Manual Therapy Journal- Volume 13, Issue 6, Pages 475-564 (December 2008)

VOLUME 13 NUMBER 6 PAGES 475– 564 December 2008

Editors

International Advisory Board

Ann Moore PhD, GradDipPhys, FCSP, CertEd, FMACP Clinical Research Centre for Health Professions University of Brighton Aldro Building, 49 Darley Road Eastbourne BN20 7UR, UK

K. Bennell (Victoria, Australia) K. Burton (Hudders¢eld, UK) B. Carstensen (Frederiksberg, Denmark) M. Coppieters (Queensland, Australia) E. Cruz (Setubal Portugal) L. Danneels (Mar|¤ akerke, Belgium) S. Durrell (London, UK) S. Edmondston (Perth, Australia) J. Endresen (Flaktvei, Norway) L. Exelby (Biggleswade, UK) D. Falla (Aalborg, Denmark) J. Greening (London, UK) C. J. Groen (Utrecht,The Netherlands) A. Gross (Hamilton, Canada) T. Hall (West Leederville, Australia) W. Hing (Auckland, New Zealand) M. Jones (Adelaide, Australia) S. King (Glamorgan, UK) B.W. Koes (Amsterdam,The Netherlands) J. Langendoen (Kempten, Germany) D. Lawrence (Davenport, IA, USA) D. Lee (Delta, Canada) R. Lee (London, UK) C. Liebenson (Los Angeles, CA, USA) L. Ma¡ey-Ward (Calgary, Canada) E. Maheu (Quebec, Canada) C. McCarthy (Coventry, UK) J. McConnell (Northbridge, Australia) S. Mercer (Queensland, Australia) D. Newham (London, UK) J. Ng (Hung Hom, Hong Kong) S. O’Leary (Queensland, Australia) L. Ombregt (Kanegem-Tielt, Belgium) N. Osbourne (Bournemouth, UK) M. Paatelma (Jyvaskyla, Finland) N. Petty (Eastbourne, UK) A. Pool-Goudzwaard (The Netherlands) M. Pope (Aberdeen, UK) G. Rankin (London, UK) D. Reid (Auckland, New Zealand) A. Rushton (Birmingham, UK) C. Shacklady (Manchester, UK) M. Shacklock (Adelaide, Australia) D. Shirley (Lidcombe, Australia) V. Smedmark (Stenhamra, Sweden) W. Smeets (Tongeren, Belgium) C. Snijders (Rotterdam,The Netherlands) R. Soames (Dundee, UK) P. Spencer (Barnstaple, UK) M. Sterling (St Lucia, Australia) P. Tehan (Victoria, Australia) M. Testa (Alassio, Italy) M. Uys (Tygerberg, South Africa) P. van der Wur¡ (Doorn,The Netherlands) P. van Roy (Brussels, Belgium) B.Vicenzino (St Lucia, Australia) H.J.M.Von Piekartz (Wierden,The Netherlands) M.Wallin (Spanga, Sweden) M.Wessely(Paris, France) A.Wright (Perth, Australia) M. Zusman (Mount Lawley, Australia)

Gwendolen Jull PhD, MPhty, Grad Dip ManTher, FACP Department of Physiotherapy University of Queensland Brisbane QLD 4072, Australia Associate Editor’s Darren A. Rivett PhD, MAppSc, (ManipPhty) GradDipManTher, BAppSc (Phty) Discipline of Physiotherapy Faculty of Health The University of Newcastle Callaghan, NSW 2308, Australia E-mail: [email protected] Tim McClune D.O. Spinal Research Unit. University of Hudders¢eld 30 Queen Street Hudders¢eld HD12SP, UK E-mail: [email protected] Editorial Committee Timothy W Flynn PhD, PT, OCS, FAAOMPT RHSHP- Department of Physical Therapy Regis University Denver, CO 80221-1099 USA Email: t£[email protected] Deborah Falla PhD, BPhty(Hons) Department of Health Science and Technology Aalborg University Fredrik BajersVej 7, D-3 DK-9220 Aalborg Denmark Email:deborahfvhst.aau.dk Masterclass Editor Karen Beeton PhD, MPhty, BSc(Hons), MCSP MACP ex o⁄cio member Associate Head of School (Professional Development) School of Health and Emergency Professions University of Hertfordshire College Lane Hat¢eld AL10 9AB, UK E-mail: [email protected] Case reports & Professional Issues Editor Je¡rey D. Boyling MSc, BPhty, GradDipAdvManTher, MCSP, MErgS Je¡rey Boyling Associates Broadway Chambers Hammersmith Broadway London W6 7AF, UK E-mail:je¡[email protected] Book Review Editor Raymond Swinkels MSc, PT, MT Ulenpas 80 5655 JD Eindoven The Netherlands E-mail: [email protected]

Visit the journal website at http://www.elsevier.com/math doi:10.1016/S1356-689X(08)00147-1

ARTICLE IN PRESS

Manual Therapy 13 (2008) 475–477 www.elsevier.com/math

Editorial

Pain, brain imaging and physiotherapy—Opportunity is knocking What is brain imaging? Brain imaging refers to a range of technologies that describe the structural and functional anatomy of the brain. There are several different technologies, each with advantages and limitations. This editorial will focus on magnetic resonance imaging (MRI). MRI uses a very strong magnetic field to align our body’s molecules and then uses a second field to invert (or ‘flip’) some of them. Because each type of molecule has a unique rate at which it flips back (or ‘relaxes’), it is possible to ‘tune’ a radiofrequency receiver within the MRI machine, to detect a particular molecule. The molecule to which the MRI is ‘tuned’ determines, after sophisticated data processing, the resultant image. Tuning it to water provides a structural image, because water content varies between tissues. Tuning it to oxygenated haemoglobin provides a functional image (fMRI) because neuronal activity causes an increase in oxygenated haemoglobin. fMRI can provide very precise information about where neuronal activity occurs (spatial resolution), but not about when it occurs (temporal resolution). Electroencephalography (EEG), for example, better tells us when, but not where. Rapid technological progress means that fMRI can now delineate areas (called ‘voxels’) smaller than 1 mm3. In fact, the spatial resolution of fMRI is limited not by technological limitations, but by the brain’s circulatory network. Bearing in mind, however, that 1 mm3 of the brain contains about 25,000 neurons and 200 trillion synapses, fMRI can only give us information about a very large number of neurons. In this sense, it is a blunt tool.

What has brain imaging told us about brain activity during experimentally induced pain? Despite its limitations, brain imaging has told us a great deal about where brain activity changes when someone is in pain. Most experiments involve healthy volunteers, who receive noxious stimuli and their brain 1356-689X/$ - see front matter r 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.10.001

activity is compared between the period directly after each stimulus and that during a control, or rest period. These studies, and there are many of them, yield a reasonably consistent result—that the dorsolateral prefrontal, insula, anterior cingulate, primary and second sensory cortices, and the thalamus, constitute a kind of human ‘pain matrix’ (Ingvar, 1999; Apkarian et al., 2005). Unfortunately, however, it is not quite that simple. First, whether the participants in these studies are in fact normal people, even though their very participation in a study that inflicts sometimes severe pain, repeatedly, might suggest otherwise, is seldom questioned. Nor is the external validity of these studies: How realistic are the stimuli and the context in which they are delivered? That is, how many patients attend for physiotherapy having sustained multiple stimuli of a known temperature and duration, delivered under the assurance that the stimuli are not, in fact, dangerous, and with the approval of a multidisciplinary ethical review board? Second, the results are variable—seldom do we get the pain matrix, the whole pain matrix and nothing but the pain matrix. So what is the truth? Our current understanding is that the neural network activated during pain is unique to the individual, although there are some areas far more likely to be activated than others. However, imaging the brain during pain is not simply neophrenology—exciting material emerges when scientists start comparing within and between people. Within-subject comparisons of brain activity associated with pain By manipulating the context of a noxious stimulus, scientists have uncovered neural correlates of, for example, attention, anxiety, expectation, empathy, placebo analgesia, mood, body ownership, somatic monitoring, and drug-induced modulation (see Tracey and Mantyh (2007) for review). The findings are consistent with the neuromatrix theory of pain: experimentally manipulating activity of the neural network representing, for example, anxiety, can modulate the neural network representing, for example, pain, via

ARTICLE IN PRESS 476

Editorial / Manual Therapy 13 (2008) 475–477

synaptic connections between the two (see Moseley (2007) for review).

interpretations for established effects, but also new effects, which in turn will provide fresh targets for physiotherapy-induced analgesia.

Between-subject comparisons of brain activity Cortical organisation As yet, no-one has tracked the changes that occur in the brain as pain persists. However, the brain of someone with persistent pain has been compared to the brain of someone with acute pain, and to that of someone who is pain free and (supposedly) normal. Three intriguing results, each with profound implications, consistently emerge from these comparative studies. First, persistent pain is associated with upregulation of the pain matrix and downregulation of endogenous antinoceptive mechanisms (Apkarian et al., 2005). This means that the relationship between pain and input from the tissues becomes less predictable. Second, the representation of body parts and movements, held within the primary sensory and motor cortices, becomes reorganised. This means that perceptual and motor mechanisms that are based on these maps, can become disrupted (see Lotze and Moseley (2007) for review). Third, the density of gray matter in several parts of the pain matrix, for example, the thalamus and dorsolateral prefrontal cortex, may decrease (Apkarian et al., 2004), which potentially means a reduction in brain performance. These changes may provide new targets for intervention, targets that physiotherapists are well resourced to pursue.

Physiotherapy, brain imaging and pain Physiotherapy-induced analgesia Many physiotherapists make a living out of decreasing pain. Aside from the staple strategies, like movement, joint mobilisation and TENS, we now also know that motor imagery and mirror therapy (Brodie et al., 2007; see also Moseley et al. (2008a) for a review of mirror therapy for analgesia and also Moseley et al. (2008b) and Gustin et al. (in press) for examples of motor imagery increasing pain), and carefully explaining the biology that underpins someone’s pain state, can decrease pain (Moseley, 2002, 2004; Moseley et al., 2004). What we do not know, however, is how the brain contributes to these effects. Collaborative studies, between brain imagers, pain scientists and physiotherapists would stand to bridge this substantial gap in our knowledge. Such studies could, for example, elucidate the relative contribution of increased inhibition and decreased facilitation, of opiate and non-opiate systems, of higher-order cognitive processes. We could tease out effects on sensory-discriminative and affective-motivational mechanisms. I contend that these are indeed exciting prospects—we stand to discover not only new

Persistent pain and cortical reorganisation are related, but we do not know whether one causes the other (see Flor et al. (2006) and Moseley (2006) for reviews). That said, there is emerging evidence that treatments that aim to normalise cortical organisation, also reduce pain and disability in people with chronic pain. For example, sensory discrimination training for phantom limb pain (Flor et al., 2001) and tactile discrimination training for complex regional pain syndrome (Moseley and Wiech, 2008), both target body maps in primary sensory cortex and both show a clear relationship between increased tactile acuity, which is a marker of primary sensory cortex organisation, and decreased pain. At present, the effects of physiotherapy interventions on body and motor maps are unknown. Clearly, however, opportunity is knocking. Again, appropriate collaborations might make us reinterpret established effects—perhaps body maps hold the key to why some patients respond better to a comprehensive physical examination than they do to the subsequent treatment. Is it possible that the physical examination, with its exhaustive and often repetitive provocation of specific joints, with specific mobilisations, which require the patient to carefully attend to and discriminate the location, quality and intensity of the percept works via similar mechanisms to discrimination training? Is it possible that learning precise and sometimes subtle motor skills, which require the patient to attend carefully to specific body parts and to discriminate the contraction of one muscle from the contraction of its immediate neighbour, has a similar effect? Is it possible that exploiting the brain’s predilection for congruent multisensory input, via manual and tactile or visual feedback, could enhance this effect? Such suggestions are speculative, but by no means are they outrageous.

Conclusion The wealth of data already uncovered concerning pain and the human brain raises questions directly relevant to physiotherapy practice. The challenge will be to establish strategic collaborations between pain scientists, brain imagers and physiotherapists. Clinicians must talk to scientists, and importantly, scientists must listen. Indeed, all stakeholders must be sufficiently open-minded to consider new explanations for old effects, and sufficiently alert to pursue new and better treatments.

ARTICLE IN PRESS Editorial / Manual Therapy 13 (2008) 475–477

References Apkarian AV, Sosa Y, Sonty S, Levy RM, Harden RN, Parrish TB, et al. Chronic back pain is associated with decreased prefrontal and thalamic gray matter density. Journal of Neuroscience 2004;24(46): 10410–5. Apkarian AV, Bushnell MC, Treede RD, Zubieta JK. Human brain mechanisms of pain perception and regulation in health and disease. European Journal of Pain 2005;9(4):463–84. Brodie EE, Whyte A, Niven CA. Analgesia through the looking-glass? A randomized controlled trial investigating the effect of viewing a ‘virtual’ limb upon phantom limb pain, sensation and movement. European Journal of Pain 2007;11(4):428–36. Flor H, Denke C, Schaefer M, Grusser S. Effect of sensory discrimination training on cortical reorganisation and phantom limb pain. Lancet 2001;357(9270):1763–4. Flor H, Nikolajsen L, Staehelin Jensen T. Phantom limb pain: a case of maladaptive cns plasticity?. Nature Reviews Neuroscience 2006;7(11):873–81. Gustin SM, Wrigley PJ, Gandevia SC, Middleton JW, Henderson LA, Siddall PJ. Movement imagery increases pain in people with neuropathic pain following complete thoracic spinal cord injury. Pain, in press. Ingvar M. Pain and functional imaging. Philosophical Transactions of the Royal Society of London B: Biological Sciences 1999; 354(1387):1347–58. Lotze M, Moseley GL. Role of distorted body image in pain. Current Rheumatological Reports 2007;9(6):488–96. Moseley GL. Combined physiotherapy and education is effective for chronic low back pain. A randomised controlled trial. Australian Journal of Physiotherapy 2002;48:297–302. Moseley GL. Evidence for a direct relationship between cognitive and physical change during an education intervention in people with chronic low back pain. European Journal of Pain 2004;8(1): 39–45.

477

Moseley GL. Making sense of s1 mania—are things really that simple?. In: Gifford L editor. Topical issues in pain, vol. 5. Falmouth: CNS Press; 2006. p. 321–40. Moseley GL. Reconceptualising pain according to its underlying biology. Physical Therapy Reviews 2007;12:169–78. Moseley GL, Wiech K. Tactile discrimination, but not tactile stimulation alone, reduces chronic limb pain. Pain 2008, in press; Online 3 December 2007. Moseley GL, Nicholas MK, Hodges PW. A randomized controlled trial of intensive neurophysiology education in chronic low back pain. Clinical Journal of Pain 2004;20(5):324–30. Moseley GL, Gallace A, Spence C. Is mirror therapy all it is cracked up to be? Current evidence and future directions. Pain 2008a, in press. Moseley GL, Zalucki N, Birklein F, Marinus J, Hilten JJv, Luomajoki H. Thinking about movement hurts: the effect of motor imagery on pain and swelling in people with chronic arm pain. Arthritis Care & Research 2008b;59(5):623–31. Tracey I, Mantyh PW. The cerebral signature and its modulation for pain perception. Neuron 2007;55(3):377–91.

G. Lorimer Moseley Department of Physiology, Anatomy & Genetics, University of Oxford, Le Gros Clark Building, South Parks Road, Oxford OX1 3QX, UK Oxford Centre for fMRI of the Brain, University of Oxford, South Parks Road, Oxford OX1 3QX, UK Prince of Wales Medical Research Institute Cnr Easy & Barker Streets Randwick 2031 Australia E-mail address: [email protected]

ARTICLE IN PRESS

Manual Therapy 13 (2008) 478–483 www.elsevier.com/math

Review

Lack of uniformity in diagnostic labeling of shoulder pain: Time for a different approach Jasper Mattijs Schellingerhout, Arianne Petra Verhagen, Siep Thomas, Bart Willem Koes Department of General Practice, Erasmus Medical Centre, Rotterdam, PO Box 2040, 3000 CA Rotterdam, The Netherlands Received 12 June 2007; received in revised form 18 March 2008; accepted 14 April 2008

Abstract Diagnostic labels for shoulder pain (e.g., frozen shoulder, impingement syndrome) are widely used in international research and clinical practice. However, about 10 years ago it was shown that the criteria to define those labels were not uniform. Since an ongoing lack of uniformity seriously hampers communication and does not serve patients, we decided to evaluate the uniformity in definitions. Therefore, we compared the selection criteria of different randomised controlled trials (RCTs). This comparison revealed some corresponding criteria, but no uniform definition could be derived for any of the diagnostic labels. Besides the lack of uniformity, the currently used labels have only a fair to moderate interobserver reproducibility and in systematic reviews none of the separate trials using a diagnostic label show a large benefit of treatment. This, altogether, seems sufficient reason to reconsider their use. Therefore, we strongly suggest to abolish the use of these labels and direct future research towards undivided populations with ‘‘general’’ shoulder pain. Possible subgroups with a better prognosis and/or treatment result, based on common characteristics that are easily and validly reproducible, can then be identified within these populations. r 2008 Elsevier Ltd. All rights reserved. Keywords: Shoulder pain; Shoulder impingement syndrome; Rotator cuff; Diagnosis

1. Introduction In the past many subgroups have been suggested in people with shoulder pain with enhancement of treatment success as one of the aims (e.g., frozen shoulder, rotator cuff tendinitis, impingement syndrome). About 10 years ago, however, it was shown that the specific criteria for each of those subgroups were not uniformly defined (Green et al., 1998). In order to systematically evaluate the efficacy and effectiveness of therapeutic interventions for shoulder pain, it is necessary to compare the results of different studies. However, if the lack of unambiguous definitions still exists today, this, would seriously hamper interstudy comparison. Corresponding author. Tel.: +31 10 404 3550; fax: +31 10 404 4766. E-mail address: [email protected] (J.M. Schellingerhout).

1356-689X/$ - see front matter r 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.04.005

Diagnostic labels are still frequently used in intervention research on shoulder pain. Therefore, this review aims to assess the uniformity of criteria used in intervention research to define diagnostic labels for subgroups of patients with shoulder pain.

2. Methods 2.1. Selection criteria Since one of the main goals of the diagnostic labels is to enhance treatment success, we focused on the main tool of intervention research: i.e., randomised controlled trials (RCTs). An RCT was included in the present review only when it concerned an intervention for shoulder pain with a specific diagnostic label. There were no restrictions on the kind of intervention or the population being studied, based on the assumption that

ARTICLE IN PRESS J.M. Schellingerhout et al. / Manual Therapy 13 (2008) 478–483

sociodemographic factors do not affect the definitions of the different diagnostic labels. Reasons for exclusion were complaints due to an external identifiable cause (e.g., trauma, surgery) or a known underlying disorder (e.g., rheumatologic disorder, neurological disorder, fracture, luxation, malignancy). Search and selection were done by one of the authors (JMS). 2.2. Search strategy Medline was searched with the medical subject headings ‘‘shoulder’’, ‘‘shoulder pain’’, ‘‘shoulder impingement syndrome’’, ‘‘rotator cuff’’ and ‘‘bursitis’’ in combination with a search strategy for RCTs and systematic reviews (Shojania and Bero, 2001; Robinson and Dickersin, 2002). The Cochrane central register of controlled trials (CENTRAL) and the Cochrane database of systematic reviews were searched using the following terms: ‘‘shoulder’’, ‘‘frozen shoulder’’, ‘‘calcifying tendinitis’’, ‘‘rotator cuff’’ and ‘‘glenoid’’. The terms were restricted to ‘‘title, abstract and keywords’’. We searched the literature for studies published from January 1990 through December 2006. Languages were restricted to English, French, German and Dutch. The only reason to search for systematic reviews was to screen their reference lists. A reference check was also performed in all the RCTs. 2.3. Evaluation of uniformity The criteria for patient selection were extracted from each trial. The separate criteria of each trial were compared with those using an equivalent diagnostic label. We aimed to identify either corresponding or contradictory diagnostic tests and features of the shoulder. Items were considered to correspond if they described the same test or feature of the shoulder (e.g., Neer’s impingement sign, or restriction of movement). They were considered to be contradictory if the item was a reason for inclusion in one article and a reason for exclusion in another (e.g., a positive test).

3. Results Our search strategy resulted in the following hits per database: Cochrane CENTRAL 1401 articles, Cochrane database of systematic reviews 40 articles, and Medline 2603 articles. Of all these articles, 66 met our criteria. A total of 13 different diagnostic labels were found (Table 1), which we combined into five main groups based on the similarities of names and the way the names are used interchangeably.

479

Table 1 Diagnostic labels used in shoulder research and the number of RCTs using the same label Label

Number of articlesa

Adhesive capsulitis Frozen shoulder Painful stiff shoulder Rotator cuff tear Shoulder tendinitis (Subacromial) bursitis Rotator cuff tendinitis Rotator cuff tendinosis Calcific tendinitis Calcifying tendinitis Tendinitis calcarea Supraspinatus tendinitis (Subacromial) impingement syndrome

18 12 3 4 5 4 5 1 6 7 1 2 15

a The total number in this column exceeds the total number of included RCTs, because some articles use more than one label for the same population.

3.1. Adhesive capsulitis/frozen shoulder This group contains ‘‘adhesive capsulitis’’, ‘‘frozen shoulder’’ and ‘‘painful stiff shoulder’’. A consistent description could not be derived from the 21 RCTs using these diagnostic labels (Jacobs et al., 1991; Rizk et al., 1991; White and Tuit, 1996; Gam et al., 1998; de Jong et al., 1998; Rovetta and Monteforte, 1998; van der Windt et al., 1998; Jones and Chattopadhyay, 1999; Dahan et al., 2000; Arslan and Celiker, 2001; Kivimaki and Pohjolainen, 2001; Sun et al., 2001; Karatas and Meray, 2002; Carette et al., 2003; Buchbinder et al., 2004a, b; Guler-Uysal and Kozanoglu, 2004; Pajareya et al,. 2004; Widiastuti-Samekto and Sianturi, 2004; Ryans et al., 2005; Vermeulen et al., 2006). All these articles stated that a restricted movement of the shoulder should be present, but they were not consistent regarding the amount of restriction (number of degrees), the kind of restriction (active and/or passive), and the direction of the restriction (e.g., abduction, external rotation). Nocturnal accentuation of shoulder pain was mentioned as an inclusion criterion in 7 RCTs (Rizk et al., 1991; White and Tuit, 1996; Gam et al., 1998; de Jong et al., 1998; Jones and Chattopadhyay, 1999; Sun et al., 2001; Karatas and Meray, 2002). 3.2. (Subacromial) impingement syndrome Thirteen of the 15 RCTs using this diagnostic label describe one or more of the following tests as an inclusion criterion: Neer’s impingement sign, Kennedy– Hawkins impingement sign, and Neer’s impingement test (Hawkins and Kennedy, 1980; Neer, 1983; Brox et al., 1993; Lindh and Norlin, 1993; Blair et al., 1996; Rahme et al., 1998; Brox et al., 1999; Bang and Deyle,

ARTICLE IN PRESS 480

J.M. Schellingerhout et al. / Manual Therapy 13 (2008) 478–483

2000; Plafki et al., 2000; Spangehl et al., 2002; Husby et al., 2003; Akgun et al., 2004; Walther et al., 2004; Haahr et al., 2005; Johansson et al., 2005). However, these three criteria were used in seven different combinations, with a maximum of four RCTs using the same combination. Therefore, no uniform definition was used. Two of the four RCTs using the same combination were actually the same study at a different time of follow-up (Brox et al., 1993, 1999). One RCT used a modified Neer’s impingement sign for inclusion (Conroy and Hayes, 1998). The remaining RCT did not clearly describe the specific diagnostic tests used for inclusion (Werner et al., 2002). 3.3. Calcifying/calcific tendinitis This group contains ‘‘calcifying tendinitis’’, ‘‘calcific tendinitis’’, and ‘‘tendinitis calcarea’’. Thirteen RCTs using one of these labels were found. Nine RCTs used the radiological classification according to Ga¨rtner as a criterion (Ga¨rtner, 1993); in all cases type I and type II were used as inclusion criterion, and in six cases type III as exclusion criterion (Rompe et al., 1998; Ebenbichler et al., 1999; Loew et al., 1999; Haake et al., 2001; Rompe et al., 2001; Gerdesmeyer et al., 2003; Perlick et al., 2003; Krasny et al., 2005; Cacchio et al., 2006). However, the studies showed differences in the minimal size of the calcification (5–15 mm) and used different additional criteria (e.g., restricted range of motion). Three of the four remaining RCTs used a radiologically proven calcification as inclusion criterion, but did not classify it according to Ga¨rtner (Ga¨rtner, 1993; Perron and Malouin, 1997; Leduc et al., 2003; Pleiner et al., 2004). The remaining trial did not report any inclusion or exclusion criteria (Rubenthaler et al., 2003). The radiological classification according to Ga¨rtner (1993) may be a basis for a uniform definition of this diagnostic label, but the populations included in the currently available trials are not comparable due to variation in additional criteria. 3.4. Rotator cuff tendinitis This group contains ‘‘rotator cuff tendinitis’’, ‘‘rotator cuff tendinosis’’, ‘‘shoulder tendinitis’’, ‘‘subacromial bursitis’’, and ‘‘supraspinatus tendinitis’’. We grouped these labels based on their name and because ‘‘subacromial bursitis’’ and ‘‘shoulder tendinitis’’ are used interchangeably. We found thirteen trials that used one of the forementioned labels (Adebajo et al., 1990; Friis et al., 1992; Vecchio et al., 1993a; Zuinen, 1993; Lecomte et al., 1994; Berrazueta et al., 1996; Itzkowitch et al., 1996;

Kleinhenz et al., 1999; Schmitt et al., 2002; Speed et al., 2002; Bertin et al., 2003; Petri et al., 2004; Alvarez et al., 2005). Five of these RCTs did not specify the physical findings for establishing the diagnosis (Friis et al., 1992; Zuinen, 1993; Lecomte et al., 1994; Schmitt et al., 2002; Bertin et al., 2003). The remaining RCTs showed some small similarities. The criteria according to Cyriax were used in three RCTs (Cyriax, 1982; Adebajo et al., 1990; Vecchio et al., 1993a; Kleinhenz et al., 1999). Three other RCTs mentioned (increasing) pain with abduction of the shoulder as an inclusion criterion (Berrazueta et al., 1996; Itzkowitch et al., 1996; Petri et al., 2004). The last similarity we found was the presence of a socalled painful arc, which was used in two articles as an inclusion criterion, but the range of the painful arc was specified in a single study only (Speed et al., 2002; Petri et al., 2004). The similarities in diagnostic criteria were not related to specific labels within this group. 3.5. Rotator cuff tear The four RCTs using this label define their diagnostic tests very poorly (Vecchio et al., 1993b; Shibata et al., 2001; Gartsman and O’Connor, 2004; MacDermid et al., 2006). Two of them mention clinical examination and/or magnetic resonance imaging to determine the diagnosis, but do not give any details (Shibata et al., 2001; MacDermid et al., 2006). A common feature of the trials discussed above is that they use diagnostic labels as an exclusion criterion without defining the properties of the labels, which further complicates the comparability between studies and patient groups.

4. Discussion Despite the frequent use of diagnostic labels for shoulder pain, there still seem to be no generally applied criteria for each of those labels. Although our search was limited to RCTs, we consider it likely that the same conclusion can be drawn from studies with a different design. This assumption is supported by a study comparing diagnostic criteria for rotator cuff tendinitis and adhesive capsulitis in epidemiologic surveys (Walker-Bone et al., 2003). The lack of uniformity may be an important reason why systematic reviews evaluating the efficacy of treatments for diagnostic subgroups do not find strong evidence for any of those treatments, since heterogeneity might reduce any effect, if present. However, another reason could be that the criteria used in those subgroups are in fact not related to treatment success. The latter

ARTICLE IN PRESS J.M. Schellingerhout et al. / Manual Therapy 13 (2008) 478–483

idea arises from the fact that the separate trials in recent systematic reviews also show no large benefit of the treatments under investigation (Desmeules et al., 2003; Ejnisman et al., 2004; Harniman et al., 2004; Buchbinder et al., 2006). The question is whether the currently used diagnostic labels have any additive value in practice and research. In theory, their main purpose is to facilitate communication and to identify subgroups of patients differing from the overall population in prognosis and/or treatment benefit. At present, however, the labels unintentionally seem to result in a Babylonian confusion of tongues and seem to be of little benefit for those with shoulder pain. Lack of a gold standard for each of the diagnostic labels makes it impossible to determine the diagnostic accuracy (sensitivity, specificity) of previously mentioned diagnostic tests. However, even if strict definitions are stated in advance, the interobserver agreement in classification of currently used subgroups, and the clinical tests leading to their diagnosis (e.g., Neer’s impingement sign, Ga¨rtner’s classification), is only fair to moderate (k ¼ 0.2–0.6) (de Winter et al., 1999; Maier et al., 2003; Ostor et al., 2004). This implies that even if the diagnostic accuracy is satisfactory, and a good face validity and content validity are ensured by stating strict definitions, the usefulness of the currently used subgroups is still undermined by lack of reproducibility of the diagnostic criteria (Buchbinder et al., 1996). Together with the other considerations, this seems sufficient reason to seriously question the use of these diagnostic subgroups. Therefore we strongly suggest to reconsider the use of these diagnostic labels. In the short term we propose to create unambiguous definitions by means of global consensus, or to abolish the use of these labels. We favour the latter approach, since previous attempts to get consensus-based definitions for frozen shoulder and rotator cuff tendinitis widely accepted, apparently have failed (Zuckerman et al., 1993; Walker-Bone et al., 2003). In the long term, we suggest to direct future research towards undivided populations with ‘‘general’’ shoulder pain. Possible subgroups with a better prognosis and/or treatment result can then be identified within these populations by means of data analysis in prospective cohort studies or randomised trials. Preferably, these new subgroups will be based on common characteristics that are easily and validly reproducible, to avoid the current problems with interobserver agreement.

Acknowledgement No financial or material support for the research and the work was received.

