonics of the I wave periodicity yield optimal facilitation of FDI responses [43]. With four near-threshold TMS pulses, a “tuning” curve (Fig. 1.8) of FDI facilitation was demonstrated, which peaked at an interstimulus interval of 1.3–1.7 ms in the four subjects tested. By contrast, facilitation tested in one of these subjects with four focal anodic stimuli was absent at 1.7 ms but prominent at 2.7 ms, indicating that the preferred timing of facilitation with TMS depends on motor cortical rather than spinal cord circuitry [12]. (b) Microelectrode stimulation at different depths within monkey motor cortex preferentially elicits late I waves superficially; early I waves are added when the stimulation is in the deeper laminae (Fig. 1.9, middle). In cat and monkey, bipolar microwire stimulation that generates an electric field with a major transverse component elicits an I1 wave when close to lamina V (Fig. 1.10), possibly through stimulation of axons in the inner line of Baillarger [10]. (c) Cooling the pial surface of monkey motor cortex selectively causes the reversible loss of late I waves without reducing the earliest I wave (Fig. 1.9, right). Taken together, (b) and (c) imply that the excitatory motor cortical neurons are vertically oriented, as was earlier inferred by Lorente de Nó [28]. The differing latencies of, e.g., FDI response with P-A versus A-P oriented TMS pose a problem because of the descriptions of corticocortical afferents as

FIGURE 1.9 Differential effects on initial versus later I waves in monkey. Middle, microelectrode stimulation at indicated depth in motor cortex elicits later I waves superficially (with D wave). Right, progressive cooling nearly abolishes later I waves, but not I1. Left, diagram of possible vertically oriented, excitatory interneuron circuit, which may underlie the multiple I waves. Left diagram reprinted from [12]; middle records from [8]; right records from [10].

Animal and Human Motor System Neurophysiology

17

FIGURE 1.10 Monosynaptic (I1) excitation in monkey and cat by bipolar microwire stimulation in deep laminae. Stimulation throughout the Teflon-insulated, adjoining 50 µm wires generated an electric field that was mainly tangentially oriented. Five responses superimposed in each trace at the indicated depth of stimulation. Left, monkey motor cortex penetrated at an angle of 55° from the radial axis. Marking lesion made at optimal site (4.9 mm) was found in the histological section to be just deep to lamina V. Middle, cat pericruciate cortex stimulated at indicated depth. Recordings from the corticospinal tract showed optimal I1 with small D wave at 1.5 mm depth; histological section indicated that this depth was below lamina V. Right, pericruciate cortex of another cat stimulated with recording from medullary pyramid. Top of marking lesion at optimal site (1.5 mm) for I1 just below lamina V. Reprinted from [10].

terminating throughout the laminae [28, 44, 45]. One possibility is that corticocortical inputs from parietal lobe and premotor cortex do not similarly terminate in all laminae; e.g., Jones et al. [46] described corticocortical afferents as terminating only in the superficial laminae of neighboring somatosensory cortex, and Chang [47] described both callosal and association corticocortical fibers as terminating in the superficial laminae. Alternatively, the classical descriptions of the site of termination may apply to motor cortex, but the synaptic efficacies of deeper termination may not be adequate for monosynaptic (I1) activation from premotor sources. A final comment on the possible function of the motor cortical interneuronal circuitry is warranted, given the present focus on human spinal cord monitoring. We have previously proposed that the multiple high-frequency I waves function in a manner analogous to a computer clock, quantizing cortical time [10]. Nowhere is the synchronization of CT discharges at a relatively fixed period so remarkable as in human I waves recorded at great conduction distance from the motor cortical circuitry that generates the period [25, 48].

18

Vahe E. Amassian

3 MUSCLE RESPONSES With the discovery by Merton et al. [2] that single-pulse TES in the awake human could elicit a muscle response in voluntarily contracting muscle, the question arises as to the necessary relationship between CT and motoneuron discharges. Brookhart [49] had earlier emphasized the role of temporal facilitation in cats and monkeys in eliciting motor responses by a train of electrical stimuli to the medullary pyramid. While anodal TES at high intensity would be expected to add I waves to the initial D discharge from a single-pulse TES (see Section 2), this seems not to be the crucial factor in eliciting the muscle response in the awake human, because the short latency of the response reflects D and not later I activation. The critical factors in the muscle response are the size of the D volley, i.e., the degree of spatial facilitation and how close to the firing level of the α-motoneurons are their membrane potentials. In the anesthetized patient, the background depolarization from voluntary activity must be replaced by another source, i.e., temporal facilitation that is achievable by a train of CT volleys. The basis of temporal facilitation by a train of CT volleys was shown by Phillips and Porter [50] in intracellular recordings from baboon α-motoneurons to result from summation of successive EPSPs that mount to the firing level. Unfortunately, in their experiments, the anodal motor cortical stimuli, in addition to the train of D volleys, elicited also I discharges because of facilitation at the cortical level [51]. (Evidently, the component of temporal facilitation at the motoneuron would best be estimated by stimulation at the level of the medullary pyramid.) Nevertheless, temporal summation of EPSPs is clearly of major importance in securing motoneuron discharge. The length of the CT train of volleys determines whether the mounting EPSPs actually reach firing level. Thus selectively reducing the latest I activity by cooling the pial surface of monkey motor cortex abolishes the muscle response (Fig. 1.11). Temporal facilitation during spinal cord monitoring could be secured by either a train of stimuli each eliciting a D volley, or potentially with fewer stimuli but each generating a D volley plus repetitive I waves. A major determinant in the human of the amplitude of I wave responses is the level of anesthesia [25, 26, 51]. Thus, under deep anesthesia, temporal facilitation can only be secured by a train of D volleys. It must be emphasized that increasing the depth of anesthesia would reduce not only I activity but also tonic facilitatory bombardment of α-motoneurons and therefore reflex responses to a given sensory stimulus. The optimal period in a train of D volleys for activating muscle is a resultant of several factors. If the period is too long, the EPSPs will have largely decayed before the next CT volley arrives; if too short, refractoriness reduces the size of the CT volley. However, refractoriness would be absent at a period of 4 ms and even at the shorter period with a long-duration (e.g., 0.5 ms) TES pulse [9].

Animal and Human Motor System Neurophysiology

19

FIGURE 1.11 The effect of cooling the pial surface of the motor cortex on paired corticospinal D with multiple I waves (above) and thenar muscle responses (below) in monkeys. Three responses superimposed in each trace. First series (bottom left) was taken immediately after the onset of cooling; second series (top right) was started 35 s later; third series (bottom right) was taken immediately after the second. Reprinted from [10].

The deeper the anesthesia, the longer would be the duration of the motoneuron membrane time constant because of the reduced resting conductance level. Thus, with an approximate decay time constant unlikely to be less than 4 ms, a 250 Hz train would still provide substantial facilitation. Under light anesthesia, when the D wave is followed by multiple I waves, optimal conditions for temporal facilitations would be determined by additional factors. The I wave period is often as brief as 1.4 ms, probably because EPSPs at the spike trigger zone overcome relative refractoriness from the previous action potential. In addition, I activation is known to occur in CT neurons that were not previously D activated [8, 18] or were activated after a double I period [14]. The optimal interstimulus period would be determined by the duration of the combined D and multiple I waves discharges, which is likely to exceed 5 ms. Little would be gained by using a shorter interstimulus period, because of occlusion between D and I waves with those from an antecedent stimulus. Another major difference between monitoring muscle versus CT response is the nonlinear relationship between TES intensity and the muscle response. In the monkey, the muscle response appears abruptly with a small increase in stimulus intensity and CT response (Fig. 1.12). As a result, the muscle response is potentially more sensitive to spinal cord damage than the CT response, not only because of the transfer function from the CT volleys to motoneurons, but

20

Vahe E. Amassian

FIGURE 1.12 The effect of increasing stimulus intensity applied to motor cortex of monkey on corticospinal D and I waves and thenar muscle EMG responses. Ten responses superimposed in each trace. Anesthesia was pentobarbital. The bipolar stimulus (duration 100 µs) was applied at the indicated current strength to the optimal cortical site for thenar muscle responses. Reprinted from [10].

also because a reduction in tonic facilitation may reduce the efficacy of the CT volleys still further. Finally, it should be noted that important aspects of the relationship of CT activity to motoneuron discharge require investigation at the cellular level. For example, it is unknown when repetitive CT volleys in the same presynaptic fibers fail to maintain quantal transmitter release, or even invade terminals as studied at group Ia endings [52]. At rates where failure occurs, increasing intensity during a TES train would, by recruiting additional presynaptic fibers, provide spatial facilitation, thereby increasing the efficiency of stimulation for a given total charge administered.

REFERENCES 1. Penfield, W., and Boldrey, E. (1937). Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation. Brain, 60, 389–443. 2. Merton, P.A., Hill, D.K., Marsden, C.D., and Morton, H.B. (1982). Scope of a technique for electrical stimulation of human brain, spinal cord and muscle. Lancet, ii, 597–598.

Animal and Human Motor System Neurophysiology

21

3. Deletis, V. (2002). Intraoperative neurophysiology and methodology for monitoring the motor system. In “Neurophysiology in neurosurgery: A modern intraoperative approach” (V. Deletis, and J. Shils, eds.), pp. 23–51. Academic Press, San Diego. 4. Levy, W.J., Amassian, V.E., Schmid, U.D., and Jungreis, C. (1991). Mapping of motor cortex gyral sites non-invasively by transcranial magnetic stimulation in normal subjects and patients. In “Magnetic motor stimulation: Basic principles and clinical experience” (W.J. Levy, R.Q. Cracco, A.T. Barker, and J. Rothwell, eds.), EEG Suppl. 4, pp. 51–75. Elsevier Science Publishers. 5. Barker, A.T., Freeston, I.L., Jalinous, R., and Jarratt, J.A. (1987). Magnetic stimulation of the human brain and peripheral nervous system: An introduction and the results of an initial clinical evaluation. Neurosurgery, 20, 100–109. 6. Levy, W.J., Amassian, V.E., Traad, M., and Cadwell, J. (1990). Focal magnetic coil stimulation reveals motor cortical system reorganized in humans after traumatic quadriplegia. Brain Res., 510, 130–134. 7. Cohen, L.G., Bandinelli, S., Findlay, T.W., and Hallett, M. (1991). Motor reorganization after upper limb amputation in man. Brain, 114, 615–627. 8. Patton, H.D., and Amassian, V.E. (1954). Single and multiple unit analysis of cortical stage of pyramidal tract activation. J. Neurophysiol., 17, 345–363. 9. Deletis, V., Isgum, V., and Amassian, V.E. (2001). Neurophysiological mechanisms underlying motor evoked potentials in anesthetized humans: Part 1. Recovery time of corticospinal tract direct waves elicited by pairs of transcranial electrical stimuli. Clin. Neurophysiol., 112, 438–444. 10. Amassian, V.E., Stewart, M., Quirk, G.J., and Rosenthal, J.L. (1987). Physiologic basis of motor effects of a transient stimulus to cerebral cortex. Neurosurgery, 20, 74–93. 11. Kaneko, K., Kawai, S., Fuchigami, Y., Morita, H., and Ofuji, A. (1996). The effect of current direction induced by transcranial magnetic stimulation on the corticospinal excitability in human brain. Electroencephalogr. Clin. Neurophysiol., 101, 478–482. 12. Amassian, V.E., and Deletis, V. (1999). Relationships between animal and human corticospinal responses. In “Transcranial magnetic stimulation” (W. Paulus, M. Hallett, P.M. Rossini, and J.C. Rothwell, eds.), Electroencephalogr. Clin. Neurophysiol., Suppl. 51, pp. 79–92. Elsevier, New York. 13. Berlin, L., and Amassian, V.E. (1965). Pyramidal tract responses during seizures. Electroencephalogr. Clin. Neurophysiol., 19, 587–597. 14. Amassian, V.E., Quirk, G.J., and Stewart, M. (1990). A comparison of corticospinal activation by magnetic coil and electrical stimulation of monkey motor cortex. Electroencephalogr. Clin. Neurophysiol., 77, 390–401. 15. Amassian, V.E., and Jayachandra, M. (1990). Aspects of conduction failure in traumatized long tract spinal axons of anaesthetized cat. J. Physiol., 426, 103P. 16. Matsuda, H., and Shimazu, A. (1989). Intraoperative spinal cord monitoring using electric responses to stimulation of caudal spinal cord or motor cortex. In “Neuromonitoring in surgery” ( J.E. Desmedt, ed.), pp. 175–190. Elsevier, Amsterdam. 17. Hern, J.E.C., Landgren, S., Phillips, C.G., and Porter, R. (1962). Selective excitation of corticofugal neurones by surface-anodal stimulation of the baboon’s motor cortex. J. Physiol., 161, 73–90. 18. Rosenthal, J., Waller, H.J., and Amassian, V.E. (1967). An analysis of the activation of motor cortical neurons by surface stimulation. J. Neurophysiol., 30, 849–858. 19. Amassian, V.E., Eberle, L.P., Maccabee, P.J., and Cracco, R.Q. (1992). Modeling magnetic coil excitation in human cerebral cortex with a peripheral nerve immersed in a brain-shaped volume conductor: The significance of fiber bending in excitation. Electroencephalogr. Clin. Neurophysiol., 85, 291–301.

22

Vahe E. Amassian

20. Maccabee, P.J., Amassian, V.E., Eberle, L.P., and Cracco, R.Q. (1993). Magnetic coil stimulation of straight and bent amphibian and mammalian peripheral nerve in vitro: Locus of excitation. J. Physiol., 460, 201–219. 21. Rothwell, J.C., Day, B.L., and Amassian, V.E. (1992). Near threshold electrical and magnetic transcranial stimuli activate overlapping sets of cortical neurons in humans. J. Physiol., 452, 109. 22. Amassian, V.E., Cracco, R.Q., and Maccabee, P.J. (1989). Focal stimulation of human cerebral cortex with the magnetic coil: A comparison with electrical stimulation. Electroencephalogr. Clin. Neurophysiol., 74, 401–416. 23. Burke, D., Hicks, R., and Stephen, J. (1992). Anodal and cathodal stimulation of the upper-limb area of the human cortex. Brain, 115, 1497–1508. 24. Boyd, S.G., Rothwell, J.C., Cowan, J.M.A., Webb, P.J., Morley, T., Asselman, P., and Marsden, C.D. (1986). A method of monitoring function in cortical pathways during scoliosis surgery with a note on motor conduction velocities. J. Neurol., Neurosurg. Psychiatry, 49, 251–257. 25. Burke, D., Hicks, R.G., and Stephen, P.H. (1990). Corticospinal volleys evoked by anodal and cathodal stimulation to the human motor cortex. J. Physiol., 425, 283–299. 26. Burke, D., Hicks, R., Gandevia, S.C., Stephen, J., Woodforth, I., and Crawford, M. (1993). Direct comparison of corticospinal volleys in human subjects to transcranial magnetic and electrical stimulation. J. Physiol., 470, 383–393. 27. Werhahn, K.J., Fong, J.K., Meyer, B.U., Priori, A., Rothwell, J.C., Day, B.L., and Thompson, P.D. (1994). The effect of magnetic coil orientation on the latency of surface EMG and single motor unit responses in the first dorsal interosseous muscle. Electroencephalogr. Clin. Neurophysiol., 93, 138–146. 28. Lorente de Nó, R. (1943). Cerebral cortex: Architecture, intracortical connections, motor projections. In “Physiology of the nervous system” ( J.F. Fulton, ed.). Oxford University Press, New York. 29. Amassian, V.E., and Weiner, H. (1966). Monosynaptic and polysynaptic activation of pyramidal tract neurons by thalamic stimulation. In “The thalamus” (D.P. Purpura, and M.D. Yahr, eds.), pp. 255–282. Columbia University Press, New York. 30. Ferbert, A., Priori, A., Rothwell, J.C., Day, B.L., Colebatch, J.G., and Marsden, C.D. (1992). Interhemispheric inhibition of the human motor cortex. J. Physiol., 453, 525–546. 31. Patton, H.D., and Amassian, V.E. (1960). The pyramidal tract: Its excitation and functions. In “Handbook of physiology” (J. Field, ed.), section 1, vol. 2, pp. 837–861. American Physiological Society, Washington, D.C. 32. Amassian, V.E., Cracco, R.Q., Maccabee, P.J., Vergara, M., Hassan, N, Eberle, L., and Rothwell, J.C. (1997). Spatial facilitation of human motor responses by near-threshold magnetic stimulation of parietal and frontal areas. J. Physiol., 504P, 115P. 33. Rothwell, J.C., Thompson, P.D., Day, B.L., Boyd, S., and Marsden, C.D. (1991). Stimulation of the human motor cortex through the scalp. Exp. Physiol., 76, 159–200. 34. Mills, K.R., Boniface, S.J., and Schubert, M. (1992). Magnetic brain stimulation with a double coil: The importance of coil orientation. Electroencephalogr. Clin. Neurophysiol., 85, 17–21. 35. Sakai, K., Ugawa, Y., Terao, Y., Hanajima, R., Furabayashi, T., and Kanazawa, I. (1997). Preferential activation of different I waves by transcranial magnetic stimulation with a figure-ofeight shaped coil. Exp. Brain Res., 113, 24–32. 36. Ziemann, U., Tergau, F., Wassermann, E.M., Wischer, S., Hildebrandt, J., and Paulus, W. (1998). Demonstration of facilitatory I wave interaction in the human motor cortex by paired transcranial magnetic stimulation. J. Physiol., 511, 181–190. 37. Amassian, V.E., Rothwell, J.C., Cracco, R.Q., Maccabee, P.J., Vergara, M., Hassan, N., and Eberle, L. (1998). What is excited by near-threshold twin magnetic stimuli over human cerebral cortex? J. Physiol., 506P, 122P.

Animal and Human Motor System Neurophysiology

23

38. Phillips, C.G. (1987). Epicortical electrical mapping of motor areas in primates: Motor areas of the cerebral cortex. In “Ciba symposium 132” (G. Bock, M. O’Connor, and J. Marsh, eds.), pp. 5–16. Wiley, Chichester. 39. Amassian, V.E., Rothwell, J.C., Ziemann, U., Meyer, B.U., Cracco, R.Q., Trompetto, C., Ashby, P., and Lalli, S.D. (1999). Do human large corticospinal neurons obey the size principle? J. Physiol., 521P, 47P. 40. Kernell, D., and Wu, C.P. (1967). Responses of the pyramidal tract to stimulation of the baboon’s motor cortex. J. Physiol., 191, 653–672. 41. Amassian, V.E., and DeVito, J.L. (1957). La transmission dans le noyau de Burdach (Nucleus cuneatus): Etude analytique par unites isolees d’un relais somatosensoriel primaire. In Colloques Internationaux du Centre National de la Recherche Scientifique 67. “Microphysiologie comparee des elements excitables.” CNRS Paris, pp. 353–393. 42. Rose, J.E., and Mountcastle, V.B. (1954). Activity of single neurons in the tactile thalamic region of the cat in response to a transient peripheral stimulus. Bull. J. Hopkins. Hosp., 94, 238–282. 43. Tokimura, H., Ridding, M.C., Tokimura, Y., Amassian, V.E., and Rothwell, J.C. (1996). Short latency facilitation between pairs of threshold magnetic stimuli applied to human motor cotex. Electroencephalogr. Clin. Neurophysiol., 101, 263–272. 44. Goldman, P.S., and Nauta, W.J.H. (1977). Columnar distribution of cortico-cortical fibers in the frontal association, limbic and motor cortex of the developing rhesus monkey. Br. Res., 122, 393–413. 45. Szentagothai, J. (1978). The neuron network of the cerebral cortex: A functional interpretation. Proc. Roy. Soc. Lond., B201, 219–248. 46. Jones, E.G., Burton, H., and Porter, R. (1975). Commissural and cortico-cortical “columns” in the somatic sensory cortex of primates. Science, 190, 572–574. 47. Chang, H.-T. (1953). Cortical response to activity of callosal neurons. J. Neurophysiol., 16, 117–144. 48. Deletis, V. (1993). Intraoperative monitoring of the functional integrity of the motor pathways. In “Advances in neurology: electrical and magnetic stimulation of the brain” (O. Devinsky, A. Beric, and M. Dogali, eds.), pp. 201–214. Raven Press, New York. 49. Brookhart, J.M. (1952). A study of corticospinal activation of motor neurons. Res. Publ. Ann. Nerv. Ment. Dis., 30, 157–173. 50. Phillips, C.G., and Porter, R. (1964). The pyramidal projection to motoneurones of some muscle groups of the baboon’s forelimbs. In “Progress in brain research” (J.C. Eccles, and J.P. Schade, eds.), vol. 12, pp. 222–245. Elsevier, Amsterdam. 51. Deletis, V., Rodi, Z., and Amassian, V.E. (2001). Neurophysiological mechanisms underlying motor evoked potentials in anesthetized humans: Part 2. Relationship between epidurally and muscle recorded MEPs in man. Clin. Neurophysiol., 112, 445–452. 52. Redman, S., and Walmsely, B. (1983). Amplitude fluctuation in synaptic potentials evoked in cat spinal motoneurons at identified group 1a synapses. J. Physiol., 343, 135–145.

CHAPTER

2

Intraoperative Neurophysiology and Methodologies Used to Monitor the Functional Integrity of the Motor System VEDRAN DELETIS Division of Intraoperative Neurophysiology, Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York

1 Intraoperative Monitoring of the Motor System: A Brief History 1.1 Penfield’s Time 1.2 Spinal Cord to Spinal Cord 1.3 Spinal Cord to Peripheral Nerve (Muscle) 2 New Methodologies 2.1 Single-Pulse Stimulation Technique 2.2 Multipulse Stimulation Technique 3 Methodological Aspects of TES During General Anesthesia 3.1 Electrode Montage Over the Scalp for Eliciting MEPs (for Single- and Multipulse Stimulation Techniques) 4 Recording of MEPs over the Spinal Cord (Epidural and Subdural Space) Using Single-Pulse Stimulation Technique 4.1 D Wave Recording Technique Through an Epidurally or Subdurally Inserted Electrode 4.2 Proper Placement of Epidural Electrodes Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

25

26

Vedran Deletis

4.3 Factors Influencing D and I Wave Recordings 4.4 Neurophysiological Mechanisms Leading to the Desynchronization of the D Wave 5 Recording of MEPs in Limb Muscles Elicited by a Multipulse Stimulating Technique 5.1 Selection of Optimal Muscles in Upper and Lower Extremities for MEP Recordings 5.2 Neurophysiological Mechanisms for Eliciting MEPs using a Multipulse Stimulation Technique 5.3 Surgically Induced Transient Paraplegia 6 Conclusion References

ABSTRACT Beginning with Penfield’s early work and covering the latest developments in the field, this chapter will present a brief history of intraoperative neurophysiology of the motor system, paying special attention to the use of motor evoked potentials (MEPs) during surgeries that place the motor system at risk of injury. The chapter will assess the advantages and disadvantages of traditional techniques previously used to monitor the corticospinal tract (CT), discuss modern methodologies for eliciting and recording MEPs (single and/or multipulse, transcranially applied, electrical stimulation, with recorded activity from either the spinal cord or from the limb muscles), and assess the neurophysiological background for both sets of techniques. Particular interest will be placed on the intraoperative changes of MEPs, their relationship to neurological outcome, and their potential neurophysiological explanations. As an example, the phenomenon of surgically induced transient paraplegia, and the changes in monitoring parameters accompanying it, will be discussed.

1 INTRAOPERATIVE MONITORING OF THE MOTOR SYSTEM: A BRIEF HISTORY

1.1 PENFIELD’S TIME When discussing the use of intraoperative electrical stimulation of the upper motoneurons in humans, it is essential to mention Wilder Penfield (1891– 1976). His publication with Edwin Boldrey in the journal Brain [1] summarized his work on the motor and somatosensory system’s organization of the cerebral cortex in humans, as explored with intraoperative electrical stimulation. Penfield’s systematic exploration of the brain with intraoperative stimulation laid the foundation for the field of intraoperative neurophysiology (ION). After Penfield—except for the work done to intraoperatively localize epileptic foci—almost half a century passed without any significant developments in

Intraoperative Neurophysiology and Methodologies

27

ION exploration of the nervous system. However, a transformation took place during the 1950s and the 1960s when clinical neurophysiology branched into three subfields: electromyography (EMG), electroencephalography (EEG), and evoked potentials. These developments helped to widen the doors of the operating room to the use of these methods intraoperatively. By the late 1970s, somatosensory evoked potentials (SEPs) became routinely used to intraoperatively assess the functional integrity of the somatosensory system in the spinal cord during surgical correction for scoliosis [2]. The same SEPs data were also routinely extrapolated to assess the functional integrity of the upper motor neuron tracts; however, as data mounted, this approach proved unreliable: (a) it provided false results when SEPs were found to be present despite postoperative motor deficits [3, 13] (see Chapter 15, Fig. 15.19, page 386); (b) it provided unreliable (low-quality) or unmonitorable (complete absence) SEPs in patients in whom certain pathologies affected the somatosensory system; and (c) because dorsal myelotomy often destroyed the dorsal column’s integrity in patients undergoing surgery for intramedullary spinal cord tumors, the ability to monitor SEPs was immediately nullified [4]. Because of these difficulties, ION was forced to search for more reliable methods to assess the motor system’s functional integrity. Initial attempts to monitor motor tracts in the spinal cord were made in both Japan and the United States. These attempts focused on two neurophysiological techniques: spinal-cord-tospinal-cord recording, and spinal-cord-to-muscle/peripheral-nerve recording.

1.2 SPINAL CORD TO SPINAL CORD This technique operates with nonselective electrical stimulation of the spinal cord and with nonselective recordings of elicited potentials from the spinal cord. It is used to record signals from the spinal cord regardless of the direction of propagation of the action potentials (either ascending, descending, or ortho/ antidromic). The type of action potential recorded depends on the position of the stimulating and recording electrodes and the direction of the traveling waves through the spinal cord with regards to the natural direction of the conducting pathways [5]. The evoked potentials recorded from the spinal cord using this technique are the electrical sum of activity from multiple pathways. Because of the different conduction properties of the various spinal cord pathways, the recorded potentials can show two distinctive wave morphologies. It has been speculated that one of these waves represents transmission in the dorsal columns (DCs) and the other by the corticospinal tract (CT). Clinical testing on a large number of patients with different and relevant pathologies has not been done to confirm this hypothesis.

28

Vedran Deletis

This method can evaluate the integrity of ascending and descending, and probably propriospinal pathways, within the spinal cord. However, specific information about the DC or CT cannot be obtained with this method. Critical reports [6] could not confirm the value of the spinal cord to spinal cord technique in monitoring motor pathways during surgery for intramedullary spinal cord tumors.

1.3 SPINAL CORD TO PERIPHERAL NERVE (MUSCLE) This technique operates with nonselective stimulation of the spinal cord and selective recordings from the peripheral nerves or muscles. Recordings from the muscle [7, 8] and peripheral nerves [9] presume that after electrical stimulation of the spinal cord, α-motoneurons are activated only by the CT tract. Therefore, compound muscle action potentials (CMAPs) in the limb muscles or electrical activity in the peripheral nerves should be generated by CT stimulation. Unfortunately, α-motoneurons can also be activated by any of the multiple descending tracts within the spinal cord after diffuse electrical stimulation of the spinal cord and/or by antidromically activated dorsal columns and their segmental branches that mediate the H reflex [10]. Electrical activity recorded from mixed peripheral nerves is a combination of α-motoneuron discharges initiated by the CT and other descending tracts. Because the sensory component of mixed peripheral nerves is a physical continuation of the dorsal columns, part of the electrical activity recorded from mixed peripheral nerves after stimulation of the spinal cord arises from the antidromically activated dorsal columns that convey traveling waves to the peripheral nerves [10]. Collision studies have challenged the widely accepted presumption that potentials recorded from peripheral nerves in the lower extremities after stimulation of the spinal cord are generated by the CT [11]. Therefore, there is convincing evidence that selective recording of the electrical activity from peripheral nerves elicited by electrical stimulation of the spinal cord does not arise from the CT [12]. Additional evidence concerning the inaccuracy of monitoring the motor pathways through potentials recorded from peripheral nerves is provided in a recent paper by Minahan et al. This paper describes two patients with postoperative paraplegia in spite of preservation of these potentials [13]. It is fair to say that both of the techniques described can grossly monitor the functional integrity of multiple pathways inside the spinal cord without being specific for any of them. In other words, these methods can indicate that certain lesions to the spinal cord have occurred, but they lack the ability to provide specific information as to which of the spinal cord pathways has been damaged. This methodology may be useful in orthopedic surgical procedures and other surgeries where lesioning of the nervous tissue within the spinal cord is diffuse in nature and where all pathways are usually affected. An exception to

Intraoperative Neurophysiology and Methodologies

29

this phenomenon involves vascular lesions of the spinal cord where selective lesioning of the anterolateral columns can occur. Unfortunately, this nonselective evaluation of multiple pathways is not sufficient during surgery of the spinal cord, during which the DCs can be independently damaged from the anterior and lateral columns [4, 14]. Furthermore, these two techniques (for methodological reasons) cannot evaluate the functional integrity of the CT from the motor cortex to the upper cervical spinal cord. Therefore, supratentorial, brainstem, foramen magnum, and upper cervical spinal cord surgeries cannot be monitored using these techniques. This is also the case in procedures involving the clipping of an intracerebral aneurysm, where the perforating branches for the CT tract in the internal capsula can be selectively damaged while leaving the lemniscal pathways intact. This results in a so-called pure motor hemiplegia (i.e., the patient is postoperatively hemiplegic while the sensory system is intact and SEPs are present) [10, 15, 16]. Since it requires the motor cortex to be surgically exposed, Penfield’s technique may not be used for monitoring motor tracts within the spinal cord.

2 NEW METHODOLOGIES Based on previous work by Hill et al. [17], Merton and Morton [18] discovered that high-voltage current applied over the skull could penetrate to the brain and activate the motor cortex and the CT. Although they produced discomfort, these methods of transcranial electrical stimulation (TES) became an additional tool used to diagnose upper motoneuron lesions in awake patients. On the basis of this work, two methodologies for monitoring the CT intraoperatively were developed, the single-pulse stimulation technique and the multipulse stimulation technique.

2.1 SINGLE-PULSE STIMULATION TECHNIQUE A single-pulse stimulating technique involves a single electrical stimulus applied transcranially or over the exposed motor cortex while the descending volley of the CT is recorded over the spinal cord as a direct wave (D wave).

2.2 MULTIPULSE STIMULATION TECHNIQUE A multipulse stimulating technique involves a short train of five to seven electrical stimuli applied transcranially or over the exposed motor cortex while muscle motor-evoked potentials (MEPs) from limb muscles in the form of

30

Vedran Deletis

FIGURE 2.1 (A) Schematic illustration of electrode positions for transcranial electrical stimulation of the motor cortex according to the International 10–20 EEG system. The site labeled “6 cm” is 6 cm anterior to CZ. (B) Illustration of grid electrode overlying the motor and sensory cortexes. (C) Schematic diagram of the positions of the catheter electrodes (each with three recording cylinders) placed cranial to the tumor (control electrode) and caudal to the tumor to monitor the descending signal after it passes through the site of surgery (left). In the middle are D and I waves recorded rostral and caudal to the tumor site. On the right is depicted the placement of an epidural electrode through a flavectomy/flavotomy when the spinal cord is not exposed. (D) Recording of muscle motor evoked potentials from the thenar and tibialis anterior muscles after being elicited with multipulse stimuli applied either transcranially or over the exposed motor cortex. Modified from [31].

CMAPs are recorded (Fig. 2.1) [28]. (This latter technique differs essentially from the Penfield technique in that it calls for only five to seven stimuli with a stimulating rate of up to 2 Hz. Penfield’s technique calls for continuous stimulation over a period of a few seconds with a frequency of stimulation of 50–60 Hz, and only in the cases when the motor cortex is surgically exposed. Furthermore, at such frequencies and train durations, seizures are easily induced.)

Intraoperative Neurophysiology and Methodologies

31

3 METHODOLOGICAL ASPECTS OF TES DURING GENERAL ANESTHESIA

3.1 ELECTRODE MONTAGE OVER THE SCALP FOR ELICITING MEPS (FOR SINGLE AND MULTIPULSE STIMULATION TECHNIQUES) The electrode placement on the skull is based on the international 10–20 EEG system (Fig. 2.1A). Note that, instead of CZ, the CZ electrode is placed 1 cm behind the typical CZ point. Some laboratories have used 2 cm in front of C3 or C4 (Z. Rodi, personal communication). For transcranial stimulation, cork screw–like electrodes (Corkscrew electrodes, Nicolet, Madison, WI) are preferable because of their secure placement and low impedance (usually 1 KΩ). Alternatively, an EEG needle electrode may be used. We do not recommend the use of EEG cup electrodes fixed with collodium since they are impractical and their placement is time-consuming. The only exception is for young children in whom the fontanel still exists. Since the CS electrodes could penetrate the fontanel during placement, the use of EEG cup electrodes is suggested. The skull presents a barrier of high impedance to the electrode current applied transcranially; therefore, we cannot completely control the spread of electrical current when it is applied. For this reason, various combinations of electrode montages may need to be explored to obtain an optimal response. The standard montage is C3/C4 for eliciting MEPs in the upper extremities and C1/C2 for eliciting MEPs in the lower extremities. With sufficient intensity of stimulation at C1/C2, MEPs are preferentially elicited in the right limb muscles while stimulation at C2/C1 elicits MEPs in the left limb muscles. With stronger electrical stimulation, the current will penetrate the brain more deeply, stimulating the CT at a different depth from the motor cortex (Fig. 2.2). On the basis of measurements of the D wave latency, it has been postulated that there are three favorable points that are susceptible to depolarization of the CT: cortex/subcortex (weak electrical stimulation), internal capsula (moderate electrical stimulation), and brainstem/foramen magnum (strong electrical stimulation). Selectivity of stimulation is possible at the level of the cortex (subcortex). Therefore, only the application of relatively weak electrical stimuli to the cortex is selective, and it activates only a small portion of the CT fibers (e.g., activating only one extremity) or only one CT. It is important to remember that during electrical stimulation of the motor cortex, the anode is preferentially the stimulating electrode. With increasing intensity of the current, the cathode becomes the stimulating electrode as well.

32

Vedran Deletis

FIGURE 2.2 D and I waves recorded after a single electrical stimulus delivered transcranially (CZ anode/6 cm anterior cathode) in 14-year-old patient with idiopathic scoliosis. When the intensity of the stimulus is increased, electrical current activates the CT deeper within the brain and the latency of the D wave becomes shorter. As current becomes stronger, more I waves are induced (100% corresponds to 750 volts of stimulator output). Modified from [10].

As an example, stimulation with the C3+/C4− will selectively activate muscles of the right arm. When stimulation intensity is increased, the cathode (C4−) becomes the stimulating electrode as well, resulting in the stimulation of the left arm. Finally, when current intensity becomes strong enough to penetrate to the internal capsule more caudally, all four extremity muscles can be activated. For anatomical reasons (deep position of the leg motor area in the interhemispheric fissure), more intense current is usually needed to obtain MEPs in the lower extremities. It is especially difficult to obtain them separately without also activating the upper extremities. Our observation has been that it can be done in certain patients, especially when using the CZ/6 cm in front montage (see Fig. 2.1). By their anatomical location, recording electrodes in the limb muscles can indicate which fibers of the CT are activated predominantly (left or right, fibers for upper or lower extremities). If one would like to activate left and right CT simultaneously to obtain D wave recordings, weak electrical stimulation should be avoided and a moderate intensity should be used. In Fig. 2.3, it is obvious that weak electrical stimulation activates fibers of the CT for the left upper extremities only. This can result in activation of only one CT while not affecting the other CT. Therefore, the intensity of electrical stimulation for eliciting a D wave should be determined by simultaneous recordings of MEPs from limb muscles (indicating which fibers of the CT have been predominantly activated), or only moderate intensities of electrical current for eliciting D waves should be used. The moderate intensity of electrical current

Intraoperative Neurophysiology and Methodologies

33

FIGURE 2.3 Transcranial electrical stimulation over the C4 anode/C3 cathode with recordings of the D wave over the C6–C7 segment (above) and the T7–T8 segment of the spinal cord (below). Stimulus intensity was 35 and 40 mA, respectively. Stronger stimuli elicit the D wave over the thoracic spinal cord, while a weaker stimulus (35 mA) elicits the D wave only over the cervical spinal cord.

will activate both CTs at the level of the internal capsule. If MEP waves have not been simultaneously recorded with D waves, the following guidelines should be followed: increase the intensity of the stimulation until D waves do not increase in amplitude (Fig. 2.2, the third trace from the top). This is a sign that most of the fast conducting neurons of CT from the left and right CT have been activated. The neurophysiological mechanism for eliciting MEPs by stimulating the motor cortex in patients under the influence of anesthetics is different from the mechanism in the awake subject. In the latter, electrical current stimulates the body of the motor neuron transynaptically over the chain of vertically oriented excitatory neurons, resulting in I waves (indirect activation of the motoneurons). At the same time, electrical current activates axons of the cortical motoneurons, directly generating D waves [19]. In anesthetized patients, anesthetics block the synapses of the vertically oriented excitatory chains of neurons terminating on the cortical motoneuron’s body. Therefore, only the D wave is generated after electrical stimulation of the motor cortex [19, 20]. Patients with idiopathic scoliosis are an exception. In this group, abundant I waves can be recorded (Fig. 2.2). We believe that this is one of the neurogenic markers of the disease present in these patients [21]. Furthermore, it has been shown that a frontally oriented cathode preferentially generates I waves because at this stimulating setting corticocortical projections of vertically oriented interneurons are optimally activated. With the cathode in the lateral position, this is not the case [22, 23] (Fig. 2.4).

34

Vedran Deletis

FIGURE 2.4 Upper thoracic epidural recordings of D and I waves in a 14-year-old female during surgery for a low cervical intramedullary tumor. The upper trace was obtained after transcranial electrical stimulation over C1 (anode) and C2 (cathode) using 140 mA stimulus intensity and a stimulus duration of 500 µs. The lower trace was obtained after anodic stimulation at CZ and cathodal stimulation at 6 cm anterior to CZ, using the same stimulus duration but at 200 mA. Note the appearance of the D and I waves with this electrode arrangement. (An upward deflection is negative.) Reprinted from [23].

4 RECORDING OF MEPs OVER THE SPINAL CORD (EPIDURAL AND SUBDURAL SPACES) USING SINGLE-PULSE STIMULATION TECHNIQUE

4.1 D WAVE RECORDING TECHNIQUE THROUGH AN EPIDURALLY OR SUBDURALLY INSERTED ELECTRODE This method is a direct clinical application of Patton and Amassian’s [19] discovery in the 1950s that electrically stimulated motor cortex in monkeys generates a series of well-synchronized descending volleys in the pyramidal tract. This knowledge of CT neurophysiology, which was collected in primates, can be applied to humans in most cases. We have to be aware that even small methodological aspects of recording D waves are of the utmost importance and should be followed in order to achieve reliable results. 4.1.1 Choice of Electrode Practically any type of catheter-type electrode designed for electrical stimulation of the spinal cord epidurally can be used for recording D and I waves. We prefer to use the JX-300 (Arrow International, Reading, PA) because of its optimal recording properties and intercontact recording electrode distance (Fig. 2.5). This electrode has three platinum-iridium recording cylinders 3 mm in length, 1.3 mm in diameter, and 18 mm apart, with recording surfaces of approximately 12.3 mm2.

Intraoperative Neurophysiology and Methodologies

35

FIGURE 2.5 Semi-rigid catheter electrode for recording MEPs (D wave) from the spinal cord, epior subdurally. The electrode has passed through a 14-gauge Touhy needle for percutaneous placement epidurally. To the left (enlarged) are two openings marked with asterisks for flushing the three cylindrical recording contacts (1, 2, 3) through the injection site (top, right).

This electrode is semi-rigid, a property that facilitates its placement either percutaneously or through flavotomy. Furthermore, it consists of a double lumen with two openings at the tip of the electrode. This allows for the injection of saline to flush the recording contact surfaces and reduce impedance. This is an important methodological detail in the case of bad electrode contact if the electrode is placed percutaneously in the epidural space (where it can face a high impedance). Once the electrode is in place, it is very difficult to reposition it. Thus an injection of saline through the outer lumen is a method of rectifying the high-impedance problem (Fig. 2.6). When the electrode is placed after laminectomy, problems with impedance and positioning of the electrode are easier to solve because the surgeons are able to reposition the lead. Most epidural electrodes are disposable. If one uses a nondisposable type, extreme care should be taken to ensure that the electrode is clean before sterilization and thus has improved electrical properties. To clean the electrode, we recommend one of the following procedures. You can immerse the electrode tip in saline and pass a 9 V DC current (regardless of polarity) through it until a bubble of gas cleans the contact surface for a period of a few minutes, or you can use an ultrasound cleaner (Branson 1210, Branson Ultrasonics Corporation, Danbury, CT) by submersing the electrode in the cleaner for 5 minutes.

36

Vedran Deletis

FIGURE 2.6 Two traces with a D wave recorded epidurally at the lower cervical spinal cord after percutaneous placement of the epidural electrode in a patient with a brain tumor. High impedance results in a large artifact (lower trace) which has been reduced (upper trace) after injection of saline into the epidural space (see Fig. 2.5).

Both techniques will remove any film or biological material remaining on the electrode from the contact surfaces and will decrease their impedance. This maneuver will diminish the stimulus artifact, which usually appears when contact surfaces have high impedance. Because of the short latency of the D wave, a large stimulus artifact in an uncleaned electrode can pose an insurmountable obstacle for D wave recording.

4.2 PROPER PLACEMENT OF EPIDURAL ELECTRODES Depending on the surgical procedure, there are two methods of electrode placement: percutaneously, or after laminectomy/laminotomy or flavotomy/flavectomy.

4.2.1 Percutaneous Placement of Catheter Electrode This technique is rather popular in Japan [5]. As used there, it is slightly different from the one we employ, since the neurosurgeon may ask for a subdural placement of the catheter electrode. We have performed this procedure to monitor the CT during brainstem and supratentorial surgeries where there is high risk of potential damage. Today, because of the increasing popularity of MEPs monitoring during procedures involving the spinal cord and brainstem, the demand (indications) for percutaneous placement of this type of electrode has diminished. When we do use percutaneous placement, a 14-gauge, thin-wall Touhy needle (T466LNRH, Becton Dickinson and Comp, Franklin Lakes, NJ; Fig. 2.5), is used for introducing the electrode into the epidural space percutaneously. Following percutaneous electrode placement, care must be taken not to withdraw the electrode while the Touhy needle is in place. Otherwise, the sharp edge of the

Intraoperative Neurophysiology and Methodologies

37

needle could shred the wall of the electrode. The optimal position for penetrating the epidural space with the Touhy needle is the upper thoracic (T1–T2) epidural space. With the needle in this region, the catheter electrode can be gently pushed up to the level of the lower cervical spinal cord. With this electrode placement we can monitor the CT for both the upper and lower extremities by recording D waves after selective stimulation of the motor cortex. Appropriate electrode placement can be confirmed either by x-ray or by recording epidural SEPs from the same electrode after stimulation of the median or ulnar nerves. In two series consisting of 57 patients [24] and 16 patients [25], no complications from the placement of the electrode occurred (e.g., bleeding, infection, or puncture of the spinal cord). This method requires skills that the anesthesiologist practiced in the epidural injection of anesthetics would typically have. 4.2.2 Placement of Electrode after Laminectomy/Laminotomy or Flavectomy/Flavotomy Our center uses this technique regularly for all procedures that require CT monitoring when a laminectomy is performed. These procedures include surgery for the removal of spinal cord tumors and different surgical interventions on the spinal cord. The surgeon places two catheter electrodes in the epi- or subdural space at the rostral and caudal edge of the laminectomy. The rostral electrode is the control electrode for nonsurgically induced changes in the D wave, while the caudal one monitors the surgically induced changes to the CT (see Fig. 2.1). Massive dural adhesions, usually from previous surgery or after spinal cord radiation, can prevent the placement of the catheter electrode. Also, placement below the T10 bony level cannot record a D wave of sufficient amplitude because of lack of sufficient CT fibers. The control (rostral) electrode cannot be placed in cases of high cervical spinal cord pathology because of the lack of space. The amplitude of the D wave recorded over the cervical spinal cord could be 60 µV or more, while over thoracic segments it may be only 10 µV. With a stimulating rate of 2 Hz, it takes two to four averaged responses to get a reliable D wave. This results in an update every second. Unfortunately, the maximal stimulating rate from commercially available TES stimulators is 1 stimulus per second. In surgical procedures in which the spine is exposed but a laminectomy is not performed (e.g., surgical corrections of scoliosis or dorsal approach to spine stabilization), the catheter electrode may be inserted through a flavotomy/flavectomy.

4.3 FACTORS INFLUENCING D AND I WAVE RECORDINGS D waves represent a neurogram of the CT which is not significantly influenced by nonsurgically induced factors. Stimulation of the CT takes place intracranially distal to the cortical motoneuron body, while recording is done caudal to the

38

Vedran Deletis

FIGURE 2.7 D waves recorded over the lower cervical spinal cord in a patient with an upper cervical intramedullary spinal cord tumor, after stimulation with CZ anode/6 cm anterior cathode. Temporary cooling of the exposed spinal cord results in delayed latency of the D and I waves. After warming of the spinal cord, the latency of the D and I waves returned to the previous values. Reprinted from [10].

surgical site but above the synapses of the CT at the α-motoneuron. Since no synapses are involved between the stimulating site and the recording site, the D wave is very stable and reliable. Therefore, we consider D wave recordings to be the “gold standard” for measuring the functional integrity of the CT. Still, there exists a few nonsurgically induced changes that will affect the D wave. Being able to correctly recognize them is essential to giving the surgeon appropriate information. If the exposed spinal cord is cooled, either by cold irrigation with saline or low operating room temperature, the latency of the D wave will be temporarily prolonged (Fig. 2.7). Sometimes during stimulation, even with a single stimulus, the epidural electrode can pick up the paraspinal muscle artifact. This would affect the I wave, but not the D wave, parameters (see Fig. 2.8). If this phenomenon occurs, it is more frequent during cervical than thoracolumbar catheter placement. In contrast to those of others [26], our data demonstrate that volatile anesthetics do not change the parameters of the D wave by influence on the membrane properties of the CT. To demonstrate this, we see that as isoflurane concentration increases (e.g., >2%), the latency of the D wave gets prolonged while the amplitude diminishes (see Fig. 2.9). However, this can be easily corrected by increasing the intensity of the current. Therefore, we believe that the mechanism by which isoflurane influences the parameters of the D wave is vasodilatation of the cortical blood vessels. Because of the vasodilatation, current between the stimulating electrodes shunts and activates the CT more superficially, resulting in longer latencies of the D wave. The smaller amplitude of the D wave results from fewer fibers of the CT being activated if current flows superficially (Fig. 2.10). A prolongation of the latency and a diminished amplitude of the D wave occur only if the CT is activated transcranially. In contrast, this phenomenon is not present when the motor cortex is stimulated directly

Intraoperative Neurophysiology and Methodologies

39

FIGURE 2.8 Epidurally recorded D and I waves over the cervical spinal cord showing a muscle artifact. After administration of the muscle relaxant, the muscle artifact disappears. The muscle artifact affects the I wave, but not the D wave, recordings. S = beginning of transcranially applied stimulus. Modified from [10].

FIGURE 2.9 Transcranial electrical stimulation (CZ anode/6 cm anterior cathode) and direct electrical stimulation of the exposed motor leg area with recording of the D wave over the lower thoracic spinal cord in two different patients. Identical concentrations of isoflurane showed a prominent effect on the amplitude and latency of the D wave (50% decrement of amplitude and 0.5 ms prolonged latency after end tidal concentration of 2% isoflurane). This effect is only evident when transcranial electrical stimulation is used. A minimal effect of isoflurane on D wave parameters was observed when electrical stimulation was applied to the exposed cortex. Reprinted from [10] and [34].

40

Vedran Deletis

FIGURE 2.10 To the left, current flow is represented schematically before (white line) and after (grey line) administration of isoflurane. Because of the vasodilatatory effects of isoflurane on the cortical blood vessels, the current between the two stimulating electrodes is shunted, flowing through the brain more superficially. This results in a prolonged latency and smaller amplitude of the D wave when compared to a D wave elicited with the same intensity of current without isoflurane (6.0 ms vs. 6.3 ms, respectively; to the right). At the same time, the disappearance of the I wave can be observed under the influence of isoflurane.

through a grid electrode with a short distance between the electrodes. All of the above observations provide evidence that changes in the D wave are due to mechanisms other than influence on the CT axon membranes.

4.4 NEUROPHYSIOLOGICAL MECHANISMS LEADING TO THE DESYNCHRONIZATION OF THE D WAVE In certain patients with spinal cord tumors (usually involving a few segments) the D wave is not recordable at the beginning of surgery [27]. At the same time, muscle MEPs are recordable, even in patients that may not necessarily have a major motor deficit (Fig. 2.11). The temporal summation of the desynchronized D waves occurs at the segmental level. The same phenomenon is present in patients who undergo radiation of the spinal cord. We believe this is a result of a desynchronization in conduction of the CT axon. In other words, fast fibers of the CT conduct D waves with different speeds over the site of the lesion or irradiation. Therefore, desynchronized D waves cannot be easily demonstrated caudal to the lesion site with the present methodology. There are different grades of desynchronization, which will be seen as low-amplitude and widebase D waves (Fig. 2.11A). A higher degree of desynchronization is represented by a nonrecordable D wave (Fig. 2.11B). Patients who do not have a recordable D wave at the beginning of surgery are challenging for the monitoring team because they represent a high-risk group of patients for injury to the CT. With the present methodology, we can only monitor them by recording MEPs from limb muscles. Because of the

Intraoperative Neurophysiology and Methodologies

41

FIGURE 2.11 (A) Recording of a D wave cranially (upper trace) and caudally (lower trace) to the intramedullary spinal cord tumor. Note the well-synchronized D wave cranially, in contrast to the desynchronized D wave caudal to the tumor. (B) Very small epidurally recorded MEPs caudal to a high cervical intramedullary tumor (due to extreme desynchronization), despite large muscle MEPs recorded from a small hand muscle elicited after a short train of six stimuli were present (to the right). Modified from [33].

possibility that transient paraplegia may occur, this is not an ideal monitoring tool. When muscle MEPs disappear during surgery in the patients who do not have a recordable D wave at baseline, it is not possible to distinguish transient from permanent motor deficit intraoperatively (see Section 5.3).

5 RECORDING OF MEPs IN LIMB MUSCLES ELICITED BY A MULTIPULSE STIMULATING TECHNIQUE

5.1 SELECTION OF OPTIMAL MUSCLES IN UPPER AND LOWER EXTREMITIES FOR MEP RECORDINGS The selection of appropriate muscles to record from is an important issue in the monitoring of MEPs. In certain patients who have deep paresis, not choosing the optimal muscles can result in “nonmonitorable” patients. The small hand muscle (e.g., abductor pollicis brevis, or APB) is one of the optimal muscles to

42

Vedran Deletis

monitor the CT for the upper extremities. It has been shown that a good alternative is the long forearm flexors [28], or even the forearm extensors. The spinal motoneurons for these muscle groups have rich CT innervation and are therefore suitable for monitoring the functional integrity of the CT. This is not the case with the proximal muscle of the arm or of the shoulder (biceps, triceps, or deltoid muscles). For the lower extremities, abductor hallucis brevis (AHB) is the optimal muscle because of its dominant CT innervation. In animal experiments, it has been shown that after CT stimulation the highest amplitude of the excitatory postsynaptic potential (EPSP) has been found in the α-motoneuron pools for the lower extremities in the small and long flexors of the foot [29]. An alternative to this muscle is the tibialis anterior muscle (TA). Our standard electrode montage for recording MEPs in the upper and lower extremities are the AHB and TA for the lower extremities and the ABP for the upper extremities.

5.2 NEUROPHYSIOLOGICAL MECHANISMS FOR ELICITING MEPS USING A MULTIPULSE STIMULATION TECHNIQUE Understanding the mechanism involved in the generation of MEPs is essential for describing their appropriate use, explaining their behavior, understanding their value, and knowing their limits during the monitoring of the CT. Generation of MEPs is more complex in nature than the generation of the D and I waves. Therefore, their interpretation, especially during anesthesia, is rather complex. Generation of MEPs and their propagation to the end organ (muscle) depends on (a) the excitability of the motor cortex and the CT tract, (b) the conductivity of CT axons, (c) the excitability level of α-motoneuron pools, (d) the role played by the supportive system of the spinal cord (helping to increase the excitability of α-motoneurons), and (e) the integrity of motor nerves, the motor endplates and muscles.

5.2.1 Recovery of Amplitude and Latency of the D Wave There is a frequency limit for the transmission of descending volleys through the CT axons to the α-motoneurons. This limit can be easily tested by applying two identical electrical stimuli transcranially with different interstimulus intervals (ISIs). This test can show the recovery time of the second D wave response.

Intraoperative Neurophysiology and Methodologies

43

FIGURE 2.12 Two diagrams showing the relationship between interstimulus interval (ISI), duration of stimuli, and recovery of the amplitude and latency of the conditioning D wave. Two identical stimuli have been applied transcranially with different ISIs. Amplitude and latency of the second D wave (D2) were compared to those of the first one (D1). Note that earlier and complete recovery of the amplitude and latency of the second D wave occurs with a stimulus duration of 500 µs and an ISI of around 4 ms. Reprinted from [30].

Using this paradigm (conditioning and test stimuli), a D wave recovery curve can be plotted relative to the amplitude and latency of the second D wave (Fig. 2.12). In a paper recently published [30], we show that the optimal ISI for complete recovery of the second D wave amplitude and latency is around 4 ms, using a moderate stimulus intensity with a duration of 500 µs. Because the αmotoneuron is optimally bombarded when the train of equal stimuli elicits D waves of equal amplitudes, the optimal ISI for muscle activation is expected to be 4 ms. Fig. 2.13 indicates that with an ISI of 4 ms, three stimuli are sufficient to elicit MEPs because of the complete recovery of each consecutive D wave (Fig. 2.13B3). Comparatively, using the identical stimulus intensity but decreasing the ISI to 2 ms, five stimuli are needed to elicit MEPs, which are of even smaller amplitude, because of incomplete recovery of the amplitude of each consecutive D wave (Fig. 2.13A5). This rule applies only if a single stimulus elicits a single D wave (see Section 5.2.3).

44

Vedran Deletis

FIGURE 2.13 Relationship between MEPs recorded epidurally and from muscle. (A) Train of five stimuli are needed with an ISI of 2 ms in order to elicit muscle MEPs in tibialis anterior muscle (A5). (B) With an ISI of 4 ms, only three stimuli are needed to elicit muscle MEPs in the tibialis anterior muscle (B3). D = D wave; I = I wave; PM = paraspinal muscle artifact. Reprinted from [31].

5.2.2 Facilitation of I Wave We have been shown that three stimuli applied transcranially over the motor cortex can elicit more than three descending volleys in lightly anesthetized patients [31]. In Fig. 2.13A3, it is clearly visible that three stimuli generate four descending volleys (D1, D2, D3, and an additional I wave). Facilitation of previously nonexisting I waves (after a single stimulus, Fig. 2.13A1) is one of the important factors underlying the potency of the multipulse stimulating technique for eliciting MEPs in lightly anesthetized patients. Furthermore, it has been shown that because of the lack of synchronicity of I waves, their recorded amplitude is only one third of their actual amplitude [32]. Certainly, if the patient is deeply anesthetized, the cortical synapses where the I wave was facilitated are completely blocked, so this phenomenon does not occur. 5.2.3 Total Number of D and I Waves As stated previously, to allow for the complete recovery of the D wave, the ISI in the multipulse train should be 4 ms. In situations where a single stimulus generates more than a single D wave, the optimal ISI should be set long enough to allow the entire set of D and I waves to recover, and in turn, to allow the next

Intraoperative Neurophysiology and Methodologies

45

FIGURE 2.14 (A) In this patient, a single stimulus delivered over the exposed motor hand area elicits a single D wave and multiple I waves. The ISI should be long enough to prevent the second set of D and I waves, elicited by a second stimulus, from falling into the CT axon refractory period resulting from the previous waves (as is the case in trace B). When the ISI is 5.9 ms (C) and 8.0 ms (D), this will not occur, resulting in a sufficient numbers of D and I waves to elicit MEPs (trace D). The stimulus is marked by an arrow and the D wave by an asterisk. Reprinted from [31].

set of D and I waves to fully develop. Therefore, the second stimulus can generate the same pattern of D and I waves (Fig. 2.14). Otherwise, the second set of D and I waves could fall into the CT axon refractory period resulting from the first set. This is the case in Fig. 2.14, where a single stimulus generates a single D wave and multiple I waves (A). In this case only two stimuli, 8 ms apart, were necessary to generate the maximum amplitude of muscle MEPs (Fig. 2.14D). If the ISI is shorter (e.g., 4.1 ms in Fig. 2.14B), partial cancellation of the D and I waves elicited by a second stimulus will occur. Consequently, the total number of D and I waves will be insufficient to bring an α-motoneuron to the firing level and MEPs will not be generated. This mechanism could be important in the lightly anesthetized patient as well as in patients with idiopathic scoliosis where a single stimulus generates multiple I waves (see Fig. 2.2).

5.2.4 Generation of Muscle MEPs Depends on Two Systems: The CT and the Supportive System of the Spinal Cord To reiterate, descending activity from the CT axons alone is not sufficient to generate muscle MEPs in anesthetized patients. The other system(s) should be activated as well. Three examples support this statement:

46

Vedran Deletis

A. If the multipulse technique (in a non-deeply anesthetized patient) with a repetition rate of 1 or 2 trains per second is performed, each consecutive response recorded from muscle will have an increasing amplitude. In cases where the intensity of stimuli is just slightly above the threshold, the first few trains will not generate muscle MEPs at all. At the same time, the D wave amplitudes remain the same (Fig. 2.15). B. In the patients with intramedullary spinal cord tumors presented in Fig. 2.16, recording of the D waves from the left and right CT generates symmetrical D waves cranially and caudally to the tumor site. Yet muscle MEPs are significantly smaller over the right TA muscle where the patient has clinical weakness. The presumption is that the current required to elicit MEPs from muscles on one side of the body is activating only one CT. Therefore, the D wave, recorded from the spinal cord using this same intensity, must predominantly belong to one CT. C. During surgery for intramedullary spinal cord tumors, muscle MEPs can completely disappear with no significant changes in the amplitude of the D wave (see further transient paraplegia, Fig. 2.17). These three examples provide convincing evidence that the generation of MEPs involves more than just the CT system (see Section 5.3.1).

FIGURE 2.15 Recordings of 10 consecutive muscle MEPs from the right abductor hallucis brevis muscle (after delivering 10 trains consisting of five stimuli, pulse width of 100 µs, intensity of 288 mA, stimulus rate of 1 Hz) over C3 anode/C4 cathode in a 60-year-old patient undergoing anterior cervical spine decompression and stabilization. Note that after the fifth train the amplitude of the muscle MEPs increases 10-fold, showing a tendency to further increase its amplitude.

Intraoperative Neurophysiology and Methodologies

47

FIGURE 2.16 Simultaneous recording of the D wave from the right and left CT, cranial and caudal to a midthoracic intramedullary spinal cord tumor (upper), showing a symmetrical amplitude of the D wave. At the same time, muscle MEPs showed significantly smaller amplitude over the right TA muscle when compared to the left, correlating with the patient’s weakness in the right leg. This recording indicates involvement of pathways other than the CT in the generation of the MEPs.

FIGURE 2.17 Muscle MEPs recorded from right and left TA muscle (left) and D wave recorded epidurally over the lower cervical spinal cord (right). During surgery, muscle MEPs completely disappeared while the D wave decreased in amplitude (less than 50%), resulting in transient paraplegia for this patient during surgery for an intramedullary spinal cord tumor. The patient recovered completely within a week. Reprinted from [33].

48

Vedran Deletis

5.3 SURGICALLY INDUCED TRANSIENT PARAPLEGIA During surgery for intramedullary spinal cord tumors in the thoracic region, MEPs in the TA muscles will frequently disappear while the D wave remains unaffected. All patients demonstrating this finding during surgery wake up paraplegic (or monoplegic if the TA MEPs disappear in one leg). In patients in whom we have observed this phenomenon, motor strength is typically recovered in a few hours to a few days following surgery. No permanent motor deficits have been observed [14, 33] (Fig. 2.17). With almost all cases of transient paraplegia, the first changes are seen in the MEPs and not in the parameters of the D wave. This gives the surgeon a warning sign and a window of time to plan to end the tumor removal. This is a critical point for intraoperative planning of the extent of tumor removal. If changes in the MEPs do not appear, tumor removal can proceed until a gross total resection is accomplished without the patients having permanent motor deficits postoperatively.

5.3.1 Neurophysiological Basis for Surgically Induced Transient Paraplegia Taking into account the previous evidence that the generation of muscle MEPs involves more than just the CT, activation of the CT and other descending systems within the spinal cord is necessary. We speculate that the propriospinal (diffuse) system of the spinal cord is activated by CT axons that are linked via synaptic connections to the propriospinal system within the spinal cord. In the case of surgically induced transient paraplegia, this system is temporarily compromised by selective surgery while the CT is left intact. After the patient wakes up, other descending systems compensate for the lack of propriospinal tonic influence on α-motoneurons. This results in the fast recovery of these patients. This suggested mechanism is speculative but from a prognostic and pragmatic point of view is critical because it correlates extremely well with clinical outcome. Comparatively, if the CT tract is damaged during surgery (complete loss of D wave or decrement of the amplitude compared with the baseline of more than 50%), a permanent motor deficit is expected [27]. Combining the information about the D wave and about the muscle MEPs during surgery for intramedullary spinal cord tumors makes this surgery safer, changes the intraoperative strategy, and significantly diminishes the occurrence of postoperative deficits (see Chapter 4).

Intraoperative Neurophysiology and Methodologies

49

6 CONCLUSION Historically, intraoperative neurophysiology has progressed by means of trial and error. Unfortunately, this has resulted in a number of different opinions as to its utility in documenting and preventing surgically induced neurological injury. In spite of this, the methodology for monitoring the functional integrity of the CT has progressed over the last 10 years into a reliable, fast, and relatively simple tool that is easily utilized intraoperatively. The development of such a solid methodology has given us reliable and specific data that highly correlate with neurological outcome postoperatively. This correlation and the published surgical outcome data demonstrate the merits of these techniques. Further developments in intraoperative neurophysiology should be directed toward developing a methodology for the functional mapping of the nervous tissue in the exposed brain, brainstem, and spinal cord during surgery. The first steps in this direction have given promising results (see mapping of the dorsal columns, Chapter 7), brainstem cranial motor nuclei (see Chapter 14), and mapping of pudendal afferents (see Chapter 9). We have recently reported the first trials using a technique for mapping the CT intraoperatively [35], and we hope that this technique will evolve into a method for identifying the CT during exposed brain and spinal cord surgery. Included with the accompanying CD is a video showing epidurally recorded D waves and muscle MEPs recorded from limb muscles during surgery for the removal of an intramedullary tumor. A CD-ROM video presentation will depict actual operating room implementations of these methods (choose Chapter 2 from the accompanying CD’s main menu).

REFERENCES 1. Penfield, W., and Boldrey, E. (1937). Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation. Brain, 60, 339–443. 2. Engler, G.L., Spielholz, N.I., Bernhard, W.N., Danziger, F., Merkin, H., and Wolff, T. (1978). Somatosensory evoked potentials during Harrington instrumentation for scoliosis. J. Bone Joint Surg., 60, 528–532. 3. Lesser, R.P., Raudzens, P., Luders, H., Nuwer, M.R., Goldie, W.D., Morris, H.H., Dinner, D.S., Klem, G., Hahn, J.F., Shetter, A.G., Ginsburg, H.H., and Gurd, A.R. (1986). Postoperative neurological deficits may occur despite unchanged intraoperative somatosensory evoked potentials. Ann. Neurol., 19, 22–25. 4. Deletis, V. (1999). Intraoperative neurophysiological monitoring. In “Pediatric neurosurgery: Surgery of the developing nervous system” (D. McLone, ed.), 4th ed., pp. 1204–1213. W.B. Saunders, Philadelphia.

50

Vedran Deletis

5. Tamaki, T., Takano, H., and Takakuwa, K. (1985). Spinal cord monitoring: Basic principles and experimental aspects. Cent. Nerv. Syst. Trauma, 2, 137–149. 6. Koyanagi, I., Iwasaki, Y., Isy, T., Abe, H., Akino, M., and Kuroda, S. (1993). Spinal cord evoked potential monitoring after spinal cord stimulation during surgery of spinal cord tumors. Neurosurgery, 33(3), 451–460. 7. Machida, M., Weinstein, S.L., Yamada, T., and Kimura, J. (1985). Spinal cord monitoring: Electrophysiological measures of sensory and motor function during spinal surgery. Spine, 10, 407–413. 8. Taylor, B.A., Fennelly, M.E., Taylor, A., and Farrell, J. (1993). Temporal summation: The key to motor evoked potential spinal cord monitoring in humans. J. Neurol. Neurosurg. Psychiatry, 56, 104–106. 9. Owen, J.H., Bridwell, K.H., Grubb, R., Jenny, A., Allen, B., Padberg, A.M., and Shimon, S.M. (1991). The clinical application of neurogenic motor evoked potentials to monitor spinal cord function during surgery. Spine, 16(8), S385–S390. 10. Deletis, V. (1993). Intraoperative monitoring of the functional integrity of the motor pathways. In “Advances in neurology: Electrical and magnetic stimulation of the brain” (O. Devinsky, A. Beric, and M. Dogali, eds.), pp. 201–214. Raven Press, New York. 11. Toleikis, J.R., Skelly, J.P., Carlvin, A.O., and Burkus, J.K. (2000). Spinally elicited peripheral nerve responses are sensory rather than motor. Clin. Neurophysiol., 111, 736–742. 12. Deletis, V. (2001). The “motor” inaccuracy in neurogenic motor evoked potentials (Editorial). Clin. Neurophys., 112, 1365–1366. 13. Minahan, R.E., Sepkuty, J.P., Lesser, R.P., Sponseller, P.D., and Kostuik, J.P. (2001). Anterior spinal cord injury with preserved neurogenic “motor” evoked potentials. Clin. Neurophysiol., 112, 1442–1450. 14. Kothbauer, K., Deletis, V., and Epstein, F. (1998). Motor evoked potential monitoring for intramedullary spinal cord tumor surgery: Correlation of clinical and neurophysiological data in a series of 100 consecutive procedures. Neurosurg. Focus (electronic journal), (4), Article 1 (http://www.aans.org/journals/online_j/may98/4-5-1). 15. Fisher, C.M., and Curry, H.B. (1965). Pure motor hemiplegia of vascular origin. Arch. Neurol., 13, 30–44. 16. Schramm, J., Koht, A., Schmidt, G., Pechstein, U., Taniguchi, M., and Fahlbusch, R. (1990). Surgical electrophysiological observations during clipping of 134 aneurysms with evoked potential monitoring. Neurosurgery, 26, 61–70. 17. Hill, D.K., McDonnell, M.J., and Merton, P.A. (1980). Direct stimulation of the abductor pollicis in man. J. Physiol., 300, 2P. 18. Merton, P.A., and Morton, H.B. (1980). Electrical stimulation of human motor and visual cortex through the scalp. J. Physiol., 305, 9–10P. 19. Patton, H.D., and Amassian, V.E. (1954). Single and multiple unit analysis of cortical state of pyramidal tract activation. J. Neurophysiol., 17, 345–363. 20. Hicks, R., Burke, D., Stephen, J., Woodforth, I., and Crawford, M. (1992). Corticospinal volleys evoked by electrical stimulation of human motor cortex after withdrawal of volatile anesthetics. J. Physiol. (Lond.), 456, 393–404. 21. Deletis, V., Kiprovski, K., Neuwirth, M., and Engler, G. (1994). Do neurogenic lesions of the spinal cord generate distinctive features of the epidurally recorded motor evoked potentials? In “Handbook of spinal cord monitoring” (S.J. Jones, S.M. Boyd, and N.J. Smith, eds.), pp. 266–271. Kluwer Academic Publishers, Dordrecht. 22. Kaneko, K., Kawai, S., Fuchigami, Y., Morieta, H., and Ofuji, A. (1966). The effect of current direction induced by transcranial magnetic stimulation on the corticospinal excitability in human brain. EEG Clin. Neurophysiol., 101, 478–482. 23. Maccabee, P., Amassian, V., Zimann, P., Wassermann, E., and Deletis, V. (1999). Emerging application in neuromagnetic stimulation. In “Comprehensive clinical neurophysiology” (K. Levin, and H. Luders, eds.), pp. 325–347. W.B. Saunders, Philadelphia.

Intraoperative Neurophysiology and Methodologies

51

24. Canter, M., and Deletis, V. (1995). Spinal epidural electrode catheter for intraoperative recording of evoked potentials (EPs) and injection of drugs. Sixth International Symposium on Spinal Cord Monitoring, abstract book, pp. 50, New York, June 1995. 25. Gokaslan, Z.L., Samudrala, S., Deletis, V., and Cooper, P.R. (1997). Intraoperative monitoring of spinal cord function using motor evoked potentials via transcutaneous epidural electrode during anterior cervical spine surgery. J. Spinal Disord., 10(4), 299–303. 26. Burke, D., Barthley, K., Woodforth, I.J., Yakoubi, A., and Stephen, P.H. (2000). The effects of a volatile anesthetic on the excitability of human corticospinal axons. Brain, 123, 992–1000. 27. Morota, N., Deletis, V., Shlomi, C., Kofler, M., Cohen, H., and Epstein, F. (1997). The role of motor evoked potentials (MEPs) during surgery of intramedullary spinal cord tumors. Neurosurgery, 41, 1327–1366. 28. Taniguchi, M., Cedzich, C., and Schramm, J. (1993). Modification of cortical stimulation for motor evoked potentials under general anesthesia: Technical description. Neurosurgery, 32(2), 219–226. 29. Jankowska, E., Padel, Y., and Tanaka, R. (1975). Projections of pyramidal tract to cells αmotoneurons innervating hind limb muscles in the monkey. J. Physiol., 249, 637–667. 30. Deletis, V., Isgum, V., and Amassian, V. (2001). Neurophysiological mechanisms underlying motor evoked potentials (MEPs) in anesthetized humans: Part 1. Recovery time of corticospinal tract direct waves elicited by pairs of transcranial stimuli. Clin. Neurophysiol., 112, 238–444. 31. Deletis, V., Rodi, Z., and Amassian, V. (2001). Neurophysiological mechanisms underlying motor evoked potentials (MEPs) elicited by a train of electrical stimuli: Part 2. Relationship between epidurally and muscle recorded MEPs in man. Clin. Neurophysiol., 112, 445–452. 32. Amassian, V.E., and Deletis, V. (1999). Relationships between animal and human corticospinal responses. In “Transcranial magnetic stimulation” (W. Paulus, M. Hallett, P.M. Rossini, and J.C. Rothwell, eds.), pp. 79–92. Elsevier, New York. Electroencephalogr. Clin. Neurophysiol. suppl. 51. 33. Deletis, V., and Kothbauer, K. (1998). Intraoperative neurophysiology of the corticospinal tract. In “Spinal cord monitoring” (E. Stalberg, H.S. Sharma, and Y. Olsson, eds.), pp. 421–444. Springer, Wien, New York. 34. Deletis, V. (1994). Evoked potentials. In “Clinical monitoring for anesthesia and critical care” (C. Lake, ed.), 2nd ed., pp. 282–314. W.B. Saunders, Philadelphia. 35. Deletis, V., Sala, F., and Morota, N. (2000). Intraoperative neurophysiological monitoring and mapping during brain stem surgery: A modern approach. Operative Techniques in Neurosurgery, 3(2), 109–113.

CHAPTER

3

Spinal Cord Surgery GEORGE I. JALLO AND FRED J. EPSTEIN Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York

1 2 3 4 5

Introduction Epidemiology and Histology Clinical Presentation Diagnostic Studies Surgical Management 5.1 Surgical Instruments 5.2 Surgical Techniques 6 Outcome Following Surgery 7 Surgical Complications 7.1 Deterioration in Functional Outcome 7.2 Cerebrospinal Fluid Leak 8 Conclusion References

ABSTRACT This chapter reviews the current surgical management for intramedullary neoplasms. The optimal management of these neoplasms remains controversial. The majority of these tumors are histologically benign, with low-grade astrocytomas being the most common in children, whereas ependymomas are the most common histology in adults. These tumors typically have a long prodrome of symptoms that include pain, motor deficit, or sensory loss. The recommended surgical approach is an osteoplastic laminotomy or laminectomy and radical resection of the tumor. This surgery is supplemented by specialized instruments such as the contact laser and ultrasonic aspirator as well as intraoperative neurophysiology. Gross total resection is feasible for ependymomas, hemangioblastomas, and cavernomas and results in a surgical cure. Although astrocytomas are typically infiltrative neoplasms, the radical resection with intraoperative neurophysiology results in a long progression-free survival. We avoid adjuvant radiotherapy for all neoplasms except the rare malignant glioma. These high-grade tumors have a dismal outcome, and surgery in these patients should be a conservative debulking with preservation of neurological function. Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

55

56

George I. Jallo and Fred J. Epstein

1 INTRODUCTION The first successful resection of an intradural tumor, a fibromyxoma, was accomplished in 1887 by Victor Horsley [1], and the first successful resection of an intramedullary spinal cord tumor was performed in 1907 by Anton von Eiselberg [2] in Austria. However, the first report of an intramedullary tumor was written in 1911 by Charles Elsberg in New York [3]. Elsberg described a two-stage strategy for the removal of these intramedullary tumors. At the initial operation a myelotomy would be performed. The surgeon would then return one week later to remove the spinal cord tumor. This technique allowed the neurosurgeon to remove only the extruded portion of an intramedullary tumor. Tumor within the spinal cord would not be removed for fear of neurological injury. Following these initial reports for spinal cord tumors, other pioneering neurosurgeons attempted spinal cord surgery. However, the complication rate, which included surgery at wrong levels, cerebrospinal fluid leaks, infection, paralysis, and death, was quite significant. Thereafter, many neurosurgeons recommended a conservative approach with biopsy, dural grafting, and radiation therapy regardless of histological diagnosis [4]. With the advent of the operating microscope, development of microsurgical techniques, imaging technology, and intraoperative neurophysiology, the strategy for these intramedullary neoplasms has further evolved. The majority of spinal cord tumors are histologically benign [5, 6], and the radical or gross total removal results in long-term survival with an acceptable morbidity [6–11].

2 EPIDEMIOLOGY AND HISTOLOGY Tumors of the intramedullary spine account for only 5 to 10% of all central nervous system tumors. A review of the computer surgical pathology database for intramedullary lesions at a single institution between 1991 to 1998 yielded 294 cases in adults and children [5]. The majority of these tumors were operated upon by the senior author (FJE). The 294 tumors included 117 removed from children under the age of 21 years, and 177 from patients 21 years and older (Table 3.1). The most common single tumor type in children was the fibrillary astrocytoma, which accounted for 45 (39%) of the tumors. There were 31 gangliogliomas, which were almost as common as the low-grade fibrillary astrocytoma in the pediatric population. In our study of 164 children with intramedullary neoplasms, the majority of the tumors were located in the cervicothoracic or thoracic spinal cord (Fig. 3.1) [6]. In contrast, ependymomas are the predominant tumor in adults, accounting for 45% of 177 intramedullary tumors [5].

57

Spinal Cord Surgery TABLE 3.1 Histological Diagnoses of Intramedullary Neoplasms at a Single Institution between 1991–1998 Tumor type

Children

Juvenile Pilocytic Astrocytoma

Adults

4

4

Fibrillary Astrocytomas Low Grade Anaplastic Glioblastoma

45 32 10 3

44 22 21 1

Ependymomas Myxopapillary Subependymoma

14 5 0

74 14 5

Oligodendroglioma

1

0

Mixed Glioma

3

5

Ganglioglioma

31

10

Miscellaneous Neuronal Hemangioblastoma Others

14 10 3 1

21 0 5 16

117

177

Total

70

Thoracic

60 50

Cervicothoracic

40 30 20

Cervical Cervicomedullary

Conus

10 0

FIGURE 3.1

A chart of tumor location in 164 children with intramedullary tumors.

3 CLINICAL PRESENTATION Intramedullary tumors may remain asymptomatic for a long time or cause nonspecific complaints that make the diagnosis difficult. The most common symptom of an intramedullary tumor in adults is pain. The pain may be diffuse or radicular in nature. There is no characteristic feature of the pain distribution in patients with intramedullary tumors, but patients with an intramedullary ependymoma tend to have dysesthetic pain as compared to astrocytomas. The diagnosis is even more difficult in children, who may not complain of pain, dysesthesias, or sensory loss. Other children may only complain of symptoms following a trivial fall or accident. Younger infants may even present with

58

George I. Jallo and Fred J. Epstein TABLE 3.2 Clinical Symptoms Prompting Radiological Investigation in the Author’s Series of 164 Children with an Intramedullary Neoplasm Symptom/sign

Percentage of children

Motor Regression

65.2%

Pain

45.7%

Gait Abnormality

37.2%

Dysesthesia

32.3%

Progessive Scoliosis

32.3%

abdominal pain and undergo extensive gastrointestinal investigations [4]. The onset of symptoms is often insidious, and symptoms are typically present for around 9 months. However, high-grade or malignant neoplasms typically have a shorter presentation than indolent low-grade tumors. Patients may also present with a motor deficit. This is the most common presentation in children with intramedullary tumors (Table 3.2). These deficits can result in clumsiness, weakness, or frequent falls. In children, this may manifest as motor regression, such as refusal to stand or crawl after having learned to walk. Scoliosis can also be a presenting complaint. This is seen in one third of children and young adults [12]. The direction of the scoliosis curve is not specific. Children with scoliosis typically have paraspinal pain, which is unusual for intramedullary tumors. Adult patients typically do not have scoliosis as a presenting complaint.

4 DIAGNOSTIC STUDIES Magnetic resonance imaging (MRI) is the imaging study of choice to identify an intradural spinal cord neoplasm. MRI scans should be performed with intravenous contrast agents (gadolinium diethylene-triamine-pentacetic acid) and in multiple planes. These images demonstrate the solid tumor component, associated cysts, and edema. Although MRI does not provide the histological diagnosis, there are some typical patterns of appearance for intramedullary tumors. Ependymomas tend to enhance brightly and homogeneously with contrast (Fig. 3.2). They are often associated with rostral and caudal cysts. These tumors are centrally located within the spinal cord. On the other hand, astrocytomas and gangliogliomas have a heterogeneous enhancement pattern (Fig. 3.3). These tumors are often eccentrically located and produce an asymmetric enlargement of the spinal cord. Intramedullary tumors such as cavernous malformations and hemangioblastomas also have distinct imaging qualities (Fig. 3.4). These lesions are typically situated near the dorsal surface of the spinal cord [11]. Hemangioblastomas, regardless of size, have a syrinx or associated edema.

Spinal Cord Surgery

59

FIGURE 3.2 MRI of a 42-year-old male who presented with several months of upper extremity weakness and dysesthesias. MRI appearance consistent with an intramedullary ependymoma. (A) T1-weighted image demonstrates a homogeneously enhancing tumor at C4-C5. (B) T2weighted image demonstrates the circumscribed tumor with rostral and caudal cysts. (C) The axial T1-weighted image demonstrates the central location of the tumor.

Computed tomography (CT) studies are reserved only for patients in which MRI is contraindicated or investigation of the bone anatomy is essential. This study is useful for the rare tumor that may involve bone and have extension to the spinal cord. With the advent of MRI, it is very unusual to diagnose intramedullary neoplasms with this imaging modality. Plain radiographs are mandatory for patients who present with scoliosis. In addition, young children who undergo an extensive laminotomy or laminectomy are at risk for developing postsurgical scoliosis and should be followed with serial radiographs.

60

George I. Jallo and Fred J. Epstein

FIGURE 3.2 (Continued)

5 SURGICAL MANAGEMENT

5.1 SURGICAL INSTRUMENTS The traditional method of suction and bipolar cautery for the removal of intramedullary neoplasms is now supplemented by specialized microinstruments. These instruments have become essential for the microsurgical resection of spinal cord tumors. The cavitron ultrasonic aspirator (CUSA; Valleylab, Boulder, CO) uses highfrequency sound waves to fragment and then suction tumor from the probe

61

Spinal Cord Surgery

FIGURE 3.2 (Continued)

tip [13, 14]. There have been significant advances in this technology. The handles have become smaller and lighter for more delicate use in the spinal cord. This allows for tumor removal with only minimal manipulation of the adjacent spinal tissue. The laser is an excellent surgical adjunct for intramedullary surgery. Although many surgeons do not use this technology, we prefer the Nd:YAG Contact Laser System (SLT, Montgomeryville, PA) to other laser systems such as the argon, CO2, or potassium titanyl phosphate (KTP) systems. The laser is used as a microsurgical instrument with its handpiece and various size and shapes of contact probes. The tip diameters of the probes for neurosurgical use vary from 200 µm to 1.2 mm (Fig. 3.5). The contact probes are useful as a scalpel to perform the myelotomy, to demarcate the glial–tumor interface, and to remove any residual fragments. Unlike other laser systems, there is minimal associated char and smoke generation. For lipomas and firm intramedullary or extramedullary tumors, the larger probes provide great precision for vaporization and internal debulking.

62

George I. Jallo and Fred J. Epstein

FIGURE 3.3 MRI of a 15-year-old male with an intramedullary astrocytoma from C3-C7. (A) T1weighted sagittal image demonstrates a heterogeneously enhancing tumor of the cervical spine. (B) Axial T1-weighted image typical for an astrocytoma.

5.2 SURGICAL TECHNIQUES The surgical approach for all intramedullary tumors is an osteoplastic laminotomy or laminectomy with the patient in the prone position. For cervical or cervicothoracic tumors, the head is fixed in a Sugita (Mizuho, Beverly, MA) or Mayfield (OMI, Cincinnati, OH) headholder. Venous hypertension, which may be significant in obese patients, is minimized using soft gel-rolls under the chest. The bone opening is done in a way that permits repositioning of the laminae when possible. Some authors advocate a laminoplasty with a threadwire saw [15].

63

Spinal Cord Surgery

FIGURE 3.3 (Continued)

However, a craniotome attachment on the Midas (Midas Rex, Fort Worth, TX) drill is an excellent alternative. In patients who have been previously operated upon, the laminae may not be present for repositioning. For these patients we perform a reopening of the laminectomy with adequate exposure of the bone anatomy. The opening, regardless of a laminotomy or laminectomy, is large enough to expose the solid component of the tumor. The rostral and caudal cysts do not need to be fully exposed. This opening is planned with intraoperative x-rays, and we seldom use fluoroscopy in localizing an intradural tumor. We prefer to use intraoperative x-rays and ultrasound, which allows us to visualize the spinal cord in two dimensions, sagittal and axial. Intramedullary astrocytomas and gangliogliomas have the same echogenicity as the spinal cord. In contrast, ependymomas tend to be hyperechogenic and can be readily differentiated from the spinal cord. This tumor tends to be more visible and in the center of the spinal cord. Ultrasound is helpful in identifying the associated tumor cyst(s). If the bone removal does not fully expose the solid tumor component, the laminectomy or laminotomy is extended prior to opening the dura. The dura is then opened in the midline. The spinal cord should be expanded and may occasionally be rotated. A surgeon should be wary of nonexpanded spinal cords. In these cases, one should suspect a nonneoplastic process. The asymmetric expansion and rotation of the spinal cord may make the identification of the midline difficult. In those exceptional cases with an asymmetric

64

George I. Jallo and Fred J. Epstein

FIGURE 3.4 MRI of 21-year-old girl with a cervical hemangioblastoma. (A) T1-weighted sagittal image with gadolinium demonstrates the enhancing small dorsally situated tumor which abuts the pial surface and the extensive syrinx. (B) Axial T1-weighted image demonstrates the small dorsally situated tumor.

tumor and rotated spinal cord, a myelotomy may be performed through the dorsal root entry zone by localizing the midline using the neurophysiological technique of “dorsal column mapping” (see Chapter 7). The neoplasm is typically located several millimeters underneath the dorsal surface. The contact laser is used to perform the myelotomy with minimal neural injury, and it does not interfere with the intraoperative neurophysiological monitoring. Intramedullary tumors have different appearances, such as texture and color, which help the neurosurgeon differentiate the tumor type.

65

Spinal Cord Surgery

FIGURE 3.4 (Continued)

5.2.1 Astrocytomas and Gangliomas Astrocytomas or gangliogliomas have a gray-yellow appearance. A true plane between tumor and normal spinal cord does not exist. The surgeon should make no effort to define this true interface because it results in hazardous manipulation of normal spinal cord tissue. Ependymomas are typically red or dark gray in color. These neoplasms have a clear margin from the surrounding spinal cord. This interface can be readily separated with a plated bayonet [16] or the scalpel probe of the contact laser. Once the tumor is exposed, a biopsy is taken for immediate histological examination. This information may be crucial in deciding the extent of tumor resection if the tumor is a malignant glioma or inflammatory process. For malignant gliomas, a more conservative approach, to limit any potential motor deficits, is undertaken. The goal is tumor debulking with preservation of motor function.

66

George I. Jallo and Fred J. Epstein

FIGURE 3.5 Photograph showing the tip of the laser probe (approximately 200 µm).

Tumor removal for low-grade astrocytomas and gangliogliomas begins after the initial myelotomy is performed. An internal debulking with the CUSA is done to reduce the tumor volume. The resection of astrocytomas is initiated at the midportion rather than the tumor poles. The rostral and caudal poles are the least voluminous, and manipulation at these locations may be the most dangerous to the normal spinal cord tissue. Then, using the suction or contact laser, the tumor is gently removed from the surrounding spinal cord tissue. These tumors do not have a cleavage plane, although in some areas a plane may exist between tumor and normal spinal tissue. These tumors tend to displace the motor tracts anteriorly or laterally. The surgeon should be aware of these pathways during tumor resection. 5.2.2 Ependymomas Ependymomas, which are more common in adults than in children, are typically located in the center of the spinal cord. They frequently have a rostral or caudal cyst. These tumors have a distinct cleavage plane that exists between the tumor and the normal spinal cord. After the myelotomy is performed, this plane can be identified most readily at the rostral or caudal pole. The contact laser is useful in defining this cleavage plane and in cutting the adhesions surrounding the tumor. The laser provides tactile sensation feedback and cuts tissue with minimal traction. Since the blood supply to this tumor comes from the ventral surface, extreme care is taken not to injure the anterior spinal artery.

Spinal Cord Surgery

67

The rostral-caudal length of the tumor does not influence the functional outcome after tumor resection. We have found the removal of a small tumor with a wide girth to be more difficult than removal of a long narrower tumor. This observation corresponds to previous reports that spinal cord atrophy is a poor prognostic factor [17]. 5.2.3 Vascular Tumors Hemangioblastomas in the spinal cord, regardless of size, are often associated with significant edema and syrinx formation. The resection of these lesions is similar to resection of their intracranial counterpart. The lesion should be resected in a circumferential fashion. The tumor surface can be coagulated to allow for the manipulation of the lesion; however, this tumor should not and cannot be debulked from within. We do not embolize these tumors, but a preoperative angiogram is still performed for large lesions. This is the only tumor type for which we do not stop surgery, regardless of information obtained from intraoperative neurophysiological monitoring, until the tumor is completely removed. Residual tumor may predispose the patient for future intraspinal hemorrhages. Cavernous malformations, similar to hemangioblastomas, are typically located on the dorsal surface of the spinal cord [11]. A bluish discoloration underneath the pial surface identifies the lesion. This vascular malformation is resected in an inside-out fashion, similar to the technique for astrocytomas. These lesions do not usually bleed during the resection; thus the CUSA or suction cautery can be safely used for their removal. Cavernous malformations are usually surrounded by a gliotic plane that permits delineation from the surrounding spinal cord tissue. These vascular malformations are quite uncommon in children. When these lesions present in the pediatric population, there is high probability for multiple intracranial lesions [11]. 5.2.4 Intramedullary Lipomas Intramedullary lipomas require a different surgical strategy than glial neoplasms. Although this tumor may appear well demarcated from the adjacent spinal tissue, these lesions are densely adherent. Thus total removal is fraught with neurological compromise. The contact laser is typically used to debulk these tumors. The laser vaporizes the fatty tissue without any surgical trauma to the spinal cord. The surgeon should only perform a debulking of this tumor, because further tumor growth is unlikely [18]. Following intramedullary tumor removal, hemostasis is obtained with warm saline irrigation and avitene. The dura is then closed primarily in a water-tight fashion. If an osteoplastic laminotomy was performed, the laminae are replaced

68

George I. Jallo and Fred J. Epstein

and secured with a nonabsorbable suture. One tissue layer must be closed in a water-tight fashion, and the muscle and fascial closure must not be under tension. A subcutaneous drain is placed only for reoperations. Patients who have had previous surgery and radiation therapy are at considerable risk for wound dehiscence and cerebrospinal fluid leak. These patients are maintained at bedrest for several days to allow for wound healing.

6 OUTCOME FOLLOWING SURGERY The major neurological hazard following intramedullary tumor surgery is paralysis. The incidence of paralysis is related to the preoperative motor status [6, 19]. Patients who have no or minimal preoperative motor deficits have less than 1% incidence of this postoperative complication. Thus today’s surgery on spinal cord tumors is relatively safe when the available microinstruments and intraoperative neurophysiological monitoring are used. Almost all patients undergoing gross total resection of intramedullary spinal cord tumors experience some immediate postoperative deterioration of neurological function. This neurological deterioration is typically temporary, and recovery occurs within a few hours, days, or weeks [20–23]. The incidence of clinical improvement after surgery is higher in patients undergoing total resection than in patients undergoing partial resection [24]. In our series the extent of resection, gross total (>95%) or subtotal resection (80–95%), did not significantly affect the long-term outcome. On the other hand, patients who underwent a partial resection or biopsy (<80%) fared significantly worse than those with radically removed tumors. These conclusions for intramedullary astrocytomas have been supported by others [25]. On the other hand, radical resection of malignant astrocytomas has failed to show any benefit [26–28]. In a review of the literature, patients with an intramedullary anaplastic astrocytoma or glioblastoma do not survive beyond 2 years, despite aggressive adjuvant radiotherapy and chemotherapy [25]. Thus, if at the time of surgery the frozen histology returns as a high-grade glioma, we prefer to perform a conservative resection of these neoplasms with preservation of motor function predicted intraoperatively by monitoring motor evoked potentials (see Chapters 2 and 4). The extent of resection is most beneficial for ependymomas. Many studies have shown substantial benefits associated with total resection of intramedullary ependymomas [8, 20, 27, 29, 30]. Survival for patients with ependymomas in one study was 219 months for patients undergoing total resection as opposed to 130 months for a subtotal removal [31]. These tumors, which have a cleavage plane, should be totally resected, and radiation therapy should be avoided.

Spinal Cord Surgery

69

Despite gross total resections for intramedullary neoplasms, residual microscopic fragments are left in the resection bed. These residual fragments may remain dormant or involute over time. There is no evidence that radiation therapy improves the outcome of low-grade astrocytomas or ependymomas [9, 21, 32, 33]. There is abundant evidence that radiation has deleterious effects on the nervous and osseous system [34–36]. There have been two studies that document alterations in motor and sensory evoked potentials in patients who have received radiation therapy [37, 38]. Some authors recommend radiotherapy for all intramedullary neoplasms [39, 40]; however, no prospective study has been performed comparing the results of radiotherapy. We recommend radiation therapy only for malignant tumors, patients with documented postoperative rapid tumor regrowth, and in those cases where substantial tumor remains and further surgery is not safely feasible. Although neuraxis radiation therapy is recommended for malignant tumors, these neoplasms invariably progress. Unfortunately, chemotherapy for these neoplasms has not been shown to be beneficial.

7 SURGICAL COMPLICATIONS

7.1 DETERIORATION IN FUNCTIONAL OUTCOME The earlier the diagnosis and radical resection of an intramedullary tumor, the greater the likelihood of preserving the patient’s neurological function [6]. Since 1987, no patient undergoing surgery who was functioning preoperatively at a modified McCormick Grade I level [41] deteriorated more than one grade. The incidence of deteriorating more than one functional grade for the other patients was 8% in our study of 164 children, and most patients who deteriorated had significant preoperative deficits. This confirms our thesis that unnecessary delays in surgery may reduce the recovery potential of the spinal cord.

7.2 CEREBROSPINAL FLUID LEAK This complication is unusual in most cases of intramedullary neoplasms, except for patients who have been previously operated on or who have received radiotherapy. In these patients there is a risk for wound dehiscence and cerebrospinal fluid fistula. Early in our experience we had several patients with poor wound healing and the development of meningitis or wound infection. We now use plastic surgical techniques for these patients [42]. We close the fascia in a water-tight fashion and without tension. This closure may require relaxing incisions to provide a tension-free closure. Subcutaneous drains are placed to allow

70

George I. Jallo and Fred J. Epstein

for the wound to heal. These patients are maintained at bed rest for several days to avoid this complication.

8 CONCLUSION Surgical resection for intramedullary neoplasms has evolved since the initial report of Elsberg. With the advent of the microsurgical technique, imaging technology, and intraoperative neurophysiology, the radical resection of intramedullary neoplasms is a safe and effective treatment. In particular, the neurophysiological monitoring of motor pathways is extremely helpful in achieving a radical resection for these intramedullary tumors. The functional outcome of surgery is best correlated with the preoperative status; thus surgery should be performed early prior to onset of severe motor deficits. The present surgical adjuncts allow us to recommend radical resection for the majority of intramedullary neoplasms. This approach should be the standard for intramedullary tumors because it provides excellent progression-free survival. Adjuvant radiation and chemotherapy should only be administered for malignant gliomas. Included with the accompanying CD is a video comparing the noise artifacts affecting the intraoperative recordings when using the bipolar cautery, as compared to the artifact-free recordings when the contact laser is being used (choose Chapter 3 from the accompanying CDs main menu).

REFERENCES 1. Gowers, W.R., and Horsley, V. (1888). Case of tumour of the spinal cord; removal; recovery. Medico-chirurgical Transactions, 53, 377–428. 2. von Eiselberg, A.F., and Ranzi, E. (1913). Ueber die chirurgische Behandlung der Hirn-Und Ruckenmarkstumoren. Arch. Klin. Chir., 102, 309–468. 3. Elsberg, C.A., and Beer, R. (1911). The operability of intramedullary tumors of the spinal cord. A report of two operations with remarks upon the extrusion of intraspinal tumors. Am. J. Med. Sci., 142, 636. 4. Wood, E.S., Berne, A.S., and Taveras, J.M. (1954). The value of radiation therapy in the management of intrinsic tumors of the spinal cord. Radiology, 63, 11–24. 5. Miller, D.C. (2000). Surgical pathology of intramedullary spinal cord neoplasms. J. Neuro.-Oncol., 47, 189–194. 6. Constantini, S., Miller, D.C., Allen, J.C., Rorke, L.B., Freed, D., and Epstein, F.J. (2000). Radical excision of intramedullary spinal cord tumors: Surgical morbidity and long-term followup evaluation in 164 children and young adults. J. Neurosurg. (Spine), 93, 183–193. 7. Greenwood, J. Jr. (1967). Surgical removal of intramedullary tumors. J. Neurosurg., 26, 276–282. 8. Guidetti, B., Mercuri, S., and Vagnozzi, R. (1981). Long-term results of the surgical treatment of 129 intramedullary spinal gliomas. J. Neurosurg., 54, 323–330.

Spinal Cord Surgery

71

9. Epstein, F., and Epstein, N. (1982). Surgical treatment of spinal cord astrocytomas of children. J. Neurosurg., 57, 685–689. 10. Epstein, F.J., and Farmer, J.P. (1990). Pediatric spinal cord tumor surgery. Neurosurg. Clin. N. Am., 1, 569–590. 11. Deutsch, H., Jallo, G., Faktorovich, A., and Epstein, F. (2000). Spinal intramedullary cavernoma: Clinical presentation and surgical outcome. J. Neurosurg. (Spine 1), 93, 65–70. 12. Yao, K., Kothbauer, K., Bitan, F., Constantini, S., Epstein, F., and Jallo, G. (2000). Spinal deformity and intramedullary tumor surgery. Child’s Nerv. Syst., 16, 530. 13. Flamm, E.S., Ransohoff, J.P., Wuchinich, D., and Broadwin, A. (1978). Preliminary experience with ultrasonic aspiration in neurosurgery. Neurosurgery, 2, 240–245. 14. Constantini, S., and Epstein, F.J. (1996). Ultrasonic dissection in neurosurgery. In “Neurosurgery” (R.H. Wilkins, and S.S. Rengachary, eds.), vol. 1, pp. 607–608. McGraw-Hill, New York. 15. Hara, M., Takayasu, M., Takagi, T., and Yoshida, J. (2000). En bloe laminoplasty performed with threadwire saw: Technical note. Neurosurgery, 48, 235–239. 16. Epstein, F.J., and Ozek, M. (1993). The plated bayonet: A new instrument to facilitate surgery for intra-axial neoplasms of the spinal cord and brain stem. Technical note. J. Neurosurg., 78, 505–507. 17. Hoshimaru, M., Koyama, T., Hashimoto, N., and Kikuchi, H. (1999). Results of microsurgical treatment for intramedullary spinal cord ependymomas: Analysis of 36 cases. Neurosurgery, 44, 264–269. 18. Lee, M., Rezai, A.R., Abbott, R., Coelho, D.H., and Epstein, F.J. (1995). Intramedullary spinal cord lipomas. J. Neurosurg., 82, 394–400. 19. Kothbauer, K., Deletis, V., and Epstein, F.J. (1997). Intraoperative spinal cord monitoring for intramedullary surgery: An essential adjunct. Pediatr. Neurosurg., 26, 247–254. 20. Brotchi, J., DeWitte, O., Levivier, M., Baleriaux, D., Vandesteene, A., Raftopoulos, C., FlamentDuran, J., and Noterman, J. (1991). A survey of 65 tumors within the spinal cord: Surgical results and the importance of preoperative magnetic resonance imaging. Neurosurgery, 29, 652–657. 21. Goh, K.Y., Velasquez, L., and Epstein, F.J. (1997). Pediatric intramedullary spinal cord tumors: Is surgery alone enough? Pediatr. Neurosurg., 27, 34–39. 22. Herrmann, H.D., Neuss, M., and Winkler, D. (1988). Intramedullary spinal cord tumors resected with CO2 laser microsurgical technique: Recent experience in fifteen patients. Neurosurgery, 22, 518–522. 23. Samii, M., and Klekamp, J. (1994). Surgical results of 100 intramedullary tumors in relation to accompanying syringomyelia. Neurosurgery, 35, 865–873; discussion 873. 24. Xu, Q.W., Bao, W.M., Mao, R.L., and Yang, G.Y. (1996). Aggressive surgery for intramedullary tumor of cervical spinal cord. Surg. Neurol., 46, 322–328. 25. Nadkarni, T.D., and Rekate, H.L. (1999). Pediatric intramedullary spinal cord tumors: Critical review of the literature. Child’s Nerv. Syst., 15, 17–28. 26. Cohen, A.R., Wisoff, J.H., Allen, J.C., and Epstein, F. (1989). Malignant astrocytomas of the spinal cord. J. Neurosurg., 70, 50–54. 27. Cooper, P.R. (1989). Outcome after operative treatment of intramedullary spinal cord tumors in adults: Intermediate and long-term results in 51 patients. Neurosurgery, 25, 855–859. 28. Houten, J.K., and Cooper, P.R. (2000). Spinal cord astrocytomas: Presentation, management and outcome. J. Neuro.-Oncol., 47, 219–224. 29. Cristante, L., and Herrmann, H.D. (1994). Surgical management of intramedullary spinal cord tumors: Functional outcome and sources of morbidity. Neurosurgery, 35, 69–76. 30. Whitaker, S.J., Bessel, E.M., Ashley, S.E., Bloom, H.J.G., Bell, B.A., and Brada, M. (1991). Postoperative radiotherapy in the management of spinal cord ependymoma. J. Neurosurg., 74, 720–728.

72

George I. Jallo and Fred J. Epstein

31. Innocenzi, G., Raco, A., Cantore, G., and Raimondi, A. J. (1996). Intramedullary astrocytomas and ependymomas in the pediatric age group: A retrospective study. Child’s Nerv. Syst., 12, 776–780. 32. Stein, B.M. (1979). Surgery of intramedullary spinal cord tumors. Clin. Neurosurg., 26, 529–542. 33. Stein, B.M. (1983). Intramedullary spinal cord tumors. Clin. Neurosurg., 30, 717–741. 34. Clayton, P.E., and Shalet, S.M. (1991). The evolution of spinal growth after irradiation. Clin. Oncol., 3, 220–222. 35. Duffner, P.K., Horowitz, M.E., Krischer, J.P., Friedman, H.S., Burger, P.C., Cohen, M.E., Sanford, R.A., Mulhern, R.K., James, H.E., Freeman, C.R., Seidel, F.G., and Kun, L.E. (1993). Postoperative chemotherapy and delayed radiation in children less than three years of age with malignant brain tumors. N. Eng. J. Med., 328, 1725–1731. 36. Marcus, R.B., and Million, R.R. (1990). The incidence of myelitis after irradiation of the cervical spinal cord. Int. J. Rad. Oncol. Biol. Phys., 19, 3–8. 37. Morota, N., Deletis, V., Constantini, S., Kofler, M., Cohen, H., and Epstein, F.J. (1997). The role of motor evoked potentials during surgery for intramedullary spinal cord tumors. Neurosurgery, 41, 1327–1336. 38. De Scisciolo, G., Bartelli, M., Magrini, S., Biti, G.P., Guidi, L., and Pinto, F. (1991). Long-term nervous system damage from radiation of the spinal cord: An electrophysiological study. J. Neurol., 238, 9–15. 39. O’Sullivan, C., Jenkin, R.D., Doherty, M.A., Hoffman, H.J., and Greenberg, M.L. (1994). Spinal cord tumors in children: Long-term results of combined surgical and radiation treatment. J. Neurosurg., 81, 507–512. 40. Minehan, K.J., Shaw, E.G., Scheithauer, B.W., Davis, D.L., and Onofrio, B.M. (1995). Spinal cord astrocytoma: Pathological and treatment considerations. J. Neurosurg., 83, 590–595. 41. McCormick, P.C., and Stein, B.M. (1990). Intramedullary tumors in adults. Neurosurg. Clin. N. Am., 1, 609–630. 42. Zide, B.M., Wisoff, J.H., and Epstein, F.J. (1987). Closure of extensive and complicated laminectomy wounds: Operative technique. J. Neurosurg., 67, 59–64.

CHAPTER

4

Motor Evoked Potential Monitoring for Intramedullary Spinal Cord Tumor Surgery KARL F. KOTHBAUER Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York

1 Introduction 2 Neurophysiology 2.1 D Waves (“Epidural MEPs”) 2.2 Muscle MEPs 3 Anesthesia 4 Safety 5 Clinical Assessment and Correlation 6 Practical Surgical Application of Intraoperative Neurophysiological Information 6.1 Feasibility and Practicality of Monitoring 6.2 Interpretation of D Wave Data 6.3 Interpretation of Muscle MEP Data 6.4 Combined Interpretation of D Wave and Muscle MEP Data 6.5 Influence of MEP Monitoring on Extent of Resection 6.6 Observations on the Behavior of MEPs During Intramedullary Spinal Cord Tumor Surgery 7 Illustrative Cases 7.1 Case 1 7.2 Case 2 7.3 Case 3 7.4 Case 4 Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

73

74

Karl F. Kothbauer

8 Summary Acknowledgments References

ABSTRACT This chapter reviews several important issues concerning intraoperative motor evoked potential (MEP) monitoring of spinal cord surgery. First, it gives an overview of the technique of MEP monitoring in a surgical environment. Second, it reviews the evidence for accurate representation of the pre- and postsurgical motor status by combined epidural and muscle MEP monitoring. Third, it addresses the intraoperative impact of the neurophysiological information on the course of the procedure (i.e., how “useful” monitoring is to the surgeon). Motor potentials are evoked by transcranial electrical motor cortex stimulation. With the single-stimulus technique, D waves are elicited that are recorded from the spinal cord. With the train-stimulus technique, muscle MEPs are elicited that are recorded from limb muscles. The amplitude of the D waves and the presence or absence of muscle MEPs are the critical parameters for MEP interpretation. The practical procedures for MEP recording fit well into a neurosurgical environment, and the intraoperative interpretation and application of neurophysiological information are fast and straightforward. MEP monitoring is almost always possible in a patient who is not severely disabled already prior to surgery. Pre- and postoperative clinical motor findings correlate with intraoperative MEP data. As a result, correct prediction of the clinical outcome at any given time during the operation is possible with considerable certainty (the sensitivity of muscle MEPs for postoperative motor deficits is next to 100%, and its specificity is about 90%). Thus, MEP data do reflect the clinical “reality” for a patient. Loss of muscle MEPs, interpreted in combination with stepwise decreases in the D wave amplitude, leaves a considerable window of warning between the first changes in recordings and a permanent injury to the essential motor pathways. Thus the surgical strategy can be adapted before irreversible neurological damage occurs. On the other hand, present and stable recordings allow complete resection of tumors with confidence about the integrity of the motor pathways. These two features make MEP monitoring during intramedullary surgery very useful. Use of this monitoring technique should become a standard for this type of surgery.

1 INTRODUCTION Neurosurgeons have great respect for the spinal cord. The resection of intramedullary tumors of the spinal cord is believed to carry a high risk for surgical damage and subsequent neurological dysfunction because the cord is a delicate structure with tightly packed essential pathways and neural circuits. Going into the cord with surgical instruments and manipulation involves the possibility of selective damage to the motor pathways, the sensory pathways, or the cord’s intrinsic neural apparatus. Since paralysis, the loss of voluntary motor control, is the most feared neurological complication, it is essential to have a tool available

Motor Evoked Potential Monitoring for Intramedullary Spinal Cord Tumor Surgery

75

that directly and selectively monitors the functional integrity of the motor system during resection of intracord lesions. Intraoperative neurophysiological monitoring with motor evoked potentials (MEPs) is exactly this tool. The first impulse to use evoked potentials to assess functional aspects of the spinal cord during surgical procedures came from orthopedic surgeons [1, 2]. Somatosensory evoked potentials (SEPs) were the only monitoring methodology available at the time. From the present “digital age” perspective, the difficulties with slow recording hardware, difficult documentation, and lack of experience must have been formidable. In addition, a serious conceptual problem comes with monitoring of SEPs: they reflect, of course, the functional integrity of the sensory pathways and therefore provide only indirect information on the motor pathways. This may be acceptable for orthopedic surgery, in which external cord compression would be the mechanism of injury in most cases. SEP monitoring is still used as a means of “overall” spinal cord monitoring in centers where MEP monitoring is not available. And SEP monitoring has indeed been beneficial in large numbers of spinal orthopedic operations [3]. However, when the resection of lesions within the cord is attempted, there is a considerable risk of selective damage to the motor tract that is not reflected by changes in SEP recordings [4, 5]. Furthermore, change in or loss of SEPs during intramedullary operations is quite common and is certainly associated with the need to enter the cord through the dorsal midline [6], but it does not correlate with the motor outcome. And, of course, there is the problem of recording delay, since with averaging times the identification of injury can lag considerably behind the progress of the surgical procedure. Because of this known shortcoming of SEPs, to this day many neurosurgeons have the somewhat unreflected view of neurophysiological monitoring in general: when changes are recognized the damage has been done, and it is too late for any intervention anyway. With Merton’s first paper on electrically evoked MEPs [7], neurosurgeons quickly understood the potential of this technique for the direct and selective intraoperative assessment of motor pathways in both the brain and the spinal cord. The early papers still reflect the methodological difficulty associated with making use of the new method in the operating room [8]. Going back to concepts of the motor system developed since the 1950s [9, 10], this direct motor monitoring technique became based on firm ground, the concept of corticospinal recordings that reflect a small but essential fiber population in the cord. This was the introduction of the D wave to the operating room [11–14]. In addition to this recording technique, which requires recording from the cord with an epidurally placed electrode, muscle recording techniques were introduced with magnetic [15] and electrical [16] stimulation of the brain. Anesthesia posed a major problem, but the development of a multipulse or train-stimulus technique [17, 18] resolved this difficulty.

76

Karl F. Kothbauer

In the 1990s the understanding of neurophysiology, interpretation, and safety increased, and the experience with practical application improved. Several series provided first evidence that intraoperative MEP monitoring really works and is indeed useful for the experienced neurosurgeon [19–23]. Our group has gained considerable experience in the combined use of D wave and muscle MEP monitoring [21, 23–27]. Based on the neurosurgical and neurophysiological experience of several hundred operations on spinal cord tumors from 1996 to 2000, and on the accumulated scientific evidence, this chapter attempts to illustrate that MEP monitoring of the functional integrity of the motor pathways during intramedullary surgery is one of the most impressive, accurate, and useful evoked potential monitoring technique in use today. MEP monitoring can be safely done, it represents the clinical “reality,” and its concept provides a warning window of reversible change. This allows for spinal cord tumor resection with much greater safety from neurological injury and with much greater confidence for radical resection during the operation.

2 NEUROPHYSIOLOGY The intraoperative MEP monitoring techniques currently used for intramedullary surgery are the result of a long and ongoing development to which many individuals have contributed [7, 9, 10, 12, 14, 16, 18–21, 24–26, 28–31]. Motor potentials are evoked with transcranial electrical stimulation. The stimulus points are C3, C4, C1, C2, Cz, and a point 6 cm in front of Cz (International 10-20 EEG electrode system [32]) (Fig. 4.1A). Electrical stimulation is performed with constant current square wave stimuli of 500 µs duration and intensities between 15 and 200 mA.

2.1 D WAVES (“EPIDURAL MEPS”) D waves, or “epidural MEPs” [9], are elicited with single stimuli. This is therefore called the “single-stimulus technique” (Fig. 4.1B). Responses are recorded from the spinal cord with an electrode (Type JX-330, Arrow International, Inc., Reading, PA) inserted into the spinal epidural space by the surgeon after the laminectomy. One electrode is placed caudally, and if the tumor location permits it, another one is placed cranially as a control (Fig. 4.1C). The signals are amplified 10,000 times, the filter bandpass is set from 1.5 to 1700 Hz, and responses are recorded within 20 ms epochs. Baseline recordings are obtained before opening of the dura. The signal usually requires no averaging but may, at times, require a few (averaged) responses. Stimulation is repeated at a rate of 0.5 to 2 Hz during the critical part of the procedure. This provides fast, “real-time,” feedback.

Motor Evoked Potential Monitoring for Intramedullary Spinal Cord Tumor Surgery

77

FIGURE 4.1 Motor evoked potential techniques. (A) Schematic view of the stimulus points for transcranial electrical stimulation of the motor cortex. (B) Schematic representation of the single stimulus and the train stimulus techniques. (C) Cranial and caudal epidural D wave recording. (D) Muscle MEP recording.

The parameter monitored in epidural recordings is the peak-to-peak amplitude of the D wave. It has been shown that a decrease of more than 50% from the baseline value is associated with a long-term motor deficit [25]. Latency changes of the D wave are rare and are due to nonsurgical influences such as temperature [24] (see also Chapter 2, Fig. 2.7, page 38). A change in

78

Karl F. Kothbauer

stimulus intensity also alters the D wave latency: higher intensities lead to shorter latencies, implying that the corticospinal tract fiber activation occurs deeper in the white matter of the brain [14].

2.2 MUSCLE MEPS Muscle MEPs are also elicited with transcranial electrical stimulation. A short train of five to seven stimuli, with a 4 ms interstimulus interval [33, 34], is used. Thus this technique is called the “train stimulus technique”(Fig. 4.1B), or multipulse technique [22]. Compound muscle action potentials (CMAPs) are recorded with needle electrodes from target muscles in all four extremities (e.g., the thenar and tibialis anterior muscles) (Fig. 4.1D). More recently, the abductor hallucis has proven to be useful, probably because this muscle has rich pyramidal innervation [35]. The signals are amplified 10,000 times and are recorded on epochs of 100 ms with a filter setting. CMAPs are amplified 10,000 times and are recorded during a 100 ms epoch with a 1.5 to 853 Hz filter setting. Baseline recordings are obtained after positioning of the patient on the operating table. Like the D waves, muscle MEP signals do not require averaging and can be repeated at a rate of 0.5 to 2 Hz. Therefore, real-time feedback is possible here as well. With the focal anode as the stimulating electrode, a montage of C1/2 or C2/1 is tried first to elicit muscle MEPs in all four extremities. In individual cases, C3/4, C4/3, or CZ/6 is used as an alternative stimulation point. Muscle MEPs are recorded in an alternating fashion with D waves. The principle of evoking muscle responses follows from the D wave concept: each individual electrical stimulus on the motor cortex, either with exposed cortex or with transcranial stimulation [13], elicits a D wave in the corticospinal tract. A train of five stimuli, with an interstimulus interval of 4 ms, elicits five D waves that travel down the corticospinal tract 4 ms apart. Thus the spinal αmotorneurons are hit by five consecutive descending volleys and their membrane potential is elevated above firing threshold. This mechanism has been shown with direct intracellular recordings from α-motorneurons in primate animal experiments [10]. The additional role of facilitation of I waves [33, 34] is discussed in a separate chapter (see Chapter 2). The parameter monitored during spine and spinal cord surgery is the presence or absence of muscle MEPs in the target muscles. The stimulus intensity typically ranges from 15 to 200 mA. This all-or-none concept has been adopted for two reasons. First, in contrast to epidural MEPs, which show little amplitude variation [36], the variability of muscle MEP amplitudes is tremendous [19, 30, 37]. Thus defining a threshold amplitude below which one expects an intraoperative injury would be extremely difficult. Second, our experience as well as the evidence in the reported series (even with varying stimulus patterns)

Motor Evoked Potential Monitoring for Intramedullary Spinal Cord Tumor Surgery

79

has indicated that a motor deficit occurred only when the muscle response was lost [16, 19, 30]. Our data do not support the idea of using an increase of muscle MEP stimulus thresholds as a monitoring parameter [22]. The thresholds for eliciting epidural MEPs fluctuate very little. On the contrary, thresholds for muscle MEPs vary considerably during an individual procedure. However, only the presence or absence of muscle responses consistently correlates with the clinical findings. Experience with monitoring supratentorial surgeries is somewhat different (see Chapter 15). All recordings in our institution are obtained with the Axon Sentinel-4 EP analyzer (Axon Systems Inc., Hauppauge, NY) equipped with a dedicated software for controlling transcranial stimulation paradigms.

3 ANESTHESIA The effects of anesthetic agents on neurophysiologic recordings is extensively covered in a separate chapter of this book (see Chapter 17). Therefore, only a brief outline of the principles of anesthesia compatible with neurophysiological monitoring is given here. An anesthesia regimen that allows for intraoperative MEP monitoring consists of a constant infusion of propofol (usually in a dose of about 100–150 µg/kg/min) and fentanyl (usually around 1 µg/kg/hr). The use of propofol for anesthesia with MEP monitoring has been reported with various stimulation techniques [38–42]. Nitrous oxide not exceeding 50 Vol% can be used. Bolus injections of both intravenous agents should be avoided because this temporarily disrupts the generation of muscle MEPs, which are particularly important during the critical resection part of the operation. Halogenated anesthetics cannot be used [17, 18]. The use of short-acting muscle relaxants must be limited to the period of intubation. Partial myorelaxation is used by some groups [30], but there is no convincing evidence that its use makes management of anesthesia easier or safer. This has been a controversial issue, and our group has been subject to considerable criticism in many discussions. Many neurosurgeons are still very reluctant to accept even the possibility of slight movement during dissection. Why should we avoid muscle relaxation? With the patient fully relaxed, muscle MEP monitoring is impossible. “Controlled” relaxation would add an uncontrollable variable in the interpretation of MEPs that would reduce the specificity of muscle MEP monitoring. On the other hand, it is still unlikely that patient movement from stimulation can be completely avoided. In other words, controlled relaxation combines poor monitoring with poor relaxation. Our experience has been that relaxation is not necessary once both surgeon and anesthesiologist are used to working without it. In our practice we have not encountered problems because muscle relaxation was not used.

80

Karl F. Kothbauer

4 SAFETY Aside from direct neural tissue damage [18, 43], the main safety concern with the use of transcranial electrical multipulse stimulation has been the issue of seizures. None of our patients have had an epileptic event. There are no reports in the literature about intraoperative seizure induction with transcranial electrical stimulation using either a single or a train stimulation technique. The term kindling has been indiscriminately used in this context. Kindling is an experimental model that refers to the induction of self-perpetuating epileptic foci in experimental animals. It requires daily repeated electrical stimulation at a rate of 50 Hz for several seconds. This paradigm is different from the MEP train stimulation paradigm of 250 Hz for 25 ms (Fig. 4.1B). In addition, kindling of an epileptic focus requires a long period of time (weeks to months), particularly in primates [44], and, to our knowledge, it has not been shown to occur in humans. Furthermore, the energy necessary to induce a seizure in electroconvulsive therapy (also with 50 Hz stimulation applied for several seconds) is two orders of magnitude higher than the overall energy used for MEP monitoring [45]. All data reported so far, as well as the theoretical concept of transcranial electrical stimulation with a short high-frequency train to elicit muscle MEPs, indicate an extremely low risk of inducing seizures. The only adverse events that we have experienced are minor laceration and hematoma of the tongue as a result of strong contraction of the masticatory muscles due to direct stimulation from the cranial stimulation electrodes. For the most part, this has been avoided by tongue protection with a padded oropharyngeal tube. Complications such as injury or infection due to electrode placement or stimulation, or spinal epidural hematomas resulting from placement of epidural electrodes, have not been reported and have not occurred in our experience.