481

References Adebajo AO, Nash P, Hazleman BL. A prospective double blind dummy placebo controlled study comparing triamcinolone hexacetonide injection with oral diclofenac 50 mg TDS in patients with rotator cuff tendinitis. Journal of Rheumatology 1990;17: 1207–10. Akgun K, Birtane M, Akarirmak U. Is local subacromial corticosteroid injection beneficial in subacromial impingement syndrome? Clinical Rheumatology 2004;23:496–500. Alvarez CM, Litchfield R, Jackowski D, Griffin S, Kirkley A. A prospective, double-blind, randomized clinical trial comparing subacromial injection of betamethasone and xylocaine to xylocaine alone in chronic rotator cuff tendinosis. The American Journal of Sports Medicine 2005;33:255–62. Arslan S, Celiker R. Comparison of the efficacy of local corticosteroid injection and physical therapy for the treatment of adhesive capsulitis. Rheumatology International 2001;21:20–3. Bang MD, Deyle GD. Comparison of supervised exercise with and without manual physical therapy for patients with shoulder impingement syndrome. Journal of Orthopaedic and Sports Physical Therapy 2000;30:126–37. Berrazueta JR, Losada A, Poveda J, Ochoteco A, Riestra A, Salas E, et al. Successful treatment of shoulder pain syndrome due to supraspinatus tendinitis with transdermal nitroglycerin. A double blind study. Pain 1996;66:63–7. Bertin P, Behier JM, Noel E, Leroux JL. Celecoxib is as efficacious as naproxen in the management of acute shoulder pain. The Journal of International Medical Research 2003;31:102–12. Blair B, Rokito AS, Cuomo F, Jarolem K, Zuckerman JD. Efficacy of injections of corticosteroids for subacromial impingement syndrome. The Journal of Bone and Joint Surgery: American Volume 1996;78:1685–9. Brox JI, Staf PH, Ljunggren AE, Brevik JI. Arthroscopic surgery compared with supervised exercises in patients with rotator cuff disease (stage II impingement syndrome). The British Medical Journal 1993;307:899–903. Brox JI, Gjengedal E, Uppheim G, Bohmer AS, Brevik JI, Ljunggren AE, et al. Arthroscopic surgery versus supervised exercises in patients with rotator cuff disease (stage II impingement syndrome): a prospective, randomized, controlled study in 125 patients with a 2 1/2-year follow-up. Journal of Shoulder and Elbow Surgery 1999;8:102–11. Buchbinder R, Goel V, Bombardier C, Hogg-Johnson S. Classification systems of soft tissue disorders of the neck and upper limb: do they satisfy methodological guidelines? Journal of Clinical Epidemio logy 1996;49:141–9. Buchbinder R, Green S, Forbes A, Hall S, Lawler G. Arthrographic joint distension with saline and steroid improves function and reduces pain in patients with painful stiff shoulder: results of a randomised, double blind, placebo controlled trial. Annals of the Rheumatic Diseases 2004a;63:302–9. Buchbinder R, Hoving JL, Green S, Hall S, Forbes A, Nash P. Short course prednisolone for adhesive capsulitis (frozen shoulder or stiff painful shoulder): a randomised, double blind, placebo controlled trial. Annals of the Rheumatic Diseases 2004b;63:1460–9. Buchbinder R, Green S, Youd JM, Johnston RV. Oral steroids for adhesive capsulitis. Cochrane Database of Systematic Reviews 2006(4):CD006189. Cacchio A, Paoloni M, Barile A, Don R, de Paulis F, Calvisi V, et al. Effectiveness of radial shock-wave therapy for calcific tendinitis of the shoulder: single-blind, randomized clinical study. Physical Therapy 2006;86:672–82. Carette S, Moffet H, Tardif J, Bessette L, Morin F, Fre´mont P, et al. Intraarticular corticosteroids, supervised physiotherapy, or a combination of the two in the treatment of adhesive capsulitis of

ARTICLE IN PRESS 482

J.M. Schellingerhout et al. / Manual Therapy 13 (2008) 478–483

the shoulder: a placebo-controlled trial. Arthritis and Rheumatism 2003;48:829–38. Conroy DE, Hayes KW. The effect of joint mobilization as a component of comprehensive treatment for primary shoulder impingement syndrome. The Journal of Orthopaedic Sports and Physical Therapy 1998;28:3–14. Cyriax J. Diagnosis of soft tissue lesions. Textbook of orthopaedic medicine, 8th ed. London: Baillie`re Tindall; 1982. Dahan TH, Fortin L, Pelletier M, Petit M, Vadeboncoeur R, Suissa S. Double blind randomized clinical trial examining the efficacy of bupivacaine suprascapular nerve blocks in frozen shoulder. The Journal of Rheumatology 2000;27:1464–9. Desmeules F, Cote CH, Fremont P. Therapeutic exercise and orthopedic manual therapy for impingement syndrome: a systematic review. Clinical Journal of Sport Medicine 2003;13:176–82. de Winter AF, Jans MP, Scholten RJ, Deville W, van Schaardenburg D, Bouter LM. Diagnostic classification of shoulder disorders: interobserver agreement and determinants of disagreement. Annals of the Rheumatic Diseases 1999;58:272–7. Ebenbichler GR, Erdogmus CB, Resch KL, Funovics MA, Kainberger F, Barisani G, et al. Ultrasound therapy for calcific tendinitis of the shoulder. New England Journal of Medicine 1999;340:1533–8. Ejnisman B, Andreoli CV, Soares BG, Fallopa F, Peccin MS, Abdalla RJ, et al. Interventions for tears of the rotator cuff in adults. Cochrane Database of Systematic Reviews 2004(1):CD002758. Friis J, Jarner D, Toft B, Christensen K, Christophersen J, Ibfeldt HH, et al. Comparison of two ibuprofen formulations in the treatment of shoulder tendonitis. Clinical Rheumatology 1992;11:105–8. Gam AN, Schydlowsky P, Rossel I, Remvig L, Jensen EM. Treatment of ‘‘frozen shoulder’’ with distension and glucorticoid compared with glucorticoid alone. A randomised controlled trial. The Scandinavian Journal of Rheumatology 1998;27:425–30. Ga¨rtner J. Tendinosis calcarea—results of treatment with needling (Tendinosis calcarea—Behandlungsergebnisse mit dem Needling). Zeitschrift fu¨r Orthopa¨die und Ihre Grenzgebiete 1993;131:461–9 [in German]. Gartsman GM, O’Connor DP. Arthroscopic rotator cuff repair with and without arthroscopic subacromial decompression: a prospective, randomized study of one-year outcomes. Journal of Shoulder and Elbow Surgery 2004;13:424–6. Gerdesmeyer L, Wagenpfeil S, Haake M, Maier M, Loew M, Wortler K, et al. Extracorporeal shock wave therapy for the treatment of chronic calcifying tendonitis of the rotator cuff: a randomized controlled trial. JAMA: The Journal of the American Medical Association 2003;290:2573–80. Green S, Buchbinder R, Glazier R, Forbes A. Systematic review of randomised controlled trials of interventions for painful shoulder: selection criteria, outcome assessment, and efficacy. The British Medical Journal 1998;316:354–60. Guler-Uysal F, Kozanoglu E. Comparison of the early response to two methods of rehabilitation in adhesive capsulitis. Swiss Medical Weekly 2004;134:353–8. Haahr JP, Ostergaard S, Dalsgaard J, Norup K, Frost P, Lausen S, et al. Exercises versus arthroscopic decompression in patients with subacromial impingement: a randomised, controlled study in 90 cases with a one year follow up. Annals of the Rheumatic Diseases 2005;64:760–4. Haake M, Deike B, Thon A, Schmitt J. Value of exact focusing of extracorporeal shock waves (ESWT) in therapy of tendinitis calcarea. A prospective randomized study (Bedeutung der exakten Fokussierung extrakorporaler Stosswellen (ESWT) bei der Therapie der Tendinitis calcarea. Eine prospektive randomisierte Studie). Biomedizinische Technik (Berl) 2001;46:69–74 [in German]. Harniman E, Carette S, Kennedy C, Beaton D. Extracorporeal shock wave therapy for calcific and noncalcific tendonitis of the rotator

cuff: a systematic review. Journal of Hand Therapy 2004;17: 132–51. Hawkins RJ, Kennedy JC. Impingement syndrome in athletes. American Journal of Sports Medicine 1980;8:151–7. Husby T, Haugstvedt JR, Brandt M, Holm I, Steen H. Open versus arthroscopic subacromial decompression: a prospective, randomized study of 34 patients followed for 8 years. Acta Orthopaedica Scandinavica 2003;74:408–14. Itzkowitch D, Ginsberg F, Leon M, Bernard V, Appelboom T. Periarticular injection of tenoxicam for painful shoulders: a doubleblind, placebo controlled trial. Clinical Rheumatology 1996;15: 604–9. Jacobs LG, Barton MA, Wallace WA, Ferrousis J, Dunn NA, Bossingham DH. Intra-articular distension and steroids in the management of capsulitis of the shoulder. The British Medical Journal 1991;302:1498–501. Johansson KM, Adolfsson LE, Foldevi MO. Effects of acupuncture versus ultrasound in patients with impingement syndrome: randomized clinical trial. Physical Therapy 2005;85:490–501. Jones DS, Chattopadhyay C. Suprascapular nerve block for the treatment of frozen shoulder in primary care: a randomized trial. British Journal of General Practice 1999;49:39–41. de Jong BA, Dahmen R, Hogeweg JA, Marti RK. Intra-articular triamcinolone acetonide injection in patients with capsulitis of the shoulder: a comparative study of two dose regimens. Clinical Rehabilitation 1998;12:211–5. Karatas GK, Meray J. Suprascapular nerve block for pain relief in adhesive capsulitis: comparison of 2 different techniques. Archives of Physical Medicine and Rehabilitation 2002;83: 593–7. Kivimaki J, Pohjolainen T. Manipulation under anesthesia for frozen shoulder with and without steroid injection. Archives of Physical Medicine and Rehabilitation 2001;82:1188–90. Kleinhenz J, Streitberger K, Windeler J, Gussbacher A, Mavridis G, Martin E. Randomised clinical trial comparing the effects of acupuncture and a newly designed placebo needle in rotator cuff tendinitis. Pain 1999;83:235–41. Krasny C, Enenkel M, Aigner N, Wlk M, Landsiedl F. Ultrasoundguided needling combined with shock-wave therapy for the treatment of calcifying tendonitis of the shoulder. Journal of Bone and Joint Surgery: British Volume 2005;87:501–7. Lecomte J, Buyse H, Taymans J, Monti T. Treatment of tendinitis and bursitis: a comparison of nimesulide and naproxen sodium in a double-blind parallel trial. European Journal of Rheumatology and Inflammation 1994;14:29–32. Leduc BE, Caya J, Tremblay S, Bureau NJ, Dumont M. Treatment of calcifying tendinitis of the shoulder by acetic acid iontophoresis: a double-blind randomized controlled trial. Archives of Physical Medicine and Rehabilitation 2003;84:1523–7. Lindh M, Norlin R. Arthroscopic subacromial decompression versus open acromioplasty. A two-year follow-up study. Clinical Orthopaedics and Related Research 1993:174–6. Loew M, Daecke W, Kusnierczak D, Rahmanzadeh M, Ewerbeck V. Shock-wave therapy is effective for chronic calcifying tendinitis of the shoulder. Journal of Bone and Joint Surgery: British Volume 1999;81:863–7. MacDermid JC, Holtby R, Razmjou H, Bryant D. All-arthroscopic versus mini-open repair of small or moderate-sized rotator cuff tears: a protocol for a randomized trial [NCT00128076]. BMC Musculoskeletal Disorders 2006;7:25. Maier M, Maier-Bosse T, Eckermann M, Maier-Bosse E, Schulz CU, Steinborn M, et al. Observer variabilities of radiological classifications of calcified deposits in calcifying tendinitis of the shoulder (Die ro¨ntgenmorphologische gestalt symptomatischer verkalkungen bei tendinosis calcarea der schulter). Unfallchirurg 2003;106: 185–9.

ARTICLE IN PRESS J.M. Schellingerhout et al. / Manual Therapy 13 (2008) 478–483 Neer CS. Impingement lesions. Clinical Orthopaedics and Related Research 1983;173:70–7. Ostor AJ, Richards CA, Prevost AT, Hazleman BL, Speed CA. Interrater reproducibility of clinical tests for rotator cuff lesions. Annals of the Rheumatic Diseases 2004;63:1288–92. Pajareya K, Chadchavalpanichaya N, Painmanakit S, Kaidwan C, Puttaruksa P, Wongsaranuchit Y. Effectiveness of physical therapy for patients with adhesive capsulitis: a randomized controlled trial. Journal of the Medical Association of Thailand 2004;87:473–80. Perlick L, Luring C, Bathis H, Perlick C, Kraft C, Diedrich O. Efficacy of extracorporal shock-wave treatment for calcific tendinitis of the shoulder: experimental and clinical results. Journal of Orthopaedic Science 2003;8:777–83. Perron M, Malouin F. Acetic acid iontophoresis and ultrasound for the treatment of calcifying tendinitis of the shoulder: a randomized control trial. Archives of Physical Medicine and Rehabilitation 1997;78:379–84. Petri M, Hufman SL, Waser G, Cui H, Snabes MC, Verburg KM. Celecoxib effectively treats patients with acute shoulder tendinitis/ bursitis. The Journal of Rheumatology 2004;31:1614–20. Plafki C, Steffen R, Willburger RE, Wittenberg RH. Local anaesthetic injection with and without corticosteroids for subacromial impingement syndrome. International Orthopaedics 2000;24:40–2. Pleiner J, Crevenna R, Langenberger H, Keilani M, Nuhr M, Kainberger F, et al. Extracorporeal shockwave treatment is effective in calcific tendonitis of the shoulder. A randomized controlled trial. Wiener Klinische Wochenschrift 2004;116:536–41. Rahme H, Solem-Bertoft E, Westerberg CE, Lundberg E, Sorensen S, Hilding S. The subacromial impingement syndrome. A study of results of treatment with special emphasis on predictive factors and pain-generating mechanisms. The Scandinavian Journal of Rehabilitation and Medicine 1998;30:253–62. Rizk TE, Pinals RS, Talaiver AS. Corticosteroid injections in adhesive capsulitis: investigation of their value and site. Archives of Physical Medicine and Rehabilitation 1991;72:20–2. Robinson KA, Dickersin K. Development of a highly sensitive search strategy for the retrieval of reports of controlled trials using Pub Med. International Journal of Epidemiology 2002;31:150–3. Rompe JD, Burger R, Hopf C, Eysel P. Shoulder function after extracorporal shock wave therapy for calcific tendinitis. Journal of Shoulder and Elbow Surgery 1998;7:505–9. Rompe JD, Zoellner J, Nafe B. Shock wave therapy versus conventional surgery in the treatment of calcifying tendinitis of the shoulder. Clinical Orthopaedics and Related Research 2001: 72–82. Rovetta G, Monteforte P. Intraarticular injection of sodium hyaluronate plus steroid versus steroid in adhesive capsulitis of the shoulder. International Journal of Tissue Reactions 1998;20: 125–30. Rubenthaler F, Ludwig J, Wiese M, Wittenberg RH. Prospective randomized surgical treatments for calcifying tendinopathy. Clinical Orthopaedics and Related Research 2003:278–84. Ryans I, Montgomery A, Galway R, Kernohan WG, McKane R. A randomized controlled trial of intra-articular triamcinolone and/or physiotherapy in shoulder capsulitis. Rheumatology (Oxford) 2005;44:529–35. Schmitt J, Tosch A, Hunerkopf M, Haake M. Extracorporeal shockwave therapy (ESWT) as therapeutic option in supraspinatus tendon syndrome? One year results of a placebo controlled study (Die extrakorporale Stosswellentherapie (ESWT) als therapeutische Option beim Supraspinatussehnensyndrom? Ein-

483

Jahres-Ergebnisse einer placebokontrollierten Studie). Orthopa¨de 2002;31:652–7 [in German]. Shibata Y, Midorikawa K, Emoto G, Naito M. Clinical evaluation of sodium hyaluronate for the treatment of patients with rotator cuff tear. Journal of Shoulder and Elbow Surgery 2001;10:209–16. Shojania KG, Bero LA. Taking advantage of the explosion of systematic reviews: an efficient MEDLINE search strategy. Effective Clinical Practice 2001;4:157–62. Spangehl MJ, Hawkins RH, McCormack RG, Loomer RL. Arthroscopic versus open acromioplasty: a prospective, randomized, blinded study. Journal of Shoulder and Elbow Surgery 2002;11: 101–7. Speed CA, Richards C, Nichols D, Burnet S, Wies JT, Humphreys H, et al. Extracorporeal shock-wave therapy for tendonitis of the rotator cuff. A double-blind, randomised, controlled trial. Journal of Bone and Joint Surgery: British Volume 2002;84:509–12. Sun KO, Chan KC, Lo SL, Fong DY. Acupuncture for frozen shoulder. Hong Kong Medical Journal 2001;7:381–91. Vecchio PC, Hazleman BL, King RH. A double-blind trial comparing subacromial methylprednisolone and lignocaine in acute rotator cuff tendinitis. British Journal of Rheumatology 1993;32:743–5. Vecchio PC, Adebajo AO, Hazleman BL. Suprascapular nerve block for persistent rotator cuff lesions. The Journal of Rheumatology 1993;20:453–5. Vermeulen HM, Rozing PM, Obermann WR, le Cessie S, Vliet Vlieland TP. Comparison of high-grade and low-grade mobilization techniques in the management of adhesive capsulitis of the shoulder: randomized controlled trial. Physical Therapy 2006;86: 355–68. Walker-Bone KE, Palmer KT, Reading I, Cooper C. Criteria for assessing pain and nonarticular soft-tissue rheumatic disorders of the neck and upper limb. Seminars in Arthritis and Rheumatism 2003;33:168–84. Walther M, Werner A, Stahlschmidt T, Woelfel R, Gohlke F. The subacromial impingement syndrome of the shoulder treated by conventional physiotherapy, self-training, and a shoulder brace: results of a prospective, randomized study. Journal of Shoulder and Elbow Surgery 2004;13:417–23. Werner A, Walther M, Ilg A, Stahlschmidt T, Gohlke F. Self-training versus conventional physiotherapy in subacromial impingement syndrome (Zentrierende Kraftigungstherapie beim einfachen subakromialen Schmerzsyndrom: Eigentraining versus Krankengymnastik). Zeitschrift fu¨r Orthopa¨die und Ihre Grenzgebiete 2002;140:375–80 [in German]. White AE, Tuit JD. The accuracy and efficacy of shoulder injections in restrictive capsulitis. Journal of Orthopaedic Rheumatology 1996;9:37–40. Widiastuti-Samekto M, Sianturi GP. Frozen shoulder syndrome: comparison of oral route corticosteroid and intra-articular corticosteroid injection. The Medical Journal of Malaysia 2004; 59:312–6. van der Windt DA, Koes BW, Deville W, Boeke AJ, de Jong BA, Bouter LM. Effectiveness of corticosteroid injections versus physiotherapy for treatment of painful stiff shoulder in primary care: randomised trial. The British Medical Journal 1998;317: 1292–6. Zuckerman JD, Cuomo F, Rokito S. Definition and classification of frozen shoulder: a consensus approach. Journal of Shoulder and Elbow Surgery 1993;3:S72. Zuinen C. Diclofenac/misoprostol vs. diclofenac/placebo in treating acute episodes of tendinitis/bursitis of the shoulder. Drugs 1993;45:S17–23.

ARTICLE IN PRESS

Manual Therapy 13 (2008) 484–491 www.elsevier.com/math

Original article

Eccentric calf muscle training compared with therapeutic ultrasound for chronic Achilles tendon pain—A pilot study Rachel Chestera,b,, Mathew L. Costac, Lee Shepstoned, Adele Cooperc, Simon T. Donellc,d a

School of Allied Health Professions, Faculty of Health, University of East Anglia, Norwich, Norfolk NR4 7TJ, UK Department of Physiotherapy, The Norfolk and Norwich University Hospital, Colney Lane, Norwich, Norfolk NR4 7UZ, UK c Institute of Orthopaedics, The Norfolk and Norwich University Hospital, Colney Lane, Norwich NR4 7UZ, UK d School of Medicine, Health Policy and Practice, Faculty of Health, University of East Anglia, Norwich, Norfolk NR4 7TJ, UK b

Received 5 September 2005; received in revised form 9 October 2006; accepted 18 May 2007

Abstract A number of studies have indicated that eccentric calf muscle training has beneficial effects in the management of Achilles tendon pain for recreational athletes. The purpose of this prospective randomised single blind pilot study was to investigate their potential effectiveness compared with therapeutic ultrasound in subjects with relatively sedentary lifestyles in an NHS hospital setting. Eleven men and five women (mean age 53721 years) with Achilles tendon pain of minimum duration 4 months were randomised to one of two treatment groups; either eccentric loading or ultrasound. Administration of ultrasound and regular supervision of exercises occurred over a period of 6 weeks, with unsupervised exercises continuing for another 6 weeks. Outcome measurements were taken prior to and after 2, 4, 6 and 12 weeks after commencing treatment. They included: pain on a visual analogue scale, functional index of the leg and lower limb, and the five question EuroQol generalised health questionnaire. The difference in mean score was calculated together with 95% confidence intervals assuming a normal distribution. There were no statistically significant differences between groups or clear trends over time. In addition there was considerable overlap between the confidence intervals. This is not unexpected given the small sample size. Both interventions proved acceptable to the patients with no adverse effects. On this basis we intend conducting a full multi-centred study. r 2007 Elsevier Ltd. All rights reserved. Keywords: Achilles; Tendinopathy; Eccentric; Physical therapy

1. Introduction Chronic Achilles tendon pain is a common and disabling condition. Until recently, there has been limited evidence for the efficacy of any non-operative treatment (McLaughlan and Handoll, 2001). However, a case series report by Alfredson et al. (1998), later supported by three randomised controlled trials (Mafi et al., 2001; Silbernagel et al., 2001; Roos et al., 2004) and one quasi-experimental study (Fahlstro¨m et al., Corresponding author. School of Allied Health Professions, University of East Anglia, Norwich, Norfolk NR4 7TJ, UK. Tel.: +44 01603 593571; fax: +44 01603 593166. E-mail address: [email protected] (R. Chester).

1356-689X/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2007.05.014

2003) have identified heavy-load eccentric calf muscle training as an effective treatment in the management of Achilles tendon pain. However, the majority of participants in these Scandinavian studies were recreational athletes. Although commonly associated with sport, a substantial number of subjects with Achilles tendon pain who are seen in a National Health Service (NHS) hospital setting lead relatively sedentary lifestyles, some of whom have a variety of additional medical conditions. Eighty-eight percent of physiotherapists in the UK (Lindsay et al., 1995), 93% in Canada (Robertson and Spurritt, 1998) and 81% in Australia (Pope et al., 1995) use ultrasound daily. There is strong supporting evidence from a number of animal and in vitro studies

ARTICLE IN PRESS R. Chester et al. / Manual Therapy 13 (2008) 484–491

485

The Local Research Ethics Committee gave full approval for this pilot study to assess the benefits of heavy eccentric loading exercises compared with therapeutic ultrasound.

from the study if they had had a history of current or previous tendon rupture or a history of lower limb musculoskeletal injury. To be included in the study, subjects had to be able to stand on the affected leg only and balance for a minimum of 10 s. Forty-six patients referred to the Orthopaedic or Physiotherapy departments at the local acute hospital, with a suggested diagnosis of Achilles tendinopathy, were approached. Subjects initially received detailed written and if interested, verbal information of the procedure and those who agreed to take part provided written informed consent. Five subjects stated they did not wish to take part in the study. Seventeen subjects did not respond to the written invitation and together with an additional three who agreed to take part, did not attend their initial Physiotherapy appointment provided, irrespective of the study. Three subjects who agreed to take part in the study were found to have alternative diagnoses on their initial assessment. Two subjects, allocated to the ultrasound group, did not attend the full course of treatment or follow up, therefore providing incomplete data. No explanation was given and therefore they are not included in the results. Sixteen subjects (11 males, 5 females) with a mean age of 53 years (range 31–76 years, SD 21 years) entered and completed the study between April 2003 and September 2004. Mean duration of symptoms was 18.5 months (range 4–36 months, SD 12 months). Table 1 outlines the demographic data for each treatment group.

2.2. Subjects

2.3. Procedure

To be included in the study subjects had to have a minimum of 3 months duration of pain arising from an area between 2 and 6 cm above the distal insertion of the Achilles tendon. Confirmation of Achilles tendinopathy during clinical examination included a palpable painful swelling in the area of pain and a negative Thompson’s (Simmonds’) test (Magee, 1997). Subjects were excluded

Subjects were randomly allocated by computer generation to one of two treatment groups over a period of six weeks; either eccentric loading exercises for the calf muscle, or ultrasound over the palpable swelling in the Achilles tendon. No additional treatments were given to the subject during treatment sessions or for the following six-week period, with the exception that those

of the positive effects of ultrasound on tendon healing (Stevenson et al., 1986; Enwemeka, 1989, 1990; Jackson et al., 1991; Ramirez et al., 1997; Saini et al., 2002; Ng et al., 2003, 2004; Demir et al., 2004). There is little support for the clinical effectiveness of ultrasound from human efficacy studies and unfortunately there are few human studies investigating the physiological effects of ultrasound in vivo (Baker et al., 2001). However, to the authors’ knowledge there are no studies investigating the beneficial effects of ultrasound on chronic Achilles tendon healing or pain resolution on human subjects. The intention is to carry out a prospective randomised trial to investigate the effectiveness of eccentric loading exercises, compared with therapeutic ultrasound in the management of Achilles tendon pain, in subjects with relatively sedentary lifestyles in an NHS hospital setting. The aim of this pilot study is to test the study protocol, assess any practical or resource implications and based on the results, perform a power study calculation to assess the number of patients required for the full study.

2. Methods 2.1. Ethics

Table 1 Characteristics of the 18 Achilles tendons in the 16 subjects with chronic Achilles tendinopathy taking part in the study Group allocation

No of subjects No of tendons Age (mean years (7 standard deviation SD) Sex Mean duration of symptoms Additional pathologies (number of presenting subjects) Recreational activity (number of presenting subjects)

Eccentric loading

Ultrasound

8 9 59 (SD 10) 4 male, 4 female 23 months (SD 13) Asthma (1), myocardial infarctions (2), general osteoarthritis (1), low back pain (2) Occasional cricket (1), cycling to work (1), occasional swimming (1), none (5)

8 9 48 (SD 12) 7 male, 1 female 14 months (SD 10) Asthma (1), general osteoarthritis (1) Occasional football referee (1), cycling to work (1), occasional swimming (1), jogging (1 above), none (4)

ARTICLE IN PRESS R. Chester et al. / Manual Therapy 13 (2008) 484–491

486

in the eccentric loading group continued their exercises for a total of 12 weeks. 2.4. Eccentric loading exercises On the first appointment, patients were given verbal, written and pictorial instruction on the prescribed exercises and their techniques checked. Subjects were instructed to stand with the balls of the feet on the edge of a step with the heels raised and the knees straight (Fig. 1a). Subjects were then instructed to ‘‘Take the unaffected leg off the step. Slowly lower your affected heel as far as you can. Hold this position for 10 s (Fig. 1b). Place the unaffected leg back onto the step. Use the unaffected leg to help raise the body back up to the starting position. Hold this position for 10 s. Repeat.’’ This involved a progression of up to three lots of 15 repetitions of eccentric loading with both a straight and if pain and strength allowed, also a bent knee (Fig. 1c) as described by Alfredson et al. (1998). The number of repetitions was based on the subjects’ physical ability and pain provocation. Subjects were instructed to ‘‘continue the exercises even if you feel pain over the Achilles tendon unless it becomes disabling. Pain should not increase in between exercise periods. Progress to an increased number of repetitions or weight once exercise pain has settled.’’ Weight was added by using a backpack. A rest period of 1 min was given in between each set. The exercises were performed once a day, 7 days a week. Feedback on performance and exercise progression took place at subsequent appointments at 2, 4 and 6 weeks. 2.5. Ultrasound Pulsed 2:8 ultrasound (Sonomed 4 art-nr, Robert Bosch GMBH) using 3 MHz at 0.5 w/cm2 was applied

for 2 min/cm2 to the symptomatic area of the painful tendon (Watson, 2000). This was repeated twice a week over a period of 6 weeks, stopping before this if the patient’s symptoms were completely resolved or aggravated by the procedure. 2.6. Outcome measurements Outcome measurements, requiring minimal assistance, were taken prior to and after 2, 4, 6 and 12 weeks of commencing treatment. These were administered by the treating physiotherapist with the exception of the last, which were collected by one of the researchers, blinded to group allocation. 2.7. Pain measurement A 100 mm horizontal visual analogue scale (VAS) was used to measure pain intensity during rest, walking and if appropriate during recreational sport. This measurement tool has good reliability and validity (Huskisson 1974; Melzack and Katz, 1999). 2.8. Functional measurement The functional index of the leg and lower limb (FILLA) was used to measure the subjects’ own assessment of their ability to perform 11 functional activities on a VAS (Anderson and Styf, unpublished data; Salen et al., 1994). A mark at one end of the line indicates no difficulties performing the activity, whilst a mark at the other end of the line indicates that the subject is unable to perform the activity at present. The scores are totalled and converted to a percentage; the higher the percentage, the greater the disability.

Fig. 1. Eccentric loading exercises: (a) starting position; (b) following eccentric loading and (c) with a bent knee.

ARTICLE IN PRESS R. Chester et al. / Manual Therapy 13 (2008) 484–491

2.9. General health measurement The EuroQol generalised health questionnaire, with good validity and reliability (Brooks, 1996), was used to assess the results in relationship to the subjects’ general health and well being. 2.10. Statistics The difference in mean pain, FILLA and EUROQOL scores were calculated together with 95% confidence intervals, assuming a normal distribution.