5 CLINICAL ASSESSMENT AND CORRELATION From 1996 to 2000, about 250 operations for removal of intramedullary spinal cord tumors were performed with the use of intraoperative neurophysiological monitoring at our institution. The majority of patients were either children or younger adults, with a median overall age of about 29 years (1–75 years). The vast majority of spinal cord tumors are low-grade astrocytomas in children and ependymomas in adults. Gangliogliomas, dermoids, hemangioblastomas, cavernomas, and other lesions also occur. Historically, malignant spinal cord tumors

Motor Evoked Potential Monitoring for Intramedullary Spinal Cord Tumor Surgery

81

are rare. The median size of the neoplasms is a segmental extension of four spinal segments (range 1–12). In our practice about one operation in five is a reoperation. Usually pre- and postoperative motor function is classified as normal (no focal motor deficit), slightly paretic (motor deficit not exceeding 4/5 and not significantly impairing the extremity’s function, walking not impaired), severely paretic (motor deficit 3/5 or worse, significantly impaired function of extremity, or inability to walk), and plegic (0/5 or 1/5). This is in principle consistent with the McCormick scale [46]. The extent of surgical resection is assessed as gross-total resection (90% resection or more), subtotal resection (50–90%), partial resection (<50%), or biopsy based upon the early postoperative MRI. In a previously published series [23] of 100 consecutive operations of spinal cord tumors, 92 of the 100 patients had a normal or slightly impaired motor status before surgery. In all of these 92 cases, muscle MEPs could be recorded at the beginning of surgery (“baseline”). Epidural MEPs were recordable in 59 of the 86 cases not involving the conus medullaris. Eight patients had severe motor deficits or were paralyzed. None of them had recordable MEPs (neither epidural nor from muscle). In no preoperatively paralyzed extremity was there ever a muscle MEP recordable. Postoperatively, a short-term motor status deterioration is noted in about every third patient (35 of 92 patients, or 38%, in the aforementioned series [23]). A severe permanent neurological dysfunction occurred as a direct result of the operation in only 2 of about 250 patients (unpublished data). Therefore, the risk of paraplegia following resection of a spinal cord tumor is lower than is commonly believed. These changes in clinical status were correctly reflected by the intraoperative MEP findings.

6 PRACTICAL SURGICAL APPLICATION OF INTRAOPERATIVE NEUROPHYSIOLOGICAL INFORMATION

6.1 FEASIBILITY AND PRACTICALITY OF MONITORING Electrodes are attached to the patient at the same time that the anesthesia preparations (intubation, intravenous and intra-arterial lines) are done. Additional time, if any, required for monitoring preparations is minimal. The epidural recording electrodes are placed by the surgeon during the operation from the surgical field to a preamplifier attached to the end of the operating table. This does

82

Karl F. Kothbauer

not disturb the dissection. Occasionally, scarring from a previous surgery prohibits insertion of a recording electrode. In terms of monitorability, practically all patients without severe preoperative motor deficits can be monitored with either D waves or muscle MEPs or both. The recordings are usually robust (between 10 and 70 µV amplitudes of epidural and up to 2 mV in muscle MEPs). Changes due to nonsurgical influences (intravenous bolus of anesthetic, temperature or blood pressure changes) can be recognized by following these parameters together with the anesthesiologist. Intraoperatively, the combined data of epidural and muscle MEPs indicate jeopardized functional integrity of the motor pathways at some point during the procedure in almost every other patient with monitorable responses. In about a third of all cases, these changes remain until the end of the operation and then correlate with a temporary motor deficit. In the remainder of cases the changes are reversible and correlate with intact motor function when the patient awakes from anesthesia.

6.2 INTERPRETATION OF D WAVE DATA Two factors are important for interpretation of intraoperative D wave recordings: the presence of the D wave and, when it is present, its peak-to-peak amplitude. The monitorability of the D wave and the significant intraoperative decline of its amplitude have been shown to be of predictive value for motor outcome after surgery for intramedullary spinal cord lesions [25]. In about two thirds of the patients with nonconus tumors, a D wave is recordable [23, 25]. (Since D waves are generated by the corticospinal tract axons, conus tumors are not monitorable with epidural MEPs.) Patients in whom the baseline recording of epidural MEPs produces no response have a higher risk of postoperative motor deficits than those with a recordable D wave [25]. It is not known whether this is due to an inherent subclinical damage and “vulnerability” of the motor tract, or to the fact that there was no monitoring support for the surgery. The mechanism of an unrecordable D wave coinciding with intact motor status (and recordable muscle MEPs) is believed to be due to chronic or inherent preexisting damage to the corticospinal tract resulting in a desynchronization of the wave [31] (see also Chapter 2, Fig. 2.11, page 41). It appears that patients with prior surgery, those with very extensive tumors, and particularly those with prior radiation therapy are in this group. The intraoperative amplitude decrease of the D wave correlates with postoperative outcome. If the D wave is unchanged, there is no permanent postoperative deficit. If it declines more than 50% of the baseline value or even disappears, the patients are likely to have a permanent motor deficit [11, 25].

Motor Evoked Potential Monitoring for Intramedullary Spinal Cord Tumor Surgery

83

6.3 INTERPRETATION OF MUSCLE MEP DATA The presence of muscle MEPs indicates that the functional integrity of the corticospinal tract is intact in all instances. Occasionally, in patients with a moderate motor deficit, it may be difficult to obtain recordings from both lower extremities. If that occurs, responses in the weaker leg usually require higher stimulation intensities. Intraoperative preservation of muscle MEPs means intact motor function postoperatively in all cases. Intraoperative loss of muscle MEPs indicates some postoperative impairment of voluntary motor control with a high (∼90%) specificity. For instance, muscle MEPs lost in one leg during the resection mean that the patient will postoperatively be unable to move this particular extremity, at least for a limited period of time. We call this a “temporary motor deficit.” Loss of muscle MEPs in both legs obviously is indicative of a bilateral motor deficit. Unilateral loss is of less concern because it has been shown in the past that unilateral motor disruption always recovers through a mechanism by which the intact side “takes over” control of the affected side [47]. These changes must be interpreted together with changes in D wave amplitude.

6.4 COMBINED INTERPRETATION OF D WAVE AND MUSCLE MEP DATA [23, 27] The D wave amplitude is a measure of the number of synchronized fast conducting fibers in the corticospinal tract. If 50% of these fibers are damaged by the procedure, the amplitude will decrease to 50% of its baseline value. From practical experience we know that the D wave decrease usually occurs in small steps, going down 15%, 20%, 30%, and so forth. Most D wave amplitude decreases will coincide with a loss of muscle MEPs. It may be, however, that muscle MEP loss occurs without D wave amplitude decrease, or that the D wave decreases without any changes in muscle MEPs. The underlying mechanisms of these observations are not understood at this time. In any event, preservation of the D wave above the 50% cutoff has been found to be predictive of long-term preservation (or recovery) of voluntary motor control. If there is loss of muscle MEPs, with preserved D wave amplitude, a temporary motor deficit can be expected postoperatively. However, in this situation it is still safe to complete a resection, or to pause and wait for recordings to improve again (which they often do). This situation is the window of warning, the window of reversible change, that allows for a change in surgical strategy before irreversible injury has occurred. This is the concept that the skeptical neurosurgeon must study before intraoperative MEP monitoring is dismissed as “useless.”

84

Karl F. Kothbauer

6.5 INFLUENCE OF MEP MONITORING ON EXTENT OF RESECTION In some cases, resection is terminated before the desired extent of resection is achieved because the acceptable limit of MEP change is reached. Although the actual number of cases is low (about 5% in the previously mentioned series [23]), this is an important factor in subsequent decision making. Depending upon the actual extent of resection on the postoperative MRI scan and the histology of the lesion, a second stage of the resection may be attempted once the patient has recovered motor control. This is of particular importance in patients with spinal cord ependymomas, for whom complete tumor resection is essential for long-term, progression-free, survival [48]. Among other criticisms, it has been claimed that MEP monitoring may be too sensitive, actually stopping resection too early [49]. Although a comparison inevitably must remain incomplete, it appears from the available data that the extent of resection increased since monitoring began at our institution [50]. However, there may still be a significant trade-off between extent of resection and preservation of motor function in some cases. On the other hand, in an equally significant number of cases, stable recordings encouraged the surgeon to proceed with tumor resection even though the anatomical situation suggested otherwise.

6.6 OBSERVATIONS ON THE BEHAVIOR OF MEPS DURING INTRAMEDULLARY SPINAL CORD TUMOR SURGERY Usually MEP changes occur toward the end of the resection. Since most spinal cord tumors are resected in an inside-out and piecemeal fashion (with the exception of some ependymomas), direct manipulation or vascular compromise occurs when the tumor-cord interface is reached. Often muscle MEPs disappear first. This may be preceded by an increase in threshold for this particular muscle response. Pausing the resection and irrigating the cavity with warm saline sometimes results in reappearance of the response. Similarly, some D wave amplitude decrease may also be reversible by pausing and irrigating. If dissection in a particular location results in MEP changes, the resection can often proceed at a different spot without further change. Sudden drops in D wave amplitude, often coinciding with sudden loss of muscle MEPs, is believed to be associated with some vascular compromise rather than with direct physical manipulation of the nervous tissue. In some patients, temporary moderate elevation of mean blood pressure has been a successful means to improve the MEPs, with a satisfactory clinical result postoperatively.

Motor Evoked Potential Monitoring for Intramedullary Spinal Cord Tumor Surgery

85

In some situations the warning provided by intraoperative MEP changes is still only of documentary value: whereas the removal of an astrocytoma can be terminated without jeopardizing the patient’s neurologic and oncologic outcome, the attempt to remove a hemangioblastoma is an all-or-none enterprise. The lesion has to be entirely removed, no matter what the MEP parameters indicate, or serious bleeding and/or swelling will lead to certain damage of the spinal cord. This, however, is a limitation imposed by the anatomic nature of the lesion, rather than a shortcoming of the monitoring technique. Other important surgical observations concern the use of specific surgical instruments and their impact on changes in MEP recordings. For instance, it appears to be our own and the observation of others that the use of the ultrasonic aspirator (CUSA, Valleylab, Boulder, Co) sometimes results in MEP deterioration. On the contrary, use of the Nd:YAG hand-held “contact” laser (SLT, Surgical Laser Technologies, Inc., Montgomery, PA) or special bipolars [51] to vaporize tumor and mobilize small fragments seems to be less damaging. Using a bipolar coagulation always disrupts MEP recordings for the time the current is active. One of the distinct advantages of the hand-held laser is that its use does not produce an electrical artifact, and therefore monitoring continues undisturbed during its use.

7 ILLUSTRATIVE CASES

7.1 CASE 1 A 14-year-old girl presented with progressive dysesthesias in the left arm and leg and a slight weakness of the left extremities. MRI disclosed an intramedullary spinal cord tumor from C3 to C7 that turned out to be an astrocytoma. During microsurgical gross total resection of the lesion, monitoring the D wave (Fig. 4.2A) showed no change in its amplitude. Muscle MEPs in the anterior tibial muscles bilaterally showed preserved responses until the end of surgery (Fig. 4.2B). The MEP data indicated preserved functional integrity of the motor system. Postoperatively the motor status was unchanged.

7.2 CASE 2 A 9-year-old girl underwent resection of a cervicothoracic intramedullary ganglioglioma. There was no preoperative motor deficit. Toward the end of the procedure, when only a small amount of tumor tissue remained, a sudden loss of muscle MEPs in the right anterior tibial muscle occurred (Fig. 4.3A). Simultaneously a drop of the D wave amplitude of about 40% was noted

86

Karl F. Kothbauer

FIGURE 4.2 Preserved functional integrity of the motor pathways. Epidural (A) and muscle MEP (B) recordings of Case 1. Stable D wave amplitude and preserved muscle MEPs correlate with intact motor control.

FIGURE 4.3 Temporary motor deficit. D wave amplitude decrease of 40% (B) and unilateral loss of muscle MEPs in the Right TA (A) correlate with monoplegia of the right leg immediately after surgery.

Motor Evoked Potential Monitoring for Intramedullary Spinal Cord Tumor Surgery

87

(Fig. 4.3B). The resection was terminated. Some residual tumor tissue lining the cavity was left. Immediately after surgery the patient had a monoplegia of the right leg, as expected from the MEP data. Recovery started on the first postoperative day with some movements of the toes. After one week she had regained antigravity muscle strength.

7.3 CASE 3 A 27-year-old woman underwent resection of a C5-Th1 ependymoma. Preoperatively she had slight leg weakness. At baseline, muscle MEPs were present in the right leg only (Fig. 4.4B). The absence of a muscle response in the other leg indicated subclinical damage to the functional integrity of the motor tracts. Early in the dissection it became clear that, morphologically, the tumor had extremely thinned the surrounding surviving cord tissue. However, stability of the epidural as well as the single side muscle MEP recording encouraged an attempt for tumor removal. The tumor was then entirely removed, and the D wave amplitude decreased but remained above the 50% limit (Fig. 4.4A). Eliciting the right leg muscle MEPs required a higher intensity and seven instead of five stimuli per train. Nevertheless, they remained present. The postoperative clinical status was not significantly changed.

FIGURE 4.4 Continual recordings as reassurance of intact motor pathways in difficult surgical dissection (case 3). The D wave amplitude (A) declined throughout the critical part of the procedure but not more than 50% of the baseline value. Muscle MEPs (B) were only present on one side at the beginning of the procedure. The response was preserved until the tumor was completely removed. The patient had no significantly increased postoperative motor deficit.

88

Karl F. Kothbauer

7.4 CASE 4 A 3-year-old boy underwent a T4 to T11 laminotomy for resection of a thoracic low-grade astrocytoma. Preoperatively he had a significant degree of scoliosis but no neurological deficits. At baseline, there was no D wave because the tumor involved the cord almost all the way to the conus. Muscle MEPs were only recordable in the toe abductor muscles but not in the tibialis anterior muscles, indicating subclinical damage to the motor system. During resection, after some minor fluctuation of stimulus thresholds, a sudden loss of muscle MEPs occurred after a bleeding tumor vessel was coagulated (Fig. 4.5A). The resection was stopped, warm irrigation was applied to the resection bed, and the blood pressure was increased from 110/80 to about 140/95. After about 5 min the MEPs reappeared in the toe abductor on the right side, albeit with a higher stimulus threshold (Fig. 4.5B). The patient started to move his right leg several hours after the operation. Thus even poor recordings at the baseline provide useful intraoperative information and correctly correlate with postoperative recovery.

FIGURE 4.5 Monitoring difficulty. In Case 4 there was no D wave because of the caudal tumor location. In addition, only the abductor hallucis muscle responses (not the tibialis anterior responses) could be elicited, indicating either some subclinical motor damage or simply the result of the patient’s young age (3 years). After coagulation of a tumor vessel, all muscle responses abruptly disappeared (A). After irrigation and slight hypertension, they reappeared on the right side (B), but not reliably on the left. The patient started to move his right leg after several hours, and his left leg on the sixth postoperative day.

Motor Evoked Potential Monitoring for Intramedullary Spinal Cord Tumor Surgery

89

8 SUMMARY The combined intraoperative monitoring of D waves and muscle MEPs elicited by transcranial electrical stimulation during intramedullary spinal cord tumor operations is based on a firm neurophysiological concept, and its use is practical and safe. Intraoperative MEP monitoring data correctly represent the clinical “reality” of the patient’s motor status: the presence of muscle MEPs always indicates intact motor function. Intraoperative loss of muscle MEPs indicates a temporary loss of motor function in the corresponding limb as long as the D wave amplitude remains above 50% of the baseline value. Further decline of the amplitude indicates permanent paraplegia or, in the case of a high cervical tumor, quadriplegia. The technique is readily implemented in a routine neurosurgical environment. It allows for the identification of impairment of the functional integrity of the motor pathways before a permanent deficit occurs. This knowledge has proven to be extremely valuable for intraoperative decision making. Included with the accompanying CD are two videos (choose Chapter 4 from the accompanying CD main menu). The first video shows a small clip of the CUSA being used to debulk a medullary tumor. The second video shows the loss of the right toe abductor MEPs during coagulation of blood vessels for hemostasis after tumor removal.

ACKNOWLEDGMENTS The author is indebted to San-San Chiang, Linda Velazquez, Ingrid KothbauerMargreiter, Matevzˇ Krzˇ an, Nobu Morota, Shlomo Constantini, George I. Jallo, Fred J. Epstein, and Vedran Deletis.

REFERENCES 1. Nash, C.L., Lorig, R.A., Schatzinger, L., and Brown, R.H. (1977). Spinal cord monitoring during operative treatment of the spine. Clin. Orthop., 126, 100–105. 2. Engler, G.L., Spielholz, N.I., Bernhard, W.N., Danziger, F., Merkin, H., and Wolff, T. (1978). Somatosensory evoked potentials during Harrington instrumentation for scoliosis. J. Bone Joint Surg., 60 A, 528–532. 3. Nuwer, M.R., Dawson, E.G., Carlson, L.G., Kanim, L.E., and Sherman, J.E. (1995). Somatosensory evoked potential spinal cord monitoring reduces neurologic deficits after scoliosis surgery: Results of a large multicenter study. Electroencephalogr. Clin. Neurophysiol., 96, 6–11. 4. Ginsburg, H.H., Shetter, A.G., and Raudzens, P.A. (1985). Postoperative paraplegia with preserved intraoperative somatosensory evoked potentials. J. Neurosurg., 63, 296–300.

90

Karl F. Kothbauer

5. Lesser, R.P., Raudzens, P., Lüders, H., Nuwer, M.R., Goldie, W.D., Morris, H.H., Dinner, D.S., Klem, G., Hahn, J.F., Shetter, A.G., Ginsburg, H.H., and Gurd, A.R. (1986). Postoperative neurological deficits may occur despite unchanged intraoperative somatosensory evoked potentials. Ann. Neurol., 19, 22–25. 6. Epstein, F.J., and Farmer, J.-P. (1990). Pediatric spinal cord tumor surgery. Neurosurg. Clin. N. Amer., 1, 569–590. 7. Merton, P.A., and Morton, H.B. (1980). Stimulation of the cerebral cortex in the intact human subject. Nature, 285, 227. 8. Levy, W.J., York, D.H., McCaffrey, M., and Tanzer, F. (1984). Motor evoked potentials from transcranial stimulation of the motor cortex in humans. Neurosurgery, 15, 287–302. 9. Patton, H.D., and Amassian, V.E. (1954). Single and multiple unit analysis of cortical stage of pyramidal tract activation. J. Neurophysiol., 17, 345–363. 10. Philips, C.G., and Porter, R. (1964). The pyramidal projection to motoneurones of some muscle groups of the baboon’s forelimb. In “Progress in brain research” ( J.C. Eccles, and J.P. Schadé, eds.), vol. 12, pp. 222–243. Elsevier, Amsterdam. 11. Boyd, S.G., Rothwell, J.C., Cowan, J.M.A., Webb, P.J., Morley, T., Asselman, P., and Marsden, C.D. (1986). A method of monitoring function in corticospinal pathways during scoliosis surgery with a note on motor conduction velocities. J. Neurol. Neurosurg. Psychiatry, 49, 251–257. 12. Katayama, Y., Tsubokawa, T., Maemjima, S., Hirayama, T., and Yamamoto, T. (1988). Corticospinal direct response in humans: Identification of the motor cortex during intracranial surgery under general anesthesia. J. Neurol. Neurosurg. Psychiatry, 51, 50–59. 13. Katayama, Y., Tsubokawa, T., Yamamoto, T., and Maejima, S. (1988). Spinal cord potentials to direct stimulation of the exposed motor cortex in humans: Comparison with data from transcranial motor cortex stimulation. In “Non-invasive stimulation of brain and spinal cord” (P.M. Rossini, and C.D. Marsden, eds.), vol. 41, pp. 305–311. Alan R. Liss, Inc., New York. 14. Burke, D., Hicks, R.G., and Stephen, J.P.H. (1990). Corticospinal volleys evoked by anodal and cathodal stimulation of the human motor cortex. J. Physiol., 425, 283–299. 15. Edmonds, H.L., Paloheimo, M.P.J., Backman, M.H., Johnson, J.R., Holt, R.T., and Shields, C.B. (1989). Transcranial magnetic motor evoked potentials (tcMMEP) for functional monitoring of motor pathways during scoliosis surgery. Spine, 14, 683–686. 16. Zentner, J. (1989). Noninvasive motor evoked potential monitoring during neurosurgical operations in the spinal cord. Neurosurgery, 24, 709–712. 17. Taniguchi, M., Schramm, J., and Cedzich, C. (1991). Recording of myogenic motor evoked potentials under general anesthesia. In “Intraoperative neurophysiologic monitoring in neurosurgery” ( J. Schramm, and Å.R. Møller, eds.), pp. 72–87. Springer, Berlin. 18. Taniguchi, M., Cedzich, C., and Schramm, J. (1993). Modification of cortical stimulation for motor evoked potentials under general anesthesia: Technical description. Neurosurgery, 32, 219–226. 19. Jones, S.J., Harrison, R., Koh, K.F., Mendoza, N., and Crockard, H.A. (1996). Motor evoked potential monitoring during spinal surgery: Responses of distal limb muscles to transcranial cortical stimulation with pulse trains. Electroencephalogr. Clin. Neurophysiol., 100, 375–383. 20. Pechstein, U., Cedzich, C., Nadstawek, J., and Schramm, J. (1996). Transcranial high-frequency repetitive electrical stimulation for recording myogenic motor evoked potentials with the patient under general anesthesia. Neurosurgery, 39, 335–344. 21. Rodi, Z., Deletis, V., Morota, N., and Vodusˇek, D.B. (1996). Motor evoked potentials during brain surgery. Pfluger’s Archiv—Euro. J. Physiol., 431, R291–292. 22. Calancie, B., Harris, W., Broton, J.G., Alexeeva, N., and Green, B.A. (1998). “Threshold-level” multipulse transcranial electrical stimulation of motor cortex for intraoperative monitoring of spinal motor tracts: Description of method and comparison to somatosensory evoked potential monitoring. J. Neurosurg., 88, 457–470.

Motor Evoked Potential Monitoring for Intramedullary Spinal Cord Tumor Surgery

91

23. Kothbauer, K., Deletis, V., and Epstein, F. (1998). Motor evoked potential monitoring for intramedullary spinal cord tumor surgery: Correlation of clinical and neurophysiological data in a series of 100 consecutive procedures. Neurosurg. Focus (electronic journal), (4), Article 1 (http://www.aans.org/journals/online_j/may98/4-5-1). 24. Deletis, V. (1993). Intraoperative monitoring of the functional integrity of the motor pathways. In “Electrical and magnetic stimulation of the brain and spinal cord” (O. Devinsky, A. Beric, and M. Dogali, eds.), pp. 201–214. Raven Press, New York. 25. Morota, N., Deletis, V., Constantini, S., Kofler, M., Cohen, H., and Epstein, F.J. (1997). The role of motor evoked potentials during surgery for intramedullary spinal cord tumors. Neurosurgery, 41, 1327–1336. 26. Kothbauer, K., Deletis, V., and Epstein, F.J. (1997). Intraoperative spinal cord monitoring for intramedullary surgery: An essential adjunct. Pediat. Neurosurg., 26, 247–254. 27. Kothbauer, K., Deletis, V., and Epstein, F.J. (1998). Motor evoked potential monitoring for spinal cord tumor surgery. J. Neurosurg., 88, 403A (Abstract). 28. Kernell, D., and Wu, C.-P. (1967). Post-synaptic effects of cortical stimulation on forelimb motoneurones in the baboon. J. Physiol., 191, 673–690. 29. Burke, D., Hicks, R., Stephen, J., Woodforth, I., and Crawford, M. (1992). Assessment of corticospinal and somatosensory conduction simultaneously during scoliosis surgery. Electroencephalogr. Clin. Neurophysiol., 85, 388–396. 30. Lang, E.W., Beutler, A.S., Chesnut, F.M., Patel, P.M., Kennelly, N.A., Kalkman, C.J., Drummond, J.C., and Garfin, S.R. (1996). Myogenic motor-evoked potential monitoring using partial neuromuscular blockade in surgery of the spine. Spine, 21, 1676–1686. 31. Deletis, V., and Kothbauer, K. (1998). Intraoperative neurophysiology of the corticospinal tract. In “Spinal cord monitoring” (E. Stålberg, H. S. Sharma, and Y. Olsson, eds.), pp. 421–444. Springer, Vienna. 32. Jasper, H.H. (1957). The ten twenty electrode system of the international federation. Electroencephalogr. Clin. Neurophysiol., 10, 371–375. 33. Deletis, V., Rodi, Z., and Amassian, V.E. (2001). Neurophysiological mechanisms underlying motor evoked potentials in anesthetized humans: Part 2. Relationship between epidurally and muscle recorded MEPs in man. Clin. Neurophysiol., 112, 445–452. 34. Deletis, V., Isgum, V., and Amassian, V.E. (2001). Neurophysiological mechanisms underlying motor evoked potentials in anesthetized humans: Part 1. Recovery time of corticospinal tract direct waves elicited by pairs of transcranial electrical stimuli. Clin. Neurophysiol., 112, 438–444. 35. Jankowska, E., Padel, E., and Tanaka, R. (1975). Projections of pyramidal tract to cells αmotoneurones innervating hind-limb muscles in the monkey. J. Physiol., 249, 637–667. 36. Burke, D., Hicks, R., Stephen, J., Woodforth, I., and Crawford, M. (1995). Trial-to-trial variability of corticospinal volleys in human subjects. Electroencephalogr. Clin. Neurophysiol., 97, 231–237. 37. Woodforth, I.J., Hicks, R.G., Crawford, M.R., Stephen, J.P., and Burke, D.J. (1996). Variability of motor-evoked potentials recorded during nitrous oxide anesthesia from the tibialis anterior muscle after transcranial electrical stimulation. Anesth. Analg., 82, 744–749. 38. Jellinek, D., Jewkes, D., and Symon, L. (1991). Noninvasive intraoperative monitoring of motor evoked potentials under propofol anesthesia: Effect of spinal surgery on the amplitude and latency of motor evoked potentials. Neurosurgery, 29, 551–557. 39. Kalkman, C.J., Drummond, J.C., Ribberink, A.A., Patel, P.M., Sano, T., and Bickford, R.G. (1992). Effects of propofol, etomidate, midazolam and fentanyl on motor evoked responses to transcranial electrical or magnetic stimulation in humans. Anesthesiology, 76, 502–509. 40. Schmid, U.D., Boll, J., Liechti, S., Schmid, J., and Hess, C.W. (1992). Influence of some anesthetic agents on muscle responses to transcranial magnetic cortex stimulation: A pilot study in man. Neurosurgery, 30, 85–92.

92

Karl F. Kothbauer

41. Taniguchi, M., Nadstawek, J., Langenbach, U., Bremer, F., and Schramm, J. (1993). Effects of four intravenous anesthetic agents on motor evoked potentials elicited by magnetic transcranial stimulation. Neurosurgery, 33, 407–415. 42. Fennelly, M.E., Taylor, B.A., and Hetreed, M. (1993). Anaesthesia and the motor evoked potential. In “Handbook of spinal cord monitoring” (S.J. Jones, S. Boyd, M. Hetreed, and N.J. Smith, eds.), vol. 1, pp. 272–276. Kluwer Academic Publishers, Dordrecht. 43. Agnew, W.F., and McCreery, D.B. (1987). Considerations for safety in the use of extracranial stimulation for motor evoked potentials. Neurosurgery, 20, 143–147. 44. Engel, J. (1989). “Seizures and epilepsy.” F. A. Davis Co., Philadelphia. 45. Barker, A.T., Freeston, I.L., Jalinous, R., and Jarratt, J.A. (1988). Magnetic and electrical stimulation of the brain: Safety aspects. In “Non-invasive stimulation of brain and spinal cord” (P.M. Rossini, and C.D. Marsden, eds.), vol. 41, pp. 131–144. Alan R. Liss, Inc., New York. 46. McCormick, P.C., Torres, R., Post, K.D., and Stein, B.M. (1990). Intramedullary ependymoma of the spinal cord. J. Neurosurg., 72, 523–532. 47. Nathan, P.W., and Smith, M.C. (1973). Effects of two unilateral cordotomies on the motility of the lower limbs. Brain, 96, 471–494. 48. Epstein, F. J., Farmer, J.-P., and Freed, D. (1993). Adult intramedullary spinal cord ependymoma: The result of surgery in 38 patients. J. Neurosurg., 79, 204–209. 49. Albright, A.L. (1998). Intraoperative spinal cord monitoring for intramedullary surgery: An essential adjunct? Pediatr. Neurosurg., 29, 112. 50. Kothbauer, K.F., Deletis, V., and Epstein, F.J. (1998). Reply. Pediatr. Neurosurg., 29, 54–55. 51. Greenwood, J. (1967). Surgical removal of intramedullary tumors. J. Neurosurg., 26, 276–282.

CHAPTER

5

Selective Spinal Cord Lesioning Procedures for Spasticity and Pain MARC SINDOU AND PATRICK MERTENS Department of Neurosurgery, Hopital Neurologique Pierre Wertheimer, University of Lyon, Lyon, France

Part I: Selective Lesioning Procedures in Spinal Roots and Spinal Cord for Treatment of Spasticity 1 Procedures 1.1 Posterior Rhizotomies 1.2 Results of Posterior Rhizotomies 1.3 Longitudinal Myelotomy 1.4 Surgery in the Dorsal Root Entry Zone (DREZ) 2 Indications 2.1 Indications for Surgery in Adults 2.2 Indications for Surgery in Children with Cerebral Palsy 3 Conclusion Part II: Selective Spinal Cord Lesioning Procedures for Treatment of Pain: DREZ Lesions 1 Microsurgical DREZotomy 1.1 Operative Procedure at the Cervical Level 1.2 Operative Procedure at the Lumbosacral Level 1.3 Neurophysiological Monitoring as an Aid to Surgery 1.4 Microelectrophysiology and Microdialysis Studies in the Dorsal Horn During Surgery 2 Radiofrequency (RF) Thermocoagulation Procedure 3 DREZ Procedures with the Laser Beam 4 Ultrasonic DREZ Procedure 5 Indications for DREZ References Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

93

94

Marc Sindou and Patrick Mertens

Part I: Selective Lesioning Procedures in Spinal Roots and Spinal Cord for Treatment of Spasticity ABSTRACT Spasticity is one of the most common sequels of neurologic diseases. In most patients, spasticity is useful in compensating for lost motor strength. However, in a significant number of patients it may become harmful and lead to further functional losses. When not controllable by physical therapy and medications, excessive spasticity can benefit from neurostimulation, intrathecal Baclofen pharmacotherapy, botulinum toxin injections, or selective ablative surgical procedures. Lesioning can be performed at the level of the peripheral nerves, spinal roots, spinal cord, or the dorsal root entry zone (DREZ). In this chapter, only selective procedures in the spinal roots, spinal cord, and DREZ will be described.

1 PROCEDURES

1.1 POSTERIOR RHIZOTOMIES After Sherrington demonstrated in 1898 that decerebrate rigidity in an animal model was abolished by section of the dorsal roots (that is, by interruption of the afferent input to the monosynaptic stretch and polysynaptic withdrawal reflexes), posterior rhizotomy for the modification of spasticity was first performed by Foerster in 1908 [1]. Its undesired effects on sensory and sphincter function have limited its application in the past. To minimize these disadvantages, several authors in the 1960s and 1970s attempted to develop more selective operations, especially for the treatment of children with cerebral palsy. 1.1.1 Posterior Selective Rhizotomy To reduce the sensory side-effects of the original Foerster method, Gros and coworkers [2] introduced a technical modification that consisted of sparing one rootlet of the five of each root, from L1 to S1. On similar principles, Ouaknine [3], a pupil of Gros, developed a microsurgical technique that consisted of resecting one third to two thirds of each group of rootlets of all the posterior roots from L1 to S1. 1.1.2 Sectorial Posterior Rhizotomy In an attempt to reduce the side-effects of rhizotomy on postural tone in ambulatory patients, Gros [4] and his pupils Privat [5] and Frerebeau [6] proposed a topographic selection of the rootlets to be sectioned. First, a preoperative

Selective Spinal Cord Lesioning Procedures for Pain and Spasticity

95

assessment of spasticity useful for postural tone (abdominal muscles, quadriceps, gluteus medius) and spasticity harmful to the patient (hip flexors, adductors, hamstrings, and triceps surae) is conducted. Mapping of the evoked motor activity of the exposed rootlets, from L1 to S2, by direct electrostimulation of each posterior group of rootlets is then carried out, and the rootlets to be sectioned are determined according to the preoperative program. 1.1.3 Partial Posterior Rhizotomy Fraioli and Guidetti [7] reported on a procedure by which the dorsal half of each rootlet of the selected posterior roots is divided a few millimeters before its entrance into the posterolateral sulcus. The authors report good results without significant sensory deficit, the latter being explained by the fact that partial section leaves intact a large number of fibers of all types. 1.1.4 Functional Posterior Rhizotomy The search for specially organized circuits responsible for spasticity led Fasano and associates [8] to propose a new method called functional posterior rhizotomy. This method is based on bipolar intraoperative stimulation of the posterior rootlets and analysis of different types of EMG reflex responses. Responses characterized by a permanent tonic contraction, an after-discharge pattern, or a large spatial diffusion to distant muscle groups were considered to belong to disinhibited spinal circuits responsible for spasticity. Functional posterior rhizotomy—which was especially conceived for children with cerebral palsy— has also been used by other outstanding surgical teams, each one having brought its own technical modifications to the method [9–12]. Our personal adaptation of these methods is illustrated in Fig. 5.1.

1.2 RESULTS OF POSTERIOR RHIZOTOMIES The results of posterior rhizotomies in children with cerebral palsy—whatever the technical modality may be—have been recently reported in several publications. We have reviewed and quoted them (46 references) on the occasion of the report of our own series [13]. Briefly, these publications show that about 75% of the patients had nearly normal muscle tone at 1 year or more after surgery without spasticity limiting the residual voluntary movements of the limbs. After a serious and persisting physical therapy and rehabilitation program, most children demonstrated improved stability in sitting and/or increased efficiency in walking. It must be noted, however, that preexisting orthopedic deformities cannot be improved with this method [13].

96 FIGURE 5.1 Lumbosacral posterior rhizotomy for children with cerebral palsy. Our personal technique consists of performing a limited osteoplastic laminotomy using a power saw, in one single piece, from T11 to L1 (left). The laminae will be replaced at the end of the procedure and fixed with wires (right). The dorsal (and ventral) L1, L2, and L3 roots are identified by means of the muscular responses evoked by electrical stimulation performed intradurally just before entry into their dural sheaths. The dorsal sacral rootlets are recognized at their entrance into the dorsolateral sulcus of the conus medullaris. The landmark between S1 and S2 medullary segments is located approximately 30 mm from the exit of the tiny coccygeal root from the conus. The dorsal rootlets of S1, L5, and L4 are identified by their evoked motor responses. The sensory roots for bladder (S2–S3) can be identified by monitoring vesical pressure. Those for the anal sphincter (S3–S4) can be identified by rectomanometry (or simply using finger introduced into the patient’s rectum) or EMG recordings. Surface spinal cord SEP recordings from tibial nerve (L5–S1) and pudendal nerve (S1–S3) stimulation may also be helpful. For the surgery to be effective, a total amount of 60% of dorsal rootlets must be cut (with a different amount of rootlets cut according to the level and function of the roots involved). Also, the correspondence of the roots with the muscles having harmful spasticity or useful postural tone must be considered in determining the amount of rootlets to be cut; in most cases, L4 (which predominantly gives innervation to the quadriceps femoris) has to be preserved.

Selective Spinal Cord Lesioning Procedures for Pain and Spasticity

97

1.3 LONGITUDINAL MYELOTOMY Longitudinal myelotomy, which was introduced by Bischof [14], was made more selective by Pourpre [15] and later on by Laitinen [16]. The method consists of a frontal separation between the posterior and anterior horns of the lumbosacral enlargement from T11 to S2 performed from inside the spinal cord after a posterior commisural incision that reaches the ependymal canal. In Laitinen’s series of 25 patients, 60% had complete relief of spasticity and 36% showed some residual spasticity in one or both legs. Within 1 year, some muscular tone returned in most patients but seldom produced troublesome spasticity. A harmful effect on bladder function was present in 27% of the patients. Longitudinal myelotomy is indicated only for spastic paraplegias with flexion spasms, when the patient has no residual useful motor control and no bladder and sexual function.

1.4 SURGERY IN THE DORSAL ROOT ENTRY ZONE (DREZ) Selective posterior rhizotomy in the dorsal root entry zone (DREZ), referred to as micro-DREZotomy (MDT), was introduced in 1972 [17] to treat intractable pain. Because of its inhibitory effects on muscular tone, it has been applied to patients with focalized hyperspasticity [18–21]. This method attempts to selectively interrupt the small nociceptive and the large myotatic fibers (situated laterally and centrally, respectively), while sparing the large lemniscal fibers which are regrouped medially. It also enhances the inhibitory mechanisms of Lissauer’s tract and the dorsal horn [22] (Fig. 5.2 left). MDT, the technique of which has been described elsewhere [23–25], consists of microsurgical incisions that are 2 to 3 mm deep and at a 35° angle for the cervical level (Fig. 5.2 right) and at a 45° angle for the lumbosacral level (Fig. 5.3), followed by bipolar coagulations performed ventrolaterally at the entrance of the rootlets into the dorsolateral sulcus, along all the cord segments selected for operation. For patients with paraplegia [24], the L2–S5 segments are approached through a T11-L2 laminectomy, whereas for the hemiplegic upper limb [25], a C4–C7 hemilaminectomy with conservation of the spinous processes is sufficient to reach the C5–T1 segments. Identification of the cord levels related to the undesirable spastic mechanisms is achieved by studying the muscle responses to bipolar electrical stimulation of the anterior and/or posterior roots. The motor threshold for stimulation of anterior roots is one third that of the threshold for posterior roots. Then, the lateral aspect of the DREZ is exposed so that the microsurgical lesioning can be performed. Lesions are 2 to 3 mm in depth and are placed at 35 to 45° angles in the

98

Marc Sindou and Patrick Mertens

FIGURE 5.2 Micro-DREZotomy (MDT). Left, organization of fibers at the DREZ in humans. The large arrow shows the proposed extent of the MDT affecting the lateral and central bundles formed by the nociceptive and myotatic fibers, as well as the excitatory medial part of the Lissauer Tract and the upper layers of the dorsal horn. Right, principle behind the MDT technique. Example of the MDT at the cervical level through a right cervical hemilaminectomy (the procedure for the lumbosacral roots is the same). The right C6 posterior root has been retracted toward the inside to make the ventrolateral region of the DREZ accessible. The incision is performed into the dorsolateral sulcus using a small piece of razor blade (upper operative view). The incision is 2 to 3 mm deep and is made at a 35° angle (at a 45° angle for the lumbosacral level). Then microcoagulations are created with a very sharp and graduated bipolar microforceps down to the apex of the dorsal horn (lower operative view).

99

FIGURE 5.3 MDT technique at the lumbosacral level. Top left, exposure of the conus medullaris through a T11–L1 laminectomy. Bottom left, approach of the left dorsolateral sulcus. For this approach, the rootlets of the selected lumbosacral dorsal roots are displaced dorsally and medially to obtain proper access to the ventrolateral aspect of the DREZ. Right, the rootlets of the selected dorsal roots are retracted dorsomedially. They are subsequently held with a specially designed ball-tip microsucker, used as a small hook to gain access to the ventrolateral part of the DREZ. After the fine arachnoidal filaments sticking the rootlets together with the pia mater are divided with curved sharp microscissors (B), the main arteries running along the dorsolateral sulcus are dissected and preserved, while the smaller ones are coagulated with a sharp bipolar microforceps (F). Then, a continuous incision is performed using a microknife (K) made with a small piece of razor blade inserted within the striated jaws of a curved razor blade holder (K). The cut is—on average—at a 45° angle and to a depth of 2 mm. The surgical lesion is completed by doing microcoagulations under direct magnified vision, at a low intensity, inside the posterolateral sulcomyelotomy down to the apex of the dorsal horn. These microcoagulations are made by means of the special sharp bipolar forceps (F), insulated except for 5 mm at the tips and graduated every millimeter.

100

Marc Sindou and Patrick Mertens

ventrolateral aspect of the sulcus all along the selected segments of the spinal cord. Intraoperative neurophysiological monitoring may be of some help for identifying cord levels, quantifying the extent of MDT, and avoiding impairing long fiber tracts. MDT is indicated in paraplegic patients, especially when they are bedridden as a result of disabling flexion spasms, and in hemiplegic patients with irreducible and/or painful hyperspasticity in the upper limb [23–27]. MDT also can be used to treat neurogenic bladder with uninhibited detrusor contractions resulting in leakage of urine around a catheter [26]. To date, our series has consisted of 45 patients with unilateral cervical (C5–T1) MDT for harmful spasticity in the upper limb, 121 patients with bilateral lumbosacral MDT (L2–S1 or S5) for disabling spasticity in the lower limbs, and 12 patients with bilateral sacral S2–S3 (S4) MDT for hyperactive neurogenic bladder only. Effects on muscular tone can be judged only after a 3-month follow-up. A “useful” effect on spasticity, allowing withdrawal of antispasmodic medications, was obtained in 78% of the patients with a spastic upper limb. A similarly useful effect was obtained in 75% of the patients with spasticity in the lower limbs. When spasms were present in paraplegic patients, they were suppressed or markedly decreased in 88% of the patients. When compared to patients with multiple sclerosis (75% with good results), the results were better in patients with spasticity (and spasms) caused by pure spinal cord lesions (80% with good results). The least improvement was observed in patients with spasticity resulting from cerebral lesions (60% with good results). Reduction in spasticity usually leads to a significant improvement in abnormal postures and articular limitations. This was achieved in about 90% of our patients. For the hemiplegic upper limb, the increase in articular amplitude was most remarkable for the elbow and shoulder (when not “frozen”) and much more limited for the wrist and fingers, especially if there was retraction of the flexor muscles and no residual voluntary motor activity in the extensors. For the lower limb(s), with abnormal postures in flexion, the increase in amplitude of joint movement was very much dependent on the degree of the preoperative retractions. When the post-MDT gains were deemed insufficient because of persistent joint limitations, complementary orthopedic surgery was indicated. With regard to the patients (n = 5) who had paraplegia with irreducible hyperextension, all were completely relieved. In the patients with some voluntary movements hidden behind spasticity, reduction in the hypertonia resulted in an improvement in voluntary motor activity. Fifty percent of the patients operated on for spasticity in the upper limb had better motor activity of the shoulder and arm, but only half of those with some preoperative distal motor function obtained additional hand prehension. Only 10% of the patients with spasticity in the lower limb(s) had significant motor improvement after surgery (because most patients in this group had no preoperative motor function). In these

Selective Spinal Cord Lesioning Procedures for Pain and Spasticity

101

severely affected patients, the main benefit was better comfort, less pain, ability to resume physical therapy, and less dependence in daily life (see [27] for pre- and postoperative assessment of patients, with details on the functional scores used). Bladder capacity was significantly improved in 85% of the 38 patients who had a hyperactive neurogenic bladder with urine leakage around the catheter. The 32 patients who improved were those in whom the detrusor was not irreversibly fibrotic. Pain, when present, was in general favorably influenced. MDT continually produced a marked decrease in sensation. Because most patients were in a precarious general and neurological state, death occurred in 5 patients (4%), resulting from respiratory problems in 4 and bed sores in 1. Two patients with multiple sclerosis (MS) presented with acute but transient increases in their preexisting neurological symptoms during the postoperative period. Two others had a new postoperative clinical manifestation of the disease. The last of the complications we have to mention concerns a patient who was operated on at the cervical level and had a persistent motor deficit in the ipsilateral leg after surgery. With rigorous selection of patients, MDT can be very effective in relieving pain and suppressing excessive spasticity. Good long-lasting relief of excess spasticity was achieved in 80% of our patients. As a result, MDT, sometimes combined with complementary orthopedic surgery, resulted in significant improvement in patient comfort and joint deformities, and even enhancement of residual voluntary motility hidden preoperatively behind hypertonicity.

2 INDICATIONS

2.1 INDICATIONS FOR SURGERY IN ADULTS 2.1.1 Spinal Cord Stimulation Provided that the spasticity is mild and the dorsal columns are still functioning, spinal cord stimulation can be useful for treating spasticity from diseases affecting the spinal cord (e.g., MS or degenerative diseases such as Strumpell-Lorrain syndrome). A percutaneous trial before a definitive implantation may be useful. 2.1.2 Intrathecal Baclofen Intrathecal baclofen administration is indicated for para- or tetraplegic patients with severe and diffuse spasticity, especially when spasticity has a spinal origin. Because of its reversibility, this method should be used before an ablative procedure is considered. But the range between excessive hypotonia with loss of strength and an insufficient effect is very narrow. An intrathecal test through a temporary access port can be useful when deciding if permanent implantation is indicated.

102

FIGURE 5.4

Marc Sindou and Patrick Mertens

Decision-making for hyperspasticity in adults.

2.1.3 Neuroablative Techniques Neuroablative techniques are indicated for severe focalized spasticity in the limbs of paraplegic, tetraplegic, or hemiplegic patients. Neurotomies are preferred when spasticity is localized to muscle groups innervated by a small number of, or a single, peripheral nerve (or nerves). When spasticity affects an entire limb, MDT is preferred. Several types of neuroablative procedures can be combined in the treatment of one patient, if needed. Whatever the situation and the etiology may be, orthopedic surgery should be considered only after spasticity has been reduced by physical and pharmacological treatments first and, when necessary, by neurosurgical procedures. Guidelines for surgical indications have been detailed elsewhere [28, 29] and are summarized in Fig. 5.4. The general rule is to tailor individual treatments as much as possible to the patient’s particular problems.

2.2 INDICATIONS FOR SURGERY IN CHILDREN WITH CEREBRAL PALSY Surgical indications depend on preoperative abilities, disabilities, and the eventual functional goals. As a means of guidance, we have adopted the six-group classification as defined by Abbott [30].

Selective Spinal Cord Lesioning Procedures for Pain and Spasticity

103

2.2.1 Independent Ambulatory Patients In independent ambulatory patients, the goal is to improve efficiency and cosmetics in walking by eliminating as many abnormally responsive neural circuits as can be identified through functional posterior rhizotomy. Surgery is best performed as soon as possible after the child has demonstrated the ability to work with a therapist, usually between ages 3 and 7 years, and frequently must be done in conjunction with operations on tendons because of concomitant shortened muscles. 2.2.2 Ambulatory Patients Dependent on Assistance Devices For ambulatory patients dependent on assistance devices (canes, crutches, rollators, walkers), the goal is to lessen that dependence. A child with poor trunk control or lack of protective reaction but with good underlying strength in the antigravity muscles can safely undergo a functional posterior rhizotomy. In children dependent on hypertonicity in the quadriceps to bear weight, a limited sectorial rhizotomy is preferable. For children who are in the process of developing ambulatory skill and need a temporary assistance device, it is important to delay surgery until they have perfected these skills. 2.2.3 Quadruped Crawlers For quadruped crawlers (or bunny hoppers) the goal is to achieve assisted ambulation during mid-childhood to early adolescence. A functional posterior rhizotomy will decrease hypertonicity in the leg musculature and allow better limb alignment in the standing position for a child with adequate muscular strength. However, a child who exhibits quadriceps weakness can be considered for a sectorial posterior rhizotomy. Children in this group can present at a young age with progressive hip dislocation. The goal is to stop the progressive orthopedic deformity by using obturator neurotomy with adductor tenotomies or functional posterior rhizotomy. 2.2.4 Commando (or Belly) Crawlers For commando (or belly) crawlers disabled by severe deficiencies in the postural control, the goal of posterior rhizotomy is only to improve functioning in the sitting position by increasing stability. 2.2.5 Totally Dependent Children In totally dependent children with no locomotive abilities, the goals are to simply improve comfort and facilitate care. As with group 4 (commando [or belly] crawlers), the preferred treatment is posterior rhizotomy, but there is also a need for exploring the efficacy of intrathecal baclofen.

104

Marc Sindou and Patrick Mertens

2.2.6 Children with Asymmetrical Spasticity For asymmetrical spasticity, selective peripheral neurotomies must be considered, especially obturator and tibial for a spastic hip or foot, respectively. For upper limb spasticity, the MDT procedure and/or selective neurotomies of the flexor muscles of wrist and fingers can be considered.

3 CONCLUSION Spasticity is usually a useful substitute for deficiencies in motor strength. Therefore, it must be preserved. Although it happens infrequently, it can lead to the harmful aggravation of a motor disability. When excessive spasticity is not sufficiently controlled by physical therapy and pharmacological agents, patients can consider surgery, especially neurosurgical procedures. By suppressing excessive spasticity, correcting abnormal postures, and relieving the frequently associated pain, surgery for spasticity allows physiotherapy to be resumed and sometimes results in the reappearance of, or improvement in, useful voluntary motility. When dealing with these patients, the surgeon must know the risks of the available treatments. To minimize those risks, the surgeon needs a strong anatomic, physiological, and chemical background, rigorous methods to assess and quantify the disorders, and the ability to work in a multidisciplinary team [29].

Part II: Selective Spinal Cord Lesioning Procedures for Treatment of Pain: DREZ Lesions ABSTRACT In the 1960s, a large number of neurophysiologic investigations proved that the dorsal horn was the primary level of modulation of pain sensation. This idea was popularized in 1965 through the gate control theory [31], which drew neurosurgeons’ attention to this area as a possible target for pain surgery. Neurostimulation of the primary afferent neurons was developed to enhance the inhibitory mechanisms of the spinal cord [32]. Conversely, in 1972 we undertook anatomical studies and preliminary surgical trials in the human dorsal root entry zone (DREZ) to determine whether a destructive procedure at this level was feasible [33, 34]. Soon after, in 1974, Nashold and his group started to develop DREZ lesions using the RFthermocoagulation as the lesion maker in the substantia gelatinosa of the dorsal horn [35] and later in the whole DREZ [36]. This has been performed especially for pain caused by brachial plexus avulsion. Later on, DREZ procedures were performed by using a laser [37, 38] and an ultrasound probe [39, 40] for pain caused by brachial plexus avulsion.

Selective Spinal Cord Lesioning Procedures for Pain and Spasticity

105

1 MICROSURGICAL DREZotomy This procedure consists of a longitudinal incision of the dorsolateral sulcus ventrolaterally at the entrance of the rootlets into the sulcus. Microbipolar coagulations are performed continuously inside the sulcus down to the apex of the dorsal horn and along all the spinal cord segments selected for surgery. The lesion, which penetrates the lateral part of the DREZ and the medial part of the tract of Lissauer (TL), extends down to the apex of the dorsal horn, which can be recognized by its brown-gray color. The average lesion is 2 to 3 mm deep and is made at a 35° angle medially and ventrally. The procedure is presumed to preferentially destroy the nociceptive fibers grouped in the lateral bundle of the dorsal rootlets, as well as the excitatory medial part of the TL. The upper layers of the dorsal horn are also destroyed if microbipolar coagulations are made inside the dorsal horn . The upper layers of the dorsal horn are known to be the site of “hyperactive” neurons, especially in the cases with peripheral deafferentation (Fig. 5.5). The procedure is presumed to at least partially preserve the inhibitory structures of the DREZ, (i.e., the lemniscal fibers reaching the dorsal column, as well as their recurrent collaterals to the dorsal horn and the substantia gelatinosa [SG] propriospinal interconnecting fibers running through the lateral part of the TL). This MDT was

FIGURE 5.5 Dorsal horn microelectrode recordings in man. The electrode was a floating tungsten microelectrode that was implanted intraoperatively free-hand under the operative microscope; it reached 5 mm in depth (in laminae IV–VI). The vertical bars are 50 µV, and the horizontal bars are 100 ms. Upper trace, normal activity. Recordings in a nondeafferented dorsal horn (spastic patient). Left, almost no spontaneous activity (3 spikes at random). Middle, spike burst discharges (arrows) evoked by regular light tactile stimulation of the corresponding dermatoma. Right, electrical stimulation of the corresponding peripheral nerve. Lower trace, deafferentation hyperactivity. Recordings in the L5 cord segment of a patient with pain caused by a traumatic section of the hemi-cauda equina from root L4 to S4. Left, spontaneous activity of the recorded unit: continuous, regular, high-frequency discharge. Middle, unit during light tactile stimulation of the L4–S1 dermatome (arrow). Right, during electrical stimulation of the tibial nerve (the arrows are two consecutive stimuli). Note the continuous regular discharges, which remain unaltered.

106

Marc Sindou and Patrick Mertens

conceived in order to prevent the complete abolition of tactile and proprioceptive sensations and to avoid deafferentation phenomena [41]. Working in the DREZ requires knowledge of the morphological anatomy of the dorsal roots corresponding to the spinal level. Details have been given in previous publications [42–45]. The axis of the dorsal horn in relation to the sagittal plan crossing the dorsolateral sulcus will condition the angulation of the DREZotomy. According to 82 measurements performed by Young (personal communication, 1991), the mean DREZ angle is 30° at C6, 26° at T4, 37° at T12, and 36° at L3. The site and extent of the DREZ lesion will also be determined by the shape, width, and depth of the TL and dorsal horn (Fig. 5.6). Surgery is performed with the patient under general anesthesia, but with only an initial short-lasting muscle relaxant to allow intraoperative observation of motor responses to bipolar electrical stimulation of the nerve roots. Stimulated ventral roots have a motor threshold at least three times lower than the dorsal roots. Standard microsurgical techniques are used with 10× to 25× magnification.

1.1 OPERATIVE PROCEDURE AT THE CERVICAL LEVEL The prone position with the head and neck flexed in the “concorde” position has the advantage of avoiding brain collapse caused by cerebral spinal fluid (CSF) depletion. The head is fixed with a three-pin head holder. The level of laminectomy is determined after identification of the prominent spinous process of C2 by palpation. A hemilaminectomy, generally from C4 to C7, with preservation of the spinous processes, allows sufficient exposure to the posterolateral aspect of the cervical spinal cord segments that correspond to the upper limb innervation, that is, the rootlets of C5 to T1 (T2). After the dura and the arachnoid are opened longitudinally, the exposed roots and rootlets are dissected free by separation of the tiny arachnoid filaments that bind them to each other, to the arachnoid sheath, and to the spinal cord pia mater. The radicular vessels are preserved. Each ventral and dorsal root from C4 to T1 is electrically stimulated at the level of its corresponding foramen to precisely identify its muscular innervation and its functional value. Responses are in the diaphragm for C4 (the response is palpable below the lower ribs), in the shoulder abductors for C5, in the elbow flexors for C6, in the elbow and wrist extensors for C7, and in the muscles of the hand for C8 and T1. Microsurgical lesions are performed at selected levels that correspond to the pain territory. The technique is summarized and illustrated in Fig. 5.2 of the

Selective Spinal Cord Lesioning Procedures for Pain and Spasticity

107

FIGURE 5.6 Variations of shape, width, and depth of the DREZ area, according to the spinal cord level (from top to bottom: cervical n° 7, thoracic n° 5, lumbar n° 4, sacral n° 3). Note how, at the thoracic level, Lissauer’s tract is narrow and the dorsal horn deep. Therefore, it is easy to understand that DREZ lesions at this level can be dangerous for the corticospinal tract and the dorsal column.

previous chapter. The incision is made with a microknife. Then microcoagulations are made in a “chain” (i.e., dotted) manner. Each microcoagulation is performed by short-duration (a few seconds), low-intensity, bipolar electrocoagulation with a special sharp bipolar forceps. The depth and extent of the lesion

108

Marc Sindou and Patrick Mertens

depend on the degree of the desired therapeutic effect and the preoperative sensory status of the patient. If the laxity of the root is sufficient, the sulcotomy is accomplished through an incision performed continuously in the dorsolateral sulcus, ventrolaterally along all of the rootlets of the targeted root. If this is not the case, a partial ventrolateral section is made successively on each rootlet of the root after the surgeon has isolated each one by separation of the tiny arachnoid membranes that hold them together. For pain due to brachial plexus avulsion, dotted microcoagulations inside the dorsal horn (at least 3 mm in depth from the surface of the cord) are performed after incision of the dorsolateral sulcus. Sharp graduated bipolar forceps are used to make the microcoagulations at the level of the avulsed roots. Selective ventrolateral DREZ lesions are extended to the root remaining above and below. In brachial plexus avulsion, dissection of the spinal cord is sometimes difficult to achieve safely because of scar tissue adhering to the cord. Atrophy and/or gliotic changes at the level of the avulsed roots can make identification of the dorsolateral sulcus hazardous. In such cases, it is necessary to start from the roots remaining above and below. The presence of tiny radicular vessels that enter the cord may help determine the site of the sulcus. Yellow areas corresponding to old hemorrhages on the surface of the cord and/or microcavities in the depth of the sulcus and the dorsal horn provide some guidance for tracing the sulcomyelotomy. When the dorsolateral sulcus is difficult to find, intraoperative monitoring of the dorsal column somatosensory evoked potentials (SEPs) evoked by stimulation of the tibial nerve is especially helpful.

1.2 OPERATIVE PROCEDURE AT THE LUMBOSACRAL LEVEL The patient is positioned prone on thoracic and iliac supports, and the head is placed 20 cm lower than the level of the surgical wound to minimize the intraoperative loss of CSF. The desired vertebral level is identified by palpation of the spinous processes or, if this is difficult, by a lateral x-ray study that includes the S1 vertebra. Interspinous levels identified by a needle can then be marked with methylene blue. A laminectomy—either bilateral or unilateral, according to pain topography—from T11 to L1 (or L2) is performed. The dura and arachnoid are opened longitudinally and the filum terminale is isolated. Roots are then identified by electrical stimulation. The L1 and L2 roots are easily identified at their penetration into their respective dural sheaths. Stimulation of L2 produces a response of the iliopsoas and adductor muscles.

Selective Spinal Cord Lesioning Procedures for Pain and Spasticity

109

Identification of L3 to L5 is difficult for many reasons: (1) the exit through their respective dural sheaths is caudal to the exposure; (2) the dorsal rootlets enter the DREZ along an uninterrupted line; (3) the ventral roots are hidden in front of the dentate ligament; and (4) the motor responses in the leg to stimulation of the roots are difficult to observe intraoperatively because of the patient’s prone position. Stimulation of L3 produces a preferential response in the adductors and quadriceps, of L4 in quadriceps, and of L5 in the tibialis anterior muscle. Stimulation of the S1 dorsal root produces a motor response of the gastrocnemius-soleus group that can be confirmed later by repeatedly checking the Achilles ankle reflex before, during, and after MDT at this level. Stimulation of the S2–S4 dorsal roots (or better, the corresponding spinal cord segments directly) can be assessed by recording the motor vesical or anal response by use of cystomanometry, rectomanometry, or electromyography of the anal sphincter (or by inserting a finger into the rectum). Because neurophysiological investigations are time-consuming to perform in the operating room, we have found that measurements at the conus medullaris can be sufficient in patients who already have severe preoperative impairment of their vesicoanal functions. These measurements, based on human postmortem anatomical studies, have shown that the landmark between the S1 and S2 segments is situated around 30 mm above the exit from the conus of the tiny coccygeal root. Microsurgical DREZotomy at the lumbosacral levels follows the same principles as at the cervical level. The technique is summarized and illustrated in Fig. 5.3 of the previous chapter. At the lumbosacral level, MDT is difficult and possibly dangerous because of the rich vasculature of the conus. The posterolateral spinal artery courses along the posterolateral sulcus. Its diameter is between 0.1 and 0.5 mm, and it is fed by the posterior radicular arteries. It joins caudally with the descending anterior branch of the Adamkiewicz artery through the conus medullaris anastomotic loop of Lazorthes. If it is freed from the sulcus, this artery can be preserved.

1.3 NEUROPHYSIOLOGICAL MONITORING AS AN AID TO SURGERY [46–48] Intraoperative monitoring of SEPs can be performed at the surface of the exposed spinal cord. Recordings of presynaptic potentials from the dorsal root and postsynaptic potentials from the dorsal horn can be useful for identification of the spinal cord segments. Potentials have a maximal amplitude in C6–C7 stimulation of the median nerve and the C8 segment for stimulation of the

110

Marc Sindou and Patrick Mertens

FIGURE 5.7 Effects of MDT on the evoked electrospinogram (EESG). Recordings from the surface of the dorsal column, medially to the DREZ at the C7 cervical (Ce) and the L5 lumbosacral (LS) segments, ipsilateral to the stimulation of the median and the tibial nerve, respectively, before (A) and after (B) MDT. The initial positive event P9 (for cervical) (P17 for lumbosacral) corresponds to the far-field compound potential originating in the proximal part of the brachial (lumbosacral) plexus. The small and sharp negative peaks N11 (N21) correspond to near-field presynaptic successive axonal events, probably generated in the proximal portion of the dorsal root, the dorsal funiculus, and the large-diameter afferent collaterals to the dorsal horn. After MDT, all of these presynaptic potentials remain unchanged. The larger slow negative wave N13 (N24) corresponds to the postsynaptic activation of the dorsal horn by group I and II peripheral afferent fibers of the median (tibial) nerves. They are diminished after MDT (in the order of two thirds). The later negative slow wave N2 (just visible in the cervical recording) corresponds to postsynaptic dorsal horn activity consecutive to the activation of group II and III afferent fibers. N2 is suppressed after MDT (Reprinted from [46]).

ulnar nerve. They have a maximum amplitude in the L5–S2 segments for stimulation of the tibial nerve, and in the S2–S4 segments for stimulation of the dorsal nerve of the penis or clitoris (see Chapter 9). Recordings of surface spinal cord SEPs can also be helpful in monitoring the surgical lesion itself. Dorsal column potentials can be monitored to check the integrity of the ascending dorsal column fibers, especially when the dorsolateral sulcus is not clearly marked (as is common in brachial root avulsion). The dorsal horn potentials can be monitored to follow the extent of MDT, particularly when good sensory functions are present before surgery (see legend of Fig. 5.7).

1.4 MICROELECTROPHYSIOLOGY AND MICRODIALYSIS STUDIES IN THE DORSAL HORN DURING SURGERY Unitary spikes generated in the dorsal horn neurons are interesting to record during DREZotomy to indicate abnormal activities, to help identify the surgical target, and to better understand the electrophysiological mechanisms underlying painful phenomena. Toward this goal, our group in Lyon has developed

Selective Spinal Cord Lesioning Procedures for Pain and Spasticity

111

special, simplified floating microelectrodes. At the beginning, these electrodes were based on the design by Merril and Ainsworth [49], and later they were developed into an original design (i.e., a double microelectrode with an enhanced ability to distinguish spikes from artifacts [50]). In this way, we have conducted recordings in 25 patients. To learn about specific patterns recorded from deafferented neurons, more patients must be studied. With approval from our Ethical Committee, our group has performed microdialysis studies in the dorsal horn of patients undergoing DREZotomy [51]. The aim of the work was to measure concentrations of some of the main neurotransmitters hypothesized from the animal experiments to be present in the human dorsal horn. The microprobe has an apical 4-mm-long tubular membrane [diameter: 0.216 mm, Cuprophane (HOSPAL Industrie, Meyzieu, France), cutoff 6000 Da]. The probe is perfused at 2 µl/min with a Ringer solution. Dialysate fractions are collected from the extracellular fluid every 5 min for about 1 hr. All the samples are frozen for later analysis [high performance liquid chromatography (HPLC) with fluometric detection]. At the present time we have made the technique feasible for identification and dosages of the following substances: glutamate, aspartate, GABA, glycine, taurine, serine, and threonine. The preliminary results indicate some differences between painful and nonpainful states. Further studies are needed before we can give conclusions.

2 RADIOFREQUENCY (RF) THERMOCOAGULATION PROCEDURE In 1976 Nashold and his group published data on a method using RF thermocoagulation to destroy hyperactive neurons in the substantia gelatinosa [35] and in 1979 in the whole DREZ region [36]. In 1981 [52] the technique was modified to produce less extensive lesioning so that the risk of encroachment into the neighboring corticospinal tract and dorsal column would be minimized. With the modified technique, the lesion is made with a 0.5 mm insulated stainless steel electrode with a tapered noninsulated 2 mm tip, designed and manufactured by Radionics Inc. (Burlington, MA). For treatment of pain after brachial plexus avulsion, the electrode penetrates the dorsolateral sulcus to a depth of 2 mm at an angle of 25–45° in the lateral– medial direction. A series of RF coagulations are made under a current of 35–40 mA (not over 75°C) for 10–15 s. The RF lesions are spaced at 2–3 mm intervals along the longitudinal extent of the dorsolateral sulcus. The lesion observed under magnification is seen as a circular whitened area that extends 1–2 mm beyond the tip of the electrode.

112

Marc Sindou and Patrick Mertens

In a recent publication, Nashold emphasizes the importance of obtaining impedance measurements from tissue during surgery [53]. Before and after each lesion is made, the impedance has to be measured. It is usually less than 1200 Ω in a damaged spinal cord. The authors state that as the transition from injured parenchyma into more normal tissue is made, impedance readings should increase and eventually reach normal levels of 1500 Ω. The authors use these numbers as a guide to stop the lesion making at the desired end.

3 DREZ PROCEDURES WITH THE LASER BEAM Levy et al. in 1983 [37] and Powers et al. in 1984 [38] advocated CO 2 and argon lasers, respectively, as lesion makers. According to Levy et al.’s description, the pulse duration of the CO2 laser is 0.1 sec and the power is adjusted to about 20 W, so that one or two single pulses create a 2 mm depression at a 45° angle in the DREZ. The lesions are probed with a microinstrument marked at 1 mm increments to ensure that the depth of the lesions (1–2 mm) is adequate. Intraoperative observations in humans and experimental studies comparing DREZ lesions performed with the RF thermocoagulation to those made with various laser beams [54] found that the laser lesions were generally more circumscribed and less variable. Walker et al. [55], on the other hand, reported on the danger of creating extensive damage and syrinx cavities with the laser (CO2). In a well-documented study evaluating the effects of DREZ lesions with RF or CO2 on the dog spinal cord, Young [56] found that the size and extent of the lesion related primarily to the magnitude of power used to make the lesion. They showed that by using any of the three techniques, the lesions could be successfully localized to the DREZ (including the layers I–VI of the dorsal horn) and the dorsal column and the corticospinal tract spared. The main difference was that with the laser, the lesion was shaped like the letter “V”, with the maximum width at the surface, whereas with RF it tended to be more spherical. The same glial reactions were observed using both methods in chronic animal models. Young [57], in his series of patients, made a comparative analysis of RF and CO2-laser procedures. With RF, 39 of the 58 patients (67%) reported good results (pain regressed by 50% or more) and with the CO2 laser, 9 out of the 20 patients (45%) reported good results. Postoperative complications with RF were noted in 26%, and with CO2 laser in 15%.

Selective Spinal Cord Lesioning Procedures for Pain and Spasticity

113

4 ULTRASONIC DREZ PROCEDURE This procedure was developed by Kandel and Dreval [39, 40] in Moscow. It has been mostly used for pain caused by brachial plexus avulsion. According to the description given by Dreval, the technique consists of a continuous longitudinal opening of the dorsolateral sulcus at the level of the avulsed roots to the depth of the microcavities and the changed spongy cord tissue. At the same time, ultrasonic destruction of the pathological tissues is done. The lesion is strictly in the projection of the dorsolateral sulcus at an angle of 25° medially and ventrally. The depth of the microcavities is the main criterion of the depth of the lesioning. After ultrasonic DREZ sulcomyelotomy, the grey color of the dorsal horn is well seen in the depth of the opened dorsolateral sulcus. The vessels crossing the sulcus are kept intact. The ultrasonic lesions are produced at a working frequency of 44 kHz, and the amplitude of ultrasonic oscillation is 15–50 µm. The lesions are placed in a “chain” manner along the sulcus.

5 INDICATIONS FOR DREZ Because of our experience with 362 patients operated on since 1972 for severe chronic pain [58] and in consideration of the literature data [59], we conclude that indications are as follows: 1. Cancer pain that is limited in extent (such as in Pancoast-Tobias syndrome). 2. Persistent neurogenic pain that is due to: A) Brachial plexus injuries, especially those with avulsion. B) Spinal cord lesions, especially for pain corresponding to segmental lesions. Pain below the lesion is not favorably influenced. Segmental pain caused by lesions in the conus medullaris and the cauda equina is significantly relieved. Pain due to cauda equina lesions can also be indications. C) Peripheral nerve injuries, amputation, and herpes zoster, when the predominant component of pain is of the paroxysmal type and/or corresponds to provoked allodynia hyperalgesia. 3. Disabling hyperspasticity with pain. Surgery in the DREZ must be considered alongside other methods belonging to the armamentarium of pain surgery. Figure 5.8 summarizes our present process of decision making for neuropathic pain [60].

114

Marc Sindou and Patrick Mertens

FIGURE 5.8 Decision making for neuropathic pain, originating from the following: upper left, peripheral nerves, plexus, roots distal to ganglion lesions; upper right, roots central to ganglion lesions; lower right, incomplete and complete spinal cord lesions; lower left, treatment for the segmental and the infralesional components of the pain are different.

REFERENCES 1. Foerster, O. (1913). On the indications and results of the excision of posterior spinal nerve roots in men. Surg. Gynecol. Obstet., 16, 463–474. 2. Gros, C., Ouaknine, G., Vlahovitch, B., and Frerebeau, P. (1967). La radicotomie sélective postérieure dans le traitement neurochirurgical de l’hypertonie pyramidale. Neurochirurgie, 13, 505–518. 3. Ouaknine, G. (1980). Le traitement chirurgical de la spasticité. Union Med. Can., 109, 1–11. 4. Gros, C. (1979). Spasticity: Clinical classification and surgical treatment. In “Advances and technical standards in neurosurgery” (H. Krayenbühl, ed.), vol. 6, pp. 55–97. Springer-Verlag, Wien. 5. Privat, J.M., Benezech, J., Frerebeau, P., and Gros, C. (1976). Sectorial posterior rhizotomy: A new technique of surgical treatment of spasticity. Acta Neurochir., 35, 181–195. 6. Frerebeau, P.H. (1991). Sectorial posterior rhizotomy for the treatment of spasticity in children with cerebral palsy. In “Neurosurgery for spasticity: A multidisciplinary approach.” (M. Sindou, A. Abbott, and Y. Keravel, eds.), pp. 145–147. Springer-Verlag, Vienna, New York. 7. Fraioli, B., and Guidetti, B. (1977). Posterior partial rootlet section in the treatment of spasticity. J. Neurosurg., 46, 618–626. 8. Fasano, V.A., Barolat-Romana, G., Ivaldi, A., and Sguazzi, A. (1976). La radicotomie postérieure fonctionnelle dans le traitement de la spasticité cérébrale. Neurochirurgie, 22, 23–24.

Selective Spinal Cord Lesioning Procedures for Pain and Spasticity

115

9. Peacock, W.J., and Aarens, L.J. (1982). Selective posterior rhizotomy for the relief of spasticity in cerebral palsy. S. Afr. Med. J., 62, 119–124. 10. Cahan, L.D., Kundi, M.S., McPherson, D., Starr, A., and Peacock, W.J. (1987). Electrophysiologic studies in selective dorsal rhizotomy for spasticity in children with cerebral palsy. Appl. Neurophysiol., 50, 459–682. 11. Abbott, R., Forem, S.L., and Johann, M. (1989). Selective posterior rhizotomy for the treatment of spasticity. Childs Nerv. Syst., 5, 337–346. 12. Storrs, B. (1987). Selective posterior rhizotomy for treatment of progressive spasticity in patients with myelomeningocele. Pediatr. Neurosci., 13, 135–137. 13. Hodgkinson, I., Berard, C., Jindrich, M.L., Sindou, M., Mertens, P., and Berard, J. (1996). Radicotomie postérieure fonctionnelle chez l’enfant IMC. Résultats à un an post-opératoire sur 18 cas. Ann. Réadaptation Med. Phys., 39, 103–111. 14. Bischof, W. (1951). Die longitudinale myelotomie. Zentralbl Neurochir., 2, 79–88. 15. Pourpre, M.H. (1960). Traitement neurochirurgical des contractures chez les paraplégiques post-traumatiques. Neurochirurgie, 6, 229–236. 16. Laitinen, L.V., and Singounas, E. (1971). Longitudinal myelotomy in the treatment of spasticity of the legs. J. Neurosurg., 35, 536–540. 17. Sindou, M. (1972). Etude de la jonction radiculo-médullaire postérieure: La radicellotomie postérieure sélective dans la chirurgie de la douleur. These med., Lyon. 18. Sindou, M., Fischer, G., Goutelle, A., Schott, B., and Mansuy, L. (1974). La radicellotomie postérieure sélective dans le traitement des spasticités. Rev. Neurol., 130, 201–215. 19. Sindou, M., Millet, M.F., Mortamais, J., and Eysette, M. (1982). Results of selective posterior rhizotomy in the treatment of painful and spastic paraplegia secondary to multiple sclerosis. Appl. Neurophysiol., 45, 335–340. 20. Sindou, M., Pregelj, R., Boisson, D., Eyssette, M., and Goutelle, A. (1985). Surgical selective lesions of nerve fibers and myelotomies for the modification of muscle hypertonia. In “Recent achievements in restorative neurology: Upper motor neuron functions and dysfunctions” (Sir J. Eccles, and M.R. Dimitrijevic, eds.), pp. 10–26. Basel, S. Kaeger. 21. Sindou, M., Abdennebi, B., and Sharkey, P. (1985). Microsurgical selective procedures in the peripheral nerves and the posterior root-spinal cord junction for spasticity. Appl. Neurophysiol., 48, 97–104. 22. Eccles, J., Eccles, R., and Magni, F. (1961). Central inhibitory action attributable to presynaptic depolarization produced by muscle afferent volleys. J. Physiol., 159, 147–166. 23. Sindou, M., Jeanmonod, D., and Mertens, P. (1991). Surgery in the dorsal root entry zone: Microsurgical DREZ-tomy (MDT) for the treatment of spasticity. In “Neurosurgery for spasticity: A multidisciplinary approach” (M. Sindou, R. Abbott, and Y. Keravel, eds.), pp.165–182. Springer-Verlag, Wien, New York. 24. Sindou, M., and Jeanmonod, D. (1989). Microsurgical-DREZ-otomy for the treatment of spasticity and pain in the lower limbs. Neurosurgery, 24, 655–670. 25. Sindou, M., Mifsud, J.J., Boisson, D., and Goutelle, A. (1986). Selective posterior rhizotomy in the dorsal root entry zone for treatment of hyperspasticity and pain in the hemiplegic upper limb. Neurosurgery, 18, 587–595. 26. Beneton, C., Mertens, P., Leriche, A., and Sindou, M. (1991). The spastic bladder and its treatment. In “Neurosurgery for spasticity: A multidisciplinary approach” (M. Sindou, R. Abbott, and Y. Keravel, eds.), pp. 193–199. Springer-Verlag, Wien, New York. 27. Sindou, M. (1997). Spinal entry zone interruption for spasticity. In “Textbook of stereotactic and functional neurosurgery” (R.R. Tasker, and P. Gildenberg, eds.), pp. 1257–1266. McGrawHill, New York. 28. Sindou, M., and Mertens, P. (1991). Indication for surgery to treat adults with harmful spasticity. In “Neurosurgery for spasticity: A multidisciplinary approach” (M. Sindou, R. Abbott, and Y. Keravel, eds.), pp. 211–213. Springer-Verlag, Wien, New York.

116

Marc Sindou and Patrick Mertens

29. Sindou, M., Abbott, R., and Keravel, Y., eds. (1991). “Neurosurgery for spasticity: A multidisciplinary approach.” Springer-Verlag, Wien, New York. 30. Abbott, R. (1991). Indications for surgery to treat children with spasticity due to cerebral palsy. In “Neurosurgery for spasticity: A multidisciplinary approach” (M. Sindou, R. Abbott, and Y. Keravel, eds.), pp. 215–217. Springer Verlag, Wien, New York. 31. Melzach, R., and Wall, P.D. (1965). Pain mechanism: A new theory. Science, 150, 971–979. 32. Wall, P.D., and Sweet, W.H. (1967). Temporary abolition of pain in man. Science, 155, 108–109. 33. Sindou, M. (1972). Etude de la jonction radiculo-médullaire postérieure: La radicellotomie postérieure sélective dans la chirurgie de la douleur. These med., Lyon, 182 pp. 34. Sindou, M., Quoex, C., and Baleydier, C. (1974). Fiber organization at the posterior spinal cord-rootlet junction in man. J. Comp. Neurol., 153, 15–26. 35. Nashold, B.S., Urban, B., and Zorub, D.S. (1976). Phantom pain relief by focal destruction of substantia gelatinosa of Rolando. In “Advances in pain research and therapy” ( J.J. Bonica, and D. Albe-Fessard, eds.), vol. 1, pp. 959–963. Raven Press, New York. 36. Nashold, B.S., and Ostdahl, P.H. (1979). Dorsal root entry zone lesions for pain relief. J. Neurosurg., 51, 59–69. 37. Levy, W.J., Nutkiewicz, A., Ditmore, M., and Watts, C. (1983). Laser induced dorsal root entry zone lesions for pain control: Report of three cases. J. Neurosurg., 59, 884–886. 38. Powers, S.K., Adams, J.E., Edwards, S.B., Boggan, J.E., and Hosobuchi, Y. (1984). Pain relief from dorsal root entry zone lesions made with argon and carbon dioxide microsurgical lasers. J. Neurosurg., 61, 841–847. 39. Kandel, E.L., Ogleznev, K.J.A., and Dreval, O.N. (1987). Destruction of posterior root entry zone as a method for treating chronic pain in traumatic injury to the brachial plexus. Vopr. Neurochir., 6, 20–27. 40. Dreval, O.N. (1993). Ultrasonic DREZ-operations for treatment of pain due to brachial plexus avulsion. Acta Neurochir., 122, 76–81. 41. Jeanmonod, D., and Sindou, M. (1991). Somatosensory function following dorsal root entry zone lesions in patients with neurogenic pain or spasticity. J. Neurosurg., 74, 916–932. 42. Sindou, M., Fischer, G., Goutelle, A., and Mansuy, L. (1974). La radicellotomie posterieure sélective: Premiers résultats dans la chirurgie de la douleur. Neurochirurgie, 20, 391–408. 43. Sindou, M., Fischer, G., Goutelle, A., Schott, B., and Mansuy, L. (1974). La radicellotomie postérieure sélective dans le traitement des spasticités. Rev. Neurol., 130, 201–215. 44. Sindou, M., Fischer, G., and Mansuy, L. (1976). Posterior spinal rhizotomy and selective posterior rhizidiotomy. In “Progress in neurological surgery” (H. Krayenbühl, P.E. Maspes, and W.H. Sweet, eds.), vol. 7, pp. 201–250. Basel, Karger. 45. Sindou, M., and Goutelle, A. (1983). Surgical posterior rhizotomies for the treatment of pain. In “Advances and technical standards in neurosurgery” (H. Krayenbül, ed.), vol. 10, pp. 147–185. Springer-Verlag, Vienna. 46. Jeanmonod, D., Sindou, M., and Mauguière, F. (1991). The human cervical and lumbo-sacral evoked electrospinogram: Data from intra-operative spinal cord surface recordings. Electroencephalogr. Clin. Neurophysiol., 80, 477–489. 47. Turano, G., Sindou, M., and Mauguière, F. (1995). SCEP monitoring during spinal surgery for pain and spasticity. In “Atlas of human spinal cord evoked potentials” (M.R. Dimitrijevic´, and J.A. Halter, eds.). Butterworth–Heinemann, Boston. 48. Sindou, M., Turano, G., Pantieri, R., Mertens, P., and Mauguière, F. (1994). Intraoperative monitoring of spinal cord SEPs, during microsurgical DREZotomy (MDT) for pain, spasticity and hyperactive bladder. Stereotact. Funct. Neurosurg., 62, 164–170.

Selective Spinal Cord Lesioning Procedures for Pain and Spasticity

117

49. Jeanmonod, D., Sindou, M., Magnin, M., and Baudet, M. (1989). Intra-operative unit recordings in the human dorsal horn with a simplified floating microelectrode. Electroencephalogr. Clin. Neurophysiol., 72, 450–454. 50. Guenot, M., Hupe, J.M., Mertens, P., Mauguière, F., Bullier, J., and Sindou, M. (1996–97). Microelectrode recordings during microsurgical DREZotomy. Stereotact. Funct. Neurosurg., 67(1–2), Abstract 210, p. 56. 51. Mertens, P., Ghaemmaghami, C., Perret-Liaudet, A., Guenot, M., Sindou, M., and Renaud, B. (1996–97). In vivo amino-acid concentrations in human dorsal horn studied by microdialysis during DREZotomy: Methodology and preliminary results. Stereotact. Funct. Neurosurg., 67(1–2), Abstract 213, p. 58. 52. Nashold, B.S. (1981). Modification of DREZ lesion technique (letter). J. Neurosurg., 55, 1012. 53. Nashold, J.R.B., and Nashold, D.S. (1995). Microsurgical DREZotomy in treatment of deafferentation pain. In “Operative neurosurgical techniques: Third edition” (H.H. Schmidek, and W.H. Sweet, eds.), pp. 1623–1636. W.B. Saunders, Philadelphia. 54. Levy, W.J., Gallo, C., and Watts, C. (1985). Comparison of laser and radiofrequency dorsal root entry zone lesions in cats. Neurosurgery, 16, 327–330. 55. Walker, J.S., Ovelmen-Levitt, J., Bullard, D.E., and Nashold, B.S. (1984). Dorsal root entry zone lesions using a CO2 laser in cats with neurophysiologic and histologic assessment. Neurosurgery, 15, 265. 56. Young, R.F., Foley, K., Chambi, I.V., and Rand, R.W. (1988). A comparison of radiofrequency and laser techniques. Personal communication. 57. Young, R.F. (1990). Clinical experience with radio-frequency and laser DREZ lesions. J. Neurosurg., 72, 715–720. 58. Sindou, M. (1995). Microsurgical DREZotomy (MDT) for pain, spasticity and hyperactive bladder: A 20 year experience. Acta Neurochir., 137, 1–5. 59. Sindou, M., and Daher, A. (1988). Spinal cord ablation procedures for pain. In “Proceedings of the Fifth World Congress on Pain” (A. Dubner, G.F. Gebbart, and M.R. Bond, eds.), pp. 477–495. Elsevier, Amsterdam. 60. Sindou, M., and Mertens, P. (1997). Dorsal root entry zone procedures: Indications and techniques. In “Proceedings of 11th International Congress of Neurological Surgery” (World Federation of Neurological Societies, ed.), vol. 1, pp. 175–181, 6–11 July 97, Monduzzi, Bologna.

CHAPTER

6

Neurophysiological Monitoring During Endovascular Procedures on the Spine and the Spinal Cord FRANCESCO SALA Section of Neurosurgery, Department of Neurological Sciences and Vision, Verona University, Verona, Italy

YASUNARI NIIMI Center for Endovascular Surgery, Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York

1 Spinal Cord Vascularization and Ischemia 1.1 Vascular Anatomy of the Human Spinal Cord 1.2 Primers on the Pathophysiology of Spinal Cord Ischemia Secondary to Spinal Cord Vascular Malformations 2 Neurophysiological Monitoring 2.1 Evoked Potentials in Spinal Cord Ischemia: Experimental and Clinical Studies 2.2 Clinical Application of Neurophysiological Monitoring for Endovascular Treatment of Spine and Spinal Cord Vascular Lesions 3 Endovascular Treatment of Vascular Malformations and Tumors of the Spine and the Spinal Cord 3.1 Indications 3.2 Angiographic Vascular Anatomy of the Spine and Spinal Cord 3.3 Spinal Angiography Neurophysiology Book Title in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, line Elsevier Science (USA). All rights of reproduction in any form reserved.

119

120

Francesco Sala and Yasunari Niimi

3.4 Angiographic Evaluation, Endovascular Treatment, and Clinical Aspect of Neurophysiological Monitoring 4 Conclusions References ABSTRACT After reviewing the vascular anatomy of the spinal cord and the pathophysiology of spinal cord ischemia, this chapter will discuss the role of neurophysiological monitoring in animal studies and during ischemic events in the human spinal cord. In order to assess the feasibility and reliability of neurophysiological monitoring, we will present data from a series of 110 consecutive endovascular procedures performed under general anesthesia with neurophysiological monitoring. The different but complementary roles of MEP and SEP monitoring will be discussed together with the critical importance of provocative tests using Amytal and Xylocaine. Finally, primers on the angiographic vascular anatomy of the spine and spinal cord and on the endovascular procedures aimed to treat these hypervascular lesions will be given. Examples of particularly challenging procedures in spinal interventional neuroradiology, and the benefit of neurophysiological assistance during these procedures, are presented.

1 SPINAL CORD VASCULARIZATION AND ISCHEMIA

1.1 VASCULAR ANATOMY OF THE HUMAN SPINAL CORD Before describing the role of neurophysiological monitoring during endovascular procedures aimed to treat spinal hypervascular lesions, an overview on the vascular anatomy of the normal spinal cord is mandatory. Although we refer the reader to classical textbooks and articles for a detailed analysis of the vascular anatomy and the wide range of variations, here we will concisely describe those aspects relevant to the discussion of intraoperative neurophysiological techniques. During early embryonal development, each somite receives one pair of socalled segmental arteries arising from the dorsal aorta. Blood supply to the neural crest is then provided by a dorsomedial division of the ipsilateral segmental artery, the dorsospinal artery. This vessel supplies the neural tube via paired ventral longitudinal arteries. Dorsally oriented branches of these ventral arteries penetrate deeply into the ipsilateral half of the neural tube. A network of capillaries around the neural tube then organizes into longitudinal arterial axes. Between the sixth week and the fourth month of uterine life, the craniocaudal formation of a more mature vascular pattern is characterized by the ventral migration and then fusion of the ventral longitudinal arterial axes to form the anterior spinal axis. On the posterior aspect of the cord, the pial network organizes into paramedian dominant axes that will give rise to two posterolateral

Neurophysiological Monitoring During Endovascular Procedures

121

spinal axes. Like the embryonal circulation of other systems, a multistage rearrangement of the original radicular feeders to the ventral and posterior axes also occurs. As a result of this process, no more than 4 to 8 anterior and 10 to 20 posterior radicular arteries remain by the end of development. In the adult, we observe three main longitudinal arterial systems. The anterior spinal artery (ASA) extends almost uninterrupted from the medulla to the filum terminale. At the cervicomedullary junction it originates from the two vertebral arteries near the vertebrobasilar junction. Caudally, the major blood supply comes from dorsal branches of intercostals and lumbar arteries. The major radiculomedullary arteries arise, at the level of the cervical enlargement, from vertebral, deep cervical, or ascending cervical arteries. The thoracolumbar territory is supplied mainly by the arteria radicularis anterior magna or “artery of Adamkiewicz.” This usually rises from the ninth to the twelfth intercostal artery, on the left side in approximately 80% of the cases. It gives off a small ascending branch and a large descending branch that anastomoses with the posterior spinal arteries to configurate the anastomotic basket surrounding the conus medullaris [1–3]. Because of its segmental vascularization, each major arterial group (cervical, upper thoracic, and Adamkiewicz) irrigates its own portion of the cord without significant anastomoses with other groups. Consequently, the spinal cord is typically vulnerable to hypoperfusion at the middle thoracic level. The paired posterior spinal arteries arise, at the cervical level, either from the vertebral arteries or, less frequently, from the posteroinferior cerebellar arteries. Caudally, these paired posterior spinal axes receive radiculopial feeders also from the vertebral, intercostal, and lumbar arteries, and they are located on the posterolateral surface of the cord adjacent to the dorsal root entry zone. The numerous anastomoses in this posterior system decrease the risk of ischemia for the posterior spinal cord. From a neurophysiological perspective, it is important to bear in mind that the direction of spinal cord blood flow (SCBF) at any level in the cord cannot be easily predicted because it depends on the location of the dominant anterior or posterior spinal artery for that segment of the cord [4]. Nevertheless, as summarized in Fig. 6.1 (see also color plate), the ASA, through perforating sulcocommissural arteries, is assumed to supply the anterior two thirds to four fifths of the cord, including the anterior column of the central grey matter, the anterior and lateral corticospinal tracts, and the anterior and lateral spinothalamic tracts. The ASA therefore accounts for vascularization of those structures involved in the propagation of motor evoked potentials (MEPs) from their cortical generators to the α-motoneurons: the anterior and lateral corticospinal tracts (CSTs) and, to a lesser extent, the propriospinal system. Conversely, the posterior spinal arteries (PSAs) supply the posterior horns of the central grey matter and the dorsal columns; although the debate is still open, these posterior columns are usually considered the main tracts for central propagation of somatosensory evoked potentials (SEPs) after peripheral stimulation [5].

122

Francesco Sala and Yasunari Niimi

FIGURE 6.1 Schematic representation of spinal vascular anatomy and its relationship with long tracts involved in the generation of somatosensory and motor evoked potentials. 1. Posterior spinal arteries. 2. Posterior spinal vein. 3. Anterior spinal artery. 4. Anterior spinal vein. 5. Spinal ventral roots. 6. Anterior corticospinal tracts. 7. Lateral corticospinal tracts. 8. Dorsal columns. (Modified from Nieuwenhuys, R., Voogd, J., and van Huijzen, C. (1988). The human central nervous system: A synopsis and atlas, rev. ed. 3. Springer Verlag, Berlin) (see also color plate).

Circumferential vessels from the ASA anastomose with the PSAs through a complex pial network, the so-called vasa corona [6], which supplies the peripheral rim of the white matter and represents a functionally relevant dorsoventral connection. Therefore, in the axial plane, the watershed zone of the cord is located in the anterior two thirds of the cord in the white matter adjacent to the anterior horn cells, where penetrating branches from the ASA and PSA meet at the circumferential pial network [7]. The complexity of both longitudinal and axial angioarchitecture of the vascular supply to the spinal cord accounts for the unpredictability of hemodynamic

Neurophysiological Monitoring During Endovascular Procedures

123

FIGURE 6.2 Left, schematic illustration of the anatomic relationship between the aorta (AO), the intercostal artery (ICA), and the radiculomedullary artery (RMA) contributing to the anterior spinal artery axis (ASA). A spinal cord (SC) arteriovenous malformation (AVM) is represented. Xylocaine injected during a provocative test is schematically represented by open circles (modified from [58] with permission from Elsevier). Right, typical example of mMEP behavior after provocative test with Xylocaine injection at different catheter positions within the ASA. Top, muscle motor evoked potentials (mMEPs) are elicited through transcranial electrical stimulation and recorded from needle electrodes inserted in the tibialis anterior muscle. Middle, when the tip of the catheter is in position 1, the injected Xylocaine will flow through vessels feeding the normal spinal cord (a). Provocative test will be positive: absence of the mMEP from the tibialis anterior muscle (TA). Bottom, when the tip of the catheter is more selectively advanced to position 2 or 3, Xylocaine will be injected only in vessels feeding the AVM (b and c) and, accordingly, the provocative test will be negative: persistence of mMEP from the TA. Reprinted from [57].

patterns in the spinal cord; the direction of SCBF becomes even more bizarre in the presence of a vascular malformation that interferes with normal patterns. Although studies on spinal cord vascular anatomy have mainly focused on the arterial circulation, it has to be emphasized that venous anatomy is equally essential, since it is dramatically involved in the pathophysiology of most vascular malformations. Although venous anatomy is even more unpredictable than its arterial counterpart, two drainage pathways are usually considered. Sulcal veins drain blood from the central portion of the cord through the anterior median fissure into the anterior median spinal vein; this receives blood from tributaries of central veins that drain the central grey matter, including the anterior horns. A dorsal spinal vein, often larger than the anterior one, drains

124

Francesco Sala and Yasunari Niimi

the posterocentral portion of the cord. The radial or coronal veins originate from capillaries at the gray–white junction, coursing centrifugally and draining the anterolateral and dorsal regions of the spinal cord. The final common pathway of spinal cord venous drainage is through the radicular veins that pierce the dura to drain into the epidural veins; these radicular veins lack valves but typically narrow at the dural penetration to prevent retrograde venous flow [8].

1.2 PRIMERS ON THE PATHOPHYSIOLOGY OF SPINAL CORD ISCHEMIA SECONDARY TO SPINAL CORD VASCULAR MALFORMATIONS More recent classifications of spinal cord vascular malformations include a number of lesions whose characteristics will be described later in this chapter. However, regardless of their specific hemodynamics, the final common pathway in the pathophysiology of these lesions is spinal cord ischemia or hemorrhage. Basically, in both the arteriovenous malformations and in the fistulas, the main mechanism is the lack of a capillary bed and a direct shunt of the arterial blood into the venous compartment. This arteriovenous shunt leads to vascular steal phenomena from the adjacent normal vasculature; the more the malformation shares its arterial supply with the normal spinal cord, the more the cord will be exposed to a vascular steal, and therefore to an ischemic injury. On the venous side, a multifactorial phenomenon leads to venous hypertension and thrombosis. Venous inflow is increased because of direct arterial feeders; venous outflow is sometimes compromised by a malfunctioning of the valve system of radicular veins when these pierce the dura (which contributes to venous engorgement). A subacute necrotizing myelopathy resulting from thrombosis of a spinal arteriovenous malformation (AVM) has been described as the Foix-Alajounine syndrome [9]. The coexistence of venous or arterial aneurysms increases the risk for subarachnoidal and parenchymal hemorrhages, which account for the acute onset of symptoms in intradural AVMs. Arachnoiditis may evolve from repeated hemorrhages.

2 NEUROPHYSIOLOGICAL MONITORING

2.1 EVOKED POTENTIALS IN SPINAL CORD ISCHEMIA: EXPERIMENTAL AND CLINICAL STUDIES Most data on the role of neurophysiological techniques in decreasing the incidence of spinal cord ischemia come from thoracoabdominal aneurysm surgery. The role of SEPs and MEPs in detecting cord ischemia and preventing irreversible neurological deficits has been investigated both in experimental animal models and in clinical studies.

Neurophysiological Monitoring During Endovascular Procedures

125

Konrad et al. tested the sensitivity of MEPs recorded epidurally from the spinal cord as well as from the peripheral nerve after direct cortical stimulation of the motor cortex in dogs. The peripheral nerve response appeared to be very sensitive to cord ischemia after cardiac arrest when compared to the late disappearance of the spinal cord response [10]. Similarly, Kai and coworkers concluded that peripheral neurogenic MEPs provide a better warning system for spinal cord ischemia than spinal MEPs and SEPs recorded from peripheral nerves after stimulation of the spinal cord; unfortunately, the spine-to-spine response is not specific, since it activates neural pathways both orthodromically and antidromically [11]. Laschinger et al. investigated spinal cord ischemia after thoracic aortic crossclamping in dogs. He documented a time- and level-dependent deterioration and loss of the spinal cord response recorded from subcutaneously inserted spinal electrodes after spinal cord stimulation at T3–T4. This suggests that ischemia begins in the most distal cord, progresses upwardly, and can be prevented by maintaining an adequate distal aortic perfusion [12]. Similar results on the higher vulnerability of the lower spinal cord were reported by Reuter and coworkers, who correlated MEPs to ischemic spinal damage after aortic occlusion in dogs; the greater portion of cord damage was confined to the grey matter of the caudal segments of the cord [13]. This might be related to a discontinuous ASA; if the lumbar cord relies only on the Adamkievicz artery for its blood supply, occlusion of this artery would not be corrected by collateral feeders and perfusion would be inadequate. These authors also confirmed the early disappearance of the peripheral nerve MEP, which was considered even too sensitive as an indicator of spinal cord damage. Conversely, they found a clear correlation between spinal MEPs after brain stimulation, spinal cord perfusion, and histopathologic findings. In the same study, SEPs appeared to be more sensitive than spinal MEPs to ischemia [13]. Concerning the different role of spinal MEPs as compared to peripheral nerve MEPs, it is noteworthy that Reuter et al. observed the presence of spinal MEPs but the absence of peripheral nerve MEPs 24 hr after cord ischemia, when animals were paraplegic. This was explained on the basis of histological findings, since the damage was primarily confined to the gray matter but did not significantly affect the white matter, where propagation of the descending volleys was preserved [13]. To elicit peripheral nerve MEPs after brain stimulation, conversely, requires the functional integrity of the anterior horns. A similar comparison of the sensitivity between spinal and muscle MEPs (mMEPs) after transcranial electrical stimulation in the detection of spinal cord ischemia was performed by de Haan et al. [14]. They concluded that mMEPs disappear earlier and are therefore more sensitive than epidural MEPs, suggesting their clinical use to assess spinal cord perfusion during surgery at risk for ischemia. Early disappearance of mMEPs is secondary to the polysynaptic transmission of this potential, so that a reduction in SCBF that affects the functional integrity of

126

Francesco Sala and Yasunari Niimi

the anterior horns will switch off neural transmission at that level. Epidurally recorded MEPs, conversely, are more robust, since no synapses are involved and white matter is more resistant to ischemia than grey matter [10, 13, 15, 16]. In the past few years, along with experimental work, there has been increasing clinical evidence of the usefulness of neurophysiological monitoring during thoracic aorta surgery. For a long time, SEPs have been used to assess the functional integrity of the spinal cord [5, 17, 18]. Unfortunately, SEPs are aimed at monitoring the dorsal column and the posterior spinal cord, but they do not reflect the functional integrity of motor pathways. Dorsal column response to ischemia, moreover, is relatively slow [18], and SEP monitoring cannot detect ischemia in time to revert the injury before irreversible neuronal damage occurs. Machida et al. [19] described a dissociation of mMEPs (after spinal cord stimulation) and SEPs (also following spinal cord stimulation) resulting from ischemic damage to the spinal cord in both an experimental setting as well as during spinal fusion with Cotrel-Dubousset instrumentation; because of the greater vulnerability of mMEPs to ischemia when elicited by this method, they suggested the use of mMEPs as a sensitive measure of anterior cord function. Similarly, de Haan et al. proposed mMEPs after transcranial stimulation as optimal tools for assessing the status of motor pathways from the cortex to the muscle [20, 21]. In their experience, mMEPs turned out to be sensitive and specific, since they correctly predicted motor outcome in all patients with no false-negative (postoperative motor deficits despite unchanged motor evoked response) or false-positive results (significant changes in intraoperative MEPs despite unchanged motor outcome). It is less likely that spinal cord surgery could damage the anterior horn grey matter while leaving the white matter intact. The opposite is true during spinal cord embolization because of the selectivity of spinal cord vascularization. For generation of epidural MEPs after transcranial electrical stimulation, only intact long motor tracts are necessary. The generation of mMEPs after transcranial electrical stimulation, however, depends on long motor tracts and the segmental level anterior horn grey matter. Therefore, mMEPs should be a better monitoring tool for endovascular embolization of the spinal cord vessels.

2.2 CLINICAL APPLICATION OF NEUROPHYSIOLOGICAL MONITORING FOR ENDOVASCULAR TREATMENT OF SPINE AND SPINAL CORD VASCULAR LESIONS Endovascular techniques are increasingly used in the treatment of hypervascular lesions in the spinal cord and surrounding structures. The injection of embolizing materials has proven useful in the devascularization of spinal cord tumors and the occlusion of intramedullary, dural, or spinal arteriovenous malformations

Neurophysiological Monitoring During Endovascular Procedures

127

or fistulas [22–29]. The occurrence of vasospasm or the unrecognized obliteration of vessels feeding the normal spinal cord, however, put the spinal cord at risk for ischemia. If ischemia is not detected in time before irreversible damage has occurred, patients can suffer from permanent neurological deficits [30–33]. Since a detailed angiographic study and the following embolization can last for several hours, these procedures are often performed under general anesthesia, which also allows for an optimal angiographic study, as discussed later. In the past, a so-called wake-up test was performed to assess the neurological status of the patient immediately after any critical maneuver. These tests, however, prolong the procedure, carry discomfort to the patient, and might take too much time before protective measures are readily available. The greater reliability of neurophysiological monitoring in detecting spinal cord ischemia, when compared to the wake-up test, has been established for spine surgery [34, 35]. Through the years, intraoperative neurophysiological monitoring has been proposed as a valid alternative to the wake-up test to assess the functional integrity of neural pathways during endovascular procedures. Somatosensory evoked potentials have been used since the mid-1980s [36–38] based on the clinical evidence that SEPs were sensitive to compromises in anterior spinal artery circulation [36]. Concerns about the reliability of SEPs in evaluating the integrity of descending motor tracts during spine and spinal cord surgery [39, 40], as well as during aortic surgery [41, 42], have then appeared in the literature. As previously discussed in this chapter, the occurrence of a motor deficit despite intraoperatively unchanged SEPs is explained by the limited ability of SEPs to assess the functional integrity of corticospinal tracts. Nevertheless, despite the advent of reliable techniques to elicit MEPs under general anesthesia [43, 44], reports on the use of MEPs during endovascular procedures in the spinal cord remain anecdotal [45–48]. In the following sections we describe the protocol we currently use at the Institute for Neurology and Neurosurgery to perform multimodal neurophysiological monitoring during endovascular treatments. 2.2.1 Patient Setup and Anesthesiological Management In order not to delay the beginning of the endovascular procedure, it is desirable to have the anesthesiologist and the neurophysiologist working together in preparing the patient. As for any intraoperative neurophysiological monitoring procedure, the greater the collaboration between these two teams, the more effective will be both the anesthesiological and neurophysiological management of the patient. Although the basic setup of the patient does not differ from that occurring during any other neurosurgical procedures, it is important to bear in mind that in the angiography suite the patient can be moved upward and downward along the operating bed and that the angiographic machine may be

128

Francesco Sala and Yasunari Niimi

rotated around the main axis for angiographic purposes. This implies careful placement of cables for evoked potentials stimulation and recording in order to avoid stretching of the wires or interference with the angiographic steps. Electrodes for intraoperative neurophysiological monitoring (IOM) are hooked up as for any other surgical procedure where IOM is used. The standard neurophysiological monitoring consists of SEPs and mMEPs. We use the Axon Sentinel-4 evoked potential system with modified software (AXON Systems, Inc., Hauppage, NY) for both stimulation and recording. SEPs are elicited by stimulation of the median nerve at the wrist (intensity up to 40 mA, duration 0.2 ms, repetition rate of 4.3 Hz) and the posterior tibial nerve at the ankle (intensity up to 40 mA, duration 0.2 ms, repetition rate of 4.3 Hz). Recordings are performed via corkscrew-like electrodes inserted subcutaneously in the scalp (CS electrode, Neuromedical Inc., Herndon, VA) at C3′/C4′–CZ′ (median nerve) and at CZ′–FZ (tibial nerve) according to the 10–20 International EEG System. The mMEPs are elicited with transcranial electrical stimulation of the motor cortex using CS electrodes. Short trains of up to seven square-wave stimuli of 500 µs duration each and interstimulus intervals of 4 ms are applied at a repetition rate of 2 Hz and intensity up to 200 mA, through electrodes placed at C1 and C2 scalp sites, according to the International 10–20 EEG System. Muscle responses are then recorded via needle electrodes inserted into the abductor pollicis brevis (APB) and hypothenar muscle for the upper extremities and into the tibialis anterior (TA) and abductor hallucis brevis (AHB) muscles for the lower extremities, bilaterally. This technique has been used at our institution to monitor over 100 spinal cord tumor cases and is described in detail elsewhere [49, 50]. A more invasive technique for monitoring MEPs was described in 1991 by Katayama et al. [46]. This author elicited MEPs through a burr hole with cortical epidural stimulation and spinal epidural recording, proposing this method as an optimal and specific tool to assess the functional integrity of the CST. At the present time there is no need for such an invasive monitoring procedure since mMEPs are easily elicitable using multipulse transcranial stimulation. We usually do not monitor spinal epidural MEPs during these procedures since these patients may receive a considerable amount of heparine by the end of the procedure and trancutaneous placement of an epidural catheter for spinal recording would therefore be hazardous. Furthermore, as previously mentioned, data from experimental studies suggest that monitoring only epidural MEPs (D waves) will not cover the functional integrity of α-motoneurons, and, because of the higher resistance of white matter to ischemia, warning signs from this monitoring modality could be too delayed to allow prompt restoration of spinal cord perfusion. According to our experience with IOM during spinal cord tumor surgery [49, 50], both muscle and epidural MEPs are required for optimal IOM to predict outcome. For endovascular procedures, however, mMEPs are sufficient. Whenever the lesion involves the lumbosacral segments of the spinal cord, we add the monitoring of the bulbocavernosus reflex (BCR) to the battery of

Neurophysiological Monitoring During Endovascular Procedures

129

neurophysiological tests [51]. This oligosynaptic reflex allows us to assess the functional integrity of both the afferent and efferent fibers of the pudendal nerves together with the reflex center located in the gray matter at S2–S4 spinal segments. For stimulation of the dorsal penile nerve (pudendal afferents), two silver/silver chloride disc electrodes are placed on the dorsal aspect of the penis with the cathode proximal. In female patients, the cathode is placed over the clitoris and the anode over the labia majora. Rectangular pulses of 0.2 to 0.5 ms duration are applied as a train of five stimuli (interstimulus intervals of 4 ms) at a repetition rate of 2.3 Hz. Stimulus intensities do not exceed 40 mA. Recordings are made from the external anal sphincter muscle using two pairs of intramuscular Teflon-coated hooked wire electrodes inserted into the anal hemisphincters. The optimal neuroanesthesiological management compatible with intraoperative neurophysiological monitoring is discussed in Chapter 17 of this book. For interventional procedures, we use a continuous infusion of propofol (100– 150 µg/kg/min) and fentanyl (1 µg/kg/h), avoiding boluses. No halogenated agents or muscle relaxants are given after intubation. These parameters, while allowing an easier management of the interventional procedure and optimization of the angiographic results, warrant an anesthesia slightly lighter than that used for major surgical procedures so that patients can be quickly and easily awakened at the end of the procedure. Baseline traces of evoked potentials are taken at the beginning of the procedure. Because of the possibility of vasospasm secondary to a vessel catheterization, it is important to obtain baselines after anesthesia induction but before any angiographic maneuver occurs. If not recognized, this event could lead to ischemic derangements of the cord (see Fig. 6.6C, right). Since MEPs could induce muscle twitches that interfere with the imaging, it is preferable not to run mMEPs while the radiologist is performing the angiography. Muscle MEPs, however, do not need averaging and can be quickly assessed immediately after any relevant angiographic step. Conversely, neither the BCR nor the SEPs induce twitches and therefore can be continuously monitored. With regard to the feasibility of evoked potentials during endovascular procedures, results from our series demonstrate that these potentials are easily elicitable in the majority of the patients, unless severe neurological deficits have already compromised the functional integrity of neural pathways. In over 110 endovascular procedures in 87 patients who were treated for spine and/or spinal cord vascular lesions between 1996 and 1999, monitorability of evoked potentials was 80% for SEPs, 85% for the BCR, and 92% for mMEPs. Monitorability is defined as the presence of a reliable response after the induction of anesthesia but before any interventional procedures. There were no significant differences in monitorability between males and females for MEPs and SEPs, while the BCR seemed more difficult to elicit in females, most likely because of technical difficulties in placing stimulating electrodes.

130

Francesco Sala and Yasunari Niimi

2.2.2 Endovascular Procedure and Provocative Tests Figure 6.3 summarizes the protocol for spinal endovascular embolization used at our institution. The first step of the procedure consists of a detailed and careful investigation of the vascular anatomy. This is necessary for successful treatment, independent of any additional neurophysiological support. More details on the angiographic protocols used for each disease category have been provided later in this chapter (see Section 3). Regardless of the pathology to be treated, a critical step of IOM during endovascular procedures is the so-called provocative test. Provocative tests are aimed at assessing the safety of a planned embolization, since they pharmacologically mimic the effects of the embolization itself. Because of the intimate relationship of the malformations with the spinal cord vascularization, the embolization of an intradural AVM is the most risky procedure and one that we will refer to in discussing the role of provocative tests. To access an intradural spinal AVM, the catheter is introduced in the pedicle artery and then superselectively advanced into the spinal cord artery (ASA or PSA) feeding the AVM. Once the catheter has reached the embolizing position, but before any embolizing material is injected, provocative tests are performed. This sort of “Wada test” [52] for the spinal cord consists of the intra-arterial injection of the short-acting barbiturates amobarbital (Amytal) and lidocaine (Xylocaine) through a microcatheter. Amytal blocks neuronal activity, and Xylocaine blocks axonal conduction [53, 54]. Therefore, a positive Amytal or Xylocaine test (i.e., more than 50% decrease in SEP amplitude and/or mMEP disappearance) indicates that the vessel distal to the tip of the microcatheter supplies the functional grey or white matter of the spinal cord, respectively (see Fig. 6.2). Identification of the ASA and PSAs supplying normal spinal cord below and above the AVM Identification of the angioarchitecture of the AVM

Selective catheterization of the pedicle artery

Superselective catheterization of the ASA or PSA feeding the AVM

Provocative tests

—

Embolization

+ No embolization from that specific catheter position

More selective catheterization or attempts to embolize through other feeding vessels

+

Provocative tests

—

Embolization

+ Abandoned embolization

FIGURE 6.3 Protocol for spinal endovascular procedures using provocative tests (see text for details).

Neurophysiological Monitoring During Endovascular Procedures

131

If a provocative test is positive, liquid embolization from that catheter position should not be performed. Instead, a superselective angiogram from the microcatheter should be carefully reviewed to assess the following possibilities: (1) to further advance the microcatheter distally to bypass the vessel supplying the normal spinal cord; (2) to protect the normal spinal cord supply before the liquid embolization, as shown in Fig. 6.7; or (3) to use particulate embolic agents. If none of these is feasible, embolization should be attempted from other feeders or completely aborted [48] (Fig. 6.7). In rare circumstances provocative tests may overestimate the effects of endovascular obliteration [55] because of local hemodynamic or nonselective catheterization of the AVM feeder. This is the case when the sizes of the embolizing material, such as particles or coils, are bigger than the diameter of the small arteries feeding the normal spinal cord. In this case Xylocaine, which is liquid, could easily diffuse through the vascular tree into those vessels feeding the AVM as much as in those feeding the normal spinal cord, and provocative tests will be positive. However, relying on the different diameter between the enlarged vessels of the AVM and the small feeders to the normal cord, in selected cases it is still possible to safely proceed with the embolization. A nonselective catheterization of the vessels feeding the malformation could also induce a spreading of the provocative drug to the normal spinal cord and therefore give a positive test result. Overall, however, these tests have proven to be very useful in enhancing the safety of endovascular procedures for spine and spinal cord vascular lesions [48, 56–58]. We have performed more than 30 provocative tests with Amytal and Xylocaine during endovascular procedures in the spinal cord. In our experience, Xylocaine tests are more often positive than are Amytal tests. This is most likely due to the different site of action of the two drugs. Xylocaine acts mainly on axonal conduction, whereas Amytal switches off neuronal activities without suppressing axonal conduction [53, 54]. Nevertheless, since these drugs test different pathways of the spinal cord, both should be used in every patient undergoing a spinal embolization. It is also noteworthy that we had no cases in which both SEPs and mMEPs were affected simultaneously after either Amytal or Xylocaine injection (Table 6.1). This suggests that to monitor only SEPs or only mMEPs would expose the patient to the risk of neurological deficits [57]. Although, in our series, provocative tests affected mMEPs more than SEPs, mMEPs should be used as a complement rather than as an alternative to SEP monitoring. In fact, the possibility of selective sensory deficits with preserved motor function during removal of a spinal AVM and its correlation with neurophysiological tests has been described [59]. The need for multimodal (mMEPs and SEPs) neurophysiological monitoring comes also from the observation that there is no correlation between the

132

Francesco Sala and Yasunari Niimi

TABLE 6.1 Results of Provocative Tests

N° of tests

Positive tests

mMEPs disappearance

31 33

1 (3%) 9 (27%)

1 7

Amytal Xylocaine

SEPs drop >50%

mMEPs disappearance & SEPs drop >50%

0 2

0 0

TABLE 6.2 Correlation between Positive Provocative Tests and Evoked Potentials

Positive tests Amytal Xylocaine

1 (3%) 9 (27%)

Vessel PSA PSA PSA PSA ASA ASA ASA ASA ASA PICA

Intraoperatively changed EP mMEPs mMEPs mMEPs SEPs mMEPs mMEPs mMEPs mMEPs SEPs mMEPs

Legend: EP = evoked potentials; PSA = posterior spinal artery; ASA = anterior spinal artery.

vascular compartment where the provocative test is performed (ASA or PSA) and the neurophysiological modality that is affected (SEPs or mMEPs) (Table 6.2). The unpredictability of provocative tests when Amytal or Xylocaine is injected in the ASA and/or PSA supports the existence of vascular anastomoses and the variability of the spinal cord flow dynamic, the latter being even more complex in the presence of an AVM. Touho et al. described motor, but not sensory, deficits as a result of Xylocaine injection in the ASA [58]. Vice-versa, we have experienced one case where injection of Amytal or Xylocaine in the ASA caused loss of SEPs while mMEPs remained unchanged. Therefore, we routinely monitor SEPs as well as mMEPs even if the endovascular procedure is limited to the ASA territory. Because any interventional procedure may acutely modify the local hemodynamic, it is also critical to repeat provocative tests for both SEPs and MEPs before each embolization procedure. We performed neurophysiological monitoring during 110 endovascular procedures in 87 patients who were treated for spine and/or spinal cord lesions between 1996 and 1999. In terms of neurological outcome, two patients woke up with a paraparesis, one moderate and one severe, that was not present before the procedure. Paradoxically, both of them belong to a group of 59 patients who were considered at low risk because they harbored extramedullary vascular lesions that

Neurophysiological Monitoring During Endovascular Procedures

133

were not in close vascular relationship with the spinal cord. In both these patients, however, provocative tests were not done, and neurophysiological monitoring could only document the disappearance of mMEPs from the lower extremities and anticipate the neurological motor deficit. Conversely, whenever provocative tests were performed, either in the low-risk patients or in those harboring intramedullary AVMs or fistulas, neurological morbidity was never documented. On the basis of our experience, we believe that our provocative test protocol helps decrease the risk of spinal cord ischemia during endovascular procedures. Since it is unsafe and unethical to expose patients to the risk of embolization after a positive test (with the few exceptions already mentioned), controlled trials are unlikely to occur. Accordingly, the ability to superselectively catheterize vessels feeding the vascular malformation is of paramount importance. In the last sections of this chapter we will focus on the classification of spinal vascular lesions, their angiographic features, and principles of endovascular treatment. Those aspects relevant to neurophysiological monitoring will be emphasized.

3 ENDOVASCULAR TREATMENT OF VASCULAR MALFORMATIONS AND TUMORS OF THE SPINE AND THE SPINAL CORD

3.1 INDICATIONS All vascular lesions of the spine, spinal cord, and surrounding tissues may be candidates for endovascular embolization. The indications include preoperative treatment to decrease vascularity (and therefore intraoperative blood loss), palliation for incurable diseases, and curative therapy when embolization alone is curative. Various embolic agents are used, depending on the purpose of the treatment and the nature of the disease. We will discuss endovascular embolization of these lesions and the implication of neurophysiological monitoring.

3.2 ANGIOGRAPHIC VASCULAR ANATOMY OF THE SPINE AND SPINAL CORD The vascular supply to the spine and paraspinal musculature arises from the main trunk of the intercostal or lumbar artery as well as the dorsospinal artery [60]. Vascular supply to the spinal dura and spinal cord is derived from the ventral division of the dorsospinal artery. There are rich longitudinal and transverse anastomoses between the adjacent segmental arteries. Longitudinal anastomotic

134

Francesco Sala and Yasunari Niimi

vessels connect branches of the segmental arteries to adjacent branches above and below. Transverse anastomotic vessels connect right and left segmental arteries across the midline. Both longitudinal and transverse anastomotic vessels can be outside or inside the spinal canal. Nerve roots and the spinal dura are supplied by the radicular artery arising from each segmental artery. Progressing caudally from the intercostal to the lumbar levels, there is increasing obliquity of both the nerve roots and the radicular arteries because of differences in the growth rate between the spine and spinal cord. If a radicular artery supplies the ASA, it is called a radiculomedullary artery, and if it supplies the PSA, it is called a radiculopial artery. A radiculomedullary artery and a radiculopial artery may have a common trunk. Because the segmental arrangement during embryonic development, particularly in the cervical region, several vessels must be angiographically evaluated to delineate the vascular supply of the spine and spinal cord. At the cervical level, the ascending cervical artery, the vertebral artery, and the deep cervical artery on both sides must be evaluated. Additionally, at the C1–C2 levels, the ascending pharyngeal and occipital arteries should also be studied. For the thoracolumbar levels, angiographic evaluation of the bilateral supreme intercostal, intercostal, and lumbar arteries is indicated. At the sacral level, bilateral lateral sacral and iliolumbar arteries arising from the internal iliac artery as well as median sacral artery may supply the sacral nerve roots, spinal cord, vertebrae, and parasacral musculature. As previously mentioned, the spinal cord derives its vascular supply from one anterior midline ASA and two posterolateral paramedian PSAs [61]. Angiographically, the radiculomedullary artery has a characteristic hairpin configuration that continues to the ASA, which appears as a midline continuous longitudinal straight vessel. The ascending limb of the ASA may be opacified from a large radiculomedullary artery. The ASA continues caudally to the filum terminale and forms anastomoses with bilateral PSAs at the level of the conus [1]. The PSA appears as a relatively small paramedian longitudinal straight vessel. The PSA axis is smaller and discontinuous compared to the ASA axis. The radiculopial artery also forms a hairpin configuration that has a more acute angle than the radiculomedullary artery because of its paramedian location. The venous drainage of the spinal cord is characterized by rich intra- and perimedullary anastomoses. The perimedullary veins form a rich venous plexus as well as two longitudinal collector veins in the midline on the anterior and posterior surfaces of the spinal cord (the anterior and posterior median spinal veins). These perimedullary veins can be angiographically opacified by injection of a large radiculomedullary artery. They drain to the extradural internal vertebral plexus via the radicular veins and then to the paravertebral veins [62–64].

Neurophysiological Monitoring During Endovascular Procedures

135

3.3 SPINAL ANGIOGRAPHY Spinal angiography and subsequent endovascular treatment are best performed under general anesthesia. This not only provides the patients with comfort, but also allows for extended periods of apnea (up to 40 s) and thus provides the opportunity to obtain high-resolution images necessary to identify small spinal cord arteries and to evaluate slow flow lesions. For lower thoracic and lumbar lesions, glucagon may also be administered to limit bowel motion which degrades the image quality. It is important to identify spinal cord arteries and differentiate them from surrounding extradural vessels supplying osteomuscular structures. In order to identify the midline ASA and paramedian PSAs, it is mandatory to perform spinal angiography in the exact posterior–anterior projection. This is sometimes impossible because of distortion of the spinal column and cord from previous treatment or a pathology, either associated with or unrelated to the target lesion. In such a case, oblique and lateral views may be necessary to identify the spinal cord vessels. Pretherapeutic spinal angiography should evaluate the vascular anatomy of both the pathology and normal surrounding spinal cord. More details of our angiographic protocol will be discussed under each disease category.

3.4 ANGIOGRAPHIC EVALUATION, ENDOVASCULAR TREATMENT, AND CLINICAL ASPECT OF NEUROPHYSIOLOGICAL MONITORING 3.4.1 Tumors Vascular tumors are classified as benign or malignant. Malignant tumors can be further classified as primary or metastatic. For the purpose of endovascular embolization, it is also useful to classify tumors as extramedullary or intramedullary. Intramedullary tumors are supplied by the spinal cord arteries, and their embolization carries higher risk than extramedullary tumors which are not supplied by the spinal cord arteries. Hemangioblastomas are the only intramedullary tumors that are candidates for embolization. These tumors are benign but usually hypervascular and are embolized primarily as a preoperative procedure depending on the vascularity, location, and size of the feeding vessels of the tumor. Angiographic assessment and endovascular treatment are performed in a manner similar to that used for intradural spinal cord vascular malformations. Intradural extramedullary tumors such as meningiomas and neurinomas are not usually very vascular and embolization is typically not indicated. In contrast, vascular tumors located in the extradural or paraspinal compartments

136

Francesco Sala and Yasunari Niimi

are frequently amenable to embolization as a preoperative or palliative measure. These tumors include benign lesions such as hemangiomas, giant-cell tumors, and aneurysmal bone cysts. Embolization may also be performed for vascular extradural malignancies such as primary sarcomas, plasmacytomas, hemangiopericytomas, and metastatic carcinomas such as those from the breast, thyroid, kidney, and stomach. We usually use particles such as polyvinyl alcohol (PVA) particles as the primary embolic agent. Coils are used to protect normal territory from inadvertent embolization. Liquid adhesive, such as n-butyl cyanoacrylate (NBCA), is not routinely used for tumors because of its higher potential risk of penetration into the spinal cord artery. It may, however, be used in highly vascular tumors to obtain a better occlusive effect. For palliative embolization of malignant tumors, ethanol may also be used as an embolic agent and results in a longlasting effect because of its cytotoxicity. Pretherapeutic angiographic assessment should address the exact location, size, and configuration of the lesion. The feeding arteries and draining veins of the lesion as well as the associated vascular dynamics (i.e., presence of arteriovenous shunting) should be assessed. It is also important to determine if there is ASA or PSA contribution at or near the level of the lesion. For this purpose, angiographic evaluation should include not only the assessment of bilateral segmental arteries at the level of the lesion, but also at least two levels above and below the lesion. As previously discussed, in the case of distortion of the spine or spinal cord from the previous treatment, the disease itself, or the existence of overlapping metallic stabilization instruments, oblique and lateral views may be helpful to identify spinal cord arteries [65, 66]. The main purpose of neurophysiological monitoring for spinal or paraspinal tumor embolization is to detect masked and unrecognized spinal cord arteries. If the existence of a spinal cord artery is suspected but uncertain, provocative testing may be performed by observing changes in SEPs or MEPs after injection of sodium Amytal and Xylocaine from a microcatheter placed within the feeding vessel of the tumor. If any changes are noted in either SEPs or MEPs, aggressive embolization from that catheter position should be avoided. Also, if a significant change in SEPs or MEPs occurs during an embolization procedure, spinal cord ischemia should be suspected and the procedure should be terminated to minimize the risk of permanent damage and maximize the possibility of recovery of the spinal cord. Improvement of SEPs or MEPs after tumor embolization is sometimes observed in association with clinical improvement. This phenomenon most often occurs in a tumor with epidural extension and spinal cord compression and is probably due to decreased mass effect secondary to tumor devascularization. This improvement is an indicator of effective embolization either as a preoperative or a palliative treatment. An example of preoperative embolization for a cervical spine tumor is presented in Fig. 6.4.

Neurophysiological Monitoring During Endovascular Procedures

137

3.4.2 Vascular Malformations Vascular malformations can be simply classified into dural/extradural or intradural lesions. This distinction is important because the risk of embolization and the role of neurophysiological monitoring are significantly different for these two categories. 3.4.2.1 Dural/Extradural Lesions Dural/extradural lesions include spinal dural arteriovenous fistulas (SDAVFs), epidural or paraspinal arteriovenous malformations (AVMs) or fistulas (AVFs), and spine AVMs or AVFs. Endovascular embolization is indicated in all of these lesions, usually as a curative treatment but also sometimes for palliative or preoperative therapy. We prefer to use a liquid adhesive, such as NBCA, as an embolic agent because of its ability to penetrate into small vessels and its permanent occlusive effect. For the permanent cure of an AVF, the liquid embolic material should penetrate into the proximal portion of the draining vein through the fistula site [67]. Insufficient penetration of the embolic material frequently results in recanalization of the lesion due to rich collateral vessels. Coils may be used as an adjunct agent in NBCA embolization to protect normal territory or may be used as a primary agent for high-flow AVFs. Particles are not used because their occlusive effect tends to be temporary, resulting in a higher rate of recanalization. The pretherapeutic angiographic protocol for these lesions is similar to that for tumors except that evaluation of the venous drainage is more important. Angiograms of the bilateral segmental arteries should be obtained at the level of the feeders as well as at least two levels above and below the level of the malformation. If there is intradural venous drainage to the perimedullary veins, it is essential to evaluate the circulation time of the ASA. For this purpose, an angiogram of the dominant radiculomedullary artery (frequently Adamkiewicz’s artery) should be obtained by injecting a large amount of contrast material with a long imaging time (e.g., 10 cc of contrast injected at a rate of 1 cc/s, imaging time 40–60 s). If the circulation time of the ASA is prolonged with no opacification of the spinal cord venous drainage, this indicates the existence of spinal cord venous hypertension and explains the most likely etiology of the patient’s neurological deficits. Spinal cord venous hypertension is the underlying cause of almost all SDAVFs, many perimedullary AVFs, and many epidural AVFs with intradural perimedullary venous drainage [30]. The primary role of neurophysiological monitoring for this disease group is to detect a masked spinal cord artery originating from the same pedicle as the feeder to the malformation. If there is an ASA or a PSA originating from the same pedicle as the feeder, endovascular embolization is contraindicated unless very distal advancement of the microcatheter within the feeding vessel can be obtained beyond the origin of the spinal cord artery. Provocative testing by

FIGURE 6.4 A 48-year-old man with a cervical spine hemangiopericytoma. (A) Right dorsocervical artery angiogram on the PA view demonstrating a hypervascular tumor stain in the right C6 hemivertebra. No definite spinal cord artery is identified on this study. (B) Postembolization control angiogram of the right dorsocervical artery demonstrating complete devascularization of the tumor stain with preservation of the anterior spinal artery (arrows). Embolization was performed using polyvinyl alcohol (PVA) particles with assistance of one provocative testing to confirm absence of a spinal cord artery distal to the tip of the microcatheter. (C) Left dorsocervical artery angiogram demonstrating hypervascular tumor stain in the left C6 hemivertebra. (D) Superselective angiogram from a branch of the left dorsocervical artery. The small arrow indicates the tip of the microcatheter, and the large arrow indicates the tip of the guiding catheter. No spinal cord artery is identified on this study. (E) Superselective angiogram of a branch of the left dorsocervical artery during embolization using PVA particles. The medium arrow shows the tip of the microcatheter. There is anastomotic opacification of the anterior spinal artery (small arrows) through the retrocorporeal anastomosis from left to right. Compare to Fig. 6.1B. Provocative test can be performed to confirm the existence of the anterior spinal artery if there is any question. The large arrow indicates the tip of the guiding catheter. (F) Postembolization control angiogram of the left dorsocervical artery demonstrating complete devascularization of the tumor. The patient was operated on 2 days later without significant blood loss and transfusion. 138

Neurophysiological Monitoring During Endovascular Procedures

139

FIGURE 6.4 (Continued)

injecting sodium Amytal and Xylocaine should be performed if there is any suspicion of the existence of an unidentified spinal cord artery from the feeding vessel to the malformation. The necessity of provocative testing in dural/ extradural lesions, however, is exceptional because, compared with tumor cases, identification of a spinal cord artery is easier because of less distortion of the spine and spinal cord and no overlapping abnormal vascularity or metallic devices. Careful analysis of the vascular anatomy of the lesion and the normal spinal cord is far more important than provocative testing. Another role of neurophysiological monitoring is early detection of possible spinal cord ischemia. If significant changes in SEPs or MEPs are detected during embolization, the procedure should be suspended until full recovery of SEPs and MEPs or terminated to minimize the permanent damage to the spinal cord. For certain diseases, neurophysiological monitoring has another promising but not yet proven role: that is, the prediction of functional recovery after embolization. Significant improvement of SEPs or MEPs after embolization is frequently observed in diseases with neurological symptoms secondary to spinal cord venous hypertension, including SDAVFs and spinal epidural fistulas with intradural venous drainage [45]. Reduction of spinal cord venous hypertension after embolization is sometimes associated with improvement of SEPs and MEPs and correlated with improvement of neurological symptoms. Although improvement of SEPs and MEPs after embolization is a favorable prognosticator for satisfactory neurological outcome in our experience, these patients will still

140

Francesco Sala and Yasunari Niimi

require intensive rehabilitation to maximize functional recovery, even if the lesion is angiographically cured. Clinical improvement typically occurs first in motor function, followed by sensory functions, and bladder and bowel functions improve last, if at all. It should be noted that improvement of potentials during neurophysiological monitoring does not promise complete cure of the disease or lasting remission of the symptoms. Symptomatic lesion recurrence is typically associated with deterioration of the neurophysiological findings. Therefore, correlation of neurophysiological improvement (SEPs and MEPs) with angiographic cure of the lesion (i.e., penetration of the embolic material into the venous side) is important for predicting permanent clinical improvement. Further accumulation of cases with detailed analysis and long-term follow-up is necessary to establish the exact role of neurophysiological monitoring as a predictor of functional recovery after endovascular treatment for these diseases. A case of endovascular treatment of an epidural fistula with intradural venous drainage is presented in the video for this chapter (choose Chapter 6 from the accompanying CD main menu). This patient had improvement of MEPs and SEPs after the embolization (Fig. 6.5) with angiographic demonstration of complete cure of the lesion and improvement of spinal cord venous hypertension. The patient experienced immediate clinical improvement after the embolization and was neurologically intact at the 3-month follow-up. 3.4.2.2 Intradural Vascular Malformations Intradural vascular malformations are further classified into spinal cord AVMs or AVFs, telangiectasias, and cavernous malformations. Endovascular embolization is indicated and is the first choice of treatment for spinal cord AVMs and AVFs. Embolization is usually curative for simple AVMs and AVFs, but palliative or rarely curative for complex or extensive AVMs. Palliative embolization is targeted to occlude dangerous structures such as aneurysms or high-flow fistulas and is performed to decrease the risk of hemorrhage or to improve neurological symptoms. Endovascular treatment for this disease group is considered as high risk because embolization is performed through the ASAs or PSAs that supply the normal spinal cord as well as the lesion. A liquid embolic agent, such as NBCA, is preferred for nidus AVMs and small fistulas because of its ability to penetrate distally and cause permanent occlusion. For large fistulas and associated aneurysms, coils are also effective. Particles are used mainly for small AVMs or AVFs for which distal catheterization through the feeder is difficult [25, 45, 68–70]. Neurophysiological monitoring, including provocative testing, is most important in this disease category because of the high-risk nature of the treatment. The main role of this monitoring is early detection of spinal cord ischemia by continuous monitoring of SEPs, MEPs, and BCRs, and prediction of the safety of embolization from a certain microcatheter position by provocative testing.

Neurophysiological Monitoring During Endovascular Procedures

141

FIGURE 6.5 A 23-year-old woman presented with progressive bilateral lower extremity weakness, numbness, and bladder and bowel dysfunction. The angiography and embolization procedure of this patient is presented in the video (choose Chapter 6 from the accompanying CD main menu). (A) SEPs from the bilateral posterior tibialis nerve (PTN) before (OP) and after (CL) embolization procedure, demonstrating significant improvement of the latency of the response. (B) MEPs from the bilateral abductor hallucis muscles (AH) before (OP) and after (CL) the embolization procedure, demonstrating significant improvement of the latency of the response. These figures demonstrate neurophysiological evidence of improvement of the spinal cord venous hypertension.

Spinal cord ischemia due to compromised spinal cord vascular supply during the procedure can occur not only by injection of embolic agents but also by catheterization of a feeder, either from blockage of the flow by the catheter itself, or spasm or dissection created by catheter manipulation [47]. An example of early detection of compromised spinal cord vascular supply by MEP monitoring and its treatment is demonstrated in Fig. 6.6.

FIGURE 6.6 A 28-year-old man presented with progressive weakness and numbness of both lower extremities and bladder dysfunction. (A) Left, normal MEPs recorded from the left tibialis anterior (TA) muscle. Right, left T11 intercostal angiogram showing an intramedullary AVM (arrows) supplied by the anterior spinal artery (ASA, arrowheads). (B) Left, disappearance of MEPs from the left TA muscle during superselective catheterization of the ASA. Middle, complete flow arrest in the ASA (arrowheads) distal to the tip of the microcatheter (arrow) was noted. Right, nonsubtracted image demonstrating opened hairpin loop of the radiculomedullary artery (arrowheads) by the microcatheter. (C) Following quick particle embolization of the AVM, the microcatheter was removed with temporary partial improvement of MEPs and flow in the ASA. Left, a few minutes later, the MEPs from the left TA muscle completely disappeared. Right, left T11 angiogram demonstrating no opacification of the ASA due to severe vasospasm. (D) This was treated by superselective infusion of papaverine into the radiculomedullary artery. Left, complete recovery of MEPs after papaverine infusion. Right, complete resolution of vasospasm with opacification of the normal ASA and minimal opacification of the remaining AVM after embolization. Modified from [47].

142

FIGURE 6.6 (Continued). 143

144

Francesco Sala and Yasunari Niimi

FIGURE 6.7 A 46-year-old man presented with progressive paraparesis, urinary and fecal incontinence, and associated lower back pain. (A) Right L2 lumbar artery angiogram demonstrating an intramedullary arteriovenous malformation (AVM) supplied by the posterior spinal artery (PSA, arrowheads). This was superselectively catheterized for embolization. (B) Superselective angiogram of the right PSA showing the AVM. No normal spinal cord supply is identified on this study. The arrow indicates the tip of the microcatheter. Provocative testing from this catheter position was positive with disappearance of SEPs from the right posterior tibialis nerve (PTN). Repeat testing was also positive, and saline injection from the same catheter position did not cause any change in SEPs. See G. (C) Superselective angiogram of the right PSA from the further advanced microcatheter (arrow), demonstrating the distal portion of the AVM as well as normal PSA on both sides (small arrowheads). There is also anastomotic opacification of the ASA (curved arrow) with deviation of its proximal portion (large arrowheads) due to the AVM. (D) Nonsubtracted image demonstrating a microcoil placed to protect the distal normal PSA (arrows). The large arrow indicates the tip of the microcatheter, which was brought back after placement of the microcoil. The arrowheads indicate NBCA cast due to prior embolization from the ASA. (E) Superselective angiogram of the PSA repeated from the catheter position in D. Repeat provocative testing was negative (see G), and embolization using NBCA was performed from this catheter position. (F) Postembolization control angiogram of the right T11 intercostal artery demonstrating small residual nidus of the AVM. The right PSA distal to the AVM (arrows) is supplied by anastomotic vessels (arrowheads) from the left PSA. No change in SEPs was seen after the embolization. See G. (G) Trace of the provocative testing using Xylocaine and SEPs from the right PTN. The patient was neurologically unchanged after the embolization. Modified from [48].

Neurophysiological Monitoring During Endovascular Procedures

FIGURE 6.7 (Continued)

145

146

Francesco Sala and Yasunari Niimi

FIGURE 6.7 (Continued)

Neurophysiological Monitoring During Endovascular Procedures

147

In the rare situation in which the vascular anatomy of the lesion is so complicated that it is difficult to differentiate feeders to the malformation from the normal spinal cord supply, provocative testing can be used as an aid for analysis of the vascular anatomy. It should be emphasized, however, that the availability of provocative testing does not decrease the importance of precise angiographic analysis of the vascular anatomy of the malformation and surrounding normal spinal cord. The correlation between clinical improvement and improvement of MEPs or SEPs after embolization of intradural AVFs and AVMs remains to be elucidated [48], and further accumulation of experience is needed. Figure 6.7 shows an example of a spinal cord AVM embolization using neurophysiological monitoring and provocative testing.

4 CONCLUSIONS To date, neurophysiological monitoring is feasible in the great majority of patients undergoing endovascular procedures for spine or spinal cord lesions. Muscle MEPs and SEPs retain their own specificity in assessing the functional integrity of motor and sensory pathways, respectively. To rely solely on either one of these monitoring modalities is not supported from a scientific background or justified from a clinical perspective. Provocative tests with both Amytal and Xylocaine are mandatory in selecting those patients amenable to a safe embolization. Neurophysiological monitoring during endovascular procedures offers a unique opportunity to investigate the spinal cord hemodynamic and to integrate functional, anatomical, and clinical data. Included with the accompanying CD is a video showing an angioembolization of an epidural fistula of the spinal cord (choose Chapter 6 from the main CD menu).

REFERENCES 1. Lazorthes, G.A., Gouaze, A., Zadeh, O., Santini, J.J., Lazorthes, Y., and Burdin, P.H. (1971). Arterial vascularization of the spinal cord: Recent studies of the anastomotic substitution pathways. J. Neurosurg., 35, 253–262. 2. Lazorthes, G.A., Gouaze, and Djindjian, R. (1973). “Vascularisation et circulation de la moelle épinière.” Masson, Paris. 3. Djindjian, M., Ribeiro, A., Ortega, E., Gaston, A., and Poirier, J. (1988). The normal vascularization of the intradural filum terminale in man. Surg. Radiol. Anat., 10(3), 201–209. 4. Krauss, W.E. (1999). Vascular anatomy of the spinal cord. Neurosurg. Clin. North. Am., 10(1), 9–15. 5. Cohen, A.R., Young, W., and Ransohoff, J. (1981). Intraspinal localization of the somatosensory evoked potential. Neurosurgery, 9(2), 157–162. 6. Herren, R., and Alexander, L. (1939). Sulcal and intrinsic blood vessels in human spinal cord. AMA Arch. Neurol. Psychiatry, 41, 678–687.

148

Francesco Sala and Yasunari Niimi

7. Rosenwasser, R.H., and Armonda, R. (1999). Spinal vascular malformations: Normal anatomy, diagnostic angiography, and angiographic classification. In “Spinal vascular malformations” (D.L. Barrow, and I.A. Awad, eds.), pp. 23–35. The American Association of Neurological Surgeons, Park Ridge, Illinois. 8. Oldfield, E.H., and Doppman, J.L. (1993). Spinal arterial malformations. Clin. Neurosurg., 40, 161–183. 9. Foix, C., and Alajouanine, T. (1926). La myélite necrotique subaique. Rev. Neurol., 2, 1–42. 10. Konrad, P.E., Tacker, W.A. Jr., Cook, J.R., and Hood, D.L. (1987). Motor evoked potentials in the dog: Effects of global ischemia on spinal cord and peripheral nerve signals. Neurosurgery, 20, 117–124. 11. Kai, Y., Owen, J.H, Allen, B.T., Dobras, M., and Davis, C. (1994). Relationship between evoked potentials and clinical status in spinal cord ischemia. Spine, 19(10), 1162–1168. 12. Laschinger, J.C., Owen, J., Rosenbloom, M., Cox, J.L., and Kouchoukos, N.T. (1988). Direct non-invasive monitoring of spinal cord motor function during thoracic aortic occlusion: Use of motor evoked potentials. J. Vasc. Surg., 7, 161–171. 13. Reuter, D.G., Tacker, W.A. Jr., Badylak, S.F., Voorhees, W.D. III, and Konrad, P.E. (1992). Correlation of motor-evoked potential response to ischemic spinal cord damage. J. Thorac. Cardiovasc. Surg., 104, 262–272. 14. de Haan, P., Kalkman, C.J., Ubags, L.H., Jacobs, M.J., and Drummond, J.C. (1996). A comparison of the sensitivity of epidural and myogenic transcranial motor-evoked responses in the detection of acute spinal cord ischemia in the rabbit. Anesth. Analg., 83(5), 1022–1027. 15. Kobrine, A.I., Evans, D.E., and Rizzoli, H. (1979). The effects of ischemia on long-tract neural conduction in the spinal cord. J. Neurosurg., 50, 639–644. 16. Elmore, J.R., Gloviczki, P., Harper, C.M., Pairolero, P.C., Murray, M.J., Bourchier, R.G., Bower, T.C., and Daube, J.R. (1991). Failure of motor evoked potentials to predict neurologic outcome in experimental thoracic aortic occlusion. J. Vasc. Surg., 14(2), 131–139. 17. Robertazzi, R.R., and Cunningham, J.N. (1998). Monitoring of somatosensory evoked potentials: A primer on the intraoperative detection of spinal cord ischemia during aortic reconstructive surgery. Semin. Thorac. Cardiovasc. Surg., 10(1), 11–17. 18. Cunningham, J.N., Laschinger, J.C., and Spencer, F.C. (1987). Monitoring of somatosensory evoked potentials during surgical procedures on the thoracoabdominal aorta: IV. Clinical observations and results. J. Thorac. Cardiovasc. Surg., 94(2), 275–285. 19. Machida, M., Weinstein, S.L., Yamada, T., Kimura, J., and Toriyama, S. (1988). Dissociation of muscle action potentials and spinal somatosensory evoked potentials after ischemic damage of spinal cord. Spine, 13(10), 1119–1124. 20. de Haan, P., Kalkman, C.J., de Mol, B.A., Ubags, L.H, Veldman, D.J., and Jacobs, M.J. (1997). Efficacy of transcranial motor-evoked myogenic potentials to detect spinal cord ischemia during operations for thoracoabdominal aneurysms. J. Thorac. Cardiovasc. Surg., 113(1), 87–101. 21. de Haan, P., Kalkman, C.J., and Jacobs, M.J. (1998). Spinal cord monitoring with myogenic motor evoked potentials: Early detection of spinal cord ischemia as an integral part of spinal cord protective strategies during thoracoabdominal aneurysm surgery. Semin. Thorac. Cardiovasc. Surg., 10(1), 19–24. 22. Breslau, J., and Eskridge, J.M. (1995). Preoperative embolization of spinal tumors. J. Vasc. Interv. Radiol., 6(6), 871–875. 23. Broaddus, W.C., Grady, M.S., Delashaw, J.B. Jr., Ferguson, R.D., and Jane, J.A. (1990). Preoperative superselective arteriolar embolization: A new approach to enhance resectability of spinal tumors. Neurosurgery, 27(5), 755–759. 24. Meisel, H.J., Lasjaunias, P., and Brock, M. (1995). Modern management of spinal and spinal cord vascular lesions. Minim. Invasive Neurosurg., 38(4), 138–145.

Neurophysiological Monitoring During Endovascular Procedures

149

25. Niimi, Y., and Berenstein, A. (1999). Endovascular treatment of spinal vascular malformations. Neurosurg. Clin. North. Am., 10(1), 47–70. 26. Niimi, Y., Berenstein, A., Setton, A., and Neophytides, A. (1997). Embolization of spinal dural arteriovenous fistulae: Results and follow-up. Neurosurgery, 40(4), 675–683. 27. Niimi, Y., Setton, A., and Berenstein, A. (1999). Spinal dural arteriovenous fistulae draining to the anterior spinal vein: Angiographic diagnosis. Neurosurgery, 44(5), 999–1004. 28. Smith, T.P., Gray, L., Weinstein, J.N., Richardson, W.J., and Payne, C.S. (1995). Preoperative transarterial embolization of spinal column neoplasms. J. Vasc. Interv. Radiol., 6(6), 863–869. 29. Vazquez, A.V., Botella, C., Beltran, A., Solera, M., and Piquer, J. (1997). Preoperative embolization of solid cervicomedullary junction hemangioblastomas: Report on two cases. Neuroradiology, 39(2), 86–89. 30. Berenstein, A., and Lasjaunias, P. (1992). Spine and spinal cord vascular lesions. In “Surgical neuroangiography: Endovascular treatment of spine and spinal cord lesions” (A. Berenstein, and P. Lasjaunias, eds.), pp. 1–109. Springer, Berlin, Heidelberg. 31. Cloft, H.J., Jensen, M.E., Do, H.M., and Kallmes, D.F. (1999). Spinal cord infarction complicating embolization of vertebral metastasis: A result of masking of a spinal artery by a high-flow lesion. Interventional Neuroradiology, 5, 61–65. 32. Halbach, V.V., Higashida, R.T., Dowd, C.F., Barnwell, S.L., and Hieshima, G.B. (1991). Management of vascular perforations that occur during neurointerventional procedures. Am. J. Neuroradiol., 12, 319–327. 33. Merland, J.J., and Reizine, D. (1987). Treatment of arteriovenous spinal cord malformations. Semin. Intervent. Radiol., 4, 281–290. 34. Ben Daviv, B., Taylor, P.D., and Haller, G.S. (1987). Posterior spinal fusion complicated by posterior column injury: A case report of a false-negative wake-up test. Spine, 12(6), 540–543. 35. Padberg, A.M., Wilson-Holden, T.S., Lenke, L.G., and Bridwell, K.H. (1998). Somatosensory and motor evoked potential monitoring without a wake up test during idiopathic scoliosis surgery: An accepted standard of care. Spine, 23(12), 1392–1400. 36. Berenstein, A., Young, W., Ransohoff, J., Benjamin, V., and Merkin, H. (1984). Somatosensory evoked potentials during spinal angiography and therapeutic transvascular embolization. J. Neurosurg., 60, 777–785. 37. Hacke, W. (1989). Evoked potentials monitoring in interventional neuroradiology. In “Neuromonitoring in surgery” ( J.E. Desmedt, ed.), pp. 331–342. Elsevier, Amsterdam. 38. Young, W., and Mollin, D. (1989). Intraoperative somatosensory evoked potentials monitoring of spinal surgery. In “Neuromonitoring in surgery” ( J.E. Desmedt, ed.), pp. 165–173. Elsevier Science, New York. 39. Ginsburg, H.H., Shetter, A.G., and Raudzens, P.A. (1985). Postoperative paraplegia with preserved intraoperative somatosensory evoked potentials. Case report. J. Neurosurg., 63, 296–300. 40. Lesser, R.P., Raudzens, P., Luders, H., Nuwer, M.R., Goldie, W.D., Morris, H.H. III, Dinner, D.S., Klem, G., Hahn, J.F., Shetter, A.G., et al. (1986). Postoperative neurological deficits may occur despite unchanged intraoperative somatosensory evoked potentials. Ann. Neurol., 19, 22–25. 41. Takaki, O., and Okumura, F. (1985). Application and limitation of somatosensory evoked potential monitoring during thoracic aortic aneurysm surgery: A case report. Anesthesiology, 63(6), 700–703. 42. Zornow, M.K., Grafe, M.R., Tybor, C., and Swenson, M.R. (1990). Preservation of evoked potentials in a case of anterior spinal artery syndrome. Electroencephalogr. Clin. Neurophysiol., 77, 137–139. 43. Jones, S.J., Harrison, R., Koh, K.H., Mendoza, N., and Crockard, H.A. (1996). Motor evoked potential monitoring during spinal surgery: Responses of distal limb muscles to transcranial cortical stimulation with pulse trains. Electroencephalogr. Clin. Neurophysiol., 100, 375–383.

150

Francesco Sala and Yasunari Niimi

44. Pechstein, U., Cedzich, C., Nadstawek, J., and Schramm, J. (1996). Transcranial highfrequency repetitive electrical stimulation of recording myogenic motor evoked potential with the patient under general anesthesia. Neurosurgery, 39, 335–344. 45. Kothbauer, K., Pryor, J.C., Bernstein, A., Setton, A., and Deletis, V. (1998). Motor evoked potentials predicting early recovery from paraparesis after embolization of a spinal dural arteriovenous fistula. Interventional Neuroradiology, 4, 81–84. 46. Katayama, Y., Tsubokawa, T., Hirayama, T., Himi, K., Koyama, S., and Yamamoto, T. (1991). Embolization of intramedullary spinal arteriovenous malformation fed by the anterior spinal artery with monitoring of the corticospinal motor evoked potential. Case report. Neurol. Med. Chir., 31, 401–405. 47. Sala, F., Niimi, Y., Krzˇan, M.J., Berenstein, A., and Deletis, V. (1999). Embolization of a spinal arteriovenous malformation: Correlation between motor evoked potentials and angiographic findings. Technical case report. Neurosurgery, 45(4), 932–938. 48. Sala, F., Niimi, Y., Krzˇan, M.J., Berenstein, A., and Deletis, V. (2000). Role of multimodality intraoperative neurophysiological monitoring during embolization of a spinal cord arteriovenous malformation: A paradigmatic case. Interventional Neuroradiology, 6, 223–234. 49. Deletis, V., and Kothbauer, K. (1998). Intraoperative neurophysiology of the corticospinal tract. In “Spinal cord monitoring” (E. Stalberg, H.S. Sharma, and Y. Olsson, eds.), pp. 421–444. Springer, Wien, New York. 50. Kothbauer, K., Deletis, V., and Epstein, F. (1998). Motor evoked potential monitoring for intramedullary spinal cord tumor surgery: Correlation of clinical and neurophysiological data in a series of 100 consecutive procedures. Neurosurg. Focus (electronic journal), (4), Article 1 (http://www.aans.org/journals/online_j/may98/4-5-1). 51. Deletis, V., and Vodusˇek, D.B. (1997). Intraoperative recording of the bulbocavernosus reflex. Neurosurgery, 40, 88–93. 52. Doppman, J.L., Girton, M., and Oldfield, E.H. (1986). Spinal Wada test. Radiology, 161, 319–321. 53. Tanaka, K., and Yamasaki, M. (1966). Blocking of cortical inhibitory synapses by intravenous lidocaine. Nature, 209, 207–208. 54. Terada, T., and Nakai, K., et al. (1991). The differential action of lidocaine and Amytal on the central nervous system as the provocative test drug. In “Proceeding of 7th annual meeting of Japanese Society of Intravascular Neurosurgery,” pp. 95–101. Mie, Japan. 55. Katsuta, T., Morioka, T., Hasuo, K., Miyahara, S., Fukui, M., and Masuda, K. (1993). Discrepancy between provocative test and clinical results following endovascular obliteration of spinal arteriovenous malformation. Surg. Neurol., 40, 142–145. 56. Sadato, A., Taki, W., Nakahara, I., Nishi, S., Yamashita, K., Matsumoto, K., Tanaka, M., and Kikuchi, H. (1994). Improved provocative test for the embolization of arteriovenous malformations: Technical note. Neurol. Med. Chir., 34, 187–190. 57. Sala, F., Niimi, Y., Berenstein, A., and Deletis, V. (2001). Neuroprotective role of neurophysiological monitoring during endovascular procedures in the spinal cord. Ann. N.Y. Acad. Sci., 939, 126–136. 58. Touho, H., Karasawa, J., Ohnishi, H., Yamada, K., Ito, M., and Kinoshita, A. (1994). Intravascular treatment of spinal arteriovenous malformations using a microcatheter. With special reference to serial Xylocaine tests and intravascular pressure monitoring. Surg. Neurol., 42, 148–156. 59. Kalkman, C.J., Drummond, J.C., and Hoi, S.U. (1994). Severe sensory deficits with preserved motor function after removal of a spinal arteriovenous malformation: Correlation with simultaneously recorded somatosensory and motor evoked potentials. Anesth. Analg., 78, 165–168. 60. Chiras, J., Morvan, G., and Merland, J. (1979). The angiographic appearance of the normal intercostal and lumbar arteries: Analysis of the anatomic correlation of the lateral branches. J. Neuroradiol., 6, 169–196.

FIGURE 6.1 Schematic representation of spinal vascular anatomy and its relationship with long tracts involved in the generation of somatosensory and motor evoked potentials. 1. Posterior spinal arteries. 2. Posterior spinal vein. 3. Anterior spinal artery. 4. Anterior spinal vein. 5. Spinal ventral roots. 6. Anterior corticospinal tracts. 7. Lateral corticospinal tracts. 8. Dorsal columns. (Modified from Nieuwenhuys, R., Voogd, J., and van Huijzen, C. (1988). The human central nervous system: A synopsis and atlas, rev. ed. 3. Springer Verlag, Berlin).

CHAPTER

7

Intraoperative Neurophysiological Mapping of the Spinal Cord’s Dorsal Columns MATEVZˇ J. KRZˇ AN Department of Neurology, Children’s Hospital, University Medical Center, Ljubljana, Slovenia

1 Introduction 1.1 Neurophysiological Generators of SEPs in the Spinal Cord 2 Methods of Intraoperative Recording of SEPs with Miniature Electrodes 2.1 Recording of SEPs with a Miniature Multielectrode 3 Results 4 Discussion 5 Conclusion References

ABSTRACT Intramedullary lesions distort the anatomical features of the spinal cord dorsum, making it difficult for the surgeon to perform a myelotomy precisely at the midline. Myelotomies not performed at the midline may damage the dorsal columns. To help establish neurophysiological landmarks on the dorsal cord surface to compensate for distorted anatomy and provide reliable guidance, a highly selective miniature mapping multielectrode was placed on the exposed dorsal cord surface in patients undergoing resection of intramedullary lesions. This miniature electrode records somatosensory evoked potentials (SEPs) after tibial and median nerve stimulation. The electrode consists of eight parallel wires spaced 1 mm apart, each having a diameter of 76 µm and an exposed recording surface of 2 mm. After each of the tibial nerves at the ankles was electrically stimulated, SEPs were recorded with the miniature electrode separately from each of the eight parallel recording sites with a reference needle electrode placed in nearby muscle. Recordings were obtained in 55

Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

153

154

Matevˇz J. Krˇzan patients and revealed an amplitude gradient across the eight recording sites with maximum amplitude toward the midline and decreasing amplitude toward the dorsal root entry zone. SEPs consisted of multispike activity lasting about 10 ms with mean amplitudes ranging from 0.7 to 43 µV. SEPs recorded over the cervical spinal cord after median nerve stimulation showed an amplitude gradient as well, but in the opposite direction of the tibial nerve SEPs. For median nerve SEPs, the site with the highest amplitude was always lateral to the site with the highest amplitude of SEPs after tibial nerve stimulation. The two recording sites with highest-amplitude SEPs after stimulation of either left or right tibial nerve identified the neurophysiological midline between the dorsal columns. In the patients in whom the anatomy had not been distorted, the neurophysiological midline corresponded with the anatomical one.

1 INTRODUCTION Lesions to nervous structures induced during surgery can be avoided using intraoperative neurophysiological techniques [1]. Mapping techniques, in contrast to those applied for continuous monitoring, are used to neurophysiologically locate specific structures in the nervous system but not to continuously monitor their functional integrity. They have been applied to map different parts of the human nervous system at critical periods of surgical procedures, usually before incision placement or lesioning, in order to confirm anatomical structures or to give guidance when anatomical landmarks are distorted or nonexistent [2]. Some of the more frequently applied intraoperative mapping techniques include phase reversal of the median nerve somatosensory evoked potentials (SEPs) [3] and exposed motor cortex stimulation [4] to locate the primary sensory and motor cortices, mapping of the cranial nerve motor nuclei within the brainstem [5], mapping of the dorsal root entry zone (DREZ) in the spinal cord [6], mapping of pudendal afferents and lumbosacral efferents [2, 7], and brachial plexus mapping [8]. Spinal intramedullary lesions are surgically approached by entering the spinal cord in the midline at the posterior median sulcus between the left and right dorsal columns. Anatomical features of the exposed spinal cord can often be distorted, making it difficult for the surgeon to choose the optimal incision placement. Severing dorsal column pathways can lead to serious postoperative neurological impairment (e.g., ataxia or sensory loss). With continuous intraoperative recording of SEPs, such lesioning can be documented but not prevented [2, 9]. In a series of 35 patients undergoing resection of intramedullary lesions at the Institute for Neurology and Neurosurgery, Beth Israel Hospital in New York, SEP recordings were lost early in the course of resection in 20 (57%), indicating a lesion of the dorsal columns during placement of the initial incision into the spinal cord (myelotomy) [10].

Intraoperative Neurophysiological Mapping of the Spinal Cord’s Dorsal Columns

155

1.1 NEUROPHYSIOLOGICAL GENERATORS OF SEPS IN THE SPINAL CORD On the surface of the spinal cord, conducted and segmental evoked potentials can be easily distinguished after stimulation of peripheral nerves or roots [11].

1.1.1 Conducted Potentials Conducted spinal evoked potentials recorded from the dorsal pial cord surface consist primarily of negative waves reflecting the compound action potentials traveling in the dorsal column fibers. Their amplitudes diminish from caudal to rostral recording sites due to dispersion of the afferent volleys caused by different conduction velocities among various fibers, but are not significantly reduced at higher stimulation frequencies. When recorded from the ventral pial cord surface, their polarity is unchanged, suggesting that a generator is oriented along the longitudinal axis of the cord, with the negative pole placed caudally and the positive pole placed cranially [11–13].

1.1.2 Segmental Potentials Segmental potentials represent the summated activity of cells of the spinal cord’s grey matter with intermixed rootlet activity. They have a maximum amplitude at the levels corresponding to the cord entry of the stimulated nerve fibers and decrease caudally and cranially [14, 15]. Following the electrical stimulus, the afferent compound action potential of the dorsal root fibers is recorded as a fast, predominantly positive wave (P1) on the dorsal surface of the exposed spinal cord. This is predominantly generated by fast-conducting Aβ and Aγ fibers, and its amplitude is not reduced by high-frequency stimulation [15]. A negative wave (N1) representing postsynaptic potentials generated in the dorsal horn neurons of the spinal cord follows. It has a longer duration than the previous one because of repetitive firing and relaying of neurons in the dorsal horns and is reduced by high-frequency stimulation. A low-voltage, slow positive wave (P2) that reflects depolarization of afferent fiber terminals follows [13–15]. When recorded from the ventral cord surface, segmental potentials invert polarity, since their generator dipoles are placed in the sagittal plane [16]. After stimulation of median/ulnar or tibial nerves, the segmental responses are P9/N13 or P17/N22, respectively [14].

156

Matevˇz J. Krˇzan

2 METHODS OF INTRAOPERATIVE RECORDING OF SEPs WITH MINIATURE ELECTRODES Intraoperative recordings from the exposed surface of the human spinal cord were first reported during surgical ablative procedures for relief of chronic pain [6, 11, 16]. Silver ball [11] or stainless steel disc electrodes [17] were used to record spinal responses after peripheral nerve stimulation. Segmental responses were recorded rostrally and caudally to identify the DREZ for lesioning. Jeanmonod et al. [11] also reported intraoperative recordings of conducted potentials with the ball electrode on the dorsal columns rostral to the surgical site. They were used for monitoring in order to avoid lesioning the dorsal columns, but not for mapping purposes [16].

2.1 RECORDING OF SEPS WITH A MINIATURE MULTIELECTRODE In order to be able to offer guidance to surgeons in performing myelotomy in the midline, we attempted to identify neurophysiological features on the exposed dorsal cord surface using conducted potentials. Prior to myelotomy, the surgeon placed a miniature multielectrode on the exposed dorsal cord surface approximately at the midline and according to available anatomical landmarks (choose Chapter 7 from the accompanying CD’s main menu to see a short video of this procedure). This multielectrode is highly selective for recording spinal SEPs from the dorsal surface of the exposed spinal cord [18]. A specially designed miniature multielectrode consisting of 8 parallel Teflon-coated stainless steel wires (with a diameter of 76 µm spaced 1 mm apart) embedded in silastic was used. Each wire was stripped of its coating along a length of 2 mm. The recording wires ran parallel to the long axis of the spinal cord, and the reference needle electrode was placed in nearby muscle. The impedance for all recording surfaces was about 20 kΩ, the filter settings were 50–1700 Hz, and the epoch length was 20 ms. To ensure reproducibility, two sets of 100 sweeps were averaged from each of the eight parallel recording surfaces after stimulation of each tibial nerve at the ankle (or median nerve at the wrist). The stimulus intensity was 40 mA or lower, stimulus duration was 0.2 ms, and the repetition rate of the stimulation was 13.3 Hz. Patients were anesthetized using a continuous intravenous infusion of propofol and fentanyl with addition of N2O. In 65 patients studied, 56 patients had tumors (ependymomas, astrocytomas, vascular tumors, lipomas, metastases, etc.), 7 patients had syrinxes, and 2 patients had inflammatory lesions. The lesion

Intraoperative Neurophysiological Mapping of the Spinal Cord’s Dorsal Columns

157

was located in the cervical spinal cord in 46 patients, 17 patients had lesions in the thoracic spinal cord, and 2 patients had lesions in the lumbar sections of the spinal cord.

3 RESULTS We obtained reproducible spinal SEPs in 55 of the 65 patients, aged 7–66 years, before myelotomy was performed. In 10 patients reproducible spinal SEPs could not be obtained because of anatomical (large lesion protruding to the cord surface, previous surgery producing adhesions), physiological (low amplitude and inconsistent responses, segmental wave contamination at the lumbar level), or technical factors. In patients in whom anatomical factors prevented recordings, the responses could usually be recorded caudally to the site of the lesion. The recordings from each of the eight parallel electrode surfaces resembled conducted spinal SEPs previously detected with the silver ball electrode or conventional epidural electrodes [11, 12, 13]. They consisted of multispike activity lasting about 10 ms with mean amplitudes ranging from 0.7 to 43 µV. An amplitude gradient of SEPs across different electrode recording sites was observed. For tibial nerve SEPs, the maximum amplitude was toward the midline and decreased toward the DREZ. The neurophysiological or functional midline was determined to lie between the two recording sites, with highest SEP amplitudes after stimulation of either left or right tibial nerve (Fig. 7.1) (see also color plate). In one patient we recorded conducted responses using two recording methods: in monopolar fashion (needle electrode versus each of the eight recording surfaces) as well as in bipolar fashion using differential recordings (recording surfaces 1–2, or 2–3, or 3–4, etc.) (Fig. 7.2). In 3 patients, spinal SEPs were obtained with the electrode turned at a right angle so that the recording wires were perpendicular to the cord axis and the direction of the fibers in the dorsal column (orthogonal position) (Fig. 7.3). In these recordings, no amplitude gradient was seen across the eight electrode surfaces. A slight latency shift from the caudal to the cranial electrode enabled the calculation of conduction velocities for some of the most prominent peaks, having a value of 45 m/s. In patients with cervical lesions, we also recorded spinal SEPs after median nerve stimulation. The recorded potentials consisted mainly of segmental responses. An amplitude gradient across eight recording sites was also observed, with highest amplitudes of SEPs laterally, close to the DREZ. Lower amplitudes were observed toward the dorsal midline (Fig. 7.4). This is the opposite of what was observed for recorded SEPs following tibial nerve stimulation.

158 FIGURE 7.1 Dorsal column mapping in a 58-year-old patient with an inflammatory lesion between the C2 and C6 segments of the spinal cord. Spinal SEP responses were obtained from the eight recording sites after left and right tibial nerve stimulation. Two sets of 100 sweeps were averaged. Note the amplitude gradient across the recording surfaces. Maximum amplitude after left-sided stimulation occurred at recording site 6, and right-sided stimulation produced the maximum amplitude at recording site 4. Between the traces is an intraoperative picture taken during the measurement. Above, schematic cross section of the cervical spinal cord showing the approximate position of the recording electrode, with the dorsal column midline under recording site 5 (see also color plate).

FIGURE 7.2 Spinal SEP responses obtained at the eight recording surfaces using monopolar (left) and bipolar (right) montage after stimulation of the same tibial nerve. Note that the highest amplitude in the responses on the left is at the identical recording site as the phase reversal for the responses on the right.

159

160

Matevˇz J. Krˇzan

FIGURE 7.3 Spinal SEP responses obtained at the eight recording sites with the electrode recording surfaces orthogonal to the long spinal cord axis. No amplitude gradient was observed. The enlarged segment of the SEP (between 28 and 32 ms) showed slight changes in latency between recording sites 1 and 8.

4 DISCUSSION Our continuous involvement with intramedullary surgery has prompted us to look for new methods in preventing intraoperative injury to the dorsal columns of the spinal cord. The posterior funiculi of the spinal cord are at high risk for damage during these procedures because the approach to deeper-lying lesions often leads through the dorsal columns. In all patients undergoing spinal cord surgery, routine multimodal neurophysiological monitoring including tibial and median nerve SEPs is used. Since SEP signals pass through the dorsal columns, we use them for mapping purposes in addition to continuous monitoring. In the cervical spinal cord, the fibers from the lower extremities that convey tibial nerve SEPs are situated in the ipsilateral fasciculus gracilis (i.e., medially), and those conveying median nerve SEPs are more lateral in the ipsilateral cuneate fascicle [11, 14]. With the multielectrode placed on the dorsal cord surface, we recorded repeatable, conducted waves after tibial nerve stimulation. The spinal SEPs showed an amplitude gradient ipsilateral to the stimulated side, with the highest amplitude of SEPs at a single electrode contact for each stimulated tibial nerve (Figs. 7.1 and 7.4) (see also color plate). Spinal SEPs after median nerve stimulation recorded with the miniature multielectrode represent mainly dorsal horn activity (segmental potentials) with a small superposition of conducted potentials. Because the dorsal horns are

161

FIGURE 7.4 Dorsal column mapping in a 43-year-old patient with an ependymoma between the C1 and C7 segments of the spinal cord. Spinal SEP responses were obtained from the eight recording sites after left and right tibial (bottom 2 traces) and median nerve (top 2 traces) stimulation. Two sets of 100 sweeps were averaged. Note the amplitude gradient of conducted potentials across the recording surfaces, with maximum amplitude after left tibial stimulation at recording site 4, whereas after right-sided stimulation it was at recording site 6. There is also an amplitude gradient of segmental potentials across the recording sites after median stimulation. The maximum amplitude after left-sided stimulation was at recording site 1, and after right-sided stimulation it was at site 7. Between the traces is an intraoperative picture taken during the measurement (see also color plate).

162

Matevˇz J. Krˇzan

laterally positioned, recorded segmental potentials still showed an amplitude gradient, having maximal amplitude laterally. In 1 patient, we recorded conducted responses in monopolar as well as bipolar fashion. The site with the highest amplitude, recorded monopolarly, showed phase reversal in the bipolar recording, demonstrating the high selectivity of the mapping multielectrode (Fig. 7.2). Contamination of recordings by activity originating from ipsilateral spinocerebellar tracts situated in the lateral columns is very unlikely because (1) the tibial nerves were stimulated at the ankles, where, according to the literature, fibers contributing to the spinocerebellar pathways are relatively rare [19]; (2) the DREZ separates the electrical activity of the dorsal columns from that of the dorsolateral columns; and (3) more laterally situated recording sites, although being closer to the spinocerebellar pathways, showed lower amplitude activity when compared to sites closer to the dorsal midline. We recorded spinal SEPs in 3 patients with the electrode positioned at a right angle to the cord axis (Fig. 7.3). In this position no amplitude gradient was recorded, but a stepwise increase in latency toward the proximal electrodes was observed. The miniature multielectrode was found to reliably record high-quality and high-amplitude tibial nerve spinal SEPs over the cervical and thoracic spinal cord. The electrode size was appropriate to the size of the dorsal columns according to anatomical studies that showed the distance between DREZs to be 6 to 7 mm at the cervical level [20]. In the lumbar area there was contamination with the higher-amplitude segmental responses arising from activity in the dorsal roots and horns. Even with the hi-pass filter set at 100 Hz, we were not able to clearly distinguish conducted waves from the segmental ones. In this area the dorsal columns are also narrower, making positioning of the electrode more difficult. Similar contamination with segmental waves was experienced in the cervical area with spinal SEPs recorded after median nerve stimulation. This was only helpful in determining the functional midline indirectly, indicating proximity to the DREZ. Therefore, it can be useful in mapping of the DREZ. Nevertheless, the midline in the cervical spinal cord could successfully be determined using tibial nerve SEPs. Besides showing that measurements of SEP amplitude gradients were possible intraoperatively, we were also able to identify the functional midline between left and right dorsal columns in 55 of 65 patients in whom mapping was attempted. We found this midline to lie between two recording sites with the highest SEP amplitudes after right or left tibial nerve stimulation. In patients in whom the anatomical midline was clearly delineated, it corresponded with the functional midline. Figure 7.5 (see also color plate) depicts the practical aspects of dorsal column mapping as used to guide the surgeon in performing the myelotomy on a patient with a cervical syringomyelic cyst. According to the mapping results, the patient’s left and right dorsal columns have been shifted to the right side within the spinal cord.

163

FIGURE 7.5 Dorsal column mapping in an 18-year-old patient with a syringomyelic cyst between the C2 and C7 segments of the spinal cord. Upper right, MRI showing syrinx. Lower middle, placement of miniature electrode over surgically exposed dorsal column; vertical bars on the electrode represent the location of the underlying exposed electrode surfaces. SEPs after stimulation of the left and right tibial nerves showing maximum amplitude between recording sites 1 and 2 (lower left and right). These data strongly indicate that both dorsal columns from the left and right lower extremities have been pushed to the extreme right side of the spinal cord. Using these data as a guideline, the surgeon performed the myelotomy through the left side of the spinal cord (upper middle) and inserted the shunt to drain the cyst. Postoperatively, the patient did not suffer from a sensory deficit. Reprinted from [1] (see also color plate).

164

Matevˇz J. Krˇzan

5 CONCLUSION In the nondistorted anatomy of the spinal cord dorsum, the determination of the midline in order to perform a midline myelotomy could be done on the basis of anatomical landmarks alone. Because of pathology, the anatomical landmarks of the spinal cord dorsum are often not visible, and the anatomy of the spinal cord can be changed. This makes midline myelotomy difficult to perform, increasing the risk of postoperative sensory deficits for the patient. The dorsal column mapping technique is a promising tool as a guide to determine the midline when anatomical landmarks are distorted. The method is also a powerful research tool for the study of human spinal cord physiology, offering new possibilities for further improvement of mapping and monitoring methodology. Included with the accompanying CD is a video demonstrating the utility of dorsal column mapping in locating the physiological midline of the dorsal columns during surgery on the cervical spinal cord for removal of an intramedullary tumor (choose Chapter 7 from the accompanying CD’s main menu).

REFERENCES 1. Deletis, V., and Sala, F. (2001). The role of intraoperative neurophysiology in the protection or documentation of surgically induced injury to the spinal cord. In “Neuroprotective agents. Fifth International Conference” (W. Slikker, Jr., and W. Trembly, eds.), vol. 939, pp. 137–144. New York Academy of Science, New York. 2. Deletis, V. (1994). Evoked potentials. In “Clinical monitoring for anesthesia and critical care” (C.L. Lake, ed.), pp. 288–314. W.B. Saunders, Philadelphia. 3. Wood, C.C., Spencer, D.D., Allison, T., McGarthy, G., Williamson, V.D., and Goff, W.R. (1988). Localization of human sensorymotor cortex during surgery by cortical surface recording of somatosensory evoked potentials. J. Neurosurg., 68, 99–111. 4. Taniguchi, M., Cedzich, C., and Schramm, J. (1993). Modification of cortical stimulation for motor evoked potentials under general anesthesia: Technical description. Neurosurgery, 32(2), 219–226. 5. Morota, N., Deletis, V., Epstein, F., Kofler, M., Abbott, R., Lee, M., and Ruskin, K. (1996). Brain stem mapping: Neurophysiological localization of motor nuclei on the floor of the fourth ventricle. Neurosurgery, 37(5), 922–930. 6. Campbell, J.A., and Miles, J. (1984). Evoked potentials as an aid to lesion making in the dorsal root entry zone. Neurosurgery, 15(6), 951–952. 7. Deletis, V., Voduˇsek, D., Abbott, I.R., Epstein, F.J., and Turndorf, H. (1992). Intraoperative monitoring of the dorsal sacral roots. Minimizing the risk of iatrogenic micturition disorders. Neurosurgery, 30(1), 72–75. 8. Deletis, V., Morota, N., and Abbott, I.R. (1995). Electrodiagnosis in the management of brachial plexus surgery. In “Hand clinics: Brachial plexus surgery” ( J.A. Grossman, ed.), vol. 11(4), pp. 555–561. W.B. Saunders, Philadelphia. 9. Young, W. (1991). Neurophysiology of spinal cord injury. In “Spinal trauma” (T.J. Errico, R. Bauer, and T. Waugh, eds.), pp. 377–414. Lippincott, Philadelphia. 10. Kothbauer, K., and Deletis, V. (1997). Comparison of motor and sensory evoked potential monitoring in surgery of intramedullary tumors. Invited lecture, annual meeting of the American Academy of Clinical Neurophysiology, Jan 29, 1997, San Francisco, CA.

Intraoperative Neurophysiological Mapping of the Spinal Cord’s Dorsal Columns

165

11. Jeanmonod, D., Sindou, M., and Mauguie`re, F. (1989). Three transverse dipolar generators in the human cervical and lumbo-sacral dorsal horn: Evidence from direct intraoperative recordings on the spinal cord surface. Electroencephalogr. Clin. Neurophysiol., 74, 236–240. 12. Ertekin, C. (1976). Studies on the human evoked electrospinogram: II. The conduction velocity along the dorsal funiculus. Acta. Neurol. Scand., 53, 21–38. 13. Halter, J.A. (1995). Spinal cord evoked potentials recorded at different vertebral levels. In “Atlas of human spinal cord evoked potentials” (M.R. Dimitrijevic´, and J.A. Halter, eds.), pp. 39–83. Butterworth-Heinemann, Boston. 14. Desmedt, J.E. (1989). Somatosensory evoked potentials in neuromonitoring. In “Neuromonitoring in surgery” ( J.E. Desmedt, ed.), pp. 1–21. Elsevier, Amsterdam. 15. Shimoji, K. (1995). Origins and properties of spinal cord evoked potentials. In “Atlas of human spinal cord evoked potentials” (M.R. Dimitrijevic´, and J.A. Halter, eds.), pp. 1–25. ButterworthHeinemann, Boston. 16. Turano, G., Sindou, M., and Mauguiere`, F. (1995). Spinal cord evoked potential monitoring during spinal surgery for pain and spasticity. In “Atlas of human spinal cord evoked potentials” (M.R. Dimitrijevic´, and J.A. Halter, eds.), pp. 107–122. Butterworth-Heinemann, Boston. 17. Nashold, B.S. Jr., Ovelmen-Levitt, J., Sharpe, R., and Higgins, A. (1985). Intraoperative evoked potentials recorded in man directly from dorsal roots and spinal cord. J. Neurosurg., 62, 680–693. 18. Krˇzan, M., Deletis, V., and Isgum, V. (1997). Intraoperative neurophysiological mapping of dorsal columns: A new tool in the prevention of surgically induced sensory deficit? Electroencephalogr. Clin. Neurophysiol., 102, 37P. 19. Halonen, J.P., Jones, S.J., Edgar, M.A., and Ransford, A.O. (1989). Conduction properties of epidurally recorded spinal cord potentials following lower limb stimulation in man. Electroencephalogr. Clin. Neurophysiol., 74, 161–174. 20. Smith, M., and Deacon, P. (1984). Topographical anatomy of the posterior columns of the spinal cord in man: The long ascending fibres. Brain, 107, 671–698.

-o ~

~

._.-..~ )

'

x

_..J'

o

.

~.o

S..~

o

8~

~

~'~

.~~

~ z s ~ ~

~

.-o~ o-

~ .

g

~

~

~0

~~'~

~.~ r~

~-

~ ~ ~ ~

m

~.~

~o~

o~.

~

~ . ~

-• ~ ~. ~

k~

~.

.

,

•

,

,

,

,

,

,

.

.n

r..i-

"1"

0

l

o

o

3

Q

•

~

!

I

'.4

Cb

qIA .Ib, r.~

I

!

I~

-~

.

w

Q. ~°

3

"1-I

0

•

3

_m.

0

~

~

~. ~

~.

o ~:~ ~-.~

~

. . . . ~-~-

~0~ . ~ 0~ ~ ' ~ "0

o~

~.~ o ~ ~ ~ .

l::r'

.9~

~.~ ~" .

,-~

~;"

0

! ~~

,~ ~ ~ "

o ~

.~ ~"~ "~ O~ ~ ~

CHAPTER

8

Intraoperative Neurophysiology of the Peripheral Nervous System LEO HAPPEL AND DAVID KLINE Louisiana State University Medical Center, New Orleans, Louisiana

1 2 3 4

Background Nerve Regeneration Equipment for Intraoperative Recording Electrodes for Intraoperative Recording and Stimulation 5 Anesthetic Considerations 6 Recording CNAPs Intraoperatively 7 Criteria for Appraising a CNAP 8 Operative Results 9 Troubleshooting 10 Conclusions References

ABSTRACT Intraoperative recordings of compound nerve action potentials (CNAPs) can provide quick, reliable information on the status of peripheral nerves at the time of surgery. The technique is straightforward and can be easily used by those without a lot of previous experience in monitoring peripheral nerves. It requires no unusual instrumentation and is very cost-effective. It does not compromise routine surgical exploration of a peripheral nerve injury. The information provided by these studies is very useful in determining the best course of action to deal with a particular peripheral nerve injury. The indications of early, successful peripheral nerve regeneration observed in these studies cannot be obtained in any other way. Thus the method should not be viewed as a “monitor” of peripheral nerve activity—the information is diagnostic and essential. We encourage the use of this technique as a means of evaluating a peripheral nerve injury and deciding on the best way to deal with it. To facilitate the application of intraoperative recordings, this chapter will include a description of the methodology, an interpretation of the findings of intraoperative Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

169

170

Leo Happel and David Kline recordings, some background in the pathophysiology of nerve injury and regeneration, and a practical section on troubleshooting to assist those who are just beginning to use this technique.

1 BACKGROUND The surgeon confronted by a neuroma in continuity has difficult decisions to make. He must determine the status of the nerve at the time of surgery and judge its potential to recover from injury. He must decide the best course of action to give the injured nerve the best prospect for an optimal recovery. Although pathologic examination can provide anatomical information on the status of the nerve, this information is not available without removing a specimen. This may further damage a nerve already undergoing regeneration and is not a reasonable solution. Few studies have focused on the functional status of the nerve. Over the course of three and a half decades, we have explored the use of operative peripheral nerve recordings to facilitate the process of decision making during exploration and repair of a peripheral nerve injury [1]. The compound nerve action potential (CNAP) has been found to be a useful tool toward this goal [1–7]. Some background knowledge on the response of nerves to injury is necessary to understand the findings of these neurophysiological studies. One objective of this chapter will be to describe the changes that occur in injured nerves and to relate these to the process of regeneration in order to gain insight into the interpretation of intraoperative neurophysiological studies. The technical difficulties associated with recording nerve action potentials often prevent those with little experience from obtaining useful recordings. Therefore, another objective will be to guide the reader with a detailed technical background. These technical issues may seem needlessly complex at times, but it is hoped that they may serve as a reference to address specific problems that may arise as one gains experience. A nerve lesion that leaves the nerve in some degree of continuity may affect some parts of the nerve more than others. Though it may be misleading, the term partial nerve injury is often applied in this circumstance. The use of this term seems to imply that some parts of the nerve fibers remain normal while other parts are affected by injury. In such a lesion it is more realistic to hold the position that none of the nerve is normal but some portions are more severely affected than others [5, 8]. Perhaps a better term for this situation would be a mixed injury to the nerve. Often, some parts of the nerve can be treated differently to improve prospects for recovery. Some fascicles can be resected and repaired, and others can simply be neurolysed. In this “split repair,” those portions of the nerve that are minimally influenced by the surgeon will show a faster functional recovery than those that had to be resected and repaired. Thus they

Intraoperative Neurophysiology of the Peripheral Nervous System

171

will also ultimately regenerate more effectively. Most injuries that are severe and yet leave the nerve in continuity affect in similar fashion the whole cross section of the nerve. However, some of these nerves have the ability to regenerate well and others do not. Operative nerve recording are equally important in these instances [8–10]. When some fascicles contain axons that are interrupted and others contain intact axons, a neuroma in continuity may develop in part of the nerve. This occurs as regrowing axons fail to project in length and fold back onto themselves. The entangled, growing neurites increase in volume and begin to compress the intact axons. This results in a progressive loss of function long after the initial insult to the nerve. We have seen many examples of large neuromas in continuity, such as that seen in Fig. 8.1B, with significant portions of the nerve still showing conduction. Operative recordings, then, have accurately shown that the proper course of surgical treatment is to do a split repair, resecting only those fascicles involved in a neuroma and sparing those that remain intact. This ensures that the patient will have the best outcome for the injury he or she has sustained. Similarly, as in Fig. 8.1A, we have seen many examples of lesions that are benign in appearance but that show no electrical conduction. Visual inspection alone might deceive the surgeon, suggesting that this lesion might regenerate without repair. In most cases such a severe lesion will not show significant regeneration, and the best course of action would be to resect and repair it. At the time of surgery the objective is to put neurophysiological findings into the context of patient history to gain some insight into the anatomy of the nerve without having to biopsy it. Preoperative neurophysiological studies may be helpful in describing the lesion, though these studies should not be done within 72 hr of injury. Within this 72 hr time period, axons distal to the point of injury may survive even if they are completely transected and may subsequently provide misleading information [5, 8, 11]. If the surgery is a primary repair performed within 72 hr of injury, preoperative neurophysiological studies may not be helpful. If the nerve is bluntly and completely transected, a delayed early repair may be planned at 3 weeks. Lesions in continuity, however, are more difficult to deal with. In the case of a secondary repair when a lesion in continuity is suspected, we will usually plan surgery at approximately 2 to 4 months post injury. In this case, initial neurophysiological studies would be routine and would evaluate the extent of the initial injury. They also provide a basis of comparison for what can be seen operatively. Surgical exploration three months after injury may also indicate whether spontaneous regeneration has begun. The operative electrophysiology supplements information that was learned preoperatively to provide a perspective of the extent of injury to the nerve. Once the extent of the injury has been determined, the optimal surgical treatment can be provided.

172

Leo Happel and David Kline

FIGURE 8.1 (A) This neuroma in continuity is small in size and appears benign. However, neurophysiological testing and subsequent pathological examination confirm that there were no significant axons passing through the lesion, which was properly resected and repaired. (B) This large neuroma in continuity looked and felt as if it was simply scar tissue. CNAPs recorded across it confirmed good conduction in many of the axons. This lesion was present in only a small part of the whole nerve, leaving much of it spared. We conducted a “split repair,” resecting and repairing only the involved fascicles.

Intraoperative Neurophysiology of the Peripheral Nervous System

173

Sunderland has classified nerve injuries into five categories ranging from mild functional change to division of the nerve [8, 12]. Operative recordings can facilitate an understanding of the degree of injury as described by Sunderland. A Sunderland grade 1 injury that is neurapraxic leaves the axon in continuity. There may be mild changes to myelin but little other anatomic change. As long as the axon remains connected to the cell body, it remains functional even though a localized conduction block exists at the point of injury. This can be determined easily at the operating table by stimulating and successfully recording from a section of axon that is distal to the point of injury. Preoperative EMG would show similar findings, and the needle EMG study would show little or no evidence of denervation. Again, these recordings should not be made within 72 hr of the initial injury, for the reasons cited previously. A functional block of conduction at the site of injury does not affect conduction in the axon distally. A demonstration of normal nerve excitation and conduction in the nerve distal to the point of injury is proof of axonal integrity. An important exception to this is the avulsive injury that divides the sensory axon proximal to the dorsal root ganglion but spares the distal axon in the nerve. The distal axon would exhibit normal excitation and conduction properties, and yet it would be disconnected from the central nervous system. Preoperative EMG studies—as well as radiographic studies—should alert the surgeon to this possibility, which should be considered in appropriate circumstances and which will be treated later in this chapter. A Sunderland grade 2 injury is axonotmetic, leading to Wallerian degeneration of the axon distal to the point of injury. This degeneration causes the distal axon to lose its properties of electrical excitability over the course of 72 hr. No peripheral nerve action potential can be seen if all of the fibers of the nerve under study have degenerated. This injury, however, is associated with little derangement of the connective tissue elements of the nerve. Spontaneous nerve regeneration is likely with this degree of injury, and, if the timing of surgical exploration is appropriate, early indications of axonal regeneration can be seen across the site of injury. This is one of the great advantages of operative recordings. Routine preoperative EMG studies would not show these early indications of regeneration. The electrical characteristics of these regenerating axons distinguish them from normal axons, as will be discussed later. Sunderland grade 3 and grade 4 injuries represent greater obstacles to nerve regeneration. In these cases the injury is neurotmetic, altering the connective tissue components of the nerve. A grade 3 injury is a mix of axonotmetic and neurotmetic injuries and is associated with mild derangement of connective tissue and mild scar formation. Examination of this injury by operative recordings at 3 months may also show indications of early spontaneous regeneration, suggesting conservative surgical treatment. However, a grade 4 injury is neurotmetic and will be associated with much greater scar formation and a formidable

174

Leo Happel and David Kline

barrier to nerve regeneration. Electrical recordings performed at 3 months after injury would not show indications of early regeneration if this scar blocks regrowing axons. Grade 3 and grade 4 injuries are the most important to differentiate, since the heavy scar of grade 4 injury will not permit spontaneous regeneration. This lesion in continuity must be resected and repaired in order to provide the best chance for optimal recovery and to remove the offensive scar that blocks the regrowth of nerve [6–8]. A grade 5 injury results in a complete transection of nerve, either from a sharp or blunt insult. Blunt injuries may include those that, by stretching, pull the nerve completely apart. In both of these cases, a complete repair of the nerve is necessary. However, with blunt injury it is often difficult to determine the length of injured nerve that should be removed. Operative recordings can be helpful in this regard, demonstrating the point on the proximal stump where viable axons remain. Then, once the nerve is sectioned at this point, one can visibly determine if a fascicular pattern remains at this level. In this way one can determine whether the entire scar has been removed before the repair is begun [8, 13].

2 NERVE REGENERATION In order to understand neurophysiological findings obtained during surgery, it is necessary to understand the process of nerve regeneration. This process is complex, and particularly so in humans. There is a significant difference in the processes of regeneration between lower mammals, such as the rat, and those processes in the human. In lower mammals nerve regeneration is much more effective and complete, so much so that in some experimental settings it even becomes difficult to prevent nerve regeneration. Rat nerves usually show significant regeneration even in the most adverse circumstances. By contrast, peripheral nerve regeneration in humans is not nearly so effective, and regenerated axons never regain the electrical properties of their original counterparts. For this reason, one needs to be particularly careful in applying the results of research conducted on lower mammals to the human [8, 13, 14]. When an axon is divided, the distal part undergoes Wallerian degeneration [11], and the proximal part seals off at the point of division. Within 36 hr, multiple sprouts of growing neurites appear at the sealed end of the proximal axon [12]. These sprouts will give rise to several small, growing axons, each attached to the single proximal axon. The point of injury becomes a branch point, and it is not uncommon to see axon counts distal to the point of injury that are higher than proximal axon counts. These growing axons are much smaller in diameter and have distinctive electrical properties [15–18]. Their thresholds are markedly higher than those of normal nerves, and they are particularly insensitive to short-duration stimulus pulses, in part because of the increased capacitance of their membranes. Their conduction velocities fall into a range that is

Intraoperative Neurophysiology of the Peripheral Nervous System

175

much lower than that of normal nerve. As the process of nerve regeneration continues, these axon sprouts elongate. During effective regeneration, some of these fine fibers will eventually die back in order to allow remaining fibers to increase in caliber [12]. If these small-diameter fibers do not increase in diameter, they are unlikely to form an effective junction with muscle, since motor axons must achieve a critical diameter in order to produce a useful motor unit. Should many fine fibers persist, the motor units formed will not lead to significant muscle strength [8, 12, 14, 19]. The presence of fine fibers may be an indication of ongoing regeneration at a very early stage but may also indicate ineffective regeneration at a later stage.

3 EQUIPMENT FOR INTRAOPERATIVE RECORDING Operative recordings of peripheral nerve action potentials can be easily accomplished with many types of commercially available EMG machines. These offer both stimulating and recording capabilities that are appropriate for the operative setting [20]. Evoked potential instrumentation can also be used, though it is usually more complex and difficult to use in such a simple setting. The stimulator used operatively should have the ability to produce pulses of short duration (0.02–0.05 ms) and intensities up to 70 V. We advocate the use of very short duration stimulus pulses to reduce stimulus artifact. Short-duration stimulus pulses may also help discriminate the types of fibers that might be present, as is illustrated in Fig. 8.2. Small-diameter fibers of normal nerve are less sensitive to

FIGURE 8.2 The strength–duration relationship for stimulation shows that for normal fibers short-duration stimulus pulses require greater intensity. This phenomenon is particularly exaggerated at extremely short stimulus pulse durations. By comparison, regenerated axons are even less sensitive to short-duration stimulus pulses. This principle can be used to help discriminate the qualities of axons found in injured nerve. By using short-duration stimulus pulses, we can selectively activate larger-diameter axons. Reprinted from [5].

176

Leo Happel and David Kline

short-duration stimulus pulses [17]. The fine fibers of regenerating axons are even less sensitive to short-duration pulses than are equivalent normal fibers. Their strength–duration relationship is different than that of normal fibers with similar size. The responses that we record to stimulation with short-duration pulses are, necessarily, from larger-diameter axons, and these may be a better indicator of effective regeneration. Additionally, these short-duration pulses reduce the amount of stimulus artifact; this subject will be discussed further in a later section [20]. The strength–duration relationship seen in Fig. 8.2 also shows that with short-duration stimulus pulses much higher stimulus intensities must be used [17, 21]. We have found that in some cases when short-duration pulses are used, stimulus intensities as high as 70 V are required to excite regenerating axons. As long as pulse duration is kept short, these intensities can be used safely. However, if long-duration pulses are used at this intensity, the energies transferred by the stimulator can become dangerous and electrical burns may be possible. This is an additional reason for using short-duration pulses. The stimulus should be properly isolated from ground (as illustrated in Fig. 8.1) in order to prevent electrical currents from leaking into the recorder or through some other part of the patient’s body. With no stimulus isolation, a potential difference applied between stimulating electrodes also represents a potential difference with any other electrode that may be connected to ground, such as the recording electrode seen in this example. The stimulator may produce a current through any other electrical contact that the patient may have with ground. Though stimulus isolation is engineered into the EMG machine, this engineering can be defeated through poor application. If the wires leading to the stimulating electrodes are shielded, the resulting capacitance to ground defeats stimulus isolation and spurious currents may result. The cable connecting stimulating electrodes to the instrumentation should not be shielded. The same process may occur if these wires are draped against a metal surface such as the operating table or next to other wires. The resulting capacitive coupling defeats the stimulus isolation engineered into the EMG machine. This may produce excessive stimulus artifact or may even put the patient at risk for accidental electrical shock. Care should also be exercised in the positioning of wires connecting the stimulator to the stimulating electrodes. When possible, suspend these wires in the air, away from any other wires or metal objects. It may also help to separate the stimulating cable from the recording cable as it is led off the sterile field to the EMG machine. Most modern recording instrumentation now employ isolation amplifiers to augment stimulus isolation. The recorder portion of the EMG machine is optically isolated from ground by isolation amplifiers that reduce stimulus artifact even more, in addition to enhancing patient safety. Each recording channel will have a positive (+) and negative (−) active input and also an isolation ground connection.

Intraoperative Neurophysiology of the Peripheral Nervous System

177

This isolated ground connection is not a true ground and would not be connected to any other part of the EMG machine. It cannot become part of a socalled ground loop. Thus, when properly connected, the patient is not attached to any true ground. The isolation ground connection on the EMG machine may safely be attached to the patient and may help to reduce electrical interference. This connection is not essential, however, and we routinely conduct studies with no ground connection at all. The recording sensitivity should initially be set to approximately 100 µV/cm or 1 mV for full-screen deflection [5, 8, 14]. At this sensitivity one should clearly see stimulus artifact at the beginning of the trace. If not, it will be necessary to troubleshoot in an effort to detect the source of the problem. Troubleshooting will be discussed in a later section. When a trace has been obtained that shows some stimulus artifact, the intensity of stimulation can be increased to a range of 6 to 8 V. If no nerve action potentials can be seen under these conditions, the recording sensitivity can be progressively increased to approximately 20 µV/cm. At this sensitivity stimulus artifact should be quite large, and one may have to inspect the tail of the stimulus artifact closely to determine if a CNAP is present. The stimulus artifact decays as an exponential curve, and the shape of this curve is dramatically affected by the settings of filters [20]. The slope of the exponential decay of the stimulus artifact is most affected by the low-frequency filter setting. We would normally begin recording with a low-frequency filter setting of about 10 Hz. At this setting the exponential decay is relatively slow and causes the tracing to be fairly flat. However, some amplifiers would saturate under these conditions, and the trace would appear flat at either the uppermost or lowermost part of the display screen. If this happens, it will be necessary to increase the low-frequency filter setting to 30 or even 100 Hz. Under these conditions, the slope of the stimulus artifact will be much steeper and the amplifier should emerge from saturation. However, this may make the CNAP difficult to see. The high-frequency filter setting that we routinely use is between 2500 and 3000 Hz. This does not usually affect the shape of the CNAP, which has an equivalent frequency of approximately 500–2700 Hz. It will remove extraneous high-frequency noise from many other sources. The high-frequency filter setting will not affect the rate at which the amplifier emerges from saturation. An important point to remember in selecting filter settings is that the CNAP should not be affected in an effort to reduce stimulus artifact or extraneous noise [20]. If an evoked potential machine is being used to record the CNAP, there may be a 60 Hz notch filter available. This should not be used under any circumstances, since it may produce an effect called “ringing.” With stimulation, a dampened oscillation will become part of the stimulus artifact. This dampened oscillation may look very much like a CNAP and confuse the observer. For this reason, most EMG machines do not contain a 60 Hz notch filter. In any case, it

178

Leo Happel and David Kline

is not advisable to use a 60 Hz notch filter when stimulating and recording from peripheral nerve.

4 ELECTRODES FOR INTRAOPERATIVE RECORDING AND STIMULATION More than 30 years ago, we developed our own electrodes for stimulating and recording from nerve tissue. During the ensuing years we have been able to refine them, adding helpful features. Examples of these can be seen in Fig. 8.3,

FIGURE 8.3 Electrodes for stimulating and recording CNAPs can be made in many sizes, according to one’s needs. Illustrated here, from left to right, are miniature, midsize, and large electrodes. The stimulating electrode contains three contacts, and the recording electrode contains two. The inset enlargement of the electrode tips illustrates the curved hooks on which exposed nerve can be suspended. The tip separation of the recording electrodes can be adjusted according to the size of the nerve from which recordings are made.

Intraoperative Neurophysiology of the Peripheral Nervous System

179

which demonstrates the simple and convenient design of these electrodes. The requirements for electrodes include durability, reliability, and functionality. The electrodes should withstand steam autoclaving and the rigors of routine handling together with other surgical instruments. They should have electrical characteristics that are conducive to safe stimulation and effective recording. The stimulating electrode contacts should never be made of silver. Although the resistance of silver wire is very low, stimulation through silver electrodes deposits silver salts that are toxic to nerve. Any metals used for tissue contact should have good tissue compatibility. The electrodes should have low electrical resistance, and they should resist tarnishing. They should also have adequate strength to hold their shape even under the weight and pull of nerves suspended on them. We have found stainless steel electrodes to be effective, and their cost is modest compared to that of noble metals such as platinum [5, 8, 14]. The recording electrodes that we use consist of a handle that is made from a high-temperature plastic (acetal), which is easy to clean. Attached to the handle are two Teflon-insulated stainless steel electrodes approximately 8 cm long. For large-sized electrodes these are 1.125 mm in diameter, for medium-sized electrodes they are 0.875 mm in diameter, and for the miniature electrodes they are 0.625 mm in diameter. The ends are blunted and bent like a shepherd’s crook and can be used to pick up and suspend the nerve. The tips of these electrodes are separated by a distance of 5 to 7 mm for the large-sized electrodes, 3 to 5 mm for the medium-sized electrodes, and 2 to 3 mm for the miniature-sized electrodes. The distance between the tips of the recording electrodes determines, in part, the amplitude of the CNAP. If the distance between the tips of the recording electrodes is too small, the size of the CNAP will be reduced and inappropriate amounts of amplification may be required to see the CNAP. At high amplification, then, stimulus artifact could become a problem. This emphasizes the need to maintain an adequate distance between electrode tips, which should always be greater than the length of active nerve during the CNAP. If both recording tips are applied to a section of nerve that is simultaneously active, the size of the CNAP may be markedly reduced. The length of active nerve is considerably larger than one might imagine based on the anatomy of nodes of Ranvier. This is due to the fact that saltatory conduction in myelinated fibers is not simply regeneration of the action potential at successive nodes of Ranvier but rather a process in which several nodes of Ranvier (two to three) are activated simultaneously. One only needs to do simple arithmetic to show this. By considering the period of time required to produce an action potential at a single node of Ranvier and also the distance between nodes of Ranvier, a theoretical conduction velocity can be calculated. This theoretical conduction velocity is only one half to one third the observed conduction velocity in myelinated fibers [15, 16]. This fact dictates that action potentials must

180

Leo Happel and David Kline

FIGURE 8.4 Saltatory conduction in myelinated axons occurs as small groups of nodes of Ranvier become active simultaneously. This conduction increases the length of active nerve and requires a greater distance between the tips of the recording electrodes. Reprinted from [5].

jump several nodes of Ranvier at a time, as is illustrated in Fig. 8.4. This process indicates that the length of active nerve is greater as a result of this phenomenon. Therefore, Fig. 8.4 illustrates that a consistent distance between the tips of the electrodes must be maintained. One of the electrode tips must lie on the part of the nerve that is not active, and the other must contact the active part of the nerve. The electrodes are soldered to a 10-ft length of flexible, Teflon-insulated wire that permits these wires to be led off of the sterile field. The Teflon insulation resists abrasion and is also unaffected by high-temperature autoclaving. Appropriate plugs are used on the ends of these wires to permit attachment to the recording instrumentation. It should be noted that soldering to stainless steel requires special soldering flux and some skill. The stainless steel electrodes are then embedded into the acetal handles using methacrylate cement. Strainrelief is provided to the wires leaving the acetal handles to prevent bending fatigue and eventual breakage of the wires. Similarly, stimulating electrodes are also fabricated with stainless steel electrodes and an acetal handle. However, in this case, three electrodes are embedded into the handle. These electrodes are also blunted and bent like a shepherd’s crook to support the suspended nerve. Tip separation is similar to that for the recording electrodes. This tripolar electrode is used to circumvent a special situation that exists when stimulating a nerve in continuity. As can be seen in Fig. 8.5, stimulation with a bipolar electrode produces two current paths, one short and one quite long. The longer current path seen in this figure is very undesirable because it leads to excessive stimulus artifact, especially when the distance between stimulating and recording electrodes is very short. In addition,

Intraoperative Neurophysiology of the Peripheral Nervous System

181

FIGURE 8.5 Stimulation of nerve in continuity presents an unusual situation in which bipolar stimulation, top, produces two current paths. There is a very short path directly between electrodes, but there is also a second, longer path through the nerve and through the forearm. This second path passes beneath the recording electrode, producing large quantities of stimulus artifact. Tripolar stimulation, bottom, breaks the longer current path and localizes the stimulus to the region of electrode application. This dramatically reduces stimulus artifact, especially when the distance between recording and stimulating electrodes is short. Modified from [5].

it may permit the spread of stimulation over long lengths of the nerve when higher intensities of stimulation are used. The tripolar electrode that we have developed breaks the longer current path and so reduces stimulus artifact and helps localize stimulation. In the tripolar electrode, the outermost electrodes are connected together so that there is no potential difference between them. There are still two current paths in this situation, but they are both short and localized to the region of contact with the nerve. There is very little spread of stimulation with the tripolar electrode [5]. The electrodes described here have been used successfully for many years and have proven their durability, reliability, and functionality. To maintain these electrodes over many years, we recommend gas sterilization for routine use, though they will withstand occasional high-temperature steam autoclaving. They can even be flash-sterilized should they become accidentally contaminated during a surgical procedure. Steam autoclaving, however, has a tendency to make plastics become brittle, and this eventually leads to a degradation of the electrodes. An occasional soaking in Instrument Milk (a cleaning and lubricating solution frequently applied to surgical instruments) will retard this degradation. The electrodes can be cleaned routinely with hot soap and water. In addition, their exposed metal tips will accumulate a protein coat of coagulum, and this should be periodically removed. Soaking the electrodes overnight in a solution of 30% bleach will soften and remove this coagulum.

182

Leo Happel and David Kline

5 ANESTHETIC CONSIDERATIONS There are few pharmacologic considerations in making operative peripheral nerve recordings. Inhalational agents and narcotics do not affect peripheral nerve function, and neuromuscular blocking agents may or may not be used, depending upon personal preference. The latter may prevent evoked muscle contractions, but this information is only useful in an ancillary way. Peripheral nerve surgery often involves surgery on a limb, and it is common to apply a tourniquet to control bleeding. We do not use a tourniquet, but if one is used, it should be released at least 20 min prior to any neurophysiological studies. If the tourniquet pressure is maintained, the nerve may not be functional and the findings of CNAP studies may be misleading. Local anesthetics placed into or close to the nerve can also block nerve conduction.

6 RECORDING CNAPs INTRAOPERATIVELY Once the level of a lesion to peripheral nerve has been determined by physical examination, patient history, and preoperative neurophysiological studies, surgical exploration can be carried out. A length of nerve is exposed that should include the site of the lesion. Over the years, we have seen many examples of lesions that appear benign yet prove to be complete and offer no indication of early successful regeneration. Similarly, we have also seen examples of large neuromas in continuity that encompass only one or two fascicles and spare adjacent fascicles or when the whole cross section is involved but it still conducts responses. The visual appearance of a lesion in continuity may be deceptive. We have successfully recorded from lengths of nerve as short as 4 cm. With this short distance, stimulus artifact can become an overpowering consideration, and longer lengths of nerve (8–10 cm) will facilitate recording. If a 4-cm length of nerve is not accessible, it may be necessary to stimulate or record percutaneously at a distant site. This can be accomplished by using skin electrodes or subdermal needle electrodes at some point down the length of the nerve. We begin by applying stimulus pulses of 0.02 ms duration and 6 to 8 V intensity. This stimulus is usually applied proximally, and recording electrodes are placed distally. When stimulation is applied distally and recordings made proximally, the size of the compound nerve action potential may be slightly reduced by fibers that are added to the nerve at proximal levels and are not subjected to the stimulating electrodes. The active fibers may thus become “buried” by the fibers that are not being stimulated and consequently do not produce action potentials. For this reason, a proximal response to distal stimulation may be reduced in size. To record potentials from the distal electrodes, the amplifiers are set at a sensitivity of 200–500 µV/cm. A time window of 0.5–0.1 ms/cm is set. Under these conditions normal nerve will produce a clear CNAP. If no response can

Intraoperative Neurophysiology of the Peripheral Nervous System

183

FIGURE 8.6 This small CNAP was recorded from a section of nerve undergoing appropriate regeneration. The low amplitude and slow conduction velocity distinguish it from the CNAP of normal nerve. The presence of such a response is an indication for conservative treatment of a lesion.

be seen, the sensitivity of the recorder will then be increased progressively down to 10 µV/cm. A small potential from regenerating nerve can be seen in Fig. 8.6. If there is still no clear CNAP, the stimulation will be increased progressively to levels of 50 V or more. If there is still no visible CNAP, this indicates the absence of significant numbers of adequate fibers and dictates resection and repair. For initial recordings, we do not use the signal-signal-averaging feature found on many EMG machines to enhance recordings of CNAPs. This technique is so sensitive that it may record very small numbers of fine fibers and indicate significant function in a segment of nerve that has no significant function. Once we do see a CNAP on each single trace, we may then average a number of traces to provide a clear, stable response for the patient’s record.

7 CRITERIA FOR APPRAISING A CNAP If the CNAP is present, it will meet the following criteria. The putative response will be phase-locked to the stimulus, causing it to appear in a fixed position on the recorder screen each time a stimulus is delivered. It will appear “frozen” on the screen with repetitive stimulation, and its amplitude will always be less than 2 mV. A response larger than 2 mV is more likely a muscle action potential. In addition to being larger, evoked muscle action potentials usually have a longer duration than the CNAP. Thus an evoked response with a duration greater than 2 ms is likely to be a muscle response. Muscle responses also tend to be polyphasic, whereas CNAPs are not. The response should exhibit threshold behavior as the stimulus intensity is raised and lowered. It should also exhibit a maximum size with increased stimulation. If visible contraction of adjacent muscle can be seen during stimulation, the stimulus intensity can be lowered until the muscle contraction stops and a small CNAP can still be seen. This may help distinguish

184

Leo Happel and David Kline

a muscle action potential from a CNAP. The duration of the CNAP should be less than 2 ms. Most CNAPs are not polyphasic, though under some particular conditions they may be. This may occur when some fascicles in the nerve are undergoing regeneration while others are recovering from a neurapraxic injury. It is helpful to begin stimulation and recordings on a segment of nerve that is presumed to be normal. This may be a portion of the nerve proximal to a visible point of injury or an adjacent nerve accessible within the operative site. By stimulating and recording from a section of nerve that was functional preoperatively, one can verify that the instrumentation is working properly and one can be comfortable with an observation of no function in a section of adjacent nerve. A great advantage of the electrodes that we use is that they can slide along the length of nerve easily. In doing this, care must be taken to maintain good contact with the nerve. If these electrodes are held perpendicular to the floor, gravity becomes an ally, pressing the nerve against all of the contacts of the electrodes evenly. This ensures appropriate stimulating and recording conditions. With this technique, proximal recordings from normal nerve can be compared to recordings made over and across a lesion in continuity and also distal to the lesion. Changes in the CNAP recorded at different levels of the nerve can then be related to the functional status of the nerve at those levels. Often, there may be little anatomical indication of a lesion along the length of a nerve, and these operative recordings can localize the problem. Again, if one recorded proximally and slides the distal stimulating electrode along the length of the nerve, the recorded CNAP would be lost at the point where axonal continuity is lost. In our experience, resection of nerve at this point shows that the specimen removed contains mostly scar and few, if any, axons. Intraoperative stimulation of peripheral nerve is often accompanied by evoked motor responses if the anesthetist has not blocked the neuromuscular junction. Although this observation may lend support to observations of peripheral nerve action potentials, it should not be used by itself as an indication of good functional connection with muscle. For example, we have seen patients with clear evoked motor activity who, preoperatively, had no voluntary control over a particular muscle following a nerve injury of long standing. With operative nerve stimulation there may be clear contractions of the muscle innervated. Collectively, these findings indicate that, with extended time, small axons may regrow and reach their target muscles. However, the motor units that they form are too small to mediate voluntary movements. When all of these motor units are synchronized by electrical stimulation of their nerve supply, they may produce a visible contraction even though such contractions cannot be produced voluntarily. Thus the manifestation of a visible contraction of appropriate muscles may not be an indication of adequate functional nerve regeneration. Even though stimulation of the nerve above the lesion will exhibit this phenomenon, the lesion should still be properly resected and repaired if it does not transmit a recordable CNAP.

Intraoperative Neurophysiology of the Peripheral Nervous System

185

For very proximal root or spinal nerve injuries, it may become necessary to stimulate spinal nerves and record from nerve trunks. In the case of a root avulsion, this preganglionic injury (between dorsal root ganglion and spinal cord) to the sensory root will produce a relatively large and rapidly conducting CNAP. If, by contrast, regeneration is occurring, the CNAP will be smaller and will have a slower conduction velocity in the range of 20–40 m/s. If there is a combination of postganglionic and preganglionic injury without effective regeneration, the recordings will be flat with no CNAP. For this type of extensive injury, the disruption of the axon proximal and distal to the dorsal root ganglion usually kills the cells of the dorsal root ganglion. For these cases we can section spinal nerve or roots proximally to prove the lack of proximal fascicular structure. We may also stimulate the exposed elements in the neck and record from the sensory cortex. In this regard, we have used somatosensory evoked potentials (SEPs) to get some indication of connection of nerve roots to the central nervous system. The complete absence of an SEP on stimulation of the nerve root indicates a complete avulsion and precludes surgical repair. The presence of an SEP upon nerve root stimulation should be viewed with some caution, however, since previous studies have shown that stimulation of even a very small number of fibers can produce a normal SEP [22]. If an SEP can be recorded following root stimulation, it should not be taken as evidence of normal function in the proximal parts of the root. Thus an absent SEP provides more definitive information than one that is present. When there is no SEP, one can accurately assume that there is no proximal connection. Such findings can often be verified by preoperative EMG studies conducted on peripheral parts of the nerve. In the case of an avulsive injury, the intact sensory axons produce a normal CNAP in the distal sensory branches. The electrical characteristics of the distal axon remain fairly normal with stimulation, and recording distal to the dorsal root ganglion will reveal the presence of these surviving axons. However, needle EMG studies will indicate profound denervation in all of the muscles supplied by this root. The distal motor axons will all have undergone Wallerian degeneration. The combination of normal sensory studies together with profound denervation of muscle indicates a very proximal, avulsive injury. In addition, the complete absence of an SEP upon stimulation of these distal axons may also demonstrate a complete avulsion.

8 OPERATIVE RESULTS Between 1965 and 1990 intraoperative recordings have been done in over 2000 patients. We have recorded through lesions in continuity in the upper and lower extremities of 877 patients, brachial plexus lesions in 432 patients, nerve tumors in 245 patients, nerve entrapments in 617 patients, and cranial nerve palsies in 62 patients. These numbers do not include patients with pelvic plexus involvement,

186

Leo Happel and David Kline

birth palsies, or injuries to smaller nerves such as digital, cutaneous antebrachii, sural, or saphenous nerves. In addition, this analysis does not reflect the fact that we often recorded from less involved or intact nerves in the neighborhood of the more seriously injured nerve. These data are not included in the analysis, since CNAP recording was not used in those cases to determine partial injury or regeneration. CNAP recording from these more intact nerves was done either out of scientific interest or to ensure that our instrument settings for recording from the more seriously injured nerve were adequate. Tables 8.1–8.5 show the results for upper- and lower-extremity nerves at different levels. Although loss associated with these lesions in continuity was usually TABLE 8.1 Criteria for Grading Whole Nerve Injury (LSUMC System) 0 (absent)

No muscle contraction. Absent sensation

1 (poor)

Proximal muscles contract but not against gravity Sensory grade 1 or 0.

2 (fair)

Proximal muscles contract against gravity, distal muscles do not contract, sensory grade if applicable was usually 2 or lower.

3 (moderate)

Proximal muscles contract against gravity and some resistance, some distal muscles contract against at least gravity, sensory grade was usually 3

4 (good)

All proximal and some distal muscles contract against gravity and some resistance. Sensory grade was 3 or better

5 (excellent)

All muscles contract against moderate resistance; sensory grade was 4 or better TABLE 8.2 In Continuity Cases Studied (1965–1990) Upper Extremity Nerves (+) NAP = Neurolysis/Result∗ Median Upper arm Elbow forearm Wrist Radial Upper arm Elbow PIN Forearm SSR

23/22 25/24 24/24 72/70

16/11• 16/12• 17/14 49/37

30/28 10/9 5/5 1/1 3/3 49/46

62/42+ 8/7 4/4 4/2 6/6 84/61

121/117 (97%) ∗

(−) NAP = Repair/Result

123/98 (79%)

Result = Those achieving a grade 3 or better result • = 1 split repair + = 2 split repairs

187

Intraoperative Neurophysiology of the Peripheral Nervous System TABLE 8.3 In Continuity Cases Studied (1965–1990) Upper Extremity Nerves (+) NAP = Neurolysis/Result∗

(−) NAP = Repair/Result

Ulnar Upper arm Elbow forearm Wrist

17/16 45/43 8/8

16/7• 14/10• 8/5+

Combined median-ulnar

40/36

48/30

Combined median-radial

11/8

5/4

Combined median-ulnarradial

4/4

8/6

125/118 (93%)

99/62 (61%)

∗

Result = Those achieving a grade 3 or better result • = 1 split repair + = 2 split repairs

complete both clinically and neurophysiologically before operation, there were some exceptions. In the upper- and lower-extremity nerve lesions, some operations were necessary for relief of severe pain or for partial but still severe functional loss distal to the lesion. Thus 70% of patients with median nerve lesions having complete functional loss distal to the lesion before operation had positive intraoperative recordings. For radial nerve lesions this figure was close to 74%, and for ulnar nerve lesions it was 68%. For lower-extremity nerves, preoperative functional loss was complete in 65% of patients. Although the majority of lower-extremity nerve lesions had complete functional loss in their distributions at the time of operation, there were important exceptions. These included injection injuries with incomplete loss but severe pain, gunshot wounds associated with partial loss and sustained pain, and a variety of other incomplete injuries affecting femoral or the more distal tibial nerve. Brachial plexus lesions are listed as having complete or incomplete functional loss at the time of their evaluation (see Table 8.4). They are also tabulated as major elements involved and evaluated, as well as cases operated on. Patients with tumors involving nerves usually had little or no functional loss preoperatively. Intraoperative CNAP recording was used to test fascicles entering and leaving intraneural tumors and to check progress of the dissection in other cases. For example, in 68% of 69 solitary neurofibromas involving nerves, major function in the innervating field of the particular nerve could be preserved despite total tumor removal by using intraoperative recordings and fascicular dissection. As can be seen in Tables 8.2–8.5, when a lesion was partial to begin with or, more frequently, had complete functional loss and yet a recordable CNAP, neurolysis alone led to an eventual recovery. Thus 93% of nerves having a

188

Leo Happel and David Kline TABLE 8.4 In Continuity Cases Studied (1965–1990) Lower Extremity Nerves (+) NAP = Neurolysis/Result Sciatic Buttock Thigh

Tibial 30/28 38/36

Peroneal 28/23 34/31

(−) NAP = Repair/Result Tibial 23/20• 37/33+

Peroneal 19/6• 43/19•

Tibial

13/12

11/10•

Peroneal

34/30

69/15+

Femoral

15/13

14/7

193/175 (90%)

216/110 (51%)

• = 1 split repair + = 2 split repairs TABLE 8.5 Brachial Plexus In Continuity Elements Studied (1965–1990) (+) NAP = Neurolysis/Result Injury Mechanism Lacer. in contin. (12)

(−) NAP = Repair/Result

Complete

Incomplete

Complete

Incomplete

10/9

18/17

15/10

4/3

GSW’s (90)

41/40

47/44

116/64∗

8/7•

Iatrogenic (30)

18/17

17/17

32/23•

6/4

Stretch/Contusion (300)

93/84

102/96

336/150•

45/28+

162/150

184/74

499/247

63/42

• = 1 split repair + = 2 split repairs ∗ = 5 split repairs

Complete = Complete loss in distribution of one or more major elements preoperatively Incomplete = Incomplete loss of function felt to be in the distribution of element tested

Cases are listed in ()

recordable CNAP that underwent subsequent neurolysis recovered to a grade 3 or better (see Table 8.1 for grading method). From another perspective, 82% of patients had a grade 4 or 5 recovery, which represents a very acceptable outcome. Of equal importance is the fact that when CNAPs were absent and a lesion in continuity was resected, pathological studies confirmed that the lesion was always neurotmetic or a Sunderland grade 4 nerve lesion. Such lesions had little or no potential for spontaneous regeneration that might lead to useful function. Optimal timing for recording varies according to the mechanism of injury. In lengthier lesions like those produced by stretch and/or severe contusion,

Intraoperative Neurophysiology of the Peripheral Nervous System

189

it takes longer for significant regeneration than can be recorded by direct CNAP studies. Thus most fracture-associated contusions and gunshot wounds can be tested operatively at 2 to 3 months post injury, whereas plexus stretch injuries are more reliably evaluated at 4 or 5 months after injury. On the other hand, recording can be done as an adjunct to tumor resection at any time and can be used as an investigative tool for entrapment or compressive neuropathies at any point in the course of these disorders. Intraoperative recording has been helpful in a relatively large number of patients with palsy of the accessory nerve. Loss of function in these patients was usually iatrogenic and due to lymph node biopsy or removal of a neck lesion and inadvertent damage to nerve distal to its innervation of sternocleidomastoid muscle. When the lesion was in continuity, as it was in 26% of cases, operative CNAP studies were done. This approach led to resection of about 50% of such accessory nerve lesions in continuity. These proved to be neurotmetic or Sunderland grade 4 nerve lesions. The other accessory nerve lesions in continuity had a neurolysis with a good outcome (average postoperative grade was 3.9). Although not essential for operative management of entrapment neuropathy, CNAP recordings were usually done and had interesting features. A direct recording was first made proximal to the presumed entrapment site. The actual entrapment site was then defined by progressively moving the recording electrodes in a distal direction toward, into, and across the presumed entrapment site. Mild degrees of decreased conduction velocities were sometimes seen well proximal to an area of more severe conduction problems. Only in a few cases did this appear to be due to separate lesions or what has been described as a “double crush syndrome.” On the other hand, operative conduction across the area of entrapment was almost always more severely affected than might have been predicted by the preoperative EMG studies. This may relate to the fact that the distance between the stimulating and recording sites was less at the time of intraoperative recordings than at the time of EMG studies. These differences were usually most obvious in patients with ulnar entrapments at the elbow and those with presumed entrapment of the peroneal nerve over the region of the head of the fibula. In only five examples of true distal cubital tunnel syndrome did ulnar nerve entrapment appear clinically or neurophysiologically at the level of the two heads of flexor carpi ulnaris and distal to the olecranon notch. On the other hand, slowing of conduction velocity with ulnar nerve entrapment at the elbow usually appeared to be maximal either just proximal to the olecranon notch or, more often, within the level of the olecranon notch itself. Thus 341 of the 350 patients studied for ulnar nerve entrapment at the level of the elbow had neurophysiological findings indicating maximal lesioning just proximal to or in the notch. Also of interest were intraoperative recordings on patients with posterior interosseus nerve entrapments. The area of maximal abnormality in

190

Leo Happel and David Kline TABLE 8.6 Other In Continuity Cases Studied (1965–1990) Intraneural Tumors of Brachial Plexus

80

Intraneural Tumors of Other Nerves

165

Entrapments

617

Cranial Nerve Injuries

62 (VII,XI,XII)

conduction, while usually beginning at the arcade of Frohse, appeared to extend beyond that level distally and beneath the actual volar head of the supinator itself. Some unusual entrapments or functional lesions to nerve have been further documented by intraoperative recording (see also Table 8.6). These have included radial nerve lesions at the level of the long head of the triceps, median as well as ulnar nerve entrapments by Struthers’ ligament, and irritative as well as compressive sciatic lesions just below the buttocks crease due to hamstring hypertrophy. There were many more lesions of plexus spinal nerves where thoracic outlet syndrome was suspected and intraoperative recordings showed conductive defects. These areas of reduced conduction velocity were more dramatic on the lower roots (especially C8 and T1) but at times were seen at C7 as well. Conductive defects in these patients began at a spinal nerve or spinal nerve to trunk level but not more distally. By comparison, conduction velocities and amplitudes recorded from C5, C6, and usually C7 roots were almost always greater than those in lower roots in the patients with “true” thoracic outlet syndrome. In these cases there was often some weakness of hand intrinsic muscles in both the median and ulnar nerve distributions.

9 TROUBLESHOOTING The operating room is generally regarded as a hostile setting for neurophysiological recording using electronic instrumentation. It is likely that those starting a program of intraoperative neurophysiology will encounter some problems, at least initially, and have to perform the task of troubleshooting. Troubleshooting involves observation of existing conditions which may be problematic and knowing how to effectively deal with them. The powers of observation cannot be overemphasized. The sources of problems vary widely, though they can be put into several general categories. They may come in the form of the spontaneous and continuous electrical noise, which prevents recordings, or, more commonly, in the form of an inability to stimulate and record from sections of nerve that are known to be normal. This section is intended to provide a basis for dealing with these problems.

Intraoperative Neurophysiology of the Peripheral Nervous System

191

With electrodes applied to the nerve and the instrumentation adjusted to the settings described previously, one should view the display of the recording equipment. With the intensity of the stimulator turned all the way down, the trace should be relatively flat. If not, and the trace displays large regular and continuous excursions, there may be several sources for the interfering signal. The most common of these is 60 Hz interference from electrical power sources. This can be readily identified by temporarily increasing the display to 10 or 20 ms per division. The most offensive devices would be those that contain electric motors. Hospital beds, pumps, and hot-air or fluid warmers are good examples. Turning these devices off may not prevent the interference, however, and they may have to be unplugged. Although older forms of fluorescent lighting were a significant problem in the past, modern fluorescent lighting rarely presents a problem. Sometimes, however, x-ray view boxes can produce an artifact, and these should be turned off when the problem is identified. Methodically unplugging, briefly, each of the devices identified as a possible source of the problem may help to eliminate the source of noise. If the source of the noise cannot be found, the electrodes should be disconnected from the EMG machine while the EMG machine is still recording. If the noise remains, it is most likely originating from the instrumentation itself, arriving there through electrical power lines. It may be necessary to plug the EMG machine into a different outlet. More commonly, however, the interference will disappear when the recording electrodes are unplugged, indicating that its source is from the recording electrodes. One should inspect the routing of the wires from the recording electrodes as they are passed off of the sterile field. If these wires are placed close to the power cords of other equipment, they may be the source of the interference. These wires should preferably be suspended in air from the operating table to the recording input; they should not be placed adjacent to any other metallic objects. In addition, these wires should not move, either from evoked muscle activity in the patient or simply from air currents around them. Movement, by itself, will produce electrical interference. Current protocols specified by the Joint Commission for the Accreditation of Hospitals ( JCAH) require that patients on the operating table must not be connected to any true ground. Thus the reference for the electrosurgical unit (Bovie) and any other electrical equipment connected to the patient should not be grounded. The operating table itself is not grounded. Attaching the isolated ground from the EMG machine to the patient may help to eliminate the source of noise under these circumstances. This may be done using a large-surface-area disposable ground pad attached to the patient’s body at a point that is as close to the recording site as is convenient. With the recording electrodes attached to the EMG machine, the surgeon should be able to make contact between tissue and both recording electrodes, as the trace of the recorder remains flat. If there is a great deal of difference in the

192

Leo Happel and David Kline

amount of noise displayed on the recorder as the surgeon touches both recording electrodes to the patient, there may be a broken or bad connection between the EMG machine and the recording electrodes. If only one of these electrodes actually makes connection with the patient, it will lead to excessive amounts of noise. This may occur if one of the wires to the electrodes is broken or if there is poor contact between a wire and the plug attached to it. Bad electrical contact between the nerve and the electrodes will lead to a similar result and may occur if the electrodes have not been properly cleaned. It may be helpful to scrape the stainless steel surface of the electrodes that contact the nerve. This will produce a low-resistance junction that facilitates stimulation and recording. Other sources of spontaneous, continuous interference include radio transmitting devices (telemetry), electrosurgical units, and spontaneous EMG. These sources of interference are high frequency and tend to fill the screen of the recorder. They may or may not be regular in appearance. In this case, it may be necessary to use the filters on the EMG machine to attenuate the noise, as was discussed previously. Occasionally, some unusual sources of interfering noise can be identified, including implanted stimulators and pacemakers. Another challenge to monitoring is artifact related to the stimulus. Excessive stimulus artifact can be caused by a loss of stimulus isolation (which has already been discussed) or by improper filter settings. Insufficient distance between stimulating and recording electrodes or insufficient distance between recording electrode tips may also lead to excessive stimulus artifact. A lack of adequate separation between stimulating and recording cables as they lead off the sterile field may capacitively produce excessive stimulus artifact. High-intensity stimulation or using long-duration stimulus pulses may also contribute to stimulus artifact problems. Although all recordings should contain some degree of stimulus artifact, it should not be so great as to prevent visualization of the CNAP immediately following it. In fact, if no stimulus artifact can be seen, it may be an indication of insufficient amplification or a failure of stimulation. As one becomes familiar with this technique, the appearance of a modest amount of stimulus artifact is comforting. Another feature of recordings that one quickly adjusts to is the sweep speed of the recording instrument. This should be adjusted to approximately 1 ms/cm or a total sweep length of approximately 10 ms. If the display is set for too long a window of time, the CNAP will be lost in the stimulus artifact at the very beginning of the trace. Those accustomed to viewing SEPs will quickly appreciate that the CNAP has a much shorter latency and is a much faster event. The sweep required to view this event must be considerably faster than that required for the SEP. In an effort to expedite troubleshooting, a flowchart is presented in Table 8.7. Following the steps of this flowchart in a methodical way should permit one to identify problems quickly.

TABLE 8.7 START

Select Nerve level (min. 4 cm. Length)

Set Stim = 6-8 V Set Rec. = 100 mV/cm

Suspend Nerve on Electrodes Check electrode connections Check stim. & Rec. settings Possible recorder malfunction

NO

Can stimulus artifact be seen at the beginning of the trace?

NO

CNAP?

YES

EVALUATION:

YES YES

Potential may be artifact or electrical noise

Is there too much stimulus artifact?

NO

“Response” always appears with exactly the same latency (appears “frozen” on screen)

Potential may be muscle activity

NO

Amplitude at maximum less than 2 mV

Potential may be stimulus artifact

NO

Does potential show threshold behavior as stimulus is increased and decreased?

Potential may be artifact or stimulus may be spreading

NO

Does potential show maximum size with increased stimulation?

Potential may be muscle action potential

NO

Can potential be seen without muscle activity near the recording electrodes?

Potential could be muscle action potential

NO

NO

Check: Appropriate nerve length? Good electrode contact with nerve? Separate Stim. & Rec. leads from electrodes? Shorten stimulus duration Check filter settings

YES

Set Stim. = 25- 35 V Set Rec. = 20 mV/cm

YES

CNAP?

YES

Go to EVALUATION

YES NO Test System: Check a segment of nerve which is known to be functional Check: Are electrode connections good? Is stimulator working touch exposed muscle; does it contract? Is recorder working some stimulus artifact at beginning of trace?

YES NO

CNAP?

YES

Previous segment shows no nerve fiber activity?

YES Retest previous segment

193

Set Stim. = 70 V Set Rec. = 10 mV/cm

No Significant Axons Present!! TESTING COMPLETE

Waveform of CNAP is less than 2 mSec duration and not polyphasic? YES

NO

CNAP? YES

Go to EVALUATION

Nerve activity indicates significant fibers present at this level EVALUATION COMPLETE PROCEED TO ANOTHER LEVEL

194

Leo Happel and David Kline

10 CONCLUSIONS Intraoperative neurophysiology is a new and exciting field that provides functional information to the surgical team. Despite the development of sophisticated new imaging techniques, these cannot provide the same kind of information that neurophysiological studies can. With respect to peripheral nerve, intraoperative neurophysiology provides diagnostic as well as prognostic information that cannot be learned in any other way. Preoperative EMG studies are very useful in evaluating the extent of a nerve injury, but even these cannot detect the electrical manifestations of very early regeneration. This can only be learned at the operating table. With this information in hand, the surgeon can decide on the proper course of action to treat the nerve injury. The assurance provided by these recordings gives him or her the proper feedback that his or her decisions are correct. The end result is that the patient will receive the benefits of surgery that will produce the best prospect for optimal recovery. As with any new procedure, there will be apprehensions with implementation. The novelty of a new technology in the operating room requires some adjustments and accommodation on the part of the surgical team. However, operative recording of peripheral nerve activity provides useful information concerning nerve function at the time of surgery, and the results are certainly worth the small amount of extra effort required to obtain them. These recordings can be made quickly and reliably and represent an effective means of assessing the status of a segment of peripheral nerve. They provide assurance to the surgeon that the difficult decisions that must be made to deal with a lesion in continuity are based on good information and are not simply guided by intuition.

REFERENCES 1. Kline, D.G., and Happel, L.T. (1993). Penfield Lecture. A quarter century’s experience with intraoperative nerve action potential recording. Can. J. Neurol. Sci., 20(1), 3–10. 2. Arai, M., Goto, T., Seichi, A., Miura, T., and Nakamura, K. (2000). Comparison of spinal cord evoked potentials and peripheral nerve evoked potentials by electric stimulation of the spinal cord under acute spinal cord compression in cats. Spinal Cord, 38(7), 403–408. 3. Carter, G.T., Robinson, L.R., Chang, V.H., and Kraft, G.H. (2000). Electrodiagnostic evaluation of traumatic nerve injuries. Hand. Clin., 16(1), 1–12, vii. 4. Grant, G.A., Goodkin, R., and Kliot, M. (1999). Evaluation and surgical management of peripheral nerve problems. Neurosurgery, 44(4), 825–839; discussion 839–840. 5. Happel, L.T., and Kline, D.G. (1991). Nerve lesions in continuity. In “Operative nerve repair and reconstruction” (R.H. Gelberman, ed.), 1st ed, vol. 1, pp. 601–616. J.B. Lippincott, Philadelphia. 6. Kline, D.G. (1990). Surgical repair of peripheral nerve injury. Muscle Nerve, 13(9), 843–852.

Intraoperative Neurophysiology of the Peripheral Nervous System

195

7. Spinner, R.J., and Kline, D.G. (2000). Surgery for peripheral nerve and brachial plexus injuries or other nerve lesions. Muscle Nerve, 23(5), 680–695. 8. Kline, D.G., and Hudson, A.R. (1995). “Nerve injuries,”1st ed. W.B. Saunders, Philadelphia. 9. Williams, H.B., and Terzis, J.K. (1976). Single fascicular recordings: An intraoperative diagnostic tool for the management of peripheral nerve injuries. Plastic and Reconstructive Surgery, 57, 562–569. 10. Oberle, J.W., Antoniadis, G., Rath, S.A., and Richter, H.P. (1997). Value of nerve action potentials in the surgical management of traumatic nerve lesions. Neurosurgery, 41(6), 1337–1342; discussion 1342–1344. 11. Chaudhry, V., and Cornblath, D.R. (1992). Wallerian degeneration in human nerves: Serial electrophysiological studies. Muscle Nerve, 15(6), 687–693. 12. Lundborg, G., and Danielson, N. (1991). Injury, degeneration, and regeneration, In “Operative nerve repair and reconstruction” (R.H. Gelberman, ed.), 1st ed., vol. 1, pp. 109–132. J.B. Lippincott, Philadelphia. 13. Lundborg, G., and Dahlin, L. (1991). Structure and function of peripheral nerve. In “Operative nerve repair and reconstruction” (R.H. Gelberman, ed.), 1st ed., vol. 1, 3–18. J.B. Lippincott, Philadelphia. 14. Kline, D.G., Kim, D., Midha, R., Harsh, C., and Tiel, R. (1998). Management and results of sciatic nerve injuries: A 24-year experience. J. Neurosurg., 89(1), 13–23. 15. Dorfman, L., and Cummins, K.L. (1981). “Conduction velocity distributions: A population approach to electrophysiology of nerve.” W.R. Liss, New York. 16. Galbraith, J.A., and Myers, R.R. (1991). Impulse conduction. In “Operative nerve repair and reconstruction” (R.H. Gelberman, ed.), vol. 1, pp. 19–45. J.B. Lippincott, Philadelphia. 17. Mogyoros, I., Kiernan, M.C., and Burke, D. (1997). Strength-duration properties of sensory and motor axons in carpal tunnel syndrome. Muscle Nerve, 20(4), 508–510. 18. Wall, E.J., Massie, J.B., Kwan, M.K., Rydevik, B.L., Myers, R.R., and Garfin, S.R. (1992). Experimental stretch neuropathy: Changes in nerve conduction under tension. J. Bone Joint Surg. Br., 74(1), 126–129. 19. Rochkind, S., and Alon, M. (2000). Microsurgical management of old injuries of the peripheral nerve and brachial plexus. J. Reconstr. Microsurg., 16(7), 541–546. 20. Tiel, R.L., Happel, L.T. Jr., and Kline, D.G. (1996). Nerve action potential recording method and equipment. Neurosurgery, 39(1), 103–108; discussion 108–109. 21. Holland, N.R., Lukaczyk, T.A., Riley, L.H. III, and Kostuik, J.P. (1998). Higher electrical stimulus intensities are required to activate chronically compressed nerve roots. Implications for intraoperative electromyographic pedicle screw testing. Spine, 23(2), 224–227. 22. Zhao, S., Kim, D., and Kline, D. (1993). Somatosensory evoked potentials induced by stimulating a variable number of nerve fibers in the rat. Muscle Nerve, 16, 1220–1227.

CHAPTER

9

Intraoperative Neurophysiological Monitoring of the Sacral Nervous System DAVID B. VODUSˇ EK University Institute of Clinical Neurophysiology, University Medical Centre, Ljubljana, Slovenia

VEDRAN DELETIS Division of Intraoperative Neurophysiology, Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York

1 Introduction 2 Functional Anatomy 2.1 Neural Control of the Lower Urinary Tract 2.2 Anorectum 2.3 Sexual Organs 3 Clinical Neurophysiological Tests in Diagnostics 4 Intraoperative Clinical Neurophysiology 4.1 Basic Technical Aspects of Stimulation for Intraoperative Sacral Monitoring 4.2 Basic Technical Aspects of Recording for Intraoperative Sacral Monitoring 4.3 Specific Sacral Neuromuscular System Monitoring Procedures 5 Discussion and Conclusions References

ABSTRACT In the first part of this chapter, the basic functional neuroanatomy of the genitourinary and anorectal systems is briefly described. These systems are involved in the socalled sacral functions (micturition, defecation, erection, ejaculation, etc.). In this section we describe clinical neurophysiological tests of the functional integrity of the Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

197

198

David B. Voduˇsek and Vedran Deletis sacral neuromuscular system used for diagnostic purposes. The second part of this chapter deals with intraoperative neurophysiological monitoring of the lumbosacral nervous system. The validity of different intraoperative monitoring techniques of this system is summarized.