Table 4 Mean pain scores after sports/recreation on the VAS (7 1 SD) for the ultrasound and eccentric loading groups, over the 12 weeks of study Ultrasound Eccentric loading Difference 95% CI Baseline 2 weeks 6 weeks 12 weeks

(16.6) (13.8) (14.6) (27.6)

70.0 65.6 63.5 61.9

(19.2) (24.1) (32.1) (23.5)

6.33 2.63 16.43 6.43

( 26.7–21.4) ( 45.5–12.6) ( 36.3–23.4)

Ultrasound Eccentric loading Difference 95% CI

Table 2 Mean pain scores during walking on the VAS (71 SD) for the ultrasound and eccentric loading groups, over the 12 weeks of study Ultrasound Eccentric loading Difference 95% CI 51.3 50.6 34.3 47.9

(25.6) (26.7) (22.3) (26.2)

50.5 50.8 42.6 48.5

(33.4) (29.3) (35.8) (23.4)

0.75 0.13 8.38 0.63

Eccentric loading

23.8 39.1 17.0 27.9

36.3 38.8 30.9 26.6

(15.1) (20.7) (19.4) (23.9)

(30.8) (27.5) (32.7) (25.9)

Difference

12.5 0.38 13.90 1.25

FIL Baseline 2 weeks 6 weeks 12 weeks

0.68 0.67 0.68 0.66 56.5 48.4 46.9 41.5

( 30.2–29.9) ( 40.4–23.6) ( 27.3–26.1)

95% CI

( 35.7–26.4) ( 42.7–14.9) ( 25.5–28.0)

(0.04) (0.22) (0.27) (0.31)

(6.9) (19.4) (20.9) (23.9)

0.65 0.61 0.61 0.64 45.2 51.0 46.0 53.0

(0.32) (0.29) (0.40) (0.31)

(22.6) (25.4) (30.0) (17.6)

0.03 0.06 0.07 0.02 11.3 2.6 0.9 11.6

( 0.22–0.33) ( 0.30–0.43) ( 0.31–0.36)

( 26.9–21.6) ( 26.8–28.7) ( 34.1–11.0)

65

EL US

55

45

35 0

Table 3 Mean pain scores during rest on the VAS (71 SD) for the ultrasound and eccentric loading groups, over the 12 weeks of study Ultrasound

EQ5D Baseline 2 weeks 6 weeks 12 weeks

FIL

Sixteen subjects completed the study. The randomisation process did result in demographic differences between the groups in a number of respects. As Table 1 indicates, the subjects in the eccentric loading group were older, had a greater proportion of women to men, had a longer duration of symptoms and had a greater number of additional pathologies than the subjects allocated to the ultrasound group. Half or more subjects in each group did not participate in any sporting or physical recreational activity. See Tables 2–5. Figs. 2–6 illustrate the findings for each outcome measure. There were some differences in baseline measurements between the groups. There was greater mean functional impairment (as measured by the FILLA), at baseline for the subjects in the ultrasound group in comparison with the subjects in the eccentric loading group. The reverse was true in terms of general

Baseline 2 weeks 6 weeks 12 weeks

76.3 63.0 47.1 55.4

Table 5 Mean Euroquol and FILLA scores (7 1 SD) for the ultrasound and eccentric loading groups, over the 12 weeks of study

3. Results

Baseline 2 weeks 6 weeks 12 weeks

487

2

6 Time (weeks)

12

Fig. 2. Mean FILLA scores for the ultrasound (US) and eccentric loading (EL) groups over the 12 weeks of study.

health measurement (EQ5D). This is supported by the additional number of existing pathologies in the eccentric loading group. Mean base line resting pain VASs were lower in the ultrasound group but pain after sports was higher. There was no difference between mean baseline pain VASs on walking. There were no statistical significant differences between groups at weeks 2, 4, 6 or 12. In addition, there was considerable overlap between the confidence intervals.

ARTICLE IN PRESS R. Chester et al. / Manual Therapy 13 (2008) 484–491

488

100

0.75

EL US

90

EL US

80 Sport Pain

EQ5D

70 0.65

60 50 40 30 20 10

0.55

0 0

2

6 Time (weeks)

12

0

2

6

12

Time (weeks)

Fig. 3. Mean EQ5D scores for the ultrasound (US) and eccentric loading (EL) groups over the 12 weeks of study.

Fig. 6. Mean pain scores following sports/recreation on the VAS for the ultrasound (US) and eccentric loading (EL) groups, over the 12 weeks of study.

100 EL US

90 80

At 12 weeks mean pain VASs had increased again in the ultrasound group and during walking in the eccentric group, although during rest and sport it continued to slowly decline in the later. Functional ability (FILLA) improved with each consecutive data collection from 0 to 12 weeks for the subjects in the ultrasound group. At 2 and 12 weeks subjects in the eccentric loading group demonstrated a decrease in self-assessed function.

Rest Pain

70 60 50 40 30 20 10 0 0

2

6 Time (weeks)

12

Fig. 4. Mean resting pain scores on the VAS for the ultrasound and eccentric loading groups, over the 12 weeks of study.

100 EL US

90 80 Walk Pain

70 60 50 40 30 20 10 0 0

2

6 Time (weeks)

12

Fig. 5. Mean walking pain scores on the VAS for the ultrasound and eccentric loading groups, over the 12 weeks of study.

Although not statistically significant at 6 weeks there were improvements in the mean pain VAS during rest, walking and sport particularly within the ultrasound group and to a lesser extent the eccentric loading group.

4. Discussion The results of this pilot study demonstrated that there was no statistically significant difference in the effectiveness of eccentric loading exercises compared with therapeutic ultrasound in the management of chronic Achilles tendon pain in subjects with relatively sedentary lifestyles in an NHS hospital setting. Mean scores for pain and function improved during the 6 weeks of ultrasound. Although pain at rest and during walking and sport had increased again at 12 weeks, function continued to improve. There were no marked changes in mean pain scores at any time during eccentric loading. Mean functional scores did, however, demonstrate a rise in dysfunction at 2 and 12 weeks. Reviews have demonstrated that although there are a number of randomised controlled trials investigating the clinical effectiveness of ultrasound these are hampered by poor methodology and therefore there is little supportive evidence for its clinical use (Van der Windt et al., 1999; Robertson and Baker, 2001; Speed, 2001). However, there is strong supporting evidence from a number of animal and in vitro studies of the positive effects of ultrasound on tendon healing (Saini et al., 2002; Ng et al., 2004), fibroblast proliferation (Ramirez et al., 1997; Tsai et al., 2005) and breaking strength (Stevenson et al., 1986; Enwemeka, 1989, 1990; Jackson

ARTICLE IN PRESS R. Chester et al. / Manual Therapy 13 (2008) 484–491

et al., 1991; Ng et al., 2003, 2004; Demir et al., 2004). However, the tendons investigated in animal studies have been surgically transected and sutured and are therefore dissimilar in composition to and do not display the degenerative changes and lack of inflammatory markers associated with chronic tendinopathy. Unfortunately, there are few human studies investigating the physiological effects of ultrasound in vivo (Baker et al., 2001). To the authors’ knowledge there are no studies investigating the beneficial effects of ultrasound on chronic Achilles tendon healing or pain resolution on human subjects. The lack of improvement for the subjects in the eccentric loading group did not concur with three previous RCTs (Mafi et al., 2001; Silbernagel et al., 2001; Roos et al., 2004) investigating the effectiveness of eccentric loading exercises. No studies were found which directly compare eccentric loading exercises with therapeutic ultrasound for patients with Achilles tendinopathy. However, Stasinopoulos and Stasinopoulos (2004) found significant improvement (po0.001) with a 4 week exercise programme, which included eccentric loading, in comparison with 4 weeks of ultrasound in terms of pain relief in a group of subjects with chronic patellar tendinopathy. Methodological factors may well have had an influence on our results. Random allocation rather than a matched subject design resulted in differences in a number of demographic features between the two treatment groups. Subjects in the eccentric loading group were older, had a greater proportion of women to men, a longer duration of symptoms and a greater number of additional pathologies. This may have biased the outcome possibilities in favour of the group receiving ultrasound in comparison with those receiving eccentric loading exercises. In addition, the subjects who received ultrasound saw the therapist more frequently than subjects in the eccentric loading group. The absence of any differences between the two groups was not unexpected given the small sample size. A study with walking pain, measured on a 100 mm VAS, as the primary end point, with an assumed standard deviation of 28 (based upon the data from this study) would require 166 subjects per group to provide 90% power, to detect a minimum clinically relevant difference in means of 10 mm. Based upon a minimum clinically relevant difference of 15 mm, 75 subjects per group would be required and based upon a minimum clinically relevant difference of 20 mm (Crossley et al., 2004; Tubach et al., 2004) 43 subjects per group would be required. Imaging and histopathology was not used to confirm a clinical diagnosis. Therefore, a diagnosis of tendinopathy rather than tendinosis (Alfredson, 2005) would be more precise for our patients. However, it is reasonable to suggest that the changes documented in the literature which are associated with tendinosis are likely to be

489

associated with signs and symptoms presented by the subjects in our study. In the absence of any signs of inflammation (A˚stro¨m and Rausing, 1995; Khan et al., 1999; Alfredson et al., 1999) a number of pathophysiological changes have been observed within the Achilles tendons of symptomatic individuals, which have not been observed in those with normal pain free tendons. Alfredson et al. (2002) identified higher concentrations of lactate within the tendons of symptomatic subjects and suggest that this may be an indication of anaerobic respiration due to ischaemic conditions within the tendon. Circulatory changes, all be it an increase, possibly in response to ischaemic changes (O¨hberg and Alfredson, 2004) has been identified as a factor in the presentation of Achilles tendinopathy. Elevated blood flow (A˚stro¨m and Westlin, 1994), neovascularisation (A˚stro¨m and Rausing, 1995; O¨hberg et al., 2001) and recently increased glutamate (Alfredson et al., 1999) have been identified in painful Achilles tendons but not in those of pain free control groups. Alterations in the structure and arrangement of collagen fibres and an increase in non-collagenous matrix have also been identified in symptomatic Achilles tendons (A˚stro¨m and Rausing, 1995; Khan et al., 1999). However, the relationship between these findings is unclear. Imaging studies have demonstrated that some of the pathophysiological changes identified in symptomatic tendons reverse following a course of eccentric loading exercises. Using various imaging techniques, O¨hberg and Alfredson (2004), and Shalabi et al. (2004) investigated the changes which occurred following a 12-week course of eccentric loading exercises, in a group of 34 subjects with Achilles tendinopathy in the former group and 25 subjects in the latter. The majority of subjects became asymptomatic and demonstrated a return to normalised Achilles tendon structure (O¨hberg and Alfredson, 2004; Shalabi et al., 2004) and a corresponding reverse in neovascularisation (O¨hberg and Alfredson, 2004). The high torques associated with eccentric loading in comparison with concentric loading may cause interfibrillar disruption and stimulate a healing response (Shalabi et al., 2004). Matrix remodelling and collagen realignment are perhaps a response to the increased demands placed upon the tendon during sustained eccentric loading. The subjects in this study led sedentary or relatively sedentary lifestyles and a number had additional medical conditions. For some subjects any form of exercise was a cultural shift and some had difficulty gaining skill acquisition of the eccentric loading. Indeed only one subject progressed to using a backpack with weights and a number of subjects were unable to progress to performing the exercise with a bent knee. Alfredson et al. (1998) describe their study population as ‘‘recreational athletes’’, with all but one (soccer), classifying ‘‘jogging’’ as their main sport. From the

ARTICLE IN PRESS 490

R. Chester et al. / Manual Therapy 13 (2008) 484–491

tabulated data presented from Silbernagel et al.’s (2001) study it appears that 39 of 40 subjects were involved in a named sport. Sixty-seven percent of the subjects in Fahlstrom et al.’s (2003) study and 65% of those in Roos et al.’s (2004) study were involved in running or a sporting activity. The rest of Fahlstrom et al.’s (2003) subjects were actively involved in walking. To be included in Roos et al.’s study activity levels were ‘‘at least equivalent to heavy household work, heavy yard work and walking on uneven ground’’. Twenty-five of the 44 subjects in Mafi et al.’s (2001) study were involved in jogging and the rest described walking as their level of activity. It is reasonable to suggest that the sedentary or relatively sedentary lifestyle of the subjects in our study in comparison with the majority of subjects in the studies above is a likely contributing factor to our results. It may be prudent to consider whether or not our subjects exercised into sufficient pain during eccentric loading. Subjects were instructed to exercise into pain unless it became disabling and to progress the exercises once pain had settled. Exercise compliance, although actively promoted during each appointment, was not formally monitored and compromised adherence to any aspect of the eccentric loading exercises may account for the differences between study groups. During full dorsiflexion, Doppler imaging has shown that blood supply in the neoneurovascular area decreases or is abolished (A˚stro¨m and Westlin, 1994; O¨hberg et al., 2001). These authors suggest that the increased pain response, which frequently occurs in the first couple of weeks, may be due to neurovascular damage, caused by repeated irritation during the exercise. This concurs with O¨hberg and Alfredson’s (2002) findings that sclerosis of neurovascular structures can reduce pain in subjects with chronic Achilles tendinopathy. Perhaps loading during eccentric exercise needs to be sufficient enough to cause injury to these neoneurovascular structures, in order to achieve a positive clinical outcome. The recommendation that eccentric loading exercises are performed into pain is, however, still largely based on hypothesis. Silbernagel et al. (2001) divided 49 subjects with Achilles tendinopathy into two groups, both of whom performed eccentric loading exercises. However, one group performed the exercises into pain whilst the other stopped just short of pain. There were no significant differences between the successful results for both groups. Whilst subjects who exercised into pain demonstrated a generally better outcome over time, a pain-free and painful version of the same exercise was not given to each group, as the exercises performed by the groups varied. Although there are hypothetical principles to support the performance of eccentric exercises into pain, the clinical evidence is still lacking. Eccentric loading exercises are becoming increasingly recognised as a potential gold standard for the manage-

ment of chronic Achilles tendinopathy. This is based on studies largely comprising of ‘‘recreational athletes’’ or subjects partaking in some form of regular exercise. Many subjects presenting with Achilles tendinopathy do not fit within such categories. Over half of our subjects did not do any form of exercise. For these subjects, eccentric exercises were both difficult to perform and progress, and particularly into pain. However monitored, compliance is likely to be an issue for these subjects, especially in the absence of any improvement after 6 weeks. The positive results in temporary mean pain and continuing functional improvement scores associated with ultrasound may be due to clinical effectiveness or placebo. Despite clear supportive evidence of the effectiveness of ultrasound for acute tendon healing and remodelling in animal studies, there is an absence of supportive physiological findings and clinical effectiveness in methodologically sound human studies for any stage of tendon pathology, but particularly in the chronic stage. However, within this small pilot study, ultrasound was of potentially more beneficial than the emerging gold standard.

5. Conclusion Although this pilot study indicates no difference in outcome measurements between heavy eccentric loading and ultrasound, for the management of Achilles tendinopathy in subjects with a relatively sedentary lifestyle, this is not unexpected given the small size. Both interventions proved acceptable to the patients, with no adverse effects.

Acknowledgement The authors thank the subjects who took part in the study and the Physiotherapists who generously gave of their time to implement the treatments required for each group. In addition, out thanks to Toby Smith and Jo Geere for proof reading the article and providing valuable feedback. References Alfredson H. The chronic painful Achilles and patella tendon: research on basic biology and treatment. Scandinavian Journal of Medicine & Science in Sports 2005;15:252–9. Alfredson H, Pietila¨ T, Jonsson P, Lorentzon R. Heavy load eccentric calf muscle training for the treatment of chronic Achilles tendinosis. American Orthopaedic Society for Sports Medicine 1998;26(3):360–6. Alfredson H, Thorsen K, Lorentzon R. In situ microdialysis in tendon tissue: high levels of glutamate, but not prostaglandin E2 in chronic Achilles tendon pain. Knee Surgery, Sports Traumatology, Arthroscopy 1999;7:378–81.

ARTICLE IN PRESS R. Chester et al. / Manual Therapy 13 (2008) 484–491 Alfredson H, Bjur D, Thorsen K, Lorentson R. High intratendinous lactate levels in painful chronic Achilles tendinosis. An investigation using microdialysis technique. Journal of Orthopaedic and Related Research 2002;20:934–8. Anderson G, Styf J. Functional index of the leg (FIL): evaluation of a new questionnaire assessing chronic leg pain induced by exercise. Unpublished data. A˚stro¨m M, Rausing A. Chronic Achilles tendinopathy. A survey of surgical and histopathologic findings. Orthopaedics and Related Research 1995;25(316):151–64. A˚stro¨m M, Westlin N. Blood flow in chronic Achilles tendinopthy. Clinical Orthopaedics and Related Research 1994;25(308): 166–72. Baker KG, Robertson VJ, Duck FA. A Review of therapeutic ultrasound: biophysical effects. Physical Therapy 2001;81(7): 1351–8. Brooks R. EuroQol: the current state of play. Health Policy 1996;37:53–72. Crossley KM, Bennell KL, Cowan SM, Green S. Analysis of outcome measures for persons with patellofemoral pain: which are reliable and valid? Archives of Physical Medicine and Rehabilitation 2004;85:815–22. Demir H, Menku P, Kirnap M, Calis M, Ikizceli I. Comparison of the effects of laser, ultrasound, and combined later and ultrasound treatments in experimental tendon healing. Lasers in Surgery and Medicine 2004;35:84–9. Enwemeka CS. The effects of therapeutic ultrasound on tendon healing. A biomechanical study. American Journal of Physical Medicine and Rehabilitation 1989;68:283–7. Enwemeka CS, Rodriguez O, Medosa S. The biomechanical effects of low intensity ultrasound on healing tendons. Ultrasound in Medicine and Biology 1990;16:801–7. Fahlstro¨m M, Jonsson P, Lorentzon R, Alfredson H. Chronic Achilles tendon pain treated with eccentric calf muscle training. Knee Surgery, Sports Traumatology, Arthroscopy 2003;11: 327–33. Huskisson EC. Measurement of pain. Lancet 1974;9:1127–31. Jackson BA, Schwane JA, Starcher BC. Effect of ultrasound therapy on the repair of Achilles tendon injuries in rats. Medicine and Science in Sports and Exercise 1991;23(2):171–6. Khan KM, Cook JL, Bonar F, Harcourt P, A˚stro¨m M. Histopathology of common tendinopathies. Update and implications for clinical management. Sports Medicine 1999;27:393–408. Lindsay DM, Dearness J, McGinley CC. Electrotherapy usage trends in private physiotherapy practice in Alberta. Physiotherapy Canada 1995;47(1):30–4. Mafi N, Lorentzon R, Alfredson H. Superior short term results with eccentric calf muscle training compared to concentric training in a randomized prospective multicentre study on patients with chronic Achilles tendinosis. Knee Surgery, Sports Traumatology, Arthroscopy 2001;9:42–7. Magee DJ. Orthopaedic Physical Assessment. Philadelphia: WB Saunders Company; 1997 636pp. [Chapter 13]. McLaughlan GJ, Handoll HHG. Interventions for treating acute and chronic Achilles tendinitis (Cochrane Review). Cochrane Library 2001(Issue 2):1–35. Melzack R, Katz J. Pain measurement in persons in pain. In: Wall PD, Melzack R, editors. Textbook of pain. 4th ed. Edinburgh: Churchill Livingstone; 1999. p. 411 [chapter 17]. Ng COY, Ng GYF, See EKN, Leung MCP. Therapeutic ultrasound improves strength of Achilles tendon repair in rats. Ultrasound in Medicine and Biology 2003;29(10):1501–6. Ng COY, Ng GYF, See EKN, Leung MCP. Comparison of therapeutic ultrasound and exercises for augmenting tendon healing in rats. Ultrasound in Medicine and Biology 2004;30(11): 1539–43.

491

O¨hberg L, Alfredson H. Ultrasound guided sclerosis of neovessels in painful chronic Achilles tendinosis: pilot study of a new treatment. British Journal of Sports Medicine 2002;36:173–7. O¨hberg L, Alfredson H. Effects on neovascularisation behind the good results with eccentric training in mid-portion Achilles tendinosis. Knee Surgery, Sports Traumatology, Arthroscopy 2004;12:465–70. O¨hberg L, Lorentzon R, Alfredson H. Neovascularisation in Achilles tendons with painful tendinosis but not normal tendons: an ultrasonographic investigation. Knee Surgery, Sports Traumatology, Arthroscopy 2001;9:233–8. Pope GD, Mockett SP, Wright JP. A survey of electrotherapeutic modalities: ownership and use in the national health service in England. Physiotherapy 1995;81(2):82–91. Ramirez A, Schwane JA, McFarland C, Starcher B. The effect of ultrasound on collagen synthesis and fibroblast proliferation in vitro. Medicine and Science in Sports and Exercise 1997;29:326–32. Robertson VJ, Baker KG. A review of therapeutic ultrasound: effectiveness studies. Physical Therapy 2001;81(7):1339–50. Robertson VJ, Spurritt D. Electrophysical agents: implications of their availability and use in undergraduate clinical placements. Physiotherapy 1998;84(7):335–44. Roos EM, Engstro¨m M, Lagerquist A, So¨derberg. Clinical improvement after 6 weeks of eccentric exercise in patients with midportion Achilles tendinopathy—a randomized trial with 1 year follow up. Scandinavian Journal of Medicine & Science in Sports 2004;14:286–95. Saini NS, Roy KS, Bansal PS, Singh B, Simran PS. A preliminary study on the effect of ultrasound therapy on the healing of surgically severed tendons in five dogs. Journal of Veterinary Medicine 2002;49:321–8. Salen BA, Sprangfort EV, Nygren AK, Nordemar R. The disability rating index: an instrument for the assessment of disability in clinical settings. Journal of Clinical Epidemiology 1994;47(12):1423–34. Shalabi A, Svensson L, Kristoffersen-Wilberg M, Aspelin P, Movin T. Tendon injury and repair after core biopsies in chronic Achilles tendinosis evaluated by serial magnetic resonance imaging. British Journal of Sports Medicine 2004;38:606–12. Silbernagel KG, Thomee´ R, Thomee´ P, Karlsson J. Eccentric overload training for patients with chronic Achilles tendon pain–a randomised controlled study with reliability testing of the evaluation methods. Scandinavian Journal of Medicine & Science in Sports 2001;11:197–206. Speed CA. Therapeutic ultrasound in soft tissue lesions. Rheumatology 2001;40:1331–6. Stasinopoulos D, Stasinopoulos I. Comparison of effects of exercise programme, pulsed ultrasound and transverse friction in the treatment of chronic patellar tendinopathy. Clinical Rehabilitation 2004;18:347–52. Stevenson JH, Pang CY, Lindsay WK, Zuker RM. Functional, mechanical and biochemical assessment of ultrasound therapy on tendon healing in the chicken toe. Plastic and Reconstructive Surgery 1986;77(6):965–70. Tsai WC, Chou YJ, Chen YJ, Hsu CC, Tang FT, Pang JHS. Ultrasound stimulation of tendon proliferation and upregulation of proliferating cell nuclear antigen. Journal of Orthopaedic Research 2005;23(4):970–6. Tubach F, Ravaud P, Baron G, Falissard B, Logeart I, Bellamy N, et al. Evaluation of clinically relevant changes in patient-reported outcomes in knee and hip osteoarthritis: the minimal clinically important improvement. Annals of the Rheumatic Diseases, published online first: 6 May 2004, doi:10.1136/ard.2004.022905. Van der Windt DAWM, van der Heijden GJMG, van den Berg SGM, ter Riet G, de Winter AF, Bouter LM. Ultrasound therapy for musculoskeletal disorders: a systematic review. Pain 1999;81:257–71. Watson T. The role of electrotherapy in contemporary physiotherapy practice. Manual Therapy 2000;5(3):132–41.

ARTICLE IN PRESS

Manual Therapy 13 (2008) 492–499 www.elsevier.com/math

Original article

An investigation into the effects of a unilaterally applied lumbar mobilisation technique on peripheral sympathetic nervous system activity in the lower limbs Jo Perry, Ann Green Department of Physiotherapy and Dietetics, Faculty of Health and Life Sciences, Coventry University, Priory Street, Coventry CV1 5FB, UK Received 7 November 2006; received in revised form 16 April 2007; accepted 23 May 2007

Abstract Physiotherapeutic management of lumbar disorders often utilises specific segmental joint mobilisation techniques; however, there is only limited evidence of any neurophysiological effects and much of this has focused on the cervical spine and upper limbs. This study aims to extend the knowledge base underpinning the use of a unilaterally applied lumbar spinal mobilisation technique by exploring its effects on the peripheral sympathetic nervous system (SNS) of the lower limbs. Using a double blind, placebo controlled, independent groups study design and based upon power calculations, 45 normal naı¨ ve healthy males were randomly assigned to one of three experimental groups (control, placebo or treatment; a unilaterally applied postero-anterior mobilisation to the left L4/5 zygopophyseal joint). SNS activity was determined by recording skin conductance (SC) obtained from lower limb electrodes connected to a BioPac unit. Validation of the placebo technique was performed by postintervention questionnaire. Results indicated that there was a significant change in SC from baseline levels (13.5%) that was specific to the side treated for the treatment group during the intervention period (compared to placebo and control conditions). This study provides preliminary evidence that a unilaterally applied postero-anterior mobilisation technique performed, at a rate of 2 Hz, to the left L4/5 lumbar zygopophyseal joint results in side-specific peripheral SNS changes in the lower limbs. r 2007 Elsevier Ltd. All rights reserved. Keywords: Lumbar spine; Physiotherapy; Joint mobilisation; Sympathetic nervous system

1. Introduction Lumbar spine disorders reportedly affect around 17.3 million people in the United Kingdom at an annual cost of £1bn to the NHS and £565m to private healthcare providers (Maniadakis and Gray, 2000). Since the publication of the recommendations of the Clinical Standards Advisory Group (CSAG, 1994), the Royal College of General Practitioners (RCGP, 1998), the UKBEAM trial (UKBEAM Trial Team, 2004) and the Clinical Guidelines for the management of Persistent Corresponding author. Tel.: +44 24 7688 7890; fax: +44 24 7688 8020. E-mail address: [email protected] (J. Perry).

1356-689X/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2007.05.015

LBP (CSP, 2006), the use of spinal manipulative therapy (SMT) techniques for patients with lumbar symptoms has received support as a management strategy with 58.9% of UK therapists (Foster et al., 1999) and 83.7% of North American therapists (Li and Bombardier, 2001; Poitras et al., 2005) utilising SMT as a preferred treatment for patients with low back pain (LBP). The available evidence regarding any neurophysiological or analgesic effects that occur as a result of treatment is not always clear (Assendelft et al., 1995). Although, the last two decades have seen the emergence of conceptual models supporting the use of SMT in the management of segmental pain and joint dysfunction (Zusman, 1986, 2004; Wright, 1995). Hypothetical mechanisms of action of SMT include: direct physiological effects on articular

ARTICLE IN PRESS J. Perry, A. Green / Manual Therapy 13 (2008) 492–499

and peri-articular structures (Wyke and Polacek, 1975), neurophysiological mechanisms at both spinal and supraspinal levels (Wright, 1995, 1999; Solly, 2004), endocrine/hormonal effects and a non-specific placebo effect (Zusman, 1986, 2004; Katavich, 1998). The dilemma facing the clinician, in determining the physiological and therapeutic effects of SMT interventions on patients, is the difficulty in accurately and quantitatively measuring the proposed effects on the key target tissue. Several researchers (Petersen et al., 1993; Vicenzino et al., 1994, 1996; Wright and Vicenzino, 1995; Chiu and Wright, 1996; Sterling et al., 2001) have explored the neurophysiological basis of specific SMT techniques in the cervical spine and upper limbs, utilising the sympathetic nervous system (SNS) as a measure of neurophysiological response. Specific SNS changes have been reported following SMT, namely; sudomotor function (Petersen et al., 1993; Slater et al., 1994; Vicenzino et al., 1994; Wright and Vicenzino, 1995; Chiu and Wright, 1996; Sterling et al., 2001); cutaneous vasomotor changes (Petersen et al., 1993) and cardiac and respiratory functions (McGuiness et al., 1997; Vicenzino et al., 1998a). Chiu and Wright (1996) compared the effects of two different frequencies/speeds of oscillation (0.5 and 2 Hz) of a cervical spine joint mobilisation technique, and observed significant SNS responses with the 2 Hz technique. This work supported that of McGuiness et al. (1997) who investigated the effects of mobilisation techniques on the magnitude of the SNS response, revealing that an oscillatory technique produced the greatest effect. This is corroborated, in part, by the findings of Moulson and Watson (2006) who applied a non-oscillatory cervical sustained natural apophyseal glide (SNAG). Despite having a sympathoexcitatory response, the magnitude of this response was not significantly greater than the placebo condition within the treatment period suggesting that the oscillatory component of the technique is an important factor in changes to SNS activity (Kenney et al., 1991; Gebber et al., 1999). Neurophysiological (SNS) effects following SMT have revealed that in humans SMT produces an immediate hypoalgesic and sympathoexcitatory effect on both asymptomatic and symptomatic subjects that are specific to mechanical nociception as opposed to thermal nociception and sympathoinhibition (Vicenzino et al., 1995, 1996, 1998b; Sterling et al., 2001). These findings have led to the concept that SMT may exert its initial effects by activating descending pain inhibitory systems (DPIS) from the peri-acqueductal gray (PAG) region of the brain and, depending on whether the response is excitatory or inhibitory could indicate whether or not it is a dorsal PAG or a ventral PAG response (Wright, 1995; Wright and Vicenzino, 1995). Kenney et al. (1991) have suggested the possibility of selective and functionally complementary coupling of

493

different post-ganglionic sympathetic nerves and, specifically, the presence of an ipsilateral preference towards observed SNS discharge. So far, these findings have not yet been applied to physiotherapy research into SMT and the lumbar spine. Indeed, studies supporting this conceptual framework have been conducted on the cervical spine (Petersen et al., 1993; Slater et al., 1994; Vicenzino et al., 1995, 1996, 1998b; Wright and Vicenzino, 1995; Chiu and Wright, 1996; Sterling et al., 2001; Moulson and Watson, 2006) and upper limbs (Vicenzino et al., 1994; Simon et al., 1997). This study aimed to explore the hypothesis that a specific mechanical mid-to-end range mobilisation technique, applied to the left Lumbar 4/5 ZP joint at a rate of 2 Hz, would result in a significant change from baseline in peripheral SNS activity (as measured by skin conductance—SC) that would be greatest in the ipsilateral leg during the intervention period compared to the contra-lateral limb, the placebo and the control conditions.

2. Methodology 2.1. Subjects The study recruited a convenience sample of 45 healthy, physiotherapeutically naı¨ ve, asymptomatic, non-smoking male volunteers (aged 18–25 years, mean 21.5 years, S.D. ¼ 1.85). All volunteers were further assessed for their suitability using inclusion and exclusion criteria as described in other studies (Vicenzino et al., 1994). Subjects were randomly assigned to one of the three subject groups using the third party, concealed randomisation method (Schulz et al., 1995). An all-male group was used, to provide a greater degree of matching of the sample groups and to negate the effects on variance that the female hormone progesterone has on electrodermal responses (Venables and Christie, 1973). Table 1 summarises the subjects anthropometric characteristics. Vicenzino et al. (1995) recorded SC values in control, placebo and treatment conditions. Based on their intrasubject standard deviation of 13.1% (control) a power analysis calculation revealed that 45 subjects (15 per sample group) would enable a difference in SC values from baseline of 7.5% to be detected at the 5% significance level with 80% power (Pocock, 1991). A 7.5% SC value difference was chosen because it was felt by the authors that this would represent a clinically significant change and has been supported by the findings of other researchers (Vicenzino et al., 1995, 1996; Sterling et al., 2001; Moulson and Watson, 2006). Ethical approval was obtained from the Coventry University research ethics committee. All volunteers received written information and gave consent prior to the experiment.

ARTICLE IN PRESS J. Perry, A. Green / Manual Therapy 13 (2008) 492–499

494

Table 1 Anthropometric characteristics of all subjects (age, weight and height) in each experimental group (control, placebo and treatment groups)

Age (years) Mean Range SD Weight (kg) Mean Range SD Height (cm) Mean Range SD

All subjects

Control group

Placebo group

Treatment group

Levene’s test of homogeneity of variance (P-value)

21.5

21.7

21.5

21.4

0.953y

18–25 71.85

18–25 71.95

18–25 71.85

18–25 71.88

72.1

72.4

72.1

71.8

62.5–82.5 74.66

65.0–80.0 74.34

62.5–82.5 75.22

62.5–80.0 74.69

172.4 165–182 74.13

172.7 165–182 74.70

172.6 165–180 73.98

172.0 165–180 73.91

0.892y

0.747y

kg: kilograms; cm: centimetres; SD: standard deviation; y Non-significant value where levels of significance were set at Po0.05.

2.2. Pilot studies Two pilot studies, using a motion analyser (Chester and Watson, 2000), established the consistency of the researcher, an appropriately experienced manipulative physiotherapist (Downey et al., 1999), regarding their ability to reproduce the grade (depth and oscillatory frequency) of the mobilisation technique. For the frequency of oscillation of the technique (2 Hz) a single-measure intra-class correlation co-efficient (ICC) value of 0.96 was obtained with force calibrations consistent with other studies (Hardy and Napier, 1991; Snodgrass et al., 2006). For repeatability of depth of mobilisation an ICC value of 0.80 was achieved. Both are considered to represent a satisfactory measure of reliability (Portney and Watkins, 2000). 2.3. Research design A double-blind, independent (matched) group, between-subjects experimental research design was used, with each subject being randomly allocated to either the control, placebo or treatment group. An independent, matched-subjects design (as opposed to a factorial or crossover design typified by the repeated measures design of many similar studies) eliminated any order effects, thus enhancing the effectiveness and feasibility of the study where order effects and dropouts are thought to be likely (Bland and Altman, 1994; Blackwood and Lavery, 1998; Sibbald and Roberts, 1998) and where the validity of the placebo condition might otherwise be challenged (Svedmyr, 1979). Internal validity of the study was enhanced by doubleblinding. The equipment was installed in a screened area adjacent to the treatment plinth, therefore blinding the

data collector to the intervention being undertaken and to the hypotheses being tested, thereby eliminating any bias due to expectations, predictions or preferred outcome. Further blinding was achieved with this setup as neither the subjects nor the researcher received any feedback from the data collector regarding SNS activity. The success of the subject blinding to the allocated intervention group was investigated via a previously validated post-trial questionnaire (Borkovec and Nau, 1972; Vincent and Lewith, 1995). 2.4. Research method and experimental interventions The study was conducted in a soundproofed, temperature-controlled laboratory (Ja¨nig and Ha¨bler, 2003). The subjects lay prone on the treatment plinth in a standardised position. Skin conductance readings were recorded using silver/silver chloride electrodes (12 mm electrode gel contact area) applied to the dorsum of the 2nd and 3rd toes of both feet simultaneously, giving an independent, single reading for each foot. These toes were selected because the L4/5 segment has a cutaneous branch, the medial plantar nerve, which supplies the plantar aspect of the 2nd and 3rd toes. Following the protocol employed by previous researchers (Petersen et al., 1993; Chiu and Wright, 1996), subjects underwent an initial 10-min stabilisation period followed by a 2-min baseline period. Subsequently, the subject received one of the three experimental conditions within the 5-min intervention period. These included either: 1. Treatment: a unilaterally applied grade III oscillatory mobilisation, at a rate of 2 Hz, to the left L4/5 facet joint.

ARTICLE IN PRESS J. Perry, A. Green / Manual Therapy 13 (2008) 492–499

2. Placebo: that replicated the same therapist-subject hand positioning as the unilateral L4/5 mobilisation technique but only light/minimal pressure was exerted, without the application of any oscillatory movement. 3. Control: identical subject positioning but without manual contact or joint oscillatory movement. The placebo and treatment conditions were each applied for 1-min, with a 1-min rest between each application, for a total of 5-min. Recording was continued for a further 5-min and this period of time was termed the final rest period. Segment files were created from the raw data recorded for each study session. At the end of each experimental, the subject was asked to complete the post-experiment questionnaire to establish the validity of the placebo technique.

495

(ipsilateral/left side). To explore differences, the data were subjected to multivariate analysis using a mixed between-within subject design involving analysis of one between-subject factor (three levels; control, placebo and treatment groups), and two within-subject factors; time (from baseline to intervention, and from baseline to final rest period) and leg (right or left). One-way ANOVA was also used to evaluate the placebo validation questionnaire. Statistically significant differences between the experimental groups were subjected to post hoc analyses to identify where differences in SC and differences in the questionnaire results lay (Tukey’s Honestly Significant Difference test and Kruskal– Wallis test, respectively). Throughout all analyses, the statistical significant P-value was set at Po0.05 (two-tailed test).

5. Results 3. Instrumentation and measurement 5.1. Laboratory conditions Physiological recording of SC was measured by a Biopac GSR100B Electro-dermal Activity Amplifier (MP30; Biopac Systems Inc., Santa Barbara, CA), employing a constant voltage technique and sampling the absolute, direct current SC at the rate of 20 samples per second. Analysis of the SC data obtained involved calculation of the ‘‘Integral Measurement’’ for baseline, intervention and final rest periods. These were normalised to the time period of each experimental phase. Intervention period and final rest period values were then converted into percentage change (PC) from baseline using the formula below (Rowland and Tozer, 1989): PC ¼

N y  100, y

where y is the baseline reading and N the new reading in the intervention or final rest period.

4. Data analysis With independent, matched-group designs, it is important to establish homogeneity of the subjects otherwise it is difficult to determine the influence that individuals may have on the final analysis (Sim and Wright, 2002). Statistical analysis of the data was designed to test the three elements of the hypothesis: (1) that the SMT treatment would result in a PC in SC values from baseline that were greater than those of the placebo and control conditions; (2) that any PC observed would be greatest during the intervention period compared to the final rest period; and (3) that any observed PC would be specific to the side treated

Room temperature was recorded at the beginning and end of each subject’s experimental session as per published guidelines (Uematsu et al., 1988) with relative constancy within each session demonstrated (mean 24.9 1C, SD 0.275, range 24.4–25.5 1C) with a maximum within subject experimental room temperature variation being no more than 0.3 1C (mean 0.2 1C, SD 0.1 1C, range 0.0–0.3 1C). 5.2. Homogeneity of the matched groups All 45 subjects completed the study (15 per group). Statistical analysis of subject variance (Levene’s Test of Homogeneity) revealed that the subjects in the experimental groups were well matched (Table 1). 5.3. Skin conductance differences between and within groups Table 2 illustrates the data for each of the experimental conditions, and for the dependent variables of PC in SC and lower limb (right and left). Fig. 1 illustrates the central tendencies and variances for SC data (PC from baseline). There was a significant difference (F ¼ 7.47, P ¼ 0.002) between the experimental conditions, the time periods and the limbs. Post hoc analysis revealed that this difference was with the SMT treatment group in the left lower limb during the intervention period (P ¼ 0.005). Statistical analysis of PC differences in the final rest period revealed that there was no statistically significant difference in PC SC values between and within the experimental groups and between the legs (within subjects: F ¼ 1.82, P ¼ 0.175; between subjects: F ¼ 0.36, P ¼ 0.701).

ARTICLE IN PRESS J. Perry, A. Green / Manual Therapy 13 (2008) 492–499

496

Table 2 Table of percentage change (PC) in SC readings from the baseline (mean)7standard error (S.E.) for each of the experimental conditions during the intervention (PC1) and final rest (PC2) periods

distinguish between the genuine treatment and the placebo condition but could recognise when they had received the control condition.

Subject group and side Control

Placebo

6. Discussion

Treatment

Right

Left

Right

Left

Right

Left

PC1 in SC S.E.

1.08 73.01

0.87 71.82

0.06 74.24

1.93 73.54

4.11 710.65

13.47 720.26

PC2 in SC S.E.

1.16 72.00

1.58 73.52

2.81 72.00

4.45 73.52

4.99 72.00

2.35 73.52

PC1: percentage change from baseline intervention period; PC2: percentage change from baseline rest period.  A statistically significant value where the level of significance is set at Po0.05.

80 Percentage Change from baseline (0%)

Right Limb Left Limb

60

40 20 0

-20 -40

N=

15

15

control

15

15

placebo

15

15

treatment

Experimental Condition Fig. 1. Cluster boxplot (including error bar) illustrating the distribution of skin conductance percentage change values (%) for the three experimental conditions during the intervention period for the right and the left limbs (*represents extreme subject values).

5.4. Post-treatment questionnaire Statistical analysis of the post-treatment questionnaire was performed and H values were calculated for each question. The results revealed that there was a significant difference in the perceptions of the subjects as to whether they had received the treatment, placebo or control condition. The results were found to be significant (P ¼ 0.001) for differences between control and placebo, and control and treatment groups. A comparison of the treatment and placebo groups revealed that there was no significant difference (P ¼ 0.388). Therefore, the subjects were unable to

The findings of this study demonstrated that a unilaterally applied antero-posterior accessory mobilisation technique administered at a rate of 2 Hz to the left side of the L4–5 segment resulted in statistically significant side-specific changes in peripheral SNS activity during the intervention period, and that this response was greater than those of the contra-lateral limb, and of both the placebo and control conditions. To the authors’ knowledge, no previous studies have investigated whether SMT applied to the lumbar spine would result in a peripheral SNS effect in the lower limbs. Overall, the subjects demonstrated a sympathoexcitatory response which corresponded with SMT application, and is consistent with previous studies performed on the cervical spine and upper limb (Ellestad et al., 1988; Petersen et al., 1993; Slater et al., 1994;Vicenzino et al., 1994, 1995, 1996, 1998b; Chiu and Wright, 1996; Simon et al., 1997; Sterling et al., 2001; Moulson and Watson, 2006). A notable difference between the current study and those of the upper quadrant is that the unilaterally applied technique resulted in greatest response on the side of treatment. This is a unique finding when compared to previous studies that reported either bilateral responses or sought to record data solely on the side treated. This may represent a difference in the design of the current study, a difference in this studies focus or a difference in the nature of the SNS response in the lower quadrant and is an area of investigation that is certainly worthy of further, future exploration. Regarding the magnitude of the response, the current study showed a PC of 13.5% (720.25) which concurs with the study by Sterling et al. (2001) who reported increases of 16% (72.96) for the treatment condition in the intervention period. Other studies have reported greater SNS changes following SMT (33% and 150% in Vicenzino et al., 1994, 1995, respectively) which may reflect regional differences in peripheral cutaneous innervation or may suggest that the treatment technique advocated in this study may not have been the optimum technique for the production of maximum SNS response. Nevertheless, it does demonstrate that SC readings could represent a proxy measure for postganglionic efferent SNS activity and could be an instrument for future studies in the quantification of neurophysiological responses to physiotherapy treatments. Central to the premise that manual therapy stimulates the SNS and purportedly may activate a descending

ARTICLE IN PRESS J. Perry, A. Green / Manual Therapy 13 (2008) 492–499

pain inhibitory system (DPIS) is the concept that SMT stimulates local receptors, which in turn are capable of directly or indirectly activating the PAG mechanisms (Zusman, 1986; Wright 1995). This study demonstrated that a unilaterally applied SMT resulted in ipsilateral sympathoexcitation, which might suggest activation of the dPAG (noradrenaline) and the DPIS (Lovick, 1991), although this requires further investigation into any associated affects on mechanical hypoalgesia in the lower quadrant. Kenney et al. (1991) and Gebber et al. (1999) may consider that the results of the current study, in part, provides limited support for a model of selective coupling of local, segmental vertebral post-ganglionic sympathetic neurones but this hypothesis requires further experimental exploration and was not within the scope of the current study. Some authors advise that the local receptors could be located within the musculoskeletal system in joints, capsule, ligaments, connective tissue and tendons (Wyke and Polacek, 1975; Yezierski, 1991). Pickar (1995) demonstrated that manipulation of cat spinal joints stimulated receptors and afferent nerve fibres within the capsule and associated connective tissues of the spinal column. Furthermore, Wyke and Polacek (1975) and Katavich (1998) suggest that stimulation of large diameter, low threshold mechanoreceptors in articular and periarticular structures by SMT may produce a local spinal cord inhibitory effect and that these effects represent predictions of the ‘Gate Control’ theory (Melzack and Wall, 1996). However, Zusman (1986) has challenged the ability of SMT to preferentially stimulate large diameter joint afferents at the expense of small diameter, high threshold afferents, arguing that the proposed hypoalgesic effects of SMT include hysteresis, a decrease in joint afferent activity following sustained or repetitive passive movement. Ultimately, it is possible that the unilaterally applied technique directly stimulated local sympathetic fibres especially as the ganglia have a close anatomical relationship with the vertebral motion segment (Slater, 2002) and therefore the observed SNS excitatory response may simply be a spinal reflex (Magoun, 1978). Sterling et al. (2001) reported changes in superficial muscle activity following SMT arguing that this may be a response to a locally induced muscle stretch and stimulation of mechanoreceptors with resultant activation of segmental myogenic spinal reflex mechanism. Indeed, it is feasible that the technique in the current study may have been a SNS response to direct compression of the lumbar musculature overlying the L4/5 segment and thus warrants future investigation. The theory that the parameters of the SMT stimulus are important in determining the magnitude of the SNS activity is, in part, supported by the findings of Chiu and Wright (1996) who found that the oscillatory aspect of a spinal treatment technique produced the greatest physiological and SNS change. Indeed, the non-oscillatory

497

SNAG performed in the study by Moulson and Watson (2006) failed to have a significant effect on SC during the treatment period, lending further support to this theory. If this were not the case, it would be reasonable to expect that placebo and mobilisation results should not be significantly different. The results of this and other studies (Petersen et al., 1993; Vicenzino et al., 1994, 1998b; Chiu and Wright, 1996; Sterling et al., 2001) suggest that SMT produces specific physiological effects which other authors have also linked to a specific hypoalgesic effect (Vicenzino et al. 1995, 1998b; Wright and Vicenzino, 1995; Simon et al., 1997) with activation of supra-spinal pathways via the PAG constituting a potential explanation. The PAG is not a discrete unifunctional structure but consists of highly specialised regions and sub-regions each serving different functions. Animal studies of dPAG regions reveal that they are modulated in medullary control nuclei with either unilateral or bilateral projections (Mouton et al., 1997). Consequently, an alternative supra-spinal explanation of the side-specific effects observed in the current study is that SNS responses elicited from the dPAG may be more somatotopically specific than general in nature with sidespecific effects being reported in previous studies (Simon et al., 1997; Sterling et al., 2001) although further development of this proposed effect is required. The use of an independent, matched-group design in the present study is, to the authors’ knowledge, unique, despite its methodological advantages over repeated measures designs. This disparity in methodological design resulted in the making of direct comparisons of SNS PC, reported in other studies, more complex. Future studies should consider this factor within the design phase. The concept of the nervous system as a dynamic continuum, responding spinally and supra-spinally to both mechanical and physiological stimuli is not a new concept, and only recently in physiotherapy research has attention been turned to quantification of a proposed neurophysiological concept through the measurement of SNS response. The findings of this study suggest that neurophysiological and anatomical inter-relationships in the lumbar region do exist and can be credibly modulated with SMT. With the instigation of further research that develops the themes of this study, physiotherapists may begin to integrate this developing knowledge-base into their clinical reasoning processes for the management of patients presenting with symptoms originating from the lumbar spine. References Assendelft WJJ, Koes BW, Knipschild PG, et al. The relation between methodological quality and conclusions in reviews of spinal manipulation. Journal of the American Medical Association 1995;274:1942–8.

ARTICLE IN PRESS 498

J. Perry, A. Green / Manual Therapy 13 (2008) 492–499

Blackwood B, Lavery G. The crossover study design and its clinical application. Nurse Researcher 1998;5:5–14. Bland JM, Altman DG. Matching. British Medical Journal 1994;309:1128. Borkovec TD, Nau SD. Credibility of analogue therapy rationales. Journal of Behavioural Therapy and Experimental Psychiatry 1972;3:257–60. Chester R, Watson MJ. A newly developed spinal simulator. Manual Therapy 2000;5(4):234–42. Chiu TW, Wright A. To compare the effects of different rates of application of a cervical mobilisation technique on sympathetic outflow to the upper limb in normal subjects. Manual Therapy 1996;1:198–203. Clinical Standards Advisory Group. In: Rosen M, editor. Back pain: report of a CSAG committee on back pain. London: HMSO; 1994. CSP. Clinical guidelines for the physiotherapy management of persistent low back pain. Part 2. Manual therapy. London: Chartered Society of Physiotherapy; 2006. Downey BJ, Taylor NF, Niere KR. Manipulative physiotherapists can reliably palpate nominated lumbar spinal levels. Manual Therapy 1999;4:151–6. Ellestad SM, Nagle RV, Boesler DR, et al. Electromyographic and skin resistance responses to osteopathic manipulative treatment for low back pain. Journal of the American Osteopathic Association 1988;88:991–7. Foster NE, Thompson KA, Baxter GD, et al. Management of nonspecific low back pain by physiotherapists in Britain and Ireland. Spine 1999;24:1332–42. Gebber GL, Zhong S, Lewis C, Barman SM. Differential patterns of spinal sympathetic outflow involving a 10-Hz rhythm. Journal of Neurophysiology 1999;82:841–54 [journal article]. Hardy GL, Napier JK. Inter and intratherapist reliability of passive accessory movement technique. New Zealand Journal of Physiotherapy 1991;12:22–4. Ja¨nig W, Ha¨bler H-J. Neurophysiological analysis of target-related sympathetic pathways—from animal to human: similarities and differences. Acta Physiologica Scandinavica 2003;177:255–74. Katavich L. Differential effects of spinal manipulative therapy on acute and chronic muscle spasm: a proposal for mechanisms and efficacy. Manual Therapy 1998;3:132–9. Kenney MJ, Barman SM, Gebber GL, Zhong S. Differential relationships among discharges of postganglionic sympathetic nerves. American Journal of Physiology 1991;29:R1159–67. Li LC, Bombardier C. Physical therapy management of low back pain: an exploratory survey of therapist approaches. Physical Therapy 2001;81:1018–28. Lovick TA. Interactions between descending pathways from the dorsal and ventrolateral periaqueductal gray matter in the rat. In: Depaulis A, Bandiler R, editors. The midbrain periaqueducatal gray matter. New York: Plenum Press; 1991. p. 101–20. Magoun HW. Discussion of impulse mechanisms. In: Korr IM, editor. Neurobiologic mechanisms in manipulative therapy. New York: Plenum; 1978. p. 269–90. Maniadakis N, Gray A. The economic burden of back-pain in the UK. Pain 2000;84:95–103. McGuiness JM, Vicenzino B, Wright A. Influence of a cervical mobilisation technique on respiratory and cardiovascular function. Manual Therapy 1997;2:216–20. Melzack R, Wall PD. The challenge of pain. 2nd ed. Harmondsworth: Penguin; 1996. Moulson KA, Watson T. A preliminary investigation into the relationship between cervical snags and sympathetic nervous system activity in the upper limb of an asymptomatic population. Manual Therapy 2006;11:214–24. Mouton L, VandeHorst V, Holstege G. Large segmental differences in spinal projections to the periaqueductal gray in the cat. Neuroscience 1997;238:1–4 [Letters].

Petersen N, Vicenzino B, Wright A. The effects of a cervical mobilisation technique on sympathetic outflow to the upper limb in normal subjects. Physiotherapy Theory and Practice 1993;9:149–56. Pickar J. Responses of mechanosensitive afferents to manipulation of the lumbar facet in the cat. Spine 1995;20:2379–85. Pocock SJ. Clinical trials: a practical approach. Chichester: Wiley Medical Publication; 1991. p. 124–9. Poitras S, Blais R, Swaine B, Rossigno M. Management of workrelated low back pain: a population-based survey of physical therapists. Physical Therapy 2005;85:1168–81. Portney LG, Watkins MP. Foundations of clinical research: applications to practice. 2nd ed. Englewood Cliffs, NJ: Prentice-Hall Publishers; 2000. Rowland M, Tozer T. Clinical pharmokinetics: concepts and applications. Philadelphia: Lea and Febiger; 1989. Royal College of General Practitioners, et al. Clinical guidelines for the management of acute low back pain: clinical guidelines and evidence review. London: Royal College of General Practitioners; 1998, and 1999. Available at: /http://www.rcgp.org.ukS. Schulz KF, Chalmers I, Hayes RJ, et al. Empirical evidence of bias: dimensions of methodological quality associated with estimates of treatment effects in controlled trials. Journal of the American Medical Association 1995;273:408–12. Sibbald B, Roberts C. Crossover trials. British Medical Journal 1998;316:1719. Sim J, Wright C. Research in health care: concepts, designs and methods. Cheltenham, UK: Nelson Thornes; 2002. Simon R, Vicenzino A, Wright A. The influence of an antero-posterior accessory glide of the glenohumeral joint on measures of peripheral sympathetic nervous system function in the upper limb. Manual Therapy 1997;2:18–23. Slater H. Sympathetic nervous system and pain: a reappraisal. In: Grant R, editor, Physical therapy of the cervical and thoracic spine. 3rd ed.; 2002. p. 295, 319 [chapter 15]. Slater H, Vicenzino B, Wright A. ‘Sympathetic Slump’: the effects of a novel manual therapy technique on peripheral sympathetic nervous system function. Journal of Manual and Manipulative Therapy 1994;2:156–62. Snodgrass SJ, Rivett DA, Robertson VJ. Manual forces applied during posterior-to-anterior spinal mobilization: a review of the evidence. Journal of Manipulative and Physiological Therapeutics 2006;29:316–29. Solly SL. Cervical postero-anterior mobilisation: a brief review of evidence of physiological and pain relieving effects. Physical Therapy Reviews 2004;9:183–7. Sterling M, Jull G, Wright A. Cervical mobilisation: concurrent effects on pain, sympathetic nervous system activity and motor activity. Manual Therapy 2001;6:72–81. Svedmyr N. The placebo effect. Scandinavian Journal of Rehabilitative Medicine 1979;11:169. Uematsu S, Jankel WR, Edwin DH, et al. Quantification of thermal symmetry. Part 1: Normal values and reproducibility. Journal of Neurosurgery 1988;69:552–5. UKBEAM Trial Team. United Kingdom back pain exercise and manipulation (UK BEAM) randomised trial: effectiveness of physical treatments for back pain in primary care. British Medical Journal 2004;329:1377–81. Venables PH, Christie MJ. Mechanisms, instrumentation, recording techniques and quantifications of responses. In: Prokasy WF, Raskin DC, editors. Electrodermal activity in psychological research. New York: Academic Press; 1973. p. 41–73. Vicenzino B, Collins D, Wright A. Sudomotor changes induced by neural mobilisation techniques in asymptomatic subjects. Journal of Manual and Manipulative Therapy 1994;2:66–74. Vicenzino B, Gutschlag F, Collins D, et al. An investigation of the effects of spinal manual therapy on forequarter pressure and

ARTICLE IN PRESS J. Perry, A. Green / Manual Therapy 13 (2008) 492–499 thermal pain thresholds and sympathetic nervous system activity in asymptomatic subjects: a preliminary report. In: Shacklock M, editor. Moving in on pain. Conference proceedings, Adelaide, Australia. Sydney: Butterworth-Heinemann; 1995. p. 185–93. Vicenzino B, Collins D, Wright A. The initial effects of a cervical spine manipulative physiotherapy treatment on the pain and dysfunction of lateral epicondylalgia. Pain 1996;68:69–74. Vicenzino B, Cartwright T, Collins D, Wright A. Cardiovascular and respiratory changes produced by the lateral glide mobilisation of the cervical spine. Manual Therapy 1998a;3(2):67–71. Vicenzino B, Collins D, Benson H, et al. An investigation of the interrelationship between manipulative therapy-induced hypoalgesia and sympathoexcitation. Journal of Manipulative and Physiological Therapeutics 1998b;21:448–53. Vincent C, Lewith G. Placebo controls for acupuncture studies. Journal of the Royal Society of Medicine 1995;88:199–202. Wright A. Hypoalgesia post-manipulation therapy: a review of a potential neurophysiological mechanism. Manual Therapy 1995;1:11–6.

499

Wright A. Recent concepts in the neurophysiology of pain. Manual Therapy 1999;4:196–202. Wright A, Vicenzino B. Cervical mobilisation techniques, sympathetic nervous system effects and their relationship to analgesia. In: Shacklock M, editor. Moving in on pain. Conference proceedings, Adelaide, Australia. Sydney: Butterworth-Heinemann; 1995. p. 164–73. Wyke B, Polacek P. Articular neurology: the present position. Journal of Joint and Bone Surgery 1975;57B:401. Yezierski R. Somatosensory input to the periacqueductal gray: a spinal relay to a descending control centre. In: Depaulis A, Bandler R, editors. The midbrain periaqueductal grey matter. New York: Plenum Press; 1991. p. 365–86. Zusman M. Spinal manipulative therapy: review of some proposed mechanisms and a new hypothesis. Australian Journal of Physiotherapy 1986;32:89–99. Zusman M. Mechanism of musculoskeletal physiotherapy. Physical Therapy Reviews 2004;9:39–49.

ARTICLE IN PRESS

Manual Therapy 13 (2008) 500–506 www.elsevier.com/math

Original article

An unstable support surface does not increase scapulothoracic stabilizing muscle activity during push up and push up plus exercises Gregory J. Lehman, Danielle Gilas, Ushma Patel Department of Graduate Studies, Canadian Memorial Chiropractic College, 6100 Leslie Avenue, Toronto, Ont., Canada M2H 3J1 Received 14 September 2006; received in revised form 12 April 2007; accepted 23 May 2007

Abstract Background: The aim of the current study is to determine if performing push up exercise variations on an unstable surface (Swiss ball) influences EMG amplitude of the scapulothoracic muscles when compared with a stable surface (Bench). Methods: Ten males were recruited from a convenience sample of college students. Surface electromyograms were recorded from the upper trapezius, lower trapezius, serratus anterior and biceps brachii while performing push up exercises with the feet or hands placed on a bench and separately on a Swiss ball. A push up plus exercise was also evaluated with hands on the different support surfaces. Results: There was no statistically significant (po0.05) difference in mean EMG amplitude on a Swiss ball when compared with the same exercise performed on a bench. Significant differences in muscle activity were seen in the upper trapezius and serratus anterior as a result of changes in foot position relative to hand position irrespective of surface stability. Intepretation: The unstable surface used in this study is not a sufficient condition to generate an increase in muscle activity in select scapulothoracic and glenohumeral muscles during push up exercise variations. Elevating the feet above the hands appeared to have a greater influence on shoulder stabilizing musculature amplitude than the addition of a Swiss ball. r 2007 Elsevier Ltd. All rights reserved. Keywords: EMG; Rehabilitation; Push ups; Exercise; Instability; Swiss ball

1. Introduction An unstable surface is often used during rehabilitation exercises in an attempt to increase muscle activation in stabilizing muscles as well as increasing the proprioceptive balance demands on a patient. Numerous authors have shown increases in muscle activity for specific trunk and leg muscles when an unstable surface is incorporated into squatting movements (Anderson and Behm, 2005), bridging exercises (Marshall and Murphy, 2005; Lehman et al., 2005a, b) and during traditional upper body strength exercises (Behm et al., 2005; Lehman et al., 2005a, b). These same experiments have also shown that not all muscles respond with increases in muscle activity. Recently, Drake et al. Corresponding author.

E-mail address: [email protected] (G.J. Lehman). 1356-689X/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2007.05.016

(2006) found that the average peak trunk muscle activity either decreased or did not change during trunk extension exercises performed on a Swiss ball compared with a mat. To date little work has investigated the influence of a Swiss ball on the muscles of the shoulder during push up exercises. Lehman et al. (2006) documented increases in the EMG of the triceps muscles when the hands were on a Swiss ball during a push up exercise compared with a push up performed on a stable surface; however, no change EMG amplitude between conditions was evident in the pectoralis major muscle. To date no studies have documented the influence of a Swiss ball on scapulothoracic and glenohumeral stabilizing musculature (upper and lower trapezius, serratus anterior, biceps brachii) during push up exercises and common push up exercise variations (push up plus). Considering the popular usage of Swiss balls in rehabilitation and resistance training and the current

ARTICLE IN PRESS G.J. Lehman et al. / Manual Therapy 13 (2008) 500–506

lack of information on the response of the scapulothoracic musculature to these exercises, this study was initiated to address this current deficit in our knowledge regarding neuromuscular function. Determining the influence of the Swiss ball on scapulothoracic muscle activity may aid in advocating certain exercises for the rehabilitation and training of the scapulothoracic musculature. The objective of this study is to determine if the addition of a Swiss ball to push up exercises influence mean scapulothoracic muscle activation levels.