1 INTRODUCTION The functions involving the genitourinary and anorectal systems are uniquely controlled by the complex interaction of the vegetative and the somatic nervous system. Insofar as it is the sacral parasympathetic and somatic systems that constitute the most important peripheral nervous structures controlling these functions, they may also be referred to as sacral functions. The functions themselves (micturition, defecation, erection, etc.) are now better understood because we have applied methods of measuring the different functional parameters (urodynamics, faecodynamics, measurements of the sexual response) that provide for better diagnosis of dysfunction. The awareness that such dysfunction is also a consequence of damage to neural structures has also greatly increased. On the other hand, it has become possible to better define the various lesions to the nervous system by electrophysiological methods. However, these methods by and large document only the somatic sacral nervous system and its central pathways [1]. Nevertheless, such information is clinically relevant because (1) the somatic nervous system plays a part in all sacral functions, and (2) the somatic and parasympathetic sacral systems are closely related, and information on the somatic system may therefore be a relevant indicator of the overall neurogenic lesion in several clinical situations. In fact, the common denominator of lower urinary tract, anorectal, and sexual functions is that many of their efferent and afferent pathways travel at least partly in close vicinity. They “share” common spinal cord regulatory segments (the upper lumbar segments—sympathetic efferents; the middle and lower sacral segments—parasympathetic and somatic efferents). Even the long pathways connecting the relevant spinal cord segments with higher levels of the central nervous system are situated close together. Thus it is not uncommon that in lesions to the nervous system, and particularly in lesions affecting the spinal cord, cauda equina, sacral plexus, and pudendal nerve, the sacral functions are affected together. Dysfunctions may arise not only from disease or trauma, but also from inadvertent lesions to the previously named structures during invasive procedures, particularly several surgeries involving the pelvic organs and the spinal canal. The consequences of lesions to the neurocontrol of sacral functions include disturbing sensory phenomena like pain, dysethesiae, urgency and frequency, loss of genital sensation, bladder or rectal fullness, and subsequent retention, obstipation, soiling, incontinence, and erectile dysfunction. Neurogenic damage may lead to lost coordination between detrusor

Intraoperative Neurophysiological Monitoring of the Sacral Nervous System

199

and sphincter function, leading to high bladder pressures and upper urinary tract dysfunction. Incontinence may—particularly in motor-disabled persons— lead to problems with hygiene, skin problems, infections, and decubiti. Although neurogenic sexual dysfunction may not be life-threatening, it can be extremely disruptive psychologically and can lead to severe emotional and interpersonal problems. All these may be particularly tragic when they are a consequence of an inadvertent intraoperative lesion. A whole array of clinical neurophysiological diagnostic methods has been modified for use in the anogenital area, including electromyographic methods and reflex, conduction, and evoked potential studies [2,3]. These methods are routinely employed in uroneurological and neuro-urological laboratories for diagnostics and follow-up in patients with (suspected) neurogenic sacral dysfunction. The authors have pioneered with trials to establish some of these neurophysiological methods in the operating room, to help the surgeon identify particular sacral nervous structures, and to monitor the function of the sacral neuromuscular system during surgery [4–7].

2 FUNCTIONAL ANATOMY Although the functional anatomy of the genitourinary and anogenital systems is highly complex, it need not be considered in detail because only the gross anatomy of the relevant somatic nervous structures can be approached by clinical neurophysiological methods that are applicable in the operating room environment. Most of the information relevant for intraoperative neurophysiology can be summarized as follows. Afferent fibers from the mucosa and skin of the genitoperineal region travel mostly with the pudendal nerves. The distally most accessible group of sensory fibers are the dorsal nerves of the penis (or clitoris). The sensory fibers from the genital, perineal, and anal region enter through the dorsal spinal roots S2–S5 into the spinal cord and synapse (through interneurons) with sphincteric motor neurons. The afferent information also ascends (the primary sensory neurons synapsing to higher-order sensory neurons at various levels) via the spinothalamic and dorsal column tracts, the lemniscal system and thalamocortical tracts, and finally to the somatosensory cortex (at its interhemispheric location) [8]. The sphincteric lower motor neurons in the midventral spinal grey matter of the second to fourth sacral spinal cord segments (the “Onuf nucleus”) are under voluntary control from the motor cortex. Somatic motor nerve fibers leave through the ventral roots and the sacral plexus, combining into the pudendal nerves; direct branches innervate the levator ani and the anal sphincter [8]. The external urethral sphincter may be innervated by sacral somatic fibers traveling via splanchnic nerves [9] or the pudendal nerve [10], or possibly both.

200

David B. Voduˇsek and Vedran Deletis

2.1 NEURAL CONTROL OF THE LOWER URINARY TRACT The lower urinary tract (LUT) is innervated by three sets of peripheral nerves. The pelvic parasympathetic nerves arise at the sacral level of the spinal cord (they excite the bladder and relax the urethra). The sympathetic nerves arise from the upper lumbar segments and inhibit the bladder body, modulate transmission in bladder parasympathetic ganglia, and excite the bladder base and urethra. The somatic efferents and afferents from the S2–S4 sacral roots innervate pelvic floor muscles (levator ani) both through direct branches and by the pudendal nerve, which innervates also the perineal muscles, including the anal and urethral sphincter. All of these nerves contain both efferent and afferent nerve fibers that are controlled by centers in the brain and particularly important centers in the brainstem. Long tracts in the spinal cord subserve the spinobulbospinal reflex pathway, which is relevant for coordinated detrusor-sphincter function and normal micturition. The dorsal pontine tegmentum is established as an essential control center for micturition (with a close anatomical relationship with the locus coeruleus). While different types of sensation of the lower urinary tract travel both in the anterolateral and the dorsal part of the spinal cord, the descending (motor) pathways lie within the lateral aspects of the spinal cord.

2.2 ANORECTUM Touch, pin-pricks, and hot and cold stimuli can be perceived in the anal canal to a level of up to 15 mm above the anal valves. The epithelium in the area from about 10–15 mm above the valves has a rich sensory nerve supply made up of both free and organized nerve endings. The sensory endings in the hairy perianal skin are similar to those in hairy skin elsewhere. The afferent nerve pathway for anal canal sensation is by the inferior hemorrhoidal branches of the pudendal nerve. Sensory pathways from the rectum and the bladder travel in the pelvic visceral nerves to the sacral cord, but some afferent information is probably also related to hypogastric nerves entering the spinal cord at the thoracolumbar level. Functionally, the most important part of the smooth musculature of the anorectum is the internal anal sphincter, which is responsible for about 85% of the resting pressure in the lumen of the canal. The smooth musculature of rectal walls (and of the detrusor) receives extrinsic motor innervation from the sacral parasympathetic outflow arising in the intermediolateral cell columns of sacral cord segments S2–S4. These first-order neurons send axons that emerge with

Intraoperative Neurophysiological Monitoring of the Sacral Nervous System

201

the ventral spinal nerve roots to synapse with second-order neurons lying within the pelvic plexus or the visceral walls. The sympathetic nerve supply arises from the thoracolumbar chain and travels in the hypogastric nerve to innervate visceral smooth muscle directly, and also via a modulatory influence on parasympathetic function at the level of the pelvic plexus. The internal anal sphincter is probably controlled both by sympathetic (hypogastric) and sacral parasympathetic pathways, but the inhibition brought about by rectal distention (the important rectoanal inhibitory reflex) is predominantly an intramural one. The external anal sphincter is innervated by the pudendal nerve and occasionally also by a perineal branch of S4. The neurons of the sphincter motor nucleus (Onuf’s nucleus) are under voluntary control via corticospinal pathways. Normal defecation is probably triggered by filling of the rectum from the sigmoid colon, and the signals from stretch receptors in the rectal wall and pelvic floor muscles are interpreted at the conscious level as a desire to defecate. The extension of the rectum causes reflex relaxation of the smooth internal sphincter muscle. Voluntary relaxation of the striated sphincter muscle permits defecation, which is assisted by colonic pressure waves and abdominal straining. If defecation is to be deferred, brief conscious contraction of the voluntary sphincter allows time for recovery of internal sphincter tone and relaxation of the rectum to accommodate filling. Conscious appreciation of the desire to defecate and intentional control over defecation are conferred by suprasacral neural influences. The precise way in which the autonomic, pyramidal, extrapyramidal, and sensory pathways integrate to achieve a reliable and predictable anorectal function is not yet fully understood [11].

2.3 SEXUAL ORGANS Of the sexual functions affected by neurogenic lesions, research has centered on the male functions, and particularly on erection. Erection can be initiated in the brain and/or follow genital stimulation; in sexual activity a combination of both is probably involved. Neurogenic erectile dysfunction due to peripheral lesions can be secondary to the disruption of sensory nerves contributing to the afferent arm of reflex erection or to the disruption of autonomic nerves that mediate arterial dilatation and trabecular smooth muscle relaxation. Erectile dysfunction can occur from disruption of the relevant pathways in centers within the spinal cord (both suprasacral and sacral), cauda equina, the sacral plexus, the pelvic plexus, the cavernosal nerves, and the pudendal nerves. Particular pelvic surgeries such as radical prostatectomy or cystoprostatectomy lead to a high percentage of

202

David B. Voduˇsek and Vedran Deletis

mostly neurogenic erectile dysfunction; the lesion occurs in the pelvic plexus or in the cavernosal nerves located in the posterolateral aspect of the prostate. Ejaculation can be abolished by a lesion to the sympathetic innervation of the bladder neck (leading to a retrograde ejaculation) and by disruption of the sensory and (particularly) motor nerves innervating the perineal muscles, whose contraction leads to expulsion of the semen. It can also be abolished by central lesions. A disturbed sexual response in females is due to (1) afferent lesions leading to loss of sensitivity of the perineal area, and (2) efferent lesions leading to a loss of lubrication, loss of clitoral erection, and pelvic floor muscle denervation.

3 CLINICAL NEUROPHYSIOLOGICAL TESTS IN DIAGNOSTICS Since the function of all the aforementioned systems relies on neural control, clinical neurophysiological tests have been introduced to support and supplement clinical evaluation in patients. The tests comprise electrophysiologic methods of testing conduction through motor and sensory pathways (both peripheral and central) and electromyographic methods. Traditionally, in testing both the lower urinary tract and anorectal function, the EMG signal obtained from sphincter muscles has been used to delineate the sphincter activity patterns in relationship to micturition or defecation. In addition to that, electromyographic methods have been used to distinguish between normal and neuropathic pelvic floor muscles. Conduction tests have been introduced to evaluate the integrity of different reflex pathways (sacral reflexes), the individual motor pathways (pudendal nerve terminal latency, MEP), and sensory pathways (penile sensory neurography, SEP). In addition, autonomic tests have also been introduced (sympathetic skin response, corpus cavernosum EMG). For diagnostic purposes a single testing is performed without knowledge of the previous status of the investigated structure. In this diagnostic situation, results have to be compared to values obtained from healthy subjects. The tests of conduction have been found to be relatively insensitive to axonal lesions because amplitudes of responses vary widely in the control population (particularly due to technical reasons), and conduction may remain normal in partial lesions. Thus, in the diagnostic situation, the ability of the concentric needle EMG to detect abnormal spontaneous activity as an indicator of denervation, and changes of motor unit potentials as indications of reinnervation, has been found to be particularly helpful. EMGs and recordings of the bulbocavernosus reflex (indicating the potency of the lower sacral reflex arc) have been proposed as the basic battery of tests for evaluation of patients with sacral dysfunctions and suspected neurogenic involvement [12]. From conduction tests, only recordings

Intraoperative Neurophysiological Monitoring of the Sacral Nervous System

203

of the sacral reflex and SEP after dorsal penile or clitoral nerve stimulation have been suggested since they have been validated by extensive clinical studies. They may be of value in selected patients with suspected peripheral (i.e., bulbocavernosus reflex testing) and central nervous system (i.e., SEP testing) lesions [3, 13]. The other neurophysiological tests have been suggested as useful in further research. The corpus cavernosum EMG is the most controversial of the tests so far described. It is not yet well clarified whether the signal really originates from penile smooth muscle; validation of the method would offer a most important source of information on penile innervation status, which is necessary for erection.

4 INTRAOPERATIVE CLINICAL NEUROPHYSIOLOGY The authors have demonstrated that with appropriate modifications of methods it is technically feasible to record in anesthetized patients (a) dorsal root action potentials (DRAPs) after pudendal nerve stimulation, (b) pudendal somatosensory evoked potentials over the conus, the spinal cord, and the scalp, (c) sphincter muscle EMG responses to sacral ventral root stimulation and motor cortex stimulation, and (d) the bulbocavernosus reflex. Only one of the aforementioned techniques has been used extensively enough to gather pertinent information regarding the practical relevance of sacral nervous system monitoring during surgical interventions. However, the techniques are expected to be valuable safeguards against inadvertent lesioning of nervous structures that would lead to some (neurogenic) dysfunction of micturition, defecation, or the sexual response. Further studies are required to clarify these issues.

4.1 BASIC TECHNICAL ASPECTS OF STIMULATION FOR INTRAOPERATIVE SACRAL MONITORING In order to obtain bioelectrical signals useful for monitoring purposes in the different segments of the sacral neuromuscular system, it is necessary to depolarize the nervous system at particular segments. Up to now only electrical stimulation has been appropriate for this purpose. Stimulation can be applied to either the sensory part or the motor part of the system (afferent versus efferent events; Fig. 9.1). At present, most intraoperative monitoring of the sacral system has relied on responses evoked from stimulation of the sensory system, apart from recording of anal sphincter muscle responses upon stimulation of ventral spinal roots or the motor cortex.

204

David B. Voduˇsek and Vedran Deletis

FIGURE 9.1 Neurophysiological events used to intraoperatively monitor the sacral nervous system. Left, “afferent” events after stimulation of the dorsal penile or clitoral nerves and recording over the spinal cord: (1) pudendal SEPs, traveling waves, (2) pudendal DRAPs, and (3) pudendal SEPs, stationary waves, recorded over the conus. Right, “efferent” events: (4) anal M wave recorded from the anal sphincter after stimulation of the S1–S3 ventral roots, (5) anal motor-evoked potentials recorded from the anal sphincter after transcranial electrical stimulation of the motor cortex, and (6) bulbocavernosus reflex obtained from the anal sphincter muscle after electrical stimulation of the dorsal penile or clitoral nerves. Reprinted from Deletis, V. (2001). Neuromonitoring. In “Pediatric neurosurgery,” 4th ed. (D. MacLeone, ed.), pp. 1204–1213. W.B. Saunders, Philadelphia.

Intraoperative Neurophysiological Monitoring of the Sacral Nervous System

205

The appropriate peripheral sensory structures that are available for stimulating purposes are the two adjacent dorsal penile or clitoral nerves. These nerves are stimulated by silver/silver chloride cup EEG-type electrodes (EEG electrodes) placed on the dorsal surface of the penis or clitoris (this electrode representing the cathode). The other electrode (anode) is placed either distally on the penis (1–2 cm apart from the proximal electrode) or on the adjacent labia (Fig. 9.2). In small children with short penises, the anode may be attached on the ventral side of the penis. The dorsal side of the penis must be scrubbed gently with Nuprep (D. O. Weaver & Co., Aurora, CO) before placing electrodes in order to avoid stimulus artifacts. The electrodes are filled with electrode cream and secured appropriately (Tegaderm; Smith and Nephew Medical Limited, Hull, England). The electrode sites are then bandaged with a few layers of

FIGURE 9.2 Lower left, position of the electrodes over the clitoris and labia majora for the stimulation of the dorsal clitoral nerves. Lower right, position of the electrode for stimulating the dorsal penile nerves. R1 = recording BCR from anal sphincter. Upper right, schematics of recording DRAPs with a hand-held hook electrode (R2) over the exposed dorsal sacral roots of the cauda equina. Upper left, intraoperative picture of DRAP recordings.

206

David B. Voduˇsek and Vedran Deletis

gauze to prevent them from being displaced when the patient is moved onto the operating table. Electrode impedances should be kept below 5 kΩ. As a general rule, stimuli of 20 mA intensity at 0.2 ms duration have been delivered in procedures involving stimulation of the penis or clitoris (at various frequencies for different measurements, up to 13.3 Hz). Stimulation can also be performed at the level of the spinal roots; a handheld sterile monopolar electrode can be placed under the appropriate roots (or rootlets, after the root is freed from neighboring roots and lifted outside the spinal canal). Square wave pulses of 1 to 2 mA intensity and 0.2 ms duration are delivered (Fig. 9.1, right).

4.2 BASIC TECHNICAL ASPECTS OF RECORDING FOR INTRAOPERATIVE SACRAL MONITORING In our practice, bioelectrical activity from the sacral neuromuscular system has up to now been recorded from the sacral dorsal spinal roots, the spinal cord, the somatosensory cortex, and the anal sphincter muscle. Recordings from dorsal spinal roots (dorsal root action potentials, DRAPs) are obtained by hand-held sterile bipolar hook electrodes (after the root is freed from neighboring roots and lifted outside the spinal canal). In this case, the electrode closer to the point of stimulation is the G1 (active) electrode. Epoch lengths of 0 to 50 ms are used for these recordings (Figs. 9.1, 9.2). Recordings of pudendal spinal somatosensory evoked responses (SSEPs, stationary wave) are obtained by a spinal epidural electrode placed over the conus (S2–S4). These potentials are generated by interneurons of the grey matter within the S2–S4 segments of the spinal cord. Typically 100 responses are averaged together; epoch lengths of 0–50 ms are used (Fig. 9.1). Recordings of pudendal spinal somatosensory evoked responses (SSEPs, traveling wave) are obtained by a spinal epidural electrode inserted anywhere over the dorsal column of the spinal cord. These potentials are generated by pudendal afferents traveling within the dorsal columns. Typically 100 responses are averaged together; epoch lengths of 0–50 ms are used (Fig. 9.1). To obtain pudendal cerebral somatosensory evoked responses (CSEPs), Screwtype recording electrodes are placed on the scalp 2 cm behind CZ (G1 or active electrode) and at FZ (G2 or reference electrode), according to the somatotopic representation in the primary somatosensory cortex related to the International 10–20 System of scalp electrode placement. The active electrode is placed in the midline because the sacral segments are represented deep within the medial longitudinal interhemispheric fissure. For CSEPs, 100–200 traces are typically averaged together. Epoch lengths of 0–200 ms are used.

Intraoperative Neurophysiological Monitoring of the Sacral Nervous System

207

To record anal sphincter (EMG) responses, either surface-type electrodes or hook wire electrodes can be used. Given the close anatomical relationship between the small sphincter muscle and neighboring larger muscle groups, recording must be selective (e.g., with intramuscular hook wire electrodes) if the stimulation technique is “nonselective” (such as in the case of ventral root stimulation, as a consequence of which neighboring muscles are also excited). When the stimulation procedure is more specific (e.g., in bulbocavernosus or pudendoanal reflex monitoring), the recording may be obtained with properly attached surface-type electrodes. Sterile hook wire recording electrodes (Teflon-coated, 76 µm diameter wire with a 3 mm bare tip) are introduced into the left and right sides of the external anal sphincter with sterile needles; these are immediately removed carefully from the sphincter (the hooked wires remaining in place). The integrity of the electrodes can be tested by passing a short train of 50 Hz current at 10 mA and observing sphincter contraction (if the patient is not paralyzed at the time of electrode placement). The electrode impedances of these electrodes should be checked, although clean recordings are usually still possible with high electrode impedances. The epoch length used for anal sphincter EMG recordings varies according to the type of response; either single or few averaged responses can be obtained upon stimulation (of course, the patient should not be under the influence of a muscle relaxant during this procedure).

4.3 SPECIFIC SACRAL NEUROMUSCULAR SYSTEM MONITORING PROCEDURES 4.3.1 Pudendal Dorsal Root Action Potentials (DRAPs) In the treatment of spasticity (e.g., in cerebral palsy), the sacral roots are increasingly being included during rhizotomy procedures (see Chapter 10). Lang [14] demonstrated that children who underwent L2–S2 rhizotomies had an 81% greater reduction in plantar/flexor spasticity compared to children who underwent only L2–S1 rhizotomies. But, as more sacral dorsal roots have been included in rhizotomies, neurosurgeons have experienced an increased rate of postoperative complications, especially with regards to bowel and bladder functions. In order to spare sacral function, we attempted to identify those sacral dorsal roots that were carrying afferents from pudendal nerves. To do this, we used recordings of dorsal root action potentials (DRAPs). Patients were anesthetized with isoflurane, nitrous oxide, fentanyl, and a short-acting muscle relaxant introduced only at the time of intubation. The cauda equina was exposed

208

David B. Voduˇsek and Vedran Deletis

through a T12–S2 laminotomy or laminectomy and the sacral roots were identified using bony anatomy. The dorsal roots were separated from the ventral ones, and DRAPs were recorded by a hand-held sterile bipolar hooked electrode (the root being lifted outside the spinal canal) (Fig. 9.2). The DRAPs were evoked by electrical stimulation of the penile or clitoral nerves. One hundred responses were averaged together and filtered between 1.5 and 2100 Hz. Each average response was repeated to assess its reliability. Afferent activity from the right and left dorsal roots of S1, S2, and S3 was always recorded, along with occasional recordings from the S4–S5 dorsal roots. DRAP recording was successfully obtained in the majority of patients. In our first publication [4], the DRAPs were present in the S2 and S3 roots bilaterally in 19 patients, whereas in 7 patients DRAPs were also present in both S1 roots (in 8 patients they were present in the S1 root unilaterally). However, the response in the Sl root was never larger than the S2 or S3 root responses. The range of amplitudes was 2.9–18.3 µV in S1 roots, 3.2–129.9 µV in S2 roots, and 4.6–333 µV in S3 roots. Of special relevance was the finding that in 7.6% of these children, all afferent activity was carried by only one S2 root (Fig. 9.3, C and F). These findings were confirmed by a later analysis of the results of mapping in 114 children (72 male, 42 female, mean age 3.8 years) [5]. Mapping was successful in 105 out of 114 patients; S1 roots contributed 4.0%, S2 roots 60.5%, and S3 roots 35.5% of the overall pudendal afferent activity. The distribution of responses was asymmetrical in 56% of the patients (Fig. 9.3, B, C, and F). Pudendal afferent distribution was confined to a single level in 18% (Fig. 9.3, A), and even to a single root in 7.6% of patients (Fig. 9.3, C and F). Fifty-six percent of the pathologically responding S2 roots during rhizotomy testing were preserved because of the significant afferent activity, as demonstrated during pudendal mapping. None of the 105 patients developed long-term bowel or bladder complications. All our results in the early series of dorsal root mapping with 19 patients [4] have been confirmed by analysis of the larger series of 105 patients [5]. With this series we showed that selective S2 rhizotomy can be performed safely without an associated increase in residual spasticity, while at the same time bowel and bladder function are preserved by performing pudendal afferent mapping [14]. Therefore, we suggest that the mapping of pudendal afferents in the dorsal roots should be employed whenever these roots are considered for rhizotomy in children with cerebral palsy without urinary retention. Preoperative neurourological investigation of the children should help in making appropriate decisions; for example, in children with cerebral palsy with hyperreflexive detrusor dysfunction, sacral rhizotomy may be considered to alleviate the problem. In any case, intraoperative mapping of sacral afferents should make selective surgical approaches possible and provide the maximal benefit for children with cerebral palsy.

FIGURE 9.3 Six characteristic examples of DRAPs showing the entry of a variety of pudendal nerve fibers to the spinal cord via S1–S3 sacral roots. (A) Symmetrical distribution of DRAPs confined to one level (S2) or three levels (D). Asymmetrical distribution of DRAPS confined to the right side (B), only one root (C or F), or all roots except right S1 (E). Recordings were obtained after electrical stimulation of bilateral penile/clitoral nerves. 209

210

David B. Voduˇsek and Vedran Deletis

Our results so far support the hypothesis that the root distribution of afferent fibers that are important for the control of micturition may be similar to the distribution of mucocutaneous afferent fibers from the pudendal nerve. Only further studies will clarify the complex functional anatomic issues involved. 4.3.2 Pudendal Spinal Somatosensory Evoked Responses (SSEPs) Recordings were performed in four children of both sexes, 2.5 to 7.0 years of age. Small-amplitude (up to 1 µV) SSEPs, which were very stable, could be recorded with subdurally placed electrodes over the thoracic spinal cord (traveling waves), while a stationary wave could be recorded with a much higher amplitude (up to 10 µV) over the conus region. The latencies of the spinal SEP over the conus region ranged from 6.0 to 10.4 ms (Fig. 9.1, left). The recordings were made as a pilot study, and thus far only demonstrate the ability to obtain such recordings intraoperatively. Further employment of this technique showed that traveling waves are difficult to record, and successful recording of a stationary wave necessitates that the electrode be placed strictly over the S2–S4 sections of the spinal cord. 4.3.3 Pudendal Cerebral Somatosensory Evoked Potentials (CSEPs) Well-formed cerebral SEPs with amplitudes of 0.5–0.7 µV were recorded on dorsal penile nerve stimulation throughout spinal neurosurgical procedures in two adult male patients, 50 and 78 years of age. Stable P40 peaks were obtained. The recordings were made as a pilot study and thus far only demonstrate the ability to obtain such recordings intraoperatively. Further employment of this method showed that this potential is very sensitive to anesthetics; a formal feasibility study has not yet been performed.

4.3.4 Anal Sphincter Motor Response Monitoring In five children of both sexes, 2.5 to 9.0 years old, anal sphincter muscle EMG responses were recorded by stimulation of the ventral spinal roots (L5, S1, S2, S3, and S4) to identify ventral roots carrying motor fibers to the sphincter muscle. Recordings were obtained by surface conductive rubber electrodes (applied para-anally) and intramuscular hooked wire electrodes. In surface recordings, no unilateral responses could be identified, and responses were also obtained on stimulation of the L5 and S1 roots (on stimulation of L5 and S1 roots no adequate responses could be discerned from simultaneous recordings from intramuscular electrodes). The surface recorded responses were recognized as “nonspecific” (derived from neighboring muscles, most probably glutei [15]).

Intraoperative Neurophysiological Monitoring of the Sacral Nervous System

211

The latency of surface recorded responses was, as a rule, shorter than the latency of responses obtained from intramuscular electrodes, which was between 5 and 8 ms. On electrical stimulation of the motor cortex, anal sphincter responses were recorded in a large group of anesthetized patients without pyramidal involvement. Because of polysynaptic connections of the corticospinal tract to the α-motoneuron of S2–S4, these responses are moderately sensitive to anesthetics (Fig. 9.1). 4.3.5 Bulbocavernosus Reflex (BCR) Monitoring Intraoperative recordings were first performed in 15 neurosurgical patients (11 males, 4 females, 2–6 years old). Patients were without sacral dysfunction and were anesthetized with fenatyl and propofol or nitrous oxide without the influence of a muscle relaxant. Recordings from the anal sphincter were obtained by hooked wire electrodes and were recorded as a single response. Very reproducible responses could be obtained on double-pulse stimulation, the optimum interstimulus interval being found to be 3 ms and the optimum stimulation rate 2.3 Hz. Continuous periods of stimulation and recording for up to 10 min were repeatedly performed with very reproducible results (Fig. 9.4). The reflex response was suppressed by the administration of isoflurane and nitrous oxide and was completely abolished by muscle relaxants [16]. After this pilot study, 119 patients were tested, 38 of which underwent surgery without risk and 81 of which underwent surgery with risk of damage to sacral structures. In all, 51 adults (19 to 64 years old, 32 male and 19 female) and 68 children (24 days to 17 years old, 30 male and 38 female) took part in the study. Patients were anesthetized with propofol, fenatyl, or nitrous oxide with a short-acting relaxant. Clinically, most patients had mild to moderate upper motor neuron deficits in the lower extremities, and no patient had major urinary problems. In all patients it was possible to record reproducible reflex response with the previously described method. In patients without risk to the sacral system, only a few minutes of the responses were recorded to test their feasibility, whereas in the patients at risk, continuous monitoring was conducted. The influence of volatile anesthetic was tested in 6 patients. In 3 patients, BCR was suppressed by the administration of 1.25% isoflurane for 15 min (Fig. 9.5). Administration of nitrous oxide (60% inspired concentration) for 15 min reduced the BCR in 3 patients, and the response was completely abolished by muscle relaxant in 2 patients [6]. Preliminary results conducted in 50 patients with a lumbosacral tethered cord showed no clear correlation between BCR and postoperative sphincter control. The authors concluded that the complexity of sphincter innervation and segmental or suprasegmental control probably accounts for this discrepancy [17].

212

David B. Voduˇsek and Vedran Deletis

FIGURE 9.4 Continuous monitoring of the BCR for a period of 10 min showing the stability of the BCR’s appearance.

Up to now, close to 250 patients have been monitored using this method. In the last 100 patients, a train of four consecutive stimuli [7] was used rather than a double stimulus, providing even more robust responses. Some problems have been encountered in female patients in whom the stimulation method is not as robust as required; this problem needs particular attention from the technician.

5 DISCUSSION AND CONCLUSIONS The impetus to include intraoperative monitoring of sacral nervous structures came from concerns over optimizing care for children with cerebral palsy. These were successfully treated for their spasticity by performing selective dorsal rhizotomy; surgeons have always sought to minimize the side-effects of this procedure while maintaining its benefits (benefits being a reduction in muscle hypertonia, the side-effects being a partial or complete loss of sensory

Intraoperative Neurophysiological Monitoring of the Sacral Nervous System

213

FIGURE 9.5 The influence of isoflurane (Iso) on the BCR. Note that the response was almost completely blocked when the concentration of 1.25% was administered for 15 min and did not recover until almost 30 min after the isoflurane was discontinued. Reprinted from [6].

modalities). The first sacral dorsal roots have always been a target for rhizotomy in this procedure, but the S2 dorsal roots have become increasingly considered for rhizotomy. Not to treat the S2 dorsal roots is to leave potentially abnormal reflexive circuits that will continue to drive spasticity in the musculature of the leg. For this reason, the lesion zone was extended to include S2 dorsal roots, and the result was a greater reduction in spasticity as compared with children

214

David B. Voduˇsek and Vedran Deletis

in whom the lesion was extended only to the S1 segment. Unfortunately, the extension to include S2 roots was also associated with disorders in micturition. In one group of patients, before DRAP mapping was introduced, 24% experienced urinary retention (which was only transient in most children). Although most of these children had both the left and right S2 roots cut, 2 of the children had only one S2 root cut, and they also experienced retention. The selective dorsal rhizotomy procedure should be performed in young children who are certainly too young to assess sexual function; they are even too young to be completely confident that all symptomatic complaints regarding micturition or defecation and perineal sensation were being relayed. It was this concern over the preservation of genitourinary afferent function that led us to develop the technique of intraoperative neurophysiological identification of the sacral roots responsible for perineal sensation. In the first 31 children, neurophysiological identification of roots and rootlets carrying afferent activity from the penile or clitoral nerves led to zero micturition disturbances and allowed for rhizotomy of S2 roots or rootlets not carrying such afferent activity. Therefore, maximum possible antispastic effects could be achieved. The particularly important lesson we learned by doing very systematic recordings in S1–S2 and S3 roots bilaterally (and in some children also S4 and S5 roots) is that although most of our patients showed evidence of pudendal afferents in S2 and S3 roots bilaterally, about half of them also showed evidence of some afferent activity in S1 (either unilaterally or bilaterally). In slightly more than half of our patients, the root potentials were symmetrical, but the pudendal afferents of many were irregularly distributed across the sacral roots. In a few of the children the afferent activity was confined to a single root (either S2 or S3 root). Since this may be the most important afferent contribution from the genitourinary area, even if only the S2 dorsal roots are checked for relevant afferent activity, they should not be sacrificed if they show any such activity. The finding of asymmetrical distribution of fibers that may be confined to a single root is consistent with the work of Junemann et al. [18], who found that the majority of motor fibers for the urethral sphincter are carried by a single variable lower sacral root. Therefore, for some patients undergoing an operation in the area of the cauda equina, the sacrifice of one single root may have dire functional consequences. For other neurosurgical procedures in the lumbosacral spinal canal (e.g., the release of a tethered cord or the removal of a tumor), other authors have proposed, and in a small series have performed, the identification of motor and sensory nerve roots. This has been achieved through continuous monitoring of electromyographic activity in muscles innervated by lumbosacral segments, and through monitoring of tibial nerve somatosensory evoked potentials [19]. In addition, monitoring pudendal SEPs should provide very relevant complementary information, and, as we have demonstrated, should not be technically demanding if the structures are preserved preoperatively. Kothbauer et al. [19]

Intraoperative Neurophysiological Monitoring of the Sacral Nervous System

215

claim that intraoperative recordings saved operating time by allowing the surgeon more rapid and decisive preparation than would be possible on an anatomical basis alone, and they also gave the impression that the procedures were safer. The recordings of spontaneous anal sphincter electromyographic activity during such operations have been described earlier [20]. To accomplish the two goals of neurophysiology (i.e., first, immediate identification of structures as functional nervous tissue and their distinction from other tissue; and second, continuous monitoring of the function of the relevant nervous structures), a battery of methods needs to be applied, and a whole set of structures needs to be assessed bilaterally. Therefore, both the lower sacral segments and sphincter muscles, and also the upper sacral segments and the lumbar segments, need to be included. Also, all these segments and several functional modalities need to be monitored more or less simultaneously. The appropriate setup for each surgical situation would need to be selected on the basis of anatomical and physiological considerations, and a compromise between the possible and the necessary would be sought. Other authors have claimed that continuous EMG recording of bursts or trains of motor unit potentials or repetitive neurotonic discharges elicited by injury to the peripheral motor fibers have correlated with postoperative transient or permanent neurological deficit [21, 22] in the area of facial nerves. The predictive value of these “manipulation-evoked” discharges is only based on empirical data, but has been proposed to also be of value in the lumbosacral segments [19]. The electrophysiological identification of motor nerves is already an integral part of cranial nerve surgery [22] and should also provide a similar service in the region of the cauda equina. Up to now, only recordings of DRAPs have been made in a large number of subjects (to identify sacral roots carrying genitourinary afferents), and the electrophysiological procedure decreased postoperative voiding disturbances [4, 5]. We propose, however, that the other intraoperative electrophysiological recordings of the sacral neuromuscular system that have been described (spinal SEP, CMAP of sphincter muscles, bulbocavernosus reflex recordings) should prove relevant in surgeries involving the sacral roots, the cauda equina, and the conus and should aid the surgeon in preventing inadvertent damage to these structures. These other procedures have not, however, been performed in an adequate number of patients, and their relevance cannot yet be appraised. The most interesting question is whether monitoring of the bulbocavernosus reflex for conus and cauda equina surgery could, as we are proposing, replace and even improve upon separate monitoring of motor and sensory fibers of the lower sacral roots. As for surgeries involving the spinal cord above the conus, the procedures of pudendal SSEPs, cortical SEPs, and (possibly) motor evoked potentials of the anal sphincter muscle might be interesting in some selected patient groups.

216

David B. Voduˇsek and Vedran Deletis

These would include groups in whom the preservation of sacral function may be particularly important (for instance, in scoliosis surgery in patients with advanced neuromuscular diseases who have heavily compromised motor function but no sacral dysfunction). In other contexts, authors have argued that recordings from the anal sphincter cannot be taken as completely adequate information on the functional status of the urethral sphincter [18]. Since innervation of both sphincters originates from the same sacral segments (which also provide innervation for the detrusor), the monitoring of sphincter ani should generally mirror the function of relevant urethral structures. As previously stated, the relatively limited experience with the monitoring of sacral structures cannot as yet prove that surgeries accompanied by such monitoring are easier and safer. For the time being, there is anecdotal evidence that difficult surgical decisions that have relied on the results of intraoperative neurophysiological measurements have not resulted in unexpected neurological deficits. Although the techniques described clearly provide the surgeon with additional information about nerve location and function, the value of these techniques must be further defined and documented. As Daube [23] suggests, it will be necessary to demonstrate that these techniques indeed save operative time, save anesthesia time, or—most importantly—improve outcomes. It will be difficult to demonstrate this without doing a randomized study of patients who are undergoing surgery with and without such monitoring techniques.

REFERENCES 1. Voduˇsek, D.B., and Fowler, C.J. (1999). Clinical neurophysiology. In “Neurology of bladder, bowel, and sexual dysfunction” (C.J. Fowler, ed.), pp. 109–143. Butterworth-Heinemann, Boston, Oxford. 2. Voduˇsek, D.B. (1996). Evoked potential testing (Urodynamics II). Urol. Clin. North Amer., 23(3), 427–446. 3. Voduˇsek, D.B., Bemelmans, B., Chancellor, M., Coates, K., van Kerrebroeck, P., Opsomer, R.J., Schmidt, R., and Swash, M. (1999). Clinical neurophysiology. In “Incontinence: First International Consultation on Incontinence” (P. Abrams, S. Khoury, and A. Wein, eds.), pp. 157–195. Health Publication Ltd., Plymouth, U.K. 4. Deletis, V., Voduˇsek, D.B., Abbott, R., Epstein, F., and Turndorf, H.H. (1992). Intraoperative monitoring of dorsal sacral roots: Minimizing the risk of iatrogenic micturition disorders. Neurosurgery, 30(1), 72–75. 5. Huang, J.C., Deletis, V., Voduˇsek, D.B., and Abbott, R. (1997). Preservation of pudendal afferents in sacral rhizotomies. Neurosurgery, 41(2), 411–415. 6. Deletis, V., and Voduˇsek, D.B. (1997). Intraoperative recording of the bulbocavernosus reflex. Neurosurgery, 40(1), 88–92. 7. Rodi, Z., and Voduˇsek, D.B. (2001). Intraoperative monitoring of bulbocavernous reflex: The method and its problems. Clin. Neurophysiol., 112, 879–883.

Intraoperative Neurophysiological Monitoring of the Sacral Nervous System

217

8. Williams, P.L., Warwick, R., Dyson, M., and Bannister, L.H. (Eds.) (1989). “Gray’s anatomy.” Churchill Livingstone, Edinburgh. 9. Donker, P.M., Droes, J., and Van Ulden, B.M. (1976). Anatomy of the musculature and innervation of the bladder and the urethra. In “The scientific foundation of urology: Volume 2” (D.I. Williams, and G.D. Chisholm, eds.), pp. 32–39. Year Book Medical Publishers, Chicago. 10. Voduˇsek, D.B., and Light, J. K. (1983). The motor nerve supply of the external urethral sphincter muscles: An electrophysiologic study. Neurourol. Urodynam., 2, 193–200. 11. Henry, M.M., and Swash, M. (1992). “Coloproctology and the pelvic floor,” 2nd ed. Butterworth Heinemann, Oxford. 12. Voduˇsek, D.B., Fowler, C.J., Deletis, V., and Podnar, S. (2000). Clinical neurophysiology of pelvic floor disorders. In “Clinical neurophysiology at the beginning of the 21st century” (Z. Ambler, S. Nevˇsímalová, Z. Kadaˇnka, and P.M. Rossini, eds.), pp. 220-227 (Clinical Neurophysiology, Suppl. 53). Elsevier Science B.V., Amsterdam. 13. Lundberg, P.O., Brackett, N.L., Denys, P., Chartier-Kastler, E., Sønksen, J., and Voduˇsek, D.B. (2000). Neurological disorders: Erectile and ejaculatory dysfunction. In “Erectile dysfunction” (A. Jardin, G. Wagner, S. Khoury, F. Guiliano, H. Padma-Nathan, and R. Rosen, eds.), pp. 593645. Health Publication, Plymouth, U.K. 14. Lang, F.F., Deletis, V., Valasquez, L., and Abbott, R. (1994). Inclusion the S2 dorsal rootlets in functional posterior rhizotomy for spasticity in children with cerebral palsy. Neurosurgery, 34, 847–853. 15. Voduˇsek, D.B., and Zidar, J. (1988). Perineal motor evoked responses. Neurourol. Urodynam., 7(3), 236–237. 16. Voduˇsek, D.B., Deletis, V., and Kiprovski, K. (1993). Intraoperative bulbocavernosus reflex monitoring: Decreasing the risk of postoperative sacral dysfunction. Neurourol. Urodynam., 12, 425–427. 17. Sala, F., Krˇzan, M.J., Epstein, F.J., and Deletis, V. (1999). Specificity of neurophysiological monitoring of the lumbosacral nervous system during tethered cord release: A preliminary report. Childs Nerv. Syst., 15, 426. 18. Junemann, K.P., Schmidt, R.A., Melchior, H., and Tanagho, E.A. (1987). Neuroanatomy and clinical significance of the external urethral sphincter. Urol. Int., 42, 132–136. 19. Kothbauer, K., Schmid, U.D., Seiler, R.W., and Eisner, W. (1994). Intraoperative motor and sensory monitoring of the cauda equina. Neurosurgery, 34(4), 702–707. 20. James, H.E., Mulcahy. J.J., Walsh, J.W., and Kaplan, G.W. (1979). Use of anal sphincter electromyography during operations on the conus medullaris and sacral nerve roots. Neurosurgery, 4(6), 521–523. 21. Daube, J.R. (1991). Intraoperative monitoring of cranial motor nerves. In “Intraoperative neurophysiologic monitoring in neurosurgery” ( J. Schramm, and A.R. Møller, eds.), pp. 246–267. Springer-Verlag, Berlin. 22. Harder, S., Daube, J.R., Ebersold, M.J., and Beatty, C.W. (1987). Improved preservation of facial nerve function with use of electrical monitoring during removal of acoustic neuromas. Mayo Clin. Proc., 62, pp. 92–102. 23. Daube, J.R. (1994). Comment on Kothbauer, K., et al. (1994). Intraoperative motor and sensory monitoring of the cauda equina. Neurosurgery, 34(4), 702–707.

CHAPTER

10

Sensory Rhizotomy for the Treatment of Childhood Spasticity RICK ABBOTT Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York

1 Introduction 2 Modern Development of Selective Dorsal Rhizotomy 3 Physiologic Basis of Current Techniques 4 Outcome 5 Conclusion References

ABSTRACT Sensory rhizotomy has been used for the past century to treat spasticity. Over the last several decades there has been an evolution in the technique, and a clearer idea of expected outcomes has been gained. This chapter will describe the evolution in surgical techniques of sensory rhizotomy as well as reported results when using these newer techniques to treat childhood spasticity.

1 INTRODUCTION Selective functional, dorsal (or posterior) rhizotomy (SDR) has become an increasingly popular means of treating children with congenital spastic diplegic and quadriplegic cerebral palsy. Sensory rhizotomies were first used in the late nineteenth century [1]. In the 1960s, serious work began on refining this technique, and SDR was introduced in the mid-1970s. Several papers have since been published describing the short- and intermediate-term benefits as well as complications seen in children receiving selective functional rhizotomies [2–8]. This chapter will describe the physiology supporting SDR as well as the complications and outcome reported in children who have been treated with it. Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

219

220

Rick Abbott

2 MODERN DEVELOPMENT OF SELECTIVE DORSAL RHIZOTOMY The bulk of SDRs performed in North America are based on the surgical technique developed by Fasano [9]. In 1976, he described a modification of earlier forms of rhizotomy that was based on the studies by DeCandia of spinal reflexes in cats [10, 11]. DeCandia had shown that stimulation of the dorsal root with trains of 15 Hz or greater would cause a progressive depression in the evoked reflex motor response in normal states of spinal reflex circuit excitability. This depression did not occur, however, in the context of an upper motor neuron injury. Rather, a one-for-one compound motor action potential (CMAP) was seen when sensory roots were stimulated in animals with upper motor neuron injuries. Fasano applied these findings to children with spasticity and found that some sensory fibers, when stimulated with 30 to 50 Hz trains, would trigger one CMAP response for each stimulus, with an associated muscle contraction. Other sensory fibers, when stimulated with an identical train, did not. Instead, there was a rapid diminution of response with an abolishment of the motor response within several pulses of initiation of the stimulus train. Fasano described the first type of response pattern as abnormal in that it identified a circuit whose controlling interneuronal pool had lost the inhibitory influence of descending upper motor neurons. Thus a circuit was identified that mediated the spasticity present (an abnormal root or rootlet). The second type of pattern identified normal roots or rootlets. He described other patterns of response as being associated with abnormal roots or rootlets (i.e., tetanic contraction lasting longer than the duration of stimulation and/or spread of muscle activation outside the myotome of the spinal segment whose sensory root was being stimulated, or to another limb) (Fig. 10.1). Fasano stimulated the sensory roots of the cauda equina, labeling them as normal or abnormal in their response pattern to a 50 Hz stimulation train. Abnormal roots were subdivided into their component rootlets, which in turn were stimulated (Fig. 10.2) (see also color plate), and abnormally responding rootlets were cut. Fasano coined the term functional dorsal rhizotomy to describe the technique. This is the rhizotomy technique that most centers in North America have used as a starting point in treating spasticity. SDR was introduced to North America by Peacock, who also can be credited with improving the procedure’s safety [12]. He noted that the procedure as performed by Fasano put the urologic system at significant risk because when operating at the level of the conus there can be difficulty in identifying the functional level of the sensory nerve roots. For that reason, he moved the site of the surgery to the lumbosacral canal, which allowed for a more secure identification of the midsacral nerve roots.

Sensory Rhizotomy for the Treatment of Childhood Spasticity

221

FIGURE 10.1 Eight-channel EMG recordings showing spreading of activity in (top to bottom) right hip adductors, quadriceps, femoris, hamstrings, and gastrocnemius muscles in response to stimulation of the right L5 sensory nerve root. (Reprinted from Abbott, R. (2001), Sensory rhizotomy for the treatment of childhood spasticity, in “Pediatric neurosurgery: Surgery of the developing nervous system,” 4th ed. (P.G. McClone, ed.), pp. 1053–1062, W.B. Saunders Company, Philadelphia.)

This concern was underscored in the early 1990s as centers began to report on occurrences of postoperative urologic dysfunction in children undergoing rhizotomies. As a consequence, Deletis et al. developed a mapping technique to identify nerve roots involved in micturition reflexes, and a decrease in postoperative urologic dysfunction was reported [13, 14] (see Chapter 9). In 1987 we reported on the utility of multichannel EMG recording in monitoring for the spread of muscle activity during nerve root stimulation [15]. As already explained, there seems to be a relationship between spasticity and hyperactive stretch reflexes. Consequently, much of the evolution in sensory rhizotomy has centered on attempting to identify sensory roots or their subdivisions that are participating in these hyperactive reflexes. In the 1950s, Magladery described the effect of a conditioning compound muscle action potential (CMAP) triggered by an afferent sensory stimulus on the amplitude of subsequent evoked CMAPs [16]. The amplitude of these CMAPs (so-called H reflexes) can be expressed as a ratio, and a H reflex recovery curve (H2/H1)

222

Rick Abbott

FIGURE 10.2 Rootlet of L5 sensory root being held by stimulator probes away from rest of sensory root (see also color plate).

can be generated. In normal individuals, an attenuation in the response to the second (or later) sensory stimuli is expected when the stimulation train is above 10 Hz due to recurrent inhibition by the interneuronal pool on the α-motoneuron. Mayer and Mosser [17] described a 82.3% reduction in responses following CMAPs with a 10-Hz stimulation train in children aged 3–7 years. Nishida et al. [18] described a surgical technique using the H2/H1 reflex recovery curve that was established using a subthreshold sensory stimulation train. They used a somewhat arbitrary figure of 50% as the amount of expected diminution in amplitude of CMAPs following the initial CMAP. Abnormal roots were those found to not have the expected 50% or greater diminution. Logigian et al. [19] described a technique of stimulating the sensory roots first near the cord and then at a distance. If the latency of the CMAP increases as the root is stimulated at greater distances from the cord, the response is felt to be reflexive [19]. These roots are then separated into their component rootlets, and these are stimulated. The amplitude of the response is then graded, and those with the greatest response are cut (typically 40–70% of the tested rootlets at a given level). There has been a tendency for centers performing SDRs to modify and introduce subtle variations to the technique. Almost uniformly, surgeons today report fewer rootlets being cut. Figures between 20 and 60% are commonly

Sensory Rhizotomy for the Treatment of Childhood Spasticity

223

cited, as opposed to 80% cited in earlier papers. A paper released in the mid1990s by Steinbok et al. [20] described the evolution of their group’s SDR technique and the rationale behind it. As their experience grew, they modified the descriptor of an abnormal CMAP response to a rootlet being stimulated. Currently, an abnormal response consists of activation of muscles in the contralateral leg or, more importantly, in the upper extremities. They made this modification after noting the pattern of muscle contraction in four nonspastic children undergoing surgery for a tethered spinal cord. The exposed sensory roots in these children were stimulated when the same protocol as their rhizotomy protocol was used. There was a spread in muscle activation to involve muscles outside of the nerve root’s myotome, but it was unusual to see contractions in the contralateral limb, and no muscle contraction was seen in the upper extremity. After using this modification, Steinbok et al. observed no change in functional outcome, although they have experienced a greater amount of postoperative intoeing, which they attribute to an imbalance in the surgery’s effect on the tone of the medial and lateral hamstrings. Phillips and Parks [21] have described a grading system of “abnormal” responses. Grade 0 is Fasano’s normal response (unsustained response to a stimulus train). Grade 1 represents a sustained response in the root’s myotome, and grade 2 represents a spread of activity to muscles innervated by roots adjacent to the one being stimulated. Grade 3 represents a spread of activity to muscles innervated by roots distant from the one being stimulated, and grade 4 represents a spread to muscles in the contralateral limb [21]. This group advocates lesioning rootlets with grade 3 and grade 4 responses, and most centers performing conventional rhizotomies do the same today.

3 PHYSIOLOGIC BASIS OF CURRENT TECHNIQUES SDR is a deafferentation procedure that is felt to alter the modulating milieu of the interneuronal pool responsible for controlling the excitability of the αmotoneurons. One of this pool’s primary functions is to modulate the pattern and reactivity of the spinal cord’s reflex circuitry [22–24]. The pool receives inhibitory innervation from Ia afferents, flexor reflex afferents, and high-threshold afferents (secondary muscle afferents, joint receptor afferents, and cutaneous receptor afferents) [23–26]. Descending fiber tracts from the brain as well as segmental spinal afferents, spinal propriospinal fibers, and Renshaw fibers provide an excitatory influence on the pool [23, 27–32; Fig. 10.3A]. When either the spinal cord or the brain is injured, the excitatory and inhibitory influences on the interneuronal pool become unbalanced, and spasticity can result if inhibitory influences outweigh excitatory ones, resulting in a net amplification

224

Rick Abbott DESCENDING CORTICOSPINAL TRACTS

INTERNEURONAL POOL

A

a-MOTONEURON

SENSORY AFFERENTS

DESCENDING CORTICOSPINAL TRACTS

INTERNEURONAL POOL

B

SENSORY

a-MOTONEURON

AFFERENTS

FIGURE 10.3 (A) Normal circuit modulating reactivity of reflex circuit. (B) Amplification of circuit as a result of loss of upper motor neurons.

of the reflexive output of α-motoneurons (Fig. 10.3B). The cutting of Ia fibers to offset the lack of descending fiber input on the α-motoneurons brings the influences on the interneuronal pool back into balance. The SDR seeks to target Ia fibers that are feeding into a segment of the interneuronal pool that is unbalanced because of a decrease in descending, excitatory influences. Unfortunately, as eloquent as this rationale seems, operative observations have not consistently supported it. We reported in 1990 that the evoked pattern of muscle contraction in response to repetitive electrical stimulation of sensory roots varies from stimulation to stimulation [33]. Weiss and Schiff [34] reported on similar findings in their paper in 1993. If the theory is correct, repeated stimulation of a sensory root after an appropriate interval for recovery should evoke the same abnormal response, but this was by no means the case. Although some of the variation seen was undoubtedly due to variation in stimulator output, subsequent reports support other conclusions. Cohen and Webster [35] reported on 22 patients undergoing SDR, stating that they were unable to elicit a “normal” response to sensory root stimulation in any of the patients.

Sensory Rhizotomy for the Treatment of Childhood Spasticity

225

In the same paper they also described the results of stimulating the sensory roots in a nonspastic patient. They were unable to record a so-called normal decremental motor response, as would be expected when stimulating a normal sensory root. Steinbok et al. [36] reported on 60 consecutive spastic patients who had 680 roots tested; 99.4% failed to demonstrate a diminution in motor response to a 50 Hz stimulus train applied to the sensory roots. Additionally, they used the same stimulation parameters on 4 nonspastic children, stimulating a total of 11 roots. They found an abnormal pattern of response in 6 out of 11, and the remaining 5 roots showed the expected diminution of response to the stimulus train. Finally, Logigian et al. [19] showed that the elicited pattern of responses in some patients was probably due to direct activation of surrounding motor neurons and that random cutting of rootlets was as efficacious as directed lesioning (although they used different criteria to determine what was cut and what was preserved). Clearly, more understanding of the normal and injured system is required before we can securely postulate what exactly we are accomplishing when we use evoked motor activity in response to sensory root or rootlet stimulation to label what should be cut and what should be preserved.

4 OUTCOME In the last decade, several papers have discussed the outcome in children with spastic cerebral palsy treated with SDR. One striking finding in all the papers is that SDR is effective in decreasing spasticity and that this reduction appears to be permanent. The experience in South Africa has been reviewed by Peter and Arens [6]. They examined 104 children 2 to 12 years after they had undergone an SDR and found that 95% experienced a long-term, persistent reduction in tone. Cahan et al. [37] used an instrumented gait analysis to examine 14 patients and found that the EMG signature of spasticity seen in preoperative testing had disappeared in the postoperative testing done 6 to 14 months after their surgery. Park et al. [4] found that SDR “always reduced spasticity” in patients, and Cohen and Webster [35] described “an immediate and significant reduction in muscle tone in the lower extremities of every patient” after they underwent an SDR. Steinbok et al. [8] used a myotome to examine 50 patients before and after undergoing an SDR and found that there was a significant decrease in tone in the lower extremities and that this decrease persisted for the first year after their surgeries. Peacock and Staudt [5] used a modified Ashworth scale to assess tone in 25 patients who had received rhizotomies for their spasticity; they found that every child demonstrated a normalization of tone and loss of clonus after surgery. We looked at 49 of our patients 6 months and 1 year after they had undergone an SDR and found that they had experienced a

226

Rick Abbott

statistically significant drop in hypertonia by 6 months after their surgery and that this drop persisted at 1 year [2]. We have since been able to examine these children at their 5 year surgical anniversary, and there have been no statistically significant changes in the tone of their lower-extremity musculature. Several children did experience transient increases in tone because of a noxious stimulus such as a dislocated hip or a viral illness. After the pain abated, however, the tone returned to its postrhizotomy, premorbid state without any further treatment. A common way to measure the impact of treatment on children with cerebral palsy is to record the available range of passive motion at a given joint (goniometric measurement), and most papers publishing outcome information in children treated with SDR have followed this convention. Peacock and Staudt [38] found that patients’ available ranges improved in the hamstrings (but this did not reach clinical significance as defined by Stuberg [39]) and in the plantar flexors. Children treated at Bristish Columbia Children’s Hospital in Vancouver experienced significant improvement in goniometric measurement of the passive range of motion of the hip adductors (hip abduction), hamstrings (knee extension), and ankle plantar flexors (ankle dorsiflexion), but the hamstring’s available range tailed off between 6 and 12 months [8]. Nishida’s patients had significant improvement 2 years after their surgeries in the available range of the hip adductors, knee flexors, hamstrings, and plantar flexors [18]. Our patients showed a statistically significant decrease in range limitation in the hip adductors and hamstrings at 1 year post rhizotomy, and this has persisted at 5 year follow-up [2]. We also found that the available range for movement of the hamstring muscle is vulnerable to deterioration over time in both the diplegic and quadriplegic groups. Although this finding may not be clinically significant in a child who is wheelchair bound, it can have a great impact on ambulation in a higher-functioning child, and it can be particularly problematic in children who are not being watched by either a therapist or a trained parent, emphasizing the importance of having an ongoing program of stretching the leg’s musculature. Although documentation of the muscle tone and passive movement is helpful in quantifying the procedure’s impact, it is the procedure’s effect on function that truly measures the utility of the procedure in the management of childhood spasticity. Early on, most facilities created in-house measurement tools to quantify change in a child’s ability to assume and maintain functionally important positions used in the children’s therapy sessions. Peacock and Staudt [5], using such a system, found that five out of nine children required less support to sit, six out of nine improved in their ability to move from a kneeling to a half-kneeling position, and four out of nine improved in their ability to rise from a seated position to standing. Twenty-nine of their patients treated in South Africa were examined by Berman 4–14 months after SDR [40].

Sensory Rhizotomy for the Treatment of Childhood Spasticity

227

She found that 27 could maintain a long sitting position better after their surgery, while 19 improved in side sitting, 24 in half-kneeling, and 15 in standing. We used a similar measure after dividing our patients into subgroups of children who preoperatively could walk (walkers), quadriped-crawl (crawlers), or at best could only drag themselves about the room (nonlocomotors) [2]. Of the 41 patients followed, 11 were walkers, 11 were crawlers, and 19 were nonlocomotors. Of the 11 walkers, 2 improved in long sitting, 1 improved in side sitting while 3 declined, 5 improved in half-kneeling, and 2 improved in standing. With regards to the crawlers, 7 improved in long sitting while 1 declined, 9 improved in side sitting, 10 improved in half-kneeling, and 9 improved in standing. Of the nonlocomotors, 4 improved in long sitting with 2 declining, 5 in side sitting with 3 declining, 8 in half-kneeling with 3 declining, and 4 in standing with 2 declining. Another way to measure the functional impact of SDR is to analyze the children’s gait pattern before and after surgery using computerized gait analysis systems. Several centers have done this and have found that the major improvement experienced by these children is an increase in the stride length due to an increased ability to extend the knee and/or hip. Adams et al. [41] found this improvement to be double that which would be expected with normal maturation over a comparable length of time. Boscarino et al. [42], in addition to documenting an increase in stride length, found that there was an increase in the amount of pelvic tilt, presumably due to an asymmetrical effect of the rhizotomy, with the hip extensors experiencing a greater reduction in tone than the hip flexors. Both of these groups also found that dorsiflexion at the ankles during the swing phase of gait improved after surgery. After documenting a significant increase in valgus deformity during the terminal stance phase of the gait cycle, Adams et al. also recommended an ankle-foot orthotic to prevent hind and forefoot eversion. He also recommended the inclusion of a dorsiflexion stop to deal with a weakness in the plantar flexors and to control the forward momentum of the tibia during the stance phase. By the mid-1990s, several validated methods had been made available to measure children’s functional capability as well as the degree of assistance they require in the management of daily activities. The Pediatric Evaluation Disability Index (PEDI) and the Functional Inventory Measure for Children (WeeFIM) are two such tools that can be used to measure a child’s independence or need for assistance in activities of daily living. Dudgeon et al. [43] followed 16 spastic diplegics for a year after their SDRs; all 16 showed improvement in their mobility scores, and 13 out of 16 improved in their self-care scores. Dudgeon also serially examined 5 spastic quadriplegics and found that the changes seen in this group were not as dramatic as those seen with the spastic diplegics. Bloom and Nazar [44] looked at 8 spastic diplegics and 8 spastic quadriplegics, also using the PEDI, and found a significant improvement in self-care scores for

228

Rick Abbott

both groups. The WeeFIM was used by Nishida et al. [18] to study a group of spastic quadriplegics who had undergone an SDR. They found that the children improved most in bowel and bladder function. They also studied a group of spastic diplegics undergoing the surgery and found that the greatest improvement was in the mobility index. Surgeons have also reported improvement in upper limb function for children after rhizotomies. Steinbok et al. [8] found that 26 out of 39 patients experienced improvement in upper limb function, and Kinghorn [45] found that for 7 children with increased and disabling tone in their upper extremities, 6 had an improvement in tone to such a degree as to allow for an improvement in function.

5 CONCLUSION It seems clear at this point that selective dorsal rhizotomies reliably reduce spasticity and that this reduction is permanent. There is increasing evidence that this reduction also translates into improvement in function. What remains unclear are the physiologic explanation for this treatment’s effect and the best way to perform the treatment. Although there can be little doubt that selective posterior rhizotomy is a valid treatment for spastic cerebral palsy, I would anticipate a further evolution in the technique as our understanding of the treatment’s effect on the nervous system’s physiology grows.

REFERENCES 1. Foerster, O. (1913). On the indications and results of the excision of posterior spinal nerve roots in men. Surg. Gynecol. Obster., 26(5), 463–475. 2. Abbott, R., Johann-Murphy, M., Shiminski-Maher, T., Quartermain, D., Forem, S.L., Gold, J.T., and Epstein, F.J. (1993). Selective dorsal rhizotomy: Outcome and complications in treating spastic cerebral palsy. Neurosurgery, 33, 851–857. 3. Fasano, V.A., Broggi, G., Zeme, S., Lo Russo, G., and Sguazzi, A. (1980). Long-term results of posterior functional rhizotomy. Acta Neurochir. Suppl. (Wien), 30, 435–439. 4. Park, T.S., Gaffney, P.E., Kaufman, B.A., and Molleston, M.C. (1993). Selective lumbosacral dorsal rhizotomy immediately caudal to the conus medullaris for cerebral palsy spasticity. Neurosurgery, 33, 929–934. 5. Peacock, W.J., and Staudt, L.A. (1991). Functional outcomes following selective posterior rhizotomy in children with cerebral palsy. J. Neurosurg., 74, 380–385. 6. Peter, J.C., and Arens, L.J. (1993). Selective posterior lumbosacral rhizotomy for the management of cerebral palsy spasticity: A 10-year experience. S. Afr. Med. J., 83, 745–747. 7. Staudt, L.A., and Peacock, W.J. (1995). Dorsal rhizotomy for spasticity. West. J. Med., 162, 260. 8. Steinbok, P., Reiner, A., Beauchamp, R.D., Cochrane, D.D., and Keyes, R. (1992). Selective functional posterior rhizotomy for treatment of spastic cerebral palsy in children: Review of 50 consecutive cases. Pediatr. Neurosurg., 18, 34–42.

Sensory Rhizotomy for the Treatment of Childhood Spasticity

229

9. Fasano, V.A., Broggi, G., and Zeme, S. (1988). Intraoperative electrical stimulation for functional posterior rhizotomy. Scand. J. Rehabil. Med. Suppl., 17, 149–154. 10. DeCandia, M., Provini, L., and Taborikova, H. (1967). Mechananisms of the reflex discharge depression in the spinal motoneurone during repetitive orthodromic stimulation. Brain Res., 4, 284–291. 11. Fasano, V.A., Barolat-Romano, G., Ivaldi, A., and Squazzi, A. (1976). La radicotomie posterieure fonctionnelle dans le traitement de la spasticite cerebrale. Neurochirurgie, 22, 23–34. 12. Peacock, W.J., Arens, L.J., and Berman, B. (1987). Cerebral palsy spasticity: Selective posterior rhizotomy. Pediatr. Neurosci., 13, 61–66. 13. Deletis, V., Voduˇsek, D., Abbott, R., Epstein, F., and Turndorf, H. (1992). Intraoperative monitoring of the dorsal sacral roots: Minimizing the risks of iatrogenic micturition disorders. Neurosurgery, 30, 72–75. 14. Huang, J.C., Deletis, V., Voduˇsek, D.B., and Abbott, R. (1997). Preservation of pudendal afferents in sacral rhizotomies. Neurosurgery, 41(2), 411–415. 15. Abbott, R., Wisoff, J., Spielholtz, N., and Epstein, F. (1987). Selective posterior rhizotomy for the treatment of spasticity: Relationship of intraoperative EMG patterns to the post-operative course, 15th Annual Meeting of the International Society for Pediatric Neurosurgery, New York. 16. Magladery, J., Teasdall, R., Park, M., and Languth, H. (1952). Electrophysiological studies of reflex activitiy in patients with lesions of the nervous system: I. A comparison of spinal motoneurone excitability following afferent nerve volleys in normal persons and patients with upper motor neruone lesions. Bull. Johns Hopkins Hosp., 91, 219–244. 17. Mayer, R., and Mosser, R. (1969). Excitability of motoneurones in infants. Neurology, 19, 932–945. 18. Nishida, T., Thatcher, S., and Marty, G. (1995). Selective posterior rhizotomy for children with cerebral palsy: A 7-year experience. Childs Nerv. Syst., 11, 374–380. 19. Logigian, E.L., Wolinsky, J.S., Soriano, S.G., Madsen, J.R., and Scott, R.M. (1994). H reflex studies in cerebral palsy patients undergoing partial dorsal rhizotomy [see comments]. Muscle Nerve, 17, 539–549. 20. Steinbok, P., Gustavsson, B., Kestle, J.R., Reiner, A., and Cochrane, D.D. (1995). Relationship of intraoperative electrophysiological criteria to outcome after selective functional posterior rhizotomy. J. Neurosurg., 83, 18–26. 21. Phillips, L., and Parks, T. (1989). Electrophysiological studies of selective posterior rhizotomy patients. In “Neurosurgery: State of the art reviews: Management of spasticity in cerebral palsy and spinal cord injury” (T.S. Park, L. Phillips, W.J. Peacock, eds.), pp. 459–469. Hanley & Belfus, Philadelphia. 22. Dimitrijevi´c, M. (1985). Spasticity. In “Scientific basis of clinical neurology” (M. Swash, and C. Kennard, eds.), pp. 108–115. Churchill Livingstone, Edinburgh. 23. Lundberg, A. (1969). Convergence of excitatory and inhibitory action on interneurons in the spinal cord. In “The interneuron” (M. Brazier, ed.), pp. 231–265. University of California Press, Los Angeles. 24. Raymer, W. (1984). Spinal mechanisms for control of muscle length and tension. In “Handbook of the spinal cord” (R. Davidoff, ed.)., vols. 2 and 3, pp. 609–646. Basel: Marcel Dekker, New York. 25. Balasubrammaniam, V., Kanaka, T.S., and Ramanujam, P.B. (1974). Stereotaxic surgery for cerebral palsy. J. Neurosurg., 40, 577–582. 26. Scheibel, A. (1984). Organization of the spinal cord. In “Handbook of the spinal cord” (R. Davidoff, ed.), vol. 2, pp. 47–77. Basel: Marcel Dekker, New York. 27. Jankowska, E., Lundberg, A., and Stuart, D. (1973). Propriospinal control of last order interneurons of spinal reflex pathways in the cat. Brain Res., 53, 227–231.

230

Rick Abbott

28. Jankowska, E., Lundberg, A., and Roberts, W. (1974). A long propriospinal system with direct effects on motoneurones and on interneurones in the cat lumbosacral cord. Exp. Brain Res., 21, 169–194. 29. Lloyd, D. (1942). Mediation of descending long spinal reflex activity. J. Neurophysiol., 5, 435–458. 30. Lundberg, A. (1964). Supraspinal control of transmission in reflex paths to motoneurones and primary afferents. In “Physiology of spinal neurones” ( J.C. Eccles, ed.), pp. 197–221. Elsevier, Amsterdam. 31. Nathan, P., and Smith, M. (1959). Fasciculi proprii of the spinal cord in man: Review of present knowledge. Brain, 82, 610–668. 32. Stelzner, D. (1982). The role of the descending systems in maintaining intrinsic spinal function: A developmental approach. In “Brainstem control of spinal mechanisms” (B. Sjolund, ed.), pp. 297–321. Elsevier, Amsterdam. 33. Abbott, R., Deletis, V., Spielholz, N.I., Wisoff, J.H., and Epstein, F.E. (1990). Selective posterior rhizotomy, pitfalls in monitoring. In “Concepts of pediatric neurosurgery” (A. Marlin, ed.), pp. 187–195. Basel: Karger. 34. Weiss, I., and Schiff, S. (1993). Reflex variability in selective dorsal rhizotomy. J. Neurosurg., 79, 346–353. 35. Cohen, A., and Webster, H. (1991). How selective is selective posterior rhizotomy. Surg. Neurol., 35, 267–272. 36. Steinbok, P., Langill, L., Cochrane, D.D., and Keyes, R. (1992). Observations on electrical stimulation of lumbosacral nerve roots in children with and without lower limb spasticity. Childs Nerv. Syst., 8, 376–382. 37. Cahan, L.D., Adams, J.M., Perry, J., and Beeler, L.M. (1990). Instrumented gait analysis after selective dorsal rhizotomy. Dev. Med. Child Neurol., 32, 1037–1043. 38. Peacock, W.J., and Staudt, L.A. (1991). Selective posterior rhizotomy: Evolution of theory and practice. Pediatr. Neurosurg., 17, 128–134. 39. Stuberg, W., Fuchs, R., and Miedaner, J. (1988). Reliability of gonometric measurements of children with cerebral palsy. Dev. Med. Child Neurol., 30, 657–666. 40. Berman, B., Vaughan, C., and Peacock, W. (1990). The effect of rhizotomy on movement in patients with cerebral palsy. Am. J. Occup. Ther., 44, 511–516. 41. Adams, J., Cahan, L., Perry, J., and Beeler, L. (1995). Foot contact pattern following selective dorsal rhizotomy. Pediatr. Neurosurg., 23, 76–81. 42. Boscarino, L., Ounpuu, S., Davis, R., Gage, J., and DeLuca, P. (1993). Effects of selective dorsal rhizotomy on gait in children with cerebral palsy. J. Pediatr. Orthop., 13, 174–179. 43. Dudgeon, B., Libby, A., McLaughlin, J., Hays, R., Bjornson, K., and Roberts, T. (1994). Prospective measurement of functional changes after selective dorsal rhizotomy. Arch. Phys. Med. Rehabil., 75, 46–53. 44. Bloom, K.K., and Nazar, G.B. (1994). Functional assessment following selective posterior rhizotomy in spastic cerebral palsy. Childs Nerv. Syst., 10, 84–86. 45. Kinghorn, J. (1992). Upper extremity functional changes following selective posterior rhizotomy in children with cerebral palsy. Am. J. Occup. Ther., 46, 502–507.

FIGURE 10.2 sory root.

Rootlet of L5 sensory root being held by stimulator probes away from rest of sen-

FIGURE 16.1 A three-dimensional artist's rendition of the structures involved in surgery for movement disorders. The light greenish blue structure on the left is the globus pallidus (GPi and GPe). The large grey structure on the right is the thalamus, and the small dark green structure is the subthalamic nuclei (STN). The medial edge of the STN is only 6.0 m m from the midline of the brain. With the trajectories that our group uses in the operating room, we encounter around 10.0 m m of GPi, 11.0 m m of VIM, and 5.0 m m of STN. Modified from [117].

CHAPTER

11

Neurophysiological Monitoring During Pedicle Screw Placement RICHARD J. TOLEIKIS Department of Anesthesiology, Rush-Presbyterian-St. Luke’s Medical Center, Rush University, Chicago, Illinois

1 Introduction 2 Techniques for Assessing Nerve Root Function and Pedicle Screw Placement 2.1 Sensory Pathway Assessment Techniques 2.2 Motor Pathway Assessment Techniques 2.3 Anesthetic Management for Motor Techniques 2.4 Factors That Can Contribute to False-Negative Findings 2.5 Impedance Testing 3 Personal Experience 4 Conclusions References

ABSTRACT The use of pedicle screws for spinal stabilization has become commonplace during various spinal surgical procedures. However, the placement of these screws is largely done blindly, and even in the hands of experienced surgeons, the incidence of misplaced pedicle screws resulting in neurological impairment has been reported to be quite high, despite the use of surgical inspection and imaging techniques. Although new imaging techniques have been developed that may help to reduce the incidence of misplaced hardware, the equipment needed to implement their usage is generally costly, and the techniques themselves are still not developed to the point where they are free from error. As a result, surgeons and clinical neurophysiologists have used various electrophysiological monitoring techniques for assessing nerve root function and pedicle screw placements. To be widely and effectively used, these techniques must meet certain criteria; the strengths and limitations of each of the techniques will be discussed in terms of these criteria. On the basis of the author’s experience and Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

231

232

Richard J. Toleikis outcome data, the combined use of spontaneous and triggered myogenic activity for intraoperative monitoring purposes satisfies all the criteria that a monitoring technique should meet. This technique is cost-effective and improves surgical outcomes.

1 INTRODUCTION Through the ages, various treatments for spinal deformity have evolved. In 1962, Harrington [1] ushered in the revolutionary use of metallic, internal fixation devices for spinal deformity when he reported on the use of a distraction rod construct for the treatment of scoliosis. In 1982, Luque [2] demonstrated how spinal deformity could be corrected by the use of segmental fixation and the application of transverse forces. In the thoracolumbar region of the spine, spinal instrumentation constructs consisting of hooks and rods have now become the standard of care for the surgical management of degenerative spinal disease and traumatic insults. In the lumbosacral region, it has become very popular to use pedicle screws rather than hooks to hold the rods in place for the purpose of segmental transpedicular fixation. Although pedicle screws can be placed in the thoracic and lumbosacral spine, they are generally placed in the most caudal segments of the spine: L2–S1. The cross-sectional area of the pedicles is smaller in the thoracic segments of the spine. Because of this, and because it is common for spinal instabilities to occur in the lumbosacral region, it is unusual for pedicle screws to be placed in the thoracic region. From a monitoring perspective, it is important to remember that the spinal cord ends in the conus medullaris at about the T12–L1 level of the spine. Therefore, the placement of pedicle screws below these levels potentially places nerve root rather than spinal cord function at risk. Proper placement of pedicle screws requires that a surgeon be extremely knowledgeable about the anatomical characteristics of the thoracic, lumbar, and sacral vertebrae. Despite the use of anatomical landmarks and fluoroscopy, the placement of pedicle screws is largely done blindly. Ideally, they should be placed so that they pass through the pedicle with about 1 mm to spare on both the medial and lateral walls and without any breach of the pedicle walls. In addition, they should be placed well into the vertebral bodies without any breach of the vertebral body walls. Nerve roots tend to position themselves near the medial and inferior aspects of the pedicles as they exit the spinal canal through the spinal foramen. Therefore, screws that are placed so that they protrude or are exposed from the medial or inferior pedicle walls can cause nerve root irritation or injury. In preparation for the placement of pedicle screws, markers are generally placed into the pedicles in order to visualize, via radiographs, the trajectories that the pedicle screws will take. These trajectories may place nerve roots at risk

Neurophysiological Monitoring During Pedicle Screw Placement

233

for injury, because both lateral and anterior–posterior radiographs are subject to reading errors. As a result, undesirable medial placements of both markers and screws may not be identified. [3, 4]. Such readings are followed by removal of the markers and tapping of the holes made by the markers. These holes can then be palpated to detect holes in the pedicle walls. The pedicle screws are then placed. This placement can result in fractures of the pedicle, breakthroughs of the pedicle walls, and/or extrusion of pedicle fragments. Even in the hands of experienced surgeons, the current literature reports pedicle cortical perforation rates that have ranged from 5.4 to 40% [5–11]. Such events may go undetected unless the pedicle walls are visualized. However, most surgeons are reluctant to routinely visualize screw placements unless this action is warranted, since doing so would require that the surgeon do multiple laminotomies. This is time-consuming, and these additional procedures by themselves could also affect postoperative outcome. However, despite the use of surgical inspection and imaging techniques, misplaced screws have still been frequently associated with neurological functional impairment. The incidence has ranged from 1% to more than 11% [12–15]. New imaging technologies aimed at reducing these incidences continue to evolve. Thus far, they have been somewhat cumbersome, costly, and timeconsuming, and the end result is that they are still not free of error [16]. Therefore, existing electrophysiological techniques have been used, as well as others that have evolved for monitoring neurological function during pedicle screw placement and for assessing these placements. They include mixed nerve somatosensory evoked potentials (SEPs), dermatomal somatosensory evoked potentials (DSEPs), and techniques that rely upon both spontaneous and triggered myogenic activity. In addition, the measurement of electrical tissue impedance has been suggested as another means for assessing placements. Other techniques, which include transcranial and spinal stimulation, can also be used to test nerve root function during pedicle screw placement; however, although feasible, these techniques are rarely used for this purpose.

2 TECHNIQUES FOR ASSESSING NERVE ROOT FUNCTION AND PEDICLE SCREW PLACEMENT All nerve roots consist of both sensory and motor fibers. The monitoring techniques that are used to assess nerve root function during surgery produce either sensory or motor responses. Sensory responses that are mediated by a single specific nerve root can be elicited by stimulating a specific body surface area known as a dermatome. Motor activity that is mediated by a single specific motor nerve root can elicit myogenic responses from a group of muscles known as a myotome. Therefore, the responses that are acquired to assess the sensory

234

Richard J. Toleikis

and motor function of a single nerve root are known as dermatomal and myotomal responses, respectively. There are certain criteria that monitoring techniques should meet if they are to be widely and effectively used to assess pedicle screw placements and preserve nerve root function. First of all, implementation of the techniques should be practical; in other words, they should not require special equipment or expertise. Otherwise, these factors will be a deterrent to their use. For economic and practical reasons, the techniques should utilize standard equipment that may already be used for monitoring purposes, they should be easy to perform, and the anesthetic requirements should not be unusual. Second, the techniques must be effective. They should provide an instantaneous indication of nerve root irritation in order to prevent injury or further damage of a nerve root that is already irritated. They should also be able to detect the presence of a misplaced screw that is not causing nerve root irritation but that may have the potential to do so. Finally, they should produce accurate results that make a difference in patient outcomes and that are cost-effective. A discussion of each of the following techniques will address these requirements.

2.1 SENSORY PATHWAY ASSESSMENT TECHNIQUES 2.1.1 Somatosensory Evoked Potentials (SEPs) Since the late 1970s, mixed nerve somatosensory evoked potentials (SEPs) have been used to monitor spinal cord function during spinal instrumentation procedures in order to minimize the probability of postoperative neurological deficits [17–20]. SEPs are elicited by stimulating a peripheral nerve at a distal site: typically the median or ulnar nerves at the wrist for acquiring SEPs from the upper extremities, and the posterior tibial nerve at the ankle or the peroneal nerve at the fibular head for acquiring lower-extremity SEPs. The ascending sensory volley that contributes to the SEP enters the spinal cord through dorsal nerve roots at several segmental levels and may ascend the spinal cord via multiple pathways. The general consensus is that the dorsal or posterior column spinal pathways [21–24] primarily mediate the SEPs. Other pathways, such as the dorsal spinocerebellar tracts [25, 26] and the anterolateral columns [27, 28], may contribute to the early SEP responses that are used for monitoring purposes. Despite the fact that SEPs are primarily mediated by the dorsal columns and therefore are a means of directly assessing only sensory and not motor pathway function, they have proven to be extremely useful as a clinical tool for detecting changes in spinal cord motor function, particularly when these changes result from mechanical insults. It is important to realize that mixed nerves receive sensory and motor fibers from multiple nerve roots. Therefore, when mixed nerves are stimulated, the

Neurophysiological Monitoring During Pedicle Screw Placement

235

FIGURE 11.1 False-negative posterior tibial nerve SEP responses from a patient who suffered intraoperative compression of the right L5 nerve root. The intraoperative responses were repeatable and demonstrated no significant changes throughout the surgical procedure. Reprinted from [34].

electrophysiological responses that result from the stimulation (known as SEPs) are mediated by more than one nerve root prior to being mediated by the spinal cord. It is not unreasonable to expect that SEP changes should occur when the function of one of the contributing nerve roots becomes abnormal. However, the usefulness of SEPs for assessing spinal root function in patients diagnosed as having cervical spondylosis [29–31] and lumbar root lesions [32, 33] has been limited. In addition, when used as a neurophysiological monitoring tool during pedicle screw placements, they appear to be totally insensitive to changes in nerve root function [34] (Figs. 11.1 and 11.2), largely because several nerve roots typically contribute to the composition of a peripheral nerve. For example, the posterior tibial nerve receives contributions from the L4, L5, S1, S2, and S3 nerve roots. As a result, a monoradicular functional abnormality may not be apparent when mixed-nerve evoked potentials are used to evaluate a patient because abnormal nerve root function may be masked by the normal activity mediated via unaffected spinal nerve roots [35–37]. Therefore, mixednerve SEPs may be insensitive to irritation or injury to a single nerve root. For this reason, they should not be used to monitor spinal nerve root function, since other techniques are much better suited for this purpose. On the other hand, since SEPs were developed and continue to be used as a technique for assessing spinal cord function, they might be useful during pedicle screw placements if spinal cord rather than nerve root function is placed at risk. However, since pedicle screws are usually placed at levels caudal to the conus medullaris (i.e., L2–S1), screw placement would not place spinal cord function at risk, and, indeed, there have been no published reports of the loss of spinal cord function during such procedures. Therefore, unless there is reason to believe that spinal cord function is at risk, there appears to be no basis for the use of SEPs as a monitoring tool during pedicle screw procedures.

236

Richard J. Toleikis

FIGURE 11.2 CT scan of the patient whose SEP responses are shown in Fig. 11.1. The pedicle screw has entered the spinal canal and is in a position where the right-sided L5 spinal nerve root (left in the photo) was found on surgical inspection to be significantly compromised. Reprinted from [34].

2.1.2 Dermatomal Somatosensory Evoked Potentials (DSEPs) A dermatome is defined as a body surface area that receives its cutaneous sensory innervation from a single spinal nerve root. It has been demonstrated that DSEPs arise from stimulation of receptors in the skin rather than from subcutaneous digital nerves [38]. As a result, they are normally elicited using some form of surface electrodes and are probably mediated via the same pathways as mixed-nerve SEPs. The first reported use of dermatomal or segmental SEPs (as they were initially named) was by European investigators [39]. Since these first studies, they have been used to assess children with myelomeningocele [40], to evaluate patients with spinal cord injuries [41], as a monitoring tool during spinal surgery to determine the adequacy of spinal nerve root decompression [42–44], and to detect nerve root functional impairment during pedicle screw placement [34, 45]. DSEPs are acquired using the same stimulation and recording techniques and equipment that are used to acquire mixed-nerve SEPs. Unlike SEPs, however, the only DSEP responses that are clinically useful are recorded from

Neurophysiological Monitoring During Pedicle Screw Placement

237

the scalp, because it is normally very difficult to record either peripheral or subcortical DSEP responses. This may be a function of the relative number of afferent nerve fibers that mediate and contribute to DSEPs as compared to mixednerve SEP responses. As a result, like mixed-nerve cortical SEP responses, DSEP responses can be very susceptible to the anesthetic drugs that are used during surgery [45]. In addition, it has been shown that the latency and amplitude of dermatomal responses are a function of the stimulation intensity [46]. Therefore, for monitoring purposes, stimulation intensity should remain constant throughout a surgical procedure so that one does not attribute response changes to surgical events. DSEP responses, at least ideally, are considered to be nerve root–specific. However, this may not always be the case. Dermatomes tend to overlap, and their spatial distributions vary from person to person. Besides their susceptibility to typical anesthetic drugs, this is another minor shortcoming of this technique. One of the two major limitations associated with the use of DSEPs is that, because of their small amplitude, they can only be acquired using an averaging technique; hence, their acquisition, like that of SEPs, may require a few minutes. During this period of time, functional changes can occur that may go undetected until another average is acquired. At that point in time, nerve root damage may have occurred and the associated functional changes may be irreversible. For pedicle screw placement, the second major limitation of the DSEP technique is that changes in dermatomal responses will only occur if a pedicle screw actually makes contact with a nerve root (Fig. 11.3). Misplaced screws that do not make contact with a nerve root may still represent a potential source of nerve root irritation or damage and will go undetected with the dermatomal technique.

FIGURE 11.3 Intraoperative responses obtained from a patient in which the right-sided L4 DSEP responses increased in latency and then disappeared during screw insertion. The surgeon was notified of the response changes and immediately removed the right L4 pedicle screw. The screw was redirected and the responses returned to normal. The patient experienced no postoperative deficits related to screw placement. Reprinted from [34].

238

Richard J. Toleikis

2.1.3 Anesthetic Management for Sensory Techniques Although the monitoring of SEPs has clearly been beneficial during many surgical procedures, the anesthesia used to facilitate these procedures produces effects that alter the evoked potentials. These effects are well documented [47–50]. They are most prominent on the cortically generated responses and less so on the subcortical and peripheral responses. They are generally dose related, and their effects on cortical SEPs tend to parallel their effects on EEG. Most of the commonly used anesthetic drugs produce dose-related SEP changes that include amplitude decreases and latency increases. The relative degree of change differs between anesthetic agents. The drug dosage that causes a 50% decrease of cortical SEP amplitude correlates with the lipid solubility of the agent and therefore its anesthetic potency [48, 49]. Therefore, when anesthetic techniques are being considered, the effect of each anesthetic agent on specific monitoring modalities must be considered. Probably the most commonly used anesthetics are the halogenated inhalational agents (desflurane, enflurane, halothane, isoflurane, sevoflurane). All these agents produce a dose-related increase in latency and reduction in amplitude of the cortically recorded SEP responses. Several studies have demonstrated that halogenated agents differ in their potency of effect on cortical SEPs. Isoflurane has been reported to be the most potent, and enflurane and halothane the least potent [49]. At steady state, the potency of sevoflurane and desflurane appears to be similar to that of isoflurane. The effects are less on the subcortical SEP responses recorded over cervical spine and are minimal on spinal responses recorded epidurally or on peripherally recorded responses. If it is essential to monitor cortical SEPs, the use of halogenated inhalational agents may need to be restricted or eliminated entirely. However, if the recording of subcortical responses is adequate for monitoring purposes, halogenated agents may be acceptable anesthetic choices. Nitrous oxide produces decreases in cortical SEP amplitude and increases in cortical SEP latencies when used alone or in conjunction with halogenated inhalational agents or opioid anesthetics. When compared to other inhalational anesthetic agents at equipotent anesthetic concentrations, nitrous oxide produces the most profound cortical SEP changes [50]. Like halogenated agents, the effects of nitrous oxide on subcortical and peripheral sensory responses are minimal. However, nitrous oxide has been reported to have a synergistic effect on cortical SEPs when used in conjunction with other inhalational agents. Although the use of DSEPs represents an improvement over the use of SEPs since one is able to detect single nerve root functional changes, both SEP and DSEP cortical responses are sensitive to standard anesthetic management that includes the use of halogenated gases and nitrous oxide; the DSEPs are relatively more sensitive to these anesthetic agents than are mixed-nerve SEPS

Neurophysiological Monitoring During Pedicle Screw Placement

239

because they are smaller in amplitude to begin with. This minor limitation can be minimized by using anesthetic agents that are administered intravenously (total intravenous anesthesia–TIVA), but these anesthetic agents tend to be more costly than the anesthetic gases. A number of factors determine the choice of anesthetic agents when monitoring is to be performed. These include (1) how anesthetic agents may interact with a patient’s pathophysiology, (2) surgical requirements (i.e., performance of a stagnara wake-up test, awake during a carotid endarterectomy procedure), and (3) the specific monitoring modalities to be used. In general, anesthetic agents produce an alteration in the evoked responses consistent with their clinical effects on the CNS. Several important generalizations can be made regarding the effects of anesthetic agents on SEPs. First, most tend to decrease neural conduction and synaptic transmission. As a result, they tend to decrease the amplitude and increase the latency of SEPs. Second, the effects of anesthetics on SEPs appear to be most prominent in regions where synaptic transmission is prominent. Therefore, their effects are most pronounced on cortically generated peaks and least effective on brainstem, spinal cord, and peripheral responses. Third, anesthetic effects appear to be dose related, although many agents have a disproportionate effect at low dosages in the range where major clinical anesthetic effects are occurring. Fourth, just as patients react differently to the same dose of an anesthetic drug, so also their SEPs are affected differently. Finally, during periods when neurological function is acutely at risk, it is important to maintain a steady state of anesthesia. Taking into consideration all these factors, an anesthetic regimen can usually be chosen that will permit effective monitoring.

2.2 MOTOR PATHWAY ASSESSMENT TECHNIQUES Myotomes are the motor complement to dermatomes, and myotomal distributions are also quite variable between individuals. Whereas a myotome is a group of muscles that receive their motor innervation from a specific spinal nerve root, most muscles receive efferent innervation from several spinal nerve root levels. The amount and type of innervation to specific muscle groups will vary from person to person. Myotomal activity can be spontaneously elicited by mechanical stimulation or triggered by electrical stimulation. Typically, the myotomal activity from several muscle groups is monitored at any given time, and the activity is recorded using either surface or subdermal needle electrodes placed over or into the various muscle groups. The selection of the muscle groups to monitor is made on the basis of which spinal nerve roots are at risk for irritation or injury. Muscles typically receive their innervation from several spinal levels, although

240

Richard J. Toleikis TABLE 11.1 Innervation to Various Muscle Groups Innervation levels

Muscles

Cervical

C2, C3, C4 C5, C6 C6, C7 C8, T1

Trapezius, Sternocleidomastoid Biceps, Deltoid Flexor Carpi Radialis Abductor Pollicis Brevis, Abductor Digit Minimi

Thoracic

T5, T6 T7, T8 T9, T10, T11 T12

Upper Rectus Abdominis Middle Rectus Abdominis Lower Rectus Abdominis Inferior Rectus Abdominis

Lumbosacral

L2, L3, L4 L4, L5, S1 L5, S1

Vastus Medialis Tibialis Anterior Peroneus Longus

Sacral

S1, S2 S2, S3, S4

Gastrocnemius External Anal Sphincter

one spinal level generally predominates in terms of the amount of innervation it provides to any given muscle group. Activity can be recorded from muscles innervated by the cervical, thoracic, lumbar, and sacral spinal nerve roots. In addition to paraspinal muscles, the muscles commonly used for these recordings and their innervation appear in Table 11.1. 2.2.1 Spontaneous Myogenic Activity The responses that are elicited when nerve roots are mechanically or electrically stimulated are summed responses from many muscle fibers known as compound muscle action potentials (CMAPs). They can be recorded using pairs of surface or needle electrodes that are placed over or into the belly of a muscle. Recordings should be made continuously throughout a surgical procedure. Assuming that excessive amounts of muscle relaxants have not been administered to a patient during surgery and that muscles are adequately unrelaxed, spontaneous activity will be elicited when mechanical activation results in nerve root irritation or injury. This spontaneous activity, suggestive of nerve root irritation, can be recorded when train-of-four testing of target muscles (i.e., muscle groups innervated by the nerve roots at risk) produces only one CMAP. The activity will typically be elicited from one or more muscle groups, depending on the activated nerve root, the muscle groups being monitored, and the placement of the recording electrodes on these muscle groups. The EMG activity from each electrode pair is recorded using differential amplification and is filtered using a wide bandpass filter (30 Hz–3 kHz).

Neurophysiological Monitoring During Pedicle Screw Placement

241

When spontaneous myogenic activity is recorded to detect mechanical nerve root irritation, the data acquisition system should be set to operate in the free run mode. In this mode, the sweep time is typically 1 s and any elicited activity can easily be visualized and evaluated. Typically, several channels of myogenic activity should be monitored simultaneously, depending on the number of channels available, but six or more should be monitored. When interpreting spontaneous activity, there are several factors to take into consideration. First of all, normal nerve roots and irritated or regenerating nerves in continuity react differently to mechanical forces. When mechanical forces are statically or rapidly applied to normal nerve roots, they induce no nerve root activity or trains of impulses of short duration [51]. When the same forces are applied to irritated or regenerating nerves, they induce long periods of repetitive impulses. Minimal acute compression of normal dorsal root ganglion also induces prolonged repetitive firing of nerve roots. When interpreting intraoperative motor nerve root activity, it is important to understand the pathophysiological mechanisms of nerve root injury and to understand the response of normal and pathological nerve to not only different types of mechanical force but also to electrical stimulation. Normally, the recordings of spontaneous activity will demonstrate the lack of activity. However, when preexisting nerve root irritation has been present, the recordings will often consist of low-amplitude periodic firing patterns. Mechanically elicited activity consists of either short bursts of activity that can last a fraction of a second or long trains of activity that can last up to several minutes (Fig. 11.4). The short aperiodic bursts of activity are common. Attention should be paid to these, but they are normally not cause for alarm and are rarely indicative of a neural insult. The long trains are more serious, may be indicative of neural injury, and are causes for alarm. The short bursts are associated with direct nerve trauma such as tugging and displacement, irrigation, electrocautery, metal-to-metal contact, or application of soaked pledgets. Train activity is commonly related to sustained traction and compression. The more sustained the activity, the greater the likelihood of nerve root damage. When train activity occurs, the surgeon must be notified so that corrective measures can immediately be taken. 2.2.2 Triggered Myogenic Activity As mentioned earlier, triggered myogenic activity can be elicited in several ways: through direct nerve root stimulation [12], indirect nerve root stimulation by means of stimulation of spinal instrumentation [9, 13, 15, 46, 52], direct spinal cord stimulation (myogenic motor evoked potentials) [53, 54], and transcranial motor evoked potentials elicited by either electrical or magnetic transcranial stimulation [55, 56].

242

Richard J. Toleikis

FIGURE 11.4 Spontaneous myogenic activity elicited from the left tibialis anterior muscle as a result of mechanical irritation of the left L5 nerve root. Short bursts of such activity are sometimes observed with manipulation of nerve roots that are not irritated or injured. Such activity is considered insignificant and rarely results in postoperative sequelae. Sustained activity (more than a second) is attributed to the manipulation of irritated nerve roots or to nerve root injury. Such activity is considered significant and should be avoided to minimize the chances of postoperative deficits. Reprinted from [52].

2.2.2.1 Transcranial and Spinal Stimulation Both spinal cord and transcranial stimulation are typically used to elicit myogenic responses from lower-extremity muscle groups in order to assess spinal cord motor function. They can also be used to assess individual nerve root function during pedicle screw placements. However, they typically are not used in this fashion, for several reasons. First, these techniques are more complex than other techniques that are currently available; they require special equipment, electrode placement skills, anesthetic management, and/or consent for their implementation. Second, functional status can only be determined when stimulation occurs, and not continuously. Third, although myogenic responses can

Neurophysiological Monitoring During Pedicle Screw Placement

243

generally be elicited from some designated target muscle groups such as the tibialis anterior muscles, the threshold stimulation intensities needed to elicit myogenic responses from other designated target muscles vary from muscle to muscle. These thresholds can vary during a procedure, and these changes may be unrelated to surgical causes. Therefore, their reliability in determining when a surgical event has caused a functional change is in question. Finally, the techniques can only detect when functional changes have already occurred; they are not able to detect potential causes of functional changes. As a result, investigators have turned to other techniques for monitoring nerve root function during pedicle screw placements. These techniques include direct nerve root stimulation and indirect nerve root stimulation by means of stimulation of spinal instrumentation. 2.2.2.2 Direct Nerve Root Stimulation Direct nerve root stimulation is sometimes used to determine the stimulation thresholds of nerve roots placed at risk during screw placement. Ideally, this technique should be used in conjunction with indirect nerve root stimulation when stimulation thresholds are in question, either as a result of chronic nerve root compression (particularly when a radiculopathy is present) [57, 58] or when disease processes are present, such as diabetes, that may effect nerve root function. It is generally assumed that when nerve roots are indirectly stimulated via the spinal instrumentation, the nerve roots that are being excited are healthy and function normally. Maguire et al. [9] reported that the constant current stimulation threshold for eliciting responses from normal nerve roots ranged from 0.2 to 5.7 mA, with an average stimulation intensity of 2.2 mA. Calancie et al. [13] reported that constant current stimulation thresholds for nerve roots ranged from 1.2 to 3.8 mA with an average of 2.1 mA. However, investigators have reported that chronically compressed nerve roots have elevated stimulation thresholds [57, 58], and Holland et al. [58] reported that stimulation thresholds greater than 20 mA may be necessary to elicit myogenic responses from such nerve roots. These findings indicate that if test parameters that have been developed from testing pedicle screw placements in patients with normal nerve root function are used to test screw placements involving chronically compressed nerve roots, false-negative findings can result. As indicated earlier, elevated stimulation thresholds may also occur in patients with metabolic disorders such as diabetes. However, to my knowledge, no reports have appeared in the literature to support this supposition. Our own experience with such patients is limited, but some of the diabetic patients we have tested have had only moderately elevated stimulation thresholds. Rather than having thresholds of 2 mA or less, the diabetic patients have been found

244

Richard J. Toleikis

to have thresholds of 4–5 mA—still in the normal range reported by other investigators [9]. One way to avoid false-negative findings is to directly stimulate each nerve root at risk to ensure that it is functioning normally before testing the placement of each pedicle screw. Although this may be a reasonable step if decompressions are being done, routine laminotomies to explore each nerve root are time-consuming and can be associated with undue risk. Therefore, most surgeons may prefer not to use direct stimulation of nerve roots in conjunction with indirect nerve root stimulation techniques. However, in patients exhibiting signs of nerve root malfunction as a result of either compression or disease processes, it is strongly suggested that direct nerve root testing be performed to establish stimulation thresholds when indirect nerve root stimulation techniques are being used. 2.2.2.3 Indirect Nerve Root Stimulation Techniques Indirect nerve root stimulation is performed by electrically stimulating bone or hardware in order to elicit nerve root responses. Some surgeons favor such testing during every aspect of pedicle screw placement. They test the probe used to make the initial hole into the pedicle for marker placement, the markers, the taps used to make the holes for the pedicle screws, the pedicle screw holes, and the pedicle screws themselves. Other surgeons may prefer to test only the screw placements. The assessment criteria are similar in all cases. Several articles have reported on the efficacy of these techniques [9, 13–16, 46, 52, 57–64]. The published stimulation parameters that have been used have varied. These studies have used either constant current or constant voltage stimulation to assess placements. Although similar, these two forms of stimulation are not equivalent. The flow of electrons, also known as current flow, is what actually causes a nerve or nerve root to depolarize. Voltage is only the driving force that causes the electrons to flow through the resistance or impedance of biological tissue. When testing pedicle screw placements, tissue impedance includes that of pedicle and vertebral body bone in addition to the impedance of muscle, vascular tissue, and blood. Although the latter impedances probably remain relatively constant between individuals, bone density and therefore bone impedance is known to vary between individuals as a result of osteoporosis and other factors. Therefore, it takes more or less voltage to cause the same current to flow in various individuals. Therefore, it would be expected that the results of using constant voltage for testing pedicle screw placements would be more variable than constant current stimulation. Several investigators have already reported this to be the case [9, 61], and our findings are in agreement (Fig. 11.5). Constant current stimulation appears to be superior to constant voltage stimulation for assessing pedicle screw placements.

Neurophysiological Monitoring During Pedicle Screw Placement

245

FIGURE 11.5 A graph of the constant voltage versus constant current stimulation intensities required to elicit myogenic responses when the same pedicle screw was stimulated. Nerve root excitation is always a function of current flow. Although a linear relationship exists between these two stimulation techniques, constant voltage stimulation appears more variable than constant current stimulation. This variability may be a function of the resistance of the pedicle bone.

FIGURE 11.6 The stimulation technique used to assess pedicle screw placements. The same technique can be used to test markers, taps, and pedicle holes. In this case, the pedicle screw has broken through the wall of the pedicle and is situated very close to an emerging nerve root. As a result, electrical current, following the path of least resistance through the pedicle screw and the breach in the pedicle wall, is expected to excite the nerve root. Excitation typically occurs at very low stimulation intensities and results in triggered myogenic responses from muscles innervated by the nerve root. Reprinted from [52].

Various stimulation parameters and techniques have been used to electrically assess pedicle screw placements. A probe of some type such as a nasopharyngeal electrode functions as the cathode and is placed within the pedicle screw holes and/or on hardware, and a needle electrode is typically placed in muscle near the surgical site (Fig. 11.6). It is used as the anode and provides a

246

Richard J. Toleikis

FIGURE 11.7 Radiographic data (A) and pedicle screw stimulation data (B, C) acquired during the same surgical procedure. (A) The lateral radiograph shows the pedicle screws that were placed into the L4 and L5 vertebrae. Although the radiograph suggests adequate screw placement at both these levels, it cannot indicate any medial screw displacement. (B, C) Triggered myogenic responses were elicited by stimulation of the left L4 (B) and left L5 (C) pedicle screws. The stimulation intensity was initially 0.0 mA and was gradually increased until triggered responses were evident. The threshold stimulation intensity (current level at which responses were first elicited) for the left L4 screw (B) was 32.6 mA, which was suggestive of adequate screw placement. The threshold stimulation intensity for the left L5 screw (C) was 4.3 mA, which was suggestive of a potentially harmful screw placement. The placement of the left L5 screw was visually inspected and confirmed to be positioned in the spinal canal. It was removed. Reprinted from [52].

Neurophysiological Monitoring During Pedicle Screw Placement

FIGURE 11.7

247

(Continued)

return path for the stimulation current. Rates of pulsatile stimulation have ranged from 1 to 5 Hz with pulse durations of 50–300 ms [61]. Typically, when testing, the intensity of the stimulation is gradually increased from 0 mA until a current threshold is reached at which a reliable and repeatable EMG response is elicited from at least one of the monitored muscle groups or a predetermined maximum stimulus intensity is reached. For safety reasons, we generally use 50 mA as a maximum stimulation intensity. If EMG responses are elicited at a stimulus intensity that is lower than a predetermined “warning threshold,” i.e., the stimulus intensity that is used to warn of a possible breach of the pedicle wall, the surgeon is advised to examine the hole or hardware placement. In such instances, radiographs may falsely suggest adequate screw placements (Fig. 11.7). These “warning thresholds” have varied between groups of investigators. Some have used stimulus intensities of 10 mA or higher, whereas others have used intensities as low as 6 mA [61].

248

Richard J. Toleikis

FIGURE 11.7

(Continued)

2.3 ANESTHETIC MANAGEMENT FOR MOTOR TECHNIQUES In order to provide an optimal surgical field, the anesthesiologist must render a patient unconscious and free from pain and must also control muscle tone. The degree of muscle relaxation is the only anesthetic factor of concern when myogenic activity is used for monitoring purposes. One way to suppress muscle tone is to suppress it at its origin, within the cerebral cortex, with deep anesthesia [65]. Although nitrous oxide does not produce muscle relaxation, the administration of the halogenated agents such as halothane, enflurane, and isoflurane does have a dose-related effect. However, because of cardiovascular depression, these agents cannot be used by themselves to produce the amount of relaxation necessary for abdominal surgery. A second means of diminishing muscle tone is to block the signals from the brain to the muscles as they traverse

Neurophysiological Monitoring During Pedicle Screw Placement

249

the spinal canal by using either spinal or epidural anesthesia. A third means of doing so is to use neuromuscular blocking agents that interfere with the transmission of signals from motor nerves to muscle fibers. In order to avoid major arterial hypotension, neuromuscular blockade is achieved by the use of a neuromuscular blocking agent in conjunction with a volatile halogenated agent. In this way, the anesthetic is used to produce only unconsciousness and analgesia and can be administered at low safe concentrations. When monitoring nerve root function using the spontaneous or triggered myogenic activity from specific muscle groups, it is imperative that these muscle groups be sensitive to changes in nerve root function resulting from traction or compression of the roots. The level of muscle relaxation significantly affects myogenic responses. The greater the degree of muscle relaxation for the muscle groups of interest, the less likely they will be to respond to changes in nerve root function. As a result, it would be ideal if no muscle relaxation were used to interfere with elicited activity. To be absolutely sure that relaxation levels play no part in determining response thresholds, some neurophysiologists insist on patients being totally unrelaxed when testing. However, in many clinical settings, this level of relaxation may be difficult if not impossible to achieve, particularly if surgeons feel that it compromises their ability to adequately perform surgery. An effective means of assessing the degree of muscle relaxation is to use a train-of-four technique, which consists of electrically stimulating a peripheral nerve four times and recording the four CMAPs (T1, T2, T3, and T4) that result from target muscle groups. For the hands, the ulnar nerve could be stimulated at the wrist with CMAPs recorded from the adductor digiti minimi muscle. For the legs, the peroneal nerve could be stimulated at the fibular head and CMAPs could be recorded from the tibialis anterior muscle. Typically, 2 Hz, 0.2 ms pulses of supramaximal stimulation intensity are used to elicit the CMAPs. The T4 CMAP disappears with a 75% blockade, the T3 with 80%, T2 with 90%, and T1 with 100% [66, 67].

2.4 FACTORS THAT CAN CONTRIBUTE TO FALSE-NEGATIVE FINDINGS When pedicle screws have breached pedicle walls, these events should be detectable when using electrical stimulation for test purposes. When this form of testing fails to detect these events, false-negative findings in the form of nerve root irritation or damage can result. Several factors, both technical and physiological, can contribute to such findings. The following is a discussion of these factors.

250

Richard J. Toleikis

2.4.1 Degree of Muscle Relaxation Probably the most important factor when testing during the placement of pedicle screws is the degree of muscle relaxation, because it can significantly influence the stimulation thresholds at which responses are elicited. Therefore, an accurate assessment of muscle relaxation is essential. Train-of-four testing is a common way for anesthesiologists to make these assessments. Although this testing technique appears to be a reasonable way to make these assessments, testing should be done by the person providing the monitoring rather than by the anesthesiologist, for several reasons. First, the anesthesiologist typically does a train-of-four assessment using a small portable battery-driven device. These devices may not always work properly and should not be relied upon. Second, the anesthesiologist’s assessment of train-of-four test results is a subjective one based on visible twitches from muscle groups that the anesthesiologist has access to—either hand or facial muscles. It is unlikely that the responses from these muscle groups will be the same as those from the leg muscles from which responses are elicited when pedicle screw testing is performed, because these muscle groups react differently to the relaxant levels. Finally, it is appropriate that the person providing the monitoring be responsible for guaranteeing that the test results are as accurate as possible by doing his or her own train-of-four testing of leg musculature. The issue of what are adequate relaxant levels for accurately assessing threshold stimulation intensities remains controversial. Clearly, no spontaneous or triggered myogenic activity will be present with 100% blockade. On the other hand, it has been reported that it is not necessary to have the absence of any blockade to effectively monitor spontaneous and triggered myogenic activity. We have observed that spontaneous activity can be elicited with one twitch present during train-of-four testing, or up to 90% neuromuscular blockade. However, when using indirect stimulation during pedicle screw placement, the assessment criteria make a determination of relaxant levels much more critical. It cannot be overstated how important it is to have the patient adequately unrelaxed when testing. If a patient is too relaxed when testing is performed, the stimulation thresholds for eliciting responses will be artificially elevated and may lead to false-negative findings. One example of this from personal experience occurred when a patient had only one large twitch during train-offour testing and pedicle screw testing was performed. The stimulation threshold was found to be 50 mA. Further screw testing was delayed until a small fourth twitch became evident during train-of-four testing. Stimulation of the same screw then resulted in a threshold of 12 mA! Clearly, the relaxant level associated with the presence of only one twitch during train-of-four testing may be adequate for eliciting spontaneous activity, but it is not adequate for testing during screw placement. It has been reported that the minimal criterion for

Neurophysiological Monitoring During Pedicle Screw Placement

251

FIGURE 11.8 Train-of-four test data. The right side of the figure indicates the four myogenic responses elicited from the left tibialis anterior muscle as a result of stimulation of the left peroneal nerve at the head of the fibula. Four 0.3 ms pulses with an intensity of 40 mA were presented at a rate of 2 Hz. On the left side of the figure, the myogenic response amplitudes resulting from the first and the fourth stimuli were measured and the amplitude ratio of the fourth over the first twitch was calculated. A minimum ratio of 0.1, based on previous findings [69], was used for testing purposes.

making such assessments is the presence of a fourth twitch from a target muscle [52, 68]. It has also been proposed that a postinduction–preinduction CMAP amplitude ratio from a hand muscle that is greater than 0.8 is a better measure for determining adequacy of relaxation [70]. Our own service is currently using an amplitude ratio of the fourth to the first twitch to determine adequacy of relaxation criteria for assessing screw placements. Based on surgical findings when visualizing screw placements after experiencing stimulation thresholds below “warning thresholds,” we now feel that the value of this ratio must be at least 0.1 (Fig. 11.8). The direct electrical stimulation of nerve roots is one means of determining if the level of muscle relaxation is adequate for using myotomal responses to

252

Richard J. Toleikis

assess nerve root function. If myogenic responses cannot be elicited using a constant current stimulation level of 2–4 mA, it is likely that the muscle relaxant level is too high to effectively monitor nerve root function using myogenic techniques. Considering the importance of proper relaxation, further studies need to be done to correlate the results of direct nerve root stimulation with noninvasive twitch monitoring. 2.4.2 Current Shunting It was pointed out earlier that excitation of a nerve root only occurs when a portion of the current being applied in a pedicle hole or to pedicle hardware is adequate to excite and depolarize the nerve root. The stimulation current that is used to test pedicle screw placements can exit the screw through several different pathways and will seek the pathways of least resistance as the current returns to the anodal electrode. When a pedicle screw breaches a pedicle wall, it provides a pathway for current to exit. The larger the breach, the lower the resistance to current flow, and the greater the amount of current that will flow through the breach. If a nerve root is located close to the breach, excitation of the nerve root will occur. However, if fluid is allowed to accumulate at the surgical site so that the fluid makes contact with the stud of a pedicle screw, that fluid will provide another low-resistance pathway for current to flow away from the screw. Less current will exit through the pedicle wall, and the amount of current needed to flow into the pedicle screw and cause depolarization of the nerve root will increase. As a result, we found that stimulation intensities needed to elicit myogenic responses generally increased between 12 and 20 mA [70] (Fig. 11.9). It is interesting that despite the low resistance of the fluid, current is not completely but only partially shunted away from the pedicle screw. Therefore, if shunting is present, it appears that it can mask the presence of a breached pedicle and result in false-negative findings when stimulation intensities are less than 30 mA. However, at stimulation intensities greater than this, such an occurrence seems unlikely. 2.4.3 Physiologic Factors The physiologic factors that can contribute to false-negative findings largely pertain to the health status of the stimulated nerve roots. The threshold criteria for all of the stimulation techniques that are used for testing purposes are based on the assumption that the nerve roots that are being excited by the pedicle screw stimulation are healthy and function normally. These nerve roots, when directly stimulated, have excitation thresholds of about 2 mA. However, it has been reported that chronically compressed nerve roots have elevated

Neurophysiological Monitoring During Pedicle Screw Placement

253

FIGURE 11.9 Data showing the effects of current shunting. Twenty-one pedicle screws were stimulated, and the threshold intensities for eliciting myogenic responses were determined with minimal fluid in the surgical field (“dry” testing) and with fluid in contact with the pedicle screws (“wet” testing). The consistent increase in stimulus intensities needed to elicit myogenic responses when “wet” testing indicates the effects of current shunting.

stimulation thresholds [57, 58]. For some of these nerve roots, thresholds may even exceed 20 mA [56]. Therefore, the use of the “warning threshold” for normal nerve roots when testing chronically compressed nerve roots may also contribute to false-negative findings. In the author’s experience, we have had two cases of chronic nerve root compression in which nerve root thresholds were determined as a result of direct nerve root stimulation and were found to be elevated. In the first case, the patient was diagnosed with spinal stenosis, and a right L5 nerve root that had been chronically compressed had a stimulation threshold of 5.5 mA. In the second case, the patient presented with a left-sided drop foot of about 1 month duration that occurred immediately after a previous back surgery. The placement of the left L5 screw was visually examined during surgery and was found to be in contact with the left L5 nerve root. Direct stimulation of the left L5 nerve root just outside the foramen resulted in a stimulation threshold of 13.7 mA. Stimulation thresholds may also be elevated when testing patients with metabolic disorders such as diabetes. Our experience with such patients is limited with regard to directly stimulating nerve roots and acquiring thresholds. Those we have tested have either exhibited normal thresholds or thresholds that have been only slightly elevated (thresholds of 4–6 mA).

254

Richard J. Toleikis

2.5 IMPEDANCE TESTING An alternative approach to pedicle screw stimulation is to use the electrical impedance of biological tissues as a means of assessing pedicle screw placements. Using a porcine model, Myers et al. [71] reported on a method for assessing pedicle wall thickness using impedance techniques. They were able to determine that the impedance of intact vertebral bone was about 400 Ω (400 ± 156 Ω) when a probe was first inserted into a pedicle and decreased as the depth of insertion increased. For an intact pedicle, the vertebral impedance decreased to 100 ± 22 Ω at maximum probe penetration. The accuracy of the technique was determined using postmortem anatomical confirmation of the pedicle probe placement and regression analysis of the impedance data. Based on this model, it was determined that impedance values below 58 Ω were associated with a 100% likelihood of a breach in the pedicle wall. These data were gathered from probing pedicle holes and measuring the impedance of the walls. Although very promising, the authors recognized that the technique’s utility needed to be demonstrated for implanted pedicle screws. Thus far, this utility has not been demonstrated. When testing implanted pedicle screws, the factors that contribute to the measured impedance become much more complex than simply measuring the tissue impedance at various points on the pedicle wall (Fig. 11.10). Other investigators [59] besides ourselves have compared the impedance measurements taken from pedicle screws to the results obtained via electrical stimulation. Our findings are in agreement that the impedance readings were very variable, with no correlation to electrical stimulation data or findings from visual observation (Fig. 11.11). At the present time, impedance measurements do not appear useful for assessing pedicle screw placements.

FIGURE 11.10 The impedance that is measured when a pedicle screw is in place is a complex measurement consisting of the combined impedance of many tissue elements.

Neurophysiological Monitoring During Pedicle Screw Placement

255

FIGURE 11.11 Impedance versus threshold stimulation intensity data. There was no correlation between the impedance readings taken from the top of the pedicle screws and the threshold current intensities needed to elicit myogenic responses or findings from visual inspection of screw placements.

Further refinements in the technique may be necessary to make this technique useful.

3 PERSONAL EXPERIENCE For many years, mixed-nerve SEPs have traditionally been used to monitor spinal cord function during spinal instrumentation procedures in order to minimize the probability of postoperative neurological deficits [17–20, 24, 25]. In the past 10–15 years, intrapedicular fixation of the thoracolumbar and lumbosacral spine by means of pedicle screw instrumentation has become increasingly popular. With increased use of pedicle screw instrumentation came varying degrees of neurological impairment. However, when the implementation of pedicle screws was increasing, the only forms of neurophysiological intraoperative monitoring that were available to avoid postoperative neurological deficits were mixed-nerve SEPs. One of the patients that was monitored in this fashion using SEPs elicited by posterior tibial nerve stimulation had complaints of paresthesia and numbness of the right great toe immediately after surgery. The scope and intensity of these symptoms increased, and the patient was clinically found to also have increasing weakness of the dorsiflexors and extensor hallucis longus muscle on the right side. A computed tomography (CT) scan was subsequently obtained that provided evidence of pedicular screw irritation of the nerve root (Fig. 11.2). Subsequent surgery revealed that the right L5 pedicular screw was located medial to the pedicle and juxtaposed to the right L5 nerve root. Review of the intraoperative monitoring tracings revealed no significant changes in the monitored responses throughout the surgical procedure. In addition, tracings acquired during postoperative testing

256

Richard J. Toleikis

were comparable to those obtained intraoperatively (Fig. 11.1). SEPs are typically mediated by several spinal nerve roots. It was surmised that the reason the monitored SEPs did not demonstrate any changes was that the functional compromise of the single L5 nerve root was masked by the normal volleys mediated by unaffected nerve roots. On the basis of this one case, it was determined that mixed-nerve SEPs might be an inappropriate tool for monitoring procedures in which nerve root rather than spinal cord function is at risk. The use of SEPs for monitoring purposes was abandoned, and other methodologies were sought for monitoring during pedicle screw procedures. As a result, DSEPs, which previous studies had already indicated were an effective means of assessing single nerve root function [35–44], were used to monitor pedicle screw procedures [34, 45]. Subsequently, the results of our experience using DSEPs as a monitoring tool during surgical intrapedicular fixation procedures were published [35]. They indicated that the loss of DSEP responses appeared to be a sensitive indicator of mechanical root compression (Fig. 11.3), whereas DSEP responses were rarely found to change to any significant degree during root decompression or even several days post surgery. However, because of the major shortcomings associated with DSEPs (i.e., the time-consuming need for averaging to acquire responses and the insensitivity to potential sources of nerve root irritation or damage), their use was later abandoned in favor of monitoring spontaneous myogenic activity in conjunction with indirect nerve root stimulation responses. Monitoring using a combination of both these techniques appears to have adequately addressed all of the DSEP shortcomings. Although the monitoring of spontaneous myogenic activity is used to safeguard nerve roots during pedicle screw placement, the actual probability of nerve root irritation or injury during these placements, although finite [35], appears to be very small. Having kept data from over 1000 surgical procedures that have involved the placement of over 5000 pedicle screws, we have never observed any sustained (longer than 2 s) spontaneous activity that was associated with the tapping of screw holes or the placement of markers or screws. Sustained spontaneous activity has been observed in less than 4% of the patients that have been monitored, and this activity has always been associated with mechanical nerve root irritation during traction or decompression. Others have reported similar findings [14]. Based on the data that were acquired from 662 patients using a “warning threshold” of 10 mA (using a stimulus duration of 0.2 ms), we have published our findings [52] on the correlation between responses elicited at or below warning threshold, the surgical findings, and the actions surgeons took based on visual inspection of the screw placements (Fig. 11.12). When EMG responses were elicited at or below 5 mA, screws were almost always removed and might be redirected. If responses were elicited at intensities at or greater

Neurophysiological Monitoring During Pedicle Screw Placement

257

FIGURE 11.12 Data from 662 patients whose screws were found to have a “warning threshold” less than or equal to 10 mA. After visual inspection, those screws with thresholds less than 5 mA were generally removed. Screws with thresholds between 5 and 8 mA were equally likely to be removed or left in place. Screws with thresholds of 8 mA or greater were generally left in place. However, some screws with thresholds as high as 10 mA were also removed. Reprinted from [52].

than 8 mA, screws were generally left in place. Between 5 and 8 mA, screws were equally likely to be removed or left in place. Therefore, despite the fact that stimulation thresholds less than 5 mA generally resulted in screw removal, screw removal also occurred at stimulation intensities up to and including the “warning threshold” of 10 mA. These findings support the findings of others [9, 10, 13–15, 46, 61, 62, 68], which indicate a close correlation between the intensity of screw stimulation needed to elicit myogenic responses and the risk for neurological injury associated with the screw placements. When electrical stimulation is used to assess hardware placement or the integrity of a pedicle hole, the stimulation current can take many pathways as it returns to the anodal needle electrode, but it will follow those pathways that provide the least resistance (Fig. 11.13). When hardware or pedicle hole stimulation is associated with low-threshold stimulation intensities, the results suggest that a path of least resistance is located near a nerve root, but one cannot tell whether the pathway is through a cracked pedicle, a thin wall of osteoporotic bone, or an exposed pedicle screw. Responses elicited at stimulation intensities below the “warning threshold” could result from current flow through any of these pathways. However, there does appear to be a relationship between threshold stimulation intensities and the exposure of a pedicle screw; i.e., cracked pedicles or a minimally exposed screw tends to be associated with stimulation thresholds greater than 7 mA, whereas exposed screws near a nerve root tend to have thresholds less than 5 mA. However, not all low-threshold

258

Richard J. Toleikis

FIGURE 11.13 The stimulation current that is used to assess pedicle screw placements can take many pathways as it returns to the anodal electrode that is placed in muscle tissue. The current will follow those pathways that provide the least resistance.

readings are associated with screw placements that represent threats of potential neurological injuries. Testing of pedicle screws in two patients resulted in each having a screw with a stimulation threshold less than 5 mA. After visual inspection, neither was removed because their placement did not appear to represent a threat of potential neurological injury. Neither patient experienced any postoperative pedicle screw–related neurological deficits. Based on the author’s experience and the results of other investigators, screw placements that are associated with stimulation thresholds greater than 10 mA are unlikely to represent a risk to neurological function if normal healthy nerve roots are involved and testing conditions are adequate. However, several factors, both technical and physiologic, can contribute to false-negative findings when stimulation thresholds exceed “warning thresholds.” These include excessive muscle relaxation, current shunting as a result of excessive fluid in the surgical site, and chronic nerve root compression. Examples of all three of these factors appear in the chapter text. Since 1995, our service has monitored well over 1000 patients during surgery in which over 5000 pedicle screws were placed. For each patient, a physical therapist and the attending physician routinely performed postoperative assessments. These practitioners were asked to inform the monitoring staff of any imaging or surgical evidence of misplaced hardware or any functional deficits that could be attributed to hardware placement. Based on their information, only one patient has experienced a postoperative neurological deficit directly attributable to a misplaced pedicle screw. In that patient, pedicle screws were placed from L3 to L5. The patient had stimulation thresholds that exceeded the “warning threshold” of 10 mA in all cases except for the left and

Neurophysiological Monitoring During Pedicle Screw Placement

259

right L3 screws, which had thresholds of 6.7 and 5.1 mA, respectively. The surgeon elected to leave these screws in place. Immediately after the operation, the patient experienced symptoms of low back pain and right leg pain. The patient was brought back to surgery, and both L3 pedicle screws were removed and replaced with sublaminar hooks at L3 and pedicle screws at S1 using Texas Scottish Rite Hospital (TSRH) instrumentation and a crosslink. Following this procedure, the patient’s leg pain completely resolved. This is considered a good example of a true-positive finding. We have also had what we consider to be one false-negative finding, although the patient did not experience any postoperative neurological symptoms as a result of the screw placement. After operation, the patient reported unilateral back and leg pain. Postoperative CT scans were not routinely performed for our patients, but in this case one was ordered and revealed a screw that was positioned medial to the pedicle in the spinal canal on the asymptomatic side. Because of the location of the screw, it was removed less than 1 week after it had been placed and before it caused any postoperative nerve root irritation. Unfortunately, in this one case, the screw was removed without any repeated testing to confirm earlier monitoring findings. A retrospective review of these findings indicated that the stimulation thresholds for the four placed screws were 40–50 mA. Routine train-of-four testing of the leg musculature performed by the monitoring staff just before and after screw placement indicated that the level of paralysis was adequate for accurate assessments because four full twitches could be elicited from the tibialis anterior muscle. Although this patient did not experience any new postoperative deficits as a result of screw placement, the electrical stimulation technique should have detected the misplaced screw. Thus this result is considered a false-negative finding. None of the factors that have been discussed earlier can explain these results. Therefore, it is possible that other factors that are not obvious to us may also contribute to false-negative findings.