2. Methods 2.1. Patient characteristics Ten male participants (height (cm) 174.7712.9, mass (kg) 83.3710.9, age 26.371.1) were recruited from the undergraduate population of the research institution. Participants were excluded if they had recent shoulder or neck pain/injury (within 1 year). Participants were required to read and sign an information and informed consent form prior to the study approved by the institution’s Research Ethics Board. 2.2. Study protocol The surface myoelectric activity of the upper trapezius, lower trapezius, serratus anterior and biceps brachii were recorded during a series of different variations of the classic push up exercise. 2.3. Data collection hardware characteristics Disposable bipolar Ag–AgCI disk surface electrodes (Bortec Biomedical, Calgary AB, Canada) with a diameter of 1 cm and interelectrode distance of 2 cm were adhered over the muscle groups parallel to their fiber orientation on the muscle belly on the right-hand side of the participant. Before the application of the electrodes, the skin was shaved with a disposable razor and abraded with a cotton swab and alcohol. The myoelectric activity of the muscles was collected with customized software (Delsys EMGWorks, Boston, MA, USA). Raw EMG was amplified between 1000 and 10,000 times depending on the subject. The amplifier had a CMRR of 10,000:1 (Bortec EMG, Calgary AB, Canada). Raw EMG was band pass filtered (10 and 1000 Hz) and A/D converted at 2048 Hz using a National Instruments data acquisition system.

501

up to a maximum contraction for each muscle against the manual resistance provided by the experimenter. Participants and experimenters were permitted to practice the MVCs. The MVC procedures and electrode positioning were based on previous work investigating shoulder muscle exercise activity and normalization procedures (Decker et al., 1999; Ekstrom et al., 2005). Electrode positioning and MVC protocol for each muscle were as follows: Upper trapezius—the electrode was placed on the muscle belly midway between the C7 spinous process and the trapezius’ insertion on the right acromioclavicular joint. The MVC was performed with the right arm abducted to 901 and the head laterally flexed 10–151 towards the abducted arm. The subject was then instructed to maximally attempt to approximate the abducted arm and head toward each other against the examiner’s matching immovable resistance. Lower trapezius—the electrode was placed on the muscle belly 1.5 cm lateral and obliquely (inferior electrode more medial) to the T6 spinous process on the subject’s right side. With the right arm fully flexed 1801 (inline with the lower trapezius), the subject attempted to continue flexing further against examiner’s resistance applied to the upper arm. Serratus anterior— the electrode was placed on the muscle belly in the midaxillary line of the right side over the fifth rib. The shoulder was flexed to 1251 as resistance was applied above the elbow and at the inferior angle of the scapula in an attempt to derotate the scapula. Biceps brachii— the electrode was placed on the right sided muscle belly midway between the cubital fossa and the neck of the humerus. With the hand supinated and the arm flexed to 901 at the elbow, the subject was instructed to maximally flex the arm against the examiner’s resistance. 2.5. Exercise protocol Following the maximal voluntary contractions, the participants were required to perform the following exercises in an order arbitrarily determined by the experimenters. Participants performed the exercises in an identical manner across exercises. Three repetitions occurred for each exercise at the same tempo. Participants began in the upright position when the EMG collection began. The eccentric (lowering) portion lasted 2 s and the concentric portion lasted 2 s with a slight pause at the bottom of the repetition. An electrical trigger (pressed by the experimenter) was used to mark the beginning of the first descent and the finish of the last repetition.

2.4. Maximum voluntary contractions (MVC) and electrode positioning

2.6. Exercise description

The myoelectric signal for all tasks was expressed as a percentage of the maximal muscle activity found during each muscle’s MVC. This required the subject to ramp

All movements were completed in a standardized position with the hands shoulder width apart with the subject’s middle finger under the acromioclavicular

ARTICLE IN PRESS 502

G.J. Lehman et al. / Manual Therapy 13 (2008) 500–506

joint. The bench height and exercise ball (65 cm uncompressed) height were the same across subjects and similar to each other. A minimum of 3 min of rest occurred between exercises to prevent the influence of fatigue on myoelectric amplitude changes. The following exercises were performed: (1) Push up with feet on an exercise bench and hands on the floor. (2) Push up with feet on exercise ball and hands on the floor. (3) Push up with hands on exercise bench and feet on the floor. (4) Push up with hands on exercise ball and feet on the floor. (5) Push up plus with hands on exercise bench (Figs. 1 and 2): starting in the push up position the participant rolls the shoulders forward (scapular protraction) and then lowers their body while allowing the shoulder blades to approximate (scapular retraction). (6) Push up plus with hands on exercise ball. Same movement as #5.

3. EMG analysis Both MVC and exercise task myoelectric data were processed in an identical manner. Using EMG analysis software (EMGWorks, Delsys, Boston, MA), the myoelectrical signal had the bias removed (removed mean from the raw signal), then, a moving average technique (mean absolute value—window of 300 data points and an overlap of 150 data points) was used to smooth the data thus providing a linear envelope of EMG activity. Using the electrical markings left by the foot switch trigger at the start and end of the movement the mean activity over the course of three repetitions of smoothed data was calculated. This mean activity was then expressed as a percentage of the peak activity found during the MVC for the corresponding muscle. 3.1. Statistical analysis A repeated-measures ANOVA with post hoc Tukey test was used to identify differences across the exercises for each muscle measured. In all tests, the 95% (p ¼ 0.05) level of confidence was used for the rejection

Fig. 1. Bottom portion (medial scapular border approximated) of the push up plus exercise performed on a bench.

ARTICLE IN PRESS G.J. Lehman et al. / Manual Therapy 13 (2008) 500–506

503

Fig. 2. Raised portion (medial scapular border protracted) of the push up plus exercise performed on a bench.

Table 1 Mean (standard deviation) of average muscle activity (expressed as % of MVC) during push up variations Muscle

Upper trapezius Lower trapezius Serratus anterior Biceps

Push up variation Handbench

Handball

Feetbench

Feetball

Plusbench

Plusball

5.2 10.5 24.2 1.77

10.5 9.5 19.7 2.2

8.4 7.01 37.9 1.6

9.7 9.6 37.5 2.5

2.3 5.7 20.7 1.1

3.8 3.9 22.6 1.8

(6.4) (12.2) (14.5) (1.1)

(6.9) (11.9) (11.5) (1.2)

of the null hypothesis. A post hoc power analysis was also performed to determine if an adequate sample size was used assuming a clinically/physiologically significant difference between exercises was 3–5% of MVC.

4. Results Table 1 depicts the mean and standard deviation for the group average activity during each exercise and for each muscle. A post hoc power analysis (using an

(8.2) (5.6) (16.4) (1.3)

(7.7) (9.3) (19.9) (2.3)

(2.4) (5.4) (7.9) (.87)

(3.2) (5.3) (10.5) (2.02)

average standard deviation of differences between pairs to be 3.1 (range: 0.32–6.89)) which assumed a biologically meaningful difference between pairs to be 3–5% of MVC revealed that a sample size of 10 had estimated statistical power at 80% to detect a difference of 2.75% MVC or more. There was no statistically significant difference in muscle activity due to the stability of the surface (bench vs. ball) for all muscles during the paired exercises (e.g., push up with hands on ball paired with push up with hands on bench).

ARTICLE IN PRESS 504

G.J. Lehman et al. / Manual Therapy 13 (2008) 500–506

For unpaired exercises, statistically significant differences were found between exercises differing in terms of hand/feet position and type of push up exercise (push up plus vs. standard push up). The upper trapezius muscle activity was less during the push up plus exercise with the hands on the bench when compared with the push ups with the hands on the ball (po0.01), feet on the bench (po0.05) and feet on the ball (po0.01). The upper trapezius activity during the push up plus with the hands on the ball was significantly less than the activity during the push ups performed with the hands on the ball (po0.05) and the feet on the ball (po0.05). There was no difference in muscle activity across all exercises for the lower trapezius muscle. For the serratus anterior muscle, the push ups with hands on the bench showed less activity then when performed with the feet on the bench (po0.05) or the ball (po0.05). The push ups with the hands on the ball was also significantly less than the push ups with the feet on the bench or ball (po0.001). Push up plus exercises on the ball (po0.01) and on the bench (po0.001) showed less serratus anterior muscle activity compared with push ups with feet on the bench. Push up plus exercises on both the ball and bench showed significantly less activity (po0.01) compared with push up exercises with the feet on the ball. The only statistically significant difference in biceps activity was seen between the push ups with feet on the ball compared with push up plus exercises with hands on the bench (po0.05).

5. Discussion The addition of a Swiss ball to the push up and push up plus exercises did not influence the muscle activity of the selected shoulder stabilizing muscles despite anecdotal claims by therapists and exercise specialists suggesting that a Swiss ball results in the greater recruitment of the ‘‘stabilizing’’ muscles around joints. While differences in muscle activity were seen across exercises there was no difference in any muscle when the only difference between the exercises was the replacement of an exercise bench with a Swiss ball (i.e., the paired exercises). However, as can be seen in Table 1, there is a trend for the upper trapezius muscle to increase its muscle amplitude when performing a push up on a Swiss ball compared with a bench. The lack of statistical significance is most likely due to the large between subject variability and the variability in the response to the unstable surface. This variability has been noted in previous work (Lehman et al., 2005a, b) and suggests that an unstable surface may not affect individuals equally. Other factors related to participant positioning (raising the height of the feet) appeared

more important in influencing activation levels. Modifying the Swiss ball push up itself (e.g., raising the feet off the ground onto a bench) may have resulted in changes in scapulothoracic muscle activity due to the possible increase in weight over the hands and the unstable surface. This influence of body height was notable in the serratus anterior muscle where greater muscle activity occurred during push ups with the feet elevated (bench or ball) compared with the push up plus which is often prescribed to recruit the serratus anterior (Decker et al., 1999; Ludewig et al., 2004). The lack of an increase in muscle activity with the addition of a Swiss ball in select muscles during the paired exercises is consistent with several previous studies (Marshall and Murphy, 2005; Lehman et al., 2005a, b; Drake et al., 2006; Lehman et al., 2006; Marshall and Murphy, 2006) and should provide caution to the therapist when assuming that the addition of a unstable surface necessitates an increase in muscle activity in all muscles. Most recently, de Oliveira et al. (2007) documented no differences in the muscle activity in the serratus anterior and biceps muscles when performing push ups on a medicine ball compared with a stable base of support. While this lack of a difference in myoelectric activity between stable and unstable conditions exists for some muscles studied it is not a generalizable phenomenon to all muscles involved in the same exercise. Instances of increases in myolectric activity in primary movers and stabilizing musculature have also been documented during upper body exercises. While de Oliveira et al. (2007) found no change in some muscles studied, the authors also found an increased muscle activity in the anterior deltoid during upper body exercises performed on a medicine ball compared with a stable base of support. The lack of difference in muscle activity between stability conditions in the current study may be due to a number of factors. The type of unstable support surface studied (Swiss ball) combined with the exercise studied may not create a sufficiently unstable condition to require increases in muscular activation. Other unstable surfaces (medicine balls, wobble boards, rocker boards, inflatable discs) may result in increases in scapulothoracic muscle activity; therefore, these findings cannot be generalized to all unstable surfaces nor all upper limb activities. More dynamic activities than a push up performed on unstable surfaces may result in increases in muscle activity not seen on stable ground. Even elevating one leg off the ground, hence only three points of surface contact may be a sufficient condition to increase muscular activation. It should also be emphasized that a lack of an increase in muscle activation does not eliminate the possible therapeutic benefits of performing push ups and push up plus exercises on the Swiss ball. Performing push ups on a Swiss ball may have additional therapeutic benefits

ARTICLE IN PRESS G.J. Lehman et al. / Manual Therapy 13 (2008) 500–506

related to proprioception and balance challenges not studied in this paper. Other well-researched exercises exist for training the scapulothoracic musculature, specifically the serratus anterior and lower trapezius, which may be important for the rehabilitation of injured shoulders (Decker et al., 1999; Kibler and McMullen, 2003). The exercises and results of the current study provide little argument in terms of replacing these exercises in rehabilitation programs in terms of muscular activation. Ekstrom and Donatelli (2003) showed that the lower trapezius exceeded maximum amplitudes in excess of 50% MVC during exercises with the arm raised overhead in line with the lower trapezius (while lying prone), prone lying shoulder external rotation at 901 flexion and full range scaption exercises (shoulder horizontal extension with external rotation). In the serratus anterior, Ekstrom and Donatelli (2003) demonstrated maximum muscle activation amplitudes greater than 50% MVC during the diagonal exercise (combined shoulder flexion, adduction and external rotation), scaption, shoulder press and shoulder external rotation at 901 while lying prone. The push up plus exercise on the floor has also been found to be an excellent exercise for recruiting the serratus anterior, while minimizing upper trapezius muscle activity (Ludewig et al., 2004). Hardwick et al. (2006) has recently documented average muscle activation levels in the range of 37–82% MVC during the wall slide and scapular plane elevation exercises when the shoulder was flexed more than 901. One limitation of this study is the lack of myoelectric information on the rotator cuff musculature which may still be influenced by the addition of an unstable surface. Concluding that all shoulder stability muscles are not influenced by surface stability used in this study and the exercises performed cannot be made. Future work investigating rotator cuff musculature and additional support surfaces is planned and should be undertaken to fully document the influence of surface instability on shoulder muscle activity during exercise. Additionally, this study also lacked information on scapular kinematics, which may have helped explain the reason for the lack of change in muscle activity. This null finding may simply be that the Swiss ball (during these specific exercises) does not result in instability of the scapula on the thorax or the humerus within the scapula. The muscles studied are proximal and the instability generated by the Swiss ball may be manifested more in the response of the muscles attempting to stabilize the distal elbow and wrist joints. This study is also limited in that it only evaluated the average activity over the course of a number of repetitions—it did not investigate specific portions of a repetition, which may illustrate changes in muscle activity between stability conditions. For example, the transition between eccentric and concentric contractions may be an area where increased

505

muscle activity occurs and could explain the perceptions amongst exercises who perceive that their muscles are working harder. Last, an attempt was made to ensure similar postural and center of mass positions during the paired exercises when a Swiss ball replaced an exercise bench. There may have been differences in joint kinetics between paired exercises suggesting that other factors besides surface stability influenced muscle activity. However, it was an aim of the study to perform exercises in a manner similar to a clinical setting. Further controls on posture may have hampered this clinical applicability.

6. Conclusion The addition of a Swiss ball to push up variations does not result in an increase in select scapulothoracic or glenohumeral muscles.

References Anderson K, Behm DG. Trunk muscle activity increases with unstable squat movements. Canadian Journal of Applied Physiology 2005(30):33–45. Behm D, Leonard A, Young W, Bonsey W, MacKinnon S. Trunk muscle electromyographic activity with unstable and unilateral exercises. Journal of Strength and Conditioning Research 2005(19): 193–201. de Oliveira AS, de Morais Carvalho M, de Brum DP. Activation of the shoulder and arm muscles during axial load exercises on a stable base of support and on a medicine ball. Journal of Electromyographic and Kinesiology 2007; Jan 9; [Epub ahead of print]. Decker MJ, Hintermeister RA, Faber KJ, Hawkins RJ. Serratus anterior muscle activity during selected rehabilitation exercises. American Journal of Sports Medicine 1999;27(6):784–91. Drake JDM, Fischer SL, Brown SHM, Callaghan JP. Do exercise balls provide a training advantage for trunk extensor exercises? A biomechanical evaluation. Journal of Manipulative and Physiological Therapeutics 2006(29):354–62. Ekstrom RA, Donatelli RA. Soderberg GLSurface electromyographic analysis of exercises for the trapezius and serratus anterior muscles. Journal of Orthopaedic and Sports Physical Therapy 2003;33(5): 247–58. Ekstrom RA, Soderberg GL, Donatelli RA. Normalization procedures using maximum voluntary isometric contractions for the serratus anterior and trapezius muscles during surface EMG analysis. Journal of Electromyography and Kinesiology 2005;15(4):418–28 [Epub 25 December 2004]. Hardwick DH, Beebe JA, McDonnell MK, Lang CE. A comparison of serratus anterior muscle activation during a wall slide exercise and other traditional exercises. Journal of Orthopaedic and Sports Physical Therapy 2006;36(12):903–10. Kibler WB, McMullen J. Scapular dyskinesis and its relation to shoulder pain. Journal of the American Academic Orthopaedic Surgeon 2003;11(2):142–51 [Review]. Lehman GJ, Gordon T, Langley J, Pemrose P, Tregaskis S. Replacing a Swiss ball for an exercise bench causes variable changes in trunk muscle activity during upper limb strength exercises. Dynamic Medicine 2005a;3(4):6.

ARTICLE IN PRESS 506

G.J. Lehman et al. / Manual Therapy 13 (2008) 500–506

Lehman GJ, Hoda W, Oliver S. Trunk muscle activity during bridging exercises on and off a Swiss ball. Chiropractic and Osteopathy 2005b;13:14. Lehman GJ, MacMillan B, MacIntyre I, Chivers M, Fluter M. Shoulder muscle EMG activity during push up variations on and off a Swiss ball. Dynamic Medicine 2006;5:7. Ludewig PM, Hoff MS, Osowski EE, Meschke SA, Rundquist PJ. Relative balance of serratus anterior and upper trapezius

muscle activity during push-up exercises. American Journal of Sports Medicine 2004;32(2):484–93. Marshall PW, Murphy BA. Core stability exercises on and off a Swiss ball. Archives of Physical Medicine and Rehabilitation 2005(86): 242–9. Marshall P, Murphy B. Changes in muscle activity and perceived exertion during exercises performed on a Swiss ball. Applied Physiology, Nutrition and Metabolism 2006;31(4):376–83.

ARTICLE IN PRESS

Manual Therapy 13 (2008) 507–512 www.elsevier.com/math

Original article

Training the cervical muscles with prescribed motor tasks does not change muscle activation during a functional activity Deborah Fallaa,b,, Gwendolen Julla, Paul Hodgesa a

Division of Physiotherapy, The University of Queensland, Brisbane, Australia Department of Health Science and Technology, Center for Sensory-Motor Interaction (SMI), Aalborg University, Denmark

b

Received 12 March 2007; received in revised form 31 May 2007; accepted 3 July 2007

Abstract Both low-load and high-load training of the cervical muscles have been shown to reduce neck pain and change parameters of muscle function directly related to the exercise performed. The purpose of this study was to investigate whether either training regime changes muscle activation during a functional task which is known to be affected in people with neck pain and is not directly related to either exercise protocol. Fifty-eight female patients with chronic neck pain were randomised into one of two 6-week exercise intervention groups: an endurance-strength training regime for the cervical flexor muscles or low-load training of the craniocervical flexor muscles. The primary outcome was a change in electromyographic (EMG) amplitude of the sternocleidomastoid (SCM) muscle during a functional, repetitive upper limb task. At the 7th week follow-up assessment both intervention groups demonstrated a reduction in their average intensity of pain (Po0.05). However, neither training group demonstrated a change in SCM EMG amplitude during the functional task (P40.05). The results demonstrate that training the cervical muscles with a prescribed motor task may not automatically result in improved muscle activation during a functional activity, despite a reduction in neck pain. r 2007 Elsevier Ltd. All rights reserved. Keywords: Neck pain; Electromyography; Exercise; Sternocleidomastoid

1. Introduction Therapeutic exercise has demonstrated efficacy in reducing pain and perceived disability in people with neck pain disorders (Bronfort et al., 2001; Jull et al., 2002; Ylinen et al., 2003; Chiu et al., 2005). In addition to a change in symptoms, therapeutic exercise has been shown to improve cervical muscle function (Bronfort et al., 2001; Ylinen et al., 2003; Chiu et al., 2005; Jull et al., 2005; Falla et al., 2006). In recent studies (Jull et al., 2005; Falla et al., 2006, 2007), we compared the effect of low-load cranioCorresponding author. Department of Health Science and Technology, Center for Sensory-Motor Interaction (SMI), Aalborg University, Fredrik Bajers Vej 7D-3, DK-9220 Aalborg, Denmark. Tel.: +45 96 35 74 59; fax: +45 98 15 40 08. E-mail address: [email protected] (D. Falla).

1356-689X/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2007.07.001

cervical flexion training and higher load endurance/ strength training of the cervical flexor muscles on various aspects of cervical muscle function in people with chronic neck pain. Following 6 weeks of training the cranio-cervical flexor muscles, patients with neck pain displayed a significant increase in deep cervical flexor electromyographic (EMG) amplitude during a test of cranio-cervical flexion (Jull et al., 2005). This was associated with decreased sternocleidomastoid (SCM) EMG amplitude and increased cranio-cervical flexion range of motion (Jull et al., 2005), indicating improved performance on the test (Jull et al., 1999). In contrast, patients who participated in 6 weeks of cervical flexor strength training did not demonstrate a similar improvement (Jull et al., 2005). This difference was identified despite a comparable reduction in pain and perceived disability between the two exercise groups. In a further study (Falla et al., 2006) it was identified that

ARTICLE IN PRESS D. Falla et al. / Manual Therapy 13 (2008) 507–512

508

6 weeks of cervical flexor endurance-strength training results in reduced myoelectric manifestations of SCM muscle fatigue and increased cervical flexion strength in people with chronic neck pain. Conversely, no effect was identified for the patients who received the lower load cranio-cervical flexion training. Thus, improvement in muscle function following training was task specific and directly related to the exercise protocol (Jull et al., 2005; Falla et al., 2006). It is unknown whether either training protocol for the cervical muscles would influence muscle activation during functional activities which are unrelated to the specific exercise protocol. The purpose of this study was to investigate whether specific training with low-load cranio-cervical flexion exercise or higher load endurance-strength exercise would change muscle activation during a functional task which is known to be altered in people with neck pain and is not directly related to the exercise protocol. This study forms part of a series of experiments to investigate the mechanisms of efficacy of cervical muscle retraining. The effects of both exercise regimes on pain and disability have been reported in our previous work (Falla et al., 2006). 2. Methods 2.1. Subjects Fifty-eight female participants with a history of chronic neck pain of greater than 3-month duration participated in this study (Table 1). Subjects were recruited by advertisements in the local press. Subjects were excluded if they scored 415 (out of a possible 50) on the Neck Disability Index (NDI) (Vernon and Mior, 1991) to minimise the limitations that higher levels of pain and disability might have on performance of the endurance-strength exercise regime. They were excluded Table 1 Baseline characteristics for patients with chronic neck pain randomised into a cranio-cervical flexion exercise intervention or endurancestrength exercise intervention

Age Length of neck pain history (years) Onset (insidious, whiplash) % whiplash Average neck pain intensity (0–10) Neck Disability Index (0–50)

Cranio-cervical flexion exercise intervention (n ¼ 28)

Endurance-strength exercise intervention (n ¼ 29)

37.7710.1 7.676.0

38.1710.7 8.377.0

17.8

13.8

3.572.0

4.772.0

9.873.3

10.173.0

Mean and standard deviation are shown.

if they had undergone cervical spine surgery, presented with any neurological signs or had participated in a neck exercise programme in the past 12 months. An examination of the cervical spine was performed to confirm the presence of palpable cervical joint tenderness (Jull et al., 1988). The subjects included in this study also formed part of another study (Falla et al., 2006). Ethical approval for the study was granted by the Institutional Medical Research Ethics Committee and all procedures were conducted according to the Declaration of Helsinki. 2.2. Exercise interventions Patients with chronic neck pain were randomised by an independent body using a computer-generated sequence of numbers into two exercise groups: endurance-strength training of the cervical flexor muscles and cranio-cervical flexion training. Exercise regimes were conducted over a 6-week period and patients in each group received personal instruction and supervision by an experienced physiotherapist once per week for the duration of the trial. None of the exercise sessions were longer than 30 min. Subjects were asked not to receive any other form of specific intervention for their neck; however, medication was not withheld from any participant. All subjects were supplied with an exercise diary and requested to practice their respective regime twice per day for the duration of the trial. The exercise occupied a period of no longer than 10–20 min per day. The exercises were to be performed without any provocation of neck pain. 2.3. Cervical flexor endurance-strength training The endurance-strength training regime consisted of a progressive resistance exercise programme for the neck flexors. Exercises were performed in supine, with the head supported in a comfortable resting position. Patients were instructed to lift their head so that cervical flexion was performed maintaining a neutral upper cervical spine. Patients were to slowly move the head and neck through full range of cervical flexion motion as possible without causing discomfort or reproduction of their symptoms. The exercise regime involved two stages. The first stage was of 2-week duration and the second was of 4-week duration as recommended for initiation of a weight programme in previously untrained individuals (McArdle et al., 1996). In stage 1, the subjects performed 12–15 repetitions with a weight that they could lift 12 times on the first training session (12 repetitions maximum) and progressed to 15 repetitions and maintained this level for the remainder of the 2-week period. In stage 2, the subjects performed three sets of 15 repetitions of the initial 12 repetitions maximum load once per day. One-minute rest intervals

ARTICLE IN PRESS D. Falla et al. / Manual Therapy 13 (2008) 507–512

were provided between sets. If repetitions were easily achieved, weighted sandbags were applied to the patient’s forehead in 0.5 kg increments as required. If the patient was unable to perform repetitions of the head lift manoeuvre, then the load on the neck flexors was reduced by decreasing the vertical component of the head weight vector. That is, the upper body (trunk and neck) was inclined from the horizontal such that the patient could perform the required repetitions of the movement. 2.4. Cranio-cervical flexion training The low-load training of the cranio-cervical flexor muscles followed the protocol described by Jull et al. (2004a). The patient was instructed to perform and hold progressively inner range positions of cranio-cervical flexion while attempting to maintain the superficial flexor muscles relaxed. The subjects were first taught to perform slow and controlled cranio-cervical flexion. They then trained to hold progressively increasing ranges of cranio-cervical flexion using feedback from an air-filled pressure sensor (StabilizerTM, Chattanooga Group, Inc., Chattanooga, TN) placed behind the neck. 2.5. Electromyography Measures were obtained at baseline and in the week immediately after treatment (week 7). The investigator was blinded to the subject group for the outcome assessments. Recordings of EMG from the sternal head of SCM were made bilaterally with Ag/AgCl surface electrodes (20 mm disc electrode, inter-electrode distance: 20 mm; Grass Telefactor, Astro-Med, Inc.) following skin preparation and using guidelines for electrode placement (Falla et al., 2002). A ground electrode was placed on the upper thoracic spine. EMG data were amplified (gain ¼ 1000), band pass filtered between 20 Hz and 1 kHz and sampled at 2 kHz. Data were sampled with Spike software (Cambridge Electronic Design, Cambridge, UK) and converted into a format suitable for signal processing with Matlab software (The MathWorks, Inc., Natick, MA, USA). 2.6. Experimental procedure Prior to commencement of the experimental trials, EMG data were collected for 10 s during a standardised manoeuvre for normalisation of EMG amplitude. The task involved a combined movement of cranio-cervical and cervical flexion to lift the head so that it just cleared the bed and was held isometrically for 10 s (Falla et al., 2004a). The reference contraction was repeated three times with a 30 s rest period between each repetition.

509

For the experimental task, the subject sat at a desk in a height adjustable office chair with their feet flat on the ground. The subject was asked to dot pencil marks in three circles in an anti-clockwise direction with their dominant hand. This task was performed to the beat of a metronome set at 88 beats/min. The subject’s nondominant hand rested motionless on the table. The 70 mm diameter circles were positioned in an equilateral triangle with a distance of 23 cm between each centre. The task was performed for 2.5 min. Myoelectric signals were recorded for 5 s epochs prior to baseline at 10, 60 and 120 s during the task and 10 s after completion of the task. In previous work it has been identified that patients with neck pain demonstrate a bilateral increase of SCM EMG amplitude (Falla et al., 2004a) compared to healthy controls during this task. Furthermore, neck pain patients demonstrate a decreased ability to relax their neck muscles after completion of the task (Falla et al., 2004a). 2.7. Data management and statistical analyses To obtain a measure of EMG signal amplitude, maximum root mean square (RMS) was calculated for 1 s using Matlab software (The MathWorks, Inc. Natick, MA, USA) after subtracting baseline. For normalisation, EMG amplitude at each stage of the task was expressed as a percentage of the maximum 1 s RMS values obtained during the reference voluntary contraction after subtracting baseline. The independent variables for this study were the subject groups (between subjects factor), and withinsubject factors, time (four measurements), and side (two measurements). The dependent variables were normalised RMS values. A repeated-measures general linear model was used to identify whether normalised RMS values were different between the two subject groups across time for the SCM muscle on each side. A value of Po0.05 was considered statistically significant. 3. Results Subject descriptives are presented in Table 1. The data for one subject in the cranio-cervical flexion training group were discarded due to errors in recordings. Baseline characteristics of subjects’ pain and disability levels were not different between the two intervention groups (all P40.05). All participants received the full six treatments. No patients reported any adverse events. Both intervention groups demonstrated a reduction in average intensity of pain (cranio-cervical flexion training, 0.972.4; endurance-strength training, 1.172.8; P4 0.05) and perceived disability (NDI: cranio-cervical flexion training, 3.774.7; endurance-strength training,

ARTICLE IN PRESS D. Falla et al. / Manual Therapy 13 (2008) 507–512

510

2.874.0; P40.05). There was no difference between groups for change in pain or perceived disability (P40.05). No significant differences were present between groups for pre-intervention SCM RMS values (P40.05; Fig. 1). Pre-intervention values were consistent with values obtained in our previous study of patients with chronic neck pain for the same functional task (Falla et al., 2004a), which showed significant differences to control subjects. Following intervention, no significant change was identified for SCM normalised RMS values for either group (P40.05; Fig. 2). 4. Discussion The results of this study demonstrate that 6 weeks of either low-load cranio-cervical flexion training or higher

load cervical flexor endurance-strength training does not change SCM muscle activation during a functional activity that is not directly related to either exercise protocol. This suggests that improved muscle function obtained with these training approaches (Jull et al., 2002, 2005; Falla et al., 2006) may not automatically transfer to a change in muscle activation during a functional, upper limb activity. This finding was observed despite the evidence that symptoms of neck pain and muscle function (related to the exercise performed) improve after both interventions (Bronfort et al., 2001; Jull et al., 2002, 2005; Ylinen et al., 2003; Falla et al., 2006). This implies that a reduction in pain alone may be insufficient to induce changes in the motor control strategy during functional upper limb tasks, such as the task described in this study.

Cranio-cervical flexion training Left Sternocleidomastoid

Endurance-strength training

Right Sternocleidomastoid

Normalised RMS Values (%)

30 25 20 15 10 5 0 10 s

60 s

120 s

10 s Post

10 s

60 s

Time

120 s

10 s Post

Time

Fig. 1. Pre-intervention normalised root mean square (RMS) values for the left and right sternocleidomastoid muscles for patients with neck pain randomised into either a low-load cranio-cervical flexion training group or an endurance-strength training group. Mean and standard deviation values are given at 10, 60 and 120 s into the performance of a repetitive upper limb task and 10 s after completion of the task (10 s post).

Cranio-cervical flexion training Endurance-strength training Left Sternocleidomastoid

Right Sternocleidomastoid

Change in Normalised RMS Values (%)

15 10 5 0 -5 -10 -15 10 s

60 s

120 s Time

10 s Post

10 s

60 s

120 s

10 s Post

Time

Fig. 2. Pre- to post-intervention change in normalised root mean square (RMS) values for the left and right sternocleidomastoid muscle for patients with neck pain randomised into either a low-load cranio-cervical flexion training group or an endurance-strength training group. Mean and standard deviation values are given at 10, 60 and 120 s into the performance of a repetitive upper limb task and 10 s after completion of the task (10 s post).

ARTICLE IN PRESS D. Falla et al. / Manual Therapy 13 (2008) 507–512

4.1. Effect of therapeutic exercise on muscle activation during functional activities A consistent finding in people with chronic neck pain is a disturbance in the activation of SCM during isometric contractions and dynamic movement of the upper limb. More specifically, increased SCM EMG amplitude has been identified during the clinical test of cranio-cervical flexion (Jull, 2000; Falla et al., 2004c; Jull et al., 2004b), sub-maximal isometric cervical flexion contractions (Barton and Hayes, 1996; Falla et al., 2004b) and the functional repetitive upper limb task described in this study (Falla et al., 2004a). It has been proposed that alterations in SCM muscle activity represent altered motor control strategies to stiffen the neck or provide compensation for reduced activation of the deep cervical muscles (Falla et al., 2004c, d). Subsequently, therapeutic exercises programmes have been designed to address these impairments. Although previous studies have examined the effect of therapeutic exercise on cervical muscle function, to our knowledge, this is the first EMG study that examined the effect of cervical muscle training on muscle activation during a functional activity which is not directly related to the training protocol. We selected to examine the effect of two different exercise regimes, which are commonly used for the rehabilitation of the cervical muscles. Although both low-load cranio-cervical flexion training and higher load training of the cervical flexor muscles have been shown to decrease neck pain (Bronfort et al., 2001; Jull et al., 2002; Ylinen et al., 2003; Chiu et al., 2005), the two approaches target different aspects of muscle performance and have been shown to have distinctly different effects on cervical muscle function. The low-load cranio-cervical flexion training regime aims to improve the activation of the deep flexors of the upper cervical region, the longus capitis and longus colli, whilst minimising activation of the superficial flexors, SCM and anterior scalene muscles, which flex the neck but not the head. In contrast, the endurance-strength protocol promotes activation of all muscles that contribute to a head lift, including the SCM, scalenes, longus capitis and longus colli and hyoid muscles. Thus, the cranio-cervical flexion training regime would appear to be a more ideal strategy to reduce augmented activity of the SCM muscle. Although cranio-cervical flexion training has been shown to reduce activity of the SCM muscle during the clinical test of cranio-cervical flexion (Jull et al., 2005), the results of this study suggest that this does not automatically transfer to a changed motor control strategy during a functional, upper limb task performed in sitting. 4.2. Methodological considerations In this study we examined the effect of therapeutic exercise on a task which has previously demonstrated

511

increased activity of the SCM muscle in people with chronic neck pain (Falla et al., 2004a). Upper limb activity in the more sedentary occupations of the modern day workforce is commonly performed under open chain conditions, for example typing, writing or performing other light activities with the hands. Therefore, the task selected in this study was designed to mimic low-load repetitious upper limb movement. However, further studies examining muscle coordination during other tasks, such as typing, may yield different results. Moreover, only the activation of SCM muscle was monitored. Other muscles such as the longus colli, longus capitis, scalenes and cervical erector spinae may have responded differently during this task following intervention with either exercise protocol. Further studies may be warranted to address these issues, the effect that longer durations of training may have on muscle activity and the whether similar findings would be observed in other neck pain populations, for example people with pain of greater severity or men. Although it can be hypothesised that altered patterns of muscle activation may increase adverse loading on sensitised cervical structures, further research is necessary to explore the consequences of changes in muscle activation on the symptoms of neck pain and recurrence rate. This is necessary if we are to fully appreciate the benefit of attempts to reverse altered cervical muscle activity in people with neck pain. 5. Conclusion This study demonstrates that 6 weeks of specific cervical flexor muscle training, which has been shown to improve parameters of muscle function and reduce the symptom of neck pain, may not automatically transfer to changes in muscle activity during an untrained functional upper limb task. These results suggest that rehabilitation of the cervical muscles should be extended to include training in functional postures and tasks. Acknowledgements This study was funded by a grant (ID 252771) received from the National Health and Medical Research Council of Australia. Deborah Falla is supported by the National Health and Medical Research Council of Australia (ID 351678). References Barton PM, Hayes KC. Neck flexor muscle strength, efficiency, and relaxation times in normal subjects and subjects with unilateral neck pain and headache. Archives of Physical Medicine and Rehabilitation 1996;77(7):680–7.

ARTICLE IN PRESS 512

D. Falla et al. / Manual Therapy 13 (2008) 507–512

Bronfort G, Evans R, Nelson B, Aker PD, Goldsmith CH, Vernon H. A randomized clinical trial of exercise and spinal manipulation for patients with chronic neck pain. Spine 2001;26(7):788–97. Chiu TT, Lam TH, Hedley AJ. A randomized controlled trial on the efficacy of exercise for patients with chronic neck pain. Spine 2005; 30(1):E1–7. Falla D, Dall’Alba P, Rainoldi A, Merletti R, Jull G. Identification of innervation zones of sternocleidomastoid and scalene muscles: a basis for clinical and research electromyography applications. Clinical Neurophysiology 2002;113(1):57–63. Falla D, Bilenkij G, Jull G. Patients with chronic neck pain demonstrate altered patterns of muscle activation during performance of a functional upper limb task. Spine 2004a;29(13):1436–40. Falla D, Jull G, Edwards S, Koh K, Rainoldi A. Neuromuscular efficiency of the sternocleidomastoid and anterior scalene muscles in patients with neck pain. Disability and Rehabilitation 2004b; 26(12):712–7. Falla D, Jull G, Hodges PW. Patients with neck pain demonstrate reduced electromyographic activity of the deep cervical flexor muscles during performance of the craniocervical flexion test. Spine 2004c;29(19):2108–14. Falla D, Jull G, Hodges PW. Feedforward activity of the cervical flexor muscles during voluntary arm movements is delayed in chronic neck pain. Experimental Brain Research 2004d;157:43–8. Falla D, Jull G, Hodges P, Vicenzino B. An endurance-strength training regime is effective in reducing myoelectric manifestations of cervical flexor muscle fatigue in females with chronic neck pain. Clinical Neurophysiology 2006;117:828–37. Falla D, Jull G, Russell T, Vicenzino B, Hodges P. Effect of neck exercise on sitting posture in patients with chronic neck pain. Physical Therapy 2007;87(4):408–17.

Jull G, Bogduk N, Marsland A. The accuracy of manual diagnosis for cervical zygapophysial joint pain syndromes. The Medical Journal of Australia 1988;148(5):233–6. Jull G, Barrett C, Magee R, Ho P. Further clinical clarification of the muscle dysfunction in cervical headache. Cephalalgia 1999;19(3): 179–85. Jull G, Trott P, Potter H, Zito G, Niere K, Shirley D, et al. A randomized controlled trial of exercise and manipulative therapy for cervicogenic headache. Spine 2002;27(17):1835–43. Jull G, Falla D, Treleaven J, Sterling M, O’Leary S. A therapeutic exercise approach for cervical disorders. In: Boyling JD, Jull G, editors. Grieve’s modern manual therapy: the vertebral column. UK: Elsevier; 2004a. Jull G, Kristjansson E, Dall’Alba P. Impairment in the cervical flexors: a comparison of whiplash and insidious onset neck pain patients. Manual Therapy 2004b;9(2):89–94. Jull G, Falla D, Hodges P, Vicenzino B. Cervical flexor muscle retraining: physiological mechanisms of efficacy. In: Proceedings of the 2nd international conference on movement dysfunction, Edinburgh, Scotland, 2005. Jull GA. Deep cervical flexor muscle dysfunction in whiplash. Journal of Musculoskeletal Pain 2000;8(1/2):143–54. McArdle W, Katch F, Katch V. Exercise physiology. Baltimore: Williams Wilkins; 1996. Vernon H, Mior S. The Neck Disability Index: a study of reliability and validity. Journal of Manipulative and Physiological Therapy 1991;14(7):409–15. Ylinen J, Takala EP, Nykanen M, Hakkinen A, Malkia E, Pohjolainen T, et al. Active neck muscle training in the treatment of chronic neck pain in women: a randomized controlled trial. The Journal of the American Medical Association 2003;289(19):2509–16.

ARTICLE IN PRESS

Manual Therapy 13 (2008) 513–519 www.elsevier.com/math

Original article

Bilateral and unilateral increases in calcaneal eversion affect pelvic alignment in standing position Rafael Z.A. Pinto, Thales R. Souza, Renato G. Trede, Renata N. Kirkwood, Elyonara M. Figueiredo, Se´rgio T. Fonseca Movement Analysis Laboratory, Department of Physical Therapy, Universidade Federal de Minas Gerais (UFMG), Av. Antoˆnio Carlos 6627, Escola de Educac- a˜o Fı´sica, Fisioterapia e Terapia Ocupacional, CEP: 31270-010, Belo Horizonte, MG, Brazil Received 25 November 2006; received in revised form 29 May 2007; accepted 3 June 2007

Abstract Excessive foot pronation has been associated with the occurrence of low back pain, possibly for generating changes in the lumbopelvic alignment. However, the influence of foot pronation (measured as calcaneal eversion) on pelvic alignment during standing has not been well established. Fourteen young healthy subjects participated in the study. A Motion Analysis System was used to obtain pelvic positions in sagittal and frontal planes and calcaneal position in the frontal plane. Volunteers were filmed in relaxed standing position during three trials, in three conditions: control; unilateral experimental with increased right calcaneal eversion and bilateral experimental with increased bilateral calcaneal eversion. Increased calcaneal eversion was obtained using wedges tilted 101 medially, unilaterally and bilaterally. Repeated measures ANOVAs with Bonferroni corrections were used for statistical analysis. Unilateral and bilateral use of medially tilted wedges produced a significant increase of calcaneal eversion (Pp0.01), on the right and left sides. Bilateral and unilateral increases of the calcaneal eversion caused average pelvic anteversion of 1.571 (P ¼ 0.003) and 1.411 (P ¼ 0.021), respectively. Unilaterally increased everted position generated an average pelvic lateral tilt of 1.461 (Po0.001). Excessive calcaneal eversion during standing changes pelvic alignment and should be considered, associated with other relevant factors, when assessing pelvic misalignments. r 2007 Elsevier Ltd. All rights reserved. Keywords: Posture; Pelvis; Calcaneus; Foot pronation.

1. Introduction The pelvic girdle is responsible for the anatomic connection and transmission of forces between upper and lower quadrants of the musculoskeletal system and, thus, affects and is affected by these segments (Snijders et al., 1993). The alignment and movement of the lumbopelvic complex depend on the integrity of the osteoligamentar system, on the adequate interaction of Corresponding author. Rua Flavita Bretas, n1. 226, apto. 1201, Luxemburgo, Belo Horizonte, Postal Code: 30380-410, Minas Gerais, Brazil. Fax: +55 31 3293 0104. E-mail address: [email protected] (R.Z.A. Pinto).

1356-689X/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2007.06.004

muscles and fascia and on an appropriate neuromuscular control (Panjabi, 1992; Pool-Goudzwaard et al., 1998). Pelvic position also depends on anatomically remote factors such as the alignment of the lower limbs joints, during activities performed in a closed kinematic chain (Gurney, 2002; Khamis and Yizhar, 2007). Lower limbs length discrepancies, as an example, generate a lateral inclination of the pelvis that may cause scoliosis and other pathological conditions in the lumbar spine (Gurney, 2002; Aebi, 2005). Thus, changes in lower limbs posture may lead to the presence of postural alterations of the lumbopelvic complex, increasing the risk of developing low back pain (Botte, 1981; Rothbart and Estabrook, 1988; Aebi, 2005).

ARTICLE IN PRESS 514

R.Z.A. Pinto et al. / Manual Therapy 13 (2008) 513–519

The posture of the feet in standing may have influence on pelvic alignment (Rothbart and Estabrook, 1988; Khamis and Yizhar, 2007) and, consequently, on the spine posture (Levine and Whittle, 1996; Legaye et al., 1998; Gurney, 2002). Pronation of the subtalar joint is a triplanar motion that, in closed kinematic chain, is characterized by adduction and plantar flexion of the talus, and eversion of the calcaneus (Rockar, 1995). The adduction of the talus leads to the internal rotation of the lower limb (Rockar, 1995) and the eversion of the calcaneus, associated to the plantar flexion of the talus, lead to a functional reduction of the lower limb length (Gurney, 2002). In the presence of an increased calcaneal eversion, a measure commonly used to assess excessive foot pronation (Donatelli et al., 1999; Haight et al., 2005; Khamis and Yizhar, 2007), the lower limb assumes an excessive internally rotated position with reduction in limb length and may alter the alignment of the pelvic girdle (Botte, 1981; Gurney, 2002; Khamis and Yizhar, 2007). The bilateral presence of excessive calcaneal eversion generates internal rotation of the hips and, consequently, may lead to increased pelvic anteversion (Khamis and Yizhar, 2007) and to the presence of lumbar hyperlordosis (Levine and Whittle, 1996). Unilateral or asymmetric presence of excessive calcaneal eversion is expected to produce a functional lower extremities length difference and, consequently, may produce a lateral tilt of the pelvis to the side with increased foot pronation, which, in its turn, may cause a certain degree of scoliosis (Gurney, 2002). Thus, the presence of excessive calcaneal eversion, bilateral or unilateral, may be related to the occurrence of pathological conditions of the lumbar spine (Botte, 1981; Rothbart and Estabrook, 1988). Khamis and Yizhar (2007) have recently investigated the effects of the bilateral calcaneal eversion increase in the sagittal plane posture of the pelvis in asymptomatic subjects. Wedges medially tilted were used in order to simulate excessive feet pronation while in the standing position. There was systematic increase of the pelvic anteversion in accordance with the increase of the calcaneal eversion, in addition to increase in internal rotation of the shank and hip. There is still a lack of evidences about the effects of increases in calcaneal eversion on the sagittal and frontal planes pelvic postures or possible changes in pelvic alignment due to the unilateral (asymmetric) increase of the calcaneal eversion. Therefore, the objectives of the present study were to investigate the effects of immediate increases in calcaneal eversion, unilaterally and bilaterally, on pelvic alignment during the maintenance of the relaxed standing position.

2. Methods 2.1. Subjects Fourteen young healthy subjects, seven men and seven women, average age 22.85 years old (SD 2.47), height 1.68 m (SD 0.11) and weight 61.89 kg (SD 11.42), took part in this study. The sample size was calculated based on a pilot study with five participants. Effect size indexes (f) were calculated for the variables of pelvic position in the sagittal and frontal planes. These values were used for determining the sample size, which would enable the study to reach a minimum statistical power of 80%, considering a significance level of 0.05. This calculation determined a number of 14 participants (Portney and Watkins, 2000). The inclusion criteria were the presence of at least 101 of eversion and 281 of inversion in the ankle joint complex (Astrom and Arvidson, 1995), and of 301 of internal rotation and 401 of external rotation in the hip joint (Magee, 2005). Furthermore, the participants could not have a structural lower limbs length difference of more than 1 cm or present any pain or pathology in the ankles, hips, pelvis and spine for at least 1 year. No subject presented pain or discomfort during data collection. 2.2. Instrumentation A motion analysis system (ProReflex, Qualisys Medical AB, Gothenburg, Sweden) with four cameras was used to obtain kinematic data about the pelvis and subtalar joint position. This is a video-based system, which enables three-dimensional reconstruction of passive reflective markers placed on specific points of the body. The data were processed by the Qualsys Track Manager 1.6.0.x software, which calculated the position of each marker in two dimensions. The coordinates of these markers were constructed in three dimensions by means of the combination of the positions captured by the four cameras. The calibration of the system was carried out using an L-shaped metallic structure, with three markers attached to the X-axis and two markers to the Y-axis. The laboratory reference coordinates were determined according to this metallic structure, defining the X-axis as anterio-posterior, the Y-axis as medio-lateral and the Z-axis as proximal-distal. The anatomic markers were positioned bilaterally on the following anatomic points: iliac tubercles; greater trochanters; lateral and medial epicondiles of the femur; lateral and medial malleoluses and heads of the first and fifth metatarsus. Clusters, containing tracking markers distributed in a non-collinear manner, were attached bilaterally to the participant’s calcaneus, shanks and pelvis. The clusters of the calcaneus consisted of three rigid stems, each containing a tracking marker on their

ARTICLE IN PRESS R.Z.A. Pinto et al. / Manual Therapy 13 (2008) 513–519

515

The present study was approved by the University’s Ethics in Research Committee and the participants signed an informed consent term agreeing to participate in the study.

Initially, the anatomic markers and the clusters were attached to the segments as previously described. The reference position was obtained with the participant in relaxed standing position, for 1 s. The objective of this trial was to enable segment’s coordinate systems construction during data processing and to determine the neutral position of the calcaneus relative to the shank. Subsequently, the anatomic markers were removed and the data of each participant were collected in three different static conditions: (1) Control Condition: the participant maintained the right and left lower limbs on two flat wedges (Fig. 1). This condition aimed at not producing changes in the calcaneal posture and keeping the participant in his/her natural standing position; (2) Unilateral Experimental Condition: the participant stood with the left lower limb on a flat wedge and the right lower limb on a 101 medially tilted wedge used to increase unilaterally calcaneal eversion (Fig. 2A); (3) Bilateral Experimental Condition: the

Fig. 1. Position of the subject and the clusters in the control condition.

Fig. 2. (A) Unilateral experimental condition and (B) bilateral experimental condition.

tip, firmly attached to molded metallic basis, which involved the calcaneus posteriorly, laterally and medially. These clusters were attached right below the insertion of the tendon of the calcaneus. The clusters of the shanks were semi-rigid, consisting of neoprene girdles with three tracking markers each. The clusters, positioned below the head of the fibula, were firmly secured to the segment by Velcros. The cluster of the pelvis consisted of a rigid basis containing three tracking markers and it was attached to the pelvis on the posterior surface of the sacrum by means of an elastic girdle, which involved the entire segment (Fig. 1). 2.3. Procedures

ARTICLE IN PRESS 516

R.Z.A. Pinto et al. / Manual Therapy 13 (2008) 513–519

participant stood on two 101 medially tilted wedges, aiming at increasing bilaterally the calcaneal eversion (Fig. 2B). The conditions sequence was randomized and the right limb was chosen for standardization of the unilateral intervention. The positions of the wedges on the ground and of the feet on the wedges were marked on papers in order to guarantee that changes would not occur in these parameters, during the different conditions of the study. Moreover, the lateral edge of foot on the tilted wedge was aligned with the lateral edge of the wedge. This procedure guaranteed that the lateral region of the plantar surface of the feet was always at the same height, since the height of the lateral edge of the tilted wedge was equal to the height of the flat wedge (Fig. 2A). Thus, postural changes were consequent only to the increase of calcaneal eversion in the experimental conditions. For each condition three trials of 5 s were carried out, with a data collection frequency of 60 Hz.

the tracking markers of the foot were attached specifically to the calcaneus, changes in foot posture relative to the shank informed about modifications in calcaneal position. Thus, calcaneal alignment was measured through calcaneal inversion and eversion at the foot–ankle complex. This variable was normalized using the calcaneal position obtained in the reference position trial. Therefore, calcaneal neutral position in the present study was defined as the relaxed calcaneal stance position. Although the motion of the ankle joint complex, specifically of the subtalar joint, is triplanar (Rockar, 1995) the frontal plane posture was chosen since its measurement is one of the most common and easiest techniques for assessment of foot pronation in clinical settings (Haight et al., 2005). Means and standard deviations of the position of the pelvis and calcaneus were computed from the three trials for each participant in each study condition. 2.5. Statistical analysis

2.4. Data reduction The data were processed through the Visual 3D Motion Analysis Software (C-Motion, Inc., Rockville, USA). Initially, the data collected with the reference position was used for creating the rigid bodies corresponding to the pelvis, shanks and feet segments. The positions of the anatomic markers were used for attributing a longitudinal axis (Z-axis) and coordinate systems for each segment, coherent to the anatomic planes and axes. The tracking markers were associated to each corresponding segment so that changes in their position would determine changes in the position of these segments. The posture of the pelvis in the frontal and sagittal planes were defined as the position of this segment in relation to the X and Y axes of the laboratory, respectively. The posture of the calcaneus was defined as the position of the foot in relation to the shank of the same lower limb, in the X-axis of the ankle joint. Since

One way repeated measures analyses of variance (ANOVA) were used to compare the following dependent variables: (1) pelvic posture in the sagittal plane; (2) pelvic posture in the frontal plane; (3) right calcaneal posture in the frontal plane and (4) left calcaneal posture in the frontal plane. Bonferroni corrections were made according to the number of ANOVAs performed, setting the a level at 0.0125. When the ANOVAs identified significant differences, pre-planned contrasts were used to locate significant differences between the control condition and each experimental condition (two comparisons). The a level considered for the contrast analyses was set at 0.05.

3. Results Means and standard deviations for each dependent variable are presented in Table 1.

Table 1 Means and standard deviations of the pelvic posture in the sagittal and frontal planes and of the calcaneal posture in the frontal plane, in the study conditions Variable

Control condition

Unilateral EC

Bilateral EC

Pelvic posture (sagittal plane) Pelvic posture (frontal plane) Right calcaneus (frontal plane) Left calcaneus (frontal plane)

2.251 (SD 8.76) 0.771 (SD 1.83) 0.251 (SD 1.04) 0.721 (SD 1.01)

3.661 0.691 6.881 1.571

3.821 (SD 8.82) 0.691 (SD 1.61) 7.751 (SD 3.07) 9.081 (SD 2.21)

(SD (SD (SD (SD

8.28) 1.96) 2.43) 1.48)

EC, experimental condition. Positive values: pelvic anteversion, pelvic tilt for the right side, calcaneal eversion. Negative values: pelvic retroversion, pelvic tilt for the left side, calcaneal inversion.  Significant differences identified by the ANOVAs (Po0.0125).

ARTICLE IN PRESS R.Z.A. Pinto et al. / Manual Therapy 13 (2008) 513–519

517

4. Discussion

Fig. 3. Graphic representation of the differences of pelvic posture values in the experimental conditions in relation to the control condition. EC: Experimental Condition. *Statistically significant difference (Po0.05). Positive values: anteversion (sagittal plane) and lateral tilt in the direction of the limb with increased calcaneal eversion (frontal plane).

3.1. Pelvic posture in the sagittal plane Bilateral and unilateral medially tilted wedges caused significant changes of pelvic posture, in comparison to the control condition, producing a mean pelvic anteversion of 1.571 (SD 1.57; 95%CI ¼ 0.70–2.49; F ¼ 13.844; P ¼ 0.003) and 1.411 (SD 1.99; 95%CI ¼ 0.25–2.55; F ¼ 6.692; P ¼ 0.021), respectively (Fig. 3). 3.2. Pelvic posture in the frontal plane Unilateral medially tilted wedges caused a significant change in pelvic posture in the frontal plane, in comparison to the control condition, generating an average lateral tilt of 1.461 in the direction of the limb with increased calcaneal eversion (SD 0.45; 95%CI ¼ 1.20–1.72; F ¼ 148.34; Po0.001). The bilateral experimental condition did not generate changes in pelvic posture in the frontal plane (95%CI ¼ 0.24 to 0.36; F ¼ 0.34; P ¼ 0.566) (Fig. 3). 3.3. Calcaneal posture Unilateral experimental condition caused a significant change in the right and left calcaneal postures, in comparison to the control condition, producing 6.621of eversion (SD 3.09; 95%CI ¼ 4.84–8.41; F ¼ 64.26; Po0.001) and 0.841 of eversion (SD 1.07; 95%CI ¼ 0.22–1.46; F ¼ 8.7; P ¼ 0.011), respectively. The bilateral experimental condition caused significant changes of the right and left calcaneal postures, in comparison to the control condition, causing eversion of 7.51 (SD 3.67; 95%CI ¼ 5.38–9.62; F ¼ 58.53; Po0.001) and 8.361 (SD 2.09; 95%CI ¼ 7.15–9.56; F ¼ 223.83; Po0.001), respectively.

The identified increase in pelvic anteversion, consequent to the bilateral increase in calcaneal eversion, is in accordance with the results found by Khamis and Yizhar (2007). However, the present study identified larger amplitudes of pelvic anteversion and calcaneal eversion than the ones observed by Khamis and Yizhar (2007), who found 0.51 of mean pelvic anteversion and approximately 1–31 of mean calcaneal eversion, when they used bilateral 101 medially tilted wedges. The present study observed a mean anteversion of 1.571 and mean calcaneal eversion of 7.51 and 8.361 at the right and left sides, respectively. It is possible that these magnitude differences have occurred due to methodology of attachment of the tracking markers on body segments. Khamis and Yizhar (2007) used markers directly attached to the skin, whereas the present study used only clusters, which reduce the artifacts of movement generated by the soft tissues improving the detection of the tracking marker on its anatomic position (Chiari et al., 2005). Furthermore, Khamis and Yizhar (2007) analyzed calcaneal position in two dimensions while the present study used a threedimension analysis, which is probably more suitable, considering that the calcaneal motion at the ankle joint complex is triplanar in oblique axes (Rockar, 1995). The present study also found a significant pelvic anteversion of 1.411 as a result of the unilateral increase in calcaneal eversion, which showed that the unilateral, or asymmetric, presence of an increased calcaneal eversion is sufficient to modify sagittal plane pelvic posture. There are possible mechanisms that may lead to the interdependence between calcaneal eversion and pelvic anteversion. The subtalar pronation, coupled with the calcaneal eversion, generates an internal rotation of the lower limb at the hip joint (Botte, 1981; Khamis and Yizhar, 2007), which may lead to a posterior location of the femur’s head and, thus, to a posterior shift of the pelvis. These postural modifications might cause the individuals to bring the trunk anteriorly, in order to regain postural balance relative to gravity, which could be achieved through a pelvic anteversion (with a possible compensatory spine extension). Furthermore, internal rotation of the hip may tension the iliopsoas muscle (Botte, 1981) and the ligaments of this joint’s capsule. The direction of the fibers of these structures determines that the tension generated by hip internal rotation produce a pelvic anteversion torque. Thus, the posterior location of the femur’s head and pelvis and the tension of iliopsoas and hip joint capsule may lead to the identified postural changes. The unilateral increase in calcaneal eversion led to lateral pelvic tilt of 1.461 due to the shortening of the ipsilateral lower limb, creating a functional limb length

ARTICLE IN PRESS 518

R.Z.A. Pinto et al. / Manual Therapy 13 (2008) 513–519

difference (Botte, 1981; Gurney, 2002). Thus, excessive asymmetric calcaneal eversion may lead to lateral tilt of the sacral base and consequently to lumbar scoliosis (Gurney, 2002; Aebi, 2005). The observed increase in calcaneal eversion, consequent to the use of medially tilted wedges, showed that the intervention (foot wedge) affected ankle joint complex’s frontal plane position in standing. Additionally, this increase can be clinically identified in subjects who present altered anatomic alignments of the feet. Donatelli et al. (1999) observed, through goniometric measurements, that subjects who have increased forefoot varus presented an increase of approximately 71 in calcaneal eversion while in the standing position, compared to subjects who have forefoot valgus. Thus, the intervention used in the present study produced postural changes, similar to that found in clinical practice. It is important to point out that the use of the relaxed calcaneal stance position to define the neutral position of the calcaneus, as carried out during the data processing, implies that the magnitudes of calcaneal eversion observed in the experimental conditions correspond to an increase in the stance calcaneal posture (5–71 of eversion) commonly found in the clinical context (Haight et al., 2005). Moreover, clinicians should be aware of the between-subjects variability of the orientation of the subtalar joint axis of movement which produces different magnitudes of lower limb internal rotation coupled with the calcaneal eversion (Nawoczenski et al., 1998), resulting different degrees of pelvic anteversion. Although there is not a clinical measure for this calcaneus–shank coupling, this issue should be taken into account for adequate interpretations of the present study’s results. The observed increase in left calcaneal eversion during the unilateral experimental condition demonstrated that there is force propagation between body sides. It is possible that the modification in pelvic posture consequent to the increase in right calcaneal eversion might have affected left lower limb posture through joint intersegmental forces (Zajac et al., 2002), showing that, besides being affected, pelvic alignment can affect foot– ankle complex’s posture. This result revealed a coupling between right and left feet postures in standing. The lumbar spine posture depends on the pelvic alignment in standing position (Legaye et al., 1998). Thus, anteversion and lateral tilt of the pelvis may lead to the presence of hyperlordosis (Legaye et al., 1998) and scoliosis (Gurney, 2002), respectively. Lumbar hyperlordosis results in an increase of the loads imposed on the facet joints and it has been associated to the occurrence of lumbar pain (Shirazi-Adl and Drouin, 1987; Steinberg, et al., 2003). Christie et al. (1995), using photographic techniques, observed that subjects with chronic low back pain have increased lumbar lordosis during standing, when compared to asymptomatic

subjects. Lumbar scoliosis generates asymmetric loads on the intervertebral discs, which contributes to the degeneration of these structures (Aebi, 2005). Steinberg et al. (2003) found a greater frequency of scoliosis, identified by X-ray, in young male army recruits with a history of non-traumatic low back pain when compared to asymptomatic ones. Thus, the changes in pelvic alignment as observed in the present study, and the possible changes occurred in lumbar posture, may contribute to the development of low back pain. There still is not a consensus about the amplitudes of pelvic anteversion and lateral tilt that might be considered as risk factors for development of lumbopelvic disorders. Christie et al. (1995) found that symptomatic subjects present an average increase of 71 in lumbar lordosis in comparison to asymptomatic subjects and Levine and Whittle (1996) observed that a pelvic anteversion of 11.41 led to a 10.81 increase in lumbar lordosis. These results and the assumption of a linear relationship between pelvic and lumbar sagittal plane postures (Legaye et al., 1998) suggest that each degree of pelvic anteversion generates approximately 11 of lumbar lordosis. Thus, the observed changes of pelvic alignment in the experimental conditions can lead to a mean increase in lumbar lordosis of 1.51, which represents about 20% of the difference between symptomatic and asymptomatic subjects (Christie et al., 1995). Therefore, it is important to stress that the small magnitudes of the postural changes observed in the present study are not likely to produce lumbopelvic disorders. The observed changes may be clinically relevant only if associated with other factors that contribute to the production of greater magnitudes of pelvic anteversion, such as increased stiffness of the hip flexors and decreased stiffness of the hip extensors and abdominal muscles (Sahrmann, 2002). It is important to point out that the results of this study are related to the immediate increase of calcaneal eversion. Hence, the findings add to clinical reasoning information related to the direction rather than to the magnitude of pelvic postural changes found in clinical settings. Long term tissue adaptations may allow greater postural changes to occur (Mueller and Maluf, 2002) leading to greater risk of pathology development. This view agrees with previously described clinical observations, which indicate the presence of excessive foot pronation in patients who presented lumbar pain and pelvic misalignments in the frontal and sagittal planes (Botte, 1981; Rothbart and Estabrook, 1988). Moreover, larger increases in calcaneal eversion can be found in clinical settings, which may lead to greater increases in pelvic anteversion. Therefore, we hypothesize that excessive calcaneal eversion may play an important role in low back pain etiology specifically when associated with other clinically relevant factors and if present in the long-term. Longitudinal studies are necessary to clarify

ARTICLE IN PRESS R.Z.A. Pinto et al. / Manual Therapy 13 (2008) 513–519

long-term effects of excessive calcaneal eversion and its possible influence on pathology development.