4 CONCLUSIONS The incidence of neurological complications associated with the placement of pedicle screws has been reported as 2–10% [68, 72, 73]. Based on these estimates of incidence, we would then expect that for every 1000 patients in which pedicle screws were placed, between 20 and 100 should have exhibited some new postoperative neurological deficits directly attributable to screw placement. Our outcome data, which indicate that only 1 patient in our population of over 1000 patients has thus far exhibited such symptoms (actually, the result of a screw with a low test threshold that was left in place), clearly suggest that our use of pedicle screw stimulation to monitor screw placements has played an important role in minimizing the incidence of such deficits. The technique

260

Richard J. Toleikis

appears to be very reliable for detecting breaches of the pedicle wall, even those that may pose no threat of causing neurological irritation or injury. It provides an easy, quick, and accurate means to assess pedicle screw placements and to safeguard neurological function. In the real world, it is also essential that monitoring be cost-effective. That is, the overall costs of monitoring should not exceed the costs associated with patient care if monitoring is not provided. In most institutions at this time, the cost of monitoring for a typical instrumented fusion involving pedicle screw placement with spontaneous and triggered myogenic techniques is generally $1000 or less. Therefore, the cost associated with monitoring 1000 procedures would be $1 million. However, as indicated earlier, the minimal expected incidence of postoperative neurological deficits resulting from the placement of pedicle screws is 2%, and it would involve at least 20 patients. Therefore, if the average medical costs to correct a patient’s postoperative outcome and to rehabilitate that patient are more than $50,000, then the monitoring is cost-effective. It is very unlikely that $50,000 would cover all of the resultant medical costs. This discussion of cost-effectiveness does not even take into account the medical and legal costs associated with each of these occurrences. Clearly, monitoring during pedicle screw placement is cost-effective. All the techniques that can be used to monitor during pedicle screw placements have some limitations. The combined use of spontaneous and triggered myogenic activity is the only technique that meets all the necessary criteria if it is to be widely and effectively used to assess pedicle screw placements and to preserve nerve root function.

REFERENCES 1. Harrington, P.R. (1962). Treatment of scoliosis: Correction and internal fixation by spine instrumentation. J. Bone Joint Surg., 44, 591–610. 2. Luque, E.R. (1982). The anatomic basis and development of segmental spinal instrumentation. Spine, 7, 270–275. 3. Farber, G.L., Place, H.M., Mazur, R.A., Jones, D.E.C., and Damiano, T.R. (1995). Accuracy of pedicle screw placement in lumbar fusions by plain radiographs and computed tomography. Spine, 18, 1494–1499. 4. Ferrick, M.R., Kowalski, J.M., and Simmons, E.D. (1997). Reliability of roentgenogram evaluation of pedicle screw position. Spine, 20, 1249–1253. 5. Gertzbein, S.D., and Robbins, S.E. (1990). Accuracy of pedicular screw placement in vivo. Spine, 15, 11–14. 6. George, D.C., Krag, M.H., Johnson, C.C., Van Hal, M.E., Haugh. L.D., and Grobler, L.J. (1991). Hole preparation techniques for transpedicle screws: Effect on pull-out strength from human cadaveric vertebrae. Spine, 16, 181–184. 7. Roberts, R.M., and Bernhardt, M. (1992). Accuracy of fluoroscopic-guided pedicle screw placement in the lumbosacral spine. Scoliosis Research Society (abstract). 8. Steinman, J.C., Herkowitz, H.N., El-Kommos, H., and Wesolowski, P. (1993). Spinal pedicle fixation: Confirmation of an image-based technique for screw placement. Spine, 18, 1856–1861.

Neurophysiological Monitoring During Pedicle Screw Placement

261

9. Maguire, J., Wallace, S., Madiga, R., Leppanen, R., and Draper, V. (1995). Evaluation of intrapedicular screw position using intraoperative evoked electromyography. Spine, 20, 1068–1074. 10. Glassman, S.D., Dimar, J.R., Puno, R.M., Johnson, J.R., Shields, C.B., and Linden, R.D. (1995). A prospective analysis of intraoperative electromyographic monitoring of pedicle screw placement with computed tomographic scan confirmation. Spine, 20, 1375–1379. 11. Castro, W.H.M., Halm, H., Jerosch, J., Malms, J., Seinbeck, J., and Blasius, S. (1996). Accuracy of pedicle screw placement in lumbar vertebrae. Spine, 21, 1320–1324. 12. Calancie, B., Lebwohl, N., Madsen, P., and Klose, K.J. (1992). Intraoperative evoked EMG monitoring in an animal model: A new technique for evaluating pedicle screw placement. Spine, 17, 1229–1235. 13. Calancie, B., Madsen, P., and Lebwohl, N. (1994). Stimulus-evoked EMG monitoring during transpedicular lumbosacral spine instrumentation: Initial clinical results. Spine, 19, 2780–2786. 14. Clements, D.H., Morledge, D.E., Martin, W.H., and Betz, R.R. (1996). Evoked and spontaneous electromyography to evaluate lumbosacral pedicle screw placement. Spine, 21, 600–604. 15. Darden, B.V. II, Wood, K.E., Hatley, M.K., Owen, J.H., and Kostuik, J. (1996). Evaluation of pedicle screw insertion monitored by intraoperative evoked electromyography. J. Spinal Disord., 9, 8–16. 16. Cotton, M.A., Jenkins, J.A., Ehle, A., Schafer, M. (1997). The efficacy of evoked EMG from direct stimulation of pedicle screws. Proceedings of the 8th annual meeting of the American Society of Neurophysiological Monitoring, Chicago. 17. Engler, G.L., Spielholz, N.I., Bernhard, W.N., Danziger, F., Merkin, H., and Wolff, T. (1978). Somatosensory evoked potentials during Harrington instrumentation for scoliosis. J. Bone Joint Surg., 60A, 528–532. 18. Jones, S.J., Edgar, M.A., Ransford, A.O., and Thomas, N.P. (1983). A system for the electrophysiological monitoring of the spinal cord during operations for scoliosis. J. Bone Joint Surg., 65B, 134–139. 19. Nash, C.L. Jr, Lorig, R.A., Schatzinger, L.A., and Brown, R.H. (1977). Spinal cord monitoring during operative treatment of the spine. Clin. Orthop., 126, 100–105. 20. Spielholz, N.I., Benjamin, M.V., Engler, G.L., and Ransohoff, J. (1979). Somatosensory evoked potentials during decompression and stabilization of the spine. Spine, 4, 500–505. 21. Cohen, A.R., Young, W., and Ransohoff, J. (1981). Intraspinal localization of the somatosensory evoked potential. Neurosurgery, 9, 157–162. 22. Cusick, J.F., Myklebust, J.F., Larson, S.J., and Sances, A. Jr. (1978). Spinal cord potentials in the primate: Neural substrate. J. Neurosurg., 49, 551–557. 23. Larson, S.J., Sances, A. Jr., and Christenson, P.C. (1966). Evoked somatosensory potentials in man. Arch. Neurol., 15, 88–93. 24. Macon, J.B., Poletti, C.E., Sweet, W.H., Ojemann, R.G., and Zervas, N.T. (1982). Conducted somatosensory evoked potentials during spinal surgery: Part 1. Clinical applications. J. Neurosurg., 57, 354–359. 25. Jones, S.J., Edgar, M.A., and Ransford, A.O. (1982). Sensory nerve conduction in the human spinal cord: Epidural recordings made during spinal cord surgery. J. Neurol. Neurosurg. Psychiatry, 45, 446–452. 26. York, D.H. (1985). Somatosensory evoked potentials in man: Differentiation of spinal pathways responsible for conduction from forelimbs vs hindlimb. Prog. Neurobiol., 25, 1–25. 27. Powers, S.K., Bolger, C.A., and Edwards, M.S.B. (1982). Spinal cord pathways mediating somatosensory evoked potentials. J. Neurosurg., 57, 472–482. 28. Simpson, R.K. Jr., Blackburn, J.G., Martin, H.F. III, and Katz, S. (1981). Peripheral nerve fibers and spinal cord pathway contribution to the somatosensory evoked potentials. Exp. Neurol., 73, 700–715. 29. El-Nagamy, and E., Sedgewick, E.M. (1979). Delayed cervical somatosensory potentials in cervical spondylosis. J. Neurol. Neurosurg. Psychiatry, 42, 238–241.

262

Richard J. Toleikis

30. Ganes, T. (1980). Somatosensory conduction times and peripheral, cervical and cortical evoked potentials in patients with cervical spondylosis. J. Neurol. Neurosurg. Psychiatry, 43, 683–689. 31. Yu, Y.L., and Jones, S.J. (1985). Somatosensory evoked potentials in cervical spondylosis. Brain, 108, 173–200. 32. Cassvan, A., and Park, Y.S. (1983). Cortical somatosensory evoked potentials following peroneal nerve stimulation in lumbosacral radiculopathies. Electromyogr. Clin. Neurophysiol., 23, 393–402. 33. Feinsod, M., Blau, D., Findler, G., Hadani, M., and Beller, A.J. (1982). Somatosensory evoked potentials to peroneal nerve stimulation in patients with herniated lumbar disc. Neurosurgery, 11, 506–511. 34. Toleikis, J.R., Carlvin, A.O., Shapiro, D.E., and Schafer, M.F. (1993). The use of dermatomal evoked responses during surgical procedures that use intrapedicular fixation of the lumbosacral spine. Spine, 18, 2401–2407. 35. Aminoff, M.J., Goodin, D.S., Barbaro, N.M., Weinstein, P.R., and Rosenblum, M.I. (1985). Dermatomal somatosensory evoked potentials in unilateral lumbosacral radiculopathy. Ann. Neurol., 17, 171–176. 36. Eisen, A., Hoirch, M., and Moll, A. (1983). Evaluation of radiculopathies by segmental stimulation and somatosensory evoked potentials. Can. J. Neurol. Sci., 10, 178–182. 37. Owen, J.H., Padberg, A.M., Spahr-Holland, L., Bridwell, K.H., Keppler, L., and Steffee, A.D. (1991). Clinical correlation between degenerative spine disease and dermatomal somatosensory evoked potentials in humans. Spine, 16, S201–205. 38. Sedgewick, E.M., Katifi, H.A., Docherty, T.B., and Nicpon, K. (1985). Dermatomal somatosensory evoked potentials in lumbar disc disease. In “Evoked potentials: Neurophysiological and clinical aspect” (C. Morocutti, and P.A. Rizzo, eds.), pp. 77–88. Elsevier Science Publications B.V. (Biomedical Division). 39. Schramm, J., and Hashizume, K. (1977). Somatosensory evoked potentials (SEP) in patients with peripheral, spinal and supraspinal lesions of the sensory system. Adv. Neurosurg., 4, 250–256. 40. Scarff, T.B., Dallmann, D.E., Toleikis, J.R., and Bunch, W.H. (1979). Dermatomal somatosensory evoked potentials in children with myelomeningocele. Z. Kinderchir., 28, 384–387. 41. Toleikis, J.R., and Sloan, T.N. (1987). A comparison study of major nerve and dermatomal somatosensory evoked potentials in the evaluation of spinal cord injured patients. In “Evoked potentials III” (C. Barber, and T. Blum, eds.), pp. 309–316. Butterworths, Stoneham, MA. 42. Cohen, B.A., and Huizinga, B.A. (1988). Dermatomal monitoring for surgical correction of spondylolisthesis. Spine, 13, 1125–1128. 43. Cohen, B.A., Major, M.R., and Huizinga, B.A. (1991). Predictability of adequacy of spinal root decompression using evoked potentials. Spine, (Suppl), 16, S379–384. 44. Herron, L.D., Trippi, A.C., and Gonyeau, M. (1987). Intraoperative use of dermatomal somatosensory-evoked potentials in lumbar stenosis surgery. Spine, 12, 379–383. 45. Coscia, M.F., Trammell, T.R., Popp, B., Gawande, S.R., Fitzgerald, J., and Scott, J.R. (1995). Effect of anesthetic variables on dermatomal somatosensory-evoked potential monitoring in elective lumbar spine surgery. J. Spinal Disord., 8, 451–456. 46. Owen, J.H., and Toleikis, J.R. (1997). Nerve root monitoring. In “The textbook of spinal surgery,” 2nd ed. (K.H. Bridwell, and R.L. DeWald, eds.), pp. 61–75. Lippincott-Raven, Philadelphia. 47. Sloan, T.B. (1998). Anesthetic effects on electrophysiologic recordings. J. Clin. Neurophysiol., 15, 217–226. 48. Angel, A., and Grafton, D.A. (1982). The effect of anesthetic agents on cerebral cortical responses in the rat. Br. J. Pharmacol., 76, 541–549. 49. Sloan, T. (1997). Evoked potentials. In “Textbook of neuroanesthesia with neurosurgical and neuroscience perspectives” (M.A. Albin, ed.), pp. 221–276. McGraw-Hill, New York.

Neurophysiological Monitoring During Pedicle Screw Placement

263

50. Nuwer, M.R. (1993). Intraoperative electroencephalography. J. Clin. Neurophysiol., 10, 437–444. 51. Howe, J.F., Loeser, J.D., and Calvin, W.H. (1977). Mechanosensitivity of dorsal root ganglia and chronically injured axons: A physiological basis for the radicular pain of nerve root compression. Pain, 3, 25–41. 52. Toleikis, J.R., Skelly, J.P., Carlvin, A.O., Toleikis, S.C., Bernard, T.N., Burkus, J.K., Burr, M.E., Dorchak, J.D., Goldman, M.S., and Walsh,T.R. (2000). The usefulness of electrical stimulation for assessing pedicle screw placements. J. Spin. Disord., 13, 283–289. 53. Adams, D.C., Emerson, R.G., Heyer, E.J., McCormick, P.C., Carmel, P.W., Stein, B.M., Farcy, J.P., and Gallo, E.J. (1993). Monitoring of intraoperative motor-evoked potentials under conditions of controlled neuromuscular blockade. Anesth. Analg., 77, 913–918. 54. Drenger, B., Parker, S.D., McPherson, R.W., North, R.B., Williams, G.M., Reitz, B.A., and Beattie, C. (1992). Spinal cord stimulation evoked potentials during thoracoabdominal aortic aneurysm surgery. Anesthesiology, 76, 689–695. 55. Glassman, S.D., Johnson, J.R., Shields, C.B., Backman, M.H., Paloheima, M.P.J., Edmonds, H.L., and Linden, R.D. (1993). Correlation of motor-evoked potentials, somatosensory evoked potentials, and the wake-up test in a case of kyphoscoliosis. J. Spin. Disord., 6, 194–198. 56. Jellinek, D., Jewkes, D., and Symon, L. (1991). Noninvasive intraoperative monitoring of motor evoked potentials under propofol anesthesia: Effects of spinal surgery on the amplitude and latency of motor evoked potentials. Neurosurgery, 29, 551–557. 57. Szkiladz, E., Calder, H.B., and Easton, R.W. (1995). Intraoperative lumbo-sacral nerve root threshold measurement. Proceedings of the sixth annual meeting of the American Society of Neurophysiological Monitoring, San Francisco. 58. Holland, N.R., Lukaczyk, T.Q., Riley, L.H., and Kostuik, J.P. (1998). Higher electrical stimulus intensities are required to activate chronically compressed nerve roots. Spine, 23, 224–227. 59. Darden, B.V., Owen, J.H., Hatley, M.K., Kostuik, J., and Tooke, S.M. (1998). A comparison of impedance and electromyogram measurements in detecting the presence of pedicle wall breakthrough. Spine, 23, 256–262. 60. Glassman, S.D., Dimar, J.R., Puno, R.M., Johnson, J.R., Shields C.B., and Linden, R.D. (1995). A prospective analysis of intraoperative electromyographic monitoring of pedicle screw placements with computed tomographic scan confirmation. Spine, 20, 1375–1379. 61. Isley, M.R., Pearlman, R.C., and Wadsworth, J.S. (1997). Recent advances in intraoperative neuromonitoring of spinal cord function: Pedicle screw stimulation techniques. Am. J. Electroneurodiagn. Technol., 37, 93–126. 62. Lenke, L.G., Padberg, A.M., Russo, M.H., Bridwell, K.H., and Gelb, D.E. (1995). Triggered electromyographic threshold for accuracy of pedicle screw placement: An animal model and clinical correlation. Spine, 22, 1585–1591. 63. Rose, R.D. (1997). Spinal cord monitoring. Curr. Opin. Orthop., 8, 49–57. 64. Rose, R.D., Welch, W.C., Balzer, J.R., and Jacobs, G.B. (1997). Persistently electrified pedicle stimulation instruments in spinal instrumentation. Spine, 22, 334–343. 65. Waud, B.E. (1988). Neuromuscular blocking agents. In “Introduction to anesthesia: The principles of safe practice” (R.D. Dripps, J.E. Eckenhoff, and L.D. Vandam, eds.), pp. 166–187. W.B. Saunders, Philadelphia. 66. Collins, V.J. (1993). General and regional anesthesia. In “Principles of anesthesia,” 3rd ed., vol. 2, pp. 850–937. Lea & Febiger, Philadelphia. 67. Viby-Mogensen, J. (1994). Neuromuscular junction monitoring. In “Anesthesia,” 4th ed., (R.D. Miller, ed.), pp. 1345–1361. Churchill-Livingston, New York. 68. Holland, N.R. (1998). Intraoperative electromyography during thoracolumbar spinal surgery. Spine, 23, 1915–1922. 69. Minahan, R.E., Riley, L.H., Luckaczyk, T., Cohen, D.B., and Kostuik, J.P. (2000). The effect of neuromuscular blockade on pedicle screw stimulation thresholds. Spine, 25, 2526–2530.

264

Richard J. Toleikis

70. Skelly, J.P., Toleikis, J.R., and Carlvin, A.O. (1999). Pedicle screw stimulation in a fluid environment. Proceedings of the 10th annual meeting of the American Society of Neurophysiological Monitoring, Denver, Colorado. 71. Myers, B.S., Hasty, C.C., Floberg, D.R., Hoffman, R.D., Leone, B.J., and Richardson, W.J. (1995). Measurement of vertebral cortical integrity during pedicle exploration for intrapedicular fixation. Spine, 20, 144–148. 72. Jacob, R.P., Mack, C.A., and Fessler, R.G. (1994). Pedicle screws: Biomechanics, uses, and current assessment of outcomes. Neurosurg. Q., 4, 39–50. 73. Matsuzaki, H., Yasuaki, T., Fujio, M., Hoshino, M., Kiuchi, T., and Toriyama, S. (1990). Problems and solutions of pedicle screw plate fixation of lumbar spine. Spine, 15, 1159–1165.

CHAPTER

12

Surgery of Brainstem Lesions ALBINO BRICOLO AND FRANCESCO SALA Section of Neurosurgery, Department of Neurological Sciences and Vision, Verona University, Verona, Italy

1 Introduction 2 Patient Selection and Rationale for Surgery 3 General Principles of the Surgical Strategy 3.1 Midbrain 3.2 Pons 3.3 Medulla and Cervicomedullary Junction 4 Postoperative Care 5 Neurophysiological Monitoring 6 Conclusion References

ABSTRACT During the decade of the brain, a more rational and constructive approach to the surgical management of neoplastic and vascular brainstem lesions has emerged. The attitude held by many in the neurosurgical community who regard the brainstem as “untouchable” has been progressively counterbalanced by increasing evidence that surgical “violation” of the brainstem is safely feasible in certain subgroups of patients. Multiplanar MRI allows for an accurate distinction between different growth patterns of brainstem mass lesions. MRI data, together with a refined neurological evaluation, support the decision-making process in selecting those patients amenable to surgical treatment and help in planning an optimal surgical strategy. Surgery of the brainstem, however, remains a challenging task that requires a detailed understanding of the microsurgical and functional correlative anatomy of the brainstem. In this chapter we will first revisit those aspects of functional neuroanatomy relevant to modern neurosurgical approaches to the brainstem. The anatomy of the brainstem is characterized by an extremely dense, although phylogenetically ordered, concentration of functionally relevant structures such as cranial nerve nuclei, motor and sensory pathways, crossing bundles, and the reticular formation. Based on results from our experience with over 250 brainstem gliomas and vascular lesions surgically treated at the Department of Neurosurgery of Verona over Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

267

268

Albino Bricolo and Francesco Sala the last 15 years, a map of safe “entry zones” is suggested that may guide the surgeon in choosing the least dangerous approach to this critical area. We will then critically review surgical approaches to the brainstem, from the midbrain to the cervicomedullary junction and the fourth ventricle. Finally, in the spirit of this book, those mapping and monitoring techniques that may play a role in enhancing the safety of the different approaches will be outlined.

1 INTRODUCTION In 1939, Bailey and colleagues described the treatment of brainstem gliomas as “a pessimistic chapter” in the history of neurosurgery [1]. Thirty years later Matson stated that “regardless of specific histology [they] must be all classified as malignant tumors, since their location in itself renders them inoperable” [2]. Traditionally, all tumors involving the brainstem were considered infiltrative with diffused glial proliferation. Therefore, in consideration of the high density of cranial nerve nuclei, fascicles, and pathways contained within the brainstem, this small part of the encephalon—approximately 6 cm high and 3.5 cm wide at the pons—has been considered untouchable by several generations of neurosurgeons. This attitude, which still prevails in a part of the neurosurgical community, has been progressively counterbalanced by increasing evidence that surgical “violation” of the brainstem is safely feasible in a subgroup of patients. Favorable results appeared in the literature starting in the early 1980s [3–7]. More recently, brainstem surgery has been promoted by a number of authors who are convinced that this delicate and difficult field of surgery can now be approached with increased confidence [8–14]. More refined imaging techniques [15–18], better surgical techniques and equipment [4–19], and more effective neuroanesthesia and postoperative intensive care [20–21] have all improved the feasibility and safety of brainstem surgery. The contribution of MRI to this field has been immeasurable, given its ability to determine that not all tumors in the brainstem are diffuse. As a result, based on MRI, brainstem gliomas may be categorized into four groups: (1) focal; (2) cervicomedullary; (3) dorsally exophytic; or (4) diffuse. These subgroups, which reflect different growth patterns, not only assist the surgeon in identifying tumor invasiveness but also aid in selecting those tumors amenable to surgical treatment and determination of prognosis [6, 11, 16, 17, 22–24]. Moreover, this grouping correlates well with tumor histology, since almost all diffuse tumors are infiltrative, highly aggressive lesions that are always malignant, regardless of histology at the time of biopsy [23]. In the other groups, the great majority of tumors are benign, low-grade astrocytomas. During the decade of the brain, a more rational and constructive approach to the surgical management of neoplastic and vascular brainstem lesions has emerged. Yet surgery of the brainstem remains a challenging task that requires deep knowledge of the microsurgical and functional correlative anatomy of the

270

Albino Bricolo and Francesco Sala

low-grade astrocytomas. Subtotal or partial removal was done in 87% of the 72 malignant gliomas, with the large majority demonstrating diffuse MRI patterns. This series strongly confirms that brainstem gliomas have to be categorized in distinct subgroups and no longer represent a homogeneous nosologic entity amenable only to nonspecific treatment. The focal brainstem glioma is an expanding mass that usually dislocates neighboring nervous structures without invading them, and tumor growth is infiltrative only in the subset of diffuse tumors [11, 23, 27].

3 GENERAL PRINCIPLES OF THE SURGICAL STRATEGY The brainstem may be defined as the part of the neuraxis located between the diencephalon and the spinal cord, into which it continues without definite anatomical demarcation. Comprising the midbrain, pons, and medulla oblongata, the brainstem is almost entirely contained in the posterior fossa, except for a small rostral portion that goes beyond the tentorial incisure and a short tract of the medulla oblongata that runs below the foramen magnum. Crowded by cranial nerve and nuclei and ascending, descending, and interconnecting fascicles, bundles, pathways, and the reticular formation, the brainstem presents a highly complex structure both anatomically and functionally. This makes it a neurological “minefield,” and surgical resection of brainstem tumors demands meticulous microsurgical technique because of the narrow routes leading to the lesion. If the tumor is exophytic or fungating out from the brainstem surface, its removal clearly begins at such an outgrowth. Thus the tumor itself creates its own entry into the brainstem, where it may be penetrated without any risk and eventually removed. Other tumors characteristically bulge without violating the brainstem surface and may be seen under the pia or ependyma. In these cases, the access route for removal is also provided by the tumor itself, but much care must be paid when widening the entry point, given the functional significance of the surrounding structures, which is vital to the application and direction of microretraction. Tumors with no surface components—the pure intrinsic tumors— require even greater care and understanding of the involved functional anatomy. With a clear mental image of the internal architecture of the brainstem and the possible deficits that may be incurred by surgical injury, the surgeon chooses the safest entry zones, avoiding those that could be more dangerous [28]. Entering the floor of the fourth ventricle, through which some tumors are reached, requires a clear understanding of the underlying structures (Fig. 12.1). Within the small concavity of the calamus scriptorius situated above the obex and usually below the striae medullares lie two triangles of great functional importance: the hypoglossal triangle and the ala cinerea or vagal triangle.

Surgery of Brainstem Lesions

271

FIGURE 12.1 The positions of the cranial nerve motor nuclei are schematically projected on the dorsal surface of the brainstem (A) and outlined on the paramedian sagittal section (B). Awareness of their location may assist in selecting the safest surgical entry zone. Modified from [11].

Immediately below the two medial triangles lie the hypoglossal nuclei, which control the muscles of the tongue. Because of the close proximity of the two nuclei, surgical injury to this area almost always results in severe tongue paralysis and atrophy. Since hypoglossal paralysis represents one of the most devastating cranial nerve deficits, even a minor injury in this area must be avoided. Lateral to the hypoglossal are the vagal triangles, and under these lie the dorsal nuclei of the vagus, from where motor fibers to the bronchi, heart, and stomach originate. Slightly deeper and laterally lies the nucleus ambiguus, which gives rise to fibers of the glossopharyngeal (IX), vagus (X), and accessory (XI) nerves supplying musculature to the palate, pharynx, and larynx. Therefore, injury to the small concavity that forms the inferior part of the rhomboid fossa, the calamus scriptorius, may result in deficits such as impaired swallowing, dysphonia, nose regurgitation, and coughing reflex loss, thus exposing the patient to the risk of aspiration pneumonia and the incapacity to eat or drink [29]. The more prominent part of the median eminence, the facial colliculus, represents a second highly dangerous brainstem “entry zone” through the rhomboid fossa [30]. Damage to this area invariably causes facial (VII) and abducens (VI) nerve paralysis, as well as lateral gaze disturbances due to parapontine reticular formation dysfunction. Injury to the medial longitudinal fascicles, which border the median sulcus and lie between the abducens and oculomotor nuclei (the

272

Albino Bricolo and Francesco Sala

so-called VI–III pathway), may cause internuclear ophthalmoplegia. Despite the high density of neural structures involved in oculomotion, one can leave the latter quite undisturbed by entering the floor of the fourth ventricle through the median sulcus above the facial colliculi, as long as the two medial longitudinal fascicles, which have no crossing fibers at this level, are not injured by retraction [11, 31]. Consequently, when surgical incision and microretraction on the floor of the fourth ventricle are required, four safer entries are advisable: the suprafacial, the infrafacial, the midline above the facial colliculus, and the acoustic area. Figures 12.2 and 12.3, respectively, present all the relatively safe entry zones to the dorsal brainstem and the dangerous areas associated with expected neurological deficits. Multiplanar MRI permits an accurate preoperative localization of the tumor and its relationship to brainstem structures. This invaluable information, together with the patient’s neurological picture and the aforementioned neuroanatomical considerations, should all help clarify the optimal approach to surgical removal of a brainstem lesion. Because of the small size of the brainstem, the surgical approach must be in all senses “minimally invasive”: a surgical microscope, often used at the highest magnification, fixed microretractors, and a few tiny instruments are therefore the standard tools for this type of surgery. In the following section, we will briefly describe the major surgical routes we routinely use to approach lesions from the upper brainstem caudally to the cervicomedullary junction.

3.1 MIDBRAIN Since almost all midbrain gliomas are focal, benign astrocytomas (Table 12.1), the planned goal of surgery in this setting must be complete removal of the tumor, which will result in permanent cure. These tumors usually arise from either the tectal plate or the tegmentum and may extend upward to the thalamus or downward to the pons, displacing but not infiltrating these structures [3, 32, 33]. The midbrain, which occupies the notch of the tentorium, consists of a dorsal part (the corpora quadrigemina or tectum), a large ventral portion (the tegmentum), and the cerebral peduncles. Posterior cerebral (PCA) and superior cerebellar (SCA) arteries encircle the midbrain and in their course are in close relationship to the oculomotor (III) and trochlear (IV) cranial nerves. Minimizing manipulation of these normal neurovascular structures is the key to successful extirpation of the lesion in such a constricted area [34]. Lesions of the dorsal mesencephalon in the tectal area of the quadrigeminal plate are easily approached through a standard infratentorial supracerebellar approach popularized by Stein [35, 36] for pineal region neoplasms. This approach may also be used successfully for removing lower median tumors in the quadrigeminal region when an infracollicular entry is chosen to avoid ocular and auditory disturbances associated with injury to the colliculi. With a meticulous opening of the cerebellomesencephalic fissure and a retractor pressing

Surgery of Brainstem Lesions

273

FIGURE 12.2 Relatively safe entry zones into the dorsal brainstem. Supracollicular (A), infracollicular (C), and lateral mesencephalic sulcus (B) are suitable entries for the removal of tectal mesencephalic tumors approached by the infratentorial-supracerebellar route. The median sulcus above the facial colliculus (D), suprafacial (E), infrafacial (F), and area acoustica (G) provides safe entry for dorsal pontine tumors approached through the floor of the fourth ventricle. The posterior median fissure below the obex (H), the posterior intermediate sulcus (I), and the posterior lateral sulcus ( J) are sites recommended for longitudinal myelotomies to approach medullary and cervicomedullary junction tumors. Reprinted from [31].

down the vermis, exposure as far down as the inferior colliculi may be obtained without splitting the anterior vermis. A particular modification of this route can be used for successful removal of lesions more ventrally placed in the tegmentum and peduncles. A semi-sitting position and suboccipital craniotomy are the standard approach, with the exception that the dural opening is extended more laterally very close to the sigmoid

274

Albino Bricolo and Francesco Sala

FIGURE 12.3 Outline of the dangerous areas for entering the dorsal brainstem. Superior colliculus (associated with visual and oculomotor disorders) (A), inferior colliculus (auditory disturbances) (B), corpora quadrigemina (C) (as in A and B), medial longitudinal fascicles (internuclear ophthalmoplegia) (D), facial colliculus (facial palsy and internuclear ophthalmoplegia) (E), facial colliculus (immobile eyes and bilateral facial palsy) (F), facial nerve (facial palsy) (G), hypoglossal and vagus nuclei (dysphagia) (H), calamus scriptorius (dysphagia and cardiorespiratory disturbances) (I), and gracilis and cuneate tubercles (ataxia) ( J). Modified from [11].

sinus on the side of the chosen entry: a careful dissection of the arachnoid membranes allows a wide opening of the cerebellomesencephalic fissure on this side. A retractor is then placed to weigh down the anterior part of the tentorial surface of the cerebellum. With this done, a full exposure of the lateral aspect of the midbrain is obtained. By moving the SCA and the fourth nerve, it is easy to identify the lateral mesencephalic vein, which courses into the lateral

Surgery of Brainstem Lesions

275

mesencephalic sulcus. The entry is posterior to this sulcus to avoid injury to the pyramidal tract in the peduncle. To avoid damage to the colliculi in the supracerebellar approach to tumors located anterior to the quadrigeminal plate, an alternative adopted by many could be the occipital transtentorial approach [37–39], which calls for a right occipital craniotomy and sectioning of the tentorium adjacent to the straight sinus. The main advantage offered by this route, compared to the supracerebellar infratentorial, is direct vision into the fissure between the superior vermis, the quadrigeminal plate, and the superior medullary velum. However, one must take into consideration that the occipitoparietal lobe retraction required for exposing the falcotentorial junction may induce sensory and visual field deficits [40]. Tumors of the midbrain located near its median ventral surface may be accessed through a pterional transsylvian route after a frontotemporal craniotomy as developed and extensively applied by Yasargil [41]. Following a wide opening of the sylvian fissure, the tentorial edge, third cranial nerve, and interpeduncular cisterns are exposed. The anatomical landmark of the target area is the emergence of the third nerve from the midbrain with the PCA above. Below the nerve, the SCA runs medial to lateral, and the safe entry zone into the midbrain is the small rectangular area outlined medially by the exit of the third nerve and the basilar artery (BA), inferiorly by the SCA, superiorly by the PCA, and laterally by the tentorial edge. This narrow but fairly safe window allows surgical access through the more medial part of the peduncle, sparing the motor tract that occupies only the intermediate three fifths or so of the peduncle [11]. Tumors involving the anterolateral aspect of the midbrain can be reached through a subtemporal transtentorial approach [42]. The tentorial incisure is divided posterior to the entry of the fourth nerve, allowing exposure of the anterolateral midbrain and upper pons. This seemingly attractive route, at least in this author’s experience, is associated with some risk because it requires temporal lobe retraction and thus may injure the vein of Labbé.

3.2 PONS Tumors that involve one side of the ventral pons and fungate into the area of the cerebellopontine angle may be reached by a standard retrosigmoid approach through a lateral suboccipital retromastoid craniectomy as used in acoustic neuroma surgery. Also, a number of diffuse gliomas appear to originate from the ventral side of the pons and grow in the direction of the pontocerebellar fibers toward the cerebellar peduncle and the cerebellopontine angle. In such cases, MRI shows the ventral pons deeply grooved by the BA and contralaterally rotated completely to the side of the tumor, resulting in a filled cerebellopontine angle. With the patient in a semi-sitting position with the head rotated toward the side of the expansion, the bulging pons is entered through the fissure

276

Albino Bricolo and Francesco Sala

between the stretched fifth and the seventh through eighth cranial nerves. The diffuse tumor is then generously debulked [11]. The uncommon focal pontine tumor that is ventrolaterally placed requires a combined petrosal approach that combines subtemporal and transtentorial presigmoid avenues as described by Al-Mefty [43] and Spetzler [44]. The main advantage of this opening is the short distance and the direct line of light to the anterolateral brainstem (Figs. 12.4 and 12.5).

FIGURE 12.4 Large cavernous malformation in the pons removed through a posterolateral transpetrosal approach. Pre- and postoperative coronal (A) and axial (B) MRIs. Note, in the early T2 postoperative axial images, the route followed for removal.

FIGURE 12.5 Intraoperative microphotographs of the patient in Fig. 12.4. The right lateral aspect of the pons, encircled by the superior cerebellar artery and the trigeminal root, is directly exposed (A) and entered. The cavernoma is easily removed without any surgical injury to the neurovascular structures involved (B, C). 277

278

Albino Bricolo and Francesco Sala

FIGURE 12.5

(Continued)

For tumors located in the dorsal part of the pons (and the open portion of the medulla), access is by a suboccipital craniotomy and trans-fourth-ventricle route. It should be noted that the temptation of splitting the vermis during this approach may result in significant postoperative disturbances, such as body truncal ataxia, a wobbling gait, oculomotor disturbances, and cerebellar mutism [45]. In our experience [11], the vermis can be preserved by positioning the patient in a semi-sitting position with the head fairly flexed, so that a wide exposure of the rhomboid fossa may be obtained through the cerebellomedullary fissure. By elevating and splitting the cerebellar tonsils and displacing the posterior inferior cerebellar arteries (PICAs), which course into the fissure itself, one can expose the tela choroidea of the roof of the fourth ventricle, cut it at the taenia at both sides, and then fold it back upward to expose both the lateral recesses, if necessary. Next, the tela choroidea can be divided longitudinally up to the anterior medullary velum, to which it is attached along with the choroid plexus. At that point, two strategically positioned retractors will keep the access open

Surgery of Brainstem Lesions

279

and a suitable angulation of the surgical microscope will yield a complete view of the floor of the fourth ventricle from the obex to the cerebral aqueduct. This route can also be used for some tumors of the dorsal mesencephalon that dip into the ventricular chamber, thereby avoiding possible injury to the colliculi. For tumors further away from the midline (i.e., those located in the area acoustica and those growing in the middle cerebellar peduncle), access can be obtained only through the ipsilateral cerebellomedullary fissure. This creates a posterolateral paratruncal route that allows for control of the region of the lateral recess and that of the perimedullary area on the same side.

3.3 MEDULLA AND CERVICOMEDULLARY JUNCTION Tumors developing in the closed part of the medulla near its posterior aspect and in the cervicomedullary junction are approached through a low midline suboccipital craniectomy (or craniotomy) extended to the posterior arch of the atlas, followed by the necessary cervical laminotomy if the lesion extends further caudally (Figs. 12.6 and 12.7). When the tumor is medial, it is accessed through a midline longitudinal myelotomy and removed with a similar technique to that normally used for a spinal cord tumor (Figs. 12.8 and 12.9). For tumors laterally placed, either the posterior intermediate or the posterior lateral sulcus is used for entrance. When the tumor is more laterally and/or ventrally located, a dorsolateral approach is recommended. This exposure, designed by Heros [46] and implemented by Spetzler and Grahm [47] for vertebrobasilar junction aneurysms, provides excellent exposure of the anterolateral aspect of the medulla, the cervicomedullary junction, the foramen of Luschka, cranial nerves IX–XII, and associated arteries. For removal of intra-axial lesions that expand the lower brainstem, it is not necessary to perform extensive bone removal of the occipital condyle and the lateral mass of C1. Adequate exposure is obtained by a restricted retrosigmoid craniotomy, a C1-hemilaminectomy, and drilling away of the posterior third of the occipital condyle.

4 POSTOPERATIVE CARE Reviewing the first 175 of more than 250 patients operated on in Verona for intrinsic brainstem lesions, we had no intraoperative mortality, and 6 patients died in the first month after surgery. Despite advancements in intra- and postoperative care that have significantly reduced mortality and morbidity related to brainstem surgery, patients often present with new deficits or a worsening of existing deficits in the early postoperative period. Most of these deficits, however,

FIGURE 12.6 Comparable pre- and postoperative sagittal (A), coronal (B), and axial (C) MRIs of a dorsally exophytic astrocytoma removed via a midline suboccipital approach. A tiny sole of non-enhancing tumor was left over the calamus scriptorius. 280

Surgery of Brainstem Lesions

281

FIGURE 12.7 Intraoperative views of the same patient as shown in Fig. 12.6. The tumor bulging at the obex is debulked (A) and then detached from the floor of the fourth ventricle (B), which is largely exposed at the end of the tumor removal (C).

282

Albino Bricolo and Francesco Sala

FIGURE 12.8 Intrinsic ependymoma of the cervicomedullary junction. Pre- and postoperative sagittal (A), preoperative axial (B), and postoperative axial (C) MRIs showing a total removal.

FIGURE 12.9 Intraoperative microphotographs of the patient in Fig. 12.8. The dramatically enlarged cervicomedullary junction is exposed (A). It is entered through a midline myelotomy and the tumor is removed with the aid of an ultrasonic aspirator during continuous lower cranial nerve monitoring (B). After radical removal of the tumor, the floor of the fourth ventricle (C) and the opened cervicomedullary junction (D) are neurophysiologically mapped. The responses from both the left (LH) and the right (RH) hypoglossal nerves are recorded through needle electrodes inserted in tongue muscles (E). 283

284

Albino Bricolo and Francesco Sala

FIGURE 12.9

(Continued)

improve within the first few weeks, and patients are stabilized or even improved at discharge. The most dangerous surgical complication is palsy of the lower cranial nerves (IX–XII), which causes dyspnea and severe dysphagia, thus requiring tracheostomy and persistent attention to prevent aspiration pneumonia [21]. To minimize these complications, as a rule, we keep patients in the intensive care unit for at least 24 hr following surgery. The oral or nasal tracheal tube is maintained with mechanical ventilation and necessary sedation. Three or 4 hours after surgery a CT scan is usually taken to identify any early blood clotting, pneumocephalus, and hydrocephalus (although such complications are very uncommon). Once one has a “clean” early postoperative CT scan, the tracheal tube may be safely removed when the patient regains consciousness and normal ventilation parameters [31].

Surgery of Brainstem Lesions

285

The removal of the tube must be done with great caution, particularly in patients operated on for lower brainstem tumors in whom dysphagia, vocal cord paresis, and loss of the cough and gag reflexes may be expected. Also, in patients having severe dysphagia, it has proven beneficial to delay tracheostomy because often with sedulous care and rehabilitation training these disturbances resolve and the patient regains the ability to swallow [48].

5 NEUROPHYSIOLOGICAL MONITORING A detailed analysis of neurophysiological monitoring techniques relevant to brainstem surgery is presented by Dr. Moller and Dr. Morota in Chapters 13 and 14 of this book. Although classical methods of neurophysiological monitoring such as somatosensory evoked potentials (SEPs) and brainstem auditory evoked potentials (BAEPs) have been extensively used in the past, it is noteworthy that these two methods can evaluate the functional integrity of less than 20% of brainstem areas [49]. Furthermore, SEPs and BAEPs are monitoring, not mapping, techniques; therefore, they cannot help in recognizing functional landmarks in the brainstem. Displacement of a classical anatomical landmark such as the facial colliculus is a common difficulty faced by neurosurgeons. Similarly, even when anatomy is not significantly distorted by the tumor, the identification of areas overlapping the lower cranial nerve motor nuclei can be challenging. Mapping techniques are now available to intraoperatively identify motor nuclei VII, X–IX, and XII on the floor of the fourth ventricle [50, 51], and patterns of cranial nerve motor nuclei displacement have also been recognized [51]. Results from direct stimulation of the brainstem may be invaluable in choosing the safest approach to intrinsic tumors with no surface components. It should be stressed, however, that mapping techniques should never replace monitoring techniques, since only these latter allow the “real-time” assessment of the functional integrity of corticobulbar pathways and should be used any time motor cranial nerves and nuclei are at risk of surgical manipulation or injury. When dealing with lesions close to the cerebral peduncle or the ventral part of the medulla, injury to the corticospinal tracts becomes a major concern to the surgeon. Like other areas in neurosurgery, brainstem surgery has significantly benefited from the recent introduction of intraoperative motor-evoked potentials (MEPs) [53, 54]. Besides continuous recording of muscle or epidural motor-evoked potentials [55], mapping of the corticospinal tract at the level of the cerebral peduncle has also become a feasible and reliable technique [56]. Neurophysiologists and neurosurgeons involved in intraoperative neurophysiology should keep in mind that the combined use of different monitoring techniques allows the most reliable and prompt evaluation of the functional

286

Albino Bricolo and Francesco Sala

integrity of brainstem structures. So, although monitoring only BAEPs, only SEPs, or only corticospinal and corticobulbar MEPs can be misleading, the rational integration of data from all these modalities will provide the best picture of “what is going on” in the brainstem. During surgery at the cervicomedullary junction, monitoring of the lower cranial nerves (X–XII) should be added to the same monitoring techniques used for cervical spinal cord tumors (see Chapter 4). When working on a more rostral lesion, such as a pontine tumor, monitoring of corticospinal and corticobulbar pathways should be integrated with SEPs, BAEPs, and mapping of the floor of the fourth ventricle. At the midbrain level, mapping of the cerebral peduncle may be added to the battery of neurophysiological tests. Neurophysiological data should never replace a detailed knowledge of brainstem functional anatomy, careful interpretation of the MRI, and an extremely gentle and refined surgical technique. Nevertheless, monitoring and mapping techniques are becoming increasingly valuable as tools in the hands of those neurosurgeons who approach this “forbidden area” of the central nervous system. From a neurophysiological perspective, postoperative swallowing and eye movement coordination deficits remain unsolved problems. It has been documented that swallowing problems may develop in spite of intraoperatively preserved corticobulbar MEPs. Besides descending cortical control of the motor nuclei of the glossopharyngeal and vagal nerves, preservation of the afferent arch of this reflex as much as of the interneurons involved in the coordinated act of swallowing is necessary for preserving normal function. Internuclear ophthalmoplegia is also far from being solved, since monitoring and mapping of the internuclear fascicles have not yet been developed.

6 CONCLUSION A hopeless attitude when facing brainstem tumors is no longer justified, at least in subgroups of patients who could significantly benefit from surgery [57, 58]. Our personal experience with over 250 tumors confirms the increasing evidence that surgical “violation” of the brainstem is safely feasible when selection criteria are followed. In our opinion, surgery should be recommended as the “first choice” for almost all focal tumors and for that subset of diffuse gliomas protruding into the cerebellopontine angle or bulging at the external aspect in a relatively silent area. Together with improvements in neuroimaging, surgical technique, and intra- and postoperative intensive care, neurophysiological monitoring has now established its role in further improving the results of this challenging surgery.

Surgery of Brainstem Lesions

287

REFERENCES 1. Bailey, P., Buchanan, D.N., and Bucy, P.C. (1939). “Intracranial tumours of infancy and versity childhood,” pp. 188–241. University Press, Chicago. 2. Matson, D.D. (1969). “Neurosurgery of infancy and childhood,” 2nd ed. Thomas, Springfield. 3. Hoffman, H.J., Becker, L.E., and Craven, M.A. (1980). A clinically and pathologically distinct group of benign brain stem gliomas. Neurosurgery, 7, 243–248. 4. Albright, A.L., and Schlabassi, R.J. (1985). Use of the cavitron ultrasonic aspirator and evoked potentials for treatment of thalamic and brainstem tumours in children. Neurosurgery, 17, 564–568. 5. Alvisi, C., Cerisoli, M., and Maccheroni, M.E. (1985). Long-term results of surgically treated brainstem gliomas. Acta Neurochir., 76, 12–17. 6. Epstein, F., and McCleary, E.L. (1986). Intrinsic brain-stem tumors of childhood: Surgical indications. J. Neurosurg., 64, 11–15. 7. Konovalov, A., and Atieh, J. (1988). The surgical treatment of primary brain stem tumours. In “Operative neurosurgical techniques,” 2nd ed. (H.H. Schmidek, and W.H. Sweet, eds.), pp. 709–737, Grune and Stratton, New York. 8. Epstein, F., and Wisoff, J.H. (1990). Surgical management of brain stem tumors of childhood and adolescence. In “Neurosurgery Clinics of North America: The role of surgery in brain tumor management.” W.B. Saunders, Toronto. 9. Bricolo, A., Turazzi, S., Cristofori, L., and Talacchi, A. (1991). Direct surgery for brainstem tumours. Acta Neurochir. (Wien) (Suppl.), 53, 148–158. 10. Pollack, I.F., Hoffman, H.J., Humphreys, R.P., and Becker, L. (1993). The long-term outcome after surgical treatment of dorsally exophytic brainstem gliomas. J. Neurosurg., 29, 164–167. 11. Bricolo, A., and Turazzi, S. (1995). Surgery for gliomas and other mass lesions of the brainstem. In “Advances and technical standards in neurosurgery” (L. Symon, ed.), vol. 22, pp. 261–341. Springer, New York. 12. Xu, Q.W., Bao, W.M., Mao, R.L., Jiang, D.J., and Yang, G.Y. (1997). Surgical treatment of solid brain stem tumors in adults: A report of 22 cases. Surg. Neurol., 77, 30–36. 13. Welch, W.C., Kornblith, P.L., and Michalopoulos, G.K., et al. (1997). Intra-axial tumors of the cervicomedullary junction: Surgical results and long-term outcome. Pediatr. Neurosurg., 19, 12–18. 14. Choux, M., Lena, G., and Do, L. (1999). Brainstem tumors. In “Pediatric neurosurgery” (M. Choux, C. Di Rocco, A. Hockey, and M. Walker, eds.), pp. 471–491. Churchill Livingstone, London. 15. Bilaniuk, L.T., Zimmerman, R.A., Littman, P., Gallo, E., Rorke, L.B., Bruce, D.A., and Schut, L. (1980). Computed tomography of brain stem gliomas in children. Radiology, 134, 89–95. 16. Stroink, A.R., Hoffman, H.J., Hedrick, E.B., and Humphreys, R.P. (1986). Diagnosis and management of pediatric brain-stem gliomas. J. Neurosurg., 65, 745–750. 17. Epstein, F., and Farmer, J.P. (1993). Brain-stem glioma growth patterns. J. Neurosurg., 78, 408–412. 18. Ruge, J.A. (1993). Mid-brain tumours in children. Crit. Rev. Neurosurg., 3, 66–71. 19. Bricolo, A. (2001). Surgical treatment of brain stem gliomas: Approaches and technique. J. Neurosurg., 94, 392A. 20. Procaccio, F., Gottin, L., Arrighi, L., Stofella, G., and Bricolo, A. (2000). Anesthesia for brain stem surgery. Operative Techniques in Neurosurgery, 3(2), 106–108. 21. Procaccio, F., Gambin, R., Gottin, L., and Bricolo, A. (2000). Complications of brain stem surgery: Prevention and treatment. Operative Techniques in Neurosurgery, 3(2), 155–157.

288

Albino Bricolo and Francesco Sala

22. Fischbein, N.J., Prados, M.D., Wara, W., Russo, C., Edwards, M.S., and Barkovich, A.J. (1996). Radiologic classification of brain stem tumors: Correlation of magnetic resonance imaging appearance with clinical outcome. Pediatr. Neurosurg., 77, 9–23. 23. Rutka, J.T., Hoffman, H.J., and Duncan, J.A. III (1996). Astrocytomas of the posterior fossa. In “Surgical disorders of the fourth ventricle” (A.L. Cohen, ed.), pp. 189–208. Blackwell, Cambridge. 24. Rubin, G., Michowitz, S., Horev, G., Herscovici, Z., Cohen, I.J., Shuper, A., and Rappaport, Z.H. (1998). Pediatric brain stem gliomas: An update. Childs Nerv. Syst. 77, 167–173. 25. Constantini, S., and Epstein, F. (1996). Surgical indication and technical considerations in the management of benign brain stem gliomas. J. Neurooncol., 77, 193–205. 26. Epstein, F., and Constantini, S. (1996). Practical decisions in the treatment of pediatric brain stem tumors. Pediatr. Neurosurg., 77, 24–34. 27. Freeman, C.R., and Farmer, J.P. (1998). Pediatric brain stem gliomas: A review. Int. J. Radiat. Oncol. Biol. Phys., 77, 265–271. 28. Kyoshima, K., Kobayashi, S., Gibo, H., and Kuroyanagi, T. (1993). A study of safe entry zones via the floor of the fourth ventricle for brain-stem lesions. J. Neurosurg., 78, 987–993. 29. Blessing, W. (1997). “The lower brainstem and bodily homeostasis.” Oxford University Press, New York. 30. Lang, J., Ohmachi, N., and Lang, J. Sr. (1991). Anatomical landmarks of the rhomboid fossa (floor of the fourth ventricle), its length and its width. Acta Neurochir. (Wien), 113, 84–90. 31. Bricolo, A. (2000). Surgical management of intrinsic brain stem gliomas. Operative Techniques in Neurosurgery, 3, 137–154. 32. May, P.L., Blaser, S.I., Hoffman, H.J., Humphreys, R.P., Harwood-Nash, D.C. (1991). Benign intrinsic tectal “tumors” in children. J. Neurosurg., 74, 867–871. 33. Vandertop, W.P., Hoffman, H.J., Drake, J.M., Humphreys, R.P., Rutka, J.T., Amstrong, D.C., and Becker, L.E. (1992). Focal midbrain tumors in children. Neurosurgery, 31, 186–194. 34. Ono, M., Ono, M., Rhoton, A.L. Jr., Barry, M. (1984). Microsurgical anatomy of the region of the tentorial incisura. J. Neurosurg., 60, 365–399. 35. Stein, B.M. (1971). The infratentorial supracerebellar approach to pineal regions. J. Neurosurg., 35, 197–202. 36. Stein, B.M. (1988). Supracerebellar approach for pineal region neoplasms. In “Operative neurosurgical techniques: Indications, methods and results,” 2nd ed. (H.H. Schmidek, and W.H. Sweet, eds.), vol. 1, pp. 401–409. Grune & Stratton, Orlando, FL. 37. Jamieson, K.G. (1971). Excision of pineal tumors. J. Neurosurg., 35, 550–553. 38. Clark, K. (1988). The occipital transtentorial approach to the pineal region. In “Operative neurosurgical techniques: Indications, methods and results,” 2nd ed. (H.H. Schmidek, and W.H. Sweet, eds.), vol. 1., pp. 411–418. Grune & Stratton, Orlando, FL. 39. Lapras, C., Bognar, L., Turjman, F., Villanyi, E., Mottolese, C., Fischer, C., Jouvet, A., and Guyotat, J. (1994). Tectal plate gliomas: Part I. Microsurgery of the tectal plate gliomas. Acta Neurochir. (Wien), 126, 76–83. 40. Stein, B.M., Bruce, J.N., and Fetell, M.R. (1990). Surgical approaches to pineal tumors. In “Neurosurgery update: I. Diagnosis, operative technique, and neurooncology” (R.H. Wilkins, and S.S. Rengachary, eds.), pp. 389–398. McGraw-Hill, New York. 41. Yasargil, M.G., Teddy, P.J., and Roth, P. (1985). Selective amygdalo hippocampectomy: Operative anatomy and surgical technique. In “Advances and technical standards in neurosurgery” (L. Symon, J. Brihaye, B. Guidetti, F. Loew, J.D. Miller, H. Nornes, E. Pasztor, B. Pertuiset, and M.G. Yasargil, eds.), vol. 12, pp. 93–123. Springer, Wien, New York. 42. Yasargil, M.G., Mortara, R.W., and Curcic, M. (1980). Meningiomas of basal posterior cranial fossa. In “Advances and technical standards in neurosurgery” (H. Krayenbuhl, J. Briahaye, F. Loew, V. Logue, S. Mingrino, B. Pertuiset, L. Symon, H. Troupp, and M.G. Yasargil, eds.), vol. 7, pp. 3–115. Springer-Verlag, New York.

Surgery of Brainstem Lesions

289

43. Al-Mefty, O. (1990). Surgical exposure of petroclival tumors. In “Neurosurgery update: I. Diagnosis, operative technique, and neurooncology” (R.H. Wilkins, and S.S. Rengachary, eds.), pp. 409–414. McGraw-Hill, New York. 44. Spetzler, R.F., Daspit, C.P., and Pappas, C.T.E. (1992). The combined supra and infratentorial approach for lesions of the petrous and clival region: Experience with 46 cases. J. Neurosurg., 76, 588–599. 45. Ersahin, Y., Mutluer, S., Saydam, S., and Barcin, E. (1997). Cerebellar mutism: Report of two unusual cases and review of the literature. Clin. Neurol. Neurosurg., 99, 130–134. 46. Heros, R.C. (1986). Lateral suboccipital approach for vertebral and vertebrobasilar artery lesions. J. Neurosurg., 64, 559–562. 47. Spetzler, R.F., and Grahm, T.W. (1990). The far-lateral approach to the inferior clivus and the upper cervical region: Technical note. BNI Quarterly, 6(4), 35–38. 48. Zald, D.H., and Pardo, J.V. (1999). The functional neuroanatomy of voluntary swallowing. Ann. Neurol., 46, 281–286. 49 Fahlbusch, R., and Strauss, C. (1991). The surgical significance of brain stem cavernous hemangiomas. Zentralbl Neurochir., 52, 25–32. 50. Strauss, C., Romstock, J., Nimsky, C., and Fahlbusch, R. (1993). Intraoperative identification of motor areas of the rhomboid fossa using direct stimulation. J. Neurosurg., 79(3), 393–399. 51. Morota, N., Deletis, V., Epstein, F.J., Kofler, M., Abbott, R., Lee, M., and Ruskin, K. (1995). Brainstem mapping: Neurophysiological localization of motor nuclei on the floor of the fourth ventricle. Neurosurgery, 37, 922–930. 52. Morota, N., Deletis, V., Lee, M., and Epstein, F.J. (1996). Functional anatomic relationship between brain stem tumors and cranial motor nuclei. Neurosurgery, 39, 787–794. 53. Taniguchi, M., Cedzich, C., and Schramm, J. (1993). Modification of cortical stimulation for motor evoked potentials under general anesthesia: Technical description. Neurosurgery, 32, 219–226. 54. Pechstein, U., Cedzich, C., Nadstawek, J., and Schramm, J. (1996). Transcranial high-frequency repetitive electrical stimulation for recording myogenic motor evoked potentials with the patient under general anesthesia. Neurosurgery, 39, 335–344. 55. Deletis, V., and Kothbauer, K. (1998). Intraoperative neurophysiology of the corticospinal tract. In “Spinal cord monitoring” (E. Stalberg, H.S. Sharma, and Y. Olsson, eds.), pp. 421–444. Springer, Wien, New York. 56. Deletis, V., Sala, F., and Morota, N. (2000). Intraoperative neurophysiological monitoring and mapping during brain stem surgery: A modern approach. Operative Techniques in Neurosurgery, 3(2), 109–113. 57. Bricolo, A. (1999). Comment on Selvapandian, S., Rajshekhar, V., and Chandy, M.J.: Brainstem glioma: Comparative study of clinico-radiological presentation, pathology and outcome in children and adults. Acta. Neurochir. (Wien), 141, 721–727. 58. Jackowski, A. (1995). Brainstem surgery. Brit. J. Neurosurg., 9, 581–589 (editorial).

CHAPTER

13

Monitoring and Mapping the Cranial Nerves and the Brainstem AAGE R. MØLLER Callier Center for Communication Disorders, University of Texas at Dallas, Dallas, Texas

1 Introduction 2 Monitoring Cranial Motor Nerves 2.1 Monitoring the Facial Nerve 2.2 Monitoring the Motor Portion of the Trigeminal Nerve (CN V) 2.3 Monitoring Cranial Nerves That Innervate the Extraocular Muscles 2.4 Monitoring Other Cranial Motor Nerves 3 Monitoring Sensory Cranial Nerves 3.1 Monitoring the Auditory Nerve 3.2 Monitoring the Optic Nerve 3.3 How Large a Change is Allowed? 4 Monitoring the Brainstem 5 Mapping the Floor of the Fourth Ventricle 6 Monitoring That Can Guide the Surgeon During an Operation 6.1 Hemifacial Spasm 6.2 Mapping the Auditory-Vestibular Nerve 6.3 Mapping the Trigeminal Nerve 7 Conclusions References

ABSTRACT Monitoring cranial nerves in skull base operations has been proven to reduce the risk of permanent postoperative neurological deficits. The most frequently monitored cranial nerves are the auditory and the facial nerves, which are at risk in operations in the cerebellopontine angle. Lower cranial nerves (CN IX, X, XI, and XII) are at risk Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

291

292

Aage R. Møller in operations affecting the lower skull base, and monitoring the cranial motor nerves that innervate the extraocular muscles is beneficial in many skull base operations. Monitoring brainstem auditory evoked potentials (BAEPs) is useful when monitoring the overall function of the brainstem in operations where the brainstem is surgically manipulated or compressed, particularly in operations of large acoustic tumors. In some operations it has been possible to use electrophysiological methods to aid the surgeon in achieving the therapeutic goal of the operation. Examples are operations for hemifacial spasm, for vestibular neurectomy, and for whenever it is important to find safe entry to the brainstem from the floor of the fourth ventricle. Monitoring the sensory part of CN V has not found general use, but mapping that nerve can help the surgeon make a selective section of the nerve. This chapter reviews methods for intraoperative localization and monitoring of cranial motor nerves. It describes methods of monitoring the neural conduction in the auditory nerve using recordings of BAEPs and evoked potentials recorded directly from the CN VIII. Methods for recording an electromyogram (EMG) from muscles innervated by cranial motor nerves are also detailed. Intracranial electrical stimulation of the respective cranial nerves is described, as is mapping of CN V, CN VIII, and the floor of the fourth ventricle for finding safe entry points to the brainstem.

1 INTRODUCTION Intraoperative neurophysiologic monitoring of cranial nerves is now used in many kinds of neurosurgical operations. This monitoring can improve surgical outcome by several means. It can help reduce the risk of surgically induced injuries, aid in proper identification of specific neural structures, and, in a few operations, can also ensure that the therapeutic goal of an operation is achieved before the operation is ended. Whereas imaging methods, such as MRI, are restricted to detecting structural changes or coarse changes in function through measurements of oxygen consumption, neurophysiologic monitoring is an inexpensive and effective method for detecting changes in functional integrity. Methods for monitoring the facial nerve were the first to be described [1–4], followed by intraoperative monitoring of the neural conduction in the auditory nerve [5–8]. These techniques came into general use in operations in the cerebellopontine angle during the 1980s [4, 9–12]. Operations on large skull base tumors were developed during the same period [13–15], and subsequently methods for intraoperative monitoring of several other cranial nerves were introduced [7, 8, 15, 16]. The work by these and other investigators has led to techniques that optimize the use of intraoperative monitoring for a growing number of surgical procedures.

2 MONITORING CRANIAL MOTOR NERVES The general principles of monitoring cranial motor nerves include the use of a hand-held electrical stimulating electrode to probe the surgical field while electromyographic (EMG) potentials are recorded from muscles that are innervated by the respective cranial nerve. That method was developed in connection with

Monitoring Monitoring and and Mapping Mapping the the Cranial Cranial Nerves Nerves and and the the Brainstem Brainstem

293 293

monitoring the facial nerve in operations for tumors of the acoustic nerve, and it makes it possible to identify regions of a tumor where there is no (motor) nerve present. The same method can be used to determine the anatomical location of cranial motor nerves in the surgical field and to preserve their functional integrity. Continuous monitoring of muscle activity helps the surgeon preserve the postoperative function of cranial motor nerves because it can detect when surgical manipulation has caused injury to the nerve.

2.1 MONITORING THE FACIAL NERVE The introduction of intraoperative monitoring of the facial nerve was prompted by the high incidence of loss of facial function in operations for tumors of the acoustic nerve. The principle of probing the surgical field with a hand-held electrode has not changed since such monitoring was introduced in the 1960s, but the way that muscle contractions are monitored has been modified. Early in the history of facial nerve monitoring, the face of the patient was observed by an assistant [1, 2], later, contractions of face muscles were converted into electrical signals using mechanotransducers [3, 9]. Now, contractions of facial muscles are usually detected by recording an electromyogram from facial muscles [4, 8, 10–12, 16, 17]. The first goal of facial nerve monitoring in operations of large acoustic tumors is to find regions of the tumor that do not contain any portion of the facial nerve so that one can rapidly remove large portions of the tumor with low risk of permanently injuring the facial nerve. That is done most efficiently by using a monopolar hand-held stimulating electrode to probe the tumor while observing EMG potentials recorded from electrodes placed in face muscles. Short, constant-voltage impulses are the most suitable stimuli [4, 8, 12, 16]. A monopolar stimulating electrode that is insulated except at the tip will activate a nerve that is located within a sphere around the tip of the electrode. The size of that sphere depends on the stimulus intensity. The larger the stimulus intensity, the larger the diameter of the sphere, and thus the larger the volume of tissue in which nerves will be activated. It is therefore important to use adequate stimulus strength so that all nerve fibers are activated in a sufficiently large volume of tumor. If the stimulus is too weak, the facial nerve may be located closer to the stimulating electrode than the surgeon has expected and the portion of the tumor that is removed may contain parts of the facial nerve. Probing the tumor with a stimulus that is too strong may give the impression that the facial nerve is closer to the stimulating electrode than it is, and that will slow down tumor removal. When a nerve is stimulated electrically, it is important that the current through the nerve is affected as little as possible by external factors such as the

294

Aage R. Møller

electrode impedance or by shunting of current by cerebrospinal fluid (CSF). When stimulating peripheral nerves through electrodes placed on the skin, it is common to use a stimulator that delivers a constant current because the electrode impedance is likely to change. Ohm’s law reveals that constant-current stimulation maintains a constant current independent of the electrode resistance; that is, the current through the nerve will be independent of changes in electrode impedance or skin resistance. The situation is different when stimulating inside the skull, where the stimulating electrode is likely to be submerged in CSF at certain times whereas the surgical field may be relatively dry at other times. Therefore, the shunting of applied electrical current will vary as the amount of CSF in the surroundings changes, but variations in the electrode impedance is of less importance. Again, Ohm’s law reveals that shunting of current will not affect the current delivered to any volume of tissue when the stimulator delivers a constant voltage. Thus a constant-voltage stimulator will deliver an electrical current to a nerve in the operative field that varies less than that delivered by a constant-current stimulator. Consequently, constant-voltage stimulation will provide a nearly constant stimulation of a nerve that is embedded in a tumor and from time to time submerged in CSF [4, 16]. For this reason, constant-voltage stimulators are preferred over constant-current stimulators for electrical stimulation of an intracranial portion of the cranial nerve in the operative field. Whereas a stimulator can be designed to deliver a constant voltage, the stimulating electrode has a certain impedance and that makes the stimulation semiconstant. Such stimulation is less dependent on current shunting than if a constant-current stimulator is used. The second goal of facial nerve monitoring is to identify all parts of the facial nerve. A monopolar stimulating electrode can be used for that purpose, but a bipolar electrode has greater spatial selectivity and is therefore more suitable for determining of the exact location of a nerve [18]. However, a monopolar stimulating electrode is more suitable for probing a tumor for the presence of the facial nerve because the volume of tissue that a bipolar electrode stimulates depends on the electrode’s orientation. Ideally, one would thus prefer to have both a monopolar and a bipolar stimulating electrode available in operations where motor nerves are at risk of being injured. It is important to use the optimal stimulus strength for the task of finding the facial nerve, particularly if a monopolar stimulating electrode is used. The stimulus strength should be kept high in the beginning, and when a response is obtained, it should be reduced to find the exact location of the facial nerve. Watching the amplitude of the EMG response while probing the surgical field can provide information about the location of the nerve relative to the stimulating electrode and thus facilitate its rapid identification. If the amplitude of the EMG response decreases when the stimulating electrode is moved, it is a sign that the stimulating electrode moved away from the facial nerve, whereas an increase in the response amplitude

Monitoring and Mapping the Cranial Nerves and the Brainstem

295

indicates that the stimulating electrode was moved toward the facial nerve. When the response reaches supramaximal amplitude, the stimulus strengths must be decreased to find the direction of the facial nerve. The third goal of facial nerve monitoring is to detect when injuries have occurred to the facial nerve. Continuous monitoring of facial EMG in the absence of electrical stimulation is useful for that purpose. Normally, no or very little EMG activity should be present in the absence of electrical stimulation of the facial nerve, but injury to the facial nerve may generate various kinds of EMG activity. Such activity may only occur when the facial nerve is surgically manipulated, or it may last for shorter or longer periods after the surgical manipulation has ended [17, 19–21]. Short periods of EMG activity may not indicate that the facial nerve has been permanently injured. However, the likelihood that a patient will have postoperative facial weakness increases if such EMG activity occurs frequently or for long periods. Mechanical stimulation of a normal (uninjured) facial nerve normally produces little or no activation of facial muscles. Slightly injured nerves are likely to become more sensitive to mechanical stimulation, whereas severely injured nerves do not respond to mechanical stimulation. Slightly injured nerves may act as impulse generators, which explains the spontaneous facial activity that can result from surgical manipulations. However, the absence of mechanically induced EMG activity does not guarantee that injury has not occurred to the facial nerve. Electrical stimulation must therefore always be used to verify that the facial nerve has not been injured. Prolonged latency and decreased amplitude of the EMG potentials evoked by electrical stimulation of the facial nerve are signs of injury. 2.1.1 Recording Facial EMG Potentials Investigators have suggested different ways of recording EMG potentials from the facial muscles. Some have used two recording channels, one connected to electrodes placed in facial muscles of the upper face and the other recording EMG potentials from muscles of the lower face. Since the purpose of facial nerve monitoring is to avoid injury to any part of the facial nerve, it is not necessary to differentiate between different parts of the face when recording EMG potentials, and a single recording channel is therefore sufficient for monitoring EMG activity. Recording from muscles of the entire face can be accomplished by placing one of the two recording electrodes that are connected to a differential amplifier in the orbicularis oculi muscles and the other in the orbicularis oris muscles (Fig. 13.1A). (EMG potentials are best recorded using needle electrodes that are secured by adhesive tape that has micropores, e.g., 3M’s Blenderm 3M Health Care, St. Paul, MN.) The motor portion (portio minor) of the trigeminal nerve may be stimulated when probing large acoustic tumors to find the facial nerve eliciting contractions

296

Aage R. Møller

FIGURE 13.1 (A) Illustration of placement of recording electrodes for recording facial EMG and EMG from the masseter muscle. (B) Responses from facial muscles and the masseter muscles to intracranial electrical stimulation of the facial and the trigeminal nerves, respectively, from electrode placement as shown in (A). Reprinted from [8].

of the muscles of mastication (masseter and temporalis muscles). Because of the spread of EMG activity, that activity from these muscles may be picked up by the recording electrodes that are placed in facial muscles and become mistaken for contractions of the facial muscles, causing misidentification of the facial nerve. The peak latency of the response to stimulation of the trigeminal nerve is less

Monitoring and Mapping the Cranial Nerves and the Brainstem

297

than 6 ms (masseter muscle), and the peak latency of the response from the facial muscles (orbicularis oris/oculi) to intracranial stimulation of the facial nerve is longer than 8 ms. The latency of the beginning of the masseter muscle’s response to intracranial stimulation is 1.5–2.0 ms for the trigeminal nerve and 5–6 ms for the facial nerve (Fig. 13.1B). If a second recording channel is available, it is best to use it to record directly from the mastication muscles such as the masseter muscle (Fig. 13.1A). Monitoring of neural conduction in the facial nerve is facilitated by making the recorded EMG potentials audible [4, 17], but it is also important to display the EMG potentials on an oscilloscope to make it possible to measure their amplitude and latency. When the recorded EMG potentials are audible, the stimulus artifact should be suppressed from reaching the loudspeaker [4]. That feature is now available in most commercial equipment. Some commercial equipment designed for facial nerve monitoring produces a tone signal when the EMG potentials reach a certain preset amplitude. However, the direct sound of the EMG signal contains much valuable information that is lost when only the tone signals are presented through the loudspeaker. EMG potentials cannot be recorded if muscle relaxants are used. It is therefore important to use an anesthesia regimen that does not include paralyzing agents. To prevent accidental muscle paralysis from the application of relaxants, the monitoring team must maintain good communication with the anesthesiologists during an operation. It has been suggested that testing of the neural conduction of the facial nerve in operations for tumors of the cerebellopontine angle might aid in making decisions regarding the feasibility of grafting the facial nerve in the same operation. Absence of an EMG response to electrical stimulation of the facial nerve at a strength that normally elicits a maximal EMG response indicates a total conduction block peripherally to the stimulated site. However, such a block of neural conduction may be caused by different kinds of injuries to the nerve. If it is caused by neurapraxia, the nerve will recover in a short time without any intervention. If the cause of the conduction block is axonotmesis, the nerve will regrow spontaneously and function will be regained after 8–12 months without any intervention. Only if the injury is caused by neurotmesis is it appropriate to graft the facial nerve, but neurotmesis cannot be distinguished from neurapraxia or axonotmesis by electrophysiologic measures. That can only be assessed by visual observation of the nerve. Injury to the facial nerve in the beginning of the operation, to the extent that blockage of neural conduction occurs, should be avoided because it will make it impossible to monitor the facial nerve during the rest of the operation, even if the cause of the neural conduction block is neurapraxia. Continuous monitoring of the facial nerve has been previously described [22]. The method described by Colletti et al. (1997) [22] makes use of recordings of antidromic potentials from the facial nerve that are elicited by electrically

298

Aage R. Møller

stimulating the marginal mandibular nerve and that are recorded from the facial nerve intracranially. This method is still in its infancy but will probably become a clinically useful method in the future.

2.2 MONITORING THE MOTOR PORTION OF THE TRIGEMINAL NERVE (CN V) The motor portion (portio minor) of the trigeminal nerve can be monitored by recording EMG potentials from the masseter muscle, as already described (Fig. 13.1). Such monitoring helps reduce the risk of injuring the trigeminal motor nerve (portio minor, CN V) in operations for large acoustic tumors and for other skull base tumors in the regions of the brain where the trigeminal nerve is present.

2.3 MONITORING CRANIAL NERVES THAT INNERVATE THE EXTRAOCULAR MUSCLES Cranial nerves IV and VI and the motor portion of CN III can be identified in the surgical field in a similar way as that described for the facial nerve: by using a hand-held electrical stimulator and recording EMG potentials from the muscles that these nerves innervate. Regions of a tumor that do not contain any parts of these cranial nerves can be identified by probing the tumor with a handheld electrical stimulating electrode. EMG potentials from the muscles that are innervated by these cranial nerves can be recorded by placing needle electrodes percutaneously so that they come in close proximity to the respective muscles [7, 8, 15] (Fig. 13.2). The CN VI is monitored by recording from the lateral rectus muscle, and CN III can be monitored by recording from the medial rectus muscle. Recordings from the superior oblique muscle can monitor neural activity in CN IV [7, 8]. Monopolar recording electrodes are used rather than bipolar electrodes because of insufficient space to place bipolar electrodes. The reference electrodes are placed on the opposite side of the head to avoid recording EMG activity from facial muscles on the operated side of the head, where they may be activated by surgical manipulation of the facial nerve (Fig. 13.2), assuming that the facial nerve on the opposite side has not been manipulated. An alternative to invasive placement of electrodes for recording EMG potentials from the extraocular muscles was described by Sekiya and coworkers [23]. Instead of using needle electrodes, these investigators placed wire loops under the eyelids to record EMG potentials. The amplitude of the recorded EMG

Monitoring and Mapping the Cranial Nerves and the Brainstem

299

FIGURE 13.2 Placement of electrodes for recording EMG responses from the extraocular muscles (CN III, CN IV, CN VI), the facial muscles (CN VII), the masseter muscle (portio minor of CN V), the tongue (CN XII), and the neck muscles (CN XI). Also shown is electrode placement for recording BAEPs and visual evoked potentials (VEP), placement of an earphone for presenting click stimuli, and a contact lens with light-emitting diodes for stimulating the eye. Reprinted from [24].

potentials is smaller than that obtained using needle electrodes but sufficiently large for direct display on an oscilloscope.

2.4 MONITORING OTHER CRANIAL MOTOR NERVES The motor portion of lower cranial motor nerves CN IX, CN X, CN XI, and CN XII can be monitored using methods similar to those described for the facial nerve. These are mixed nerves containing sensory and autonomic fibers in addition to motor fibers. It is not practical to monitor these other populations

300

Aage R. Møller

of nerve fibers, but it is assumed that if motor fibers are not injured the sensory and autonomic fibers are also intact. Thus EMG potentials from muscles innervated by mixed nerves can be assumed to represent the entire nerve, and the lower cranial nerves can be monitored by recording EMG potentials from muscles innervated by these nerves. The motor portion of the glossopharyngeal nerve (CN IX) is monitored by placing needle electrodes in the soft palate [2, 8, 16, 24]. The vagus nerve (CN X) can be monitored by placing receding electrodes in the vocal folds [16] to record EMG potentials from laryngeal muscles. This must be done with a laryngoscope and requires expert assistance. Some tracheal tubes have electrodes built in for recording EMG potentials from larynx muscles, which can also be recorded from needle electrodes that are placed percutaneously in the larynx muscles [25]. The spinal accessory nerve (CN XI) is monitored by recording EMG potentials from neck muscles (e.g., the trapezius muscle). The hypoglossal nerve (CN XII), the main function of which is to control the tongue, is at risk in operations of tumors of the clivus. A lesion of the hypoglossal nerve is serious, and bilateral loss of that nerve is devastating. It is a very small nerve that is often totally obscured in the surgical field. The technique described for finding the facial nerve is equally effective in finding the hypoglossal nerve. Recording EMG potentials from the tongue [16, 24] (Fig. 13.2) using needle electrodes is suitable for monitoring EMG potentials evoked by intracranial stimulation of the CN XII. Although the amplitude of the recorded EMG potentials obtained from these different muscles varies, it is usually sufficient to directly observe the EMG potentials on an oscilloscope. Monitoring many cranial motor nerves requires simultaneous display of the EMG potentials recorded from one muscle for each of the nerves that are monitored. It is useful to make the EMG audible, but only one channel can be made audible at one time. There are some risks involved in stimulating motor nerves electrically. For example, stimulating the vagus nerve in this way may affect vital organs such as the heart. Since it is particularly risky to use a high stimulating rate, the stimulus rate should be kept low (2–4 pps). Electrical stimulation of motor nerves such as CN XI that innervate large muscles also involves risk because such stimulation may activate many motor units in synchrony, which does not occur when the muscles are activated voluntarily. Electrical stimulation of a motor nerve can therefore produce more muscle force than do the usual voluntary movements and can possibly injure the muscle or its tendons. Also, the natural feedback from tendon organs that normally prevents excessive forces is not functional when the motor nerve is stimulated directly. It is therefore important that caution is exercised whenever one is electrically stimulating motor nerves to large muscles.

Monitoring and Mapping the Cranial Nerves and the Brainstem

301

3 MONITORING SENSORY CRANIAL NERVES The auditory nerve is the most commonly monitored cranial sensory nerve. The sensory portion of the trigeminal nerve (portio major) or the sensory part of mixed nerves is rarely monitored. As described previously, it is the motor portions of mixed cranial nerves that are usually monitored because it is technically easier than monitoring sensory nerves.

3.1 MONITORING THE AUDITORY NERVE The auditory nerve is especially at risk during operations in the cerebellopontine angle. It may be injured from surgical manipulations or from heating from electrocoagulation, which results in changes in the neural conduction of the nerve. Several monitoring methods are currently used for detecting such changes in neural conduction in the auditory nerve. Recording of far-field auditory evoked potentials (brainstem auditory evoked potentials, BAEPs), also known as auditory brainstem responses (ABRs), is the most commonly used method for such monitoring [5, 8, 26–33]. Recording BAEPs requires little preparation, but it takes a relatively long time to obtain an interpretable record because the amplitude of these potentials is very small. In the other method used for monitoring neural conduction in the auditory nerve, the auditory evoked potentials are recorded directly from the exposed CN VIII [34–36] or from the surface of the cochlear nucleus [37, 38]. Direct recordings from CN VIII and the cochlear nucleus have the advantage of providing almost instantaneous information about changes in neural conduction in the auditory nerve, but these methods can only be used when the appropriate structure is exposed surgically, and they require that the surgeon position the recording electrode. 3.1.1 Techniques for Recording BAEPs The technique used for recording BAEPs in the operating room is similar to that used in the clinic, with some important differences. It is important to obtain an interpretable record in a short time in the operating room, and the electrical noise that typically occurs in operating rooms needs to be addressed to monitor BAEPs intraoperatively. The following can reduce the time it takes to obtain an interpretable record: 1. 2. 3. 4.

Reduce electrical interference Use optimal electrode placement Use optimal filtering Use optimal stimulus intensity and repetition rate

302

Aage R. Møller

The ratio between the amplitude of the recorded potentials and the background noise (the signal-to-noise ratio) determines how many responses must be added in order to obtain an interpretable record. Reduction of electrical interference can therefore reduce the number of responses that must be added and thus the time it takes to detect a change in neural conduction time of the auditory nerve. The ongoing electrical activity from the brain (EEG) and the activity from muscles (EMG) are also recorded by the electrodes used to record BAEPs. These biological potentials reduce the signal-to-noise ratio and thus increase the number of responses that must be added to obtain an interpretable record. Optimal electrode placement can reduce such interference. The use of optimal filtering can decrease the time it takes to obtain an interpretable record because it reduces the noise more than the signal. Using optimal stimulus intensity ensures the highest amplitude of the BAEP. 3.1.1.1 Reducing Interference Signal averaging (adding responses) is a relatively slow way to improve the signal-to-noise ratio because the ratio only improves by the square root of the number of responses that are added. This makes it imperative to reduce the electrical interference as much as possible, and that can only be done by meticulously studying the operating room and its electrical installations. The best time to do this is when no operation is being performed, such as often occurs in the late afternoon. Then it is possible to switch equipment on and off while watching its effect on evoked potentials recorded from a dummy or from a volunteer who is wired up in a similar way as a patient. An oscilloscope connected directly to the output of the physiologic amplifiers (before the signal is routed to the analog to the digital converter) is the necessary tool for such work. The use of an antenna connected to an amplifier with an oscilloscope attached is also an effective means of identifying sources of electrical interference [8]. Magnetic interference may occur from transformers in the power supply, for instance, for the light source of microscopes. This interference affects recording because it induces electrical current in the electrode leads. The source of magnetic interference may be identified in a similar way as that used for identifying static electric interference, by using a coil connected to an amplifier [8]. To reduce the interference from magnetic fields, the electrode wires should be kept as short as possible and the wires should be twisted. That will reduce the pickup of both electrical and magnetic interference. Wire loops can pick up magnetic interference and should be avoided. Interference may appear unexpectedly during monitoring. To detect such interference, the recorded potentials should be displayed directly on an oscilloscope that is connected to the output of the physiologic amplifiers. To identify the nature of the interference, which is a prerequisite for eliminating the interference, one must inspect the displayed waveform of interference. Display of the

Monitoring and Mapping the Cranial Nerves and the Brainstem

303

averaged potentials does not provide such information, and it is difficult to identify the interference by observing the waveform of the averaged potentials. If not watched directly on an oscilloscope, the interference becomes apparent only by an increase in the number of rejected responses. 3.1.1.2 Electrode Placement Recording electrodes should be placed so that they record the BAEPs with the highest amplitude while recording as little EEG and muscle EMG as possible. Recording electrodes placed on the vertex and earlobe fulfill that requirement. It may be advantageous to record BAEPs in two planes [8]. Two pairs of electrodes are used, one pair placed at the vertex and upper neck, and the other pair placed at the earlobe. The electrodes are connected to separate amplifiers, and the responses are averaged and displayed separately. The earlobe–earlobe electrodes will record peaks I–III of the BAEPs with a higher amplitude than the vertex–neck electrodes, which in turn will record peak V with a higher amplitude than the earlobe–earlobe electrodes. The BAEP recorded in patients with hearing loss usually has a lower amplitude than that recorded in individuals with normal hearing, and the waveform is often less well defined. The BAEPs in individuals with acoustic tumors have a low amplitude and prolonged latencies. It is therefore a greater challenge to obtain an interpretable BAEP in individuals with hearing loss than it is in individuals who have normal hearing preoperatively. It takes a longer time to obtain an interpretable record in patients with tumors of the acoustic nerve than in other patients, and therefore changes in neural conduction in the auditory nerve are detected later. Also, injuries to the auditory nerve affect the BAEP more than do injuries to the cochlea. As an operation proceeds, temporary or permanent injury to the auditory nerve may occur; that will further impair the quality of the BAEP and prolong the time it takes to detect a change in neural conduction in the auditory nerve. 3.1.1.3 Stimulation Clicks are the common form of stimuli used both in the clinic and in the operating room, and the stimulus intensity should be as high as possible without involving a noticeable risk for noise-induced hearing loss. Approximately 105 dB PeSPL, corresponding to approximately 65–70-dB hearing level (HL) when presented at a rate of 20 pps, is appropriate. The optimal repetition rate is 30–40 pps [8], but often a much lower rate is used. When recording BAEPs under the best circumstances, i.e., in individuals with normal hearing when the electrical interference is low, one must add at least 1000 responses to obtain an interpretable record. Using a repetition rate of 10 pps, as is common in the clinic, means that it would take 100 s to obtain an interpretable record. Using a stimulus rate of 30 pps would reduce the time to 30 s.

Monitoring and Mapping the Cranial Nerves and the Brainstem

305

without introducing phase shift (zero phase filters), which is unavoidable when using electronic filtering. Phase shifts cause changes in the latencies of responses such as BAEPs.

3.1.2 Recording Directly from the Exposed Auditory Nerve The compound action potential (CAP) that is recorded in response to clicks from the exposed intracranial portion of the CN VIII has been used for monitoring in operations for tumors of the acoustic nerve [36, 41] and in microvascular decompression operations [30, 35]. When recorded with a monopolar electrode, these potentials have much larger amplitudes than BAEPs [34, 35, 42]. The CAP recorded directly from the auditory nerve can therefore be viewed directly on an oscilloscope, or an interpretable record can be obtained after only a few responses have been added. That means that such recordings provide nearly instantaneous monitoring of neural conduction in the portion of the auditory nerve that is located peripherally to the recording electrode. The waveform of the recorded CAP is similar to that recorded from a long nerve using a monopolar recording electrode, i.e., an initial positive deflection followed by a large negative peak, which is followed by a smaller positive potential (Fig. 13.4). The waveform of the CAP is affected by hearing loss [43] and is typically more complex in individuals who are hard of hearing than in individuals with normal hearing. Stretching the auditory nerve typically reduces its conduction velocity which is reflected by an increased latency of the negative peak of the CAP. Conduction block in some nerve fibers results in decreased amplitude of the negative peak of the CAP, while the amplitude of the initial positive peak increases (Fig. 13.4). A total conduction block results in the total absence of a negative peak, and the CAP consists of a single positive deflection. This is known as a “cut end” potential. Recordings directly from the exposed CN VIII can only be done when the auditory nerve is exposed, and the recording electrode must be kept in place on the nerve during the operation [8]. CN VIII is more sensitive to mechanical manipulation than other cranial nerves in the cerebellopontine angle because it is covered with central myelin (oligodendrocytes) over its entire intracranial course and has no perineurium. Thus CN VIII’s glial transition zone (ObersteinerRedlich zone) is located in the internal auditory meatus, just inside the porus acusticus, whereas in other cranial nerves of the cerebellopontine angle the transition zone between peripheral and central myelin is located close to the brainstem, leaving only a few millimeters of the nerve covered with central myelin [44]. In addition, there are indications from animal experiments that manipulations of CN VIII intracranially that cause the nerve to be pulled in a

306

Aage R. Møller

FIGURE 13.4 Examples of changes in the waveform of CAP recorded from the proximal portion of the intracranial part of the auditory nerve during a microvascular decompression operation. (A) Normal waveform. (B) Beginning coagulation showing how spread of heat to the nerve gradually impairs the response. Reprinted from Møller, A.R. (1988). “Evoked potentials in intraoperative monitoring.” Williams & Wilkins, Baltimore, MD.

medial direction may cause injury to the nerve deep within the internal auditory meatus. The recording electrode must therefore be soft and exert minimal pressure on the nerve [8]. We have used a multistrand, Teflon-insulated silver wire to which uninsulated tips of cotton wick are sutured [8, 34, 35]. One major problem is that the electrode may easily be dislocated during surgical manipulations. Many of the problems associated with recording directly from the exposed eighth nerve can be eliminated, however, by instead recording from the surface of the cochlear nucleus [38].

3.1.3 Recording Directly from the Cochlear Nucleus The cochlear nucleus is the floor of the lateral recess of the fourth ventricle, and recordings from the cochlear nucleus can therefore be made by placing an electrode in the lateral recess of the fourth ventricle [35, 38, 45] (Fig. 13.5A). A cotton wick electrode similar to the one used to record from the exposed

FIGURE 13.5 (A) Illustration of placement of the electrode for recording evoked potentials from the cochlear nucleus. (B) Waveform of recorded potentials. Top tracings: BAEP; middle tracings: CAP recorded directly from the intracranial portion of CN VIII; bottom tracings: response recorded from the surface of the cochlear nucleus as shown in (A). The stimuli for all recordings were clicks, 105 dB PeSPL. Solid lines: rarefaction clicks; dashed line: condensation clicks. Reprinted from [38]. 307

308

Aage R. Møller

eighth nerve is suitable for such recordings. The electrode may be pushed through the foramen of Luschka to reach the floor of the lateral recess. Recording from the cochlear nucleus yields potentials of similar amplitude to those recorded from the exposed eighth nerve (Fig. 13.5B), but the recording electrode is out of the surgical field and is not disturbed by surgical manipulations in operations in the cerebellopontine angle. Useful recordings can also be obtained by placing the electrode on the CN X where it enters the brainstem.

3.2 MONITORING THE OPTIC NERVE Recordings of visual evoked potentials from electrodes placed on the scalp are used in operations where the optic nerve is at risk. It is also technically possible to record CAPs from the optic nerve during operations in which it is exposed [7]. The choice of stimulus to evoke the visual response in the operating room is limited to brief light flashes. Such stimuli evoke a clearly recordable response, but the changes in such visual evoked potentials to flash stimulation that may occur intraoperatively are not directly related to injuries of the optic nerve or the optic tract [46]. This is probably because the visual system is not specifically responding to the time pattern of light stimuli. This fact has been recognized clinically, and therefore diagnostic tests using visual evoked potentials follow a checkerboard pattern of changing contrasts as stimulus. Such a stimulus cannot be used in the anesthetized patient because it requires a focused pattern on the patient’s retina. Recently, it has been suggested that the responses to high-intensity light flashes are better indicators of injuries to the optic nerve than responses to less intense light flashes [47].

3.3 HOW LARGE A CHANGE IS ALLOWED? The question of how large a change in the latency of peak III or peak V of the BAEP is “safe” and does not require any action from the surgeon has been debated extensively. Some authors have stated that there is no need to take action even when large changes in the BAEP occur [27], while others [8] have expressed the opinion that any small change in the latency implies a certain (small) risk of change in hearing postoperatively. Measurements of changes in neural conduction of the auditory nerve only provide information about the likelihood of acquiring a permanent hearing impairment. That means that a certain increase in latency of recorded evoked

Monitoring and Mapping the Cranial Nerves and the Brainstem

309

potentials means that there is a likelihood of postoperative hearing loss of a certain magnitude. Therefore, the allowable change in latency depends on how large a number of individuals can be allowed to acquire a permanent hearing loss of a certain magnitude. Information about the relationship between latency shift and the likelihood of permanent hearing loss is difficult to obtain because a large number of patients must be studied. It is not known if there is a safe range of latency changes in which there is no risk of permanent hearing loss. The time during which the latency is increased most likely also plays a role, and it is likely that a concomitant decrease in amplitude influences the outcome [28]. As a further complication, the meaning of “safe” varies among investigators, and so does the definition of “no noticeable hearing loss” (i.e., the amount of allowable hearing loss). Our experience from monitoring a large number of patients who were operated on for disorders of cranial nerves V, VII, and VIII (due to vascular compression) leads us to believe that there is no threshold. This means that the risk of permanent postoperative hearing loss increases from very small shifts in latency of the BAEP. Thus intervention by the surgeon is justified when any detectable change (i.e., changes that are larger than those that occur when no operation is performed) in the latency occurs. This emphasizes the importance of reversing even small changes in intraoperative recordings. It is noteworthy that the quality of life was reduced in a large fraction of patients who had no postoperative neurological deficits that were detectable using objective testing [48]. This would suggest that detrimental changes did occur that were beyond the scope of the objective tests used.

4 MONITORING THE BRAINSTEM Intraoperative monitoring of BAEPs can provide information about the general condition of the brainstem, and BAEPs are generally monitored in operations where the brainstem is being manipulated [7, 8, 15, 49]. The changes in the BAEP that occurred as a result of surgical manipulations of the brainstem were found to be more consistent and occurred, on average, earlier than changes in cardiovascular signs could be detected [50]. Monitoring of BAEPs is therefore valuable in operations for large tumors of the acoustic nerve and other operations where the brainstem may be manipulated. When used in operations on tumors of the acoustic nerve, the BAEP is elicited from the ear opposite to the tumor. It is useful to consider the neural generators of the different components of the BAEP [8, 51] when interpreting the recordings because that can provide information about the anatomical location of the changes.

310

Aage R. Møller

5 MAPPING THE FLOOR OF THE FOURTH VENTRICLE Intraoperative mapping of the floor of the fourth ventricle can help the surgeon find safe entry points to the brainstem for removal of intrinsic tumors or vascular malformations. Mapping makes use of recordings of EMG responses from muscles that are innervated by cranial nerves VII and XII while the floor of the fourth ventricle is being probed with a bipolar electrical stimulating electrode [52, 53] (Fig. 13.6). This technique has mainly been able to identify the CN VII, but the location of CN XII can also be determined. For the location of CN XII, recordings are made from the genioglossal muscle or from the tongue. Other locations can also be identified using such mapping. Thus the location of the abducent and hypoglossal

FIGURE 13.6 Recordings of EMG potentials from muscles innervated by CN VII and CN XII when bipolar electrical stimulation was done at different locations on the floor of the fourth ventricle. (A) Bipolar stimulation of the right facial colliculus and recordings from the genioglossal (CN XII) and orbicularis muscles (CN VII) on both sides. The stimulus current was 0.5 µA. (B) Bipolar stimulation at the left trigone of the hypoglossal (CN XII) nerve. (C) Bipolar stimulation of the left facial colliculus in the same patient who had a left peripheral facial paresis. The stimulus strength required to evoke a response was 2 mA because of the facial paresis. Reprinted from [53].

Monitoring and Mapping the Cranial Nerves and the Brainstem

311

nerves, the motor nucleus of the fifth cranial nerve, and the ambiguous nucleus can be identified using these techniques [54] (see Chapter 15 for details).

6 MONITORING THAT CAN GUIDE THE SURGEON DURING AN OPERATION The use of electrophysiologic methods in the operating room can increase the success rate and reduce failures in a few operations.

6.1 HEMIFACIAL SPASM Neurophysiologic monitoring can increase the success rate of operations for hemifacial spasm (HFS) [55]. HFS is a rare disorder (incidence of approximately 0.8 per 100,000 in the United States [56]) that can be cured by microvascular decompression (MVD) of the intracranial portion of the facial nerve. In this operation, a blood vessel is moved off the facial nerve and a soft implant is placed between the vessel and the nerve. The basis for the neurophysiologic monitoring is the finding that abnormal muscle responses that seem to be characteristic of HFS [57] disappear when the vessel that is associated with the spasm is moved off the facial nerve [8, 58] (Fig. 13.7). The abnormal muscle

FIGURE 13.7 EMG recordings from a patient undergoing MVD to relieve HFS. Each graph shows consecutive recordings (beginning at top) from the mentalis muscle in response to electrical stimulation of the zygomatic branch of the facial nerve. As indicated, the recordings at the bottom of the left column were made when the vessel was lifted off the nerve. Reprinted from [58].

312

Aage R. Møller

FIGURE 13.8 Electrode placement for monitoring abnormal muscle response in a patient undergoing MVD to relieve HFS. Reprinted from [8].

response can be demonstrated by recording EMG potentials from a face muscle while electrically stimulating a branch of the facial nerve different from the branch that innervates the muscle from which recording is made (Fig. 13.8). By monitoring that response during an MVD operation for HFS, it is possible to identify the vessel that causes the spasm [55]. That vessel may be an artery, a vein, or a very small artery. Such monitoring has increased the success rate of the MVD operation, which, under ideal circumstances, can be more than 95%

Monitoring and Mapping the Cranial Nerves and the Brainstem

313

successful [55]. In some cases more than one vessel is in close contact with the facial nerve. Before such intraoperative neurophysiological monitoring is introduced [55], some patients have to be reoperated on because vessels involved in causing the symptoms of HFS have remained in contact with the facial nerve.

6.2 MAPPING THE AUDITORY-VESTIBULAR NERVE The auditory and the vestibular portions of the eighth cranial nerve are located close together in the posterior fossa, and it is necessary to identify the cleavage plane between these two portions of the auditory vestibular nerve to perform selective sectioning of the vestibular nerve. Although this can often be done using visual criteria, electrophysiological methods are of great help in this task [59]. Direct recording of sound-evoked CAPs from the surface of the eighth cranial nerve using a bipolar recording electrode is a useful method for identifying the border between the auditory and the vestibular nerves. Bipolar recordings have a greater spatial selectivity than monopolar recordings and are the method of choice for this task. A hand-held bipolar electrode made of Teflon-insulated silver wires that are twisted and cut with an intertip distance of 1–2 mm [60] is suitable for probing the nerve (Fig. 13.9).

FIGURE 13.9

Bipolar electrode used to record from the exposed CN VIII. Reprinted from [60].

314

Aage R. Møller

FIGURE 13.10 Bipolar recordings from the intracranial portion of CN VIII. The stimuli were clicks with an intensity that was 25 dB above the threshold for BAEPs. Reprinted from [59].

When CN VIII is probed by a bipolar recording electrode, the response to sound stimulation is only present when the recording electrode is placed on the auditory nerve (Fig. 13.10). It has been shown that a low stimulus intensity is important for achieving the best spatial selectivity [59]. Recording from the exposed auditory-vestibular nerve using such a bipolar electrode requires caution because it can injure the nerve.

6.3 MAPPING THE TRIGEMINAL NERVE Mapping the intracranial portion of the sensory portion of the trigeminal nerve is useful in operations for selective posterior fossa trigeminal rhizotomy to treat trigeminal neuralgia. Similar techniques as already described for mapping CN VIII can be used to localize fibers of individual subdivisions of the trigeminal nerve as well as to ensure that the selective trigeminal rhizotomy is complete [61]. The intracranial portion of the trigeminal sensory nerve is mapped by stimulating the exposed trigeminal nerve electrically in the posterior fossa while recording the CAP from the individual branches of the trigeminal nerve where they emerge from their respective foramina. The recording electrodes are needle electrodes (Grass-type E2 subdermal needles; Grass Instruments, Astro-Med, Inc., West Warwick, RI) placed in each of the supraorbital, infraorbital, and mental foramina. In some such operations, the respective reference electrodes were placed subdermally 5 mm adjacent to the recording electrodes. The antidromic CAPs were observed while the trigeminal nerve was stimulated

Monitoring and Mapping the Cranial Nerves and the Brainstem

315

FIGURE 13.11 Compound action potentials recorded from the trigeminal nerve by monopolar electrodes placed in each of the three foramina. Reprinted from [61]. A: supraorbital foramen; B: infraorbital foramen; C: mental foramen.

electrically with a bipolar electrode similar to the one used for mapping the eighth cranial nerve (Fig. 13.11).

7 CONCLUSIONS Many studies have shown that intraoperative monitoring of the integrity of the facial nerve in operations for acoustic tumors is helpful in reducing the risk of loss or impairment of facial function. Similar techniques can be used to monitor other cranial motor nerves, and such monitoring is an effective tool for reducing postoperative neurological deficits. Intraoperative neurophysiologic monitoring also helps the surgeon in other ways, such as to confirm the anatomical location of specific structures, and it gives the surgeon a feeling of security that can make operating less stressful. Mapping the floor of the fourth ventricle can benefit operations of the brainstem by allowing the surgeon to find safe entry to the brainstem. Mapping the eighth cranial nerve is important for hearing preservation in operations for sectioning the vestibular portion in the cerebellopontine angle. Mapping the trigeminal nerve can facilitate selective sectioning of branches of the trigeminal nerve.

316

Aage R. Møller

REFERENCES 1. Rand, R.W., and Kurze, T.L. (1965). Facial nerve preservation by posterior fossa transmeatal microdissection in total removal of acoustic tumours. J. Neurol. Neurosurg. Psychiatry, 28, 311–316. 2. Hilger, J. (1964). Facial nerve stimulator. Trans. Am. Acad. Ophthalmol. Otolaryngol., 68, 74–76. 3. Sugita, K., and Kobayashi, S. (1982). Technical and instrumental improvements in the surgical treatment of acoustic neurinomas. J. Neurosurg., 57, 747–752. 4. Møller, A.R., and Jannetta, P.J. (1984). Preservation of facial function during removal of acoustic neuromas: Use of monopolar constant-voltage stimulation and EMG. J. Neurosurg., 61, 757–760. 5. Grundy, B.L. (1983). Intraoperative monitoring of sensory evoked potentials. Anesthesiology, 58, 72–87. 6. Grundy, B.L. (1985). Evoked potentials monitoring. In “Monitoring in anesthesia and critical care medicine” (C.D. Blitt, ed.), pp. 345–411. Churchill-Livingstone, New York. 7. Møller, A.R. (1987). Electrophysiological monitoring of cranial nerves in operations in the skull base. In “Tumors of the cranial base: Diagnosis and treatment” (L.N. Sekhar, and V.L. Schramm Jr., eds.), Chap. 7, pp. 123–132. Futura Publishing, Mt. Kisco, New York. 8. Møller, A.R. (1995). “Intraoperative neurophysiologic monitoring.” Harwood Academic Publishers, Luxembourg. 9. Silverstein, H., Smouha, E., and Jones, R. (1988). Routine identification of facial nerve using electrical stimulation during otological and neurotological surgery. Laryngoscope, 98, 726–730. 10. Prass, R.L., Kinney, S.E., Hardy, R.W., Hahn, J.F., and Lueders, H. (1987). Acoustic (loudspeaker) facial EMG monitoring: Part II. Use of evoked EMG activity during acoustic neuroma resection. Otolaryngol. Head Neck Surg., 97, 541–551. 11. Harner, S.G., Daube, J.R., Ebersold, M.J., and Beatty, C.W. (1987). Improved preservation of facial nerve functions with use of electrical monitoring during removal of acoustic neuromas. Mayo Clin. Proc., 62, 92–102. 12. Yingling, C.D., and Gardi, J.N. (1992). Intraoperative monitoring of facial and cochlear nerves during acoustic neuroma surgery. Otolaryngol. Clin. N. Am., 25, 413–448. 13. Sekhar, L.N., and Schramm, V.L. Jr. (Eds.) (1987). “Tumors of the cranial base: Diagnosis and treatment.” Futura Publishing, Mt. Kisco, New York. 14. Al-Mefty, O., Holoubi, A., Rifai, A., and Fox, J.L. (1985). Microsurgical removal of suprasellar meningiomas. Neurosurgery, 16(3), 364–372. 15. Sekhar, L.N., and Møller, A.R. (1986). Operative management of tumors involving the cavernous sinus. J. Neurosurg., 64, 879–889. 16. Yingling, C.D. (1994). Intraoperative monitoring in skull base surgery. In “Neurotology” (R.K. Jackler, and D.E. Brackmann, eds.), pp. 967–1002. Mosby Year Book, Inc., St. Louis. 17. Prass, R.L., and Lueders, H. (1986). Acoustic (loudspeaker) facial electromyographic monitoring. Part I. Neurosurgery, 19, 392–400. 18. Babin, R.M., Jai, H.R., and McCabe, B.F. (1982). Bipolar localization of the facial nerve in the internal auditory canal. In “Disorders of the facial nerve: Anatomy, diagnosis, and management” (M.D. Graham, and W.F. House, eds.), pp. 3–5. Raven Press, New York. 19. Prass, R.L., and Lueders, H. (1985). Constant-current versus constant-voltage stimulation. J. Neurosurg., 62, 622–623. 20. Daube, J.R., and Harper, C.M. (1989). Surgical monitoring of cranial and peripheral nerves. In “Neuromonitoring in surgery” (J.E. Desmedt, ed.), pp. 115–138. Elsevier Science Publishers, Amsterdam. 21. Daube, J.R. (1991). Intraoperative monitoring of cranial motor nerves. In “Intraoperative neurophysiologic monitoring in neurosurgery” ( J. Schramm, and A.R. Møller, eds.), pp. 246–267. Springer-Verlag, Heidelberg, Germany.

Monitoring and Mapping the Cranial Nerves and the Brainstem

317

22. Colletti, V., Fiorino, F., Policante, Z., and Bruni L. (1997). Intraoperative monitoring of facial nerve antidromic potentials during acoustic neuroma surgery. Acta Otolaryngol., 117, 663– 669. 23. Sekiya, T., Hatayama, T., Iwabuchi, T., and Maeda, S. (1992). A ring electrode to record extraocular muscle activities during skull base surgery. Acta Neurochir., (Wien), 117, 66–69. 24. Møller, A.R. (1990). Intraoperative monitoring of evoked potentials: An update. In “Neurosurgery update I: Diagnosis, operative technique, and neuro-oncology” (R.H. Wilkins, and S.S. Rengachary, eds.), Chap. 14, pp. 169–176. McGraw-Hill, New York. 25. Stechison, M.T. (1995). Vagus nerve monitoring: A comparison of percutaneous versus vocal fold electrode recording. Am. J. Otol., 16, 703–706. 26. Fischer, C. (1989). Brainstem auditory evoked potential (BAEP) monitoring in posterior fossa surgery. In “Neuromonitoring in surgery” ( J.E. Desmedt, ed.), pp. 191–218. Elsevier Science Publishers, Amsterdam. 27. Friedman, W.A., Kaplan, B.J., Gravenstein, D., and Rhoton, A.L. (1985). Intraoperative brainstem auditory evoked potentials during posterior fossa microvascular decompression. J. Neurosurg., 62, 552–557. 28. Hatayama, T., and Møller, A.R. (1998). Correlation between latency and amplitude of peak V in brainstem auditory evoked potentials: Intraoperative recordings in microvascular decompression operations. Acta Neurochir. (Wien), 140, 681–687. 29. Linden, R.D., Tator, C.H., Benedict, C., Charles, D., Mraz, V., and Bell, I. (1988). Electrophysiological monitoring during acoustic neuroma and other posterior fossa surgery. Le Journal Canadien des Sciences Neurologiques, 15, 73–81. 30. Møller, A.R., and Møller, M.B. (1989). Does intraoperative monitoring of auditory evoked potentials reduce incidence of hearing loss as a complication of microvascular decompression of cranial nerves? Neurosurgery, 24, 257–263. 31. Radtke, R.A., Erwin, W., and Wilkins, R.H. (1989). Intraoperative brainstem auditory evoked potentials: Significant decrease in post-operative morbidity. Neurology, 39, 187–191. 32. Raudzens, P.A. (1982). Intraoperative monitoring of evoked potentials. Ann. N.Y. Acad. Sci., 388, 308–326. 33. Watanabe, E., Schramm, J., Strauss, C., and Fahlbusch, R. (1989). Neurophysiologic monitoring in posterior fossa surgery: II. BAEP waves I and V and preservation of hearing. Acta Neurochir. (Wien), 98, 118–128. 34. Møller, A.R., and Jannetta, P.J. (1981). Compound action potentials recorded intracranially from the auditory nerve in man. Exp. Neurol. 74, 862–874. 35. Møller, A.R., and Jannetta, P.J. (1983). Monitoring auditory functions during cranial nerve microvascular decompression operations by direct recording from the eighth nerve. J. Neurosurg., 59, 493–499. 36. Silverstein, H., Norrell, H., and Hyman, S. (1984). Simultaneous use of CO2 laser with continuous monitoring of eighth cranial nerve action potential during acoustic neuroma surgery. Otolaryngol. Head Neck Surg., 92, 80–84. 37. Møller, A.R., and Jannetta, P.J. (1983). Auditory evoked potentials recorded from the cochlear nucleus and its vicinity in man. J. Neurosurg., 59, 1013–1018. 38. Møller, A.R., Jho, H.D., and Jannetta, P.J. (1994). Preservation of hearing in operations on acoustic tumors: An alternative to recording BAEP. Neurosurgery, 34, 688–693. 39. Møller, A.R. (1988). Use of zero-phase digital filters to enhance brainstem auditory evoked potentials (BAEPs). Electroencephalogr. Clin. Neurophysiol., 71, 226–232. 40. Doyle, D.J., and Hyde, M.L. (1981). Analogue and digital filtering of auditory brainstem responses. Scand. Audiol. (Stockholm), 10, 1–89. 41. Colletti, V., Bricolo, A., Fiorino, F.G., and Bruni, L. (1994). Changes in directly recorded cochlear nerve compound action potentials during acoustic tumor surgery. Skull Base Surg., 4, 1–9.

318

Aage R. Møller

42. Hashimoto, I., Ishiyama, Y., Yoshimoto, T., and Nemoto, S. (1981). Brainstem auditory evoked potentials recorded directly from human brainstem and thalamus. Brain, 104, 841–859. 43. Møller, A.R., and Jho, H.D. (1991). Effect of high-frequency hearing loss on compound action potentials recorded from the intracranial portion of the human eighth nerve. Hear. Res., 55, 9–23. 44. Lang, J. (1982). Über Bau, Länge und Gefässbeziehungen der “zentralen” und “peripheren” Strecken der intrazisternalen Hirnnerven. Zentralbl. Neurochirur., 43, 217–258. 45. Kuroki, A., and Møller, A.R. (1995). Microsurgical anatomy around the foramen of Luschka with reference to intraoperative recording of auditory evoked potentials from the cochlear nuclei. J. Neurosurg., 82, 933–939. 46. Cedzich, C., Schramm, J., Mengedoht, C.F., and Fahlbusch, R. (1988). Factors that limit the use of flash visual evoked potentials for surgical monitoring. Electroencephalogr. Clin. Neurophysiol., 71, 142–145. 47. Pratt, H., Martin, W.H., Bleich, N., Zaaroor, M., and Schacham, S.E. (1994). A high-intensity, goggle-mounted flash stimulator for short-latency visual evoked potentials. Electroencephalogr. Clin. Neurophysiol., 92, 469–472. 48. Nikolopoulos, T.P., Johnson, I., and O’Donoghue, G.M. (1998). Quality of life after acoustic neuroma surgery. Laryngoscope, 108(9), 1382–1385. 49. Kalmanchey, R., Avila, A., and Symon, L. (1986). The use of brainstem auditory evoked potentials during posterior fossa surgery as a monitor of brainstem function. Acta Neurochir. (Wien), 82, 128–136. 50. Angelo, R., and Møller, A.R. (1996). Contralateral evoked brainstem auditory potentials as an indicator of intraoperative brainstem manipulation in cerebellopontine angle tumors. Neurol. Res., 18, 528–540. 51. Møller, A.R. (1994). Neural generators of auditory evoked potentials. In “Principles and applications in auditory evoked potentials” (J.T. Jacobson, ed.), Chap. 2., pp. 23–46. Allyn & Bacon, Boston. 52. Strauss, C., Lutjen-Drecoll, E., and Fahlbusch, R. (1997). Pericollicular surgical approaches to the rhomboid fossa: Part I. Anatomical basis. J. Neurosurg., 87(6), 893–899. 53. Strauss, C., Romstock, J., Nimsky, C., and Fahlbusch, R. (1993). Intraoperative identification of motor areas or the rhomboid fossa using direct stimulation. J. Neurosurg., 79, 393–399. 54. Strauss, C., Romstock, J., and Fahlbusch, R. (1999). Pericollicular approaches to the rhomboid fossa: Part II. Neurophysiological basis. J. Neurosurg., 91, 768–775. 55. Møller, A.R., and Jannetta, P.J. (1987). Monitoring facial EMG responses during microvascular decompression operations for hemifacial spasm. J. Neurosurg., 66, 681–685. 56. Auger, R.G., and Whisnant, J.P. (1990). Hemifacial spasm in Rochester and Olmsted County, Minnesota, 1960 to 1984. Arch. Neurol., 47, 1233–1234. 57. Esslen, E. (1957). Der spasmus facialis—Eine Parabioseerscheinung. Dtsch. Z. Nervenh., l76, l49–l72. 58. Møller, A.R., and Jannetta, P.J. (1985). Microvascular decompression in hemifacial spasm: Intraoperative electrophysiological observations. Neurosurgery, 16, 612–618. 59. Rosenberg, S.I., Martin, W.H., Pratt, H., Schwegler, J.W., and Silverstein, H. (1993). Bipolar cochlear nerve recording technique: A preliminary report. Am. J. Otol., 14, 362–368. 60. Møller, A.R., Colletti, V., and Fiorino, F.G. (1994). Neural conduction velocity of the human auditory nerve: Bipolar recordings from the exposed intracranial portion of the eighth nerve during vestibular nerve section. Electroenceph. Clin. Neurophysiol., 92, 316–320. 61. Stechison, M.T., Møller, A.R., and Lovely, T.J. (1996). Intraoperative mapping of the trigeminal nerve root: Technique and application in the surgical management of facial pain. Neurosurgery, 38, 76–82.

CHAPTER

14

Brainstem Mapping NOBUHITO MOROTA Department of Neurosurgery, National Children’s Medical Center, National Center for Child Health and Development, Tokyo, Japan

VEDRAN DELETIS Division of Intraoperative Neurophysiology, Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York

FRED J. EPSTEIN Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York

1 Introduction 1.1 What is Brainstem Mapping? 2 Methodology of BSM 2.1 Anesthesia Regimen 3 Results of BSM 4 Surgical Implications of BSM 5 Clinical Limitations of BSM 6 Representative Case of BSM 7 Clinical Application of BSM 8 Summary References ABSTRACT Brainstem mapping is a neurophysiological method of locating the cranial nerve motor nuclei (CMN) on the floor of the fourth ventricle. The motor nuclei of the cranial nerves are usually located in the vicinity of specific anatomical landmarks on the floor of the fourth ventricle. Because of the distorting effects of a tumor on the local anatomy, these landmarks are not evident in most patients. Even in patients without a tumor, specific anatomical landmarks are often not visible. Different points of the surgically exposed floor of the fourth ventricle were electrically stimulated by the surgeon using a hand-held probe. Electromyographic responses were recorded with electrodes inserted in the muscles of the head that are innervated by cranial motor nerves. This technique was found to be useful for locating cranial nerve motor nuclei before tumor resection and enabled the surgeon to avoid damaging the nuclei when entering the brainstem. Furthermore, intraoperative neurophysiological localization of the CMN showed specific patterns of displacement by brainstem tumors. Pontine tumors displaced the CMN of the nerve VII around the edge of the tumor, and medullary tumors ventrally displaced the low CMN. Understanding the patterns of

Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

319

320

Nobuhito Morota, Vedran Deletis, and Fred J. Epstein CMN displacement can help in establishing a surgical plan that minimizes the risk of damaging the CMN and allows for safer surgery for brainstem tumors.

1 INTRODUCTION Brainstem mapping (BSM) is emerging as an addition to the neurophysiological armamentarium available during surgery for brainstem lesions that often places cranial nerve motor nuclei (CMN) at risk for injury [1–4]. Historically, auditory brainstem responses (ABRs) and somatosensory evoked potentials (SEPs) were among the most commonly used neurophysiological techniques for monitoring during brainstem surgery [5]. However, responses obtained from ABRs and SEPs give only indirect information regarding the functional integrity of the CMN. These two methods cover only 20% of the brainstem area [6] and thus play a limited role in the effort to preserve the functional integrity of the CMN during brainstem surgery.

1.1 WHAT IS BRAINSTEM MAPPING? BSM is a neurophysiological technique that helps to localize the CMN on the surgically exposed floor of the fourth ventricle (Fig. 14.1). During surgery for a brainstem tumor, the surgeon must know the location of the CMN to avoid damaging this area. Up to now, surgeons were guided by anatomical landmarks [7] in the operating field of the fourth ventricle when trying to prevent injury to the CMN and other structures. Using these landmarks, safe entry zones to the brainstem through the floor of the fourth ventricle have been determined that include the supra- and infrafacial triangles. Their anatomical importance has been discussed elsewhere [8, 9]. The problems the surgeon encounters using anatomical landmarks as guidelines for safe entry to the brainstem are twofold: first, normal anatomy is usually distorted by tumor, and second, anatomical landmarks on the floor of the fourth ventricle (facial colliculus and striae medullares) can not be easily recognized in some patients with brainstem tumors. In our previous study [2], we showed that the facial colliculi could be visualized in 3 out of 12 patients. This study was done in 7 patients with medullary tumors that did not influence the anatomy of the facial colliculi and in 5 patients with a pontine tumor. In 9 out of 14 patients, the striae medullares could be visualized (9 patients with medullary tumors and 5 patients with pontine tumors). BSM was used to localize the CMN and to provide the surgeon with anatomical guidance, even when anatomical landmarks were not visible on floor of the fourth ventricle [2, 3, 10, 11]. Thus, it provided guidance as to

321

FIGURE 14.1 Mapping of the brainstem cranial nerve motor nuclei. Upper left, drawing of the exposed floor of the fourth ventricle with the surgeon’s hand-held stimulating probe in view. Upper middle, sites of insertion of wire hook electrodes for recording muscle responses. Far Upper right, compound muscle action potentials recorded from the orbicularis oculi and oris muscles after stimulation of the upper and lower facial nuclei (upper two traces) and from the pharyngeal wall and tongue muscles after stimulation of the motor nuclei of cranial nerves IX, X, and XII (lower two traces). Lower left, photograph obtained from the operating microscope showing the hand-held stimulating probe placed on the floor of the fourth ventricle (F). A: aqueduct. Modified from [2].

322

Nobuhito Morota, Vedran Deletis, and Fred J. Epstein

where to make incisions on the floor of fourth ventricle and/or when to stop tumor resection at the bottom of the tumor cavity.

2 METHODOLOGY OF BSM After endotracheal intubation, EMG wire recording electrodes were inserted into the appropriate muscles under direct laryngoscopy. They were inserted bilaterally into the posterior pharyngeal wall to record responses from the CMN IX and X and into the lateral aspect of the tongue to record the response of CMN XII. Wire electrodes were inserted into the obicularis oris and oculi muscles to record responses from CMN VII (Fig. 14.1). Electrodes can be safely placed in the extraocular muscles for mapping CMN III and CMN VI [12]. In this case, pairs of electrodes were used with an interelectrode distance of 5 mm. The recording electrodes were custom-made Teflon-coated wire electrodes (type 316SS 3T; Medwire, Mount Vernon, NY) with 2 mm bare, hooked tips. The electrode end was encased in a 27-gauge needle, which allowed it to be manipulated with a long needle holder. After insertion into the muscle, the needle was withdrawn, leaving the tip of the electrode in place. After surgical exposure of the floor of the fourth ventricle, electrical stimulation was delivered using a hand-held monopolar stimulation probe with a modified tip (Xomed, #82-25100; Medtronic Xomed Surgical Products, Inc., Jacksonville, FL). A corkscrew electrode (Nicolet, Madison, WI), placed at FZ (10–20 International EEG System) was used as reference (anode). Stimuli of 0.2 ms duration were delivered at a stimulating rate of 4 Hz for a few seconds. An initial stimulus intensity of 1.5–2.0 mA was used. Later, when muscle responses were obtained, the intensity was reduced to determine the threshold (usually 0.3–2.0 mA). Mapping the upper and lower CMN took about 5 min. The EMG responses were recorded within an epoch length of 20 ms, were amplified 10,000 times, and were filtered between 50 and 2133 Hz.

2.1 ANESTHESIA REGIMEN Anesthetics used for general anesthesia have little or no effect on the lower motor neurons. Since we were stimulating the lower motor neurons (motor nuclei of cranial nerves or their intramedullary roots), any type of anesthesia that didn’t include a long-lasting relaxant was compatible with BSM. In the first series of 18 patients, we used fentanyl, thiopental, and a shortacting, nondepolarizing relaxant. Anesthesia was maintained with fentanyl infusion, 0.4% isoflurane, and a 70% nitrous oxide and oxygen mixture. Once BSM was established as a routine intraoperative procedure, our standard anesthesia regimen included propofol, fentanyl, a nitrous oxide and oxygen mixture, and a short-acting muscle relaxant (prior to intubation only).

Brainstem Mapping

323

3 RESULTS OF BSM In our study published previously [2], we neurophysiologically localized the CMN of the facial nerve prior to tumor resection in 8 out of 10 patients who underwent surgery for brainstem tumors. The facial nucleus was typically located close to the facial colliculus. CMN XII was often localized close to or at the obex, and CMN IX or X was localized at the area rostral and lateral to the obex. Large pontine tumors displaced the facial nuclei laterally (at the edge of the tumor). Contrary to the mapping of the facial nuclei, localization of the CMN of IX or X and XII in patients with medullary tumors was not always possible before tumor resection. The tumor mass displaced those nuclei ventrally, preventing electrical current from reaching the nuclei during stimulation. Our data, collected in 20 patients with brainstem tumors and cervicomedullary spinal cord tumors, showed displacement of the CMN in consistent patterns, depending on the location of the tumor (Fig. 14.2) [3]. Those patterns are as follows. In pontine tumors, the CMN of the facial nerves were displaced around the tumor edge on the floor of the fourth ventricle. An upper pontine tumor displaced the CMN of the facial nerve caudally, and a lower pontine tumor displaced them rostrally. If the tumor was predominantly occupying one side, both facial nuclei were displaced to the opposite side. We did not find displacement of the lower CMN in patients with pontine tumors and no displacement of the facial nuclei in patients with medullary tumors. In patients with medullary tumors, one or more lower CMN were displaced ventrally to the tumor. Therefore, mapping of lower CMN in those cases was only possible at the end of tumor resection at the bottom of the tumor cavity, and unsuccessful BSM of low CMN conveyed important information suggesting that the CMN could be located ventral to the tumor. In patients with cervicomedullary spinal cord tumors, the lower CMN were displaced rostrally, sometimes involving the rostral displacement of the striae medullares [3]. In patients with small tumors, no displacement of the lower CMN was found. The varying modality of tumor compression to the CMN based on tumor location seems to be derived from variations in tumor biology. It is well known that most pontine tumors are malignant and grow invasively in an intrinsic fashion, thus displacing the CMN VII bilaterally at the edge of the exposed tumor. On the other hand, the medullary tumor is likely to be low grade in malignancy and tends to grow in a more exophytic fashion. As the tumor extends into the fourth ventricle, the lower CMN can be pushed ventrally beneath it. The cervicomedullary junction spinal cord tumor behaves much like a medullary tumor. Since low-grade tumors are common at the cervicomedullary junction, they displace the lower CMN rostrally during their growth into the fourth ventricle [14]. Electrical stimulation of the floor of the fourth ventricle with a low stimulation rate and low current intensity is safe [2].

324

Nobuhito Morota, Vedran Deletis, and Fred J. Epstein

FIGURE 14.2 Typical patterns of cranial nerve motor nuclei displacement by brainstem tumors in different locations. Upper and lower pontine tumors typically push the facial nuclei around the edge of the tumor, suggesting that precise localization of the facial nuclei before tumor resection is necessary to avoid their damage during surgery. Medullary tumors typically grow more exophytically and compress the lower cranial nerve motor nuclei ventrally; these nuclei may be located on the ventral edge of the tumor cavity. Because of the interposed tumor, in these cases mapping before tumor resection usually does not allow identification of cranial nerve IX, X, and XII motor nuclei. However, responses could be obtained close to the end of the tumor resection, when most of the tumoral tissue between the stimulating probe and the motor nuclei had been removed. At this point, repeated mapping is recommended because the risk of damaging motor nuclei is significantly higher than at the beginning of tumor debulking. Cervicomedullary junction spinal cord tumors simply push the lower cranial nerve motor nuclei rostrally when extending into the fourth ventricle. Reprinted from [3].

In a series of 18 patients, we have had 1 patient with transient, premature ventricle contraction during mapping of CN IX or X, and other patients have had a mild increase in blood pressure. Both of these side-effects disappeared after the stimulation ceased. Other authors [4, 15] have not reported any side effects or complications during BSM. In an experimental animal study, Suzuki et al. [16] reported that transient hypotension and bradycardia developed in adult dogs when the floor of the fourth ventricle was electrically stimulated with greater than 2 mA intensity. Respiratory arrest was observed with 3 mA. The administration of atropine sulfate prior to stimulation decreased the intensity of these side-effects.

Brainstem Mapping

325

4 SURGICAL IMPLICATIONS OF BSM The surgical implications of BSM are important [3]. The pontine tumor often pushes the CMN VII around the tumor edge, suggesting that precise localization of the CMN VII before tumor resection is mandated to avoid direct damage by compression or incision on the floor of the fourth ventricle. The surgeon is required to perform an elaborate dissection at the tumor edge on the floor of the fourth ventricle where the CMN VII are in close proximity. If the tumor is in the upper pons and the CMN VII are displaced caudally, incision on the floor of the fourth ventricle should be directed rostrally. If the tumor is in the lower pons and the CMN VII are displaced rostrally, incision should be directed caudally (Fig. 14.3). As mentioned previously, the medullary tumor tends to grow exophytically and compress some of the lower CMN ventrally. Care should be taken near the end of tumor resection if the CMN were unmapped before tumor resection. The unmapped lower CMN may be present at the bottom of the tumor cavity. The tumor resection should be approached cautiously at the tumor base to preserve the lower CMN. Leaving a thin layer of tumor at the bottom of the tumor cavity is recommended, considering the fact that most medullary tumors are low grade (Fig. 14.4).

FIGURE 14.3 Schematic of the different displacements of the facial nuclei on the floor of the fourth ventricle in upper and lower pontine tumors. Arrows indicate the direction of the initial incision on the floor of the fourth ventricle.

326

Nobuhito Morota, Vedran Deletis, and Fred J. Epstein

FIGURE 14.4 Schematic of displacement of the lower CMN in patients with medullary tumors. Because the lower CMN are displaced ventrally, extreme care should be taken to prevent damaging them during resection of the tumor base.

FIGURE 14.5 Schematic of rostral displacement of the lower CMN by a cervicomedullary junction spinal cord tumor. If the rostral end of the tumor is directly approached through the floor of the fourth ventricle, the lower CMN are placed at high risk for injury.

Large cervicomedullary junction spinal cord tumors often grow and extend into the fourth ventricle. Displacing the lower CMN rostrally, the rostral end of the tumor may extend beneath the lower part of the floor of the fourth ventricle and lift it up slightly. In such a case, the surgeon needs to undermine the lower part of the floor of the fourth ventricle when the tumor resection reaches the rostral end. If the rostral end of the tumor is directly approached through the floor of the fourth ventricle, the lower CMN are placed at high risk for injury (Fig. 14.5).

Brainstem Mapping

327

5 CLINICAL LIMITATIONS OF BSM BSM has proven valuable for preventing direct damage to the CMN during surgery of brainstem tumors and other pathologies in and around the brainstem [2, 3, 6]. However, it does have some limitations. First of all, this is a mapping technique, not a monitoring technique. BSM is a procedure to localize the CMN, and as such is performed intermittently during tumor resection to confirm the location of the CMN. It is not a procedure to continuously monitor the integrity of the CMN throughout the tumor resection. Any damage induced during tumor resection cannot be prevented by this method. Second, BSM cannot detect damage to the corticonuclear tract originating from the motor cortex and ending on the CMN. Therefore, supranuclear motor paralysis may occur even if the integrity of lower motor neurons is preserved (from the CMN to the cranial muscle). Third, the response recorded at the end of the tumor resection does not mean that the CMN were preserved. Theoretically, the response could be obtainable from the cut end of the intramedullary roots of the cranial nerve that were mapped. However, this has not been our experience. Furthermore, it seems very difficult to preserve other neural elements involved in the swallowing and coughing reflex, despite the use of BSM. This is understandable, since BSM can only test the efferent part of those reflex arcs. Damage to the intramedullary afferent roots on neural connections between nuclei is undetectable using this technique. Difficulties with swallowing or coughing (with an absence of the gag reflex) may develop even though EMG responses from the lower CMN have been maintained after BSM at the end of tumor resection. Damage to the corticobulbar tract is a possible condition undetected by BSM. However, damage is less likely to happen because of the anatomy of the corticobulbar tract. According to Krieg’s anatomy of the corticobulbar tract [17], there are seven major branches of the corticobulbar tract. All of them run in the ventrodorsal direction toward the CMN (Fig. 14.6). Thus, as long as the brainstem tumor is approached through the floor of the fourth ventricle, the corticobulbar tract remains relatively isolated from possible trauma. Furthermore, because of the complexity of CMN innervation by the corticobulbar tract, it would be unlikely that all branches of the corticobulbar tract ending up on the CMN would be damaged at once. Although there is a possibility that some branches of the corticobulbar tract could be damaged during surgery, their function will be taken over by other corticobulbar tracts later. Permanent damage to the corticobulbar tract by brainstem surgery seems exceptional as long as the brainstem lesion is approached from the fourth ventricle.

328

Nobuhito Morota, Vedran Deletis, and Fred J. Epstein

FIGURE 14.6 Schematic of three major branches of the corticobulbar tract (midbrain, pontine, and medullary; drawn according to Krieg’s anatomy book), with their divisions (a–g). Note that the corticobulbar tract innervates the cranial nerve motor nuclei from the ventral side. Therefore, it is less likely that they will be damaged during surgery for brainstem tumors.

6 REPRESENTATIVE CASE OF BSM The advantages and limitations of using BSM during surgery on a brainstem lesion when anatomy is distorted are described in the following case (patient 1). A 54-year-old woman had repeated episodes of brainstem hemorrhage resulting in mild left hemiparesis, sensory disturbance, dysarthria, and bilateral facial weakness (House & Brackmann grade 2). MRI revealed a large brainstem hematoma caused by a brainstem cavernous angioma (Fig. 14.7). During surgery, swelling of the right upper half of the floor of the fourth ventricle was so prominent that the left upper half of the floor was not visible (Fig. 14.8). No anatomical landmarks were observed on the upper half of the floor of the fourth ventricle. The striae medullares were displaced caudally. An abnormal vessel was observed on the swelled right upper part of the floor of the fourth ventricle. A part of the hematoma was exposed rostral to the abnormal vessel. By using BSM, we located the right facial colliculus caudal to the abnormal vessel (Fig. 14.9). An incision on the floor of the fourth ventricle was directed caudally from the exposed hematoma and rostral to the abnormal vessel (Fig. 14.10). The hematoma and cavernous angioma were extensively removed (Fig. 14.11). BSM following the hematoma removal was consistent with responses obtained prior to hematoma removal. Postoperatively, the patient showed no neurological

FIGURE 14.7

Preoperative MRI of patient 1. A large hematoma caused by a cavernous angioma occupies the pons predominantly on the right side.

329

330 FIGURE 14.8 Left, intraoperative photograph of patient 1. The right pontine part of the floor of the ventricle bulges onto the left side. Abnormal vessels cross the right part of the pontine floor of the fourth ventricle. Part of the hematoma is exposed rostrally to the abnormal vessel. Right, intraoperative photograph of BSM using a hand-held stimulating probe.

Brainstem Mapping

331

FIGURE 14.9 Results of BSM from patient 1. The largest response from the right orbicularis oculi and oris was recorded from the area caudal to the abnormal vessel (B). The amplitude of the response is smaller rostral (A) and caudal (C) to the point where the maximum response was obtained.

deterioration except mild ataxia. Facial weakness disappeared soon after the surgery (Fig. 14.12). This example illustrates how BSM can show the surgeon where to perform the initial incision and prevent injury to the CMN of the facial nerves.

7 CLINICAL APPLICATION OF BSM The methodology of BSM can be applied to locate CMN other than CMN VII, IX or X, and XII. Mapping of the oculomotor nuclei (CMN III) and trochlear nuclei (CMN IV) can be achieved in surgery on a midbrain lesion if it is approached dorsally. The same is true in surgery for quadrigeminal plate, tectal, and pineal region tumors [12].

332

Nobuhito Morota, Vedran Deletis, and Fred J. Epstein

FIGURE 14.10 Left, schematic of results of BSM from patient 1. Right, localized area of the right facial nucleus is superimposed on the floor of the ventricle. Surrounding grey zone indicates the area where smaller responses from the facial muscles were obtained by BSM. No response was obtained from the area rostral to the abnormal vessel. Based on the result of BSM, an incision was made on the hematoma cranial to the abnormal vessel. Rt FN: right facial nucleus.

8 SUMMARY Recent advances in neurophysiological techniques have influenced several challenging neurosurgical procedures [18]. Despite the advent of microneurosurgery and MRI, which clearly depicts the relationship between the surgical lesion and the surrounding anatomy [8, 19–22], surgery on the brainstem still holds the possibility of surgical morbidity, mainly because of the distorting effects that a tumor can have on the local anatomy. The neurophysiological technique of BSM can be used to recognize the functional anatomy of the CMN when this anatomy has been distorted. Mapping critical neural structures during surgery has become an important concept in intraoperative neurophysiology. Brainstem lesions in the midbrain, pons, and medulla can now be surgically resected using BSM as a technique to locate the CMN on the floor of the fourth ventricle. Using BSM, the surgeon is guided to enter the brainstem through a “silent area,” thus avoiding direct damage to the CMN.

333

FIGURE 14.11 Left, CT scan taken immediately after surgery demonstrated the path of hematoma removal. Right, MRI taken 2 weeks after surgery showed a small residual hematoma with sufficient brainstem decompression.

334

FIGURE 14.12 weakness.

Nobuhito Morota, Vedran Deletis, and Fred J. Epstein

Postoperative facial appearance of patient 1 showed complete recovery from facial

REFERENCES 1. Katsuta, T., Morioka, T., Fujii, K., and Fukui, M. (1993). Physiological localization of the facial colliculus during direct surgery on an intrinsic brain stem lesion. Neurosurgery, 32, 861–863. 2. Morota, N., Deletis, V., Epstein, F.J., Kofler, M., Abbott, R., Lee, M., and Ruskin, K. (1995). Brain stem mapping: Neurophysiological localization of motor nuclei on the floor of the fourth ventricle. Neurosurgery, 37, 922–930. 3. Morota, N., Deletis, V., Lee, M., and Epstein, F.J (1996). Functional anatomic relationship between brain stem tumors and cranial motor nuclei. Neurosurgery, 39, 787–794. 4. Strauss, C., Romstock, J., Nimsky, C., and Fahlbusch, R. (1993). Intraoperative identification of motor areas of the rhomboid fossa using direct stimulation. J. Neurosurg., 79, 393–399. 5. Albright, A.L., and Sclabassi, R.J. (1985). Use of the Cavitron ultrasonic surgical aspirator and evoked potentials for the treatment of thalamic and brain stem tumors in children. Neurosurgery, 17, 564–568. 6. Fahlbusch, R., and Strauss, C. (1991). The surgical significance of brainstem cavernous hemangiomas. Zentrabl. Neurochir., 52, 25–32. 7. Lang, J., Ohmachi, N., and Lang Sen, J. (1991). Anatomical landmarks of the rhomboid fossa (floor of the 4th ventricle), its length and its width. Acta Neurochir. (Wien), 113, 84–90. 8. Bricolo, A., and Turazzi, S. (1995). Surgery for gliomas and other mass lesions of the brainstem. In “Advances in Technical Standards Neurosurgery, 22” (L. Symon, ed.), pp. 261–341. Springer, New York. 9. Kyoshima, K., Kobayashi, S., Gibo, H., and Kuroyanagi, T. (1993). A study of safe entry zones via the floor of the fourth ventricle for brain-stem lesions: Report of three cases. J. Neurosurg., 78, 987–993. 10. Grabb, P.A., Albright, A.L., Sclabassi, R.J., and Pollack, I.F. (1997). Continuous intraoperative electromyographic monitoring of cranial nerves during resection of fourth ventricular tumors in children. J. Neurosurg., 86, 1–4.

Brainstem Mapping

335

11. Strauss, C., Romstock, J., and Fahlbusch, R. (1999). Pericollicular approaches to the rhomboid fossa: Part II. Neurophysiological basis. J. Neurosurg., 91, 768–775. 12. Sekiya, T., Hatayama, T., Shimamura, N., and Suzuki, S. (2000). Intraoperative electrophysiological monitoring of oculomotor nuclei and their intramedullary tracts during midbrain tumor surgery. Neurosurgery, 47, 1170–1177. 13. Epstein F.J., and Farmer, J.P. (1993). Brain-stem glioma growth patterns. J. Neurosurg., 78, 408–412. 14. Epstein, F., and Constantini, S. (1996). Practical decisions in the treatment of pediatric brain stem tumors. Pediatr. Neurosurg., 24, 24–34. 15. Eisner, W., Schmid, U.D., Reulen, H.J., Oeckler, R., Olteanu-Nerbe, V., Call, C., and Kothbauer, K. (1995). The mapping and continuous monitoring of the intrinsic motor nuclei during brain stem surgery. Neurosurgery, 37, 255–265. 16. Suzuki, K., Matsumoto, M., Ohta, M., Sasaki, T., and Kodama, N. (1997). Experimental study for identification of the facial colliculus using electromyography and antidromic evoked potentials. Neurosurgery, 41, 1130–1136. 17. Krieg, W.J.S. (1957). “Brain mechanisms in diachrome,” Brain Books, 2nd edition, pp. 287–290. Bloomington, IL. 18. Deletis, V. (1993). Intraoperative monitoring of the functional integrity of the motor pathways. In “Electrical and magnetic stimulation of the brain and spinal cord.” (O. Devinsky, A. Beric, M. Dogali, eds.), pp. 201–214. Raven Press, New York. 19. Abbott, R., Shiminski-Maher, T., and Epstein, F.J. (1996). Intrinsic tumor of the medulla: Predicting outcome after surgery. Pediatr. Neurosurg., 25, 41–44. 20. Epstein, F., and McCleary, E.L. (1986). Intrinsic brain stem tumors of childhood: Surgical indications. J. Neurosurg., 64, 11–15. 21. Pierre-Kahn, A., Hirsch, J.F., Vinchon, M., Payan, C., Sainte-Rose, C., Renier, D., LelouchTubiana, A., and Fermanian, J. (1993). Surgical management of brain stem tumors in children: Results and statistical analysis of 75 cases. J. Neurosurg., 79, 845–852. 22. Pollack, I.F., Hoffman, H.J., Humphreys, R.P., and Becker, L. (1993). The long-term outcome after surgical treatment of dorsally exophytic brain stem gliomas. J. Neurosurg., 78, 859–863.

CHAPTER

15

Intraoperative Neurophysiological Mapping and Monitoring for Supratentorial Procedures GEORG NEULOH AND JOHANNES SCHRAMM Department of Neurosurgery, University of Bonn, Germany

1 Introduction 2 Somatosensory Evoked Potentials 2.1 Technique 2.2 Principles of Clinical Application 3 Intraoperative and Perioperative Neurophysiological Functional Mapping 3.1 SEP Phase Reversal 3.2 Extraoperative Mapping with Grid or MultipleStrip Electrodes for Motor, Sensory, and Speech Localization 3.3 Intraoperative Electrical Stimulation of Motor Cortex and White Matter 3.4 MEP Mapping 4 MEP Monitoring 4.1 Technique 4.2 Principles of Clinical Application 5 Safety and Anesthesia 5.1 Safety 5.2 Anesthesia 6 Value and Limitations from a Clinical Perspective 6.1 Aneurysms 6.2 Aneurysms—Summary Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

339

340

Georg Neuloh and Johannes Schramm

6.3 AVMs—Summary 6.4 Centrally Located Tumors 6.5 Centrally Located Tumors—Summary Acknowledgments References

ABSTRACT The goal of delineation through mapping and monitoring of eloquent cortical areas and subcortical pathways is to achieve a more radical cytoreduction while still preserving unimpaired function. The use of different neurophysiological methods in the operation room has a long history, but monitoring with evoked potential technology is a development of the last 25 years; in particular, motor tract monitoring is a very recent development. In this chapter we summarize the experience from our center in intraoperative neurophysiological monitoring of more than 1500 patients over the last 10 years. This includes patients operated on for cerebral aneurysms, arteriovenous malformations, central and insular tumors, and deep-seated cerebral lesions, as well as patients who have undergone surgery for epilepsy. The methods employed for monitoring were somatosensory evoked potential phase reversal; motor pathway monitoring; mapping of motor, sensory, and speech areas; and zones of ictal and interictal spiking in patients with chronic epilepsy. The experience in monitoring supratentorial lesions includes over 60 vascular malformations, 170 tumors, and over 200 patients with epilepsy. We discuss the interpretation of results, safety and anesthesia issues, and the limitations and drawbacks of the method. Intraoperative neurophysiological monitoring of the functional integrity of the cortical sensory and motor areas and pathways is noninvasive and does not interfere with the ongoing surgical procedure. It gives immediate feedback about functional impairment of the monitored structure. In some cases, this feedback provides the opportunity to rectify this impairment.

1 INTRODUCTION From the end of the 1970s to the middle of the 1980s, intraoperative evoked potential monitoring developed from an experimental technique into a reliable tool used during different types of cranial and spinal surgeries involving acoustic, motor, and sensory evoked potential modalities. The intraoperative use of neurophysiological methods, in particular evoked potentials as a monitoring tool, was independently conceived by a Japanese orthopedic surgeon [1] and a group of American neurosurgeons [2]. The idea to use evoked potential monitoring was first applied to spine surgery and later on to posterior fossa surgery. The first papers about evoked potentials from our group concerned experimental spinal cord injury, but clinical applications soon followed [3–8]. Some of the early diagnostic applications for spinal lesions thought to be important were soon overshadowed by the advent of MRI technology, which is more proficient for pinpointing intramedullary lesions (e.g., in multiple sclerosis [9]). The early years were devoted to working out the technical principles and

Intraoperative Neurophysiological Mapping and Monitoring

341

optimizing anesthesia, electrode montages, and criteria of abnormality. One of the first meetings on intraoperative monitoring was an informal one in St. Louis in 1977. This was followed shortly afterwards by the First International Spinal Cord Monitoring Symposium held in Tokyo and organized by T. Tamaki [10]. Some of the historical views discussed are quite interesting in retrospect, such as the definition of criteria for abnormal intraoperative changes in evoked potentials. Although these criteria were arbitrarily defined, they were based on a lengthy experience of observing SEPs vary and change during surgery, sometimes being lost and sometimes recovering again. These criteria were tentatively based on experiences that were not scientifically justified or on statistically examined series that used varying parameters. Twenty years later, much of what was intuitively defined in those days has remained valid: 1. Changes in amplitudes are much more variable than latency changes. 2. Amplitude changes have to be much more significant than latency changes to be relevant. 3. Waveform alterations are not really important, with a few exceptions. When the first papers on intraoperative monitoring were presented, skeptics asked questions such as the following: What are you going to do when the potential is lost? Are you going to leave the tumor in? This last question has since been answered. A surgeon’s reaction may involve checking retractors, vessels, clips, and blood pressure; stopping coagulation; applying papaverine; stopping dissection; and changing the area and style of dissection. In short, monitoring has become an educational tool. Evoked potential monitoring during surgery, first with somatosensory evoked potentials (SEPs) and now also with motor evoked potentials (MEPs), has become an accepted technique for monitoring the integrity of motor and sensory pathways within the spinal cord, the medulla, the brainstem, and the brain. For many years MEPs could not be reliably assessed intraoperatively. Many of the so-called false-positive and false-negative results of the past, when SEP monitoring was the only available methodology (particular during aneurysm surgery) [11–16], can now be judged using both modalities. It has become quite clear that monitoring using SEPs, MEPs, and auditory evoked potentials (AEPs) improves safety standards for intracranial surgery. Certain lesions, such as large multilobulated aneurysms in the anterior cerebral circulation, should not be operated on without neurophysiological monitoring. Monitoring reassures the surgeon when MEPs and SEPs remain stable with a clipped internal carotid artery (ICA), and it clarifies necessary action if potentials are lost with a clip on the ICA. There are many other instances in which monitoring has proven extremely useful. In our hospital service, certain procedures, such as treatment of central and insular gliomas, acoustic neurinomas with preserved hearing, complex aneurysms, and central arteriovenous malformations (AVMs), would

342

Georg Neuloh and Johannes Schramm

not be done without monitoring. This chapter reviews our current experience using intraoperative neuromonitoring techniques during surgeries for supratentorial lesions.

2 SOMATOSENSORY EVOKED POTENTIALS SEP recording has become a well-established method in intraoperative neuromonitoring. Extensive literature exists on SEP monitoring during spinal cord and posterior fossa surgeries for the following services: neurosurgery [4, 5, 7, 8, 17–26], orthopedic surgery [27, 28], cardiovascular surgery [29–31], and interventional neuroradiology [32]. SEP monitoring with supratentorial procedures is mainly applied for functional mapping purposes with lesions adjacent to the central region (see Section 3) and for continuous monitoring purposes in aneurysm surgery [11, 15, 16, 33–40]. The latter application has become standard in a growing number of neurosurgical centers over the last decade. It is based on the well-established observation of a close relationship between cortical cerebral blood flow and SEP changes [41–44]. With aneurysm dissection and clipping, SEPs may reflect intentional or inadvertent vessel occlusion or compromise. In addition, clinical experience and experimental work have shown that SEPs are sensitive to local factors such as pressure or heat, as well as to systemic parameters like blood pressure, body temperature, or metabolic changes. They may therefore also be applied with other lesions, such as vascular malformations or tumors. This chapter sketches SEP stimulation and recording techniques, safety and anesthesia requirements, and principles and specific aspects of their clinical application as appears necessary for understanding SEP monitoring in supratentorial surgical procedures.

2.1 TECHNIQUE For intraoperative monitoring purposes, SEPs are elicited and recorded according to common international standards previously published [7, 15–17, 45, 46]. A comprehensive review of the underlying functional neuroanatomy and physiology as well as detailed considerations regarding SEP stimulation and recording in the neurosurgical setting can be found in Intraoperative Neurophysiologic Monitoring by A. Møller [26]. Here we will present the basic technical and physiological information. The basic principle behind SEP recording is stimulation of a peripheral nerve with recording of the resulting evoked activity from the central nervous system, which typically arises from nuclei of the somatosensory pathways, the cortex, and the axonal fibers [26]. In neuromonitoring of supratentorial lesions, cortical activity must obviously be included in the recording

Intraoperative Neurophysiological Mapping and Monitoring

343

scheme, although subcortically evoked SEP components may give valuable additional information. 2.1.1 Stimulation SEPs are typically elicited by electrical bipolar stimulation of a peripheral nerve [46]. The most commonly stimulated nerves are the median nerve, if sensory modalities and cortical regions representing the upper extremities are to be monitored, and the posterior tibial nerve for the lower extremities. SEP recording from these nerves is easiest to obtain and provides the most stable results [23]. Under special circumstances, such as amputated limbs or preexisting nerve damage, other (e.g., ulnar or peroneal) nerves may be chosen for stimulation. The cathode is placed proximally to avoid anode block of the ascending action potential. Monopolar pulses of 200 to 500 µs duration are delivered as constant-current or constant-voltage stimuli through sterile subcutaneous needle electrodes or surface electrodes spaced about 2 cm apart. We prefer needle electrodes for the posterior tibial nerve and surface electrodes for median nerve stimulation. Ground electrodes (saline-soaked cloth wrapped around the stimulated limb) are placed close to the stimulation site to keep the stimulating artifact low. With simultaneous stimulation of several limbs, which is often the case in intraoperative monitoring, a single ground must be used. The actual stimulation parameters (such as frequency and intensity) need to be adjusted to intraoperative requirements, particularly regarding to the effects of general anesthesia. 2.1.1.1 Stimulus Intensity As with awake patients, the stimulus intensity is set slightly above the motor threshold in nonrelaxed patients. Otherwise, the lowest intensity at which the maximum amplitude of early SEP components can be evoked is chosen. It may be considerably higher in anesthetized than in awake patients, but 20–25 mA with median nerve stimulation and 25–30 mA with posterior tibial nerve stimulation are rarely exceeded (Table 15.1). Unnecessarily strong stimulation must be avoided, both for safety considerations and to maintain minimal muscle twitching in nonrelaxed patients, which can be disturbing during subtle microsurgical supratentorial procedures. 2.1.1.2 Stimulus Duration A stimulus duration of 300 µs is usually sufficient, representing a compromise between safety considerations and the effort to obtain a stable signal. In cases of difficult SEP recording (e.g., with sensory polyneuropathy), a stimulus duration of 500 µs or more may be adequate.

344

Georg Neuloh and Johannes Schramm TABLE 15.1 SEP Stimulation and Recording Parameters for Intraoperative Monitoring of SEPs

Pulse width Intensity

Median nerve

Tibial nerve

300 (200–500) µs

300 (200–500) µs

max. 20–25 mA

max. 25–30 mA

Stimulation rate

5.3 Hz

5.3 Hz

Recording sites

C3′, C4′ (C5) vs Fpz

Cz (lower cervical) vs Fpz

50 ms post stim

100 ms post stim

Sweep length Averages

150–250

150–250

Band pass

30–3000 (500) Hz

30–3000 (500) Hz

2.1.1.3 Stimulus Frequency SEP recording requires the averaging of 200–500 (or more) trials. Thus stimulation must be repeated at a relatively fast frequency. Stimulation frequencies of 1–10 Hz are commonly used in awake patients. With intraoperative monitoring, a fast SEP update is obviously important in order to provide real-time feedback and to allow for early reaction to relevant changes. No significant differences in latency and amplitude occur with stimulation frequencies between 1 and 3 Hz, and a stimulus frequency below 5 Hz has been recommended for routine recording of early cortical SEPs [46]. In our experience, a stimulation frequency around 5 Hz (e.g., 5.3 Hz) represents a good trade-off between a fast update and high-amplitude, stable potentials. Stimulation at exactly 5 Hz may lead to interference with line-frequency (50 or 60 Hz) artifacts, which would be particularly disturbing since 50 or 60 Hz notch filters should not be applied with a recording of early SEP components (see Section 2.1.2). In general, an even-numbered Hz value, or harmonic parts of 50 or 60 Hz, should be avoided to minimize artifacts. If SEPs are recorded bilaterally, stimulation may be performed by alternating between sides to preserve update speed. With such a stimulus rate, and particularly in combination with a current anesthetic regimen, an SEP update up to three times per minute or even more frequently may be obtained, which comes close to real-time monitoring [47] (Fig. 15.1).

2.1.2 Recording 2.1.2.1 Electrode Montage For intracranial procedures, the recording electrode montage should include a cervical and a cortical site to differentiate between central and peripheral causes of SEP latency changes. This setup also helps in differentiating technical problems

Intraoperative Neurophysiological Mapping and Monitoring

345

FIGURE 15.1 Fast median nerve SEP updates were recorded in a patient under propofol anesthesia. Twenty updates were obtained within 7 minutes, allowing for nearly real-time monitoring.

from relevant changes that may occur intraoperatively. The cervical electrode should be placed at about C5–C7 spine level, whereas the scalp electrodes for recording of cortical potentials are placed at parietal positions 2–3 cm posterior to the C3–C4 (upper extremity) or the Cz (lower extremity) sites of the International 10–20 System. Recording is performed in a referential way with the reference electrode placed at Fpz, although other montages are also used (e.g., with extracranial references) [26, 46, 48]. Since intraoperative recording of spinal components of SEPs may often prove difficult (in particular with lower extremity SEPs) and acute peripheral changes are unlikely to occur, scalp electrodes alone can be considered sufficient in many cases (see Section 2.2.1). The skin incision for craniotomy may necessitate some adjustment of electrode placement, although this problem is more common with MEP recording, since SEPs are usually monitored in cases requiring standard frontotemporal craniotomy, e.g., in cerebral aneurysms. Intraoperative SEP recording is best performed through sterile subdermal needle electrodes. It is recommended to place at least two electrodes at each recording site—one as a reserve—since electrodes may easily slip during the disinfection and scrub procedure. Impedance measured at 100–500 Hz should be kept below 5–10 kΩ.

346

Georg Neuloh and Johannes Schramm

2.1.2.2 Filtering The signal is amplified and (digitally) filtered. A band-pass of about 10–3000 Hz is standard in common outpatient SEP recording [46], and equivalent filtering parameters are usually applied with intraoperative recording. However, the signal components most relevant for intraoperative purposes, namely, the spinal and the early cortical responses, do not have much power above 100– 200 Hz. Therefore, a considerably tighter low-pass filter can usually be applied without losing much relevant information. High-frequency artifacts are common in the operating room setting and may require aggressive filtering in order to enhance the signal-to-noise ratio. According to our experience, low-pass filtering as low as 500 Hz or even lower can be used without significant compromise of early cortical SEP recording. On the other hand, very tight high-pass filtering should not be used and will rarely be necessary above 30 Hz. Line-frequency notch filters, available with most current recording systems, should not be applied or applied only in cases with a severe artifact problem, since the signal of interest has its main power within this frequency range. Ringing, an artifact caused by the filter, may also occur. 2.1.2.3 Sweep Length and Averaging A poststimulus time window of 50–100 ms with median or tibial nerve stimulation is usually sufficient to register the relevant early SEP components. The number of sweeps averaged depends on the signal-to-noise ratio, which can be quite low intraoperatively because of the usually high noise level and the often low SEP amplitude under general anesthesia. With an adequate anesthetic regimen and with properly chosen stimulation and recording parameters, averaging of 100–150 sweeps or even fewer is sufficient in many cases [47]. In intraoperative situations requiring a particularly fast update, the averaging procedure can be stopped manually as soon as an assessable potential can be recognized. The amplitude gain through this method (the averaging procedure leads to a slight amplitude reduction due to the natural latency jitter of the signal) will be insignificant in most cases. 2.1.2.4 Signal A large variety of positive and negative deflections can be recorded from numerous spinal and cranial recording sites with peripheral nerve stimulation [26, 49]. However, for intraoperative monitoring purposes, only a few SEP components are relevant and can be reliably assessed. If spinal components can be obtained, the most prominent negative deflections in healthy subjects occur at approximately 13 ms (N13; upper extremity SEP) and 30 ms (N30; lower extremity SEP) after the stimulus . For median nerve SEPs, they most probably

Intraoperative Neurophysiological Mapping and Monitoring

347

represent dorsal horn postsynaptic activity triggered by ascending volleys from the dorsal nerve roots, which are reflected by smaller positive deflections preceding the N13/22 [49]. The N30 recordable at C5–C7 vs. Fpz with tibial nerve stimulation probably arises at the pontomedullary junction. The SEP components most commonly evaluated intraoperatively are the early cortical components N20 (upper extremity SEP) and P40 (lower extremity SEP), which are recorded from the (contralateral) parietal postrolandic electrode position described previously. They probably correspond to activity elicited in the posterior bank of the central sulcus (Brodmann area 3b) and are followed by P25/N50-components, which are thought to be generated in Brodmann areas 1 and possibly 3a/4 [50, 51]. Later components are probably generated in secondary association cortices. There is little experience with those late potentials in the intraoperative setting, in particular since they are more sensitive to anesthetic effects. They are not routinely evaluated for intraoperative monitoring purposes. Earlier components immediately preceding the N20/P40 may arise from subcortical structures like the basal ganglia and could therefore be of particular interest with supratentorial procedures [52], but they are too small and delicate to allow for reliable intraoperative evaluation. With all SEP components, amplitude and latency are the most important parameters that can be easily and objectively evaluated, the potential shape being more difficult to assess. Latencies are usually determined from stimulus to potential peak or between potential peaks (e.g., N13/N30–N20/P40 for central conduction time; for a definition of central conduction time, see [42]), and amplitudes are measured in a base-to-peak or peak-to-peak fashion. Common criteria for assessing SEP parameters intraoperatively are discussed in following text. Table 15.1 summarizes common stimulation and recording parameters.

2.2 PRINCIPLES OF CLINICAL APPLICATION SEP monitoring can obviously only provide direct information about the somatosensory pathways and regions. However, indirect conclusions about other brain regions can be drawn from SEP observation. Several different principles apply. Local proximity (e.g., of central lesions) makes it possible that locally interfering factors affect both the target region and adjacent somatosensory structures. Adjacent or remote regions sharing a common vascular supply with somatosensory structures can be indirectly assessed with vascular procedures (e.g., in middle cerebral artery [MCA] aneurysm surgery). Here the M2/M3 branches may supply the motor and sensory cortex simultaneously, and thus SEP changes would also indicate some possibility of a major motor deficit. Smaller perforators will more likely affect only one tract system. Therefore, if perforator manipulation is likely to occur [e.g., large posterior communicating

348

Georg Neuloh and Johannes Schramm

artery (PCOM) or ICA bifurcation aneurysm], combined SEP and MEP monitoring is advisable. SEP monitoring may also reflect the general functional state of the brain and can give important information in case of increasing intracranial pressure caused by remote events (bleeding), general hypoxia, or other systemic changes. This fact is most important when trying to determine the prognosis for comatose patients on the neurosurgical intensive care unit [53–56], but it may also be of use intraoperatively. With every application, only clinical experience can reveal the validity of SEP findings, i.e., the degree to which SEP changes reflect significant physiological phenomena in the target structure. 2.2.1 Interpretation and Practical Aspects The relevant parameters and SEP components have been described (Section 2.1.2). At present, there is no definite criterion for when to consider SEP changes significant with regard to impending neural impairment. Normal fluctuations, technical conditions, the influence of systemic factors like anesthesia or blood pressure (Fig. 15.2), and interindividual differences must be taken into consideration when interpreting intraoperative SEP findings.

FIGURE 15.2 Effect of systemic blood pressure on SEP. Median nerve SEPs were recorded during dissection of a right temporobasal AVM. During bleeding from the AVM, mean arterial blood pressure (ABP) was lowered to 60 mmHg for more than 1 hr, eventually leading to a significant impairment of SEP amplitude. This was resolved when ABP was raised above 80 mmHg after the successful resection of the AVM.

Intraoperative Neurophysiological Mapping and Monitoring

349

Although they still remain arbitrary to a certain extent, the experimental and clinical literature converges toward some criteria concerning SEP interpretation. A latency increase of earlier components by 10% is usually considered significant, and an amplitude decrease by less than 30–50% is considered within normal range [11, 15, 33, 34, 36–42, 57]. It is worthwhile to check sudden drops of amplitude against recent anesthetic bolus, blood pressure decreases, or surgical maneuvers. Changes of SEP waveforms are much more difficult to assess objectively and have been found to be of little value. SEP changes must be determined by comparing the latest curve with some reasonable baseline measurement. Usually, a first baseline potential is obtained after induction of anesthesia, before the operation has started. However, craniotomy and opening of the dura may lead to marked SEP changes before the relevant intracranial procedure has started. Thus it is often necessary to update the baseline values. These postdurotomy changes may be due to brain prolapse with ischemia at the rim or air between cortex and skull [58]. Moreover, nonsurgical factors like fluctuation of anesthesia, change in technical conditions, or previous transient or incomplete neural impairment may also lead to intraoperative SEP changes. These must be differentiated from truly current pathologic events so that further relevant changes can be interpreted with regard to the latest state considered stable. In principle, SEP monitoring makes sense with virtually all supratentorial procedures with regard to possible systemic or remote events, even if the somatosensory system is not affected. Given limits on time, equipment, and staff, a more specific use of SEP monitoring appears desirable. The next section presents some considerations on how to use a simple yet sensitive setup with specific applications. A general aspect to consider is whether to record SEPs bilaterally. In principle, bilateral recording is useful even with a clearly lateralized lesion, since monitoring of the unaffected side provides a biological control, which helps to differentiate specific (i.e., localized technical or iatrogenic) from systemic effects (e.g., anesthetic). In midline lesions such as anterior communicating artery (ACOM) aneurysms, the necessity of bilateral recording is obvious. Another issue is whether to record more than cortically evoked responses. Many published results refer to the central conduction time (CCT) [11, 15, 33, 34, 36–40]. In order to determine CCT, cervical potentials need to be recorded as well. The presence of spinal potentials also helps to ensure that proper stimulation and conduction through the peripheral nerves are being maintained. This helps to distinguish between peripheral failure and central failure when changes in cortical SEPs occur. We have seen two cases with unilateral cortical and spinal SEP loss due to acute plexus traction from an unnoticed dropped arm in the sitting position. However, in the operating room these subcortical potentials may be very difficult to record, much more so than cortical activity, in particular with tibial nerve stimulation. In fact, reliable recording of cervical potentials cannot be obtained in many cases.

350

Georg Neuloh and Johannes Schramm

2.2.2 Specific Applications 2.2.2.1 Aneurysms In supratentorial surgery, continuous SEP monitoring is mainly used with cerebral aneurysms [11, 15, 16, 33–40]. The primary goal of recording SEPs is to detect impairment of cortical perfusion by intended or inadvertent vessel occlusion. SEP changes are expected to also reflect perfusion of nonsensory eloquent areas like the motor or speech cortex, as well as local events (pressure or traction from retractor) and systemic influences. This necessitates that the manipulation of larger arteries is monitored (i.e., of ICA, M1, or A1). Principally, SEP monitoring in supratentorial aneurysms is only reliably used with lesions of the anterior circulation. The territory supplied by the posterior cerebral artery is neither overlapping with, nor adjacent to, the somatosensory cortex. The situation is different with posterior fossa aneurysms, since vascular supply of infratentorial pathways stems from the basilar and vertebral arteries. Tables 15.2 and 15.3 summarize the correlation of SEPs, surgical events, and clinical outcome. TABLE 15.2 Correlation of Intraoperative SEP Changes with Surgical Events in an Earlier Series of 282 Aneurysms from Our Service [16]

Event

N

Unchanged

SEP Deterioration

Loss

Accidental occlusion

22

18

2

2

Intentional occlusion

10

8

1

1

Temp. occlusion

51

29

14

8

Vessel retraction

6

4

2

—

Cerebellar retraction

1

—

1

—

Intent. narrowing

2

2

—

—

Others

11

11

—

—

Total

103

72

20

11

TABLE 15.3 SEP Findings versus Clinical Outcome with Temporary Vessel Occlusion in Aneurysm Surgery a Vessel occlusion (46/148)

SEP change (18/46)

No SEP change (28/46)

Clinical impairment (8)

6/18 (33%)

2/28 (7%)

No clinical impairment (38)

12/18 (67%)

26/28 (93%)

a

In a series of 148 aneurysms with 46 vessel occlusions, intraoperative SEP changes after vessel occlusion were correlated with postoperative clinical outcome. With SEP changes occurring independently from time of vessel occlusion, postoperative clinical impairment was significantly more frequent than with unchanged SEPs. Modified from [16].

Intraoperative Neurophysiological Mapping and Monitoring

351

For monitoring aneurysms of the upper posterior circulation system, visual evoked potentials would, in principle, be suitable. However, it has been shown that the current intraoperative technique of flash evoked visual potentials does not allow for reliable monitoring of the visual cortex or pathways [60–63]. With aneurysms of the anterior cerebral artery (ACA), tibial nerve SEPs are most likely to pick up compromise of the vascular supply of ACA territory, whereas median nerve SEPs will not reflect cortical ischemia in this area. Since ACA aneurysms are usually located close to the midline, bilateral tibial nerve stimulation is highly recommended, in particular with ACOM lesions. In the latter case, median nerve SEP monitoring may in theory also be useful, since perforating arteries branching off the ACA (such as Heubner’s recurrent artery) may supply territories comprising other than lower extremity motor or sensory fibers. However, median nerve monitoring is rarely performed with ACA or ACOM aneurysm clipping at our institution. MEP monitoring has been introduced at our institution with all anterior circulation aneurysms to pick up compromise of perforating vessels, which may lead to motor impairment not reflected by SEPs [12–14]. Contrary to the ACA, the branches of the MCA supply the sensorimotor hand area on the hemispheric convexity. Therefore, median nerve SEP (and upper extremity MEP) monitoring is performed with MCA aneurysms. Although bilateral recording is desirable, contralateral SEPs will provide most of the relevant information with a limited number of recording channels. Clipping of aneurysms of the ICA proximal to its bifurcation will be sufficiently monitored with median nerve SEPs, which are to be preferred over lower extremity SEPs because of their technical advantages. With aneurysms of the ICA bifurcation, tibial nerve SEPs may also be monitored if the ACA or its perforators are involved or if sufficient proximal control cannot be obtained. As with MCA aneurysms, unilateral SEP recording is not optimal, but sufficient in most cases of ICA lesions. If perforators are endangered, however, MEP monitoring is a useful addition (e.g., in large PCOM or giant ICA aneurysms). See Figs. 15.3 and 15.4 for examples of intentional and inadvertent vessel occlusion.

2.2.2.2 Vascular Malformations Surgery or endovascular embolization of vascular malformations located near the central region or close to the sensorimotor pathways can be usefully monitored with SEPs [64–66]. Local factors like retractor placement or heat from electrocoagulation may damage those structures directly, and SEP recording may provide early warning signs. Moreover, vascular malformations may share blood supply with adjacent brain areas, including sensory regions, and occlusion of superficial or deep feeders may compromise perfusion of normal brain tissue. SEP recording will allow for test occlusion of those vessels even in the

352

Georg Neuloh and Johannes Schramm

FIGURE 15.3 Surgeon’s reaction triggered by SEP loss. Median nerve SEPs were monitored during clipping of a middle cerebral artery (MCA) aneurysm. The aneurysm ruptured during repositioning of a first clip, and the M1 was temporarily closed. A significant reduction of SEP amplitude occurred almost immediately. Since clipping and suture of the avulsed aneurysm neck were not accomplished within 10 min, the temporary clip was readjusted to close only that M2 branch of the M1 bifurcation carrying the aneurysm. SEP recovery provided reassurance to proceed with the closure of the aneurysm base, which lasted a further 20 min. Postoperatively there was a slight transient hemiparesis. Modified from Fig. 2 in [143].

anesthetized patient. As with aneurysms, the type of SEPs that should be recorded depends on the location and vascularization of the lesion and the involvement of perforating vessels. Future monitoring setups will routinely include MEP monitoring in case motor function is endangered. 2.2.2.3 Tumors When feasible, tumor removal from within or adjacent to the central region or sensorimotor pathways mainly requires MEP monitoring, since impending sensory impairment alone will generally not be accepted as a criterion to stop tumor resection. Only with tumors close to or involving the basal ganglia (except the internal capsule), SEPs alone may provide the criterion for how far to proceed with the resection. With insular tumors extending deeply toward the

Intraoperative Neurophysiological Mapping and Monitoring

353

FIGURE 15.4 Inadvertent M2 closure detected by SEP change. Median nerve SEPs were monitored during clipping of an MCA aneurysm. Shortly after a seemingly successful placement of the clip, a significant SEP change occurred. The surgeon was informed, and thorough examination of the situs revealed an inadvertent closure of one M2 branch. After readjustment of the clip, SEP amplitudes recovered fully within a few minutes.

internal capsule, MEPs will be more useful than SEPs to determine resection borders (see Section 3). However, the tumor may involve vessels in the Sylvian fissure, or the transsylvian approach may lead to significant vasospasm [67], and SEPs may provide valuable information complementary to MEP monitoring in such cases. For complex deep-reaching and vessel-encroaching tumors, the combined use of SEP and MEP monitoring may be considered. If resources are limited, however, MEPs should be preferred.

3 INTRAOPERATIVE AND PERIOPERATIVE NEUROPHYSIOLOGICAL FUNCTIONAL MAPPING In supratentorial procedures, functional neurophysiological mapping basically comprises median nerve SEP phase reversal techniques for the localization of the central sulcus, or direct stimulation techniques used extraoperatively with implanted grid electrodes and intraoperatively in awake craniotomies. The most

354

Georg Neuloh and Johannes Schramm

common way to use the median nerve SEP phase reversal technique is as a first step during the resection of a lesion around the motor area. Median nerve phase reversal aids in planning the cortical approach, in determining the precise location of the motor strip, and in using the ideal stimulation site for motor tract monitoring during surgery in these areas. Extraoperative mapping via implanted grid and strip electrodes is mostly used in cases of epilepsy surgery. In addition to giving the same information as a simple phase reversal technique with median nerve SEPs, it allows additional information to be gained regarding sensory areas and in particular speech areas. Extraoperative mapping via a grid or multiple strips is therefore useful in all lesions encroaching on one of the two speech areas or encroaching on both the speech and the motor strip. From our experience with over 200 of these patients in our epilepsy surgery series, we are increasingly applying this technique to a series of regular tumor surgeries when the complexity, the location, or other factors seem to warrant such information.

3.1 SEP PHASE REVERSAL This technique is useful to point out the location of the central sulcus. The classic triphasic negative-positive-negative configuration of the early components of median nerve SEPs with N20 over the sensory cortex changes shape into a mirror image with N20 becoming P25 over the motor cortex. This phenomenon was initially described by Woolsey et al. [68] and has been used by other authors during epilepsy surgery [69] and tumor surgery [70–72]. Provided that the strip electrode is positioned in an adequate angle across the central sulcus, the switch in polarity for the early cortical median nerve SEP can easily be detected. Since both intrinsic and extrinsic brain tumors can easily displace cortical structures by 1.5–2.0 cm (which is more than the breadth of a gyrus), this can be important information. Figure 15.5 shows a typical case of a lesion around the central sulcus requiring use of the intraoperative phase reversal method (see also Fig. 15.8). 3.1.2 Technique Standard stimulation setup and recording parameters for median nerve SEPs are used. The reference electrode is fixed to the skull with its usual reserve electrode. The trepanation is planned in such a manner that the lesion can be approached and that there is enough space to place a two-row strip electrode across the presumed central sulcus. In principle, the SEP phase reversal technique can be applied both intra- and extraoperatively. The strip electrode may be pushed underneath the dura and away from the rim of the trepanation. Having ensured good contact with the brain surface by saline irrigation, one should check the electrode impedance to ensure that the electrode has settled

355

FIGURE 15.5 Central sulcus mapping by the SEP phase reversal method. Median nerve SEP phase reversal was recorded in a patient with a left precentral AVM (Spetzler-Martin grade II). The first precentral electrode (closest to the central sulcus) was selected for intraoperative MEP stimulation and elicitation of MEPs (see Fig. 15.11).

356

Georg Neuloh and Johannes Schramm

and that artifacts are minimized. If a four to eight channel averager is available, evoked potentials from the two rows of eight contact electrodes can be recorded within 5–10 min, depending on the number of channels and the level of noise. Since electrodes are located 1 cm apart and the sulci can sometimes easily be detected or dissected underneath the veins, a delineation of the position of the central sulcus is usually achievable in well over 90% of cases. Cedzich et al. have summarized some of our experiences [72]. One important result was that the presumed localization of the central sulcus as calculated on the basis of preoperative MRIs was found to be inaccurate as judged by intraoperative SEP phase reversal in 12% of all cases [72]. Even with modern-day MRI technology, it may be difficult to precisely and correctly locate the central sulcus, especially if normal anatomy has been distorted by tumor or overshadowed by pathological vascularization. In about 5% of these cases, the redetermination of the lesion site was from postcentral to precentral or from central to precentral (Fig. 15.6). Even if a classical phase reversal cannot be obtained, the central sulcus can often be located by the absence of the classical negative-positive-negative configuration of median nerve SEPs across the suspected sulcus (Fig. 15.7). It is also important that the direction of the two electrode strips pushed across the sulcus is at least 45° (and optimally 90°) to the assumed direction of the sulcus (on-axis vs. off-axis recording [73]). Obviously, the best median nerve potentials could also be obtained in the area representing the hand, which means the electrode should be neither immediately parasagittal nor just briefly above the Sylvian fissure. Sometimes if orientation is difficult, a four-row strip electrode should be used. The neurophysiology technician then marks the position of the phase reversal potential between the respective electrodes on a schematic drawing for the surgeon and presents it to the surgeon. The surgeon can then either memorize the position of the sulcus in this field or can mark it with a sterile cotton strip

FIGURE 15.6 Comparison of preoperative (MRI) data versus intraoperative (SEP phase reversal) localization. In a series of 67 cases from our service [72], the localization of pericentral lesions as determined according to preoperative MRI and intraoperative SEP phase reversal was compared. In a significant number of patients (8 patients, 12%) the lesion was intraoperatively found to be located differently than was suggested by preoperative imaging, which had obvious consequences in terms of approach and extent of resection.

Intraoperative Neurophysiological Mapping and Monitoring

357

FIGURE 15.7 Alternative localization of the central sulcus by absence of N20 instead of true phase reversal. In this patient with a parietotemporal tumor, median nerve SEPs were recorded via an (epidural) electrode strip placed across the central region. Although no “classical” phase reversal could be obtained, absence of the precentral SEPs allowed for identification of the central sulcus, which correlated well with anatomy.

placed directly on the cortex. The position with the best P25, which most likely corresponds to the hand area of the motor cortex, is then used for placement of the stimulation electrode for motor tract monitoring. For this purpose, the already placed strip electrode can be connected to the stimulator and used for stimulating purposes, or it can be replaced by a single 1 cm cup electrode for motor tract stimulation.

3.2 EXTRAOPERATIVE MAPPING WITH GRID OR MULTIPLE-STRIP ELECTRODES FOR MOTOR, SENSORY, AND SPEECH LOCALIZATION This technique is applied in lesions in or adjacent to speech areas and motor and sensory cortices. It is mostly used in surgery for chronic and drug-resistant epilepsy, but it is also applied in vascular and tumor-related lesions if the risk is considered high for a standard surgery. The technique of extraoperative

358

Georg Neuloh and Johannes Schramm

mapping implies two different procedures: the implantation of the grid followed by the neurophysiological evaluation, and as a second step the resection of the lesion. The alternative method to this two-step procedure is a one-step procedure whereby the electrophysiological functional mapping is done intraoperatively with the patient being awake during craniotomy. This intraoperative mapping technique was pioneered by Foerster, was used extensively by Penfield, and was later used extensively by Goldring, Ojeman, and the Montreal group. It has recently been revived for glioma surgery in several centers [71, 74–77]. Both methods have strengths and weaknesses. The main weakness of the two-step method is that two surgical procedures with anesthesia are needed. For review of advantages, see [78]. The advantage of this method is that testing can be done in a patient who is not under stress and who is not sedated. Most importantly, testing can be expanded as long as needed. In our settings, usually two sessions of approximately 3 hr are needed to get a reliable map of the speech area, the motor cortex, the sensory cortex, and the areas with interictal spikes for the epilepsy surgery cases. Proponents of the awake craniotomy procedure maintain that the patients do not find it stressful and are cooperative, and that generalized seizures are rare. We prefer extraoperative mapping mostly because of its more reliable nature and the chance to reproduce results on a second day of testing. It also has the advantage that in patients with epilepsy, lengthy EEG recordings to detect zones of interictal spiking, as well as ictal video-EEG recordings, are easy to perform. For several reasons, extraoperative functional mapping is obviously superior to acute intraoperative stimulation mapping in terms of reliability and precision of results (see [78] for review). On the other hand, it entails the risks of an additional operation, and it lengthens the inpatient hospitalization considerably if the electrode implantation is performed for reasons of functional mapping alone.

3.2.1 Technique When planning the trepanations, it is important to keep in mind that they must be large enough to accommodate the grid electrodes, which are usually 8 by 8 cm in size, or several strip electrodes, which come in sizes of 2 by 8, 2 by 4, or 4 by 8 cm. For this application, the larger rectangular subdural grid electrodes with up to 8 by 8 contacts are particularly well suited (see [79] for review). In the last 2 years it has been our routine to do a digital photograph of the exposed brain surface with and without the electrode in situ. This has the advantage that during a second stage the electrode grid can be removed, and the orientation will be easily maintained using the two photographs without having an electrode in the surgeon’s way or having to cut the electrode in order to delineate

Intraoperative Neurophysiological Mapping and Monitoring

359

the area to be preserved or resected. Craniotomies are placed in such a fashion that both the lesion and the functional areas of interest can easily be reached. The electrodes should usually be placed so that all the areas are open to direct visual observation. This implies that craniotomies are larger than for the standard removal of a tumor or a lesion in another brain area. Electrode cables are led outside through separate small incisions in the dura, led around the bony borders, which have been smoothed by rongeur, and then led through the scalp again via a separate small incision. Both the dural exit and the scalp exit are closed with an encircling pouch suture. Watertight sutures cannot always be achieved, but they are possible with a very careful suturing technique in most cases. The patient is referred to the electrophysiological unit the day after surgery, and a first recording session is done, usually lasting at least 2.5–3 hr. Since all the procedures have been explained to the patient in advance, cooperation is usually very high. Extraoperative mapping can include the standard median nerve SEP phase reversal, but also recording for interictal spikes and, of course, mapping of areas where stimulation induces dysnomia or speech arrest, or motor and sensory phenomena. In the presurgical evaluation of epilepsy, chronically implanted subdural or depth electrodes are widely applied for invasive seizure recording and hence for the determination of the epileptogenic area. Circumscribed electrical stimulation may induce clinical or subclinical seizures, especially in the zone of onset of spontaneous seizures. However, the localizing value of those induced seizures has been controversial [80, 81]. In epilepsy surgery programs worldwide, stimulation mapping via chronically implanted grid electrodes is applied whenever an overlap of the epileptogenic area with eloquent cortices is suspected. Three types of eloquent cortices are of major interest: 1. The primary motor cortex 2. The primary sensory cortices, especially the visual cortex and the somatosensory cortex 3. Association cortices with more or less circumscribed representations of higher cognitive functions, particularly the language cortices and the left perisylvian association cortex In the primary motor cortex, single stimuli typically elicit single clonic unilateral movements according to the topography of the motor homunculus in the precentral gyrus. High-frequency (50 to 60 Hz) stimulus series are responded to by slower, rather tonic contralateral movements. In the primary sensory cortices, high-frequency trains of stimuli also elicit positive responses according to the respective modality, namely, visual or auditory pseudohallucinations and somatosensory phenomena. On the other hand, in higher association

360

Georg Neuloh and Johannes Schramm

FIGURE 15.8 Lesions at the deeper part of the precentral sulcus, undermining the precentral gyrus. Schematic presentation of a grid with small electrodes of 0.5 cm distance center to center. The position of the central sulcus is most easily determined by initial phase reversal of median nerve SEPs for the hand area, and it can be confirmed by combined use of stimulation mapping and direct visualization following the direction or path of the identified sulcus.

cortices, mainly negative phenomena (i.e., specific loss of some aspect of cognitive behavior) are observed. Because of the different organization of those functional networks, the represented functions are acutely impaired during stimulation.

Intraoperative Neurophysiological Mapping and Monitoring

361

3.2.1.1 Mapping Using the Stimulation Technique In the University of Bonn epilepsy surgery program, electrical stimulation is performed and documented in a video-electrocorticography setting. Stimulation is performed with a Grass S88 stimulator and Grass constant-current unit (Grass Instruments, Astro Med, West Warwick, RI) by way of a step-by-step stimulation of neighboring electrodes in a bipolar montage. On the second day, the whole procedure, which may last up to 3 or 4 hr, is repeated with a montage orthogonal to the preceding one in order to improve spatial resolution and to confirm the previous results. If the data are contradictory or inconclusive, the procedures will be repeated. Stimulus intensity ranges from 1 to 15 mA, depending on the threshold of after-discharges. For intraoperatve stimulation parameters, see Table 15.4. The stimulation rate ranges between 10 and 60 Hz, with a train duration of 3–8 s, depending on the task (see following text). Motor and sensory cortices are mapped with a short stimulus train. For mapping of language and associated perisylvian functions (reading, calculation, writing, etc.), a train of stimuli below the threshold producing after-discharges is delivered for the duration of patient instruction and performance. After prior instruction, specially selected short tasks are given repetitively with and without stimulation. For language testing, the following subfunctions are evaluated: naming, continuous speech, repetition, body commands, reading, and sentence comprehension. Additionally, simple tasks from the Token Test [82] are performed under stimulation. Repeated testing yields a sufficiently reliable map of eloquent areas. In epileptic patients, this documentation is supplemented by an overlayed map of ictal onset sites, and additional electrophysiological data are recorded with the very same electrode array. It needs to be mentioned that the functional inhibitory effects of electrical stimulation, such as dysnomia or speech arrest, do not always and precisely indicate that the brain directly underlying that particular electrode represents the speech center. Spreading effects to adjacent cortical areas (e.g., by shortreaching U-fibers) are possible. In Fig. 15.9, for example, the stimulation effects with sensory speech phenomena are much bigger than Wernicke’s area as known from functional MRI testing. Another argument underlining this cautionary remark is the wide-spread disinhibition of effects on speech described by Ojemann, which reach anteriorly on the temporal lobe. If these stimulation effects would represent the true speech cortex, standard temporal lobectomy should produce speech deficits in a considerable proportion of cases. This is not the case, however, as shown by Hermann and Wyler in a prospective comparative study on temporal lobe resection with and without language mapping [83, 84]. As for rivaling methods, the localization of primary sensory and motor areas can be determined noninvasively by means of functional MRI with increasing validity [85]. The primary motor area can be identified indirectly using the

362

Georg Neuloh and Johannes Schramm

FIGURE 15.9 A patient with parietal cortical dysplasia. Motor, sensory, and speech phenomena can be identified. Stimulation in the region of the angular gyrus elicits motor and sensory speech phenomena. The dispersion of speech phenomena is marked and may be induced by cortical dysplasia. Occasional motor phenomena may be recorded following stimulation of the postcentral gyrus. This case demonstrates that conventional knowledge about cortical function may be misleading.

intraoperative SEP phase reversal technique [72]. Results from intracarotid amobarbital testing correspond well with electrical language mapping [86], but the Wada test is insufficient for determining the intrahemispheric topography of language areas. For the depiction of language areas, functional MRI has recently been shown to be an increasingly valuable method as well [87]. However, neurophysiological mapping must still be considered an indispensable procedure in presurgical evaluation, particularly in epilepsy surgery programs.

3.3 INTRAOPERATIVE ELECTRICAL STIMULATION OF MOTOR CORTEX AND WHITE MATTER An alternative method for functional mapping of the central region in anesthetized patients is to directly stimulate motor structures. SEP phase reversal essentially identifies the motor cortex as it relates to the somatosensory hand or foot area. With increasing distance from these areas, and with distortion of

Intraoperative Neurophysiological Mapping and Monitoring

363

the local geometry, mapping results may become unreliable. In this connection, SEP phase reversal may not be obtained, and identification of the motor cortex may only be possible with regard to anatomical criteria. Despite clear advantages of the phase reversal method, direct motor mapping is desirable in such cases. In fact, it has been repeatedly shown that a combination of direct motor mapping and SEP phase reversal allows for reliable identification of the motor area with a success rate close to 100% [72, 88, 89]. Two major methods for direct mapping of the motor cortex via electrical stimulation are currently applied. The first and more common method uses bipolar stimulation (i.e., the Penfield technique), with observation of movement during craniotomy in an awake patient. Or, if the patient is anesthetized, EMG can be used. This method has been used for many years [60, 71, 74, 75, 77, 88, 90–95]. However, it has been found to have conceptual and practical weaknesses and disadvantages that limit its use. The second method, using monopolar stimulation with recording of MEPs, has been developed over the last 10 years [72, 89, 91, 96]. Although it is not yet as commonly used as the traditional technique, it seems to lack some of the drawbacks and provides the additional possibility of continued MEP monitoring throughout the operation. Both methods are described only broadly here. A detailed overview of more recent techniques is given in Sections 2 and 4. 3.3.1 Bipolar Cortex Stimulation (Penfield’s Technique) Penfield [97] established the method of eliciting movement using a bipolar electrode to stimulate the exposed motor cortex with 50 to 60 Hz frequency stimulation. This method has been developed further by different groups over the last 50 years [68, 71, 74, 75, 77, 88, 90–95]. Whereas the original method relies on observation of tonic movements after cortical stimulation, recording of electromyographic (EMG) responses has been introduced as an improvement in sensitivity and safety [93]. 3.3.1.1 Stimulation The stimulation technique is based on the activation of cortical circuitry by repetitive application of electrical pulses. The current paradigm [71, 91] uses bipolar, biphasic rectangular pulses of 1 ms duration and 2 to 20 mA intensity, delivered for 1–4 s at a rate of 50–60 Hz via electrode strips or a hand-held stimulator with spheric steel tips spaced at about 5 mm apart. Stimulus intensity is increased in 1 to 2 mA steps, starting from a low level of 3–5 mA, until a motor response is recorded. The stimulus train is delivered for 1–4 s, until a response is obtained (if present) with a given stimulus intensity. Sites of successful stimulation can be marked with sterile tokens.

364

Georg Neuloh and Johannes Schramm

3.3.1.2 Recording Motor responses can be observed as limb or face movements. The limbs need to be sufficiently exposed so that the entire side contralateral to the craniotomy can be observed. With sufficient stimulation, brief but tonic movements are elicited that allow for a determination of the stimulated cortex’s specific motor function. The movement usually is small so that positioning of the patient and surrounding equipment will not be disturbed. As an alternative to the observation of movements, EMG recording with this stimulation paradigm has been recently described [89, 93]. Subdermal needle electrodes are placed into representative face, trunk, and limb muscle groups. The recording scheme depends on the location of the lesion. If precise mapping is the main goal, as many different muscles as possible should be recorded from. If subtle cortical mapping is less important than a greater extent of the activation given the limited number of amplifier channels, the two electrodes of each channel may be placed to connect related muscles groups, e.g., elbow flexors and extensors [93]. Neither movement observation nor EMG recording can provide quantitative information. 3.3.2 Clinical Application 3.3.2.1 Cortical Mapping The most common application of this mapping method is identification and delineation of the motor cortex with lesions within or adjacent to the central region. In a recent series of 66 such operations [93], motor responses could be registered by either EMG recordings or observation of movement in 79% of cases. In another recent study [89], motor responses were obtained in 95% of motor cortex stimulations, but were also present in almost 30% of premotor and in only 9% of postcentral stimulations. Whereas the sensitivity of the method appears somewhat low in the former series, the latter result indicates a doubtful specificity of the method with regard to motor cortex identification. 3.3.2.2 Subcortical Stimulation There are anecdotal reports on the use of this method for stimulation of subcortical motor tracts [92–94]. These concern its use for both defining resection borders with tumors approaching deep motor (and other functional) pathways, and for determining the extent of functioning tissue within the boundaries of infiltrative gliomas. However, the latter applications were not easy to replicate at our institution. Obtaining a muscle response by 50 to 60 Hz electrical stimulation of the motor cortex is based on the activation of motor cortex circuitry, which is a very powerful mechanism for reaching the α-motoneuron firing threshold. Producing the same effect by subcortical (axonal) stimulation requires much higher stimulation frequencies than 50 or 60 Hz (see following text).

Intraoperative Neurophysiological Mapping and Monitoring

365

Nevertheless, a reliable technique for stimulation of subcortical motor structures would obviously be a very attractive intraoperative aid to avoid lesioning the motor tracts. 3.3.2.3 Advantages and Disadvantages The method is virtually the oldest intraoperative electrophysiological method applied for routine use, and much experience has been gathered. A number of high-quality publications enable a reliable assessment of the adequacy in a given situation. However, the limited specificity of the method, its lack of quantifiable results and continuous monitorability, as well as safety issues (see Section 5), trigger the search for an alternative, and less problematic, technique.

3.4 MEP MAPPING The major part of the stimulation and recording paradigm for mapping purposes is equivalent to the monitoring technique presented in Section 4.1. Therefore, only the basic concepts, as far as they are important for an understanding of the mapping method, will be presented here. MEPs recorded from limb muscles are elicited by short trains of electrical pulses delivered to the exposed motor cortex. Because of the somatotopic organization of the motor cortex, motor responses recorded as evoked potentials from muscle groups provide a functional map of the stimulated cortical area. Continuous monitoring of motor tract functional integrity would, in principle, be possible with recording of spinal volleys alone, which are not significantly affected by anesthesia. Conversely, motor mapping requires specific identification of cortical regions beyond mere identification of the motor strip. This cannot be achieved by a recording of spinal activity, since a differentiated mapping of the spinal motor somatotopy is not yet possible. Therefore, recording directly from relevant muscle groups must be included in the mapping paradigm. 3.4.1 Technique 3.4.1.1 Stimulation Typically, trains of five (four to nine) monopolar anodal, rectangular electrical pulses of 300 (200–500) µs duration, with an interstimulus interval of 2–4 ms, are delivered to the exposed cortex via a hand-held stimulator or, more commonly, via an electrode grid covering the area of interest. The reference electrode is placed at Fpz or some other suitable remote cephalic site. For mapping purposes, as low as possible suprathreshold stimulus intensities are desirable to preserve the specificity of the method, since strong stimulation may lead to activation of neurons not located directly beneath the

366

Georg Neuloh and Johannes Schramm

stimulation site. Threshold intensity is usually below 15–20 mA. When muscle MEPs are being recorded, extrapyramidal motor fibers may theoretically be activated as well, but the actual mono- or oligosynaptic corticospinal connections descending from the primary motor cortex are obviously more sensitive to electrical stimulation. 3.4.1.2 Recording Although phasic muscle twitching is often elicited, observation of movements is not practical. MEPs, or compound muscle action potentials (CMAPs), are recorded by differential surface electrodes or, more commonly, by sterile subdermal needle electrodes. Impedances should be below 10 kΩ. Electrodes are placed in a belly-tendon fashion if single muscles or muscle groups are to be mapped. If a comprehensive sampling of all relevant muscle groups is required with a limited number of recording channels available, electrodes from two adjacent muscle groups can be connected to one channel. The recorded signal can be quantitatively evaluated in terms of latency, amplitude, width, and shape. Whereas this is important with continuous MEP monitoring, it is less relevant with mapping, although differences in amplitude between different stimulation sites may contribute to the specificity of the method. 3.4.2 Clinical Application 3.4.2.1 Cortical Mapping The main goal of using this technique is identification and delineation of the motor cortex, as well as determination of the lesion’s position with respect to it. For superficial lesions within or directly adjacent to the motor cortex, resection borders may be determined with the aid of the mapping procedure. With deeply located lesions, the transcortical approach can be planned with greater safety. This kind of motor mapping is usually performed in combination with recording of SEP phase reversal. Recent reports [72, 89, 91] showed a sensitivity and specificity of over 90% of combined MEP and phase reversal recording (as well MEP recordings alone) with regard to motor cortex identification [91]. MEPs were elicited in 97% of primary motor cortex stimulation, but in only 14.6% of premotor and in 8% of postcentral stimulation. The latter results are better when compared with the traditional stimulation method [91] (see Section 3.3). 3.4.2.2 Advantages and Disadvantages In principle, both methods presented here are suitable for mapping the motor cortex. The major advantages and disadvantages of the traditional method are discussed in Section 3.2.1.1. The more recent method of MEP mapping is not supported by clinical experience. Nevertheless, it appears superior in terms of

367

Intraoperative Neurophysiological Mapping and Monitoring TABLE 15.4 Stimulation and Recording Parameters for Intraoperative Mapping of the Motor Cortex

Polarity

Bipolar stimulation

MEP mapping

Bipolar

Monopolar anodal

Biphasic/monophasic

Monophasic

Frequency

50–60 Hz

200–500 Hz

Amplitude

2–20 mA

5–25 mA

Phase

Pulse width Number of pulses

1 ms

50–500 µs

Sustained stim 1–4 s

4–9 pulses

Motor response

Movement, EMG

CMAP

Filtering

10 Hz–3 (10) kHz

10 Hz–3 (10) kHz

Averages

None

None/few

sensitivity, specificity, and reliability [91]. At the same time, it is safer [96] and allows for quantitative evaluation and continuous intraoperative monitoring. Table 15.4 summarizes stimulation and recording parameters for both methods.

4 MEP MONITORING The preservation of motor function is a major concern with surgery in many supratentorial lesions. At the same time, the most radical resection possible is desirable with regard to postoperative quality of life [98–100]. It is therefore desirable to obtain a clear delineation of motor cortex and a continuous monitoring of the motor tracts’ functional integrity. Although preoperative functional imaging may help to identify the structures [101], those methods still lack the required precision at this point of time. Furthermore, they cannot be used for continuous intraoperative functional assessment. In fact, no imaging or mapping technique is conceivable that may replace preoperative direct recording of the actual neural activity. The intraoperative monitoring of MEPs with supratentorial procedures has only developed over the last decade, and its status has not yet reached that of a standard method such as SEP monitoring. In fact, this application must still be considered developmental. Beyond preliminary reports, no large series has yet been published [20, 72, 96, 102] on the clinical use of MEPs with supratentorial surgery. The situation is different with spinal cord MEP monitoring, where an increasing number of clinical reports converge on some general principles regarding recording technique and clinical interpretation [22, 103–107]. Based on these principles, this method is routinely applied in many institutions that perform spinal, medullary, or aortic surgery. Our method of MEP recording with certain supratentorial procedures, as well as some considerations and experiences with regard to its clinical application, is now described.

368

Georg Neuloh and Johannes Schramm

4.1 TECHNIQUE 4.1.1 Development of Technique Experimental work performed over the last 50 years [108, 109] has led to a rather detailed, though still very incomplete, understanding of the physiological processes occurring after electrical stimulation of the cerebral cortex (see Chapter 1 of this book). Since the introduction of transcranial electro- and magnetic motor cortex stimulation [110, 111] with recording of spinal and muscle evoked activity (MEPs), various attempts were made to use this technique for intraoperative motor monitoring purposes, mainly in spinal surgery [103, 112–117]. However, general anesthesia seemed to prevent reliable intraoperative recording of myogenic MEPs. In 1993, Taniguchi et al. from Bonn [96] proposed a modification of the stimulus paradigm that turned out to be suitable for both direct cortical [72, 91, 96] and transcranial [118, 119] stimulation under general anesthesia. The method exploits the fact that the inhibitory effect of most anesthetics at the α-motoneuron can be overcome by temporal summation [120] of excitatory postsynaptic potentials elicited by a train of descending volleys. This is directly elicited via corticospinal tract (CT) axons activated by transcranial stimulation. Only with this facilitation does reliable intraoperative recording of muscle motor activity become possible. Other methods of facilitation had previously been tried (e.g., H reflex). 4.1.2 Stimulation The current flow necessary to activate the motor tract can be induced by electric [108, 111] or magnetic [110] pulses, the latter also being applied transcranially. Both direct cortical and transcranial stimulation are suitable with monitoring for supratentorial lesions. If the motor cortex is exposed, direct cortical stimulation is preferable. With electric pulses delivered directly to the exposed cortex, descending pyramidal axons are activated at one of the first internodes [108], while the site of activation appears to be deeper in the white matter with transcranial stimulation [121]. Cortical circuitry may also be activated with direct electrical stimulation, leading to a variable series of corticospinal volleys (I waves) following the first volley from direct axonal excitation (D wave) [108]. Conversely, transcranial magnetic stimulation, with the coil axis oriented tangentially to the cortex, seems to activate only (or predominantly) cortical circuitry [122]. Therefore, the inhibition in the cerebral cortex induced by anesthesia is easily seen if MEPs are exerted with magnetic, rather than with electrical, stimulation. In practice, MEPs elicited by magnetic stimulation start to deteriorate during induction of anesthesia. Since it does not provide other advantages but does have a number of disadvantages (e.g., depression by anesthetics, strong magnetic field, bulky equipment, slow repetition rate

Intraoperative Neurophysiological Mapping and Monitoring

369

of magnetic stimulators, overheating of the coil, etc.), we will focus here on electrical stimulation. Most parameters applied with transcranial stimulation correspond to those employed with direct cortical excitation. Since transcranially elicited MEPs are usually more difficult to obtain, and since the recordings tend to be less stable in general, it is important to determine the optimum parameters, which can vary between patients. 4.1.2.1 Polarity Monopolar anodal stimulation has turned out to be the most efficient for cortical stimulation [108, 121]. With an anodal stimulus applied to the cortex, current is assumed to enter at the apical dendrites, leading to depolarization at the proximal Ranvier internodes of the CT axons [108]. Conversely, current from cathodal stimulation is thought to lead to hyperpolarization, which prevents excitation. In practice, cathodal cortical stimulation at a somewhat higher intensity level also leads to reliable motor activation in most cases. 4.1.2.2 Stimulation Site Direct cortical stimulation is applied with lesions that require exposure of the motor cortex, or if craniotomy enables the subdural advance of surface electrodes over the central region. In order to perform successful stimulation, the motor cortex has to be identified, which is often not possible by mere anatomic criteria, given the limited view. Therefore, some kind of functional mapping must precede the actual MEP recording. It is also possible to perform preoperative functional MRI and feed the data into a navigation system. The functional MRI data may then be transposed onto the cortex during surgery. This, however, is an expensive way to replace the quick phase reversal technique. In principle, the mapping procedures described in Section 3.3 could be used to search systematically for the optimum stimulation site. The point of best response (highest amplitude MEPs or lowest threshold for eliciting them) would be chosen. In practice, however, this procedure would be relatively timeconsuming and might even give unreliable results. For these reasons, SEP phase reversal is always performed preceding cortical stimulation (see Section 3.1). The point of the best precentral response (i.e., highest amplitude of P25/N40) is considered closest to the motor hand/foot area (N40) and is selected for motor cortex stimulation, while the second electrode (cathode) is placed at Fpz. In most cases, the craniotomy allows for median nerve SEP recording. The electrode is placed as close as possible to the sensorimotor hand area. With lesions located predominantly at the paramedian central region, tibial nerve SEP phase reversal recorded from the paramedian motor cortex is advantageous. If the recording electrode cannot be placed optimally, full phase reversal

370

Georg Neuloh and Johannes Schramm

may not be obtained. In particular, if the electrode strip is positioned off axis (i.e., does not equally cover sensory and motor hand area) and crosses the central sulcus, only an absence of potentials from the precentral gyrus, after attempts to record with the strip electrode, would indicate that motor cortex has been identified (Fig. 15.7) (for details see [69]). With reference to the local anatomy, it is usually still possible to reliably identify the central sulcus and the motor cortex [72, 89]. Similar problems may arise with extra-axial tumors like meningiomas or unusual constellations of vasculature preventing subdural advancement and correct positioning of the electrode strip. Alternatively, MEPs can still be successfully recorded by looking for the best MEP amplitude using an electrode strip and trying various electrodes. However, if the local geometry is significantly distorted (e.g., with a truly central lesion), unambiguous identification of the motor strip may not be possible with mere SEP phase reversal. The optimum site for transcranial stimulation has not yet been determined. Several different stimulation electrode montages have been described: The stimulating anode can be placed at Cz+2cm to Fpz or Fz (Cz+2cm indicating an anterior shift of about 2 cm from the standard Cz position of the International 10–20 System in order to obtain placement over the motor cortex). Use of a circumferential cathode from linked electrodes at Fz, F3/4, and A1/2 has proved to be more efficient [124]. Other electrode montages include positions C3+2cm/ C4+2cm and C1+2cm/C2+2cm (as before, 2 cm anterior to the standard positions of the International 10–20 System), corresponding to the motor hand area and positions somewhat more medial [118, 123, 125]. Stimulation of C3+2 cm versus C4+2cm, with the cathode placed over the contralateral hemisphere with upper-extremity MEP recording and over the ipsilateral hemisphere with lowerextremity MEPs, has been shown to be most efficient but produces pronounced twitches from contractions of the back musculature. The twitches are considerably lower with stimulation sites C1+2cm/C2+2cm or if the cathode is placed at Cz+2cm, but the stimulus efficiency is usually lower as well. Nevertheless, particularly with recording of MEPs from the lower extremities, which requires high-intensity stimulation, positions C1+2cm/C2+2cm may provide the best trade-off between high efficiency and low twitching. If the patient is in a sitting position, subdural air may collect intracranially, preventing transcranial stimulation with the standard electrode montage. In those cases, stimulation electrodes may be placed laterally/caudally from C3+2cm/C4+2cm positions [59]. 4.1.2.3 Intensity The threshold intensity at which a motor response can be recorded obviously depends on various factors, such as electrode position, anesthetic regimen, and application and degree of relaxation. With direct cortical stimulation, 10–15 mA

Intraoperative Neurophysiological Mapping and Monitoring

371

corresponding to 40–60 V at typical impedances around 2.5–6.0 kΩ is usually sufficient. With optimal electrode placement, 5 mA can be sufficient. Even with suboptimal electrode placement, preexisting paresis, partial muscle relaxation, or other interfering factors, 25–30 mA is very rarely to be exceeded. With transcranial stimulation, about 10–20% of the applied current reaches the motor cortex [126]. For eliciting upper extremity MEPs, 50–100 mA is usually sufficient, while for lower extremity MEPs, a considerably stronger transcranial stimulation intensity of up to 150 mA or even higher is often required. The common commercially available stimulating devices do not provide such highintensity output at sufficiently long duration of stimuli and high-stimulation repetition rate. In practice, stimulus intensity appears to be more important than stimulus duration, and pulses lasting 50–100 µs are acceptable, although a significant superiority of stimuli lasting 500 µs has been shown [127]. With constant voltage stimulation, up to 1000 V output can be obtained from current monitoring hardware using a stimulus duration of 50 µs. For mapping of the exposed cortex, stimulation at threshold may be adequate. For continuous monitoring purposes, an intensity somewhat above threshold is usually selected to obtain stable MEP monitoring. Since it could lead to a misinterpretation of MEP behavior, stimulation intensities higher than that are not recommended, both in order to not deliver unnecessary charge load to the brain, and to avoid excitation of the descending CT fibers caudal to the lesion. For eliciting MEPs from remote sites of the cortex after direct cortical stimulation (e.g., simultaneously from face and the lower extremities), stronger stimulation may be necessary to obtain reliable responses. However, the lesion is always deeply located in such situations. 4.1.2.4 Pulse Duration In our experience, pulse width is not a very important factor with direct cortical stimulation. With this technique, pulses lasting 200–300 µs are always sufficient, and the use of longer pulse width may mean unnecessary load to the brain. With transcranial stimulation, pulse width seems to play a more significant role. 4.1.2.5 Number of Pulses in Train Since a higher number of pulses will lead to summation of more EPSPs at the α-motoneuron, it will in principle lead to better motor responses. Depending on the frequency at which the pulses are delivered (interstimulus interval within the train), there is a limit with regard to how many successive excitatory postsynaptic potentials (EPSPs) can actually overlap and summate for a more efficient depolarization. However, it has been shown that sustained synaptic

372

Georg Neuloh and Johannes Schramm

depolarization at the motoneuron leads to repetitive firing of the neuron [128], so that a longer train of stimuli may eventually lead to a better motor response, without increased depolarization. With direct cortical stimulation, more than four to six pulses will rarely be necessary. 4.1.2.6 Interstimulus Interval (ISI) Within Train The findings on ISI are somewhat controversial [96, 118, 123, 125]. While in a number of clinical studies the highest MEP amplitudes were obtained at a train frequency of 350–500 Hz (ISI of 2.0–2.8 ms) [96, 118, 125], it has been argued from experimental findings that a frequency of about 250–350 Hz may be most efficient [123]. According to our clinical experience, all train stimulation frequencies between 250 and 500 Hz are about equally efficient with cortically elicited MEPs, while transcranial stimulation may be more successful with longer ISIs (lower train stimulation frequencies). 4.1.2.7 Train Stimulation Rate Another fundamental question is at which rate to repeat stimulation. Obviously, frequently recording MEPs will increase the sensitivity of the method for detecting impending motor damage. In our experience a repetition every 15–30 s is sufficient with many supratentorial procedures because of the slow pace at which such an operation usually proceeds. In certain situations (e.g., when approaching motor tracts during resection or after application of a clip to a blood vessel), the stimulation rate may be increased up to 1 Hz or even faster for a short period. During monitoring of MEPs in supratentorial lesions, preserving them is very important to predict a good motor outcome. Conversely, convincing data exist with spinal cord monitoring that show that the mere preservation or loss of muscle MEPs correlates with clinical outcome, whereas changes in amplitude and latency do not influence prognosis [22]. Figure 15.10 shows the influence of varying stimulation parameters on MEP latency and amplitude. 4.1.3 Recording Responses after transcranial or direct cortical stimulation can be recorded from the spinal cord as well as from peripheral nerves and muscles as spinal, neurogenic, and muscle MEPs. Eliciting movement after motor cortex stimulation under conditions of general anesthesia is not possible in a significant number of patients, and it does not provide quantifiable information [89, 93]. Recording of axonal, synaptic, or neuromuscular electric activity (MEPs) has greatly replaced the observation of movements with most stimulation paradigms.

Intraoperative Neurophysiological Mapping and Monitoring

373

FIGURE 15.10 Influence of varying stimulation parameters on MEPs recorded from the thenar muscle and elicited by transcranial stimulation (C3+2cm vs. C4+2cm). The following parameter constellation was varied systematically: five anodal pulses of 300 ms duration and 100 mA intensity were applied with an interstimulus interval of 2 ms (500 Hz) using constant current.

4.1.3.1 MEPs Recorded from the Spinal Cord (D and I Waves) and Neurogenic MEPs (recorded from Peripheral Nerves) Theoretically, motor evoked activity over the spinal cord could be recorded for monitoring supratentorial motor pathways. During spinal and intramedullary surgeries, recording of epidural MEPs has become a standard monitoring procedure in a number of institutions [123, 129], and semiquantitative correlations between intraoperative MEP findings and motor outcome have been established for those applications [22]. In practice, however, invasive recordings for D waves from the spinal cord are rarely used for monitoring the supratentorial portion

374

Georg Neuloh and Johannes Schramm

of the CT [102, 130]. Transcranial or exposed motor cortex stimulation elicits a series of volleys descending through the CT (D and I waves; see Chapter 2 in this book). If converging corticomotoneuronal output leads to a sufficient depolarization of the α-motoneurons, further volleys descend via the peripheral nerve toward the neuromuscular endplate. Axonal activity can be recorded as D and I waves via epidurally placed catheter electrodes [123]. The signal-to-noise ratio is usually good enough to allow for unaveraged recording. Occasionally, a few averages, rarely more than 3–10, are required. The main advantage of this recording technique is the relative refractoriness of D waves to the effects of general anesthetics, contrary to I waves, which are more easily suppressed by many anesthetics. With surgeries that do not provide access to the epidural spinal space, in particular with supratentorial lesions, the recording electrodes must be inserted via Touhy needles, which adds minimal but still additional risk. Although this may be the main drawback of the method, there are further disadvantages of spinal MEP recording with respect to supratentorial monitoring. Epidural recording does not reflect the somatotopy of the motor cortex; thus it allows for neither functional mapping nor selective monitoring for upper and lower extremities, as is possible with MEP recording from muscles. Corticonuclear tract volleys cannot be recorded at all, so facial paresis of central origin will remain undetected. Nevertheless, identification and global monitoring of large parts of the motor cortex are possible by spinal MEP recording [102, 130]. A success rate of only 80% for spinal MEP recording after direct cortical stimulation has been described [102]. In this series, the recording electrodes were inserted the day before surgery. For patients in whom reliable monitoring was achieved, a good correlation between spinal MEP behavior and postoperative outcome was found. Recording neurogenic MEPs (from the peripheral nerves) has been advocated with spinal procedures [131, 132], mainly after direct spinal rather than cortical stimulation. Application of this technique has not been described as yet for supratentorial procedures. The major potential advantage as compared with myogenic MEPs is the insensitivity toward muscle relaxants. Theoretically, a more differentiated monitoring may be possible with recording activity from peripheral nerves rather than spinal cord activity. However, only a few major nerves supplying upper and lower extremities are suitable for this rather tricky method, which requires the averaging of about 100 additional responses due to the low signal-to-noise ratio. Lateralized recording is possible, however. Other than spinal responses, neurogenic MEPs have the same sensitivity toward anesthesia as muscle potentials. Therefore, successful recording requires the same modified stimulation paradigm as described in Section 4.1. In summation, neurogenic MEPs do not seem to provide significant advantages over spinal MEPs in monitoring for supratentorial surgeries. (Note: A critical review of this method can be found in Chapter 2).

Intraoperative Neurophysiological Mapping and Monitoring

375

4.1.3.2 Muscle MEPs Recording MEPs from the muscle as the actual end organ of the motor system makes possible a valid assessment of the system’s overall functional state. Unlike spinal recordings, placement of recording electrodes in the muscle is minimally invasive. With the current anesthetic regimen and the adequate stimulation paradigm, reliable recording of muscle responses has become feasible. Recorded potentials represent CMAPs, or more specifically, muscle motor evoked potentials (mMEPs). 4.1.3.3 Electrode Montage for Recording mMEPs Muscle MEPs are most readily recorded from distal limb muscles, the corticomotoneuronal supply of which includes the highest proportion of mono- or oligosynaptic connections. Thus a typical electrode montage includes distal muscles of upper and lower extremities (e.g., forearm flexor, thenar, or hypothenar muscles and tibialis anterior and possibly extensor digitorum muscles). Other muscle groups typically monitored are the orbicularis oculi and orbicularis oris, the trapezoid, the biceps and triceps brachii, the quadriceps femoris, and the soleus muscles. Which muscle groups are selected for monitoring in a given patient depends on various factors, including the location of the lesion, the necessity of detailed intraoperative motor mapping, and technical constraints (number of available recording channels, etc.). With increased sensitivity (more muscles monitored) and unchanged specificity, the rate of falsepositive responses will increase. According to the experience with MEP monitoring at our institution, a montage including the orbicularis oris, thenar or forearm flexor, and tibialis anterior muscle will pick up basically all motor impairment that can be found on postoperative examination. Depending on the exact location of the target lesion, additional muscles may be included in the recording scheme. Regardless of the location of the lesion, merely recording from one limb may miss impending damage elsewhere. 4.1.3.4 Filtering Muscle MEPs may contain a large degree of high-frequency activity. Therefore, filtering must be performed with great care, avoiding too tight bounds. A 10 Hz to 3–10 kHz band-pass is usually suitable. The signal-to-noise ratio is usually quite excellent in terms of high-frequency noise, since MEP amplitude is often in the mV range. Conversely, a strong low-frequency electric stimulus artifact may occur that cannot usually be filtered out without affecting the signal. Thorough grounding of the patient is important with regard to both safety and recording quality.

376

Georg Neuloh and Johannes Schramm

4.1.3.5 Sweep Length and Averaging Sweeps of 70 ms will be sufficient to record motor responses from both upper and lower extremity muscles. With preexisting paresis, MEP latency may increase significantly, and sweep length may have to be adapted. Because of the high amplitude of a normal mMEP response, averaging is not necessary in most cases. With partial neuromuscular block, inappropriate anesthesia, or preexisting paresis, MEP amplitude may be decreased significantly, and averaging of a few (usually not more than 3–10) MEP responses may be helpful in overcoming the poor signal-to-noise ratio. High trial-to-trial variability of motor responses must be taken into consideration when deciding whether or not to perform averaging. 4.1.3.6 Signal The typical muscle MEP is an oligophasic response of 10 µV to 10 mV amplitude, occurring 15–30 (35) ms post stimulus with a duration of 10–15 ms. The latency depends on the recording site and varies greatly between individuals. Both latency and amplitude are subject to strong fluctuations in anesthetized patients [121, 133]. In particular, with transcranial electrical stimulation, amplitudes tend to vary to an extent that may prevent true quantitative assessment. Latencies of responses are determined as the time of the first significant deflection of the isoelectric line after stimulation. Amplitudes are usually measured in a peak-to-peak fashion as the difference between the most negative and the most positive component within one response. The duration of the response does not play an important role in monitoring practice. Central motor conduction time, an important parameter in perioperative and intensive care unit diagnostics, is not routinely measured in the operating room, since it complicates the recording procedure without adding much essential information.

4.2 PRINCIPLES OF CLINICAL APPLICATION The literature on the application of MEP monitoring with supratentorial procedures remains sparse [20, 59, 72, 89, 102]. In fact, the method must still be considered to be “in statu nascendi,” since only preliminary principles on how to interpret intraoperative MEP data with this application are emerging. The considerations and observations outlined here are based on our own clinical experience gathered with about 250 patients with supratentorial lesions. MEP monitoring followed transcranial or direct motor cortex stimulation. MEPs can provide direct information about the functional state of motor pathways. Although indirect conclusions about other brain regions can be drawn according to principles like local proximity and common vascular supply, the high sensitivity of MEPs to all kinds of interfering factors suggests a more specific use. Nevertheless, MEP monitoring can provide complementary information to SEP data in most cases where SEP monitoring is performed.

Intraoperative Neurophysiological Mapping and Monitoring

377

4.2.1 Interpretation of MEPs and Practical Considerations The relevant mMEP parameters are described in Section 4.1.3.2. There are no definite criteria for considering which MEP changes are significant for indicating impairment in supratentorial lesions. Our practical experience in using them intraoperatively is described in following text. As with SEPs, normal fluctuations, technical conditions, the influence of systemic factors like anesthesia, and interindividual differences must be taken into consideration. According to our experience, the application of latency and amplitude criteria analogous to interpretation of SEP findings leads to clinically useful conclusions. While a latency increase by 10% can be considered significant, an amplitude decrease by less than 50% may be considered within acceptable range. The following semiquantitative relations seem to apply. Thus far, potential loss is most important: irreversible MEP loss is associated with permanent new postoperative motor deficits, while preservation of MEPs denotes preserved motor function (Fig. 15.11).

FIGURE 15.11 Functional testing after placement of the test clip on a feeder in a patient with an AVM at the precentral sulcus. Upper extremity MEPs were recorded during removal of a left precentral AVM. After induction of hypotension and application of a test clip on a major feeder, no significant changes of the parameters of the MEPs were recorded. This ruled out ischemia to the motor cortex and enabled the surgeon to proceed with this potentially hazardous procedure. There was no postoperative neurologic deficit.

378 TABLE 15.5

Georg Neuloh and Johannes Schramm Correlation Between Cortically Evoked MEPs and Clinical Outcomea

MEP findings

Clinical outcome (monitored limb) Permanent new deficit

Irreversible loss Irreversible alteration

Transient new deficit

No new deficit

always frequently

frequently

rarely

Reversible loss

rarely

frequently

frequently

Reversible alteration

rarely

frequently

frequently

Unaltered

always

a

This is a schematic summary presenting our long-standing experience in about 170 patients in whom MEP monitoring was applied for supratentorial lesions and intraoperative MEP findings were correlated with postoperative motor outcome.

Irreversible MEP alteration, or reversible loss or alteration, is associated with either permanent or transient motor impairment, the former being rare with reversible MEP loss or alteration. Reversibility of MEP changes is usually due to some surgical reaction to an early warning of MEP impairment. These principles differ from the current state of mMEP interpretation in spinal procedures [22, 123], resembling an all-or-nothing paradigm. Changes of MEP waveform are more difficult to assess objectively. However, dispersion phenomena with impaired motor conduction may lead to a more polyphasic potential, which may therefore be interpreted as a warning sign. Central motor conduction time, routinely evaluated in preoperative and diagnostic MEP measurements, does not play a significant role in intraoperative MEPs. As a matter of fact, MEP latency increase is rarely found without concomitant amplitude decrease, so that isolated latency assessment is hardly ever necessary. Table 15.5 summarizes current principles of MEP interpretation. Figures 15.12–15.15 show typical MEP behavior due to lesion resection or other surgical events. MEP changes must be determined by comparing the most recent potential with some reasonable baseline measurement. A first baseline potential should be obtained after induction of anesthesia, before the operation has started. However, craniotomy and opening of the dura may lead to significant MEP changes before the actual intracranial procedure has started, so that it may be necessary to update the baseline values. Nonsurgical factors such as fluctuation of anesthesia or change in technical conditions (e.g., poorer electrode contact during direct cortical stimulation due to brain shift) may lead to intraoperative MEP changes, necessitating the interpretation of further MEP changes with regard to the latest state considered stable. Analogous to SEP monitoring, bilateral MEP recording is mandatory with midline lesions and desirable with lateralized approaches. However, bilateral recording is usually only feasible with transcranial stimulation, since direct cortical stimulation requires exposure of the motor cortex.

Intraoperative Neurophysiological Mapping and Monitoring

379

FIGURE 15.12 MEP loss due to overheating from an electrocoagulation instrument. Upper extremity MEPs were monitored during removal of a partially thrombosed postcentral-central AVM (Spetzler-Martin grade III) that had bled 2 weeks before, leading to a severe hemiparesis. Extensive intraoperative use of bipolar coagulation during removal of the bleeding AVM led to transient MEP deterioration and subsequent complete loss. At the end of the AVM resection, MEPs returned to the baseline parameters. There was no postoperative increase in weakness, and the preoperative hemiparesis slowly improved during follow-ups.

MEPs should be recorded from both upper and lower extremities as well as from facial muscles, since motor deficit may occur in muscle groups or limbs not monitored with otherwise stable MEPs. Given the usually limited number of recording channels, the selection of recording sites depends on the specific situation in a given patient. Some basic considerations are sketched in following text.

380

Georg Neuloh and Johannes Schramm

FIGURE 15.13 MEP loss due to vasospasm. MEPs were recorded from forearm flexor muscles during resection of an insular glioma via a transsylvian approach. After tumor removal, a significant reduction (and eventually a complete disappearance) in MEP amplitude occurred. Warning was given to the neurosurgeon. Papaverine was applied to the spastic sylvian vessels, and MEPs recovered within 6 min.

4.2.2 Specific Applications 4.2.2.1 Central Tumors and AVMs The indication for MEP monitoring with central lesions adjacent to motor pathways is obvious. Direct motor cortex stimulation is performed via an electrode strip or grid. Figure 15.16 shows a typical case. Where to focus the recording depends on the exact location of the lesion. The more deeply seated the lesion, the more comprehensive a recording scheme is required. With a small superficial lesion of the lateral frontal convexity, MEPs from the arm and face muscles may be sufficient, while a frontomedial lesion will also require lower extremity MEP recordings. Concomitant SEP recordings are seldom useful even with postcentral lesions, since resection will usually not be stopped because of SEP changes with preserved MEPs. In arteriovenous malformations (AVMs), motor tract monitoring is useful if the AVMs are located very close to the motor cortex or the motor tract. An example is found in temporomesial locations when the AVM encroaches on the brainstem and the tangle of vessels needs to be dissected from those vessels that encircle the

Intraoperative Neurophysiological Mapping and Monitoring

381

FIGURE 15.14 Permanent MEP deterioration with permanent new paresis. Upper extremity MEPs were recorded during removal of a precentral malignant glioma. During resection, a significant MEP deterioration occurred without complete loss close to the posterior margin of the tumor. When a further MEP deterioration was indicated during removal of the posterior lesion wall, resection was halted. There was a permanent new hemiparesis postoperatively.

brainstem and give off perforators through the basal surface of the brain. In this situation, a test clipping of a vessel can be useful. The same holds during a dissection of an AVM when the surgeon wants to temporarily exclude vascular branches to the motor cortex. Lack of significant MEP and SEP changes helps support the surgeon in a dangerous step of dissection or sacrifice of a vessel. In our experience, combined use of MEPs and SEPs can be helpful if the vascular malformation is located close to the brainstem, underneath the insular cortex, or very close to the central sulcus. In some occasions potential loss was observed due to vasospasm, which could be reversed by the application of papaverine. In several instances SEPs or MEPs changed because of the systemic effects of low blood pressure (BP). Significant lowering of mean arterial BP is a well-known aid in handling the deep-seated, thin-walled vessels in the white matter at the deep apex of the AVM. Usually there are no changes at a mean BP of 60 mm Hg, but the combined effects of retraction, compression of the AVM toward the surrounding brain, and lowered BP can lead to MEP loss. Although the number of AVMs lying close to the motor pathways is not high enough for a systematic evaluation of the results

382

Georg Neuloh and Johannes Schramm

FIGURE 15.15 Permanent MEP deterioration with transient new deficit. Upper extremity MEPs were recorded during resection of a temporoinsular benign oligoastrocytoma invading the basal ganglia. A significant MEP impairment occurred when the medial portion of the tumor was being resected. Resection was stopped in this area. Postoperatively, there was a slight transient worsening of the preoperative hemiparesis, which resolved completely until discharge.

of MEP monitoring, a number of rewarding observations support the notion that monitoring for AVMs may be very helpful in avoiding neurological deficits. More clinical perspectives and experiences are summarized in Section 6. 4.2.2.2 Insular Lesions Insular tumors may extend deeply toward the basal ganglia and the internal capsule [134]. The temporal part and, on the right side, the frontal part do not pose insurmountable problems. The greatest threat is posed by a tumor growing beneath the M2 and M3 branches, which are covered by the frontal and temporal opercula, and the tumor has replaced the insular cortex, the claustrum, and the external capsule and has invaded the lateral putamen (especially its most dorsal extension). If the surgeon uses a transsylvian approach, all these MCA branches have to be dissected away and the tumor resection has to proceed beyond these arteries. One will also encounter the small perforating vessels that are supplying the lateral putamen and parts of the tumor. This is the

Intraoperative Neurophysiological Mapping and Monitoring

383

FIGURE 15.16 MEP deterioration that convinced the surgeon to stop a resection of a precentrally located tumor. Upper extremity MEPs were monitored during removal of a precentral low-grade glioma that had partially invaded the precentral gyrus. Tumor resection was halted when MEPs deteriorated during surgical approach to the motor strip. This led to full MEP recovery. Postoperatively there was transient motor aphasia without paresis, corresponding to an incomplete supplementary motor area (SMA) syndrome. Postoperative MRI revealed the remains of the tumor anterior to and within the precentral gyrus.

most significant technical problem in the resection of insular gliomas, and, in our experience, MEP recording has been found to be useful in such cases. It provides a valuable criterion for proceeding with the resection, in particular if intraoperative imaging is not available, as is the case in most neurosurgical centers. Our experience with some 40 patients with insular tumors and MEP monitoring shows that maximum resection without significant postoperative motor impairment is obtained if resection is stopped at the first signs of significant and persistent MEP deterioration (Fig. 15.17). Postoperative imaging typically reveals resection in the putamen close to the internal capsule, while slight transient paresis may be found in the postoperative clinical evaluation. With insular

384

Georg Neuloh and Johannes Schramm

FIGURE 15.17 MEP loss indicating resection of tumor near the internal capsule. Upper extremity MEPs were recorded during resection of a grade II insular glioma extending deeply toward the internal capsule. While the surgeon was working deep under the level of the insular cortex, MEPs were impaired, and resection was halted within this portion of the tumor. There was no new motor deficit postoperatively, and early postoperative MRI showed gross total (>90%) tumor resection, with the small dorsal portion of the tumor removed.

lesions, the size and location of the craniotomy determine whether transcranially or cortically elicited MEPs are performed. In any case, a comprehensive recording montage is advisable. SEP monitoring may also be useful with large lesions that require extensive dissection of sylvian vessels that are possibly supplying nonmotor parietal territories. We have now had experience with well over 40 patients with insular tumors monitored with MEPs; the tumors included various types of astrocytic tumors. Nearly half of the patients had a tumor on the left side, and in about 40% of them, complete resection could be achieved as assessed on postoperative MRI. The policy of continuing the resection until significant changes of MEPs are observed has proven to be extremely useful. Even transient MEP losses are

Intraoperative Neurophysiological Mapping and Monitoring

385

acceptable if surgery stops immediately. The likelihood of permanent deficits in this group is less than 10%. The introduction of MEP monitoring was a strong supporting element in our decision to develop a strategy for total resection of huge fronto-temporo-insular gliomas where the whole MCA tree frequently has to be skeletonized. More clinical perspectives and experiences are summarized on Sections 6.4 and 6.5. 4.2.2.3 Aneurysms SEP monitoring does not pick up impending motor impairment in a certain proportion of aneurysms, obviously because of a poor sensitivity for subcortical ischemia induced by compromise of perforating vessels (see Section 2) [13]. At our institution, MEP monitoring in parallel with SEP monitoring has been introduced, and has proven useful in predicting postoperative subcortical stroke with unchanged SEPs (Figs. 15.18 and 15.19). It is not yet clear, however,

FIGURE 15.18 Test occlusion of ICA with no change in the parameters of SEPs and MEPs. MEPs recorded from the finger flexor and median nerve SEPs were monitored during clipping of a left ACI aneurysm. No changes in the parameters of the evoked potentials during test occlusion enabled the surgeon to proceed with surgery.

386

Georg Neuloh and Johannes Schramm

FIGURE 15.19 New motor deficit from subcortical stroke following perforator manipulation reflected by transient MEP loss and subsequent MEP deterioration despite stable SEPs. Thenar MEPs and median nerve SEPs were recorded during dissection and clipping of an aneurysm of the basilar artery and superior cerebellar artery and a posterior communicating artery (PCOM) infundibulum via a common pterional approach. MEP impairment with no changes in the parameters of SEPs occurred during manipulation of perforators from the PCOM. Postoperatively, the patient experienced a new slight hemiparesis, and brain CT scans revealed a small basal ganglia infarction.

whether MEP monitoring can reliably help to prevent such lesions, which seem to occur with inadvertent compromise of perforating arteries at dissection of the aneurysm. The dissection will be performed with maximum care anyway and perforators will be inspected. They may appear open but kinked, and the compromised flow through these small vessels may be irreversible even if it is not due to placement of the clip (e.g., by vasospasm). Given the more basal location of the craniotomy usually required for aneurysm surgery, MEPs are usually elicited by transcranial electrical stimulation. Recording MEPs from the upper extremities and face muscles may be

Intraoperative Neurophysiological Mapping and Monitoring

387

sufficient for vertebral basilar artery and MCA aneurysms. In surgery for ICA and ACA aneurysms, lower extremity SEPs are more important than lower extremity MEPs which are more difficult to elicit. Cortical ischemia seems to be reliably picked up by SEP monitoring. The perforating arteries branching from the ACA (e.g., the Heubner’s recurrent artery) supply motor pathways, including corticonuclear efferents and fibers descending toward cervical motoneurons. Thus, if lower extremity MEPs cannot be reliably obtained with ACA aneurysms, or exceedingly strong muscle twitches incompatible with microsurgical dissection would be unavoidable, one may switch to the more readily recordable hand MEPs. A summary of clinical aspects and limitations is given in Sections 6.1 and 6.2.

5 SAFETY AND ANESTHESIA

5.1 SAFETY Neurophysiological monitoring, when properly performed with reliably functioning equipment and according to the technical standards described in these chapters, does not seem to carry significant health risks to the patient. The placement of subdermal needle electrodes in a sterile manner for stimulation or recording has never been associated with inflammation or severe bleeding in our experience with over 1500 patients monitored intraoperatively. The only adverse effect is slight local venous bleeding after removal of the needle electrodes in some cases. The use of electrical equipment and the application of current for stimulation purposes requires optical grounding of all patient contacts and thorough grounding of the patient to prevent inadvertent current flow. 5.1.1 SEP Monitoring This type of monitoring requires almost continuous stimulation, often for several hours at a relatively strong intensity (as compared with the SEP eliciting in awake patients). This has never led to adverse effects in a series of over 900 monitored patients at our institution. Muscle twitching induced by peripheral nerve stimulation in nonrelaxed patients does not seem to cause postoperative consequences and rarely interferes with microsurgical dissection. 5.1.2 Motor Mapping and Monitoring The main concern with high-voltage electric pulses applied to the brain is excessive charge density, charge per phase, and “damage threshold” values for neural injury that have been deduced from animal models [126, 135]. It has been argued that electrical stimulation for eliciting MEPs, presented in Section 4,

388

Georg Neuloh and Johannes Schramm

meets the limits set by these conservative criteria [96]. Conversely, the sustained repetitive stimulation at 50–60 Hz via bipolar ball-tip stimulators (Penfield’s technique) may exceed those limits by many thousand of times [71, 77, 88]. As a matter of fact, no neural damage has been found in well over 350 cases of direct cortical or transcranial stimulation performed at our institution, or in other large series of transcranial MEP monitoring [22, 107, 125]. Even with the much more aggressive exposed cortex bipolar stimulation method (Penfield’s technique), no neural injury has been reported [68, 71, 74, 75, 77, 88, 90–95]. However, the release of gas bubbles has been observed in animal experiments with only a slight modification of the method (rectified stimuli) [68]. With the bipolar stimulation method (Penfield’s technique), overt seizures (tonic movement beyond stimulation) were found in 11% of cases in a recently published series [93] and prolonged EMG activity after stimulation in another 14%. The concern, however, that the method may induce epilepsy by some kind of kindling effect [136] has not been confirmed. Conversely, epileptogenicity of the MEP stimulation method is very low because of the high-frequency trains, which have a lower tendency to trigger autonomous cortical circuitry than frequencies around 50 Hz. An additional factor is the length of the train: 25 ms in the MEP stimulation method versus 3–4 s in Penfield’s technique using 50 Hz stimulation. In our experience with both direct cortical and transcranial electrical stimulation using a short train of stimuli, intraoperative seizures are induced or facilitated in very rare cases (below 1%) that require immediate application of neuromuscular block or barbiturates. Other adverse effects that can potentially interfere with the ongoing microsurgical procedure include strong muscle twitching by inadvertent activation of back muscles. Sometimes this requires a brief halt in monitoring during the lesion dissection.

5.2 ANESTHESIA SEP and MEP amplitudes, waveforms, and latencies are affected by various systemic factors, including body temperature, blood pressure, blood oxygenation, and PaCO2 [137, 138]. Although the effects can be significant, careful preservation of homeostasis by the anesthesiologist and the surgeon will minimize such influences. On the contrary, the effects of anesthestics may be more discrete in some cases, but they obviously cannot be eliminated and must be taken into consideration with intraoperative evoked potential (EP) assessment. Latency or amplitude alterations occurring with fluctuations in the depth of anesthesia may mask or simulate changes induced by surgical events. This may lead to false-negative or false-positive interpretations of the neurophysiological signals. The effects of anesthesia will often be more prominent with preexisting EP alterations (e.g., due to neural damage) or with more difficult recording

Intraoperative Neurophysiological Mapping and Monitoring

389

conditions (e.g., with lower extremity SEPs or MEPs). Coordination with the neuroanesthesiological team should be established. In addition to sharing details about anesthesia in a given case, the surgeon often has to guide the anesthesiologist with regard to the anesthetic regimen. This method should be discussed with the anesthesiologist to obtain optimal anesthesia for the patient. 5.2.1 Anesthesia for Motor Cortex Mapping and MEP Monitoring 5.2.1.1 MEPs Homeostasis and adequate anesthesia are of paramount importance with muscle MEP monitoring. The great importance of anesthesia for successful EP recording is attested to by the fact that a whole chapter (Chapter 17) is devoted to this subject. Here, we will restrict ourselves to some basic descriptions regarding MEPs. Muscle MEP monitoring is mainly hampered by the inhibitory effects of general anesthesia at the corticomotoneuronal junction. Although the short train stimulation paradigm has led to a major improvement in eliciting MEPs in terms of stability and overall recordability, the anesthetic regimen still exerts great influence on MEP properties (e.g., signal-to-noise ratio and trialto-trial variability). Since semiquantitative evaluation of responses is helpful in supratentorial MEP monitoring, amplitude stability is a major issue. Total intravenous anesthesia with propofol and opioids has proven suitable and is applied in many institutions that perform MEP monitoring [123, 125, 139]. Volatile anesthetics, in particular halogenated agents, may have a stronger depressant effect on motor responses and are therefore avoided with MEP monitoring when possible [139]. In practice, however, MEP monitoring is usually still possible under inhalational anesthesia if total intervenous anesthetic (TIVA) use is not feasible. The greater variability of responses must be taken into consideration when interpreting intraoperative changes. Partial neuromuscular blockade adds to the variability of MEP responses but may still preserve recordability [140]. Although relaxation should be avoided with MEP monitoring to obtain optimum recording conditions, the muscle twitching may be so disturbing that only a partial neuromuscular block will enable continued parallel MEP monitoring and microsurgery. In our experience, a blockade of up to two responses in the train-of-four paradigm is compatible with MEP monitoring. This corresponds to the findings of other groups [140, 141]. 5.2.1.2 Bipolar Cortex Stimulation Technique (Penfield’s technique) The compatibility of mapping the exposed motor cortex using the bipolar stimulation method in the presence of anesthesia has not been formally tested. However, it has proven compatible with an anesthetic regimen including

390

Georg Neuloh and Johannes Schramm

nitrous oxide, isoflurane, fentanyl, and propofol [93]. To reliably pick up movement or EMG motor responses, relaxation cannot be applied during the mapping procedure. Even residual relaxation after intubation should be excluded to avoid false-negative responses when trying to identify motor cortex areas.