5. Conclusions Bilateral and unilateral increases of calcaneal eversion led to small but significant changes in pelvic alignment. The bilateral and unilateral conditions caused increases of pelvic anteversion and the unilateral condition led to lateral pelvic tilt. These results show that the presence of excessive calcaneal eversion at the foot–ankle complex may be considered as a contributing factor for the production of pelvic misalignments during the maintenance of standing position. Thus, the position of the calcaneus should be taken into account, associated with other clinically relevant factors, when assessing pelvic posture in subjects with lumbopelvic disorders related to postural problems. References Aebi M. The adult scoliosis. European Spine Journal 2005;14:925–48. Astrom M, Arvidson T. Alignment and joint motion in the normal foot. Journal of Orthopeadic and Sports Physical Therapy 1995; 22(5):216–22. Botte RR. An interpretation of the pronation syndrome and foot types of patients with low back pain. Journal of the American Podiatry Association 1981;71(5):243–53. Chiari L, Della Croce U, Leardini A, Cappozzo A. Human movement analysis using stereophotogrammetry. Part 2: instrumental errors. Gait and Posture 2005;21:197–211. Christie HJ, Kumar S, Warren SA. Postural aberrations in low back pain. Archives of Physical Medicine and Rehabilitation 1995; 76:218–24. Donatelli R, Wooden M, Ekedahl SR, Wilkes JS, Cooper J, Bush AJ. Relationship between static and dynamic foot postures in professional baseball players. Journal of Orthopeadic and Sports Physical Therapy 1999;29(6):316–25. Gurney B. Leg length discrepancy. Gait and Posture 2002;15:195–206. Haight HJ, Dahm DL, Smith J, Krause DA. Measuring standing hindfoot alignment: reliability of goniometric and visual measurements. Archives of Physical Medicine and Rehabilitation 2005; 86:571–5.

519

Khamis S, Yizhar Z. Effect of feet hyperpronation on pelvic alignment in a standing position. Gait and Posture 2007;25(1):127–34. Legaye J, Duval-Beaupere G, Hecquet J, Marty C. Pelvic incidence: a fundamental pelvic parameter for three-dimensional regulation of spinal sagittal curves. European Spine Journal 1998;7:99–103. Levine D, Whittle MW. The effects of pelvic movement on lumbar lordosis in the standing position. Journal of Orthopeadic and Sports Physical Therapy 1996;24(3):130–5. Magee DJ. Orthopedic physical assessment, 4th ed. Philadelphia: W.B. Saunders Company; 2005. Mueller MJ, Maluf KS. Tissue adaptation to physical stress: a proposed ‘‘physical stress theory’’ to guide physical therapist practice, education, and research. Physical Therapy 2002;82: 383–403. Nawoczenski DA, Saltzman CL, Cook TM. The effect of foot structure on the three-dimensional kinematic coupling behavior of the leg and rear foot. Physical Therapy 1998;78(4):404–16. Panjabi MM. The stabilizing system of the spine: I. Function, dysfunction, adaptation, and enhancement. Journal of Spinal Disorders 1992;5(4):383–9. Pool-Goudzwaard AL, Vleeming A, Stoeckart R, Snijders CJ, Mens JM. Insufficient lumbopelvic stability: a clinical, anatomical and biomechanical approach to ‘a-specific’ low back pain. Manual Therapy 1998;3(1):12–20. Portney LG, Watkins MP. Foundations of clinical research applications to practice. 2nd ed. New Jersey: Prentice-Hall; 2000. Rockar Jr. PA. The subtalar joint: anatomy and joint motion. Journal of Orthopaedic and Sports Physical Therapy 1995;21(6):361–72. Rothbart BA, Estabrook L. Excessive pronation: a major biomechanical determinant in the development of chondromalacia and pelvic lists. Journal of Manipulative and Physiological Therapeutics 1988;11:373–9. Sahrmann SA. Diagnosis and treatment of movement impairment syndromes. St. Louis: Mosby; 2002. Shirazi-Adl A, Drouin G. Load-bearing role of facets in a lumbar segment under sagittal plane loadings. Journal of Biomechanics 1987;20:601–13. Snijders CJ, Vleeming A, Stoeckart R. Transfer of lumbosacral load to iliac bones and legs. 1: biomechanics of self-bracing of the sacroiliac joints and its significance for treatment and exercise. Clinical Biomechanics 1993;8:285–94. Steinberg EL, Luger E, Arbel R, Menachem A, Dekel S. A comparative roentgenographic analysis of the lumbar spine in male army recruits with and without lower back pain. Clinical Radiology 2003;58:985–9. Zajac FE, Neptune RR, Kautz SA. Biomechanics and muscle coordination of human walking. Part I: introduction to concepts, power transfer, dynamics, and simulations. Gait and Posture 2002;16:215–32.

ARTICLE IN PRESS

Manual Therapy 13 (2008) 520–528 www.elsevier.com/math

Original article

Measuring the posteroanterior stiffness of the cervical spine Suzanne J. Snodgrassa,, Darren A. Rivetta, Val J. Robertsonb a

Discipline of Physiotherapy, The University of Newcastle, Box 24, Hunter Building, Callaghan, NSW 2308, Australia b School of Health Sciences, The University of Newcastle, Australia Received 27 October 2006; received in revised form 14 June 2007; accepted 22 July 2007

Abstract An essential part of improving manual therapy treatment for cervical spine disorders is the identification of the mechanical effects of manual techniques. The aims of this research were to develop a reliable and safe instrument for measuring cervical spine stiffness, and to document stiffness in a group of asymptomatic individuals. A device for measuring cervical spine stiffness was designed and tested. The stiffness of the cervical spine of 67 asymptomatic individuals was measured at C2 and C7 on one or more occasions. Stiffness was defined as the slope of the linear region of the force–displacement curve (coefficient K). For C2, the linear region of the force–displacement curve was from 7 to 40 N, and for C7, 20–70 N. The mean stiffness (coefficient K) on the first measurement occasion at C2 was 4.58 N/mm (95% CI 4.30–4.85), and at C7 was 7.03 N/mm (95% CI 6.50–7.57). ICC(2,1) for repeated measurements was 0.84 (95% CI 0.74–0.90). Stiffness measurements in the cervical spine were generally lower than those previously reported for the lumbar spine. Age was positively associated with C2 stiffness (p ¼ 0.01). Males were stiffer at C7 than females (po0.001). This research provides a basis for future studies investigating the effects of manual techniques on cervical spine stiffness, potentially leading to improved outcomes for patients treated by manual therapy. r 2007 Elsevier Ltd. All rights reserved. Keywords: Cervical vertebrae; Manipulation; Spinal; Spine; Stiffness

1. Introduction Manual therapy techniques are often recommended to treat mechanical disorders of the cervical spine. Of these techniques, the posterior-to-anterior (PA) mobilisation as described by Maitland et al. (2005) is one of the most commonly used (Jull, 2002; Magarey et al., 2004). The dosage of mobilisation is guided by a grading system which is based on a therapist’s perception of the stiffness of the vertebra being treated (Maitland et al., 2005). To improve the selection and delivery of manual treatments, the mechanical effects of manual treatments need to be investigated.

Corresponding author. Tel.: +61 2 49212089; fax: +61 2 49217902.

E-mail address: [email protected] (S.J. Snodgrass). 1356-689X/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2007.07.007

Stiffness in the cervical spine, unlike the lumbar spine, has not been quantified. Doing this is the first stage in attempting to measure the effects of specific techniques on different presenting clinical problems. The first aim of this research was to develop a reliable and safe instrument for measuring cervical spine stiffness. The second aim was to document stiffness in a group of asymptomatic individuals. 2. Methods 2.1. Equipment The apparatus for collecting cervical spine stiffness data is modelled on three lumbar spine stiffness assessment devices described previously (Lee and Svensson, 1990; Latimer et al., 1996b; Edmondston et al., 1998). Specifically, the equipment was designed

ARTICLE IN PRESS S.J. Snodgrass et al. / Manual Therapy 13 (2008) 520–528

to simultaneously measure the excursion at a point on the cervical spine while applying a mechanical force at a constant speed, and the resistance to that force. The cervical spine stiffness assessment device uses a 12 V direct current (DC) motor (Model No. BM 4023MA2, Shinko Electric Co. Ltd., Minato-Ku, Tokyo, Japan) to power a gear drive (Fig. 1). This produces forward and backward movement of a stainless steel rod used as an indenter on the skin of the neck overlying a spinous process. The indenting end of the rod has a head made of firm plastic that is 15 mm in diameter (flat portion), with a 2.5 mm tapered edge (Fig. 1, D). Movement of the rod is measured with a DC-operated linear variable differential transformer (LVDT, Model No. DC-EC 1000, SchaevitzTM Sensors, Lucas Control Systems, Hampton, VA, USA). Resistance to movement is measured using a compression and tension load cell (Model No. UMM-K050, Dacell Co., Ltd., Chungbuk, South Korea). Voltage output from the LVDT and load cell is routed through an amplifier (Strain Gauge Signal Conditioner, Model RM-044, Applied Measurement Australia, Sydney) to a Powerlabs system (ADInstruments, Castle Hill, Australia) at a transfer rate of

521

100 Hz. Appropriate voltage conversions for the LVDT and load cell are programmed in Powerlab’s Chart software (Version 4.2.4, ADInstruments, Castle Hill, Australia), which records output from each instrument simultaneously. The entire apparatus is mounted on a frame which allows for variations in the sagittal angle of applied force, and changes in the height and position of the device along the length of the treatment table. A power supply (Model 53/2B, Statronics, Hornsby, NSW, Australia) operates the motor that turns the gear drive. It is also linked to an electronic motor controller which allows for variations in the application of the device and safety controls. The motor controller is used to adjust the speed of movement of the rod, the voltage supplied to the motor, and the maximum force that can be applied. It has modes for manual operation and data collection. In manual mode, the motor controller allows the operator to move the rod towards or away from the subject’s skin. In the data collection mode, the indenter rod moves forward and back for five cycles, monitored by an electronic counter which is reset by the operator prior to data collection. There are remote safety switches, one held by the subject being tested and one

Fig. 1. Cervical spine stiffness assessment device positioned for C7 testing. (A) rotary mechanism for positioning device on subject; (B) mechanical stop, positioned to allow maximum displacement; (C) gear drive and motor; (D) indenter probe.

ARTICLE IN PRESS 522

S.J. Snodgrass et al. / Manual Therapy 13 (2008) 520–528

by the operator, which when pressed, automatically move the indenter rod to the position that is furthest away from the subject. A variable oscillator controls the frequency at which the rod oscillates forward and backward. The available oscillation frequencies range from 0.25 to 3 Hz, in 0.25 Hz increments (Squires et al., 2000). Adjusting the voltage supplied to the motor changes the distance the indenter rod travels while oscillating at the set frequency. The distance the rod travels is also affected by the force exerted back as the tissues are compressed by the indenter head (i.e., the resistive force). If a higher resistive force is encountered, then the traverse of the rod is less, despite the voltage supplied to the motor being the same. This safety feature prevents the rod from continuing to drive the spinous process further into the subject’s neck if strong resistance to movement occurs. To collect comparable values for a group of individuals, the voltage supplied to the motor for all tests is set to a specified level. This needs to be sufficient to collect the required force and displacement data and tolerable for the majority of a range of test subjects. Following pilot trials the voltage used in the present cervical spine stiffness tests was set at 85% of the motor’s maximum, 10.2 V. This level was based on the reported comfort of subjects during trials of different voltages. A physiotherapist with post-graduate manual therapy qualifications and extensive teaching experience indicated the level which he reported felt similar to a physiotherapist applying a grade III mobilisation. This voltage level was selected so that the force applied to the neck would be similar to a clinically applied force, without being excessive. It was then trialled on four asymptomatic subjects, one physiotherapist and three others. All tolerated it well. At the selected voltage level, the rod moves 14 mm both forwards and backwards against a resistance equal to 70 N. If there is a lower resisting force the indenter travels further within a preset limit determined by the operator; if greater, it does not travel as far. For safety reasons, the maximum force and displacement possible can be limited using controls set by the operator. If the motor controller senses the specified maximum force (by detecting the corresponding voltage), the indenter does not move any further towards the subject. The indenter remains stationary for the remainder of the forward cycle (duration depends on the oscillation frequency, e.g., less than 0.5 s if 1 Hz), before reversing. The maximum safe force was set at 80 N. This level was selected because it was well within the range of force applied to the cervical spine by physiotherapists in a pilot study (Snodgrass et al., 2007), and was considered high enough to collect the necessary data. Previous studies measuring lumbar stiffness have calculated stiffness coefficients in ranges

up to 80 N (Edmondston et al., 1998), 90 N (Shirley et al., 2002) and 200 N (Latimer et al., 1998). The maximum displacement possible is controlled by manually adjusting a mechanical stop (Fig. 1, B), which contacts an electronic switch if the rod traverses the maximum pre-set distance. Contact with the switch causes the motor to reverse away from the subject and data collection to cease. The maximum displacement available on the existing equipment is 28 mm. This was used in all trials reported here because pilot testing demonstrated large movements occur in the cervical spine without the subjects reporting any discomfort. Previous stiffness devices used for the lumbar spine have reportedly been able to measure maximum displacements of 15 mm (Edmondston et al., 1998; Allison et al., 2001) and 22 mm (Latimer et al., 1996b). 2.2. Data collection 2.2.1. Reliability testing To test the reliability of the stiffness assessment device, repeated measurements were obtained using eight different combinations of foam of varying densities. Foams were selected because their stiffness measurements were in the range of those recorded on the cervical spine in early trials. The foam was positioned under the indenter head and stiffness was measured with the sagittal angle of inclination set to zero degrees. This was repeated without moving the foam being tested. 2.2.2. Cervical spine stiffness measurement Sixty-seven asymptomatic individuals were recruited and their cervical spine stiffness measured at C2 and C7 on one or two occasions. Ethical approval for the study was granted by the University and local health service Human Research Ethics Committees. Subjects were eligible if they were between 18 and 50 years of age, and they had not had neck pain or headaches for which they sought treatment in the previous 12 months. Subjects were excluded if they had been diagnosed with any condition where PA cervical mobilisation might be contraindicated, such as inflammatory or infectious diseases affecting the neck, nerve root pain, instability, or symptoms potentially related to the vertebrobasilar system such as dizziness or nausea. Prior to the collection of cervical spine stiffness data, each subject’s C2 and C7 spinous processes were pre-marked by an experienced physiotherapist researcher using commonly recommended clinical methods (Hoppenfeld, 1976; Palmer and Epler, 1998; Gross et al., 2002). The C2 and C7 vertebrae were then pre-conditioned by applying five manual PA oscillations of force to the spinous process, as a clinician might apply when assessing a joint. During the stiffness test, each subject lay prone with their cervical spine in a neutral position while their

ARTICLE IN PRESS S.J. Snodgrass et al. / Manual Therapy 13 (2008) 520–528

head rested on a custom-made piece of foam with a cutout for the face (Dunlop utility foam AA23-130). In each case C7 was tested first and then C2. For C7 tests, the sagittal angle of inclination was standardised at 201 caudad, and for C2, 141 cephalad. These angles were based on the average angle of inclination for 252 individuals without craniocervical symptoms, as measured by radiographs in a previous study (Harrison et al., 1996). For testing each level, the stiffness measurement device was positioned by winding the mechanical stop to zero and moving the indenter rod to its starting position using the electronic manual mode. The mechanical stop was then positioned to allow maximal movement (28 mm) in testing mode. Next, the device was manually aligned with the subject. The indenter head was positioned on the mark over the spinous process by sliding the frame in the caudad–cephalad plane and securing it, then positioning the device on the subject’s skin using the rotatory mechanism at its top (Fig. 1, A) which allowed the whole device to move along a coiled thread. Indenter positioning was standardised by moving it toward the subject until it was touching the skin, stopping only when a light indentation first became visible (Fig. 2). The stiffness measurement was taken after the subject exhaled a deep breath. The subject was instructed to remain relaxed and to hold their breath (at functional residual capacity) for 5 s while data were collected. Subjects were warned that they may feel their head or neck move but to stay relaxed without resisting or tensing any muscles. In lumbar spine stiffness testing, breathing (Shirley et al., 2003) and muscle contractions (Lee et al., 1993) can affect measurements. The implications for cervical spine stiffness measurements

523

are unknown, but at least one set of local muscles, the scalenes, is active during breathing (De Troyer and Estenne, 1984). Stiffness measurements were taken with the oscillatory frequency set at 1 Hz, which corresponds with the mean frequency applied by a group of physiotherapists performing cervical mobilisations (Snodgrass et al., 2006). Two streams of data were collected for each subject: displacement by time from the LVDT and force by time from the load cell. These data were saved as text files and then extracted using a custom-written program in IDL software (Version 6.2, ITT Visual Information Solutions, Boulder, CO, USA). At this stage, prerecorded friction (from the linear bearings holding and guiding indenter rod movement) was subtracted from the results, and a force (y-axis) by displacement (x-axis) curve created representing the forward (towards subject) movement for each of the five oscillation cycles of applied force (Fig. 3). The stiffness measurement (coefficient K) at a single vertebral level was calculated as the mean of the slopes of the linear portions of the force–displacement curves for cycles two through five. The first repetition has usually been excluded in previously reported lumbar stiffness research as it is consistently different than the subsequent four cycles (Latimer et al., 1998; Shirley et al., 2002; Shirley, 2004). The linear portion of the curve is used because force– displacement curves representing spinal stiffness typically have an initial non-linear region at lower applied forces, and a linear region as higher forces are applied (Shirley, 2004). The initial non-linear or toe region is thought to represent the indenter pressing into the skin and soft tissues; essentially ‘taking up the slack’ of the

Fig. 2. Subject positioned for stiffness testing with the cervical spine in neutral and the indenter probe on the spinous process of C7.

ARTICLE IN PRESS 524

S.J. Snodgrass et al. / Manual Therapy 13 (2008) 520–528

Fig. 3. Typical force–displacement graph illustrating five cycles of applied force to C7. The bold portion of each force–displacement curve represents the stiffness device moving in the posterior-to-anterior direction. This was used for calculating spinal stiffness, which equated to the mean of the slopes of the linear regions of the force–displacement curves for cycles two through five.

soft tissues by compression prior to movement of the spinal unit (Latimer et al., 1996a, b). 2.3. Analysis 2.3.1. Reliability testing Stiffness coefficient K for foam types was defined as the mean of the slopes of the linear regions of the force– displacement curves for cycles two through five of applied force. The linear region of the force–displacement curve for each foam type was determined and used for calculating K. Reliability between repeated measures was calculated using the intra-class correlation coefficient, ICC(2,1). 2.3.2. Stiffness calculation For stiffness measurements at C2, stiffness coefficient K was calculated from 7 to 40 N, and for C7 measurements from 20 to 70 N. These selected force ranges represent the linear region, determined by calculating the linearity of consecutive data sets along each force–displacement curve. The selected values are the lower bound of the 95% confidence interval of the mean of the force values at the start of the linear region calculated for each force–displacement cycle, and the upper bound of the 95% confidence interval of the mean of the maximum forces applied for each cycle. Stiffness coefficients for cycles two through five were analysed for consistency using one-way ANOVAs with Bonferroni post hoc tests. Cycles 2 through 5 for each measurement occasion were averaged to determine the stiffness value for each spinal level of each subject. Repeatability of cervical spine stiffness measurements was determined using paired t-tests, standard error of measurement

(SEM), and ICC (2,1). Linear regression was used to determine associations between cervical spine stiffness and age, gender, height and weight of subjects. Analysis was performed in SPSS 14.0 (Chicago, IL).

3. Results 3.1. Reliability testing The stiffness assessment device was reliable for repeated measures of stiffness of inert materials. The ICC(2,1) for repeated measurements was 0.99 (95% CI 0.93–1.00). 3.2. Stiffness calculation Table 1 contains descriptive data comparing cycles two through five. For C7 measurements, the slopes for individual oscillatory cycles were not significantly different (F[3, 264] ¼ 0.17, p ¼ 0.915). For C2 measurements, cycle two was different to cycles four and five (mean difference o0.83 N/mm, po0.05), but cycles three through five were not significantly different (F[2,198] ¼ 2.56, p ¼ 0.08). A reliability analysis of C2 measurements determined that the mean of cycles 2–5 was very similar to the mean of cycles 3–5 (ICC[2,1] ¼ 0.98, 95% CI 0.87–0.99). Therefore, cycles two through five were used to calculate the average stiffness value for all tests, consistent with previous research measuring lumbar spine stiffness (Latimer et al., 1996a; Shirley et al., 2002). Subjects are described in Table 2. Using data from the first occasion of measurement for all 67 subjects, the

ARTICLE IN PRESS S.J. Snodgrass et al. / Manual Therapy 13 (2008) 520–528

525

Table 1 Mean stiffness coefficient K for the cervical spine calculated by repetition cycle of applied oscillatory force (n ¼ 67, first occasion of measurement)

Table 3 Stiffness coefficient K (N/mm) measured on two occasions at C2 and C7 with number of days between repeated measurements for each subject (n ¼ 31)

Repetition

Mean

SD

95% CI

Minimum

Subject numbera

Days between measurements

C2 trial 1

C2 trial 2

C7 trial 1

C7 trial 2

C2 measures 2 3 4 5

4.10 4.47 4.72 4.93

1.27 1.19 1.14 1.16

3.79–4.41 4.18–4.76 4.44–5.00 4.64–5.21

1.49 1.54 2.06 2.03

6.91 7.12 7.02 7.12

C7 measures 2 3 4 5

6.92 7.08 7.21 7.12

2.34 2.49 2.46 2.44

6.34–7.49 6.47–7.68 6.61–7.81 6.52–7.72

3.18 3.20 2.96 3.22

14.17 14.58 14.70 14.42

2 3 4 5 6 7 9 10 12 14 15 16 17 23 25 31 34 38 39 44 47 50 51 53 54 55 56 58 60 63 64

20 14 22 28 0.5 56 124 6 13 7 5 6 5 6 14 22 71 62 54 21 44 20 13 14 16 26 35 20 14 13 12

6.07 4.54 4.28 3.93 3.92 3.90 4.38 2.30 5.28 6.94 5.40 2.72 5.01 3.60 4.69 5.94 4.59 5.53 5.20 4.92 5.42 1.95 3.62 3.62 4.10 4.99 3.26 2.36 5.76 2.76 4.05

5.82 4.54 4.50 4.64 3.92 6.56 5.78 4.98 4.96 5.95 5.68 2.75 5.37 5.37 5.26 5.13 4.50 5.68 4.79 5.24 3.35 2.38 3.57 3.83 5.66 4.52 2.87 3.82 5.11 3.58 4.35

8.43 6.04 6.37 5.92 7.66 4.50 9.16 9.08 11.30 9.83 5.86 5.77 4.67 4.83 8.41 7.99 5.77 7.02 7.09 9.56 4.81 13.63 7.80 7.28 8.90 6.07 6.97 5.01 8.17 4.91 5.28

9.47 8.55 6.23 5.90 7.14 5.97 7.56 8.02 6.44 9.74 6.81 6.98 4.92 4.89 7.06 6.22 6.78 10.33 6.97 7.42 6.87 8.89 7.80 7.05 8.46 6.27 6.62 6.44 8.50 5.84 5.28

Maximum

Table 2 Description of subjects

N Number female Age (years) Height (cm) Weight (kg)

All mean (SD)

Subjects completing repeated testing mean (SD)

67 41 30.0 (9.5) 170.5 (9.0) 73.6 (15.8)

31 17 29.2 (9.4) 170.8 (8.4) 71.6 (13.7)

mean stiffness coefficient K at C2 was 4.58 N/mm (SD 1.13, range 1.95–6.94). At C7 it was 7.03 N/mm (SD 2.20, range 4.15–14.16). Stiffness was reasonably consistent over time. There was no significant difference between measurements for the 31 subjects who returned for repeated testing (Table 3). The mean difference for C2 measurements was 0.30 N/mm (95% CI 0.67 to 0.06, p ¼ 0.102) and for C7 0.09 (95% CI 0.55 to 0.73, p ¼ 0.783). The SEM between the two occasions was 0.53 N/mm for C2 measurements and 0.83 N/mm for C7. ICC(2,1) for all cervical measurements (both levels combined) was 0.75 (95% CI 0.62–0.84). There were five outliers that were 72 SD from the mean (three measurements at C2 and two at C7). Excluding these outliers, ICC(2,1) was 0.84 (95% CI 0.74–0.90). Linear regression indicated that the age of the subject was positively associated with C2 stiffness (p ¼ 0.01, regression coefficient 0.04, 95% CI 0.01–0.07, r2 ¼ 0.099), but gender, height and weight were not. For C7 stiffness, males were stiffer than females (mean difference 2.20 N/mm, 95% CI 1.23–3.17, po0.001, r2 ¼ 0.239). Height, weight and age were not associated with C7 stiffness. Assumptions of normality, linearity and homoscedasticity were satisfied for the linear regression models.

a

Includes only subjects who returned for repeated testing.

4. Discussion The main findings from this study were that the cervical spine responds differently to mechanical force than the lumbar spine, resulting in lower stiffness values; cervical spine stiffness differs between individuals, though an individual’s stiffness remains relatively consistent over two sessions; and cervical spine stiffness is associated with gender and age. Cervical spine stiffness values (coefficient K) were lower than those measured in the lumbar spine (Shirley et al., 2002) due to increased displacement per unit of applied force compared to that reported for the lumbar spine (Latimer et al., 1996a; Shirley et al., 2002). In contrast to the lumbar spine, the linear region of the force–displacement curve began earlier in the range; on some occasions the entire force–displacement curve was linear, with virtually no toe region observed. A possible reason may be differences in the soft tissue covering the cervical and lumbar spines. Another dissimilarity was

ARTICLE IN PRESS 526

S.J. Snodgrass et al. / Manual Therapy 13 (2008) 520–528

the force range representing the linear region and used for calculating coefficient K. A lower force range was used for cervical spine measurements compared to that reported for lumbar measurements, because lower forces were generated during cervical testing due to increased displacement. This might relate to the fact that the lumbar spine is anchored by the pelvis and ribcage which have been shown to affect stiffness values (Chansirinukor et al., 2001, 2003), whereas the cervical spine has considerably fewer restraints. Cervical spine stiffness measurement using this device appears reliable. For the device itself, the accuracy of repeated measurements of inert materials is very high. When using it on the cervical spine of asymptomatic subjects, the absolute differences between repeated measurements are reasonably small (Table 3). On 90% of occasions the differences between repeated C2 measurements was o1.5 N/mm and for C7 o2.1 N/mm. This level of agreement is similar to that previously reported for lumbar stiffness measurements (Latimer et al., 1996a; Shirley et al., 2002). The SEM of 0.53 and 0.83 N/mm, for C2 and C7 measurements, respectively, is comparable to the SEM reported for two instruments used to measure lumbar spine stiffness, 1.03 N/mm (Shirley et al., 2002) and 0.26–1.05 N/mm (Edmondston et al., 1998; Allison et al., 2001). The ICC(2,1) of 0.84 for repeated cervical spine stiffness measurements indicates excellent reliability, according to Fleiss (1986). When compared to previously reported ICC values for lumbar spine stiffness measurements, the cervical spine values are similar (Lee and Svensson, 1990; Viner et al., 1997; Allison et al., 2001; Shirley et al., 2002) to slightly lower (Latimer et al., 1996a; Edmondston et al., 1998). However, only two of these studies reported reliability for measurements taken on different days (Lee and Svensson, 1990; Shirley et al., 2002), whereas the others reported measurements taken minutes apart without the subject moving from the testing surface. The repeated cervical measurements in the current study were recorded a number of days apart (Table 3), which may account for some of the difference in ICC values compared to some previous studies. A future study should investigate the reliability of cervical stiffness measurements taken consecutively on the same day, as well as a standard number of days apart. C2 stiffness was less than C7, likely because C7 is closer to structures that might provide some additional support or restriction to movement, such as the ribs or the soft tissues about the shoulders and chest. C7 stiffness was associated with gender, with males being stiffer. Studies investigating gender differences in joint mobility agree that females usually display greater joint mobility (Russek, 1999; Seow et al., 1999; Didia et al., 2002). However, a previous study found lumbar spine stiffness was not significantly different between males

and females (Lee and Evans, 1992), and gender was not associated with C2 stiffness in the current study. C2 stiffness was associated with older age. This might be expected because the prevalence of osteoarthritis increases with age (van Saase et al., 1989), and symptoms of osteoarthritis are associated with stiffness (Kornaat et al., 2006). However, the increase in stiffness with age was very small (0.037 N/mm per year older, 95% CI 0.009–0.065), and this association with age was not observed in C7 measurements. A potential confounding factor when comparing C2 and C7 stiffness measurements was that C7 measurements were performed before C2 on each occasion. The order of measurements may have had an effect on the amount of preconditioning at each spinal level. Additionally, familiarisation may have reduced possible apprehension about the testing procedure, reducing potential muscle activity in some subjects. Both of these issues might have potentially affected the stiffness values. There are several possible limitations to the cervical spine stiffness data in this present study. First, the stiffness measurement was designed to quantify the stiffness sensations palpated by a clinician when performing a PA mobilisation, so should not be directly compared to pure segmental stiffness as measured on cadavers (Sran et al., 2005) or in vivo spinal flexibility measurements (McClure et al., 1998). To mimic clinical practice, subjects lay prone on a foam support with no additional stabilisation of the head or neck. This likely resulted in both angular rotation (sagittal plane extension) and segmental movement during testing, as would occur when a therapist clinically performs a PA mobilisation (McGregor et al., 2001; Lee et al., 2005). Therapists use information from PA motion assessment to guide manual treatment choices (Maitland et al., 2005). The cervical spine stiffness measurement attempts to quantify this, similar to the way lumbar spine stiffness has been evaluated (Shirley, 2004). In clinical terms, the stiffness measurement objectively quantifies the relationship between the amount resistance to manually applied force and the movement produced as a result of that force. Second, measurements were taken with subjects lying on a padded plinth using a custom-made piece of foam to standardise head position. Plinth padding has been shown to decrease stiffness measurements in the lumbar spine (Maher et al., 1999), so the current cervical spine stiffness measurements may be low because of this. A padded plinth was used to replicate the clinical situation. Other potential sources of error were friction within the apparatus and the positioning of the indentor rod on the skin prior to data collection. Error in positioning the rod was minimised by using a standardised process which included observing the probe touching the skin, and then winding it towards the skin until it just lightly indented the skin (Fig. 2). The

ARTICLE IN PRESS S.J. Snodgrass et al. / Manual Therapy 13 (2008) 520–528

amount of friction remained constant, so it would not affect comparisons between subjects or between repeated measurements. However, it could potentially affect comparisons between stiffness values measured by the machine and physiotherapist stiffness assessment. Lastly, there was the potential for error in the palpation of C2 and C7 by the operator. To limit this possibility, the same experienced physiotherapist operator identified the levels in each subject using standardised palpation methods. 5. Conclusion This study introduces a safe and reliable method for measuring stiffness in the cervical spine. Cervical spine stiffness measurements were shown to be less than those in the lumbar spine, with greater displacement per amount of applied force during testing. Cervical stiffness varied between individuals, with a positive association between male gender and C7 stiffness and between older age and C2 stiffness. This information forms the basis for future research that could investigate changes in stiffness as a result of manual therapy treatments and lead to improved patient outcomes. Acknowledgements The authors would like to thank Dr Colin Waters, School of Mathematical and Physical Sciences, The University of Newcastle, for writing the computer programs used in IDL, and Trevor White, Dean Jeffs and Darren Gorton from the Faculty of Health Workshop, The University of Newcastle, for contributing to the design and constructing the stiffness assessment machine.