6 VALUE AND LIMITATIONS FROM A CLINICAL PERSPECTIVE For all neurosurgical groups that have been involved with SEP monitoring— such as ours—the advent of reliable motor tract monitoring has been a longawaited event. MEP monitoring has been possible on a routine basis for about 7 years now and has been applied not only for aneurysm surgeries but also for centrally located gliomas, centrally located AVMs, insular gliomas, difficult-toaccess cavernomas, deep-seated metastases (if not treated by Gamma-knife), and lesions in the midbrain. Simultaneous monitoring of MEPs and SEPs was found to be a valuable addition to the armamentarium that is available in modern microneurosurgery. This section summarizes our experience with 250 supratentorial surgeries monitored with MEPs.

6.1 ANEURYSMS Many experienced surgeons do not believe that there is a need to monitor aneurysms with SEPs, let alone with MEPs. Is it worthwhile to perform both methods and increase the expense of monitoring 100 patients to avoid one or two hemipareses? In our experiences, it is. In our first publication on aneurysms and monitoring with SEPs, we concluded that monitoring should be done in larger lobulated or complicated aneurysms of the anterior circulation [15]. We still believe that this is true. All neurosurgeons who have started aneurysm surgery during the last 10 years in our service have chosen to use this modality whenever it was available. In certain aneurysms, such as very large ones in the MCA bifurcation or large aneurysms of the carotid bifurcation or of the PCOM, we would delay surgery if monitoring was not available (even for more than a day if it was elective surgery). Such has become the attitude of the neurovascular fellows in our group. Small MCA aneurysms, especially if they have not been bleeding, can be done without monitoring. Aneurysms of the basilar tip or at the origin of the superior cerebellar artery (SCA) need to be done with simultaneous monitoring of SEPs and MEPs. We have currently gathered experience with about 60 aneurysms using simultaneous SEP and MEP monitoring, since SEPs alone would be unreliable. It was found that if new

Intraoperative Neurophysiological Mapping and Monitoring

391

motor deficits occurred that had not been detected with the SEPs and could be reconstructed after the surgery from postoperative MRI or CT, small perforating vessels usually accounted for those insults that led to MEP deterioration during surgery or shortly after it. Obviously, the avoidance of all motor defects from surgery is the aim. But some of these defects were obviously due to vasospasm or to the insufficient flow in a seemingly unaffected or only mildly distorted perforating vessel. Thus MEP monitoring helps more than SEP monitoring alone, but it cannot prevent all motor deficits. Although no formal evaluation has yet taken place, a number of cases in which clips were reapplied or the course of surgery was influenced by motor tract monitoring appear to be similar to the experience with SEP monitoring. It is an interesting experience in aneurysm surgery when one observes potential loss, inspects the situs, and suddenly detects the kinking of a perforator 2–3 mm beyond the tip of the clip due to indirect traction exerted by tension on an arachnoid string. The use of intraoperative ultrasound is easily possible on branches such as M2 or the carotid main trunk. However, it can be impossible for the smaller perforators, especially if they are obstructed from view or covered by the bulk of the aneurysm.

6.2 ANEURYSMS—SUMMARY The absence or presence of SEP changes following surgical maneuvers may be equally useful indicators. SEP changes led to a modification in surgical strategy in 8.1% of patients in whom SEP changes were monitored. Surgical reaction consisted of removing offending traction, removing temporary vessel clips, replacing aneurysm clips, and elevating blood pressure [16]. Whereas in the past aneurysms were only monitored with SEPs, complicated aneurysms have now been simultaneously monitored using MEPs and SEPs in over 60 patients. The most important difference from single-modality SEP monitoring was the increased detection of motor impairment stemming from manipulation of small perforators. Not all these motor deficits could be reversed by the surgeon’s reaction, but there are reasons beyond the immediate capability of the surgeon to rectify them. The fact that not all motor deficits can be avoided should not be taken as a serious argument against motor tract monitoring, since every motor deficit avoided is worth the effort. We observed that necessary narrowing of an M2 branch with reduced flow confirmed by intraoperative ultrasound testing could be accepted since the functional monitoring with MEPs and SEPs did not show changes. Neurophysiological monitoring was thus helpful in deciding which degree of narrowing was acceptable and which was not.

392

Georg Neuloh and Johannes Schramm

6.3 AVMS—SUMMARY (SEE ALSO SECTION 2.2.2.2) Over 30 AVMs and over 30 patients with cerebral cavernomas have been monitored. In vascular surgery the absence of changes is as important as changes or the disappearance of the potentials. Test clipping without changes in evoked potentials can be of significant reassurance for the surgeon and thus help in speeding up the procedure. In addition, systemic factors such as arterial blood pressure that may play a more important role than in tumor cases can also be assessed. With many vascularly induced changes, reversibility can be achieved. Therefore, monitoring in patients with vascular pathology appears to be particularly helpful.

6.4 CENTRALLY LOCATED TUMORS Centrally located tumors provide particular problems: The normal geometry may be distorted to such a degree that it is impossible to even guess whether the motor cortex has been pushed backward or anteriorly or whether it is lying underneath the lesion. In patients with cavernomas, it may be of the utmost importance to detect precisely whether the motor tract fibers converging toward the knee of the internal capsule are being pushed anteriorly or posteriorly by the lesion. Intraoperative use of the phase reversal technique or extraoperative mapping with a grid has been found to be helpful in planning the approach (either more anteriorly or posteriorly) as well as for guidance during the resection. The same has been found true for gliomas lying superficially or on the mesial hemispheric cortex. Even if the surgeon thinks he or she is staying strictly within the tumor, he or she can reach the healthy surrounding tissue and unknowingly affect the motor tracts. Because removal of tumor tissue is indirectly exerting traction forces on the surrounding healthy brain, MEP or SEP alterations may occur shortly before one reaches the motor fibers. Particularly dangerous is a situation where one reaches a sulcus 2 or 2.5 cm below the surface with the chance of injuring a vessel that supplies the motor cortex. In our experience, it has been impossible to avoid all new motor deficits, and in a strict scientific way we have not even tried to ensure that the rate of new motor deficits is lower with monitoring compared to surgery without monitoring. In essence, the number of amplitude deteriorations and complete potential losses that have been observed have convinced us that monitoring of MEPs is useful. This is particularly true for insular gliomas. No insular glioma will be operated on without monitoring. The schematic relationship given in Table 15.5 has thus far been consistent. If some evoked potentials remain at the end of

Intraoperative Neurophysiological Mapping and Monitoring

393

surgery, then there is a very good chance that the motor deficit is either not present or, if it has been present, it will recover. It remains more doubtful whether MEP monitoring is necessary with meningeomas bulging into the brain substance. It may help with the preservation of ACA branches that have been pushed away by the meningeoma, but usually the dissection of the border between meningeoma and brain is not so difficult. In insular gliomas we have repeatedly seen cases where vasospasm led to loss of potentials and recovery following application of papaverine. We are convinced that the removal of large type B insular gliomas would have been associated with a much higher morbidity than in our previously published paper [134]. Since that study we have altered our strategy and no longer operate on grade IV gliomas in older patients. With this new policy the incidence of motor deficits has decreased significantly, and it appears that the use of MEP monitoring has contributed to this. We continue the resection following the outline of the tumor (now with the use of a neuronavigation system) as long as there is no significant reduction in amplitude of evoked potentials and as long as there is no loss of the potentials (even in the deeper sections of the tumor cavity). If loss of potentials occurs, we apply papaverine to the vessels and wait a certain amount of time. If there is no recovery (making it a transient loss), resection is stopped. We frequently find the posterior part (2–3 cm) of the insular cortex infiltrated by the tumor left behind. Encouraged by our experiences with surgery for drug-resistant epilepsy, in certain centrally located gliomas without drug-resistant epilepsy we have started to apply preoperative mapping of motor, sensory, and speech function. In these cases the procedure is explained to the patient, and if he or she agrees to the second procedure of grid implantation, one can approach the tumor with a precise map of functional areas as it relays to the cortical surface. Transposition of the map is facilitated by the use of digital photography of the brain surface before and after grid application. In benign lesions, such as cavernomas, MEP and SEP monitoring is used if the lesions are close to the motor tract. These cases are rare, but they have occurred, and MEP and SEP monitoring was able to provide good feedback to the surgeon. Therefore, the general attitude in our department is that benign tumors should not be removed without combined MEP and SEP monitoring. An interesting case has been published of a midbrain cavernoma that was approached through a transventricular subforniceal approach and was removed with monitoring of MEPs and SEPs. Obviously, brainstem cavernomas are also monitored with both modalities [142]. With approximately 350 MEP monitoring cases, including spinal surgeries and infratentorial surgeries, we have gained enough experience to develop a feeling for the application of MEP monitoring and for criteria for reacting to MEP changes.

394

Georg Neuloh and Johannes Schramm

6.5 CENTRALLY LOCATED TUMORS—SUMMARY Neurophysiological mapping has corrected the presurgical estimation of the central sulcus location in 12% of the patients. It also helps identify the motor stimulation point for monitoring MEPs. In a series of 140 operations for central and insular lesions, most cases being gliomas and a minority being meningiomas, metastases, and other lesions, important observations could be made. In the vast majority of patients operated on, there were no new motor deficits. Many new deficits were related to supplementary motor area syndrome; thus they were totally reversible. Of the permanent new deficits, around 10% were severe, and around 29% occurred in muscle groups that were not included in the monitoring setup. Therefore, if only monitored muscle groups or limbs are considered, only about 9% new permanent deficits were observed in this group of central and insular space-occupying lesions. Considering the MEP results in this group, 58% had no significant change in potentials. In the group without significant changes, continuously stable potentials were seen in 75%, the rest showing only variations in amplitude. In the group with significant changes, irreversible loss occurred in 9% of the patients and reversible loss in 17%. The rest were reversible or irreversible alteration of potentials only. Of the group with reversible potential loss, permanent deficits occurred in only 10% of patients and transient deficits in 45%. In all patients in which irreversible MEP loss occurred, however, long lasting new motor deficits were observed postoperatively. In one third of cases with irreversible MEP alterations, transient deficits occurred, but permanent deficits were more frequent in this group (see Table 15.5).

ACKNOWLEDGMENTS We thank M. Kurthen, M.D., Department of Epileptology, Bonn, for help with the section on the extraoperative mapping technique in epilepsy, and U. Pechstein, M.D., for his help in designing some figures while he was a coworker in the monitoring group in Bonn. Thanks also to all former coworkers from the Bonn monitoring group: M. Taniguchi, M.D.; U. Pechstein, M.D.; J. Zentner, M.D.; and C. Cedzich, M.D.

REFERENCES 1. Tamaki, T., Yamashita, T., Kobayashi, H., et al. (1972). Spinal cord monitoring. Jpn. J. Electroencephalogr. Electromyograph., 1, 196. 2. Croft, T.J., Brodkey, J.S., and Nulsen, F.E. (1972). Reversible spinal cord trauma: A model for electrical monitoring of spinal cord function. J. Neurosurg., 36, 402–406.

Intraoperative Neurophysiological Mapping and Monitoring

395

3. Schramm, J., Mokrusch, T., Fahlbusch, R., et al. (1985). [Intra- and perioperative acoustic evoked brain stem potentials in cerebellopontile angle operations]. Hno, 33, 495–498. 4. Schramm, J., Mokrusch, T., Fahlbusch, R., et al. (1988). Detailed analysis of intraoperative changes monitoring brain stem acoustic evoked potentials. Neurosurgery, 22, 694–702. 5. Schramm, J., Watanabe, E., Strauss, C., et al. (1989). Neurophysiologic monitoring in posterior fossa surgery: I. Technical principles, applicability and limitations. Acta Neurochir., 98, 9–18. 6. Schramm, J., Hashizume, K., Fukushima, T., et al. (1979). Experimental spinal cord injury produced by slow, graded compression: Alterations of cortical and spinal evoked potentials. J. Neurosurg., 50, 48–57. 7. Schramm, J., and Hashizume, K. (1977). Somatosensory evoked potentials (SEP) in patients with peripheral, spinal and supraspinal lesions of the sensory system. Adv. Neurosurg., 4, 250–256. 8. Schramm, J. (1985). Spinal cord monitoring: Current status and new developments. Cent. Nerv. Syst. Trauma, 2, 207–227. 9. Huk, W.J., and Gademann, G. (1984). Magnetic resonance imaging (MRI): Method and early clinical experiences in diseases of the central nervous system. Neurosurg. Rev., 7, 259–280. 10. Homma, S., and Tamaki, T., Eds. (1984). “Fundamentas and clinical application of spinal cord monitoring.” Saikon, Tokyo. 11. Mizoi, K., and Yoshimoto, T. (1991). Intraoperative monitoring of the somatosensory evoked potentials and cerebral blood flow during aneurysm surgery—Safety evaluation for temporary vascular occlusion. Neurol. Med. Chir. (Tokyo), 31, 318–325. 12. Little, J.R., Lesser, R.P., and Luders, H. (1987). Electrophysiological monitoring during basilar aneurysm operation. Neurosurgery, 20, 421–427. 13. Holland, N.R. (1998). Subcortical strokes from intracranial aneurysm surgery: Implications for intraoperative neuromonitoring. J. Clin. Neurophysiol., 15, 439–446. 14. Krieger, D., Adams, H.P., Albert, F., et al. (1992). Pure motor hemiparesis with stable somatosensory evoked potential monitoring during aneurysm surgery: Case report. Neurosurgery, 31, 145–150. 15. Schramm, J., Koht, A., Schmidt, G., et al. (1990). Surgical and electrophysiological observations during clipping of 134 aneurysms with evoked potential monitoring. Neurosurgery, 26, 61–70. 16. Schramm, J., Zentner, J., and Pechstein, U. (1994). Intraoperative SEP monitoring in aneurysm surgery. Neurol. Res., 16, 20–22. 17. Schramm, J., Romstöck, J., Thurner, F., et al. (1985). Variance of latency and amplitude in SEPs monitored during spinal operations with and without cord manipulation. In “Spinal cord monitoring” ( J. Schramm, and S.J. Jones, eds.), pp. 186–196. Springer-Verlag, Berlin, Heidelberg, New York. 18. Schramm, J., Romstock, J., and Watanabe, E. (1986). [Intraoperative monitoring of the spinal cord. Personal results and an inventory of the current status]. Z. Orthop. Ihre. Grenzgeb., 124, 726–731. 19. Schramm, J., and Watanabe, E. (1988). [Possibilities and limits of intraoperative functional assessment of the spinal cord]. Klin. Wochenschr., 66, 19–26. 20. Deletis, V. (1993). Intraoperative monitoring of the functional integrity of the motor pathways. In “Advances in neurology: Electrical and magnetic stimulation of the brain” (O. Devinsky, A. Beric, and M. Dogali, eds.), pp. 201–214. Raven Press, New York. 21. Herdmann, J., Deletis, V., Edmonds, H.L., Jr., et al. (1996). Spinal cord and nerve root monitoring in spine surgery and related procedures. Spine, 21, 879–885. 22. Kothbauer, K., Deletis, V., and Epstein F.J., (1997). Intraoperative spinal cord monitoring for intramedullary surgery: An essential adjunct. Pediatr. Neurosurg., 26, 247–254.

396

Georg Neuloh and Johannes Schramm

23. Mauguiere, F. (1999). Utility of somatosensory evoked potentials (SEPs) in spinal cord lesions and functional surgery of pain and spasticity. Electroencephalogr. Clin. Neurophysiol. Suppl., 50, 31–39. 24. Schramm, J., Jones, S.J., eds. (1985). “Spinal cord monitoring” Springer-Verlag, Berlin, Heidelberg, New York. 25. Schramm, J., Møller, A.R., eds. (1991). “Intraoperative neurophysiologic monitoring in neurosurgery.” Springer-Verlag, Berlin, Heidelberg, New York. 26. Møller, A. (1995). “Intraoperative neurophysiologic monitoring.” Harwood Academic Publishers, Luxembourg. 27. Forbes, H.J., Allen, P.W., Waller, C.S., et al. (1991). Spinal cord monitoring in scoliosis surgery: Experience with 1168 cases. J. Bone Joint Surg. Br., 73, 487–491. 28. Nuwer, M.R., Dawson, E.G., Carlson, L.G., et al. (1995). Somatosensory evoked potential spinal cord monitoring reduces neurologic deficits after scoliosis surgery: Results of a large multicenter survey. Electroencephalogr. Clin. Neurophysiol., 96, 6–11. 29. Guerit, J.M., Witdoeckt, C., Verhelst, R., et al. (1999). Sensitivity, specificity, and surgical impact of somatosensory evoked potentials in descending aorta surgery. Ann. Thorac. Surg., 67, 1943–1946; discussion 1953–1958. 30. Galla, J.D., Ergin, M.A., Lansman, S.L., et al. (1999). Use of somatosensory evoked potentials for thoracic and thoracoabdominal aortic resections. Ann. Thorac. Surg., 67, 1947–1952; discussion 1953–1958. 31. Wan, I.Y., Angelini, G.D., Bryan, A.J., et al. (2001). Prevention of spinal cord ischaemia during descending thoracic and thoracoabdominal aortic surgery. Eur. J. Cardiothorac. Surg., 19, 203–213. 32. Berenstein, A., Young, W., Ransohoff, J., et al. (1984). Somatosensory evoked potentials during spinal angiography and therapeutic transvascular embolization. J. Neurosurg., 60, 777–785. 33. Sako, K., Nakai, H., and Kawata, Y., et al. (1998). Temporary arterial occlusion during anterior communicating or anterior cerebral artery aneurysm operation under tibial nerve somatosensory evoked potential monitoring. Surg. Neurol., 49, 316–322; discussion 322–323. 34. Sako, K., Nakai, H., Takizawa, K., et al. (1995). [Aneurysm surgery using temporary occlusion under SEP monitoring]. No Shinkei Geka, 23, 35–41. 35. Matsuda, M., Kidooka, M., Nakazawa, T., et al. (1991). Intraoperative monitoring of somatosensory evoked potentials in patients with cerebral aneurysm: Correlation between central conduction time and postoperative neurological status. Neurol. Med. Chir. (Tokyo), 31, 13–17. 36. Friedman, W.A., Chadwick, G.M., Verhoeven, F.J., et al. (1991). Monitoring of somatosensory evoked potentials during surgery for middle cerebral artery aneurysms. Neurosurgery, 29, 83–88. 37. McPherson, R.W., Niedermeyer, E.F., Otenasek, R.J., et al. (1983). Correlation of transient neurological deficit and somatosensory evoked potentials after intracranial aneurysm surgery: Case report. J. Neurosurg., 59, 146–149. 38. Buchthal, A., Belopavlovic, M., and Mooij, J.J. (1988). Evoked potential monitoring and temporary clipping in cerebral aneurysm surgery. Acta Neurochir., 93, 28–36. 39. Mooij, J.J., Buchthal, A., and Belopavlovic, M. (1987). Somatosensory evoked potential monitoring of temporary middle cerebral artery occlusion during aneurysm operation. Neurosurgery, 21, 492–496. 40. Lopez, J.R., Chang, S.D., and Steinberg, G.K. (1999). The use of electrophysiological monitoring in the intraoperative management of intracranial aneurysms. J. Neurol. Neurosurg. Psychiatry, 66, 189–196. 41. Symon, L. (1980). The relationship between CBF, evoked potentials and the clinical features in cerebral ischaemia. Acta. Neurol. Scand. Suppl., 78, 175–190.

Intraoperative Neurophysiological Mapping and Monitoring

397

42. Hargadine, J.R., Branston, N.M., and Symon, L. (1980). Central conduction time in primate brain ischemia: A study in baboons. Stroke, 11, 637–642. 43. Coyer, P.E., Simeone, F.A., and Michele, J.J. (1987). Latency of the cortical component of the somatosensory evoked potential in relation to cerebral blood flow measured in the white matter of the cat brain during focal ischemia. Neurosurgery, 21, 497–502. 44. Momma, F., Wang, A.D., and Symon, L. (1987). Effects of temporary arterial occlusion on somatosensory evoked responses in aneurysm surgery. Surg. Neurol., 27, 343–352. 45. Burke, D., Nuwer, M.R., Daube, J., et al. (1999). Intraoperative monitoring: The International Federation of Clinical Neurophysiology. Electroencephalogr. Clin. Neurophysiol. Suppl., 52, 133–148. 46. Mauguiere, F., Allison, T., Babiloni, C., et al. (1999). Somatosensory evoked potentials: The International Federation of Clinical Neurophysiology. Electroencephalogr. Clin. Neurophysiol. Suppl., 52, 79–90. 47. Taniguchi, M., Nadstawek, J., Pechstein, U., et al. (1992). Total intravenous anesthesia for improvement of intraoperative monitoring of somatosensory evoked potentials during aneurysm surgery. Neurosurgery, 31, 891–897; discussion 897. 48. Mauguiere, F., and Restuccia, D. (1991). Inadequacy of the forehead reference montage for detecting abnormalities of the spinal N13 SEP in cervical cord lesions. Electroencephalogr. Clin. Neurophysiol., 79, 448–456. 49. Jones, S.J. (1993). Somatosensory evoked potentials: I. Methodology, generators and special techniques. II. Clinical observations and applications. In “Evoked potentials in clinical testing” (A.M. Halliday, ed.), pp. 383–466. Churchill Livingstone, Edinburgh. 50. Cakmur, R., Towle, V.L., Mullan, J.F., et al. (1997). Intra-operative localization of sensorimotor cortex by cortical somatosensory evoked potentials: From analysis of waveforms to dipole source modeling. Acta Neurochir., 139, 1117–1124. 51. Lueders, H., Lesser, R.P., Hahn, J., et al. (1983). Cortical somatosensory evoked potentials in response to hand stimulation. J. Neurosurg., 58, 885–894. 52. Stohr, M., Dichgans, J., Voigt, K., et al. (1983). The significance of somatosensory evoked potentials for localization of unilateral lesions within the cerebral hemispheres. J. Neurol. Sci., 61, 49–63. 53. Zentner, J., and Rohde, V. (1994). SEP and MEP in comatose patients. Neurol. Res., 16, 89–92. 54. Zentner, J., and Rohde, V. (1992). The prognostic value of somatosensory and motor evoked potentials in comatose patients. Neurosurgery, 31, 429–434. 55. Chiappa, K.H., and Hill, R.A. (1998). Evaluation and prognostication in coma. Electroencephalogr. Clin. Neurophysiol., 106, 149–155. 56. Carter, B.G., and Butt, W. (2001). Review of the use of somatosensory evoked potentials in the prediction of outcome after severe brain injury. Crit. Care Med., 29, 178–186. 57. Branston, N.M., Strong, A.J., and Symon, L. (1977). Extracellular potassium activity, evoked potential and tissue blood flow. Relationships during progressive ischaemia in baboon cerebral cortex. J. Neurol. Sci., 32, 305–321. 58. Watanabe E., Schramm, J., Schneider, W. (1989). Effect of a subdural air collection on the sensory evoked potential during surgery in the sitting position. Electroencephalogr. Clin. Neurophysiol., 74, 194–201. 59. Kombos, T., Suess, O., Pietila, T., et al. (2000). Subdural air limits the elicitation of compound muscle action potentials by high-frequency transcranial electrical stimulation. Br. J. Neurosurg., 14, 240–243. 60. Cedzich, C., Schramm, J., Mengedoht, C.F., et al. (1988). Factors that limit the use of flash visual evoked potentials for surgical monitoring. Electroencephalogr. Clin. Neurophysiol., 71, 142–145. 61. Cedzich, C., and Schramm, J. (1990). Monitoring of flash visual evoked potentials during neurosurgical operations. Int. Anesthesiol. Clin., 28, 165–169.

398

Georg Neuloh and Johannes Schramm

62. Chacko, A.G., Babu, K.S., and Chandy, M.J. (1996). Value of visual evoked potential monitoring during trans-sphenoidal pituitary surgery. Br. J. Neurosurg., 10, 275–278. 63. Jones, N.S. (1997). Visual evoked potentials in endoscopic and anterior skull base surgery: A review. J. Laryngol. Otol., 111, 513–516. 64. Carter, L.P., Raudzens, P.A., Gaines, C., et al. (1984). Somatosensory evoked potentials and cortical blood flow during craniotomy for vascular disease. Neurosurgery, 15, 22–28. 65. Grundy, B.L., Nelson, P.B., Lina, A., et al. (1982). Monitoring of cortical somatosensory evoked potentials to determine the safety of sacrificing the anterior cerebral artery. Neurosurgery, 11, 64–67. 66. Paulsen, R.D., Steinberg, G.K., Norbash, A.M., et al. (1999). Embolization of rolandic cortex arteriovenous malformations. Neurosurgery, 44, 479–484; discussion 484–486. 67. Schaller, C., and Zentner, J. (1998). Vasospastic reactions in response to the transsylvian approach. Surg. Neurol., 49, 170–175. 68. Woolsey, C.N., Erickson, T.C., and Gilson, W.E. (1979). Localization in somatic sensory and motor areas of human cerebral cortex as determined by direct recording of evoked potentials and electrical stimulation. J. Neurosurg., 51, 476–506. 69. Goldring, S., and Gregorie, E.M. (1984). Surgical management of epilepsy using epidural recordings to localize the seizure focus: Review of 100 cases. J. Neurosurg., 60, 457–466. 70. Wood, C.C., Spencer, D.D., Allison, T., et al. (1988). Localization of human sensorimotor cortex during surgery by cortical surface recording of somatosensory evoked potentials. J. Neurosurg., 68, 99–111. 71. Berger, M.S., Kincaid, J., Ojemann, G.A., et al. (1989). Brain mapping techniques to maximize resection, safety, and seizure control in children with brain tumors. Neurosurgery, 25, 786–792. 72. Cedzich, C., Taniguchi, M., Schafer, S., et al. (1996). Somatosensory evoked potential phase reversal and direct motor cortex stimulation during surgery in and around the central region. Neurosurgery, 38, 962–970. 73. Allison, T. (1982). Scalp and cortical recordings of initial somatosensory cortex activity to median nerve stimulation in man. Ann. N. Y. Acad. Sci., 388, 671–678. 74. Berger, M.S. (1995). Functional mapping-guided resection of low-grade gliomas. Clin. Neurosurg., 42, 437–452. 75. Berger, M.S. (1996). Minimalism through intraoperative functional mapping. Clin. Neurosurg., 43, 324–337. 76. Ojemann, G., Ojemann, J., Lettich, E., et al. (1989). Cortical language localization in left, dominant hemisphere: An electrical stimulation mapping investigation in 117 patients. J. Neurosurg., 71, 316–326. 77. Ebeling, U., and Huber, P. (1992). Localization of central lesions by correlation of CT findings and neurological deficits. Acta Neurochir., 119, 17–22. 78. Ojemann, G.A. (1997). Intraoperative methods. In “Epilepsy: A comprehensive textbook” ( J. Engel, and T.A. Pedley, eds.), pp. 1777–1783. Lippincott-Raven, New York. 79. Jayakar, P., and Lesser, R.P. (1997). Extraoperative methods. In “Epilepsy: A comprehensive textbook” ( J. Engel, and T.A. Pedley, eds.) pp. 1785–1793. Lippincott-Raven, New York. 80. De Salles, A.A., Swartz, B.E., Lee, T.T., et al. (1994). Subdural recording and electrical stimulation for cortical mapping and induction of usual seizures. Stereotact. Funct. Neurosurg., 62, 226–231. 81. Schulz, R., Luders, H.O., Tuxhorn, I., et al. (1997). Localization of epileptic auras induced on stimulation by subdural electrodes. Epilepsia, 38, 1321–1329. 82. De Renzi, E., and Vignolo, L.A. (1962). The token test: A sensitive test to detect receptive disturbances in aphasics. Brain, 85, 665–678. 83. Hermann, B.P., Wyler, A.R., Somes, G., et al. (1994). Dysnomia after left anterior temporal lobectomy without functional mapping: Frequency and correlates. Neurosurgery, 35, 52–56; discussion 56–57.

Intraoperative Neurophysiological Mapping and Monitoring

399

84. Hermann, B.P., Wyler, A.R., and Somes, G. (1991). Language function following anterior temporal lobectomy. J. Neurosurg., 74, 560–566. 85. Lehericy, S., Duffau, H., Cornu, P., et al. (2000). Correspondence between functional magnetic resonance imaging somatotopy and individual brain anatomy of the central region: Comparison with intraoperative stimulation in patients with brain tumors. J. Neurosurg., 92, 589–598. 86. Wyllie, E., Luders, H., Murphy, D., et al. (1990). Intracarotid amobarbital (Wada) test for language dominance: Correlation with results of cortical stimulation. Epilepsia, 31, 156–161. 87. Binder, J.R., Swanson, S.J., Hammeke, T.A., et al. (1996). Determination of language dominance using functional MRI: A comparison with the Wada test. Neurology, 46, 978–984. 88. King, R.B., and Schell, G.R. (1987). Cortical localization and monitoring during cerebral operations. J. Neurosurg., 67, 210–219. 89. Kombos, T., Suess, O., Funk, T., et al. (2000). Intra-operative mapping of the motor cortex during surgery in and around the motor cortex. Acta Neurochir., 142, 263–268. 90. Ebel, H., Ebel, M., Schillinger, G., et al. (2000). Surgery of intrinsic cerebral neoplasms in eloquent areas under local anesthesia. Minim. Invasive Neurosurg., 43, 192–196. 91. Kombos, T., Suess, O., Kern, B.C., et al. (1999). Comparison between monopolar and bipolar electrical stimulation of the motor cortex. Acta Neurochir., 141, 1295–1301. 92. Berger, M.S., and Rostomily, R.C. (1997). Low grade gliomas: Functional mapping, resection strategies, extent of resection, and outcome. J. Neurooncol., 34, 85–101. 93. Yingling, C.D., Ojemann, S., Dodson, B., et al. (1999). Identification of motor pathways during tumor surgery facilitated by multichannel electromyographic recording. J. Neurosurg., 91, 922–927. 94. Skirboll, S.S., Ojemann, G.A., Berger, M.S., et al. (1996). Functional cortex and subcortical white matter located within gliomas. Neurosurgery, 38, 678–684; discussion 684–685. 95. Ebeling, U., Schmid, U.D., and Reulen, H.J. (1989). Tumour-surgery within the central motor strip: Surgical results with the aid of electrical motor cortex stimulation. Acta Neurochir., 101, 100–107. 96. Taniguchi, M., Cedzich, C., and Schramm, J. (1993). Modification of cortical stimulation for motor evoked potentials under general anesthesia: Technical description. Neurosurgery, 32, 219–226. 97. Penfield, W., and Boldrey, E. (1937). Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation. Brain, 60, 389–443. 98. Berger, M.S., Deliganis, A.V., Dobbins, J., et al. (1994). The effect of extent of resection on recurrence in patients with low grade cerebral hemisphere gliomas. Cancer, 74, 1784–1791. 99. Ammirati, M., Vick, N., Liao, Y.L., et al. (1987). Effect of the extent of surgical resection on survival and quality of life in patients with supratentorial glioblastomas and anaplastic astrocytomas. Neurosurgery, 21, 201–206. 100. Ciric, I., Ammirati, M., Vick, N. et al. (1987). Supratentorial gliomas: Surgical considerations and immediate postoperative results. Gross total resection versus partial resection. Neurosurgery, 21, 21–26. 101. Roberts, T.P., Ferrari, P., Perry, D., et al. (2000). Presurgical mapping with magnetic source imaging: Comparisons with intraoperative findings. Brain Tumor. Pathol., 17, 57–64. 102. Horikoshi, T., Omata, T., Uchida, M., et al. (2000). Usefulness and pitfalls of intraoperative spinal motor evoked potential recording by direct cortical electrical stimulation. Acta Neurochir., 142, 257–262. 103. Kakimoto, M., Inoue, S., Sakamoto, T., et al. (2000). [Intraoperative motor evoked potential monitoring: A review of 115 cases]. Masui, 49, 240–244. 104. Lee, W.Y., Hou, W.Y., Yang, L.H., et al. (1995). Intraoperative monitoring of motor function by magnetic motor evoked potentials. Neurosurgery, 36, 493–500. 105. Nagle, K.J., Emerson, R.G., Adams, D.C., et al. (1996). Intraoperative monitoring of motor evoked potentials: A review of 116 cases. Neurology, 47, 999–1004.

400

Georg Neuloh and Johannes Schramm

106. Calancie, B., Harris, W., Broton, J.G., et al. (1998). “Threshold-level” multipulse transcranial electrical stimulation of motor cortex for intraoperative monitoring of spinal motor tracts: Description of method and comparison to somatosensory evoked potential monitoring. J. Neurosurg., 88, 457–470. 107. Meylaerts, S.A., Jacobs, M.J., van Iterson, V., et al. (1999). Comparison of transcranial motor evoked potentials and somatosensory evoked potentials during thoracoabdominal aortic aneurysm repair. Ann. Surg., 230, 742–749. 108. Amassian, V.E., Stewart, M., Quirk, G.J., et al. (1987). Physiological basis of motor effects of a transient stimulus to cerebral cortex. Neurosurgery, 20, 74–93. 109. Gorman, A.L. (1966). Differential patterns of activation of the pyramidal system elicited by surface anodal and cathodal cortical stimulation. J. Neurophysiol., 29, 547–564. 110. Barker, A.T., Jalinous, R., and Freeston, I.L., (1985). Non-invasive magnetic stimulation of human motor cortex. Lancet, 1, 1106–1107. 111. Merton, P.A., and Morton, H.B. (1980). Stimulation of the cerebral cortex in the intact human subject. Nature, 285, 227. 112. Zentner, J. (1991). Motor evoked potential monitoring during neurosurgical operations on the spinal cord. Neurosurg. Rev., 14, 29–36. 113. Zentner, J. (1989). Noninvasive motor evoked potential monitoring during neurosurgical operations on the spinal cord. Neurosurgery, 24, 709–712. 114. Ghaly, R.F., Stone, J.L., Lee, J.J., et al. (1999). Monitoring spinal cord motor and somatosensory evoked potentials in anesthetized primates. Neurol. Res., 21, 359–367. 115. Levy, W.J., and York, D.H. (1986). Evoked potential monitoring of the motor system. Biomed. Sci. Instrum., 22, 51–54. 116. Boyd, S.G., Rothwell, J.C., Cowan, J.M., et al. (1986). A method of monitoring function in corticospinal pathways during scoliosis surgery with a note on motor conduction velocities. J. Neurol. Neurosurg. Psychiatry, 49, 251–257. 117. Burke, D., Hicks, R.G., and Stephen, J.P. (1990). Corticospinal volleys evoked by anodal and cathodal stimulation of the human motor cortex. J. Physiol., 425, 283–299. 118. Pechstein, U., Cedzich, C., Nadstawek, J., et al. (1996). Transcranial high-frequency repetitive electrical stimulation for recording myogenic motor evoked potentials with the patient under general anesthesia. Neurosurgery, 39, 335–343; discussion 343–344. 119. Rodi, Z., Deletis, V., Morota, N., et al. (1996). Motor evoked potentials during brain surgery. Pflugers. Arch., 431, R291–292. 120. Taylor, B.A., Fennelly, M.E., Taylor, A., et al. (1993). Temporal summation: The key to motor evoked potential spinal cord monitoring in humans. J. Neurol. Neurosurg. Psychiatry, 56, 104–106. 121. Cedzich, C., Pechstein, U., Schramm, J., et al. (1998). Electrophysiological considerations regarding electrical stimulation of motor cortex and brain stem in humans. Neurosurgery, 42, 527–532. 122. Amassian, V.E., Quirk, G.J., and Stewart, M. (1990). A comparison of corticospinal activation by magnetic and electrical stimulation of monkey motor cortex. Electroencephalogr. Clin. Neurophysiol., 77, 390–401. 123. Deletis, V., and Kothbauer, K. (1998). Intraoperative neurophysiology of the corticospinal tract. In “Spinal cord monitoring” (E. Staalberg, H.S. Sharma, Y. Olsson, eds.) pp. 421–444, Springer, Wien, New York. 124. Ubags, L.H., Kalkman, C.J., Been, H.D., et al. (1996). The use of a circumferential cathode improves amplitude of intraoperative electrical transcranial myogenic motor evoked responses. Anesth. Analg., 82, 1011–1014. 125. Jones, S.J., Harrison, R., Koh, K.F., et al. (1996). Motor evoked potential monitoring during spinal surgery: Responses of distal limb muscles to transcranial cortical stimulation with pulse trains. Electroencephalogr. Clin. Neurophysiol., 100, 375–383.

Intraoperative Neurophysiological Mapping and Monitoring

401

126. Agnew, W.F., and McCreery, D.B. (1987). Considerations for safety in the use of extracranial stimulation for motor evoked potentials. Neurosurgery, 20, 143–147. 127. Deletis, V., Isgum, V., and Amassian, V.E. (2001). Neurophysiological mechanisms underlying motor evoked potentials in anesthetized humans: Part 1. Recovery time of corticospinal tract direct waves elicited by pairs of transcranial electrical stimuli. Clin. Neurophysiol., 112, 438–444. 128. Granit, R., Kernell, D., and Lamarre, Y. (1966). Synaptic stimulation superimposed on motoneurones firing in the “secondary range” to injected current. J. Physiol., 187, 401–415. 129. Burke, D., and Hicks, R.G. (1998). Surgical monitoring of motor pathways. J. Clin. Neurophysiol., 15, 194–205. 130. Katayama, Y., Tsubokawa, T., Maejima, S., et al. (1988). Corticospinal direct response in humans: Identification of the motor cortex during intracranial surgery under general anaesthesia. J. Neurol. Neurosurg. Psychiatry, 51, 50–59. 131. Padberg, A.M., and Bridwell, K.H. (1999). Spinal cord monitoring: Current state of the art. Orthop. Clin. North. Am., 30, 407–433, viii. 132. Pereon, Y., Bernard, J.M., Fayet, G., et al. (1998). Usefulness of neurogenic motor evoked potentials for spinal cord monitoring: Findings in 112 consecutive patients undergoing surgery for spinal deformity. Electroencephalogr. Clin. Neurophysiol., 108, 17–23. 133. Woodforth, I.J., Hicks, R.G., Crawford, M.R., et al. (1996). Variability of motor-evoked potentials recorded during nitrous oxide anesthesia from the tibialis anterior muscle after transcranial electrical stimulation. Anesth. Analg., 82, 744–749. 134. Zentner, J., Meyer, B., Stangl, A., et al. (1996). Intrinsic tumors of the insula: A prospective surgical study of 30 patients. J. Neurosurg., 85, 263–271. 135. McCreery, D.B., Agnew, W.F., Yuen, T.G., et al. (1990). Charge density and charge per phase as cofactors in neural injury induced by electrical stimulation. IEEE Trans. Biomed. Eng., 37, 996–1001. 136. Majkowski, J. (1999). Kindling: Clinical relevance for epileptogenicity in humans. Adv. Neurol., 81, 105–113. 137. Browning, J.L., Heizer, M.L., and Baskin, D.S. (1992). Variations in corticomotor and somatosensory evoked potentials: Effects of temperature, halothane anesthesia, and arterial partial pressure of CO2. Anesth. Analg., 74, 643–648. 138. Gugino, V., and Chabot, R.J. (1990). Somatosensory evoked potentials. Int. Anesthesiol. Clin., 28, 154–164. 139. Pechstein, U., Nadstawek, J., Zentner, J., et al. (1998). Isoflurane plus nitrous oxide versus propofol for recording of motor evoked potentials after high frequency repetitive electrical stimulation. Electroencephalogr. Clin. Neurophysiol., 108, 175–181. 140. Lang, E.W., Beutler, A.S., Chesnut, R.M., et al. (1996). Myogenic motor-evoked potential monitoring using partial neuromuscular blockade in surgery of the spine. Spine, 21, 1676–1686. 141. Kalkman, C.J., Drummond, J.C., Kennelly, N.A., et al. (1992). Intraoperative monitoring of tibialis anterior muscle motor evoked responses to transcranial electrical stimulation during partial neuromuscular blockade. Anesth. Analg., 75, 584–789. 142. Cedzich, C., Pechstein, U., Zentner, J., et al. (1999). Minimally invasive stereotacticallyguided extirpation of brain stem cavernoma with the aid of electrophysiological methods. Minim. Invasive. Neurosurg., 42, 41–43. 143. Schramm, J., and Taniguchi, M. (1991). Value of stable and changing somatosensory evoked potentials (SSEP) during aneurysm surgery. In “Intraoperative neurophysiologic monitoring in neurosurgery” ( J. Schramm, and A.R. Møller, eds.), pp. 151–161. Springer-Verlag, Berlin, Heidelberg, New York.

CHAPTER

16

Neurophysiological Monitoring During Neurosurgery for Movement Disorders JAY L. SHILS Division of Intraoperative Neurophysiology and Department of Neurosurgery, Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York

MICHELE TAGLIATI Department of Neurology, Beth Israel Medical Center, New York

RON L. ALTERMAN Department of Neurosurgery, Hyman-Newman Institute for Neurology and Neurosurgery, Beth Israel Medical Center, New York

1 Introduction 2 History and Theory 2.1 Historical Notes 2.2 Theoretical Basis for Surgery in the Basal Ganglia 2.3 Modern Movement Disorder Surgery: General Overview 3 Operating Room Environment and Basic Equipment 3.1 Operating Room 3.2 Recording Electrodes 3.3 Amplification 3.4 Stimulation 4 Technique for Movement Disorder Surgery 4.1 General Stereotactic Technique 5 Conclusion References ABSTRACT During stereotactic procedures for the treatment of medically refractory movement disorders, intraoperative neurophysiology shifts its focus from simply monitoring the effects of surgery to actually guiding the surgeon’s actions. The small size, poor visualization, and physiological nature of these deep brain targets compel the surgeon to rely on some Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

405

406

Jay L. Shils, Michele Tagliati, and Ron L. Alterman form of physiological confirmation of proper anatomical targeting, whether a neuroablative or deep brain stimulation procedure is being performed. This chapter reviews the history of movement disorder surgery; the current physiological rationale for surgical intervention in cases of medically refractory Parkinson’s disease, dystonia, and essential tremor; and the various neurophysiological monitoring methods employed during these so-called functional neurosurgical procedures. Finally, the authors discuss the most common surgeries currently employed and detail their preferred surgical techniques.

1 INTRODUCTION The full potential of intraoperative neurophysiology is realized during the performance of so-called functional neurosurgical procedures. During these interventions therapeutic lesions or stimulating electrodes are stereotactically placed within deep brain structures to treat movement disorders such as Parkinson’s disease (PD), essential tremor (ET), dystonia, affective disorders, and chronic neuropathic pain. The deep location of these structures precludes direct surgical approaches. Instead, surgeons rely on a combination of image-guided stereotactic techniques and intraoperative neurophysiology to place the therapeutic lesions or stimulating electrodes with acceptable accuracy and safety. Unlike tumors, which are relatively large and easily identified on CT or MRI, functional neurosurgical targets typically are small and poorly visualized with current imaging modalities. Moreover, because these are physiologic as much as anatomic targets, image-based targeting may incompletely identify the desired location. Consequently, intraoperative recording and stimulation techniques have been developed to aid target localization. These techniques complement anatomical targeting by providing real-time electrophysiological data concerning probe position and the surgical target. The surgeon and physiologist use these data to “fine-tune” their anatomic targeting before completing the therapeutic intervention. Thus employed, intraoperative neurophysiology does not simply monitor surgical activity; it guides it. In this chapter, we provide an historic overview of intraoperative monitoring for movement disorder surgery and a detailed account of our approach to these surgeries, which has evolved over the course of more than 500 interventions.

2 HISTORY AND THEORY

2.1 HISTORICAL NOTES 2.1.1 Surgery for Movement Disorders Sir Victor Horsely is reported to have performed the first neurosurgical procedure for a movement disorder when, in the late 1800s, he resected part of the precentral gyrus in a patient with athetoid movements. The surgery halted the abnormal movements but caused dyspraxia and paralysis of the limb [1].

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

407

The first successful basal ganglia surgery is credited to Meyers [2–4], who reported improvement in a patient with postencephalitic parkinsonism in 1939. Prior to this landmark report, surgery within the basal ganglia was avoided because it was believed that human consciousness resided in these structures. Despite the high mortality rates (10–12%) that plagued these “open” procedures (i.e., via craniotomy) [2–5], Meyers demonstrated the potential benefits of basal ganglia surgery and opened the door for the application of less invasive stereotactic approaches to these deep brain structures. He also provided the first accounts of human basal ganglia physiology, describing the frequency, phase, and amplitude of neuronal signals from the striatum, pallidum, corpus callosum, internal capsule, subcallosal bundle, and dorsal thalamus in patients with and without movement disorders [3, 4, 6]. Meyers quickly realized the potential value of the accumulated data, which he ultimately employed to help localize specific deep brain structures during movement disorder surgery. Robert Clarke designed the first stereotactic frame in 1908 [7]. His frame employed skull landmarks to target deep brain structures in small animals, a technique that could not be translated to clinical use because of the more varied and complex shape of the human skull and brain. Consequently, it was not until 1947, after the introduction of ventriculography, that Spiegel and Wycis performed the first human stereotactic surgeries, for psychiatric illness [8] and Huntington’s chorea [9]. In following years a number of human stereotactic atlases were published, and standard meridia (e.g., the intercommissural line) from which stereotactic coordinates could be determined were established. Effective targets for stereotactically guided neuroablation were discovered empirically. For example, Cooper stumbled upon the beneficial effects of globus pallidus lesioning by accidentally ligating the anterior choroidal artery of a PD patient [10] while performing a pedunculotomy. He later adopted stereotactic approaches to pallidal lesioning, reporting favorable results and reduced surgical mortality rates (∼3%) as compared to open procedures [11–14]. Laitinen described how Leksell further improved the results of pallidotomy by placing the lesion more posteriorly and ventrally within the internal segment of the globus pallidus (GPi) [15], that portion of the nucleus that we now know is responsible for sensorimotor processing [16, 17]. In 1963, Spiegel et al. [18] described campotomy, in which the fibers of the pallidofugal, rubrothalamic, corticofugal, and hypothalamofugal pathways are interrupted within the H fields of Forel. They reported promising results in 25 patients with tremor and 25 with rigidity. In the end, however, thalamotomy emerged as the most commonly performed movement disorder procedure in the pre-levodopa era because of the consistent tremor control it provided. Though most surgery for PD ceased after the introduction of levodopa in 1967, small numbers of thalamotomies were performed for medically refractory tremor during the next 25 years, until the reintroduction of Leksell’s pallidotomy by Laitinen et al. in 1992 [15].

408

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

2.1.2 Neurophysiology and Movement Disorder Surgery Most early electrophysiologic studies of the human thalamus and basal ganglia were performed with macroelectrode techniques that yielded relatively crude, EEG-like responses [19–23]. Electrodes and recording techniques were refined over subsequent decades, culminating in the development of single-cell microelectrode recording. Of note is the work of Albe-Fessard, who refined microelectrode techniques for experimental purposes and paved the way for their intraoperative use [24, 25]. It was her belief that micro-electrode recording (MER) would “provide a powerful tool in improving stereotactic localization and that it would furthermore reduce the risk due to anatomical variability” [25]. In recent years, Madame Albe-Fessard’s vision has been realized as MER has gained in popularity and ready-to-use recording systems have become commercially available. The history of electrical brain stimulation begins with Fritsch and Hitzig, who in 1870 elicited limb movement in dogs by stimulating the frontal cortex, and then defined the limits of the motor area electrophysiologically [26]. Intraoperative cortical stimulation studies by Penfield and colleagues from the late 1920s through the late 1940s contributed seminal information concerning the somatotopic organization of the cerebral cortex by defining the motor and sensory “homunculi.” In 1950, Spiegel et al. described the use of stimulation during surgery at the H fields of Forel to both “test the position of the electrode and to avoid proximity to the corticospinal pathways ventrally, the sensory thalamic-relay nuclei dorsally, and the third nucleus posteriorly” [18]. Other neurophysiological techniques, such as impedance monitoring [20, 27, 28] and evoked potential recordings [29–33] also have been employed as localization tools; however, these techniques serve predominantly as adjuncts to recording and stimulation. Perhaps the most significant advance in functional neurosurgery in the last decade has been the introduction of chronic electrical stimulation (termed “deep brain stimulation” or DBS) as a therapeutic alternative to neuroablation. Deep brain stimulation provides three potential advantages when compared to neuroablation: 1. DBS is reversible. If stimulation induces an unwanted side-effect, one simply turns the stimulator off or adjusts parameters. Thus the risk of permanent adverse neurological events is reduced. 2. Stimulation parameters may be customized to each patient, potentially enhancing therapeutic efficacy. 3. Access to the surgical target is maintained via the implanted electrode and programmable pulse generator. Therefore, therapy may be modified over time through simple stimulation adjustments, potentially increasing the longevity of response.

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

409

FIGURE 16.1 A three-dimensional artist’s rendition of the structures involved in surgery for movement disorders. The light greenish blue structure on the left is the globus pallidus (GPi and GPe). The large grey structure on the right is the thalamus, and the small dark green structure is the subthalamic nuclei (STN). The medial edge of the STN is only 6.0 mm from the midline of the brain. With the trajectories that our group uses in the operating room, we encounter around 10.0 mm of GPi, 11.0 mm of VIM, and 5.0 mm of STN. Modified from [117] (see also color plate).

Thus far, two studies that compared thalamic DBS to thalamotomy for the treatment of tremor have been published. Both studies found DBS to be the superior treatment modality in large part because of the ability to adjust stimulation parameters in the event of symptom recurrence. Presently, movement disorder surgery is focused on three structures: the ventrolateral (VL) nucleus of the thalamus, the globus pallidus pars internus (GPi), and the subthalamic nucleus (STN) (Fig. 16.1) (see also color plate). Each of these structures can be targeted for ablation in procedures that are, respectively, termed thalamotomy [26, 34–47], pallidotomy [18, 33, 39, 46, 48–86], and subthalamotomy [87, 88]. Alternatively, each can be targeted for chronic electrical stimulation [33, 47, 52, 77, 89–115]. The choice of target is based largely on clinical diagnosis and the symptoms to be treated. The interested reader is directed to a detailed account of our patient and target selection criteria [116].

410

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

2.2 THEORETICAL BASIS FOR SURGERY IN THE BASAL GANGLIA Our current understanding of the functional organization of the basal ganglia and PD pathophysiology is based predominantly on data derived from the study of primates with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-induced Parkinsonism [118–123]. Microelectrode techniques also have contributed greatly to this body of knowledge. Though incomplete, the current model of basal ganglia function is partly responsible for the rebirth of movement disorder surgery, providing a scientific basis for selecting those deep brain structures that are currently targeted for therapeutic interventions. The model is depicted in Fig. 16.2. The basal ganglia are composed of two principal input structures (the corpus striatum and the STN), two output structures (GPi and substantia nigra pars reticulata [SNr]), and two intrinsic nuclei (external segment of the globus pallidus [GPe] and substantia nigra pars compacta [SNc]) [124]. Five parallel basal ganglia-thalamo-cortical circuits (motor, oculomotor,

FIGURE 16.2 Diagrammatic representation of the basal ganglia circuit, showing the direct and indirect pathways proposed by DeLong and colleagues [121, 122]. The light grey lines represent excitatory pathways, and the darker lines show inhibitory pathways.

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

411

two prefrontal, and limbic) have been described [123]. While surgical interventions target the motor circuit, it is likely that lesioning and stimulation also impact other circuits as well. The corpus striatum, which is composed of the caudate and putamen, is the largest nuclear complex of the basal ganglia. The striatum receives excitatory (glutamatergic) input from several areas of the cerebral cortex as well as inhibitory input from the dopaminergic cells of the SNc. Cortical and nigral inputs are received via the “spiny” neurons. One subset of these cells projects directly to the GPi, forming the “direct pathway,” while another subset projects to the GPe, the first relay station of a complementary “indirect pathway,” that passes through the STN before terminating at GPi. The antagonistic actions of the direct and indirect pathways regulate the neuronal activity of GPi, which, in turn, provides inhibitory input to the pedunculopontine nucleus (PPN) and the VL nucleus of the thalamus. The VL nucleus projects back to the primary and supplementary motor areas [125, 126], completing the cortico-ganglio-thalamo-cortical loop. The direct pathway inhibits GPi, resulting in a net disinhibition of the motor thalamus and facilitation of the thalamo-cortical projections. The indirect pathway, via its serial connections, provides excitatory input to the GPi, inhibiting the thalamo-cortical motor pathway. In PD, loss of dopaminergic input to the striatum leads to a functional reduction of direct pathway activity and a facilitation of the indirect pathway. These changes result in a net increase in GPi excitation and a concomitant hyperinhibition of the motor thalamus. The excessive inhibitory outflow from GPi reduces the thalamic output to supplementary motor areas that are critical to the normal execution of movement. This model accounts well for the negative symptoms of PD (i.e., rigidity and bradykinesia) and supports both GPi and STN as rational targets for surgically treating PD. The model is incomplete, however, because it does not fully account for hyperkinetic features of PD such as tremor and levodopa-induced dyskinesias, physiological phenomena that are poorly understood. Tremor activity is consistently detected in the VL nucleus of patients with PD or ET, and the VL nucleus continues to be the primary surgical target for treating medically refractory tremor. However, it is unclear if the motor thalamus is the primary generator of tremor activity or merely participates in the transmission of tremor-generating signals. Moreover, the evidence that both pallidotomy [52, 60] and STN DBS [96] also control parkinsonian tremor suggests that intervention at many points within the tremor-generating loop may suppress this symptom. Levodopa-induced dyskinesias (LIDs) are involuntary movements of the limbs or trunk that are temporally associated with levodopa administration [127, 128].

412

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

These movements are typically choreiform or dystonic in nature and are easily distinguished from the tremor of PD. Pharmacodynamic factors related to chronic exogenous dopaminergic stimulation probably play a fundamental role in levodopa-induced dyskinesia. According to the model, pallidotomy should worsen LID by reducing pallidal inhibition of the VL nucleus, a hypothesis that is supported by the experimental observation that STN lesions, which reduce the excitatory output from STN to GPi, cause dyskinesias in primates that are indistinguishable from LID [129]. On the contrary, LID is the most responsive symptom to pallidotomy, a consistently observed phenomenon [130]. It has been hypothesized that sensitization of dopamine receptors by exogenously administered levodopa may cause aberrant neuronal firing patterns with consequent disruption of the normal flow of information to the thalamus and the cortical motor areas [131]. It follows that pallidotomy may improve LID by disrupting this aberrant flow.

2.3 MODERN MOVEMENT DISORDER SURGERY: GENERAL OVERVIEW There is no one best method for performing movement disorder surgery. Rather, stereotactic surgeons modify general approaches to target localization to suit their personal preferences and to take advantage of their institution’s strengths. Currently accepted technique involves frame-based anatomical localization supported by intraoperative physiological confirmation of proper targeting. An overview of the various anatomic and physiologic techniques currently in use follows.

2.3.1 Anatomical Targeting Techniques In the pre-levodopa era, positive contrast and air ventriculography were employed to localize the foramen of Monro and the anterior and posterior commissures. The stereotactic coordinates of therapeutic targets were then determined [98, 132] based on their relationship to these structures as described in various stereotactic atlases. Targeting accuracy was therefore limited by the inaccuracies of these atlases, which were typically generated from just one or a few specimens whose true dimensions were distorted by formalin fixation and by anatomical distortions created by the intraventricular injection of air or contrast. Today, CT- and MRI-based techniques, which demonstrate the brain parenchyma noninvasively, have supplanted ventriculography as the primary means of anatomically localizing stereotactic targets. Nevertheless,

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

413

ventriculography is still employed by many stereotactic surgeons and therefore remains an important technique. The introduction of CT [133] revolutionized the diagnosis and treatment of neurologic diseases and encouraged changes in stereotactic frame design, expanding the uses of frame-based stereotaxis to include tumor biopsy and resection [134, 135]. Soon after the introduction of MRI, Leksell et al. [136] demonstrated its applicability to stereotactic systems. MRI provides superior resolution as compared to CT, as well as multiplanar images [137, 138] with minimal frame-related artifact [137]. Nonreformatted MRI beautifully demonstrates the commissures, the thalamus, and most basal ganglia structures [138]. These features permit direct stereotactic localization of the surgical target in some instances [104, 139–143]; however, indirect targeting, based on accurate localization of the commissures, may still yield the most reliable target coordinates [143]. The most significant drawback to targeting with MRI is the potential for image distortion introduced by nonlinearities within the magnetic field [144]. Distortions can be generated by a number of factors, including the presence of ferromagnetic objects within the field, imperfections in the scanner’s magnets, and, most commonly, patient movement [138, 144, 145]. Walton et al. demonstrated that targeting errors are greater in the periphery than in the center of the magnetic field and stereotactic space [146, 147]. MRI distortion may also be related to the pulse sequence(s) employed. For example, it has been suggested that fast spin-echo inversion recovery sequences resist imaging distortions secondary to magnetic susceptibility better than other image acquisition methods [148, 149]. In contrast to MRI, CT maintains linear accuracy, thereby reducing imageinduced targeting errors [138, 150]. However, metallic artifact can impede visualization of the commissures, CT tissue resolution is inferior to MRI, and axial images alone are provided. Commercially available targeting software packages can fuse CT and MRI images, but to our knowledge there are no studies to suggest that such image fusion techniques improve targeting accuracy.

2.3.2 Physiological Targeting: Recording Techniques The four most commonly employed techniques for physiologic localization during movement disorder surgery are: (1) impedance measurements; (2) macroelectrode recordings and stimulation; (3) semimicroelectrode recording (and/or stimulation); and (4) microelectrode recording (with or without stimulation). Evoked potentials have also been employed at times [29–33], but at present these are primarily used as an adjunct to stimulation during thalamic interventions.

414

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

2.3.2.1 Impedance Techniques Changes in electrical impedance can accurately demarcate the boundaries of neural structures and may therefore be used to define the borders of a surgical target [27, 28]. Impedance measurements can be performed with monopolar electrodes that are referenced to the scalp [57, 81] or with concentric bipolar electrodes employing the outer ring as the reference [56, 57, 152]. Employing a test frequency of 1 KHz, impedances of 400 Ω or greater are recorded in the deep grey matter while white matter can be greater or less depending upon orientation. The major advantages of this technique are the ease with which it is performed and the fact that the same electrode can be used during both the localization and the lesioning phases of an ablative procedure. The major disadvantage is the relative crudeness of the physiological information provided. Moreover, impedance measures may not adequately distinguish borders between adjacent nuclei and work best when there are clear grey matter–white matter boundaries to be defined. Therefore, impedance recordings primarily are used for the localization of large white matter bundles and nuclear groups [41, 152]. We perform impedance measurements during ablative procedures only after the final target is selected via microelectrode recording (see following text), and simply to ensure that the lesioning electrode has not strayed from its desired trajectory and is located within grey matter. 2.3.2.2 Macroelectrode Recording Macroelectrode (ME) recording, defined as any low-impedance (1–100 kΩ) recording that generates either multiunit potentials or neural background noise, provides somewhat more detailed physiologic information as compared to impedance measurements [20, 21, 41, 153]. The electrode tip may be as small as 50 µm and may be configured in a bipolar concentric fashion with an intertip distance of 200–300 µm, or as a single active tip referenced to the cortical surface via the insertion cannula or to the scalp via a surface electrode. The main advantage of ME recording is the ease and speed with which data are collected as compared to microelectrode techniques. The obvious disadvantage is that EEG-like field potentials lack the discrimination necessary to characterize single-unit firing features within the surgical target (Fig. 16.3). Consequently, physiologic detail regarding the surgical target is lacking. 2.3.2.3 Semi-Microelectrode Technique Electrodes that have small tip diameters (<50 µm) and impedances of 100–500 kΩ are referred to as semi-microelectrodes. These electrodes provide more detailed information than do macroelectrodes, but they still do not yield

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

415

FIGURE 16.3 This is a poor semi-microelectrode recording from a substantia nigra pars reticulata cell. Note the multiple amplitude activity and the depth of EEG quality. This cell was recorded from an electrode that had an impedance of around 50 kΩ. The diameter of the electrode was around 50 µm (5 s epoch).

FIGURE 16.4 Three semi-microelectrode recordings in which single units can be distinguished. What differentiates these from pure microelectrode recordings is the fact that they contain more than one clearly distinguishable unit (5 s epoch).

single-unit recordings. Semi-microelectrodes detect the responses of a few cells (∼10−100) (Fig. 16.4) localized to a small area around the recording tip (∼10− 100 µm). These so-called field potentials are more refined than the EEG-like recordings provided by macroelectrodes but lack the detail provided by microelectrode techniques.

416

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

2.3.2.4 Microelectrode Techniques Microelectrodes provide the most detailed picture of the neural elements encountered during movement disorder surgery [20, 34, 38, 53, 71, 73, 74, 102, 111, 154–159]. Microelectrode tips have diameters of 1–40 µm and impedances of ∼1 MΩ. By recording individual neuronal activity (Fig. 16.5), microelectrodes provide real-time information concerning the physiological characteristics of the recorded neuron and thereby the nucleus within which the cell is located. The major drawback to microelectrode recordings is the time and expertise required to perform the technique well. The sophisticated electronics equipment is expensive and must be maintained expertly. Thus the investment in machinery and personnel can be prohibitive to some centers. It is sometimes difficult to acquire a useful signal because of electrical noise in the operating room, and even in the best circumstances, recording tracts may take 20–40 min to complete. Finally, interpreting single-cell recordings is a skill that is mastered only with experience and patience. It is our experience, however, that, once mastered, microelectrode recording can be performed efficiently and yields invaluable data concerning electrode position. For example, Alterman et al. demonstrated that in 12% of 132 consecutive pallidotomies, final lesion placement, as guided by microelectrode recording, was more than 4 mm removed from the site that was originally selected by the surgeon based on stereotactic MRI [68]. This distance is considered significant, since it is equivalent to the diameter of the typical pallidotomy lesion.

FIGURE 16.5 A set of microelectrode recordings. Note that only a single unit is being recorded. Each spike has relatively the same amplitude and shape (5 s epoch).

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

417

3 OPERATING ROOM ENVIRONMENT AND BASIC EQUIPMENT

3.1 OPERATING ROOM 3.1.1 Electrical Noise (recording only) It is difficult to record low-amplitude neural signals reliably in the electrically harsh operating room environment, which can affect even more robust, easily recorded signals, such as the EKG. Anesthesia equipment, electric cautery, lighting, radios, telemetry equipment, and countless other electronic devices can all negatively impact recording quality. While the surgical team can control the use of these devices within their own operating room, external electrical influences, such as ongoing construction, poor wiring, and the use of large pieces of equipment in adjacent operating rooms, may also erode recordings. In order to control for these external influences fully, movement disorder surgery procedures ideally should be performed in an electrically shielded operating room. Of course, few facilities possess such an expensive facility, so we make the following recommendations: 1. Minimize any stray electrical switching noises. Typically, this type of noise derives from two sources: lighting fixtures that are equipped with dimmers and poorly shielded computer equipment. In our experience, a properly grounded recording head stage can be operated with minimal switching interference when the dimmers are set either all the way on or all the way off. Fluorescent lighting may also interfere with the recording equipment, but such 60-Hz signals are attenuated easily with a notch filter. Computer monitors should be fitted with static screen covers that can be grounded. If the monitor is part of the recording system, it can be grounded to the common system ground. Otherwise, it should be grounded to the operating room grounding system. 2. Employ battery-powered anesthesia and monitoring equipment. Alternatively, position anesthesia equipment in such a way as to reduce electrical interference. Turn down audible indicators. One can reduce cross-talk by keeping monitoring and neural recording cables on opposite sides of the patient. Newer anesthesia systems are equipped with cathode ray tube (CRT), liquid crystal display (LCD), and/or plasma displays, the electromagnetic (EM) leakage from which can be bothersome. If the interference from such monitors becomes overpowering, a simple aluminum foil shield can be placed between the monitor and the recording stage and connected to the system ground. 3. Turn off and unplug all electrical equipment that is not in use during recording. Electric cautery, electrically controlled operating tables, and patient

418

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

warmers generate very powerful electromagnetic radiation. Fortunately, these devices are not necessary during recording and can be unplugged. 4. Employ isolated power supplies for the recording equipment. Electrical equipment used in adjacent operating rooms may interfere with recording due to poor operating room wiring schemes. Employing isolated power supplies and grounded EM shields can minimize this interference. Proper planning will help minimize most sources of noise, but noise will occur despite the most prudent planning. It is important that both the surgeon and the neurophysiologist are prepared for these occasional frustrations. The patient should also be informed of the possibility of delays during the surgery should electrical noise be encountered. Taking the aforementioned preventive steps minimizes the risk of encountering noise and provides a framework from which one can troubleshoot noise problems when they occur.

3.1.2 Electrical Noise (internal system influences) Sources of electrical noise from within the recording system include: (1) the microelectrode transducer, which detects the neural activity; (2) the preamplifier, which is located close to the recording structure; (3) the amplifier; (4) signal conditioners; (5) the visual display; and (6) auditory processors (Fig. 16.6). However, electrical noise primarily enters the system proximal to the first stage of the preamplifier. The amplitude of the recorded signals is small (range: 100 µV to 100 mV) so that failure of any real-time component can severely compromise the integrity of the signal and, in turn, the accuracy of the mapping. Poorly designed equipment is the most common cause of intrasystem noise; poor system maintenance is second. Connectors must be cleaned or replaced regularly to combat oxidation, particularly in high-humidity environments. Cables must also be inspected regularly and replaced when worn.

3.2 RECORDING ELECTRODES Lenz [36, 37] has previously described the construction of recording microelectrodes, and Geddes [160] provides a useful description of electrode properties. Microelectrode tips may be composed of a number of materials, including stainless steel and tungsten, but the authors prefer the platinum-iridium etched tip, which is glass coated. The tip diameter ranges from 1 to 40 µm and is beveled to a maximum diameter of 350–400 µm. The tip is coated with a thin

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

419

FIGURE 16.6 A representation of the signal flow through the intraoperative recording system. The microelectrode (or transducer) converts the cellular chemical potentials to a pure electrical signal that is then passed though the amplification system. From there the data pass through a digitizer or audio processing system. The data are then displayed on a computer, amplified and played through audio speakers, and stored for off-line analysis.

layer of glass to make the maximum diameter between 400 and 450 µm. The electrode tip is connected to a stainless steel wire (diameter: 500 µm) and a glass soldered bead. Alternatively, Epoxylite (The Epoxylite Corporation, St. Louis, MO) is used to seal the junction. An outer insulating sheath is placed over the stainless steel wire, making the total shaft diameter 600–700 µm. An electrode (including the tip) is typically around 300 mm in length. The last 15–20 mm of insulation is removed in order to connect the electrode to the amplifier. The electrodes exhibit a low-frequency roll-off below 1000 Hz (Fig. 16.7). The resulting reduction in transmitted power (frequency range: 100–2000 Hz) can be as much as 17.9 dB [161]. Even though cellular firing rates range from 5 to 500 Hz, it is the high-frequency components that are most important for auditory discrimination. The microelectrodes exhibit adequate response characteristics in these higher frequencies. Semi-microelectrodes are usually made of either stainless steel or tungsten with tip diameters of less than 50 µm; however, tip impedance and geometry impact recording discrimination (i.e., field potentials vs. single unit recordings) more than tip diameter [160]. Semi-microelectrodes are technically easier to

420

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

FIGURE 16.7 The gain versus frequency of the recording system. The recording system acts as a high-pass filter. Below 1000 Hz there is a reduction in the system’s gain. This reduction is acceptable because most of the spike energy is contained in the higher frequencies of the spike. Reprinted from [161].

produce than microelectrodes because they can be made from existing fine wire, while microelectrode tips must be electrolytically etched.

3.3 AMPLIFICATION The preamplifier is the first active component of the recording system. Either referential or differential amplification techniques are employed to measure voltage variations at both the active and referential inputs. Referential amplifiers reference the active input to a second input that is either located far from the active input and/or has a larger surface area than the active input. The variations measured by the active input are independent of the relatively inactive reference input, permitting discrimination of the true signal. In reality, large amplitude signals in the reference electrode may conceal smaller voltage variations at the active electrode, masking signal. This possibility should be kept in mind if extensive noise is observed on the recording display. Another disadvantage of referential recording is the possibility of amplifying noise that is mistakenly interpreted as signal. Differential amplification is superior in this regard.

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

421

Differential amplifiers use two active inputs and electrically subtract signals that are common to both. The transmission cables of both inputs run to the amplifier side by side. The amplifier receives two of the same input signal, but one is 180° out of phase from the other (i.e., one is positive while the other is negative). The two active signals are then subtracted. Noise that is externally induced on the transmission cable is subtracted out since the same noise is theoretically induced in both the transmission and reference cables. Differences in the signals are also accentuated (SA1 + N − (−SA1 + N) = 2 SA1). The commonmode-rejection-ratio (CMRR) defines the ability of a differential amplifier to exclude common input noise. The larger this number, the greater is the reduction of the induced cable noise. We employ a differential amplifier (Model MDA4I, Atlantic Research Systems, Inc., Dothan, AL) for intraoperative single-unit recording. The ground and the active input are interconnected on a large ground plane to minimize voltage variations, which are typically close to zero. The ground plane includes the head stage, the cerebral cortex, and the base of the isolated amplifier. Isolation is important for safety. These amplifiers must have a very high input impedance (∼200 MΩ) to enhance signal transfer from the high-impedance electrode (∼1 MΩ). Considering the electrode and amplifier as a voltage divider, we see that the voltage at the amplifier is determined by the equation: V AMP =

VIN (200 ) = 0.995 VIN 200 + 1

As the amplifier input impedance approaches the electrode impedance, signal transfer decreases. The amplifier output impedance is 10 Ω. The amplifier consists of two sections: the preamplifier and a built-in impedance test circuit. The preamplifier attaches to the head stage and serves not only as the first stage of the amplification section but also as a switchbox that is used to switch between recording, stimulation, and impedance testing modes. Since the preamplifier is isolated from the main amplifier by an optical connection, the preamplifier is powered from two 9-V batteries that are located in the main amplifier. Amplification control is possible via the main amplifier. The gain of the entire amplification system varies from 100 to 10,000 times with a CMRR of 80 dB at 1000 Hz. The noise floor level of the system is 5 µV when the inputs are shorted. The maximum input to the amplifier is ±15 V, while the maximum linear output is 20 Vp-p. The amplifier has built-in high- and low-pass variable single-pole filters (range: 1–500 Hz and 1–10 kHz, respectively). The second component of the amplifier is a built-in impedance test circuit. This circuit passes a 30-nA (max.) current through the electrode to ground and has a range of 10 kΩ to 5 MΩ. Our standard settings are as follows: gain: approximately 4000 times; high pass filter: 100 Hz; low pass filter: 10 kHz.

422

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

3.4 STIMULATION A number of stimulation techniques may also be performed during movement disorder surgery. Stimulation may be delivered via macro- or microelectrodes and may be used either to assess proximity to structures one wishes to avoid (e.g., internal capsule, optic tract) or to assess the potential clinical effects of chronic stimulation. Many microelectrode recording systems allow the surgical team to switch between recording and stimulation modes. This permits direct comparison of recording and stimulation data; however, stimulation leads to a more rapid degradation of the microelectrode, so a new electrode may be required for each recording tract. Moreover, the volume of tissue that can be affected with microelectrode stimulation is so small that gross clinical changes are rarely observed with this technique, in our experience. We therefore prefer to stimulate with macroelectrodes, employing either the Radionics (Burlington, MA) stimulator and lesion generator prior to performing a neuroablation, or the DBS lead itself when performing a DBS procedure. Single- and dual-channel “screener boxes” (Fig. 16.8) are commercially available for this purpose (Models 3625 [single lead] and 3628 [dual lead]; Medtronics Inc., Minneapolis, MN).

FIGURE 16.8 The Medtronic’s screener boxes. The unit on the left is a dual channel stimulator and allows for testing two leads simultaneously. These devices are used in the operating room to test the location of the DBS electrode before final implantation. The screener boxes can also be used with the lead externalized while the patient is in the hospital. This gives the movement disorder team time to test parameters without permanently implanting the whole system.

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

423

During stimulation, a train of impulses is passed through the region of interest and the clinical effects are noted. The stimulus can be delivered in either a mono- or biphasic fashion. A monophasic stimulus varies from the reference by the signal amplitude and then returns to the reference. The rate of change can be edge, ramp, or sinusoidal in nature. A biphasic stimulus varies from the reference in both the positive and negative directions. Typically, the amplitude of the change is the same in both directions, but this is not always the case. The Medtronics, Inc. implantable neural stimulators generate a biphasic pulse with a positive component that is less intense than the negative component. Stimulation may also be mono- or bipolar in nature. Monopolar stimulation is generated at the active tip and is referenced to some distant point. With bipolar stimulation, the active and reference electrodes are in close proximity so that current flows within a tightly defined space. The concentric ring electrode is a commonly employed bipolar stimulation configuration where the inner tip is the active electrode and the outer ring is the reference electrode. Chronically implanted DBS leads are equipped with four contacts arranged in series, allowing for either mono- or bipolar stimulation employing any one or combination of contacts. In order to deliver a monopolar stimulus, the active contact(s) is(are) referenced to the pulse generator case. Bipolar stimuli are conducted between any combination of contacts. Table 16.1 demonstrates some of the important specifications for stimulators.

TABLE 16.1

Stimulator Specifications

Feature

First type

Second type

Output Polarity

Bi-Phasic – Deviations in both the positive and negative directions from the reference point

Mono-Phasic – Single deviation from the reference point

Constant Measure

Constant Current – The current of the device is set by the user, and the stimulator adjusts the voltage to compensate for impedance deviations

Constant Voltage – The voltage of the device is set by the user, and the stimulator adjusts the current to compensate for the impedance deviations.

Pulse Width

The width of each pulse

Frequency

The number of pulses per second

Train Length

The time that the stimulator presents a set of pulses

Amplitude

The strength of the stimulus

Wave Shape

The type of waveform. Most stimulators used for these procedures generate square pulses.

424

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

4 TECHNIQUE FOR MOVEMENT DISORDER SURGERY A detailed account of our surgical technique is beyond the scope of this book. Instead, we will provide a brief overview of our general technique for performing movement disorder surgery and detail the physiological localization methods we employ for each surgical target. The interested reader is directed to a more detailed description of our preferred technique [131].

4.1 GENERAL STEREOTACTIC TECHNIQUE The stereotactic headframe is applied on the morning of surgery with local anesthetic (Fig. 16.9). Care is taken to center the head within the frame and to align the base ring of the frame with the orbitomeatal line, which approximates the orientation of the AC-PC line. In this way, axial images obtained perpendicular to the axis of the frame will run parallel to the AC-PC plane. The patient is transferred to radiology, where a stereotactic MRI is performed. We employ

FIGURE 16.9 The stereotactic frame with the MRI localizer box. The plastic box is used to add coordinate points the surgeon can use to locate objects in the frame’s three-dimensional space.

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

425

axial fast spin-echo inversion recovery MRI to localize the commissures and determine their stereotactic coordinates. We then derive the coordinates of the midcommissural point (MCP) by averaging the coordinates of the commissures and calculate the coordinates of our surgical target based on its relationship to the commissures and/or the MCP. The calculations employed for the most commonly targeted sites are given in Table 16.2. The patient is returned to the operating room (Fig. 16.10 shows the room layout that we employ at our center) and is positioned supine on the operating table, which is configured as a reclining chair for the patient’s comfort. The target

TABLE 16.2

Initial Target Coordinates

Target

Medial lateral coordinate

Anterior-posterior coordinate

Ventral-dorsal coordinate

GPi

20–23 mm from midline

2–3 mm anterior to MCP

VIM

13–15 mm from midline

5–6 mm anterior to PC

0 mm from AC-PC

STN

12 mm from midline

2 mm posterior to MCP

6 mm ventral to AC-PC

FIGURE 16.10 noise.

6 mm ventral to AC-PC

Layout of our operating room. This particular setup has been found to minimize

426

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

coordinates are set on the frame, bringing the presumptive target to the center of the operating arc. The operation is performed through a 14-mm burr hole that is positioned approximately 1 cm anterior to the coronal suture and 2–3 cm lateral of the midline. The dura mater is opened and microelectrode recording is begun. The microdrive adapter and the X-Y adjustment stage are mounted onto the operating arc. The microelectrode is back-loaded into the microdrive and zeroed to the guide tube. The electrode is withdrawn into the cannula (∼5 mm) for safe insertion. An insertion cannula is advanced through the frontal lobe to a point that is 20 mm anterosuperior to the presumptive target. The guide tube containing the recording electrode is inserted into the insertion cannula and the microdrive apparatus is mounted to the X-Y adjustment stage. At this point the guide tube, to the end of which the electrode tip position is zeroed, is flush with the end of the insertion cannula. Thus recording begins 20 mm anterosuperior to the presumptive target. The electrode is driven 3.0 mm into the brain and the impedance of the electrode–tissue system is measured. In our experience, impedances of 700 KΩ to 1.2 MΩ provide the best single-unit recordings. Even with conditioning of the electrode and stimulation testing, these starting impedances allow for sufficient current passage without degradation of the recording electrode surface. If there is a large impedance drop following electrode conditioning, the electrode is deemed unacceptable and is replaced. We correct any noise problems at this time and then proceed to data acquisition. At the conclusion of each recording trajectory, the collected data are mapped onto scaled sagittal sections derived from the Schaltenbrand-Wahren stereotactic atlas [164], and a determination is made as to tract location and orientation employing a “best fit” model (see data organization section). When the data suggest that our targeting is correct, we proceed either to test stimulation and ablation or DBS lead insertion. Detailed discussions of microelectrode recording and macroelectrode stimulation, as we employ them for each of the three primary movement disorder targets, are described in the following sections. 4.1.1 GPi Procedures1 Posteroventral pallidotomy and GPi deep brain stimulation are reported to improve tremor, rigidity, and LID in patients with medically refractory, moderately advanced PD. Though the published experience is limited, preliminary 1 To hear single-unit examples: (1) go to Chapter 16 from the main menu; (2) select the target of interest (GPi, VIM, STN) with the left mouse button; (3) select the single-unit example with the left mouse button.

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

427

results suggest that GPi stimulation yields results that are similar to pallidotomy, with the added benefit that bilateral stimulation can be performed more safely than bilateral pallidotomy. Profound improvements have also been reported in patients with DYT1associated primary dystonia in whom GPi stimulation was performed. The authors have performed seven of these procedures, noting dramatic improvements in tone, posture, and overall motor function. Of course, further study is required before the full benefit of this surgery in primary and secondary dystonias is known. Successful pallidal interventions require targeting of the sensorimotor region of GPi, which lies posterior and ventral in the nucleus. When recording in this region, three key nuclear structures must be recognized: the striatum, the GPe, and the GPi (Fig. 16.11) (see also color plate). Our typical trajectory passes at a 60–70° angle above the horizontal of the AC-PC line, and at a medial-lateral angle of 90° (i.e., true vertical). By employing this purely parasagittal trajectory, we can more readily fit the operative recording data to the parasagittal sections provided in human stereotactic atlases. The first cells encountered during recording are in the corpus striatum (caudate and putamen; colored blue in Fig. 16.11). They exhibit characteristic

FIGURE 16.11 Sagittal slice through the globus pallidus, taken 20.0 mm from the midline. The color shading is referenced in the text. Reprinted from [164] (see also color plate).

428

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

FIGURE 16.12 Representative tracings of cellular activity that may be encountered during a GPi recording trajectory. Each tracing is 5 s in length, except for trace GG, which is 1 s in length. (A) (Sound 1) Low frequency, and sparse single spikes of the striatum. (B) (Sound 2) Boarder cell. (C) (Sound 3) GPe pauser cell. (D) (Sound 4) GPe burster cell. (E) (Sound 5) The X-cell represents a cell that is dying. (F) (Sound 6) A GPi tremor cell. (G and GG) (Sound 7) A high-frequency cell from GPi. (H) (Sound 8) The entry of the microelectrode into the optic tract. The point at which the amplitude starts to increase represents the optic tract entry.

low-amplitude action potentials, which sound like corn popping (Fig. 16.12A, CD-GPi sound 1). Cellular activity in this area is extremely scanty, and the background is quiet. The electrode may also traverse some quiet regions that represent small fingerlike projections of the internal capsule into the striatum. Either the detection of a border cell or an increase in background activity marks entry into the GPe, the next structure to be encountered. Border cells (Fig. 16.12B, CD-GPi sound 2) exhibit very low frequencies (between 2 and 20 Hz) that are highly periodic and high-amplitude spikes with moderate to wide firing times. Though rare in this region, border cells greatly facilitate localization of the boundaries within the globus pallidus. Two major cell types are found within the GPe: pausers (Fig. 16.12C, CDGPi sound 3) and bursters (Fig. 16.12D, CD-GPi sound 4). Pauser cells fire arrhythmically at a frequency of 30–80 Hz. They exhibit moderate to high amplitude discharges, a shorter time period, and lower amplitude than the

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

429

border cells. They are distinguishable by their staccato-type, asynchronous pauses. An extremely small number of pauser cells (<5%) may demonstrate somatotopically organized kinesthetic responses. As their name implies, burster cells are distinguished by short bursts of high-frequency discharges, achieving rates as high as 500 Hz. Amplitudes vary but are usually less than the amplitudes of the pauser cells. It is important to differentiate bursters from what we refer to as X-cells (Fig. 16.12E, CD-GPi sound 5). X-cells exhibit high-frequency discharges (near 500 Hz) with a time-related (<30 s) decrease in amplitude, representing death of the cell. We may encounter anywhere from 4 to 8 mm of GPe during one recording tract. Border cells are again encountered at the inferior border of GPe and are more plentiful in this region. A quiet laminar area (Fig. 16.11) is encountered upon exit from the GPe, marked by a steep dropoff in background activity. Border cells are again encountered upon entry into the GPi, and again, two classes of neurons predominate within the nucleus: tremor-related cells and high-frequency cells. Tremor cells (Fig. 16.12F, CD-GPi sound 6) fire rhythmically in direct relation to the patient’s tremor. Single-unit recordings show a frequency modulation pattern, while semi-microelectrode recordings show a frequency and amplitude modulation pattern. The firing rate of these cells is between 80 and 200 Hz. High-frequency cells (Fig. 16.12G, CD-GPi sound 7) are characterized by firing rates that are similar to the tremor cells (80–100 Hz), but are much more stable, exhibiting consistent amplitude and frequency. Many of these cells respond to active or passive range of motion of a specific joint or extremity. Guridi et al. have physiologically defined a somatotopic organization of the kinesthetic cells in the GPi, with the face and arm region located ventrolaterally and the leg dorsomedially [69]. Taha et al. found a slightly different arrangement, with the leg sandwiched centrally between the arm in both the rostral and caudal areas [74]. Vitek et al. have found the leg to be medial and dorsal with respect to the arm, and the face more ventral [16]. The GPi is subdivided into external and internal segments, labeled GPie (external GPi) and GPii (internal GPi), respectively. Both regions exhibit similar cellular recording patterns, but GPie may exhibit less cellularity than GPii. Total GPi recordings normally span from 5 to 12 mm. A steep dropoff in background activity denotes exit from the GPi inferiorly. Three important white matter structures border the GPi and may be encountered during recording. The ansa lenticularis (AL), which emerges from the base of the GPi, carries motor-related efferents from the GPi to the ventrolateral thalamus, merging with its sister pathway, the lenticular fasciculus at the H field of Forel. The AL is an electrically quiet region, although rare cells of relatively low amplitudes and firing frequencies can be recorded. It has been

430

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

proposed that lesioning within the AL generates the best results from posteroventral pallidotomy. The optic tract (OT) lies directly inferior to the AL (Fig. 16.11, CD-GPi sound 8), accounting for the high rate of visual field complications reported in the early modern pallidotomy literature [17, 60]. With quality recordings, it is possible to hear the microelectrode tip enter the OT, the sound of which is reminiscent of a waterfall. Upon hearing this background change, one may confirm entry into the optic tract by turning off the ambient lights and shining a flashlight in the patient’s eyes. This will increase the recorded signal if the electrode is within the OT. Finally, one may encounter the internal capsule. Background recordings within the capsule are similar to those of the OT. Movement of the mouth or contralateral hemibody will generate a swooshing sound that is correlated to the movement. Obviously, one wishes to avoid the posterior capsule when making a lesion or placing a DBS lead, since a hemiparesis or hemiplegia may result. Macroelectrode stimulation is performed prior to lesioning to ensure that the electrode is a safe distance from the internal capsule and the OT. We conduct test stimulation with the Radionics 1.1-mm by 3-mm exposed-tip stimulating and lesioning electrode (Radionics, Burlington, MA), employing a stimulation frequency of 60 Hz and a pulse width of 0.2 ms at 0–10 V. Stimulation of contralateral muscular contractions at less than 2.5 V suggests that the lesioning electrode is too close to the internal capsule and should be adjusted laterally. Induction of phosphenes at less than 2.0 V suggests that the electrode is too close to the OT and should be withdrawn slightly. Test stimulation should be performed at 2- to 3-mm intervals beginning 6–8 mm above the base of GPi as defined by MER. Decreasing voltage trends in the induction of muscular contractions and/or phosphenes should be monitored. If stimulation is begun inferiorly, one risks creating a tract through which current may leak, resulting in persistently low thresholds for the stimulation of phosphenes that will cause the lesioning probe to be withdrawn too far. A suboptimal lesion may result. Details of this technique have been published previously [76]. Employing this technique, one of the authors (RLA) has performed more than 110 pallidotomies without inducing visual field abnormalities or hemiparesis. If stimulation indicates that the targeted region is a safe distance from the internal capsule and OT, the therapeutic lesion is placed. Ablation begins at the base of the GPi and progresses upward in 2-mm increments, creating a cylindrical lesion that encompasses the span of GPi as defined by MER. A test lesion is initially performed at 40°C for 40 s, after which the patient’s visual fields and basic motor function are checked. If there are no adverse visual field or motor changes, a permanent lesion is performed at 80°C for 60 s. Ideally, lesions should not encroach upon the GPe, because the working model of basal ganglia physiology suggests that GPe lesioning may worsen parkinsonism.

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

431

Excellent pallidotomy results also have been reported without the use of microelectrode recording and with the performance of ablations of varying size ranges. To date, no correlation between lesion size and surgical outcome has been made.

4.1.2 VIM Procedures Therapeutic neuroablation or chronic high-frequency electrical stimulation within the ventral intermediate nucleus of the thalamus (VIM; Fig. 16.13) (see also color plate) suppresses parkinsonian and essential tremor without adversely affecting voluntary motor activity to a significant degree (thalamotomy may be associated with some loss of fine dexterity). Thalamic interventions are extremely gratifying to perform because of the immediacy of the results and the well-defined physiology of the motor and sensory thalamic nuclei [41]. When targeting VIM, our standard angles of approach are 60–70° relative to the AC-PC line, and 5–10° lateral of the true vertical. Pure parasagittal trajectories cannot be employed as they are in globus pallidus procedures due to the medial location of the target and a desire to avoid the ipsilateral lateral ventricle.

FIGURE 16.13 Sagittal slice through the thalamus taken 14.5 mm from the midline. The color shading is referenced in the text. Reprinted from [164] (see also color plate).

432

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

FIGURE 16.14 Representative tracings of cellular activity that may be encountered during a VIM recording trajectory. Each tracing is 5 s in length. (A) (Sound 9) Sparse dorsal thalamic cells. (B) (Sound 10) Nontremor VIM cell. (C) (Sound 11) VIM tremor cell. (D) (Sound 12) Nonsensory VC cell. (E) (Sound 13) Finger VC sensory cell. Note the increase in firing rate as a light bristle paint brush is dabbed against the finger.

Transit through the ventricle may increase the risk of hemorrhage and typically leads to more rapid loss of cerebral spinal fluid (CSF) with resulting brain shift and loss of targeting accuracy. Recording begins in the dorsal thalamus, where cells characterized by low amplitudes and sparse firing patterns are encountered. Bursts of activity and small-amplitude single spikes (Fig. 16.14A, CD-VIM sound 9) are typical findings in this region. Upon exiting the dorsal thalamus, the electrode enters the VL nucleus, which is composed of nucleus ventralis oralis anterior (VOA), ventralis oralis posterior (VOP), and VIM. The dorsal third of the VL nucleus is sparsely populated such that cellular recordings in this area are similar to those of the dorsal thalamus. As the electrode passes ventrally within the VL complex, cellular density increases and cells with firing rates of 40–50 Hz (Fig. 16.14B, CD-VIM sound 10) are encountered. Kinesthetic cells with discrete somatotopic representation are routinely encountered. This organization permits an assessment of the mediolateral position of the electrode. The homunculus of the

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

433

ventrocaudal (Vc) and VIM nuclei are virtually identical: representation of the contralateral face and mouth lies 9–11 mm lateral of midline, the arm is represented lateral to this at 13–15 mm lateral of midline, and the leg is more lateral still, adjacent to the internal capsule. Thus, if one encounters a cell that responds to passive movement of the ankle, one knows that one has targeted too laterally to treat an upper-extremity tremor and should adjust the mediolateral position accordingly. In addition to kinesthetic neurons, one will routinely encounter “tremor” cells (Fig. 16.14C, CD-VIM sound 11) within the VIM of tremor patients. These cells exhibit a rhythmic firing pattern that can be synchronized to EMG recordings of the patient’s tremor [37]. Lenz et al. demonstrated that these cells are concentrated within VIM, 2–4 mm above the AC-PC plane, a site that is empirically known to yield consistent tremor control [162]. The recording electrode may exit VIM inferiorly, passing into the zona incerta (ZI) with a resulting decrease in background signal, or it will enter Vc, the primary sensory relay nucleus of the thalamus. Entry into Vc is marked by a change in the background signal. Cells in this region are densely packed, exhibit high amplitudes, and respond to sensory phenomena (e.g., light touch) with a discreet somatotopic organization, which mirrors that of VIM and may also be used to assess target laterality (Fig. 16.14D, CD-VIM sound 12). A typical cell, which responds to lightly brushing the patient’s finger, is featured in Fig. 16.14E (CD-VIM sound 13). Note the increase in firing rate as a light bristle paintbrush is dabbed against the finger. The bars represent the times that the brush is being dabbed against the finger. If Vc is encountered early in the recording trajectory, the electrode may be targeted posteriorly and should be adjusted anteriorly. The nucleus ventrocaudalis parvocellularis (VCpc) rests inferiorly to Vc. Recordings within this nucleus are similar to those of Vc; however, stimulation in this location may yield painful or temperature-related sensations. Single-unit recordings in this area will respond to both painful and temperature-related stimuli applied within the cell’s receptive field. Stimulation within the thalamus for the purposes of localizing therapeutic lesions may be performed with constant-voltage or constant-current devices, and with micro- or macroelectrodes. When stimulating with constant current, we employ 60 µs and 1 ms pulse widths at a frequency of 180 Hz. Regardless of technique, the reference is a cautery ground pad that is placed on the back of the thigh ipsilateral to the side of the stimulation. We consider a motor stimulation threshold of 1 mA or 3 V safe for placing a thalamotomy lesion.2 2 Note that the threshold values described in this chapter are specific to the stimulators, stimulator parameters, and electrode geometries that we employ.

434

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

When performing VIM DBS, we use the lead itself to perform test stimulation. In such cases bipolar stimulation is performed so that a reference pad is unnecessary. In our experience, a properly positioned DBS lead results in tremor arrest at <3 V (pulse width: 60 µs; frequency: 180 Hz). Transient paresthesias are common with a properly positioned electrode; however, persistent paresthesias, which are induced at low voltages, indicate that the electrode is positioned posteriorly, near or within Vc. Failure to suppress tremor or induce paresthesias, even at 5 V, suggests that the electrode is positioned anteriorly within VOA. Muscular contractions (typically of the contralateral face and/or hand) suggest that the lead is positioned too laterally and stimulation is affecting the internal capsule. It has been our experience that microelectrode stimulation may not suppress tremor at sites where macroelectrode stimulation is effective.

4.1.3 STN Procedures Bilateral STN DBS appears to be the most effective treatment for PD since levodopa, which was introduced more than a generation ago. Subthalamic DBS improves all of the cardinal features of PD, dampens the severity of “on–off” fluctuations, alleviates freezing spells, and dramatically reduces medication requirements. The STN is approached at an angle of 70° relative to the AC-PC line and 10–15° lateral of the true vertical. Microelectrode recording begins in the anterior thalamus and passes sequentially through the ZI, Forel’s field H2, the STN, and the substantia nigra pars reticulata (SNr) (Fig. 16.15) (see also color plate). In the thalamus, one encounters cells that fire with low amplitude and frequency. Two patterns of activity may be identified: (1) bursts of activity (Fig. 16.16A, CD-STN sound 1) and; (2) irregular, low-frequency (1–30 Hz) activity (see Chapter 16, CD-video segment 1) (Fig. 16.16B, CD-STN sound 2). The density of cellular activity varies in this region. For example, we have observed that VOA is more cellular than the reticular thalamus. The border between the thalamus and ZI (Fig. 16.16C, CD-STN sound 3) may be very distinct, but not in all cases. Developmentally, the ZI is a continuation of the reticular nucleus of the thalamus, and the transition from one to the other may not be clear. The ZI can be differentiated electrophysiologically from the thalamus in two ways. First, cellular activity is more muffled or “muddy” in the ZI. By this we mean that the cellular firing rates slow and become a little more asynchronous, and the amplitudes decrease in intensity. These changes are subtle and can be missed by inexperienced observers. The second indication of transition from thalamus to ZI is a change in the background recordings.

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

435

FIGURE 16.15 Sagittal slice through the STN taken 12.0 mm from the midline. Reprinted from [164] (see also color plate).

Whereas the background of the thalamus proper is somewhat active, the ZI background is much quieter. Typically, the recording electrode will exit the thalamus 6–10 mm anterosuperior to our presumptive target and will pass through 2.5–4.0 mm of ZI before entering H2. If more than 4 mm of relative “quiet” is encountered, a trajectory that is anterior or posterior to the STN should be suspected. A decrease in background activity demarcates entry into Forel’s field H2, which lies immediately superior to the STN, 10–12 mm lateral of midline. Sparse cellular activity is detected over a span of 1–2 mm. Background activity increases as the recording electrode enters STN. Additionally, dense cellular activity is now encountered. Two patterns of cellular activity are observed within STN: (1) tremor activity (Fig. 16.16D, CD-STN sound 4) similar to that encountered in VIM or GPi; and (2) single-cell activity (Fig. 16.16E, CD-STN sound 5) with frequencies that vary from ∼25 Hz to 45 Hz. Cells in the dorsal segments of the STN exhibit slower firing rates than those of the ventral STN (A. Beric, personal communication). Kinesthetic related activity (see Chapter 16 CD-video segments 2 and 3) is often observed, but a clear somatotopy is not evident. Upon exiting the STN, the microelectrode may pass through a thin quiet

436

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

FIGURE 16.16 Representative tracings of cellular activity that may be encountered during a STN recording trajectory. Each tracing is 5 s in length, except for trace FF, which is 1 s in length. (A) (Sound 14) Thalamic burster cell and single cell. (B) (Sound 15) Thalamic single cell. (C) (Sound 16) ZI cellular activity. (D) (Sound 17) STN tremor cell. (E) (Sound 18) Nontremor STN cell from the ventral half of the STN nucleus. (F) (Sound 19) SNr cell.

zone or will pass directly into the SNr. Entry into the SNr is demarcated by significant increases both in background neural activity and in cellular firing rates (Fig. 16.16F, CD-STN sound 6), which are usually greater than 60 Hz. Up to 7 mm of SNr may be encountered, depending on the anteroposterior position of the trajectory. We require 4–6 mm of STN, preferably with evidence of kinesthetic activity, for implantation of the DBS lead. This large a span allows for two of the four electrode contacts to be placed within the nucleus, leaving the other two above the nucleus in the ZI and H2. Additionally, this large a span of STN recording ensures that the electrodes are implanted solidly within the nucleus and not near a border. The primary goal of test stimulation at the STN is to check for stimulationinduced adverse events (AEs) because, aside from tremor arrest and some modest reductions in rigidity, positive STN stimulation effects may not be observed for hours or days. Test stimulation is performed in bipolar configuration

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

437

with the implanted DBS lead and Medtronic’s single lead screener (model 3625, Medtronic, Minneapolis, MN). Parameters are: 60 µs, 180 Hz, 0–4 V. We do not stimulate higher than 4.0 V for fear of inducing hemiballism. Moreover, we have yet to employ amplitudes greater than 4 V to achieve clinical benefit at this target. Transient paresthesias are frequently encountered with the onset of stimulation. Persistent paresthesias indicate stimulation of the medial lemniscal pathway, which lies posterolateral to the nucleus. Stimulation-induced contractions of the contralateral hemibody and/or face indicate anterolateral misplacement of the lead. Finally, abnormal eye movements may be encountered if the lead is positioned too medially or deep to the nucleus. The first test stimulation is performed using contacts 0−, 1+ up to a voltage of 4.0 V. If no significant adverse effects are encountered with this focal test, we proceed to test stimulation employing all four contacts (i.e., 0−, 1−, 2+, 3+ up to a voltage of 4.0 V). This test covers the full contact space of the electrodes and focuses on identifying stimulation-induced adverse events in the ventral aspect of the stimulation field. This is the area where most AEs have occurred in our experience. The final stimulation is performed using contacts 0+, 1+, 2−, 3− up to a voltage of 4.0 V. This examines the dorsal aspect of the stimulation field.

4.1.4 Data Organization The data from each microrecording tract are plotted on scaled graph paper (1.0 cm: 1.0 mm) [17, 163]. The borders of each encountered structure are marked, and the span of each region is represented by a different color for easy differentiation. In order to accurately account for our angle of approach, a line that is parallel to the intercommissural line is also drawn. The plotted tract is then traced onto a transparent plastic sheet. The transparency is placed on scaled maps (10:1) derived from the Schaltenbrand-Wahren human stereotactic atlas [164] (see Chapter 16, CD-video segment 4 for a STN procedure and Fig. 16.17 (see also color plate) for a GPi procedure) in order to determine to which map the trajectory best fits. The accuracy of the fit is dependent upon the number of trajectories, the number of structures encountered along each trajectory, and finally upon how well the patient’s anatomy fits the atlas, which is derived from a single human specimen. It can be difficult to find one place to which a single tract fits best, especially when performing pallidal or thalamic interventions. When mapping the STN, the many structures encountered along a single trajectory make fitting it to the atlas a little more straightforward. If there is any question about the proper fit of the data, we perform another recording tract. Knowing the spatial relationship between each tract, we can better fit all of the data to the atlas with each subsequent trajectory.

438

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

FIGURE 16.17 Once the recording data are transferred to 1:10 scaled graph paper, the trajectories are transferred to a transparency. The angle of the trajectory relative to the AC-PC line is added to the transparency, and the trajectory is then fitted to scaled atlas sections. This figure shows two trajectories during a GPi lesion surgery. The green lines represent the GPe part of the trajectory, and the red lines represent the GPi part of the trajectory. By overlaying two atlas maps, a threedimensional picture of the trajectories can be formed. Original non-scaled or enhanced tracing from [164] (see also color plate).

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

439

5 CONCLUSION The fine details of these procedures vary from center to center, but the neurophysiological techniques used by each center can be divided into the following categories: (1) microrecording; (2) semi-microrecording; (3) stimulation; and (4) evoked response testing. In the over 1,500 trajectories performed by the authors, we feel that the information gathered with microrecording is of great benefit when performing these surgeries. Microrecording has been shown to be as safe as other stereotactic procedures [165] when done properly. With these surgeries we are trying to modify the physiology of a target structure; therefore, microrecording gives specific physiologic data to help determine the optimal placement. In most cases (43–88%, depending on the study [69, 114, 139]), this physiological target corresponds to the anatomic target, but in the 12–67% of cases that is not the case. At present there is no way of knowing which of these patients will fall into either category before the surgery. The neurophysiologic techniques used in the operating room require trained and skilled personnel, not only to acquire but also to interpret the data. If everything goes perfectly, the data are relatively easy to interpret, but when the signals are not textbook cases, this interpretation needs to be done by very experienced personnel. Up until the mid-1990s, centers had to put their own microelectrode recording systems together and build their own microelectrodes, since there were no commercially available systems. At the present time there now exist about 10 companies (internationally) that produce microrecording systems, and the first FDA-approved microelectrodes were placed on the market in 2000. The key points to get the best signals at are the microelectrode, preamplifier, and amplifier. The main feature of reliable microelectrode systems for neurophysiological targeting of deep brain structure is the quality of the recorded signal. This is more important than any of the fashionable features that many manufactures offer. No software-based interpretation scheme is going to replace the skilled human interpreter when the recordings are difficult. As already stated, the operating room is very harsh electrically. The more we learn about the areas of interest, the faster and smoother each of these procedures will go. Included with the CD at the end of the book is a short video that demonstrates the recording of an anterior thalamic cell, the kinesthetic response of the STN to passive and active movements of the patient’s right wrist, and a single trajectory showing the relationship of the recordings for that tract to the scaled atlas maps. The CD also includes examples of single-unit recordings from the various structures described in this chapter.

440

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

REFERENCES 1. Horsley, V. (1909). The function of the so-called motor area of the brain. Br. Med. J., 2, 125– 132. 2. Meyers, R. (1940). Surgical procedure for postencephalitic tremor, with notes on the physiology of the premotor fibers. Arch. Neurol. Psychiatry, 44, 455–459. 3. Meyers, R. (1942). Surgical interruption of the pallidofugal fibers: Its effect on the syndrome of paralysis agitans and technical considerations in its application. N.Y. State J. Med., 42, 317–325. 4. Meyers, R. (1968). The surgery of the hyperkinetic disorders. In “The handbook of clinical neurology” (P.J. Vinken, and G. Bruyn, eds.), vol. 6, pp. 844–878. North Holland Publishers, Amsterdam. 5. Meyers, R. (1941). The modification of alternating tremors, rigidity and festination by surgery of the basal ganglia. Res. Pub. Ass. Nerv. Ment. Dis., 21, 602–665. 6. Meyers, R. (1941). Cortical extinction in convulsions. J. Neurophysiol., 4, 250–256. 7. Fotlstad, H., Hariz, M., and Ljunggren, B. (1991). History of Clarke’s stereotactic instrument. Stereot. Funct. Neurosurg., 57, 130–140. 8. Spiegel, E.A., Wycis, H.T., and Marks, M. (1947). Stereotactic apparatus for operations the human brain. Science, 106, 349–350. 9. Spiegel, E.A. (1996). Development of stereoencephalatomy for extrapyramidal disease. J. Neurol., 24, 433–439. 10. Cooper, I.S. (1953). Ligation of the anterior choroidal artery for involuntary movements of parkinsonism. Psychiatry, 27, 317–319. 11. Cooper, I.S., and Bravo, G.O. (1958). Anterior choroidal artery occlusion, chemopallidectomy and chemothalarnectomy in parkinsonism: A consecutive series of 700 operations. In “Pathogenesis and treatment of parkinsonism”(W.S. Fields, ed.), pp. 325–352. Thomas, Springfield. 12. Cooper, I.S. (1958). Chemopallidotomy. J. Neurosurg., 15, 244–250. 13. Narabayashi, H., and Okuxna, T. (1953). Procaine-oil blocking of the globus pallidus for the treatment of rigidity and tremor of parkinsonism. Proc. Jpn. Acad., 29, 134–137. 14. Guiot, G., and Brion, S. (1953). Traitement des mouvements anormaux par la coagulation pallidae. Technique et results. Rev. Neurol., 89, 578–580. 15. Laitinen, L.V., Bergenheim, A.T., and Hariz, M.I. (1992). Leksell’s posteroventral pallidotomy in the treatment of Parkinson’s disease [see comments]. J. Neurosurg., 76, 53–61. 16. Yoshida, S., Nambu, A., and Jinnai, K. (1993). The distribution of the globus pallidus neurons with input from various cortical areas in the monkeys. Brain Res., 611(1), 170–174. 17. Vitek, J.L., Bakay, R.A.E., Hashimoto, T., Kaneoke, Y., Mewes, K., Yu Zhang, J., Rye, D., Starr, P., Baron, M., Turner, R., and DeLong, M.R. (1998). Microelectrode-guided pallidotomy: Technical approach and its application in medically intractable Parkinson’s disease. J. Neurosurg., 88, 1027–1043. 18. Spiegel, E.A., Wycis, H.T., Szekely, E.G., Adams, J., Flanaga, M., and Baird, W. (1963). Campotomy in various extrapyramidal disorders. J. Neurosurg., 20, 871–884. 19. Hughes, J.R. (1996). The electroencephalogram in parkinsonism. J. Neurosurg., 24, 369– 376. 20. Tasker, R.R., Organ, L.W., and Hawrylyshyn, P.A. (1982). “The thalamus and midbrain of man. A physiological atlas using electrical stimulation.” Charles C. Thomas, Springfield, Illinois. 21. Nakajima, H., Fukamachi, A., Isobe, I., Miyazaki, M., Shibazaki, T., and Ohye, C. (1978). Estimation of neural noise. Appl. Neurophysiol., 41, 193–201. 22. Bertrand, C., Poirier, L., Martinez, N., and Gauthier, C. (1958). Pneumotaxic localization, recording, stimulation, and section of basal brain structures in dyskinesia. Neurology, 8, 783–786.

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

441

23. Bertrand, G., Jasper, H., and Wong, A. (1967). Microelectrode study of the human thalamus: Functional organization in the ventro-basal complex. Confin. Neurol., 29, 81–86. 24. Albe-Fessard, D., Arfel, G., Guiot, G., Derome, P., Hertzog, E., Vourc’h, G., Brown, H., Aleonard, P., De La Herran, J., and Trigo, J.C. (1966). Electophysiological studies of some deep cerebral structures in man. J. Neurosci., 3, 37–51. 25. Albe-Fessard, D. (1973). Electrophysiological methods for the identification of thalamic nuclei. Z. Neurol., 205, 15–28. 26. Penfield, W., and Rasmussen, T. (1957). “The cerebral cortex of man: A clinical study of localization and function.” The Macmillian Co., New York. 27. Organ, L.W., Tasker, R.R., and Moody, N.F. (1967). The impedance profile of the human brain as a localization technique in stereoencephalotomy. Confin. Neurol., 29, 192–196. 28. Laitinen, L.V., and Johansson, G.G. (1967). Locating human cerebral structures by the impedance method. Confin. Neurol., 29, 197–201. 29. Albe-Fessard, D., Tasker, R., Yamashiro, K., Chodakiewitz, J., and Dostrovsky, J. (1986). Comparison in man of short latency averaged evoked potentials in thalamic hand and scalp zones of representation. Electroencephalogr. Clin. Neurophysiol., 65, 405–415. 30. Hassler, R., Riechert, T., Mundinger, F., Umbach, W., and Ganglberger, J.A. (1960). Physiologic observations in stereotaxic operations in extrapyramidal motor disturbances. Brain, 83, 337–350. 31. Narabayashi, H., and Ohye, C. (1974). Nucleus ventralis intermedius of human thalamus. Trans. Am. Neurol. Ass., 99, 232–233. 32. Fukamachi, A., Ohye, C., and Narabayashi, H. (1973). Delineation of the thalamic nuclei with a microelectrode in stereotaxic surgery for parkinsonism and cerebral palsy. J. Neurosurg., 39, 214–225. 33. Yokoyama, T., Imamura, Y., Sugiyama, K., Nishizawa, S., Yokota, N., Ohta, S., and Uemura, K. (1999). Prefrontal dysfunction following unilateral posteroventral pallidotomy in Parkinson’s disease. J. Neurosurg., 90, 1005–1010. 34. Bakay, R.A., DeLong, M.R., and Vitek, J.L. (1992). Posteroventral pallidotomy for Parkinson’s disease [letter; comment]. J. Neurosurg., 77, 487–488. 35. Lenz, F.A., Tasker, R.R., Kwan, H.C., Schider, S., Kwong, R., Dostrovsky, J.O., and Murphy, J.T. (1987). Selection of the optimal site for the relief of parkinsonian tremor on the basis of spectral analysis of neuronal firing patterns. Appl. Neurophysiol., 50, 338–343. 36. Lenz, F.A., Dostrovsky, J.O., Tasker, R.R., Yamashiro, K., Kwan, H.C., and Murphy, J.T. (1988). Single-unit analysis of the human ventral thalamic nuclear group: Somatosensory responses. J. Neurophysiol., 59, 299–316. 37. Lenz, F.A., Tasker, R.R., Kwan, H.C., Schneider, S., Kwong, R., Murayama, Y., Dostrovsky, J.O., and Murphy, J.T. (1988). Single unit analysis of the human ventral thalamic nuclear group: Correlation of thalamic tremor cells with the 3–6 Hz component of parkinsonian tremor. J. Neurosci., 8, 754–764. 38. Lenz, F.A., Kwan, H.C., Dostrovsky, J.O., Tasker, R.R., Murphy, J.T., and Lenz, Y.E. (1990). Single unit analysis of the human ventral thalamic nuclear group. Brain, 113, 1795–1821. 39. Levy, A. (1967). Stereotaxic brain operations in Parkinson’s syndrome and related motor disturbances. Comparison of lesions in the pallidum and the thalamus with those in the internal capsule. Supplement to confina neurologica, 29, 5–68. 40. Lenz, F.A., Dostrovsky, J.O., Kwan, H.C., Tasker, R.R., Yamashiro, K., and Murphy, J.T. (1988). Methods for microstimulation and recording of single neurons and evoked potentials in the human central nervous system. J. Neurosurg., 68, 630–634. 41. Tasker, R.R., and Kiss, Z.H.T. (1995). The role of the thalamus in functional neurosurgery. Neurosurg. Clin. N. Amer., 6, 73–99. 42. Dostrovsky, J.O., Davis, K.D., Lee, L., Sher, G.D., and Tasker, R.R. (1993). Electrical stimulation-induced effects in the human thalamus. In “Electrical and magnetic stimulation of the brain and spinal cord” (O. Devinsky, A. Beric, and M. Dogali, eds.). Raven Press, New York.

442

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

43. Yoshida, M. (1989). Electrophysiological characteristics of human subcortical structures by frequency spectral analysis of neural noise (field potential) obtained during stereotactic surgery. Stereotac. Funct. Neurosurg., 52, 157–163. 44. Narabayashi, H., Yokochi, F., and Nakajima, Y. (1984). Levodopa-induced dyskinesia and thalamotomy. J. Neurol., Neurosurg., Psychiatry, 47, 831–839. 45. Sano, K., Mayanagi, Y., Sekino, H., Ogahiwa, M., and Ishijima, B. (1970). Results of stimulation and destruction of the posterior hypothalamus in man. J. Neurosurg., 33, 687–707. 46. Lang, A.E. (1998). Surgical treatment of dystonia. In “Dystonia 3: Advances in neurology” (S. Fahn, C.D. Marsden, and M. DeLong, eds.), vol. 78, pp. 185–198. Lippincott-Raven, Philadelphia. 47. Vitek, J.L. (1998). Surgery for dystonia. Neurosurg. Clin. N. Amer., 9, 345–366. 48. Sterio, D., Beric, A., Dogali, M., Fazzini, E., Alfaro, G., and Devinsky, O. (1994). Neurophysiological properties of pallidal neurons in Parkinson’s disease. Ann. Neurol., 35, 586–591. 49. Gillingham, F.J., Watson, W.S., Donaldson, A.A., and Naughton, J.A. (1960). The surgical treatment of Parkinsonism. J. Br. Med., 2, 1395–1402. 50. Baron, M.S., Vitek, J.L., Bakay, R.A., Green, J., Kaneoke, Y., Hashimoto, T., Turner, R.S., Woodard, J.L., Cole, S.A., McDonald, W.M., and DeLong, M.R. (1996). Treatment of advanced parkinson’s disease by posterior GPi pallidotomy: 1-year results of a pilot study. Ann. Neurol., 40, 355–366. 51. Sutton, J.P., Couldwell, W., Lew, M.F., Mallory, L., Grafton, S., DeGiorgio, C., Welsh, M., Apuzzo, M.L.J., Ahmadi, J., and Waters, C.H. (1995). Ventroposterior medial pallidotomy in patients with advanced Parkinson’s disease. Neurosurgery, 36, 1112–1117. 52. Lang, A.E., Lozano, A.M., Montgomery, E., Duff, J., Tasker, R., and Hutchinson, W. (1997). Posteroventral medial pallidotomy in advanced Parkinson’s disease. N. Engl. J. Med., 337, 1036–1042. 53. Dogali, M., Fazzini, E., Kolodny, E., Eidelberg, D., Sterio, D., Devinsky, O., and Beric, A. (1995). Stereotactic ventral pallidotomy for Parkinson’s disease. Neurology, 45, 753–761. 54. Iacono, R.P., Shima, F., Lonser, R.R., Kuniyoshi, S., Maeda, G., and Yamada, S. (1995). The results, indications, and physiology of posteroventral pallidotomy for patients with Parkinson’s disease [see comments]. Neurosurgery, 36, 1118–1127. 55. Iacono, R.P., Lonser, R.R., Oh, A., and Yamada, S. (1995). New pathophysiology of Parkinson’s disease revealed by posteroventral pallidotomy. Neurol. Res., 17, 178–180. 56. Iacono, R.P., Kuniyoshi, S.M., and Schoonenberg, T. (1998). Experience with stereotactics for dystonia: Case examples. In “Dystonia 3: Advances in neurology” (S. Fahn, C.D. Marsden, and M. DeLong, eds.), vol. 78, pp. 221–226. Lippincott-Raven, Philadelphia. 57. Kondziolka, D., Bonaroti, E., Baser, S., Brandt, F., Kim, Y.S., and Lundsford, L.D. (1999). Outcomes after stereotactically guided pallidotomy for advanced Parkinson’s disease. J. Neurosurg., 90, 197–202. 58. Lozano, A.M., Lang, A.E., Galvez-Jimenez, N., Miyasaki, J., Duff, J., Hutchinson, W.D., and Dostrovsky, J.O. (1995). Effect of GPi pallidotomy on motor function in Parkinson’s disease. Lancet, 346, 1383–1387. 59. Laitinen, L.V., and Hariz, M.I. (1990). Pallidal surgery abolishes all Parkinsonian symptoms. Movement Disorders, 5, 82. 60. Laitinen, L.V., Bergenheim, A.T., and Hariz, M.I. (1992). Ventroposterolateral pallidotomy can abolish all parkinsonian symptoms. Stereotact. Funct. Neurosurg., 58, 14–21. 61. Laitinen, L.V. (1994). Ventroposterolateral pallidotomy. Stereotact. Funct. Neurosurg., 62, 41– 52. 62. Laitinen, L.V. (1995). Pallidotomy for Parkinson’s disease. Neurosurg. Clin. N. Amer., 6, 105–112. 63. Krauss, J.K., Jankovic, J., Lai, E.C., Rettig, G.M., and Grossman, R.G. (1997). Posterioventral medial pallidotomy in levodopa-unresponsive Parkinsonism. Arch. Neurol., 54, 1026–1029. 64. Fazzini, E., Dogali, M., Sterio, D., Eidelberg, D., and Beric, A. (1997). Stereotactic pallidotomy for Parkinson’s disease. Neurology, 48, 1273–1277.

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

443

65. Krack, P., Pollak, P., Limousin, P., Hoffmann, D., Benazzouz, A., Le Bas, J.F., Koudsie, A., and Benabid, A.L. (1998). Opposite motor effects of pallidal stimulation in Parkinson’s disease. Ann. Neurol., 43, 180–192. 66. Honey, C.R., Phil, D., Stoessl, A.J., Tsui, J.C.K., Schulzer, M., and Calne, D.B. (1999). Unilateral pallidotomy for reduction of parkinsonian pain. J. Neurosurg., 91, 198–201. 67. Ghika, J., Ghika-Schmid, F., Fankhauser, H., Assal, G., Vingerhoets, F., Albanese, A., Bogousslavsky, J., and Favre, J. (1999). Bilateral contemporaneous posteroventral pallidotomy for the treatment of Parkinson’s disease: Neuropsychological and neurologic side effects. J. Neurosurg., 91, 313–321. 68. Alterman, R.L., Sterio, D., Beric, A., and Kelly, P.J. (1999). Microelectrode recording during posterioventral pallidotomy: Impact on target selection and complications. Neurosurgery, 44, 315–323. 69. Guridi, J., Gorospe, A., Ramos, E., Linazasoro, G., Rodriguez, M.C., and Obeso, J.A. (1999). Stereotactic targeting of the globus pallidus internus in Parkinson’s disease: Imaging versus electrophysiological mapping. Neurosurgery, 45, 278–289. 70. Hassler, R., Mundinger, F., and Riechert, T. (1965). Correlations between clinical and autopic findings in stereotaxic operations of parkinsonism. Confin. Neurol., 25, 282–290. 71. Dogali, M., Beric, A., Sterio, D., Eidelberg, D., Fazzini, E., Takikawa, S., Samelson, D.R., Devinsky, O., and Kolodny, E.H. (1994). Anatomic and physiologic considerations for Parkinson’s disease. Stereotact. Funct. Neurosurg., 62, 53–60. 72. Justesen, C.R., Penn, R.D., Kroin, J.S., and Egel, R.T. (1999). Stereotactic pallidotomy in a child with Hallervorden-Spatz disease. J. Neurosurg., 90, 551–554. 73. Taha, J.M., Favre, J., Baumann, T.K., and Burchiel, K.J. (1996). Functional anatomy of the pallidal base in Parkinson’s disease. Neurosurgery, 39, 1164–1168. 74. Taha, J.M., Favre, J., Baumann, T.K., and Burchiel, K.J. (1996). Characteristics and somatotopic organization of kinesthetic cells in the globus pallidus of patients with Parkinson’s disease. J. Neurosurg., 85, 1005–1012. 75. Taha, J.M., Favre, J., Baumann, T.K., and Burchiel, K.J. (1997). Tremor control after pallidotomy in patient with Parkinson’s disease: Correlation with microrecording findings. J. Neurosurg., 86, 642–647. 76. Beric, A., Sterio, D., Dogali, M., Alterman, R., and Kelly, P. (1996). Electrical stimulation of the globus pallidus preceding stereotactic posteroventral pallidotomy. Stereotact. Funct. Neurosurg., 66, 161–169. 77. Limousin, P., Greene, J., Pollak, P., Rothwell, J., Benabid, A.L., and Frackowiak, R. (1997). Changes in cerebral activity pattern due to subthalamic nucleus or internal pallidum stimulation in Parkinson’s disease. Ann. Neurol., 42, 283–291. 78. Svennilson, E., Torvik, A., Lowe, R., and Leksell, L. (1960). Treatment of parkinsonism by stereotactic thermolesions in the pallidal region: A clinical evaluation of 81 cases. Acta Psychiatr. Neurol. Scand., 35, 358–377. 79. Kelly, P.J. (1995). Pallidotomy in Parkinson’s disease [editorial; comment]. Neurosurgery, 36, 1154–1157. 80. Young, M.S., Triggs, W.J., Bowers, D., Greer, M., Friedman, W.A. (1997). Stereotactic pallidotomy lengthens the transcranial magnetic cortical stimulation silent period in Parkinson’s disease. Neurology, 49, 1278–1283. 81. Kazumata, K., Antonini, A., Dhawan, V., Moeller, J.R., Alterman, R.L., Kelly, P., Sterio, D., Fazzini, E., Beric, A., and Eidelberg, D. (1997). Preoperative indicators of clinical outcome following stereotaxic pallidotomy. Neurology, 49, 1083–1090. 82. Vitek, J.L., Zhang, J., Evatt, M., Mewes, K., DeLong, M.R., Hashimoto, T., Triche, S., and Bakay, R.A. (1998). GPi pallidotomy for dystonia: Clinical outcome and neuronal activity. In “Advances in neurology” (S. Fahn, C.D. Marsden, and M. DeLong, eds.), vol. 78, pp. 211–219. Lippincott-Raven, Philadelphia.

444

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

83. Lozano, A.M., Kumar, R., Gross, R.E., Giladi, N., Hutchison, W.D., Dostrovsky, J.O., Lang, A.E. (1997). Globus pallidus internus pallidotomy for generalized dystonia. Movement Disorders, 12, 865–870. 84. Lin, J.-J., Lin, S.-Z., Lin, G.-Y., Chang, D.-C., and Lee, C.-C. (1998). Application of bilateral sequential pallidotomy to treat a patient with generalized dystonia. Eur. Neurol., 40, 108–110. 85. Lin, J.-J., Lin, S.-Z., and Chang, D.-C. (1999). Pallidotomy and generalized dystonia. Movement Disorders, 14, 1057–1059. 86. Lin, J.-J., Lin, G.-Y., Shih, C., Lin, S.-Z., Change, D.-C., and Lee, C.-C. (1999). Benefit of bilateral pallidotomy in the treatment of generalized dystonia. J. Neurosurg., 90, 974–976. 87. Alvarez, L., Macias, R., Guridi, J., Lopez, G., Alvarez, E., Maragoto, C., Teijeiro, J., Torres, A., Pavon, N., Rodriguez-Oroz, M.C., Ochoa, L., Hetherington, H., Juncos, J., DeLong, M.R., and Obeso, J.A. (2001). Dorsal subthalamotomy for Parkinson’s disease. Movement Disorders, 16, 72–78. 88. Andy, O.J., Jurko, M.F., and Sias, F.R. (1963). Subthalomotomy in treatment of parkinsonian tremor. Presented at meeting of Harvey Cushing Society, Philadelphia, PA. 89. Benabid, A.L., Pollak, P., Gao, D., Hoffmann, D., Limousin, P., Gay, E., Payen, I., and Benazzouz, A., (1996). Chronic electrical stimulation of the ventralis intermedius nucleus of the thalamus as a treatment of movement disorders. J. Neurosurg., 84, 203–214. 90. Koller, W., Pahwa, R., Busenbark, K., Hubble, J., Wilkinson, S., Lang, A., Tuite, P., Sime, E., Lazano, A., Hauser, R., Malapira, T., Smith, D., Tarsy, D., Miyawaki, E., Norregaard, T., Kormos, T., and Olanow, C.W. (1997). High-frequency unilateral thalamic stimulation in the treatment of essential and parkinsonian tremor. Ann. Neurol., 42, 292–299. 91. Hubble, J.P., Busenbark, K.L., Wilkinson, S., Pahwa, R., Paulson, G.W., Lyons, K., and Koller, W.C. (1997). Effects of thalamic deep brain stimulation based on tremor type and diagnosis. Movement Disorders, 12, 337–341. 92. Hubble, J.P., Busenbark, K.L., Wilkinson, S., Penn, R.D., Lyons, K., and Koller, W.C. (1996). Deep brain stimulation for essential tremor. Neurology, 46, 1150–1153. 93. Geny, C., Nguyen, J.P., Pollin, B., Feve, A., Ricolfi, F., Cesaro, P., and Degos, J.D. (1996). Improvement of severe postural cerebellar tremor in multiple sclerosis by chronic thalamic stimulation. Movement Disorders, 11, 489–494. 94. Whittle, I.R., Hooper, J., and Pentland, B. (1998). Thalamic deep-brain stimulation for movement disorders due to multiple sclerosis. Lancet, 351, 109–110. 95. Krack, P., Limousin, P., Benabid, A.L., and Pollak, P. (1997). Chronic simulation of subthalamic nucleus improves levodopa-induced dyskinesias in Parkinson’s disease. Lancet, 350, 1676. 96. Krack, P., Pollak, P., Limousin, P., Benazzouz, A., and Benabid, A.L. (1997). Stimulation of subthalamic nucleus alleviates tremor in Parkinson’s disease. Lancet, 350, 1675–1676. 97. Krack, P., Hamel, W., Mehdorn, H.M., and Deuschl, G. (1999). Surgical treatment of Parkinson’s disease. Curr. Opin. Neurol., 12, 417–425. 98. Limousin, P., Krack, P., Pollak, P., Benazzouz, A., Ardouin, C., Hoffmann, D., and Benabid, A.L. (1998). Electrical stimulation of the subthalamic nucleus in advanced Parkinson’s disease. N. Engl. J. Med., 339, 1105–1111. 99. Houeto, J.L., Damier, P., Bejjani, P.B., Staedler, P.B., Bonnet, A.M., Arnulf, I., Pidoux, B., Dormont, D., Cornu, P., and Agid, Y. (2000). Subthalamic stimulation in Parkinson disease: A multidisciplinary approach. Arch. Neurol., 57, 461–465. 100. Molinuevo, J.L., Valldeoriola, F., Tolosa, E., Rumia, J., Valls-Sole, J., Roldan, H., and Ferrer, E. (2000). Levodopa withdrawal after bilateral subthalamic nucleus stimulation in advanced Parkinson disease. Arch. Neurol., 57, 983–988. 101. Pinter, M.M., Alesch, F., Murg, M., Seiwald, M., Helscher, R.J., and Binder, H. (1999). Deep brain stimulation of the subthalamic nucleus for control of extrapyramidal features in advanced idiopathic Parkinson’s disease: One year follow-up. J. Neur. Trans., 106, 693–709.

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

445

102. Hutchison, W.D., Allan, R.J., Opitz, H., Levey, R., Dostrovsky, J.O., Lang, A.E., and Lozano, A.M. (1998). Neurophysiological identification of the subthalamic nucleus in surgery for Parkinson’s disease. Ann. Neurol., 44, 622–628. 103. Gross, C., Rougier, A., Guehl, D., Boraud, T., Julien, J., and Bioulac, B. (1997). High-frequency stimulation of the globus pallidus internalis in parkinson’s disease: A study of seven cases. J. Neurosurg., 87, 491–498. 104. Bejjani, B., Damier, P., Arnuff, I., Bonnet, A.M., Vidailhet, M., Dormont, D., Pidoux, B., Cornu, P., Marsault, C., and Agid, Y. (1997). Pallidal stimulation for Parkinson’s disease. Two targets? Neurology, 49, 1564–1569. 105. Tronnier, V.M., Fogel, W., Kronenbuerger, M., and Steinvorth, A. (1997). Pallidal stimulation: An alternative to pallidotomy? J. Neurosurg., 87, 700–705. 106. Kumar, R., Dagher, A., Hutchison, W.D., Lang, A.E., and Lozano, A.M. (1999). Globus pallidus deep brain stimulation for generalized dystonia: Clinical and PET investigation. Neurology, 53, 871–874. 107. Lang, A.E. (2000). Surgery for Parkinson disease: A critical evaluation of the state of the art. Arch. Neurol., 57, 1118–1125. 108. Burchiel, K.J., Anderson, V.C., Favre, J., and Hammerstad, J.P. (1999). Comparison of pallidal and subthalamic nucleus deep brain stimulation for advanced Parkinson’s disease: Results of a randomized, blinded pilot study. Neurosurgery, 45, 1375–1384. 109. Vayssiere, N., Hemm, S., Zanca, M., Picot, M.C., Bonafe, A., Cif, L., Frerebeau, P., Coubes, P. (2000). Magnetic resonance imaging stereotactic target localization for deep brain stimulation in dystonic children. J. Neurosurg., 93, 784–790. 110. Coubes, P., Roubertie, A., Vayssiere, N., Hemm, S., Echenne, B. (2000). Treatment of DYT1generalized dystonia by stimulation of the internal globus pallidus. Lancet, 355, 2220–2221. 111. Bakay, R.A.E., Vitek, J.L., and DeLong, M.R. (1992). Thalamotomy for tremor. In “Neurosurgical operative atlas” (S.S. Rengachary, and R.N. Wilkins, eds.), vol. 2, pp. 299–312. Williams and Wilkins, Baltimore. 112. Limousin, P., Pollak, P., Benazzouz, A., Hoffmann, D., Le Bas, J.F., Broussolle, E., Perret, J.E., and Benabid, A.L. (1995). Effect on parkinsonian signs and symptoms of bilateral subthalamic nucleus stimulation. Lancet, 345, 91–95. 113. Limousin, P., Pollak, P., Hoffmann, D., Benazzouz, A., Perret, J.E., and Benabid, A.L. (1996). Abnormal involuntary movements induced by subthalamic nucleus stimulation in parkinsonian patients. Movement Disorders, 11, 231–235. 114. Forster, A., Eljamel, M.S., Varma, T.R., Tulley, M., and Latimer, M. (1999). Audit of neurophysiological recording during movement disorder surgery. Stereotact. Funct. Neurosurg., 72, 154–156. 115. Bejjani, B.P., Dormont, D., Pidoux, B., Yelnik, J., Damier, P., Arnulf, I., Bonnet, A.M., Marsault, C., Agid, Y., Philippon, J., and Cornu, P. (2000). Bilateral subthalamic stimulation for Parkinson’s disease by using three dimensional stereotactic magnetic resonance imaging and electrophysiologic guidance. J. Neurosurg., 92, 615–625. 116. Tagliati, M., and Alterman, R.L. (2001). Guidelines for patient selection for ablative and deep brain stimulation surgery. Semin. Neurosurg., 12, 161–168. 117. Kretschmann, H.J., and Weinrich, W. (1999). “Neurofunctional systems” (CD-ROM). Thieme, New York. 118. Miller, W.C., and DeLong, M.R. (1967). Altered tonic activity of neurons in the globus pallidus and subthalamic nucleus in the primate MPTP model of parkinsonism. In “The basal ganglia II” (M.B. Carpenter, and A. Jayarman, eds.), pp. 415–427. Plenum, New York. 119. Filion, M., and Tremblay, L. (1991). Abnormal spontaneous activity of globus pallidus neurons in monkeys with MPTP-induced parkinsonism. Brain Res., 547, 142–151. 120. DeLong, M.R. (1990). Primate models of movement disorders of basal ganglia origin. Trends in Neuroscience: Special Issue: Basal Ganglia Research, 13, 281–285.

446

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

121. DeLong, M.R., Crutcher, M.D., and Georgopoulos, A.P. (1985). Primate globus pallidus and subtahalmic nucleus: Functional organization . J. Neurophysiol., 53, 530–543. 122. Alexander, G.E., and Crutcher, M.D. (1990). Functional architecture of basal ganglia circuits: Neural substrates of parallel processing [see comments]. Trends Neurosci., 13, 266–271. 123. Alexander, G.E., Crutcher, M.D., and DeLong, M.R. (1990). Basal ganglia-thalamocortical circuits: Parallel substrates for motor, oculomotor, prefrontal and limbic functions. Prog. Brain Res., 85, 119–146. 124. Mink, J.W. (1996). The basal ganglia: Focused selection and inhibition of competing motor programs. Prog. Neurobiol., 50, 381–425. 125. DeLong, M.R., and Wichmann, T. (1993). Basal ganglia-thalamocortical circuits in Parkinsonian signs. Clin. Neurosci., 1, 18–26. 126. Kandel, E.R., Schwartz, J.H., and Jessell, T.M. (1991). “Principles of neural science.” Appleton & Lange, Norwalk. 127. Obeso, J.A., Linazasoro, G., Gorospe, A., et al. (1997). Complication associated with chronic levodopa therapy in parkinson’s disease. In “Beyond the decade of the brain” (C.W. Olanow, and J.A. Obeso, eds.), vol. 2, pp. 11–31. Wells Medical Limited. Kent, U.K. 128. Nutt, J.G. (1990). Leva-dopa–induced dyskinesia: Review, observations, and speculations. Neurology, 40, 340–345. 129. Crossman, A.R., Sambrook, M.A., and Jackson, A. (1984). Experimental hemichorea/ hemiballismus in the monkey. Studies on the intracerebral site of action in a drug-induced dyskinesia. Brain, 107, 579–596. 130. Pal, P.K., and Samii, A., Kishore, A., et al. (2000). Long term outcome of unilateral pallidotomy: Follow-up of 15 patients for 3 years. J. Neurol, Neurosurg. Psychiatry, 69, 337–344. 131. Arle, J.E., and Alterman, R.L., (1999). Surgical options in Parkinson’s disease. Med. Clin. N. Ame., 83, 483–498. 132. Siegfried, J., and Lippitz, B. (1994). Bilateral chronic electrostimulation of ventroposterolateral pallidum: A new therapeutic approach for alleviating all parkinsonian symptoms. Neurosurgery, 35, 1126–1130. 133. Ambrose, J. (1974). Computerized x-ray scanning of the brain. J. Neurosurg., 40, 679–695. 134. Lunsford, L.D., Latchaw, R.E., and Vries, J. (1983). Stereotaxic implantation of deep brain electrodes using computed tomography. Neurosurgery, 13, 280–286. 135. Lunsford, L.D., and Martinez, A.J. (1984). Stereotactic exploration of the brain in the era of computed tomography. Surg. Neurology., 22, 222–230. 136. Leskell, L., Herner, T., and Leskell, D. (1985). Visualisation of stereotactic radiolesions by nuclear magnetic resonance. J. Neurol. Neurosurg. Psychiatry., 48,19–20. 137. Lunsford, L.D., Martinez, A.J., and Latchaw, R.E. (1986). Stereotaxic surgery with a magnetic resonance and computed tomography-compatible system. J. Neurosurg., 64, 872–878. 138. Kondziolka, D., Dempsey, P.K., Lunsford, L.D., Kestle, J.R., Dolan, E.J., Kanal, E., Tasker, R.R. (1992). A comparison between magnitude resonance imaging and computed tomography for stereotactic coordinate determination. Neurosurgery, 30, 402–407. 139. Alterman, R.L., Reiter, G.T., Shils, J., Skolnick, B., Arle, J.E., Lesutis, M., Simuni, T., Colcher, A., Stern, M., and Hurtig, H. (1999). Targeting for thalamic deep brain stimulation implantation without computer guidance: Assessment of targeting accuracy. Stereotact. Funct. Neurosurg., 72, 150–153. 140. Taren, J.A., Ross, D.A., and Gebarski, S.S. (1993). Stereotactic localization using fast spin-echo imaging in functional disorders. Acta Neurochir., 58 (suppl), 59–60. 141. Starr, P.A., Vitek, J.L., DeLong, M., and Bakay, R.A.E. (1999). Magnetic resonance imagingbased stereotactic localization of the globus pallidus and subthalamic nucleus. Neurosurgery, 44, 303–314.

Neurophysiological Monitoring During Neurosurgery for Movement Disorders

447

142. Reich, C.A., Hudgins, P.A., Sheppard, S.K., Starr, P.A., and Bakay, R.A.E. (2000). A highresolution fast spin-echo inversion-recovery for preoperative localization of the internal globus pallidus. Amer. J. Neuroradiol., 21, 928–931. 143. Zonenshayn, M., Rezai, A.R., Mogilner, A.Y., Beric, A., Sterio, D., and Kelly, P.J. (2000). Comparison of anatomic and neurophysiologic methods for subthalamic nucleus targeting. Neurosurgery, 47, 282–294. 144. Sumanaweera, T.S., Adler, J.R., Napel, S., and Glover, G.H. (1994). Characterization of spatial distortion in magnetic resonance imaging and its implications for stereotactic surgery. Neurosurgery, 35, 696–704. 145. Gerdes, J.S., Hitchon, P.W., Neerangun, W., and Torner, J.C. (1994). Computed tomography versus magnetic resonance imaging in stereotactic localization. Stereotact. Funct. Neurosurg., 63, 124–129. 146. Walton, L., Hampshire, A., Forster, D.M.C., and Kemeny, A.A. (1995). Stereotactic localization using magnetic resonance imaging. Stereotact. Funct. Neurosurg., 64 (suppl 1), 155– 163. 147. Walton, L., Hampshire, A., Forster, D.M.C., and Kemeny, A.A. (1996). A phantom study to assess the accuracy of stereotactic localization, using T1-weighted magnetic resonance imaging with the Leksell stereotactic system. Neurosurgery, 38, 170–178. 148. Alterman, R.L., Shils, J.L, Tagliati, M., and Rogers, J. (2000), Relative accuracy of inversion recovery and T2-weighted MRI for targeting the subthalamic nucleus. Presented at the XIVth European Society for Stereotactic and Functional Neurosurgery, London, U.K. 149. Shils, J.L., and Alterman, R.L. (2000). Comparative accuracy of stereotactic coordinates derived from various MRI acquisition techniques. Presented at the XIVth European Society for Stereotactic and Functional Neurosurgery, London, U.K. 150. Holtzheimer, P.E., Roberts, D.W., and Darcey, T.M. (1999). Magnetic resonance imaging versus computer tomography for target localization in functional stereotactic neurosurgery. Neurosurgery, 45, 290–298. 151. Axer, H., Stegelmeyer, J., and von Keyserlingk, D.G. (1999). Comparison of tissue impedance measurements with nerve fiber architecture in human telencephalon: Value in identification of intact subcortical structures. J. Neurosurg., 90, 902–909. 152. Organ, W.L., Tasker, R.R., and Moody, N.F. (1967). The impedance profile of the human brain as a localization technique in stereoecephalotomy. Confin. Neurol., (Madrid), 29, 192–196. 153. Yoshida, M., Yanagisawa, N., Shimazu, H., Givre, A., and Narabayashi, H. (1964). Physiological identification of the thalamic nucleus. Arch. Neurol., 11, 435–443. 154. Guridi, J., and Lozano, A.M. (1997). A brief history of pallidotomy. Neurosurgery, 41, 1169–1180. 155. Ohye, C., and Narabayashi, H. (1979). Physiological study of presumed ventralis intermedius neurons in the human thalamus. J. Neurosurg., 50, 290–297. 156. Hardy, J. (1966). Electrophysiological localization and identification. J. Neurosurg., 24, 410–414. 157. Lozano, A., Hutchison, W., Kiss, Z., Tasker, R., Davis, K., and Dostrovsky, J. (1996). Methods for microelectrode-guided posteroventral pallidotomy. J. Neurosurg., 84, 194–202. 158. Hirai, T., Miyazaki, M., Nakajima, H., Shibazaki, T., and Ohye, C. (1983). The correlation between tremor characteristics and the predicted volume of effective lesions in stereotaxic nucleus ventralis intermedius thalamotomy. Brain, 106, 1001–1018. 159. Gillingham, F.J. (1966). Depth recording and stimulation. J. Neurosurg., 24, 382–387. 160. Geddes, L.A. (1972). “Electrodes and the measurement of bioelectric events.” Wiley-Interscience, New York.

448

Jay L. Shils, Michele Tagliati, and Ron L. Alterman

161. Shils, J.L., Patterson, T., and Stecker, M.M. (2000). Electrical properties of metal microelectrodes. Amer. J. Electrodiag. Technol., 40, 71–82. 162. Lenz, F.A., Tasker, R.R., Kwan, H.C., Schnider, S., Kwong, R., Murayama, Y., Dostrovsky, J.O., and Murphy, J.T. (1988). Single unit analysis of the human ventral thalamic nuclear group: Correlation of thalamic “tremor cells” with the 3–6 Hz component of parkinsonian tremor. J. Neurosci., 8, 754–764. 163. Starr, P.A., Vitek, J.L., and Bakay, R.A.E. (1998). Deep brain stimulation for movement disorders. Neurosurg. Clin. N. Amer., 9, 381–402. 164. Schaltenbrand, G., and Wahren, W. (1977). “Atlas for stereotaxy of the human brain.” Thieme, New York. 165. Alterman, R.L., and Kelly, P.J. (1998). Pallidotomy technique and results. Neurosurg. Clin. N. Amer., 9, 337–343.

FIGURE 10.2 sory root.

Rootlet of L5 sensory root being held by stimulator probes away from rest of sen-

FIGURE 16.1 A three-dimensional artist's rendition of the structures involved in surgery for movement disorders. The light greenish blue structure on the left is the globus pallidus (GPi and GPe). The large grey structure on the right is the thalamus, and the small dark green structure is the subthalamic nuclei (STN). The medial edge of the STN is only 6.0 m m from the midline of the brain. With the trajectories that our group uses in the operating room, we encounter around 10.0 m m of GPi, 11.0 m m of VIM, and 5.0 m m of STN. Modified from [117].

,,,,,

SUPERIOR/ DORSALI

M EDIAL

ANTERIOR

POSTERIOR

LATERAL

~r INFERIOR/VENTRAL

FIGURE 16.11 Sagittal slice through the globus pallidus, taken 20.0 mm from the midline. The color shading is referenced in the text. Reprinted from [164].

AL

SUPERIORI DORSAL

) ANTERIOR i

POSTERIOR

~r INFERIOR/ VENTRAL

LATERAL

FIGURE 16.13 Sagittal slice through the thalamus taken 14.5 mm from the midline. The color shading is referenced in the text. Reprinted from [164].

DIAL

SUPERIORI DORSAL

T ANTERIOR L

POSTERIOR

LATERAL INFERIORI VENTRAL~1~ FIGURE 16.15

~i~

Sagittal slice through the STN taken 12.0 mm from the midline. Reprinted from [164].

FIGURE 16.17 Once the recording data are transferred to 1:10 scaled graph paper, the trajecto m ries are transferred to a transparency. The angle of the trajectory relative to the ACmPC line is added to the transparency, and the trajectory is then fitted to scaled atlas sections. This figure shows two trajectories during a GPi lesion surgery. The green lines represent the GPe part of the trajectory, and the red lines represent the GPi part of the trajectory. By overlaying two atlas maps, a threedimensional picture of the trajectories can be formed. Original non-scaled or enhanced tracing from [164].

CHAPTER

17

Anesthesia and Motor Evoked Potential Monitoring TOD B. SLOAN Department of Anesthesiology, University of Texas Health Science Center, San Antonio, Texas

1 Introduction 2 Overview 3 Effects of Specific Anesthetic Agents 3.1 Halogenated Inhalational Agents 3.2 Intravenous Analgesic Agents 3.3 Muscle Relaxants 4 Conclusion: Anesthetic Choice for Motor Tract Monitoring References

ABSTRACT Anesthesia used to conduct surgery where motor evoked potential (MEP) monitoring is used has marked effects on the ability to record responses. This review will focus on both the theory and the practical issues of the effects of anesthetic agents. In theory, the type of interaction should be predictable, based on the mechanism of action of the drugs involved. Unfortunately, we do not have a thorough understanding of the mechanisms of anesthesia. However, the major target of anesthetic action appears to be at the gaba amino butyric acid (GABA) and the n-methyl-daspartic acid (NMDA) receptors mediating electrolyte channels (Na+, Cl−, Ca2+) at synapses, so that synaptic transmission is hampered. In addition, halogenated inhalational agents and ketamine appear to hinder axonal conduction. As a result, the major anesthetic impact on neurological pathways used for monitoring appears to be at the synaptic connections, with an additional minor component based on the length of the pathway. Varying locations of action as well as differences related to drug dosage make marked differences between agents when related to motor-evoked responses. Further, neurological disease appears to make the responses more difficult to record under anesthesia, increasing the challenge of monitoring in the very patients where it may be most important. Neurophysiology in Neurosurgery: A Modern Intraoperative Approach Copyright 2002, Elsevier Science (USA). All rights of reproduction in any form reserved.

451

452

Tod B. Sloan

1 INTRODUCTION The anesthesia that is used during surgery where motor evoked potential (MEP) monitoring is used can markedly affect the ability to record some types of responses. This review will focus on both the theory and practical aspects of these anesthetic agents. In theory, the type of interaction between the anesthesia and the motor responses should be predictable from knowledge of the mechanism of action of the anesthetic drugs involved. Unfortunately, we do not have a thorough understanding of the mechanism of anesthesia, but the major target for anesthetic action appears to be at neural synapses, especially the gaba amino butyric acid (GABA) and n-methyl-d-aspartic acid (NMDA) receptors, which mediate electrolyte channels (Na+, Cl−, Ca2+). Hence synaptic transmission is hampered. In addition, ketamine appears to hinder axonal conduction. As a result, these two effects suggest that the major anesthetic impact on neurological pathways will be at synaptic connections, with an additional minor component based on the length of the pathway [1]. This is consistent with the observed effects (see following text) and is corroborated by the observation that muscle recorded responses from transcranial magnetic stimulation are altered by anticonvulsant medications that do not produce anesthesia [2] but produce postsynaptic enhancement of GABA receptors. Different anesthetic agents produce unique effects, depending on the specific locations of anesthetic action, drug type, and drug potency. Finally, the presence of neurological disease may enhance the effects of anesthesia drugs as well as change the relative effects on different pathways. As discussed in following text, providing anesthesia during monitoring of the motor pathway can pose a very significant challenge for the anesthesiologist in the very patients where it may be most important.

2 OVERVIEW Since the anesthetic effects will vary with the methodology, it is important to review the three basic techniques that have been employed to attempt monitoring of motor tracts. The first technique involves stimulation of the motor cortex using transcranial electrical (tcEMEP), direct cortical electrical, or transcranial magnetic stimulation (tcMMEP). The descending electrical volley in the motor tracts that is produced can be measured in the epidural space or by recording the compound muscle action potential (CMAP) produced in the muscles (“myogenic” responses). The second method involves the stimulation of the spinal cord by epidural or percutaneously placed perispinal electrodes and recording in the muscles or in the peripheral nerves (“neurogenic” responses). One popular technique of spinal stimulation has been pioneered by Owen [3] and has been termed the “neurogenic motor-evoked potential”

Anesthesia and Motor Evoked Potential Monitoring

453

(NMEP), based on early experiments that suggested that the response so measured was a motor tract response. As discussed in the following text, this technique (as with any form of spinal stimulation) is probably the result of stimulation of both sensory and motor tracts and therefore may not be an appropriate technique for monitoring pure motor tracts. Recent reports by Minahan and colleagues show that this technique monitors only the functional integrity of the dorsal columns [4, 5]. To date it is believed that transcranial motor evoked potentials (tcMEP) by electrical or magnetic stimulation produce a pure motor tract response. The third technique of motor monitoring involves stimulation of the peripheral motor system such as the nerve roots (as with pedicle screw monitoring) or the use of reflex arcs synapsing in the caudal spinal cord. For motor cortex stimulation (transcranial or direct), two basic mechanisms produce a descending electrical volley. First, they produce a direct activation of the pyramidal cells producing a D (direct) wave. Second, activation of internuncial pathways produces a series of I waves that follow the D waves down the descending motor pathways in 1.3 to 2.0 ms intervals. D waves appear to originate from the trigger zone of the motor cortex pyramidal cells by direct stimulation, whereas I waves appear to be produced by transsynaptic activation of tangentially oriented corticocortical interconnections of lamina V as well as corticocortical projections from the precentral and premotor cortex. Since magnetic stimulation induces a tangentially oriented electric current in the brain, weaker magnetic stimulation may preferentially produce I waves. Hence, although electrical and high-magnetic-field impulses directly stimulate the pyramidal cells, weaker magnetic impulses appear to depend on synaptic activation for production of a response. Since no synapses are involved, the production of D waves will be relatively immune to anesthetic effects on the motor cortex. However, since the production of I waves involves synapses, the production of these waves will be reduced with anesthetic agents that depress synaptic function. The situation may be a bit more complex, since synaptic function (and the ability to activate I waves via synaptic stimulation) is probably the result of a delicate balance of inhibitory and excitatory influences from adjacent neural pathways. Therefore, it is possible that anesthetic agents that block inhibitory influences may lesson the anesthetic impact by making internuncial synapses more easily activated. Likewise, anesthetics that block excitatory influences may worsen the anesthetic impact at the internuncial synapses. Once activated, the electrical responses that travel down the spinal cord reach the α-motoneuron, where, after sufficient stimulation has occurred, a peripheral nerve response results. For a single stimulation of the motor cortex, D and I waves both appear necessary for bringing the α-motoneuron to firing threshold for production of a peripheral nerve response. Likewise, techniques that involve stimulation of the spinal cord or sensory reflex arcs will also traverse

454

Tod B. Sloan

the α-motoneuron. Methods that involve stimulation of the nerve root or motor component of the peripheral nerve will not involve this synapse. As the second synaptic system in the motor pathway, corticospinal tract (CT) α-motoneurons synapse in a location potentially susceptible to anesthetic effects. Anesthetic action here may have two effects. First, partial synaptic blockade may compound a loss of I waves, making it more difficult for cortical stimulation to bring the α-motoneuron to firing threshold. This may explain why cortical stimulation with weak magnetic fields (tcMMEP) is more susceptible to inhalational anesthesia than is electrical stimulation (tcEMEP). At higher anesthetic doses, synaptic blockade may inhibit synaptic transmission regardless of the composition of the descending spinal cord volley of activity. Of note is that stimulation of the pathways that lead to peripheral motor response (peripheral sensory stimulation or stimulation of the sensory pathways of the spinal cord that result in descending antidromic volleys) may pass through other synapses and thus make the anesthetic effects more complex. Likewise, anesthetic effects that alter the excitatory or inhibitory influences on the α-motoneuron may also alter the anesthetic effect on the α-motoneuron. One method that has somewhat overcome the anesthetic effect at the α-motoneuron is multipulse stimulation of the motor cortex. In this case, multiple D waves are produced such that anesthetic depression of the internuncial pathways in the cortex has little effect. Thus, for this technique, the major anesthetic effect may be at the α-motoneuron. This may allow better myogenic responses at low anesthetic doses. However, at higher doses of anesthetics, these multiple D waves may be insufficient to overcome the depression effectively blocking the response. Clearly, the interstimulus interval of the multipulse stimulation will also interact with the effectiveness of the technique. With widely spaced stimuli, decay of the effect at the α-motoneuron may prevent effective summation. With closely spaced stimuli, the cortical neurons may not have recovered effectively from the previous response to produce an adequate response to subsequent stimuli. Hence an optimal interstimulus interval (ISI) should be found. It is possible that anesthetic influences may alter the optimal ISI if they alter the recovery characteristics of the motor cortex or spinal cord for the production of spinal cord responses. The same is true if anesthetic influences alter the decay characteristics of the α-motoneuron that effect summation. The third major synaptic location for anesthetic effect in the motor pathway is at the neuromuscular junction. Fortunately, with the exception of neuromuscular blocking agents and drugs that alter acetylcholine transmission, anesthetic drugs have little effect at the neuromuscular junction. The converse is also true, since neuromuscular blocking agents have little effect in synaptic transmission and axonal conduction in motor pathways other than at the neuromuscular junction. This means that neuromuscular blockade will need to be

Anesthesia and Motor Evoked Potential Monitoring

455

carefully controlled when myogenic responses are monitored. Finally, it should be noted that anesthetic drugs might have an indirect effect on motor-evoked responses by virtue of alterations in the physiological factors that provide nutrient supply to the neural tracts. For example, anesthetic agents typically lower blood pressure, which can contribute to neural ischemia and alteration in motor responses. Similarly, anesthesia or anesthetic management can cause changes in cerebral or spinal cord blood flow through changes in vascular tone mediated by the anesthetic directly or through carbon dioxide and tissue pH. Of interest is that synaptic function may be the most highly vulnerable region of the neural tracts to ischemia and physiological changes because of its high dependence on energy metabolism.

3 EFFECTS OF SPECIFIC ANESTHETIC AGENTS Given this theoretical background, we can review the effects of individual agents on the various components of the motor-evoked responses. Since agents may differ as to which synaptic transmitter they interact with, the actual effect on the evoked responses may also differ. Further, as already stated, if they interact primarily at excitatory or inhibitory responses, they may produce a spectrum of effects at different concentrations due to changes in the balance of excitatory and inhibitory contributions to the motor pathway. Since the electroencephalogram (EEG) is produced by synaptic activity, the effect of anesthetic drugs on MEP often parallels the effects on the EEG [6, 7].

3.1 HALOGENATED INHALATIONAL AGENTS The most common anesthetics in use today, the halogenated inhalational agents (desflurane, enflurane, halothane, isoflurane, sevoflurane), have been extensively studied with motor-evoked responses [8–17]. These agents produce synaptic inhibition as revealed by reduction in frequency and amplitude in the EEG until electrocerebral silence occurs. As such, one would predict anesthetic depression at the α-motoneuron as well as in the internuncial neurons of the motor pathway. Depression of the neuromuscular junction does not appear to be a major effect on the MEP. These agents depress the EEG to different degrees at equipotent anesthetic concentrations. For the MEP, several studies support differences in the potency of halogenated inhalational agents on transcranial MEPs. The relative order seen is isoflurane (most potent), enflurane, and halothane (least potent) [10]. Studies with the newer agents sevoflurane and desflurane suggest that these agents are similar to isoflurane at steady state. However, because of their more rapid onset and offset of effect (because of their

456

Tod B. Sloan

FIGURE 17.1 The effect of increasing isoflurane concentrations on the compound muscle action potential (CMAP) response to transcranial electrical motor cortex stimulation (tcEMEP) in a ketamine-anesthetized baboon. As can be seen, the amplitude decreases progressively with increasing concentrations.

relative insolubility), they may appear to be more potent during periods when concentrations are increasing. Also, because of their lower solubility, their use allows a more rapid adjustment during anesthesia (i.e., the concentration can be raised or lowered more rapidly if needed). Single-pulse transcranial stimulation with MEPs recorded in muscle appears to be so easily abolished by inhalational agents that the MEPs are often unrecordable in the presence of these agents [9, 14, 18, 19]. When recordable, the major effect appears to occur at low concentrations (e.g., less than 0.2–0.5% isoflurane) [12, 16, 20, 21]. This effect is likely due to depression of the αmotoneuron synapse as well as loss of I waves caused by anesthetic effects in the internuncial synapses [15, 22]. Changes in the H reflex confirm an effect of halogenated inhalational agents at the spinal level [23]. Figure 17.1 shows the loss of amplitude of the tcEMEP CMAP as isoflurane concentration is increased in the ketamine anesthetized baboon. In contrast to myogenic responses, the D response recorded from the epidural space is highly resistant to the effects of these agents and is easily recordable at high concentrations [19, 20, 24, 25] and can be used for monitoring. This has suggested that the most prominent anesthetic effect on tcMEP is at the α-motoneuron level [11, 13]. However, the loss of I waves from a cortical effect may be sufficient to block myogenic responses, even in the absence of anesthetic effects at the α-motoneuron. This is because repetitive I waves appear to be necessary for producing myogenic responses in the unanesthetized state [26]. Figure 17.2 shows the tcEMEP epidural response in a baboon as isoflurane concentration is increased from 0.3 to 2.1%. Note that although the D wave is maintained, the I waves are lost.

Anesthesia and Motor Evoked Potential Monitoring

457

FIGURE 17.2 The effect of increasing isoflurane concentrations on the epidural response to transcranial electrical motor cortex stimulation (tcEMEP) in a ketamine-anesthetized baboon. Note that although the D wave is maintained, the I waves are lost.

Studies comparing tcMMEP and tcEMEP suggest that the magnetic technique can be more sensitive to the inhalational agents [10], probably because magnetic stimulation (especially weaker field strengths) rely more on transsynaptic activation of the CT. High-magnetic-strength tcMMEP (which produces D waves) appears to overcome this cortical difference. The difference between tcEMEP and tcMMEP likely also relates to the type of current pulse driving the magnetic coil. Since biphasic or rapidly attenuated sine wave pulses may be more effective than monophasic pulses, anesthetic effect may be more pronounced in the latter technique [27, 28]. Because the D wave is resistant to anesthetic depression, the anesthetic effect at the α-motoneuron can be partially overcome at low concentrations by multiple-pulse transcranial stimulation [33, 34]. In this circumstance the multiple D waves formed (and I waves if produced) summate at the α-motoneuron, resulting in a peripheral nerve and motor response when cortical stimuli are placed at an ISI interval of 1–2 ms optimally, but also effectively to 10 ms [33]. Alternatively, the anesthetic effect can also be partially overcome by activation of the H reflex through peripheral nerve stimulation combined with transcranial stimulation [35]. Hence, low concentrations of inhalational agents appear acceptable when high-frequency transcranial stimulation is used (trains of stimuli with ISI of 2–5 ms [29, 30]). As predicted, higher concentrations of these agents eliminated the myogenic responses from this stimulation. Clinical experience (noted in following text) suggests that avoiding the inhalational agents may still be desirable for monitoring during high-frequency stimulation [30]. It appears that the total intravenous anesthesia (TIVA) technique also may produce superior responses with high-frequency stimulation [30–32]. As indicated previously, the optimal interstimulus interval may vary with the anesthetic effect [36]. This has been noted with isoflurane and is depicted in

458

Tod B. Sloan

FIGURE 17.3 A smoothed plot of CMAP amplitude at various combinations of isoflurane (0.2–1.0%) and interstimulus interval (1–6 ms) for dual-pulse transcranial electrical stimulation. As can be seen, the largest amplitude with low concentrations is 2–5 ms. As the isoflurane concentration is increased, the optimal ISI increases to the higher level, and at the highest concentration the largest amplitude is with an ISI of 1 ms.

Fig. 17.3, which shows that a relatively broad ISI (1–5 ms) is effective at low concentrations (0.2% isoflurane); however, a wider interval appears better at higher concentrations (e.g., 4–5 ms at 0.4–0.6% isoflurane). At even higher concentrations (1% isoflurane), the most effective ISI was 1 ms. These data suggest that if inhalational agents are used with the multipulse technique, a “tuning” of the stimulation ISI may improve the effectiveness of the monitoring. Studies with spinal or epidural stimulation show minimal effects of anesthesia on neurogenic or myogenic responses, suggesting that the neurophysiology of the electrical activity arriving at the α-motoneuron is different than from cortical stimulation [3, 37]. However, the anesthetic effects in the spinal cord at all of the synapses involved (sensory and motor pathways) may change the mixture of orthodromic motor and antidromic sensory contributions to the recorded responses. Mochida studied the responses in the peripheral nerve and muscle following epidural stimulation in the cat [34]. He noticed that single-pulse

Anesthesia and Motor Evoked Potential Monitoring

459

stimulation produced a response that was eliminated by pentobarbitol, by lowdose isoflurane, and by posterior column transection (but not lateral column transection). When a pair of stimuli was used (ISI 1–5 ms), a new complex in the peripheral nerve response was seen. This complex and the CMAP were eliminated only by high-dose isoflurane or by lateral spinal cord transection. Mochida’s study suggests that the type of spinal cord stimulation and the anesthetic used may alter the balance of sensory and motor contributions to the peripheral nerve and muscle response of spinal stimulation. Of interest is that the sensory tracts were more easily stimulated than motor tracts. Recent studies suggest that with isoflurane anesthesia, the motor component is preferentially blocked, perhaps by interaction at the synapses at the α-motoneuron or by differential effects on conduction in the spinal tracts in humans [20]. Based on these studies, it is conceivable that spinal stimulation techniques may monitor a mixture of sensory and motor pathways that may change with the type and dosage of the anesthetic agents used. 3.1.1 Nitrous Oxide Despite its weak anesthetic profile, studies with tcMMEP [8] and tcEMEP [38] show that nitrous oxide produces depression of myogenic tcMEP. When compared at equipotent anesthetic concentrations, nitrous oxide produces more profound changes in myogenic tcMEP than any other inhalational anesthetic agent [1]. Like halogenated agents, the effects on the epidurally recorded MEP are minimal. Despite the depressant effect of nitrous oxide, it has been used with recording of myogenic responses, particularly when combined with opioids (“nitrousnarcotic” anesthetic technique). It has also been used to supplement intravenousbased anesthetics with opioids combined with propofol [30, 32] or etomidate [32, 39–42]. It has been used in concentrations of <52% [42–43], 50–60% [30, 44–47], 60–65% [48], 65–66% [39, 40], and 70–75% [18, 25, 32]. Since nitrous oxide is rather insoluble in tissues, its concentration and the depressant effect can be titrated rather quickly, so that if chosen as an anesthetic technique it can be reversed rapidly [41]. Figures 17.4 (CMAP responses) and 17.5 (epidural responses) show the effect of increasing inspired nitrous oxide from 0 to 79% on tcEMEP. The effects appear to mimic the effects of isoflurane (i.e., loss of CMAP and decreased number of I waves at higher concentrations). Studies suggest that nitrous oxide may actually be “context sensitive” in its effects, similar to its effects on the EEG (i.e., the actual effect may vary depending on the other anesthetics already present). Studies of equi-anesthetic mixtures of isoflurane and nitrous oxide have demonstrated that the mixture has a more potent effect on cortical SSEP than would be predicted by adding the

460

Tod B. Sloan

FIGURE 17.4 The effect of increasing nitrous oxide concentrations on the CMAP response to tcEMEP in a ketamine-anesthetized baboon. As can be seen, the amplitude is progressively decreased with increasing concentrations, similar to isoflurane.

FIGURE 17.5 The effect of increasing nitrous oxide concentrations on the epidural response to tcEMEP in a ketamine-anesthetized baboon. Note that although the D wave is maintained, the number of I waves is decreased, and their latencies are prolonged in a way that is similar to the effects of isoflurane.

effects of each agent [49]. This suggests that the mechanism of action of nitrous oxide may be different from that of isoflurane.

3.2 INTRAVENOUS ANALGESIC AGENTS Since the inhalational agents and nitrous oxide are poor choices for anesthesia when myogenic responses of tcMEP are desired, anesthetic techniques have focused on intravenous anesthetic agents for clinical monitoring. If the inhalational agents need to be completely avoided, then intravenous agents can be combined to produce a total intravenous anesthetic (TIVA). Fortunately, because

Anesthesia and Motor Evoked Potential Monitoring

461

the mechanism of action of intravenous agents appears to be different than that of inhalational agents, these agents differ in their effects on MEP such that they can be more favorable for intraoperative monitoring. 3.2.1 Opioid Agents Since analgesia (pain relief) is a primary component of anesthesia, the opioids (fentanyl, alfentanil, sufentanil, and remifentanil) are the intravenous agents most frequently chosen when inhalational agents must be avoided or used in low concentrations. As with minimal depression of the EEG (a dose-related decline in frequency of the EEG in the delta range while maintaining amplitude), opioid effects on MEP are less than those of inhalational agents. Studies with myogenic responses of tcMEP from electrical and magnetic stimulation show only mild amplitude decreases and latency increases that usually permit recording [21, 50–52]. The observed effects are reversed with naloxone, suggesting that this effect is related to mu receptor activity [53–55]. As with systemic opioids, the spinal application of morphine or fentanyl for postoperative pain management produces minimal changes in the H reflex [56, 57] suggesting that effects on motor-evoked responses should be minimal. In addition to having minimal effects on the motor pathways, fentanyl has been suggested to be useful in reducing background spontaneous muscle contractions and associated motor unit potentials, which may improve CMAP recordings. The effects of opioids appear to be related to drug concentration, since maximal changes occur at the same time drug concentrations peak, after bolus drug delivery. One study of fentanyl suggests that the effect on sensory evoked responses may be minimized by using a drug infusion to avoid transient bolus effects [58]. Remifentanil, a rapidly metabolized opioid, may be well suited for use by infusion since its concentration and effect can be rapidly changed. Because opioids have less effect than inhalational agents, opioid-based anesthesia has usually been used when myogenic tcMEPs are monitored [18, 24, 25, 30–32, 39, 40, 42, 44–48, 59–68]. 3.2.2 Ketamine Ketamine is a less frequently used analgesic but a valuable component of anesthetic techniques for recording responses that are easily depressed by anesthesia. This is because ketamine is an excitatory agent (probably through its interaction at the NMDA receptor) that may heighten synaptic function rather than depress it. For example, ketamine produces high-amplitude theta activity in the EEG, with an accompanying increase in beta activity that appears to represent activation of thalamic and limbic structures. It has been reported to

462

Tod B. Sloan

provoke seizure activity in individuals with epilepsy but not in normal individuals. In addition, ketamine has been reported to increase cortical SSEP amplitude [56] and increase the amplitude of muscle and spinal recorded responses following spinal stimulation [21, 69]. This latter effect on muscle responses may be mediated by the same mechanism that potentates the H reflex [70]. Minimal effects were observed in myogenic tcMEP with ketamine [21, 71, 72]. Muscle responses and spinal recorded responses to spinal stimulation are also enhanced at doses that do not produce spike and wave activity in the EEG [65, 69]. As such, ketamine has become a valuable adjunct during some TIVA techniques for recording muscle responses. In these techniques it has been combined with opioids [42, 59, 66, 67] or methohexitol [66]. High dosages, however, produce depression of the myogenic response, which is consistent with its known property of spinal axonal conduction block [73]. These effects have made ketamine a valuable adjunct to anesthesia with tcMEP [74]; however, its hallucinatory potential and known increase in intracranial pressure with intracranial pathology have led to a reluctance to use it in anesthesia. 3.2.3 Sedative-Hypnotic Drugs Intravenous sedative agents are frequently used to induce or supplement general anesthesia, particularly with opioids or ketamine, when inhalational agents are not used. This is because fentanyl is primarily an analgesic, and, even with high doses, sedation, anxiolysis, or amnesia cannot be ensured (i.e., intraoperative awareness may be present). Although ketamine doses produce some dissociative effects in addition to analgesia, supplementation of ketamine with sedative drugs can reduce the risk of excitatory events, including hallucinations. Hence, a TIVA usually includes an opioid or ketamine for analgesia combined with a sedative-hypnotic agent. Like opioids and ketamine, the sedative-hypnotic agents (except droperidol) can be used by infusion to reduce transient changes in the monitored responses. In studies where the different drugs have been compared, marked differences in recording myogenic tcMEP have been observed [18, 21, 65]. In general, thiopental, midazolam, and propofol produced marked depression in bolus doses. Because of the slower metabolism of these drugs, the authors concluded that thiopental and midazolam were poor drugs for induction of anesthesia because their effects may linger into the surgical procedure. 3.2.4 Barbiturates Popular drugs for induction of general anesthesia, barbiturates are similar to inhalational agents in their effect on the EEG, producing mild activation (fast activity) at low doses and a depressant effect leading to burst suppression and

Anesthesia and Motor Evoked Potential Monitoring

463

electrical silence at higher doses. Not surprisingly, myogenic responses of tcMEP are unusually sensitive to barbiturates. Further, the effect appears quite prolonged; in one study, induction eliminated the tcMMEP response for a period of 45–60 min [21], suggesting that barbiturates may be a poor induction choice when monitoring with this modality. For this reason, most anesthetic protocols do not use thiopental for induction of anesthesia. However, it has been successfully used in some anesthetic regimes [44, 46, 68] and given as intermittent boluses during the anesthetic [68]. Given newer, better agents (e.g., propofol and etomidate), the use of barbiturates has largely been eliminated during tcMEP monitoring. One exception, methohexitol, has different characteristics from thiopental. It is rapidly metabolized, so that it is short acting and rapidly titratable. In addition, since it is known to enhance seizure activity at low doses, it may reduce the inhibitory influences on the motor pathway. Although not commonly used, one TIVA protocol for myogenic tcMEP successfully used methohexitol infusions with opioids and ketamine [66]. Fortunately, this drug is more rapidly metabolized and appears to have excitatory properties (low doses can be used to identify seizure foci during cortical mapping of epilepsy). 3.2.5 Benzodiazepines The benzodiazepines, notably midazolam, have been advocated as supplements to TIVA in routine surgery because of their excellent sedation and amnesic qualities (particularly to reduce the chance of hallucinogenic activity with ketamine). However, at higher doses they produce generalized slowing of the EEG into the theta and delta range without burst suppression, suggesting marked synaptic inhibition via GABA channel action. Midazolam has been used as intermittent boluses during recording of myogenic tcMEP [68], but as with thiopental, it produces prolonged marked depression of myogenic tcMEP [51, 65, 75, 76]. This has been interpreted as inhibition of cortical pyramidal cell neurons. Like barbiturates, the benzodiazepines have not gained favor for induction or as a component of TIVA during myogenic tcMEP recordings. 3.2.6 Etomidate As opposed to the barbiturates and benzodiazepines, etomidate can enhance synaptic activity at low doses, possibly by changing the balance of inhibitory and excitatory influences on motor pathways. At low doses (0.1 mg/kg), etomidate may produce seizures in patients with epilepsy [77], and marked myoclonic activity is often seen with anesthesia induction. However, at higher doses it can produce a flat EEG. Since etomidate is rapidly metabolized, its concentration can be rapidly adjusted to take advantage of the enhancing activity

464

Tod B. Sloan

FIGURE 17.6 The effect of increasing doses of etomidate on the CMAP response to tcEMEP in a ketamine-anesthetized baboon. As can be seen, the amplitude is progressively decreased with increasing concentrations, similar to isoflurane. Note an initial increase in CMAP amplitude at low doses.

or reduce the depressant effects seen at high concentrations. This effect has been used to enhance amplitude in both sensory and motor-evoked responses [78, 79]. Fortunately, the enhancing activity occurs at doses that are consistent with the desired degree of sedation and amnesia needed for TIVA. Studies with tcMEP have suggested that etomidate is an excellent agent for induction and monitoring of these modalities [21, 51, 65, 67, 80, 81]. Of several intravenous agents studied, etomidate had the least degree of amplitude depression after induction doses or continual intravenous infusion [21]. Latency (onset) changes were not observed, and amplitude enhancement of muscle responses was not observed except at high dosages [80]. Because of the prolonged effect of thiopental, etomidate has been used for induction of anesthesia during monitoring [41, 42, 45, 59, 67]. As a component of TIVA, infusions of etomidate have been combined with opioids [40, 41, 60, 67]. Figures 17.6 (CMAP) and 17.7 (epidural), showing recordings from tcEMEP with increasing concentrations of etomidate, demonstrate that etomidate behaves differently than inhalational agents or propofol (see next section). Note an initial increase in CMAP amplitude at low doses (an effect more prominent in tcMMEP than tcEMEP) and an increase in I waves rather than a loss. 3.2.7 Propofol As the newest sedative-hypnotic agent, propofol has been extensively studied. It produces dose-dependent depression of the EEG reminiscent of the barbiturates and can produce burst suppression and electrical silence at high doses. This is consistent with the postulated site of anesthetic action of propofol on the cerebral cortex [82]. However, the drug is very rapidly metabolized, so that the

Anesthesia and Motor Evoked Potential Monitoring

465

FIGURE 17.7 The effect of increasing doses of etomidate on the epidural response to tcEMEP in a ketamine-anesthetized baboon. Note that although the D wave is maintained, the I waves are lost (similar to isoflurane). Note an increase in I waves rather than a loss.

FIGURE 17.8 The effect of increasing doses of propofol on the CMAP response to tcEMEP in a ketamine-anesthetized baboon. As can be seen, the amplitude decreases progressively with increasing concentrations (similar to isoflurane). Note the pattern is similar to that of inhalational agents, with loss of CMAP at higher concentrations.

drug concentration can be titrated down to levels compatible with adequate TIVA and MEP recording. Studies with tcMEP have demonstrated a depressant effect on myogenic response amplitude, also consistent with a cortical effect [51, 65, 83]. As a component of TIVA, induction of anesthesia can include propofol [30], and infusions of propofol have been combined with opioids [30–32, 48]. Not unexpectedly, propofol has been used in tcEMEP when the recordings are epidural [13]. Figures 17.8 (CMAP) and 17.9 (epidural) show recordings from tcEMEP with increasing concentrations of propofol. Note that the pattern is similar to that of inhalational agents, with loss of CMAP and decrease of the number of I waves at higher concentrations.

466

Tod B. Sloan

FIGURE 17.9 The effect of increasing doses of propofol on the epidural response to tcEMEP in a ketamine-anesthetized baboon. Note that the pattern is similar to that of inhalational agents, with decreased number of I waves at higher concentrations.

3.2.8 Droperidol Droperidol has little effect on the EEG when used alone. However, it is known to lower seizure threshold, probably by dopamine antagonism. It does not appear to produce neuroexcitatory phenomena or induce seizures in epileptic patients. When combined with fentanyl (“neurolept anesthesia”), it increases EEG alpha activity at low doses. At higher doses, it produces high-amplitude beta and delta activity. It appears to have minimal effects on myogenic tcMMEP when combined with opioids [60, 65]. However, since its effect is long-lasting, it is not suitable for use by infusion, and many anesthesiologists would prefer to use a more rapidly metabolized sedative hypnotic for TIVA.

3.3 MUSCLE RELAXANTS Since muscle relaxants have their major site of action at the neuromuscular junction, they have little effect on electrophysiological recordings that do not derive from muscle activity. In fact, they may improve or be essential for some types of recordings where the muscle activity near the recording electrode may be unwanted noise. This is true for epidural or peripheral nerve recordings where the activity of overlying muscle obscures the response from transcranial or spinal stimulation. For recording of epidural or neurogenic responses, complete or near complete neuromuscular blockade is highly desirable [25, 61]. Figure 17.10 shows recording from the epidural space from tcEMEP with (below) and without (top) muscle relaxation. Note that the muscle artifact obscures the identification of I waves.

Anesthesia and Motor Evoked Potential Monitoring

467

FIGURE 17.10 Recordings from the epidural space from tcEMEP with (below) and without (top) muscle relaxation. Note that the muscle artifact obscures the identification of I waves.

Certainly, complete neuromuscular blockade will prevent recording of muscle responses (CMAPs) during MEP recording. However, partial neuromuscular blockade has the benefit of reducing a substantial portion of the movement that accompanies the testing, and it may facilitate some surgical procedures in which muscle relaxation is needed for retraction of tissues. In these cases, careful monitoring of the blockade of the neuromuscular junction is critical. Two methods are customarily used to assess the degree of neuromuscular blockade [85]. The method that best quantitates the blockade involves measuring the amplitude of the CMAP (T1) produced by supramaximal stimulation of a peripheral motor nerve (M response). When neuromuscular monitoring is conducted this way, successful monitoring of myogenic responses has been accomplished at 5–15% [61], 10% [63], 10–25% [64], 10–25% [47], 20% [40, 41, 45, 59, 60], 25% [42], and 30–50% [24, 43, 67] of T1 compared to baseline. Clinically, anesthesiologists often assess neuromuscular blockade by counting the number of twitches resulting from four motor nerve stimuli delivered at a rate of 2 Hz (called a train-of-four response). Measured this way, acceptable CMAP monitoring has been conducted with only two out of four responses remaining [31, 48]. For comparison of the two techniques, only one response of four is present when T1 is less than 10%, two twitches are present at 10–20%, and three twitches at 20–25% of the baseline T1 response [84]. When intense neuromuscular blockade is required (e.g., recording of epidural or neurogenic responses), T1 response less than 10% [85], or no more than two out of four twitches [37], has been recommended. Many clinicians use closed-loop control systems to monitor the twitch and control the infusion so that excessive blockade does not eliminate the ability to record or mimic loss of the response with neural injury [40, 43, 60, 64, 86, 87].

468

Tod B. Sloan

Because of varying muscle sensitivity to muscle relaxants, the neuromuscular blockade may need to be evaluated continuously in the same muscle groups used for monitoring. It is important to note that the use of neuromuscular blockade is controversial during monitoring of muscle responses from mechanical stimulation of nerves, and partial paralysis may reduce the ability to record these responses (e.g., facial nerve monitoring or monitoring for pedicle screw placement). Although recording of myogenic responses is possible with partial neuromuscular blockade, the amplitude of the CMAP will be reduced by the blockade. Studies suggest that the actual reduction varies from a linear reduction paralleling the percent T1 effect to a slightly decreased rate of reduction [88, 89]. As a consequence of the amplitude reduction, the ability to record with partial neuromuscular blockade will be dependent on other factors that reduce the myogenic response amplitude, such as anesthesia or neurologic disease. Hence, amplitude reduction with initially small responses or with anesthetic choices that markedly reduce amplitude may make the use of blockade more difficult. Fortunately, the CMAP amplitude is usually quite large. It is also important to recognize that the use of amplitude criteria for warning of impending neurological injury may not be possible because inevitable fluctuations in the degree of blockade may obscure the application of strict criteria.

4 CONCLUSION: ANESTHETIC CHOICE FOR MOTOR TRACT MONITORING These studies suggest that for monitoring of epidural D responses from transcranial stimulation, the sole anesthetic consideration is the use of adequate muscle relaxation to prevent paraspinal muscles from obscuring the epidural recordings. Because of its resistance to anesthesia, the D wave response should be remarkably stable if the anesthetic state fluctuates so that both amplitude and latency criteria are usable for determining neurophysiological change. If maintenance of I waves is desired, then the anesthetic choices are limited, as discussed in the previous text for recording of myogenic responses. To date, the use of I waves has not been described, although their loss might be indicative of ischemia in the motor cortex. Anesthesia for monitoring of peripheral muscle responses to spinal cord nerve root stimulation should also be unaffected by anesthetic choice, with the sole exception of neuromuscular blockade. If the responses are dependent on sensory tract stimulation (e.g., monitoring of reflex activity through the spinal cord), then anesthetic choice must consider the effects on the α-motoneuron and the synapses involved in the pathway.

Anesthesia and Motor Evoked Potential Monitoring

469

Anesthesia choice for recording of neurogenic or myogenic responses from spinal stimulation has been described clinically. Neuromuscular blockade is clearly important (as with epidural recording) to reduce the influence of overlying muscle activity. The studies presented here, particularly those of Mochida [34], suggest that anesthesia may play a very important role in determining the contributions of sensory and motor pathways to monitoring responses following spinal stimulation. Since the type and intensity of stimulation may vary between different clinically used protocols, it is difficult to make anesthetic recommendations that would allow preferential recording of motor tract responses. Clearly, the choice of anesthesia makes a marked difference in the ability to record myogenic (and presumably neurogenic) responses following transcranial stimulation of the motor tracts. Because these responses are exquisitely sensitive to a large variety of anesthetic agents, it appears that the best technique for monitoring is a total intravenous technique. Current drug combinations usually include opioids with ketamine, etomidate, or closely titrated propofol infusions [30, 31, 66, 40–42, 59]. Although neuromuscular blockade reduces the amplitude of the muscle response, a controlled degree of blockade (10–20% of single twitch remaining, or two out of four twitches remaining in a train-offour response) is highly desirable to reduce patient motion and facilitate some procedures. A tightly controlled muscle relaxant infusion is needed to accomplish this to avoid excessive blockade, which would hamper monitoring. In the circumstances of anesthetic and neuromuscular blockade reduction in amplitude, the amplitude of the myogenic response will inevitably fluctuate during the procedure. Hence, warning criteria may need to be less dependent on amplitude and more dependent on onset-latency or the simple presence or absence of the response. Perhaps newer transcranial stimulation or response facilitation techniques will allow a more liberal anesthesia use. Although high-frequency stimulation would appear to allow the use of depressant agents (notably, low-dose inhalational agents), the authors of clinical studies using this technique recommend TIVA. Clearly, this is an area of anesthesia and monitoring that awaits advances to allow for a wider application of this monitoring technique.

REFERENCES 1. Sloan, T. (1997). Evoked potentials. In “Textbook of neuroanesthesia with neurosurgical and neuroscience perspectives” (M.A. Albin, ed.), pp. 221–276. McGraw-Hill, New York. 2. Hufnagel, A., Elger, C., Marx, W., and Ising, A. (1990). Magnetic motor-evoked potentials in epilepsy: Effects of the disease and of anticonvulsant medication. Ann. Neurol., 28, 680–686. 3. Owen, J. (1997). Applications of neurophysiological measures during surgery of the spine. In “The adult spine: Principles and practice,” 2nd ed. (J.F. Frymoyer, ed.), pp. 673–702. Lippincott-Raven Publishers, New York.

470

Tod B. Sloan

4. Minahan, R.E., Sepkuty, J.P., Lesser, R.P., Sponseller, P.D., and Kostuik, J.P. (2001). Anterior spinal cord injury with preserved neurogenic “motor” evoked potentials. Clin. Neurophysiol., 112, 1442–1450. 5. Deletis, V. (2001). The “motor” inaccuracy in neurogenic motor evoked potentials [editorial]. Clin. Neurophysiol., 112, 1365–1366. 6. Winters, W.D., Mori, K., Spooner, C.E., and Bauer, R.O. (1967). The neurophysiology of anesthesia. Anesthesiology, 28, 65–80. 7. Winters, W.D. (1976). Effects of drugs on the electrical activity of the brain: Anesthetics. Ann. Rev. Pharm. Toxicol., 16, 413–426. 8. Firsching, R., Heinen-Lauten, M., and Loeschke, G. (1991). Effects of halothane and nitrous oxide on transcranial magnetic evoked potentials. Anasthesiol. Intensivmed. Notfallmed. Schmerzther, 26, 381–383. 9. Stone, J., Ghaly, R., Levy, W., Kartha, R., Krinsky, L., and Roccaforte, P. (1992). A comparative analysis of enflurane anesthesia on primate motor and somatosensory evoked potentials. Electroencephalogr. Clin. Neurophysiol., 84, 180–187. 10. Sloan, T., and Angell, D. (1993). Differential effect of isoflurane on motor evoked potentials elicited by transcortical electric or magnetic stimulation. In “Handbook of spinal cord monitoring” (S.S. Jones, S. Boyd, M. Hetreed, and N.J. Smith, eds.), pp. 362–367. Kluwer Academic Publishers, Hingham, MA. 11. Zentner, J., Albrecht, T., and Heuser, D. (1992). Influence of halothane, enflurane, and isoflurane on motor evoked potentials. Neurosurgery, 32, 298–305. 12. Haghighi, S., Madsen, R., Green, K., Weston, P.F., Boyd, S.G., and Hall, G.M. (1990). Suppression of motor evoked potentials by inhalation anesthetics. J. Neurosurg. Anesth., 2, 73–76. 13. Loughnan, B., Anderson, S., Hetreed, M., Weston, P.F., Boyd, S.G., and Hall, G.M. (1989). Effects of halothane on motor evoked potentials recorded in the extradural space. Br. J. Anaesth., 63, 561–564. 14. Kalkman, C., Drummond, J., and Ribberink, A. (1991). Low concentrations of isoflurane abolish motor evoked responses to transcranial electrical stimulation during nitrous oxide/opioid anesthesia in humans. Anesth. Analg., 73, 410–415. 15. Hicks, R., Woodforth, I., Crawford, M., Stephen, J.P., and Burke, D.J. (1992). Some effects of isoflurane on I waves of the motor evoked potential. Br. J. Anaesth., 69, 130–136. 16. Haghighi, S., Green, K., Oro, J., Drake, R.K., and Kracke, G.R. (1990). Depressive effect of isoflurane anesthesia on motor evoked potentials. Neurosurgery, 26, 993–997. 17. Calancie, B., Klose, K.J., Baier, S., and Green, B.A. (1991). Isoflurane-induced attenuation of motor evoked potentials caused by electrical motor cortex stimulation during surgery. J. Neurosurg., 74, 897–904. 18. Schmid, U.D., Boll, J., Siechti, S., Schmid, J., and Hess, C.W. (1992). Influence of some anesthetic agents on muscle responses to transcranial magnetic cortex stimulation: A pilot study in humans. Neurosurgery, 30, 85–92. 19. Yamada, H., Transfeldt, E.E., Tamaki, T., Torres, F., and Iaizzo, P.A. (1994). The effects of volatile anesthetics on the relative amplitudes and latencies of spinal and muscle potentials evoked by transcranial magnetic stimulation. Spine, 19, 1512–1517. 20. Deletis, V. (1993). Intraoperative monitoring of the functional integrity of the motor pathways. In “Advances in neurology: Electrical and magnetic stimulation of the brain” (O. Devinsky, A. Beric, and M. Dogali, eds.), pp. 201–214. Raven Press, New York. 21. Glassman, S.D., Shields, C.B., Linden, R.D., Zhang, Y.P., Nixon, A.R., and Johnson, J.R. (1993). Anesthetic effects on motor evoked potentials in dogs. Spine, 18, 1083–1089. 22. Thompson, P.D., Day, B.L., Crockard, H.A., Calder, I., Murray, A.M., Rothwell, J.C., and Marsden, C.D. (1991). Intra-operative recording of motor tract potentials at the cervicomedullary junction following scalp electrical and magnetic stimulation of the motor cortex. J. Neurol. Neurosurg. Psychiatry, 54, 618–623.

Anesthesia and Motor Evoked Potential Monitoring

471

23. Mavroudakis, N., Vandesteene, A., and Brunko, E. (1994). Spinal and brain-stem SEPs and H reflex during enflurane anesthesia. Electroencephalogr. Clin. Neurophysiol., 92, 82–85. 24. Gugino, L.D., Aglio, L.S., Segal, N.E., Gonzalez, A.A., Kraus, K.H., and Woodard, E.J. (1997). Use of transcranial magnetic stimulation for monitoring spinal cord motor paths. Semin. Spine Surg., 9, 315–336. 25. Stephen, J.P., Sullivan, M.R., Hicks, R.G., Burke, D.J., Woodforth, I.J., and Crawford, M.R. (1996). Cotrel-Dubousset instrumentation in children using simultaneous motor and somatosensory evoked potential monitoring. Spine, 21, 2450–2457. 26. Amassian, V.E., Stewart, M., Quirk, G.J., and Rosenthal, J.L. (1987). Physiological basis of motor effects on a transient stimulation to cerebral cortex. Neurosurgery, 20, 74–93. 27. Loughnan, B.A., and Fennelly, M.E. (1995). Spinal cord monitoring. Anaesthesia, 50, 101–102. 28. Taniguchi, M., Cedzich, C., and Schramm, J. (1993). Modification of cortical stimulation for motor evoked potentials under general anesthesia: Technical description. Neurosurgery, 32, 219–221. 29. Kawaguchi, M., Sakamoto, T., Ohnishi, H., Shimizu, K., Karasawa, J., and Furuya, H. (1996). Intraoperative motor evoked potentials induced by direct electrical stimulation of the motor cortex under isoflurane and sevoflurane. Anesth. Analg., 82, 593–599. 30. Pechstein, U., Nadstawek, J., Zentner, J., and Schramm, J. (1998). Isoflurane plus nitrous oxide versus propofol for recording of motor evoked potentials after high frequency repetitive electrical stimulation. Electroencephalogr. Clin. Neurophysiol., 108, 175–181. 31. Pechstein, U., Cedzich, C., Nadstawek, J., and Schramm, J. (1996). Transcranial high-frequency repetitive electrical stimulation for recording myogenic motor evoked potentials with the patient under general anesthesia. Neurosurgery, 39, 335–344. 32. Jones, S.J., Harrison, R., Koh, K.F., Mendoza, N., and Crockard, H.A. (1996). Motor evoked potential monitoring during spinal surgery: Responses of distal limb muscles to transcranial cortical stimulation with pulse trains. EEG Clin. Neurophysiol., 100, 375–383. 33. Taylor, B.A., Fennelly, M.E., Taylor, A., and Farrell, J. (1993). Temporal summation: The key to motor evoked potential spinal cord monitoring in humans. J. Neurol., 56, 104–106. 34. Mochida, K., Shinomiya, K., Komori, H., and Furuya, K. (1995). A new method of multisegment motor pathway monitoring using muscle potentials after train spinal stimulation. Spine, 20, 2240–2246. 35. Taniguchi, M., Schramm, J., and Cedzich, C. (1991). Recording of myogenic motor evoked potential (mMEP) under general anesthesia. In “Intraoperative neurophysiological monitoring” ( J. Schramm, and A.R. Moller, eds.), pp. 72–87. Springer-Verlag, Berlin. 36. Sloan, T.B., and Rogers, J.N. (1996). Inhalational anesthesia alters the optimal interstimulus interval for multipulse transcranial motor evoked potentials in the baboon. J. Neurosurg. Anesth., 8, 346. 37. Schwentker, M.C., Russell, G.B., Rodichok, L.D., Segal, L.S., Schwentker, E.P., and Blackburn, T.W. (1995). Myogenic response distortion of neurogenic motor evoked potential morphology. Anesthesiology, 83, 616–619. 38. Jellinek, D., Platt, M., Jewkes, D., and Symon, L. (1991). Effects of nitrous oxide on motor evoked potentials recorded from skeletal muscle in patients under total anesthesia with intravenously administered propofol. Neurosurgery, 29, 558–562. 39. Kalkman, C.J., Drummond, J.C., Kennelly, N.A., Patel, P.M., and Partridge, B.L. (1992). Intraoperative monitoring of tibialis anterior muscle motor evoked responses to transcranial electrical stimulation during partial neuromuscular blockade. Anesth. Analg., 75, 584–589. 40. Lang, E.W., Beutler, A.S., Chestnut, R.M., Patel, P.M., Kennelly, N.A., Kalkman, C.J., Drummond, J.L., and Garfin, S.R. (1996). Myogenic motor-evoked potential monitoring using partial neuromuscular blockade in surgery of the spine. Spine, 21, 1676–1686. 41. Lang, E.W., Chestnut, R.M., Butler, A.S., Kennelly, N.A., and Renaudin, J.W. (1996). The utility of motor-evoked potential monitoring during intermediary surgery. Anesth. Analg., 83, 1337–1341.

472

Tod B. Sloan

42. Bags, L.H., Kalkman, C.J., Been, H.D., and Drummond, J.C. (1996). The use of a circumferential cathode improves amplitude of intraoperative electrical transcranial myogenic motor evoked responses. Anesth. Analg., 82, 1011–1014. 43. Lee, W., Hou, W., Yang, L., and Lin, S. (1995). Intraoperative monitoring of motor function by magnetic motor evoked potentials. Neurosurgery, 36, 493–500. 44. Zentner, J. (1989). Noninvasive motor evoked potential monitoring during neurosurgical operations on the spinal cord. Neurosurgery, 24, 709–712. 45. Glassman, S.D., Zhang, P., Shields, C.B., Johnson, J.R., and Linden, R.D. (1995). Transcranial magnetic motor-evoked potentials in scoliosis surgery. Orthopedics, 18, 1017–1023. 46. Zentner, J. (1991). Motor evoked potential monitoring during neurosurgical operations on the spinal cord. Neurosurg. Rev., 14, 29–36. 47. Shields, C.B., Paloheimo, M.P.J., Backman, M.H., Edmonds, H.L. Jr., and Johnson, J.R. (1990). Intraoperative use of transcranial magnetic motor evoked potentials. In “Magnetic stimulation in clinical neurophysiology” (S. Chokroverty, ed.), Chap. 13, pp. 173–184. Butterworths, Boston. 48. Calancie, B., Harris, W., Broton, J.G., Alexeeva, N., and Green, B.A. (1998). “Threshold-level” multipulse transcranial electrical stimulation of motor cortex for intraoperative monitoring of spinal motor tracts: Description of method and comparison to somatosensory evoked potential monitoring. J. Neurosurg., 88, 457–470. 49. Sloan, T., Rogers, J., Rogers, J., and Sloan, H. (1995). MAC fractions of nitrous oxide and isoflurane are not electrophysiologically additive in the ketamine anesthetized baboon. J. Neurosurg. Anesth., 7, 314. 50. Gugino, L.D., Kraus, K., Heino, R., Aglio, L.S., Levy, W.J., Cohn, L., and Maddi, R. (1992). Peripheral ischemia as a complicating factor during somatosensory and motor evoked potential monitoring of aortic surgery. J. Cardiovasc. Anesth., 6, 715–719. 51. Kalkman, C.J., Drummond, J.C., Ribberink, A.A., Patel, P.M., Sano, T., and Bickford, R.G. (1992). Effects of propofol, etomidate, midazolam, and fentanyl on motor evoked responses to transcranial electrical or magnetic stimulation in humans. Anesthesiology, 76, 502–509. 52. Ghaly, R.F., Stone, J.L., Levy, W.J., Krinsky, L., and Asokan, A. (1991). The effect of neuroleptanalgesia (droperidol-fentanyl) on motor potentials evoked by transcranial magnetic stimulation in the monkey. J. Neurosurg. Anesth., 3, 117. 53. Lee, V.C. (1994). Spinal and cortical evoked potential studies in the ketamine-anesthetized rabbit: Fentanyl exerts component-specific, naloxone-reversible changes dependent on stimulus intensity. Anesth. Analg., 78, 280–286. 54. Velasco, M., Velasco, F., Castaneda, R., and Sanchez, R. (1984). Effect of fentanyl and naloxone on human somatic and auditory-evoked potential components. Neuropharmacology, 23, 359–366. 55. Chi, O.Z., McCoy, C.L., and Field, C. (1987). Effects of fentanyl anesthesia on visual evoked potentials in humans. Anesthesiology, 67, 827–830. 56. Schubert, A., Licina, M.G., and Lineberry, P.J. (1990). The effect of ketamine on human somatosensory evoked potentials and its modification by nitrous oxide. Anesthesiology, 72, 33–39. 57. Chabal, C., Jacobson, L., and Little, J. (1988). Effects of intrathecal fentanyl and lidocaine on somatosensory evoked potentials, the h reflex, and clinical responses. Anesth. Analg., 67, 509–513. 58. Pathak, K.S., Brown, R.H., and Nash, C.L. Jr. (1983). Continuous opioid infusion for scoliosis fusion surgery. Anesth. Analg., 62, 841–845. 59. de Haan, P., Kalkman, C.J., de Mol, B.A., Ubags, L.H., Veldman, D.J., and Jacobs, M.J. (1997). Efficacy of transcranial motor-evoked myogenic potentials to detect spinal cord ischemia during operations for thoracoabdominal aneurysms. J. Thorac. Cardiovasc. Surg., 113, 87–101.

Anesthesia and Motor Evoked Potential Monitoring

473

60. Herdmann, J., Lumenta, C.B., and Huse, K.O. (1993). Magnetic stimulation for monitoring of motor pathways in spinal procedures. Spine, 18, 551–559. 61. Levy, W.J., York, D.H., McCaffrey, M., and Tanzer, F. (1984). Motor evoked potentials from transcranial stimulation of the motor cortex in humans. Neurosurgery, 15, 287–302. 62. Morota, N., Deletis, V., Constantini, S., Kofler, M., Cohen, H., and Epstein, F.J. (1997). The role of motor evoked potentials during surgery for intramedullary spinal cord tumors. Neurosurgery, 41, 1327–1336. 63. Nagle, K.J., Emerson, R.G., Adams, D.C., Heyer, E.J., Roye, D.P., Schwab, F.J., Weidenbaum, M., McCormick, P., Pile-Spellman, J., Stein, B.M., Farcy, J.P., Gallo, E.J., Dowling, K.C., and Turner, C.A. (1996). Intraoperative monitoring of motor evoked potentials: A review of 116 cases. Neurology, 47, 999–1004. 64. Stinson, L.W., Murray, M.J., Jones, K.A., Assef, S.J., Burke, M.J., Behrens, T.L., and Lennon, R.L. (1994). A computer-controlled, closed-loop infusion system for infusing muscle relaxants: Its use during motor-evoked potential monitoring. J. Cardiothorac. Vasc. Anesthesia, 8, 40–44. 65. Taniguchi, M., Nadstawek, J., Langenbach, U., Bremer, F., and Schramm, J. (1993). Effects of four intravenous anesthetic agents on motor evoked potentials elicited by magnetic transcranial stimulation. Neurosurgery, 33, 407–415. 66. Watt, J.W.H., Frazer, M.H., Soni, B.M., Sett, P.K, and Clay, R. (1996). Total i.v. anaesthesia for transcranial magnetic evoked potential spinal cord monitoring. Br. J. Anaesth., 76, 870–871. 67. Yang, L.H., Lin, S.M., and Lee, W.Y. (1994). Intraoperative transcranial electrical motor evoked potential monitoring during spinal surgery under intravenous ketamine or etomidate anaesthesia. Acta. Neurochir. (Wien), 127, 191–198. 68. Zentner, J. (1991). Motor evoked potential monitoring in operations of the brainstem and posterior fossa. In “Intraoperative neurophysiol monitoring” ( J. Schramm, and A.R. Moller, eds.), pp. 95–105. Springer-Verlag, Berlin. 69. Kano, T., and Shimoji, K. (1974). The effects of ketamine and neuroleptanalgesia on the evoked electrospinogram and electromyogram in man. Anesthesiology, 40, 241–246. 70. Shimoji, K., and Kano, T. (1973). Evoked electrospinogram: Interpretation of origin and effects of anesthetics. In “Brain unit activity during behavior” (M.I. Phillips, ed.), pp. 171–190. Charles C. Thomas, Springfield. 71. Ghaly, R.F., Stone, J.L., Aldrete, J.A., and Levy, W.J. (1990). Effects of incremental ketamine hydrochloride doses on motor evoked potentials (MEPs) following transcranial magnetic stimulation: A primate study. J. Neurosurg. Anesth., 2, 79. 72. Kothbauer, K., Schmid, U.D., Liechti, S., and Rosler, K.M. (1993). The effect of ketamine anesthetic induction on muscle responses to transcranial magnetic cortex stimulation studied in man. Neurosci. Lett., 154, 105–108. 73. Lida, H., Dohi, S., Tanahashi, T., and Takenaka, M. (1997). Spinal conduction block by intrathecal ketamine in dogs. Anesth. Analg., 85, 106–110. 74. Kalkman, C.J., Drummond, J.C., and Patel, P.M. (1994). Effects of droperidol, pentobarbital, and ketamine on myogenic transcranial magnetic motor-evoked responses in humans. Neurosurgery, 35, 1066–1071. 75. Ghaly, R.F., Stone, J.L., Levy, W.J., Kartha, R., Miles, M.L., and Jaster, H.J. (1990). The effects of etomidate or midazolam hypnotic dose on motor evoked potentials in the monkey. J. Neurosurg. Anesthesiol., 2, 244. 76. Schonle, P.W., Isenberg, C., Crozier, T.A., Dressler, D., Machetanz, J., and Conrad, B. (1989). Changes of transcranially evoked motor responses in man by midazolam, a short acting benzodiazepine. Neurosci. Lett., 101, 321–324. 77. Rampil, I.J. (1997). Electroencephalogram. In “Textbook of neuroanesthesia with neurosurgical and neuroscience perspectives” (M.A. Albin, ed.), Chap. 6, pp. 193–220. McGraw-Hill, New York.

474

Tod B. Sloan

78. Kochs, E., Treede, R.D., and Schulte am Esche, J. (1986). Increase of somato sensorially evoked potentials during induction of anaesthesia with etomidate. Anaesthetist, 35, 359–364. 79. Sloan, T.B., Ronai, A.K., Toleikis, J.R., and Koht, A. (1988). Improvement of intraoperative somatosensory evoked potentials by etomidate. Anesth. Analg., 67, 582–585. 80. Sloan, T., and Levin, D. (1993). Etomidate amplifies and depresses transcranial motor evoked potentials in the monkey. J. Neurosurg. Anesth., 5, 299. 81. Lumenta, C.B. (1991). Effect of etomidate on motor evoked potentials in monkeys. Neurosurgery, 29, 480–482. 82. Angel, A., and LeBeau, F. (1992). A comparison of the effects of propofol with other anaesthetic agents on the centripetal transmission of sensory information. Gen. Pharmac., 23, 945–963. 83. Keller, B.P., Haghighi, S.S., Oro, J.J., and Eggers, G.W.N. Jr. (1992). The effects of propofol anesthesia on transcortical electric evoked potentials in the rat. Neurosurgery, 30, 557–560. 84. Law, S.C., and Cook, D.R. (1990). Monitoring the neuromuscular junction. In “Clinical monitoring” (C.L. Lake, ed.), Chap. 21, pp. 719–755. W.B. Saunders, Philadelphia. 85. Adams, D.C., Emerson, R.G., Heyer, E.J., McCormick, P.C., Carmel, P.W., Stein, B.M., Farcy, J.P., and Gallo, E.J. (1993). Monitoring of intraoperative motor-evoked potentials under conditions of controlled neuromuscular blockade. Anesth. Analg., 77, 913–918. 86. Kalkman, C.J., Drummond, J.C., and U, H.S. (1994). Severe sensory deficits with preserved motor function after removal of a spinal arteriovenous malformation: Correlation with simultaneously recorded somatosensory and motor evoked potentials. Anesth. Analg., 78, 165–168. 87. Tabaraud, F., Boulesteix, J.M., Moulies, D., Longis, B., Lansade, A., Terrier, G., Vallat, J.M., Dumas, M., and Hugont, J. (1993). Monitoring of the motor pathway during spinal surgery. Spine, 18, 546–550. 88. Sloan, T.B., and Erian, R. (1993). Effect of atracurium induced neuromuscular blockade on cortical motor evoked potentials. Anesth. Analg., 76, 979–984. 89. Sloan, T.B., and Erian, R. (1993). Effect of vecuronium induced neuromuscular blockade on cortical motor evoked potentials. Anesthesiology, 78, 966–973.

INDEX

The letter f appended to a page reference indicates a figure; the letter t refers to a table. 1-methyl-4phenyl-1,2,3,6-tetrahydropyridine (MPTP), 410

A α-motoneuron anesthetic effects on, 368, 454 excitability level of, 42, 44, 47 and corticospinal tract, 15, 18, 28, 211 Abducens nerve, 271 Abductor hallucis brevis, use in recording MEPs, 42 Abductor pollicis brevis, use in recording MEPs, 41 AC-PC, 425t AC/PC line, 424, 427, 431, 438f ACA, 351, 387, 393 Accessory nerve, palsy of, 189 Acoustic nerve, 292, 303, 309 Acoustic tumor, 295, 298, 304, 315 Adamkiewicz’s artery, 137 Affective disorders, 406 Albe-Fessard, 408 Alfentanil, effects on intraoperative monitoring, 461 Amplification differential, 420–421 impedance testing and, 421 preamplification, 420–421 referential, 420

Amplitude gradient, 157, 158f, 160–162 Amytal test, for spinal cord ischemia, 130–132 Anal M wave, 204f Anatomic target, 406 Anesthesia agents of, 238, 241, 455–468 and [alpha]-motoneuron synapses, 454 and bipolar cortex stimulation technique, 389–390 effects on intraoperative monitoring, 31–34, 238–239, 389 effects on motor cortical stimulation, 453–454 effects on neuromuscular junction, 454–455 effects on spinal stimulation, 452–453 electrode montage in, 30f, 31–33 halogenated inhalational agents, 455–458 intravenous agents, 459–466 methods of, 79 for motor techniques, 248–249 muscle relaxants, 466–468 nitrous oxide, 459–460 target for action of, 452 Aneurysmal bone cysts, embolization therapy for, 136 475

476 Aneurysms MEP monitoring in surgery for, 385–387, 385f, 386f SEP monitoring in surgery for, 350–351, 351t, 385 value of intraoperative monitoring in, 390–391 Angiography anatomy for, 133–134 anesthesia for, 135 of dural/extradural lesions, 137–140 of intradural lesions, 140–147 neurophysiological monitoring and, 35–147 of vascular tumors, 135–136 Anodal stimulation, 9, 369 Anorectum, innervation of, 200–201 Ansa lenticularis (AL), 429 Anterior spinal artery (ASA) effects on MEP’s, 121–122, 123f and angiographic evaluation, 135–136, 141f, 144f Arteriovenous malformations (AVMs) MEP monitoring in surgery for, 380–382 value of intraoperative monitoring for, 392 Arteriovenous malformation (AVM), 123f, 130–141, 140–142, 143f Ascending cervical artery, 134 Ascending occipital artery, 134 Ascending pharyngeal artery, 134 Astrocytoma appearance of, 65 case studies of, 85, 88 clinical presentation of, 57 diagnosis of, 58, 62–63f incidence of, 56, 57f in midbrain, 274–275, 280–281f prognosis after surgery, 68 surgical treatment of, 65–66, 274–275 Auditory nerve monitoring, 292 directly from cochlear nucleus, 306, 307f, 308 directly from exposed nerve, 305–306, 306f electrode placement for, 303 filtering for, 304–305 interference reduction in, 302–303 rationale for, 301 stimuli for, 303 techniques for, 301–302 Auditory-vestibular nerve, mapping of, 313–314

Index

B Baclofen, to treat spasticity, 101, 103 Barbiturates, effects on intraoperative monitoring, 462–463 Basal ganglia surgery history of, 407 theoretical basis for, 410–412, 410f Benzodiazepines, effects on intraoperative monitoring, 463 Bilateral lateral sacral artery, 134 Bilateral supreme intercostal artery, 134 Bipolar cortex stimulation technique advantages and disadvantages of, 365 adverse effects of, 388 anesthesia and, 389–390 applications of, 364–365 cortical mapping by, 364 recording in, 364, 367t stimulation in, 363, 367t subcortical stimulation by, 364–365 Blocked impulses characteristics of, 5f, 6–8 from cord trauma, 8, 9f Border cell, 428–429 Brachial plexus, 108, 118, 154, 185 Brain shift, 379, 432 Brainstem anatomy of, 269–271, 271f entry routes into, 272f tumor localization in, 272–274 Brainstem auditory evoked potentials (BAEPs) allowable variation in, 308–309 enhancement of, 304f filtering of, 304–305 monitoring of, 309 recording of, 301–305 Brainstem lesion surgery anatomic background for, 270 approaches to, 274–279 complications of, 284–285 dangers of, 272–273, 273f history of, 268–269 neurophysiological monitoring in, 285–286 patient selection for, 269 postoperative care in, 279, 284–285 rationale for, 269 strategy for, 269–274

Index Brainstem mapping (BSM) anesthesia in, 322 case study of, 328, 329f–330f, 331, 331f, 332f, 333f, 334f clinical application of, 331–332 clinical limitations of, 327 described, 320, 321f emergence of, 320 results of, 323–324 surgical implications of, 325–326, 325f technique of, 322 Bulbocavernosus reflex (BCR), 203 monitoring of, 128–129, 211–212, 215 Burster cell, 428f, 436f

C Campotomy, 407 Carotid artery, 341 Cathodal stimulation, 7, 10, 369 Cavernomas, 393 Cavernous malformations, 142 diagnosis of, 56 surgical treatment of, 67 Cavitron ultrasonic aspirator, 60–61 Central sulcus, localization of, 354, 355f, 356, 357f Central tumors difficulties posed by, 392–393 MEP monitoring in surgery for, 380–382 SEP monitoring in surgery for, 352–353 value of intraoperative monitoring in surgery for, 394 Cerebral palsy baclofen to treat, 103 classification of, 103 intraoperative sacral monitoring in, 212–216 MDT to treat, 104 outcome assessment for, 226–228 rhizotomy to treat, 95, 96f, 103 SDR rhizotomy to treat, 94–95, 96f, 219–228 Cerebrospinal fluid, postsurgical leakage of, 69–70 Cervicomedullary junction, 121, 279, 282f Cervicomedullary tumors, cranial nerve motor nucleus displacement by, 323, 324f, 326, 326f CMAPs (compound muscle action potentials), 28, 220 intraoperative use of, 221–223

477 monitoring in pedicle screw placement, 240–241 CNAPs (compound nerve action potentials), 170, 183f anesthetics and, 182 characteristics of, 184 electrodes for recording and stimulating, 178–181 evaluation of, 183–185 to grade nerve injury, 185–187 intraoperative recording of, 182–183, 187–190 in spinal nerve injuries, 185 Cochlear nucleus, 301, 306f, 308 Common mode rejection ratio (CMRR), 421 Conducted impulses characteristics of, 5–6, 5f generation of, 8–12, 155 Constant current, 76, 243–244, 245f, 373f Constant voltage, 244, 245f, 294, 423t Contact laser, 60, 63–66 Cooper, 407 Corpus striatum, and Parkinson’s disease, 411 Cortical mapping using MEP mapping, 366 using Penfield’s technique, 364 Corticobulbar tract, 327, 328f Corticospinal tract (CT) waves, 4–5 D waves, 5–6, 5f I waves, 5f, 6 monitoring, 7–8 stimulus intensities and, 10f Cranial motor nerve monitoring benefits of, 315 of CN III, IV, VI, 298–299 of CN IX–XII, 299–300 of facial nerve, 293–298 risks of, 300 of trigeminal nerve, 298 Cranial motor nerves, displacement by tumor, 323, 324f Cranial nerve III, motor portion of, monitoring of, 298–299 Cranial nerve IV, monitoring of, 298–299 Cranial nerve VI, monitoring of, 298–299 Current shunting, 252, 253f, 294 CUSA, 59, 65–66, 89 Cystoprostatectomy, adverse effects of, 203–204

478 D D volley, 18 D waves, 5–6, 5f, 75 anesthesia and, 453 desynchronization of, 40–41, 41f eliciting, 7f, 8–12, 15f, 76 factors influencing recording of, 37–38, 39f, 40, 40f generation of, 45–46, 47f interpretation of, 82–83 latency of, 43 neurophysiology of, 76–77 number of, 44–45 recording, 7–8 recording using single-pulse TES, 34–36 recovery of, 42–43, 42f site of, 9 Deep brain stimulation (DBS), 408 advantages of, 408 studies on, 409 Deep cervical artery, 134 Dermatomal SEPs (DSEPs), 236–237 to aid pedicle screw placement, 255–259 Desflurane, effects on intraoperative monitoring, 453 Desynchronization, 8, 40, 41f, 82 Differential amplifier, 421 Direct electrical stimulation, 251, 368 Direct nerve root stimulation, of MEPs, 243–244 Direct pathway, 410f, 411 Distal cubital tunnel syndrome, 189 Dorsal columns mapping of, 157, 158f, 161f, 163f midline determination, 154, 164 Dorsal horn electrode recordings of, 105, 110–111 surgery on. See MDT Dorsal rhizotomy. See Posterior rhizotomy Dorsal root entry zone (DREZ), 97, 98f indications for surgery on, 113–114 laser surgery on, 112 microsurgery on. See MDT (microDREZotomy) RF thermocoagulation in, 111–112 ultrasonic procedures on, 113 variations in, 107f Dorsospinal artery, 134 Double crush syndrome, 189 DRAPs (dorsal root action potentials) elicitation of, 206

Index of pudendal nerve, 203, 207–208, 209f, 210, 215 Droperidol, effects on intraoperative monitoring, 466 Dural/extradural vascular malformations angiography of, 137 treatment of, 137, 139–140 types of, 137 Dystonia, 406 globus pallidus procedures to treat, 427

E EEG (electroencephalography) intraoperative use of, 27 Eighth nerve, 306, 308 Electrical noise compensating for, 190–191 sources of, 191–192 Electrodes percutaneous placement of, 36–37 placement after laminectomy or flavectomy, 37 for recording D and I waves, 34–36, 39f, 40f for peripheral nerve monitoring, 178–181 for recording, 418–420, 420f for sacral monitoring, 207 EMG (electromyography) of facial nerve, 295–298 for intraoperative monitoring, 5, 108 origin of, 27 Endovascular embolization anatomy for, 133–134 angiographic evaluation for, 135–137 indications for, 133 for tumors, 135–136, 140f–141f for vascular malformations, 137–147 Enflurane, effects on intraoperative monitoring, 455 Entrapment neuropathy, assessment for, 189 Ependymoma appearance of, 65 case study of, 87 clinical presentation of, 57 diagnosis of, 58, 59–61f incidence of, 56, 59f in midbrain, 282f–283f prognosis after surgery, 68 surgical treatment of, 66–67 Epidural electrode, 35, 38, 80, 157, 206

479

Index Epidural MEPs, 76–77, 373 Epilepsy, stimulation mapping and, 357, 359, 361–362 Erectile dysfunction diagnostic testing of, 202–203 neurogenic causes of, 201–202 Essential tremor, 406 Ethanol, as embolic agent, 136 Etomidate, effects on intraoperative monitoring, 463–464, 464f,465f Extraoperative function mapping with grid or multiple-strip electrodes indications for, 357–358 using stimulation technique, 361–362 techniques for, 358–360 types of, 358

F Facial nerve monitoring, 292 goals of, 293–295 neural conduction, 297 recording EMG potentials, 295–298 Fentanyl, effects on intraoperative monitoring, 461 Fibrillary astrocytoma, incidence of, 57f Fifth nerve, 311 Filtering, 301–305, 367t First dorsal interosseus muscle, 14–15 Flavectomy, electrode placement after, 37 Flavotomy, electrode placement after, 37 Floor of fourth ventricle mapping of, 310–311, 310f, 315 recording from, 306, 307f, 308 stimulation of, 323–324 Fluoroscopy, 62, 232 Forearm flexors, use in recording MEPs, 42 Forel’s Field H (H2), 434–435 Functional Inventory Measure for Children (WeeFIM), 227–228 Functional mapping, 353–354 extraoperative, with grid or multiple-strip electrodes, 357–362 intraoperative, of motor cortex and white matter, 362–365 SEP phase reversal in, 354–357

G Ganglioglioma appearance of, 65 case study of, 85, 87

diagnosis of, 58 incidence of, 56, 57f surgical treatment of, 65–66 Giant-cell tumor, embolization therapy for, 136 Glioblastoma incidence of, 57f prognosis after surgery, 68 Glioma brainstem, 268 operability of, 269 Glioma, 383f, 384f incidence of, 57f insular, 392–393 prognosis after surgery, 68 Globus pallidus pars externa, cell types of, 428 Globus pallidus pars interna deep brain stimulation of, 426 and Parkinson’s disease, 411 postventral pallidotomy, 426–427 recording of, 428–431, 428f Glossopharyngeal nerve (CN IX), motor portion of, monitoring of, 300 GPe, 409f, 410–411, 427–430 GPi, 409f, 410–411, 425t, 426–430 Grid electrode, 30f, 353, 358 Ground, 176–177, 191, 343, 375

H H reflex, 28, 221, 456 Halogenated anesthetic agents effects on intraoperative monitoring, 456–459, 456f, 457f,458f types of, 455–456 Halogenated inhalational agents, 456 Halothane, effects on intraoperative monitoring, 455 Hamstring hypertrophy, nerve entrapment caused by, 190 Hemangioblastoma diagnosis of, 58, 64–65f incidence of, 57f surgical treatment of, 67 vascularization of, 135 Hemangioma, embolization therapy for, 136 Hemangiopericytoma cervical spinal, 138f–139f embolization therapy for, 136

480 Hemifacial spasm incidence of, 311 intraoperative monitoring for, 311–313, 311f, 312f Hemiplegia, MDT to treat, 100 Hypoglossal nerve (CN XII), monitoring of, 300

I I waves, 5f, 6 anesthesia and, 453–454 eliciting by extrinsic inputs, 12–15 eliciting by intrinsic inputs, 12, 15–18, 16f, 17f facilitation of, 44, 44f factors influencing recording of, 37–38, 39f number of, 44–45 Iliolumbar artery, 134 Impedance, 11, 35–36, 206, 423t Impedance change monitoring, 414 Impedance testing, 254, 421 Indirect nerve root stimulation, of MEPs, 244–245, 247 Indirect pathway, 410f, 411 Insular lesions, 382, 394 Insular tumors, 392–393 MEP monitoring in surgery for, 382–385 Intercostal artery (ICA), 123f, 134 Internal capsule (IC), 32–33, 352–353, 384f, 407, 428 Internal iliac artery, 134 Interstimulus interval, 15f, 43f, 128, 211, 371–372, 373f Interstimulus interval (ISI), 15f, 42, 43f Intradural vascular malformations classification of, 140 neurophysiological monitoring for, 140–141, 141f, 142f–146f Intramedullary lipoma, surgical treatment of, 67–68 Intramedullary neoplasms clinical presentation of, 55–56 diagnostic studies of, 58–59 epidemiology of, 56 prognosis after surgery, 68–69 radiotherapy for, 69 surgical management of, 60–68 symptoms of, 58t

Index Intraoperative monitoring amplification in, 420–421 anesthesia and, 451–468 anesthetic choice for, 468–469 data organization for, 437, 448f electrodes for, 418–420 EMG for, 5 evolution of, 338–339 flowchart for, 193 indications for, 311–315 macroelectrode recording in, 414 microelectrode recording in, 416, 416f noise issues in, 191, 417–418 of peripheral nervous system, 175–181 recording sweep speed, 192 of sacral nervous system, 203–216 safety of, 387–388 semi-microelectrode technique in, 414–415, 415f stimulation techniques for, 422–423 stimulus artifact in, 192 targeting techniques in, 412–413 transcranial stimulation for, 4–5 troubleshooting of, 190–192 value and limitations of, 390–395. See also MEP monitoring; SEP monitoring Intraoperative neurophysiology (ION), 26–27 Intravenous analgesics, 460–461 barbiturates, 462–463 benzodiazepines, 463 droperidol, 466 etomidate, 463–464 ketamine, 461–462 opioids, 463 propofol, 464–465 sedative-hypnotic drugs, 463 Intubation, 79–81, 207, 322 Ischemia, spinal cord, 142f–143f assessment of, 127 endovascular correction of, 130–133 neurophysiological findings in, 124–126 preoperative procedures in, 127–129 provocative testing of, 130–132, 130f, 132t surgical treatment for, 126–133 Isoflurane, effects on intraoperative monitoring, 455, 456, 456f, 457f, 458f

Index

J Juvenile pilocytic astrocytoma, incidence of, 57f

K Ketamine, effects on intraoperative monitoring, 461–462 Kindling, 79

L Laminectomy electrode placement after, 37 technique for, 62–64 Laminotomy electrode placement after, 37 technique for, 62–64 Lasers in DREZ surgery, 112 in spinal cord surgery, 61 Leksell, 407, 413 Levodopa (L-Dopa), 407, 411–412 Levodopa-induced dyskinesia, 411–412 pallidotomy as treatment for, 412 Lipoma, intramedullary, 67–68 Localizer box, 424f Long forearm flexors, use in recording MEPs, 42 Longitudinal myelotomy, 97 Lower urinary tract, neural control of, 200 Lumbar artery, 134

M Macroelectrode recording, 414 Macroelectrode stimulation, 426, 430, 434 Mayfield, 61 MDT (micro-DREZotomy) anesthesia for, 106 cervical-level operative procedure, 106–108 described, 105–106 effects on SEPs, 109–111, 110f indications for, in adults, 101–102 indications for, in children, 102–104 lumbosacral–level operative procedure, 108–109 microdialysis in dorsal horn during, 111 radiofrequency thermocoagulation in, 111–112 results of, 100–101 technique of, 97, 98f, 99f, 100

481 to treat hemiplegia, 100 to treat neurogenic bladder, 101 to treat paraplegia, 97, 100 to treat spasticity, 97, 100, 102 Median sacral artery, 134 Medulla, and cervicomedullary junction, 279–285 Medullary pyramid, 17f, 18 Medullary tumor, 320, 323, 324f Medullary tumors, 280f–283f approaches to, 279 cranial nerve motor nucleus displacement by, 323, 324f, 325, 326f Meningioma, vascularization of, 135 MEP mapping advantages and disadvantages of, 366–367 applications of, 366 for cortical mapping, 366 recording for, 366, 367t safety of, 387–388 stimulation for, 365–366, 367t MEP monitoring, 34–42, 367 anesthesia and, 389–390, 451–468 anesthetic choice for, 468–469 in brainstem surgery, 285–286 case studies of, 85–88 clinical applications of, 376–387 complications of, 80 difficulties in, 88, 88f electrodes for, 375 history of, 75–76, 341–342 interpretation of, 82–83 in pedicle screw placement, 241–247 polarity of stimulus and, 369 practicality of, 81–82 pulse characteristics for, 371–372 safety of, 80, 387–388 selection of muscles for, 41–42 signal filtering for, 375 in spinal ischemia, 124–127, 128 stimulation for, 368–369 surgical decisions based on, 84, 87f sweep length and averaging in, 376 techniques of, 77f, 368 MEP recording, 372, 373f electrode placement for, 375 filtering of signal, 375 sweep length and averaging in, 376 MEP recording using, 41–42

482 MEP stimulation, 368–369 intensity of, 370–371 interstimulus interval for, 372 polarity of, 369 pulse duration for, 371 pulse frequency for, 371–372 site for, 369–370 train stimulation rate for, 372 MEPs (motor-evoked potentials) assessment based on, 75, 80–81, 86f characteristics of, 376 correlated to clinical outcome, 378, 378t direct nerve root stimulation of, 243–244 effects of surgical events on, 378, 379f, 380f, 381f, 382f eliciting during general anesthesia, 30f, 31–33, 79 epidural, 76–77, 373 generation of, 45–47, 49f during intramedullary surgery, 84–85 indirect nerve root stimulation of, 244–245, 247 interpretation of, 377–379, 377f muscle, 78–79, 375–376 neurogenic, 373–374 neurophysiology of, 42–42, 76–79 spinal generation of, 242–243, 373–374 transcranial generation of, 242–243 Metastatic carcinoma, embolization therapy for, 136 Methohexitol, effects on intraoperative monitoring, 463 Meyers, 407 Microdrive, 426 Microelectrode recording, 416, 416f Microelectrode stimulation, 13, 16, 434 Midas, 62 Midazolam, effects on intraoperative monitoring, 462, 463 Midbrain tumors occipital transtentorial approach to surgery in, 275 pterional transsylvan route of entry into, 275 standard infratentorial supracerebellar approach to surgery in, 274 subtemporal transtentorial approach to surgery in, 275 Miniature multielectrode, 156, 162 Mixed nerve injury, 170 Motor homunculus, 408

Index Movement disorder surgery amplification techniques in, 420–421 anatomy of, 409, 409f deep brain stimulation and, 408–409 globus pallidus procedures, 426–431 history of, 406–407 impedance change monitoring in, 414 indications for, 406 macroelectrode recording in, 414 microelectrode recording in, 414, 416f monitoring as adjunct to, 428–431 neurophysiology and, 408–409 operating room environment in, 417–418, 425 recording electrodes for, 418–420 recording techniques for, 413–416 semi–microelectrode technique in, 414–415, 415f stereotactic technique for, 424–438 stimulation techniques in, 422–423 STN procedures, 434–437 targeting techniques in, 412–413 VIM procedures, 431–434 MRI and brainstem gliomas, 268–269, 269t and brainstem hematomas, 328, 328f and central sulcus localization, 356 and intramedullary tumors, 58, 59f for intraoperative mapping, 340, 362, 412 Multiphase TES, 29–30 Multiple sclerosis (MS), 101 Multipulse technique, 46, 78, 458 Multiunit, 414 Muscle artifact, 38, 39f, 466 Muscle MEPs, 78–79 interpretation of, 83 optimizing signal of, 375 recording of, 375 Muscle relaxants effects on intraoperative monitoring, 467–468 site of action of, 464 Myelotomy, longitudinal, 97 Myogenic activity, 233, 241, 242f Myorelaxation, 79

N NBCA (n-butyl cyanoacrylate), as embolic agent, 136, 137, 140 Nd:YAG lasers, 61

Index Nerve injury classification of, 173–174 Nerve regeneration, 174–175 Nerve root stimulation, 221, 241–244 LSUMC grading of, 186t–187t nature of, 170–171 partial, 170 regeneration after, 174–175 from trauma, 189 Neurapraxia, 297 Neurinoma, vascularization of, 135 Neuroablation (lesioning), 407–408, 431 Neurogenic bladder, MDT to treat, 100, 101 Neuromas, 172f nature of, 170–171 Neuromuscular blockade assessment of, 466–467 importance of, 469 Neuronal neoplasms, incidence of, 57f Neuropathic pain, 406 Neurophysiological mapping, 353, 362, 394 Neurophysiological monitoring, 66–67, 136–137 Neurotmesis, 297 Neurotomy, to treat spasticity, 102 Nitrous oxide effects of, 459 effects on intraoperative monitoring, 459–460, 460f Noise effects in the operating room, 417, 425f effects on the CNAP, 178 effects on EMG machine, 191–192 effects on intraoperative recording, 74, 301 signal-to-noise-ratio, 346, 374–376

O Oligodendroglioma, incidence of, 57f Opioids, effects on intraoperative monitoring, 461 Optic nerve, monitoring of, 308 Optic tract, and pallidotomy, 430

P Pain decision making regarding, 114 spinal cord surgery to remedy, 104–114 Pallidotomy complications of, 430 postventral, 426–427

483 Pallidum, 407 Paralysis, postsurgical, 68 Paraplegia MDT to treat, 97 surgically induced transient, 48 Paraspinal arteriovenous fistula (AVF), 137 Paraspinal arteriovenous malformation (AVM), 137 Paravertebral veins, 134 Parkinson’s disease, 406 etiology of, 411 globus pallidus procedures to treat, 426–427 and Pars compacta (SNc), 410 and Pars reticulata (SNr), 410, 415f STN procedures to treat, 434–437 studies on, 410 VIM procedures to treat, 431 Partial nerve injury, 170 Partial posterior rhizotomy, 95 Pauser cell, 428f, 429 Pediatric Evaluation Disability Index (PEDI), 227 Pedicle screw placement anesthesia in, 238–239, 248–249 assessment of nerve root function for, 233–234 CMAP recording in, 240–241 complications of, 259–260 current shunting and, 252, 253f direct nerve root stimulation for, 243–244 DSEPs to aid, 236–237, 237f, 255–259 false negative findings of complications of, 249–253 history of, 232 impedance testing of, 254–255 indirect nerve root stimulation for, 244–245, 245f, 247 MEP monitoring in, 241–248 motor path assessment techniques for, 239–248 muscle relaxation and assessment of, 250–252 physiologic factors and assessment of, 252–253 proper placement of, 232–233, 246f, 247f SEPs to aid, 234–235, 236f transcranial and spinal stimulation for, 242–243

484 Penfield, Wilder, 26 Penfield’s technique advantages and disadvantages of, 365 adverse effects of, 388 anesthesia and, 389–390 applications of, 364–365 cortical mapping by, 364 recording in, 364, 367t stimulation in, 363, 367t subcortical stimulation by, 364–365 Perimedullary veins, 134 Peripheral nervous system electrodes for recording and stimulating impulses in, 178–181 injuries to, 170–174 nerve regeneration in, 174–175 recording of action potentials in, 175–178, 175f, 182–183 Peroneal nerve entrapment, 189 Phase reversal, 154, 159f, 162, 353 Physiologic target, 439 Plasmacytoma, embolization therapy for, 136 Polarity, 5, 7, 9, 35, 155, 367t Polyvinyl alcohol, as embolic agent, 136 Pons, 268–270, 276f Pontine tumors combined petrosal approach to, 276, 276f, 277f–278f cranial nerve motor nucleus displacement by, 323, 324f, 325, 325f retrosigmoid approach to, 273 suboccipital craniotomy and trans-fourthventricle route to, 278–279 Postcentral, 356, 362f, 366 Posterior communicating artery (PCOM), 348, 351, 386f Posterior rhizotomy functional, 95, 97f partial, 95 selective, 94 results of, 95, 225–228 sectorial, 94–95 techniques of, 223–225 Posterior spinal artery (PSA), 121–122, 130f, 132t, 133–137, 144f Precentral, 369–370, 377f, 453 Preoperative mapping, transcranial stimulation for, 4 Programmable pulse generator, 408

Index Propofol, 462 effects on intraoperative monitoring, 464–465, 465f, 466f Prostatectomy, adverse effects of, 201–202 Provocative tests, 123f, 130–132, 132t PT neurons, membrane potential of, 10f Pudendal SEP, 204f, 214 Pudendal nerve cerebral SEPs of, 206, 210, 215 dorsal root action potentials (DRAPs) of, 203, 207–208, 209f, 210, 215 spinal SEPs of, 203, 208, 210, 215

R Radicular arteries, 134 Radicular veins, 134 Radiculomedullary arteries, 134 Radiculopial arteries, 134 Radiofrequency (RF) thermocoagulation, in DREZ region, 111–112 Radiotherapy, for intramedullary neoplasms, 69 Recording electrode, 7, 27, 32, 81, 158f, 181f, 191, 314, 322, 435, 466 Remifentanil, effects on intraoperative monitoring, 461 Rhizotomy functional posterior, 95, 96f history of, 220–223 indications for, 219 partial posterior, 95 posterior selective, 94 results of, 95, 225–228 sectorial posterior, 94–95 techniques of, 223–225 Rootlets, 94–99, 106, 206 Roots, 94–99, 106, 206

S Sacral monitoring, 203 Sacral nervous system of anorectal region, 200–201 diagnostic tests of, 202–203 functional anatomy of, 199–202 functions of, 198 of lower urinary tract, 200 neurophysiology of, 224f recording techniques for, 206–207

Index of sexual organs, 201–202 stimulation techniques, 203–206 types of problems of, 198–199 Sarcoma, embolization therapy for, 136 Sectorial posterior rhizotomy, 94–95 Sedative-hypnotic drugs, effects on intraoperative monitoring, 462 Segmental potentials, generation of, 155 Selective functional posterior rhizotomy (SDR), 95, 96f, 222f history of, 220–223 indications for, 219 monitoring in, 220–221, 221f responses to, 223 results of, 225–228 techniques of, 223–225 variations in, 222–223 Semi-microelectrode recording, 414–415, 415f Sensory cranial nerve monitoring of auditory nerve, 292, 301–308 of optic nerve, 308 Sensory homunculus, 408 SEP monitoring clinical application of, 347–353 to facilitate pedicle screw placement, 234–235, 236f indications for, 342 interpretation of, 348–350 intraoperative, 27, 156, 352f, 353f with miniature multielectrodes, 156–157, 158f, 159f, 160 1f, 162, 162f, 163f recording for, 344–347 safety of, 387 stimulation for, 343–344 technique of, 342–353 SEP phase reversal indications for, 354, 355f technique for, 354, 356 SEP recording electrode placement for, 344–345 filtering in, 346 signal in, 346–347 sweep length and averaging in, 346 SEP stimulation duration of, 343 electrode placement for, 343 frequency of, 344

485 intensity of, 343 parameters for, 344t SEPs (somatosensory evoked potentials, SSEPs), 27, 75 dermatomal, 236–237, 255–259 effect of blood pressure on, 348f effects of MDT on, 109–110, 110f generators of, 155 and peripheral nervous system, 185 in spinal cord ischemia, 125–126, 127 Sevoflurane, effects on intraoperative monitoring, 455 Sexual organs, innervation of, 201–202 Signal averaging, 302 Single unit, 415f, 416f, 419 Single-pulse TES, 29 D wave recording in, 34–41 Spasticity decision tree for treatment for, 102 intrathecal baclofen to treat, 101 longitudinal myelotomy to treat, 97 MDT to treat, 97, 100, 102 neurotomy to treat, 102 sacral rhizotomy for, 207, 212–214 SDR rhizotomy to treat, 94–95, 96f, 219–228 Speech localization, 357 Sphincter muscles, EMG of, 203, 207, 210–211, 216 Spiegel, 407–408 Spinal accessory nerve (CN XI), monitoring of, 300 Spinal arteriovenous fistula, 137 Spinal arteriovenous malformation, 137 Spinal cord angiographic vascular anatomy of, 133–134 arterial systems of, 120–121, 123f direction of blood flow in, 121 dorsal columns, mapping of, 153–164 ischemia of, 124–133, 142f–143f neurophysiological monitoring of, 124–133 vascular anatomy of, 120–123, 122f vascular malformations of, 137–147 vascular tumors of, 135–136, 138f–139f venous systems of, 122–123 Spinal cord AVFs, 140 Spinal cord AVMs, 140, 142f–146f

486 Spinal cord lesioning DREZ surgery, 97–101, 104–114 indications for, in adults, 101–102 indications for, in children, 102–104 longitudinal myelotomy, 97 posterior rhizotomies, 94–97 to treat pain, 104–114 to treat spasticity, 94–104 Spinal cord stimulation, 101, 125, 241, 459 Spinal cord surgery case studies of, 85–88 complications of, 69–70, 74 history of, 56 indications for, 57–58 instruments for, 60–61 lasers in, 61 MEP behavior during, 84–85 outcomes after, 68–69 techniques of, 62–68 Spinal deformities, pedicle screws to correct, 232–233 Spinal dural arteriovenous fistula (SDAVF), 137 Spinal-cord-to-peripheral-nerve recording, 28–29 Spinal-cord-to-spinal-cord recording, 27–28 Spinal roots, 199, 203, 206, 210 Stereotactic, 406–407, 412–413 Stereotactic atlas, 412, 426–427, 437 Stereotactic frame, 407, 413, 424f Stereotactic technique, 406 Stimulating electrodes, 7, 39, 40f, 129, 176 Stimulus artifact, 192 Striae medullares, 270, 320, 328 Striatum, 407, 410–411, 427 Strumpell-Lorrain syndrome, 101 Subcortical stimulation, using Penfield’s technique, 364–365 Substantia gelatinosa, thermocoagulation of, 111–112 Subthalamic nucleus (STN) anatomy of, 435f electrical behavior of, 434–435, 436f and Parkinson’s disease, 411, 434 test stimulation of, 436–437 Subthalamic nucleus (STN), 409, 411, 425t, 434–437 Subthalamotomy, 409 Sufentanil, effects on intraoperative monitoring, 461

Index Sugita, 61 Sunderland classification, 173–174 Supportive system, 42, 45 Supratentorial surgery functional mapping in, 353–367 MEP monitoring in, 367–388 SEP monitoring in, 350–351 Surgical instruments, 74, 179, 181 Surgically induced transient paraplegia, 48 Sylvian fissure, 275, 353, 356

T Target localization, 406, 412 Telangiectasias, 140 TES (transcranial electrical stimulation) compared to TMS, 13–14, 13f in general anesthesia, 31–34 multiphase stimulation, 29–30, 41–47 muscle responses to, 18–20, 19f, 20f new methods of, 29–31 single-pulse stimulation, 29 uses of, 4 Thalamotomy, 409, 431, 473 Thalamus cell electrophysiology of, 434 ventral intermediate nucleus of, 431–434 Thermocoagulation, in DREZ region, 111–112 Thiopental, effects on intraoperative monitoring, 462 Thoracic outlet syndrome, plexus nerve lesions due to, 190 Tibialis anterior muscle (TA), 30f, 42, 44f, 78, 123f, 242f, 249, 251f, 375 TMS (transcranial magnetic stimulation) compared to TES, 13–14, 13f intensity-related effects of, 7, 8f uses of, 4–5 Total intravenous anesthesia (TIVA), 239, 389, 457, 462 Train stimulation, 80, 372, 389 Transient paraplegia, 41, 46, 47f Tremor cell, 428f, 429, 432f Trigeminal nerve (CN V) mapping of, 314–315 monitoring of motor portion of, 298 stimulation of, 295–297 Trigeminal neuralgia, 314 Trochlear cranial nerve, 272

487

Index

U Ulnar nerve entrapment, 189–190 Ultrasonic aspirator, 60–61 Ultrasonic DREZ sulcomyelotomy, 113 Urinary tract, 198, 200

V Vagal nerve (CN X), monitoring of, 300 Vagus nerve, 300 Vascular malformations dural/extradural lesions, 137–140 intradural lesions, 140–147 SEP monitoring in surgery for, 351–352 Vascular tumors classification of, 135 Ventralis lateralis (VL), 13, 411, 432 Ventralis oralis anterior (VOA), 432 Ventralis oralis posterior (VOP), 432 Ventriculography, 407, 412–413 Ventrocaudal Nucleus (VC), 433 Vertebral artery, 134

VIM (ventral intermediate nucleus of the thalamus) and Parkinsonism, 431–432 recording of, 432–433, 432f

W Wada test, 130 Wallerian degeneration, 173 regeneration after, 174 WeeFIM (Functional Inventory Measure for Children), 227–228 Wycis, 407

X Xylocaine test, for spinal cord ischemia, 130, 131, 132

Z Zona incerta (ZI), 433–435, 436f