References Allison G, Edmonston S, Kiviniemi K, Lanigan H, Simonsen AV, Walcher S. Influence of standardized mobilization on the posteroanterior stiffness of the lumbar spine in asymptomatic subjects, Physiotherapy Research International 2001;6(3):145–56. Chansirinukor W, Lee M, Latimer J. Contribution of pelvic rotation to lumbar posteroanterior movement. Manual Therapy 2001;6(4):242–9. Chansirinukor W, Lee M, Latimer J. Contribution of ribcage movement to thoracolumbar posteroanterior stiffness. Journal of Manipulative and Physiological Therapeutics 2003; 26(3):176–83. De Troyer A, Estenne M. Coordination between rib cage muscles and diaphragm during quiet breathing in humans. Journal of Applied Physiology: Respiratory, Environmental and Exercise Physiology 1984;57(3):899–906. Didia BC, Dapper DV, Boboye SB. Joint hypermobility syndrome among undergraduate students. East African Medical Journal 2002;79(2):80–1. Edmondston SJ, Allison GT, Gregg CD, Purden SM, Svansson GR, Watson AE. Effect of position on the posteroanterior stiffness of the lumbar spine. Manual Therapy 1998;3(1):21–6.

527

Fleiss JL. The design and analysis of clinical experiments. New York: Wiley; 1986. p. 7 [chapter 1]. Gross JM, Fetto J, Rosen E. Musculoskeletal examination. 2nd ed. Massachusettes: Blackwell Science, Inc; 2002. p. 43 [chapter 4]. Harrison DD, Janik TJ, Troyanovich SJ, Holland B. Comparisons of lordotic cervical spine curvatures to a theoretical ideal model of the static sagittal cervical spine. Spine 1996;21(6):667–75. Hoppenfeld S. Physical examination of the spine and extremities. Norwalk, CT: Appleton-Century-Crofts; 1976. p. 109 [chapter 4]. Jull G. Use of high and low velocity manipulative therapy procedures by Australian manipulative physiotherapists. Australian Journal of Physiotherapy 2002;48:189–93. Kornaat PR, Bloem JL, Ceulemans RYT, Riyazi N, Rosendaal FR, Nelessen RG, et al. Osteoarthiritis of the knee: association between clinical findings and MR imaging findings. Radiology 2006;239(3): 811–7. Latimer J, Goodsell MM, Lee M, Maher CG, Wilkinson BN, Moran CC. Evaluation of a new device for measuring responses to posteroanterior forces in a patient population, part 1: reliability testing. Physical Therapy 1996a;76(2):158–65. Latimer J, Lee M, Goodsell M, Maher CG, Wilkinson BN, Adams R. Instrumented measurement of spinal stiffness. Manual Therapy 1996b;1(4):204–9. Latimer J, Lee M, Adams RD. The effects of high and low loading forces on measured values of lumbar stiffness. Journal of Manipulative and Physiological Therapeutics 1998;21(3): 157–63. Lee M, Svensson NL. Measurement of stiffness during simulated spinal physiotherapy. Clinical Physics and Physiological Measurement 1990;11(3):201–7. Lee M, Esler M-A, Mildren J, Herbert R. Effect of extensor muscle activation on the response to lumbar posteroanterior forces. Clinical Biomechanics 1993;8(3):115–9. Lee RYW, Evans JH. Load–displacement–time characteristics of the spine under posteroanterior mobilisation. Australian Journal of Physiotherapy 1992;38(2):115–23. Lee RYW, McGregor AH, Bull AMJ, Wragg P. Dynamic response of the cervical spine to posteroanterior mobilisation. Clinical Biomechanics 2005;20(2):228–31. Magarey ME, Rebbeck T, Coughlan B, Grimmer K, Rivett DA, Refshauge KM. Pre-manipulative testing of the cervical spine: review, revision and new clinical guidelines. Manual Therapy 2004;9(2):95–108. Maher CG, Latimer J, Holland MJ. Plinth padding confounds measures of posteroanterior spinal stiffness. Manual Therapy 1999;4(3):145–50. Maitland GD, Banks K, English K, Hengeveld E. Maitland’s vertebral manipulation, 7th ed. Oxford: Butterworth-Heinemann; 2005. p. 171–181 [chapter 7]. McClure P, Siegler S, Nobilini R. Three-dimensional flexibility characteristics of the human cervical spine in vivo. Spine 1998; 23(2):216–23. McGregor AH, Wragg P, Gedroye WMW. Can interventional MRI provide an insight into the mechanics of a posterior–anterior mobilisation? Clinical Biomechanics 2001;16:926–9. Palmer ML, Epler ME. Fundamentals of musculoskeletal assessment techniques. 2nd ed. Philadelphia: Lippincott Williams & Wilkins; 1998. p. 230 [chapter 9]. Russek LN. Hypermobility syndrome. Physical Therapy 1999;79(6): 591–9. Seow CC, Chow PK, Khong KS. A study of joint mobility in a normal population. Annals of the Academy of Medicine 1999;28(2):231–6. Shirley D. Manual therapy and tissue stiffness. In: Boyling JD, Jull GA, Twomey LT, editors. Grieve’s modern manual

ARTICLE IN PRESS 528

S.J. Snodgrass et al. / Manual Therapy 13 (2008) 520–528

therapy. 3rd ed. Edinburgh: Churchill Livingstone; 2004. p. 381–90 [chapter 26]. Shirley D, Ellis E, Lee M. The response of posteroanterior lumbar stiffness to repeated loading. Manual Therapy 2002;7(1): 19–25. Shirley D, Hodges PW, Eriksson AEM, Gandevia SC. Spinal stiffness changes throughout the respiratory cycle. Journal of Applied Physiology 2003;95:1467–75. Snodgrass SJ, Rivett DA, Robertson VJ. Manual forces applied during posterior to anterior spinal mobilization: a review of the evidence. Journal of Manipulative and Physiological Therapeutics 2006; 29(4):316–29. Snodgrass SJ, Rivett DA, Robertson VJ. Manual forces applied during cervical mobilization. Journal of Manipulative and Physiological Therapeutics 2007;30(1):17–25.

Squires MC, Latimer J, Adams RD, Maher CG. Indenter head area and testing frequency effects on posteroanterior lumber stiffness and subjects’ rated comfort. Manual Therapy 2000;6(1):40–7. Sran MM, Khan KM, Zhu Q, Oxland TR. Posteroanterior stiffness predicts sagittal plane midthoracic range of motion and threedimensional flexibility in cadaveric spine segments. Clinical Biomechanics 2005;20(8):806–12. van Saase JLCM, van Romunde LKJ, Cats A, Vandenbrouke JP, Valkenburg HA. Epidemiology of osteoarthritis: Zoetermeer survey. Comparison of radiological osteoarthritis in a Dutch population with that of 10 other populations. Annals of the Rheumatic Diseases 1989;48:271–80. Viner A, Lee M, Adams R. Posteroanterior stiffness in the lumbosacral spine: the correlation between adjacent vertebral levels. Spine 1997;22(23):2724–9.

ARTICLE IN PRESS

Manual Therapy 13 (2008) 529–535 www.elsevier.com/math

Original article

Manual landmark identification and tracking during the medial rotation test of the shoulder: An accuracy study using three-dimensional ultrasound and motion analysis measures D. Morrisseya,, M.C. Morrisseyb, W. Driverc, J.B. Kinga, R.C. Woledgec a

Centre for Sports and Exercise Medicine, Mile End Hospital, London E14DG, UK Division of Applied Biomedical Research, King’s College London, London SE11UL, UK c Centre for Applied Biomedical Research, King’s College London, London SE11UL, UK

b

Received 19 October 2006; received in revised form 15 June 2007; accepted 22 July 2007

Abstract Palpation of movement is a common clinical tool for assessment of movement in patients with musculoskeletal symptoms. The purpose of this study was to measure the accuracy of palpation of shoulder girdle translation during the medial rotation test (MRT) of the shoulder. The translation of the gleno-humeral and scapulo-thoracic joints was measured using both three-dimensional ultrasound and palpation in order to determine the accuracy of translation tracking during the MRT of the shoulder. Two movements of 11 normal subjects (mean age 24 (SD ¼ 4), range 19–47 years) were measured. The agreement between measures was good for scapulo-thoracic translation (r ¼ 0.83). Gleno-humeral translation was systematically under estimated (p ¼ 0.03) although moderate correlation was found (r ¼ 0.65). These results indicate that translation of the measured joints can be tracked by palpation and further tests of the efficacy of palpation tracking during musculoskeletal assessment may be warranted. r 2007 Elsevier Ltd. All rights reserved. Keywords: Motion; Shoulder; Ultrasound; Accuracy

1. Introduction The shoulder medial rotation test (MRT) has been proposed as a means of deciding whether a patient presenting with shoulder dysfunction exhibits movement patterns indicative of impingement or instability pathology. The MRT has been developed by observation and analysis of patient presentations together with consideration of the research relating to shoulder girdle movement patterns (Comerford and Mottram, 2001). Stated normal values and features of movement based on this cumulative experience are described (Table 1) but have not been investigated quantitatively. Corresponding author. Tel.: +44 20 8223 8459; fax: +44 20 8983 6500. E-mail address: [email protected] (D. Morrissey).

1356-689X/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2007.07.009

The test consists of manual assessment of the accessory translation of the gleno-humeral and scapulothoracic articulations that occurs when a subject in the supine position actively medially rotates the shoulder while the shoulder is placed in 901 of abduction (scapular plane). The accessory translation of interest at the gleno-humeral joint is anterior translation of the humerus in relation to the glenoid. The movement of interest at the scapulo-thoracic articulation is the total translation/rotation as detected by palpation of the coracoid process in relation to the thorax (Fig. 1). For simplicity, this will be described as translation. Assessment of these movement patterns is used to aid clinical decision-making regarding diagnosis and rehabilitation priorities. When the primary site of ‘give’ or relative flexibility is palpated as being scapulo-thoracic, the MRT is regarded as being indicative of sub-acromial

ARTICLE IN PRESS 530

D. Morrissey et al. / Manual Therapy 13 (2008) 529–535

Table 1 Shoulder medial rotation test interpretation Normal: The subject achieves 701 of total medial rotation at 901 abduction in the scapular plane without greater then 4 mm glenohumeral or 6 mm scapulo-thoracic anterior translation. Scapulo thoracic give and impingement risk: The subject demonstrates significant translation (46 mm) at the scapulo-thoracic articulation before 701 of medial rotation. Gleno-humeral give and instability risk: The subject demonstrates significant anterior translation (44 mm) at the gleno-humeral articulation before 701 of medial rotation.

movements with the palpating fingers. To date, this accuracy has not been described in the literature. The purpose of this study was to quantify the accuracy of palpation for tracking coracoid and anterior–posterior humeral head translation during the MRT manoeuvre.

2. Method 2.1. Subjects A convenience sample of 11 subjects (three males) was recruited whose mean (SD) age was 24 (4) years with a range of 19–47 years. Subjects were included if they had no history of shoulder pain or injury requiring medical attention, no history of cervical spine area pain in the previous 6 months, were neither pregnant nor lactating and did not suffer from orthopnea or systemic joint disease. The King’s College London research ethics committee granted ethical approval and each subject provided written informed consent. One shoulder of each subject was measured by random assignment (six left, five right, six non-dominant, five dominant with dominance defined by preferred writing and throwing arm). Two movements were measured for each subject. 2.2. Instrumentation of the MRT

Fig. 1. Palpation of bony landmarks during the shoulder MRT.

impingement risk, while primary gleno-humeral ‘give’ is thought to be indicative of instability risk. Accuracy of landmark identification and tracking is critical to the accuracy of the test. An overt, detailed description of the method of palpation is necessary to optimise this accuracy. The coracoid is tracked irrespective of direction or amplitude of movement by the proximally placed palpating finger. The humeral anterior translation is tracked in relation to the coracoid position by the distally placed finger (Fig. 1). The examiner attempts to keep the fingers a fixed distance apart in the anatomically coronal plane of the body while aiming to track the gleno-humeral anterior– posterior movement with respect to coracoid position. The gleno-humeral medial-lateral and inferior-posterior movements are therefore dictated by the scapulothoracic movement while the anterior–posterior movements, the gleno-humeral movements of interest, are independently tracked. The effectiveness of the MRT is critically dependent on the examiner’s ability to accurately track the bone

Subjects lay supine with the humerus supported by a tripod-mounted gutter that allowed individual positioning of the humerus in the scapular plane at 901 abduction (Fig. 2). Movement was measured using the CODA MPX30 motion tracking system (Charnwood Dynamics, Rothley, UK). The system uses active infrared LED markers to measure positions within a 2  2  3 m volume. The precision of the instruments has been shown to be 0.2 mm in the default X (range) and Z (height) directions and 0.3 mm in the Y direction (perpendicular to the instrument) in our own and

Fig. 2. A subject is shown with the arm supported in a tripod-mounted gutter, allowing positioning in the scapular plane.

ARTICLE IN PRESS D. Morrissey et al. / Manual Therapy 13 (2008) 529–535

Fig. 3. A three-dimensional graph of the three-dimensional ultrasound derived reconstruction of the humeral head is shown in red, with the fitted co-ordinates of the sphere shown in black and the centroid in green. The scales are in mm.

400 350 300 Z axis (mm)

factory calibration tests. Marker positions were captured at 200 Hz. Motion tracking markers were attached to the anterior chest and acromium to track movement of the thorax and scapula, respectively. Further markers on the forearm were used to measure total medial rotation. Compliance with the International Society of Biomechanics (ISB) recommendations for upper limb motion tracking were ensured by digitising otherwise hidden landmarks prior to data collection and inserting these as virtual markers into the relevant co-ordinate systems. The landmarks identified on the thorax were the C7 and T8 spinous processes. The landmarks identified on the scapula were the medial end of the spine and inferior angle. These virtual markers were used to construct co-ordinate systems for the thorax and scapula that conformed with the ISB recommendations (van der Helm, 2002; Wu et al., 2005). Palpation is used in this paper to mean identification and tracking of bony landmarks. Markers were attached to the nail beds of the examiner’s palpating fingers in order to track finger movement and therefore quantify the manual landmark tracking. The subject repeatedly actively medially rotated the shoulder around the long axis of the humerus and back at a set speed for 20 s with the instruction being to move as far as was ‘comfortably possible’. A metronome guided the speed, with each movement from starting position and back taking 6 s. Typically, this resulted in collection of data from three complete movement cycles. The second movement of each data set was used for further analysis in each case. All measurements were made by an experienced physiotherapist with a postgraduate manual therapy qualification (primary author).

531

250 200 150 100 50

2.3. Three-dimensional ultrasound measures 0

A custom-built three-dimensional ultrasound (3DUS) system was used to measure the starting position of the humeral head centroid and the tip of the coracoid prior to measurement of the MRT in each case. Two series of perpendicular scans of the humeral head were made. In post-processing, the head of the humerus was manually segmented in each image and the 3D anatomy reconstructed. A sphere was then fitted to the reconstructed head of the humerus and the centroid of this sphere taken to be the centroid of the head of the humerus (Fig. 3). Prior calibration experiments in our laboratory using this method on a cadaveric humerus demonstrated the root mean square (RMS) accuracy of the 3DUS to be less than 2 mm in identifying the humeral head centroid. The tip of the coracoid was also identified. The coracoid tip and humeral centroid co-ordinates were mathematically inserted as virtual markers into the data set and their movement measured during the subsequent MRT manoeuvre (Fig. 4).

-100

-50

0

50

100

150

200

250

300

Y axis (mm)

Fig. 4. A ZY stick figure view of the movement of a left humeral head in the scapular co-ordinate system during one MRT manoeuvre is shown. The scapula is red and the forearm is brown with the path taken by the forearm markers shown in magenta. The track of the palpated humeral head point is shown in blue and the path of the centroid of the humerus derived from the 3DUS is black. The data is rotated about the Y-axis in order to show the scapula stick figure clearly.

2.4. Analysis The primary comparison between the 3DUS and motion analysis data was the total translation of the coracoid and head of humerus. In addition, the accuracy of motion tracking of the coracoid was measured by calculating the Euclidian distance (non-vectorial measure of distance) between the palpating finger and

ARTICLE IN PRESS D. Morrissey et al. / Manual Therapy 13 (2008) 529–535

532

relevant bony feature at each time epoch of data collection (Fig. 4). The total data set and all subsets were shown to be normally distributed using the Kolmogorov–Smirnov test and P–P plots (cumulative expected probabilitycumulative observed probability). Scatter plots for pair-wise variables were generated and examined in order to detect any non-linear relationships. Pearson correlation tests were used to calculate agreement, and paired t-tests used to test for differences, between data sets. Linear regression models were used to describe the predictive power of palpation for measurement of a given movement. Potential covariates and confounding factors were assessed using the enter method of stepwise regression. All results were calculated using SPSS 12.0.1 for windows (SPSS, Chicago, USA) with the level of significance being set at po0.05.

3. Results 3.1. Scapular measures 3.1.1. Coracoid position detection and tracking The Pearson test of correlation showed the agreement between the palpation measures of total coracoid movement and the 3DUS measures to be high (r ¼ 0.83, po0.01) (Fig. 5). A paired t-test confirmed that the palpation measures of total coracoid movement were not significantly different from the 3DUS measures (p ¼ 0.41). The mean (SD) 3DUS total coracoid anterior translation measure was 17 mm (10) and the

mean palpation total coracoid anterior translation measure was 16 mm (8). Linear regression showed the predictive ability of palpation for determining the true (3DUS) value of coracoid movement to be high with the palpation accounting for approximately 70% (R2 ¼ 0.69) of the variance in the model. The addition of data concerning arm length, height, body mass, age, and range of movement did not improve the predictive value of the model. The formula for measurement of true translation at the scapulo-thoracic articulation was derived from the model and is: Movementtrue ¼ 1.002(Movementpalpated)+0.143. One outlier was removed from the data used to calculate the formula as the residual was found to be greater than three SDs greater than the mean residual. Confidence intervals for the linear regression model were found to be nonuniform as is usually the case with this type of analysis (Sincich, 1993). The intervals are relevant within the range of measured values (5–43 mm) and were 71 mm at the mean value of measured movement (16 mm) and 75 mm at 1 SD above and below the mean (10 and 21 mm, respectively). The consistency of landmark tracking during movement was described by the SD of the Euclidian distance between the palpation and 3DUS-derived co-ordinates at each data collection epoch (Fig. 6). Sampling was at 200 Hz for approximately 3 s for each movement. The mean Euclidian distance between the palpating finger 40

35 50

Series 1 Series 2

30

25 30

mm

Palpation measure (mm)

40

20

20

15

10

10

5

0 0

10

20

30

40

50

3DUS measure (mm)

Fig. 5. A scatter plot of the palpation and 3DUS measures of total movement of the coracoid process for 22 MRT movements are presented. Where two measures were close together on the graph, the icons are shown with a vertical slash through them. The line of best fit is shown.

0

Fig. 6. The mean Euclidian difference in position detection of the coracoid process between the 3DUS and palpation methodologies across each collected time epoch in each of 22 MRT movements is shown in series 1. Series 2 shows the mean SD of this position detection for the 22 movements. The error bars represent the SD of each mean.

ARTICLE IN PRESS D. Morrissey et al. / Manual Therapy 13 (2008) 529–535

and the 3DUS measure of position was 26.9 (6.8) (range 13.1–38.8 mm). The mean (SD) of this position was 3.6 mm (1.4) (range 1.2–5.9). Once the thickness of soft tissue over the coracoid and the height of the marker on the finger have been taken into account these values indicate that the coracoid process was clearly identified and accurately tracked in each subject. A further indication of the accuracy of the palpatory tracking was given by calculating the correlation between the individual co-ordinates of the palpating finger marker and 3DUS-derived virtual markers rather than the Euclidian distance. The correlation in the X-axis was found to be 0.88, Y-axis 0.85 and Z-axis 0.82 demonstrating consistent tracking in all planes. 3.1.2. Humeral measures The Pearson test showed the correlation between the palpation measures of anterior translation and the 3DUS measures (r ¼ 0.65, po0.01) to be moderate (Fig. 7). The mean 3DUS anterior translation measure was 11 mm (SD ¼ 5) and the mean palpation measure was 8 mm (SD ¼ 3). A paired t-test showed that the palpation measures of humeral movement were significantly smaller than those derived from the 3DUS (p ¼ 0.03). This showed that palpation systematically underestimated the degree of humeral anterior translation. Linear regression showed the predictive ability of palpation for determining the true (3DUS) value of humeral centroid movement was moderate with the palpation accounting for just over 40% (R2 ¼ 0.42) of the variance in the model. Addition of arm length, height, body mass, age and range of movement did not improve the predictive value of the model. The formula for prediction of true translation at the gleno-humeral

Palpation measure (mm)

30

20

10

0

0

10

20

30

3DUS measure (mm)

Fig. 7. The scatter plot of the modelled palpation and 3DUS measures of anterior movement of the head of the humerus is shown for 22 MRT movements in 11 subjects. The line of best fit is shown.

533

articulation was developed from the model: Movementtrue ¼ 1.059(Movementpalpated)+3.765. Confidence intervals for this model were calculated, as for the coracoid tracking data, being relevant within the range of measured values (0.98–15 mm). The 95% confidence intervals were 77.01 mm at the mean value of measured movement (7.9 mm) and 714.4 mm at 1 SD below and above the mean (6.2 and 10.6 mm, respectively).

4. Discussion 4.1. Scapular measures 4.1.1. Scapular position identification and motion tracking The coracoid process of the scapula can be difficult to palpate, particularly in larger subjects with more muscle, fat or breast tissue. Nonetheless, the coracoid process was identified to within 35 mm (mean ¼ 26.9 mm, SD ¼ 6.8 mm, Fig. 6) at the starting position of the MRT. Given that the skin, fat, muscle and fascial soft tissues overlying the coracoid were interposed between the palpating finger and the tip of the coracoid, the measured values were considered to be acceptably accurate. Placement of the marker on the fingers above the actual contact area of the finger added a further offset. Perhaps most importantly, the palpating finger stayed a consistent distance from the coracoid, as represented by the relatively low SD of positional error throughout movement in relation to the 3DUS-derived marker. The consistent distance may be an indication that the soft tissues did not change dimension significantly during the MRT, in contrast to those overlying the humerus, discussed below. Palpation showed high correlation with the 3DUSderived virtual marker on the coracoid both in the Euclidian distance (r ¼ 0.83) and in each individual coordinate (r ¼ 0.82–0.88). There was no systematic error in the measurement as demonstrated by the negative paired t-test. The correlation in Euclidian distance was the most relevant accuracy measure, as the aim in palpation is to follow the coracoid process irrespective of direction. The high correlations in all three axes for both movements of each subject were a further measure of accuracy indicating that no one co-ordinate was poorly tracked, nor that systematic error in one or more co-ordinates was leading to a false positive measure of total translation. The high accuracy of scapular motion tracking is a critical finding. The method employed during the MRT relies on accurate scapular motion tracking not only to describe scapular movement but also as a base from which humeral tracking occurs. Two planes of the position of the finger palpating the humerus are dictated by the position of the finger palpating the coracoid with

ARTICLE IN PRESS 534

D. Morrissey et al. / Manual Therapy 13 (2008) 529–535

only the anterior translation being independent. Had the scapular motion tracking been inaccurate, the humeral tracking would have been inherently compromised. A formula to describe the real movement of the scapula has been derived from the linear regression, with associated confidence intervals which could be of use in patient studies using this method and examiner. Other examiners and methods of palpating movement may yield other conversion formulae, but the correlation between measures in this experiment is evidence of the potential accuracy of palpation of this movement. 4.1.2. Actual values No in-vivo or in-vitro studies have previously measured scapular movement during the MRT manoeuvre or similar upper limb movements. Cadaveric studies typically immobilise the scapula in order to measure gleno-humeral movement (Abboud and Soslowsky, 2002). The values of Euclidian translation of the scapula were, however, comparable with those found during elevation in normal subjects where up to an average of 20 mm of anterior scapular translation was demonstrated (Barnett et al., 1999). These studies did not specifically measure the medial rotation movement. The values of scapular translation reported in this study are therefore the first reported for the movement of interest, and will serve as the basis for comparison with subsequent patient-based studies. 4.2. Humeral measures 4.2.1. Actual values and motion tracking The motion tracking of the humerus by palpation with comparison to the 3DUS-derived measures showed the humerus to be only moderately well tracked and systematically under-estimated. This was an unsurprising finding for two reasons: the complexity of the movement and the difficulty in tracking a single coordinate. Palpation showed a mean (SD) anterior humeral translation of 8 mm (3) during the MRT manoeuvre, while the 3DUS mean was 11 mm (5). These values of humeral translation with both methodologies are surprisingly high when the anatomy of the joint is considered. The mean anterior–posterior radius of the adult glenoid, not including the labrum, is 12 mm so a translation of 11 mm from a centred resting position could potentially sublux the joint (De Wilde et al., 2004). Werner et al. (2004) found 12 mm (SD 8 mm) of obligate humeral head translation associated with medial rotation when the shoulder is abducted 901 in the scapular plane in cadavers, similar to the results in the present study. This comparison must be made with caution as no muscle forces were included in the model, significant end of range forces were applied and a previous cadaveric study has shown 1 mm (2.4) of anterior

translation at 01 of abduction as compared to 6 mm (4) at 01 in Werner’s study (Harryman, 1990). It is likely that the measured translation is accurate. The anterior translation that occurs is due to a combination of ligamentous and muscle forces (Itoi et al., 2004) from a relaxed starting position in which the scapula and distal humerus are supported, whereas the humeral head is not. Gravity may have resulted in an initial posteriorly translated starting position from which a greater than expected anterior translation occurred. In addition, the slow anti-gravity movement may have resulted in selective activation of the shoulder external rotators acting eccentrically therefore minimising the stabilising function of the anterior rotator cuff and maximising anterior translation. Teres minor and infraspinatus have been shown to exert a significant anterior shear force that is balanced by a posterior shear generated by subscapularis and supraspinatus in 901 of abduction (Lee et al., 2000). The humeral translation measured could therefore be a reflection of the passive posterior and dynamic anterior translation available in 901 of abduction in the scapular plane, produced by the static stabilisers and dynamic decelerators of the movement. The geometric centre of the head of the humerus was used as the humeral centre of rotation in this study. Veeger (2000) showed this to be an accurate assumption in a study of four cadavers where the geometric and kinematic axes of movement were shown to coincide, in a range of movements including medial rotation using the same co-ordinate systems as this study. The kinematic centre of the humerus has not been defined in-vivo although the neutral zone, flexibility and ROM have been kinematically measured (Novotny et al., 2000). The definition of the humeral geometric centre has been shown to be accurate using this 3DUS system with cadaveric phantoms used for development. The accuracy of the 3DUS in describing the geometric centroid of the humerus has not, however, been fully established in-vivo. It may be that differences in the speed of ultrasound in different tissues and sampling of the head of the humerus including the greater tuberosity of the humerus led to systematic offset of the measured geometric centroid from the actual centroid therefore amplifying the measured translation. Palpation may also have shown a systematic over-estimation as the greater tuberosity moves to a position under the palpating finger during the MRT manoeuvre. Changes in the width of the soft tissue mass over the greater tuberosity during the test manoeuvre was not thought to be a significant factor as the tuberosity is first covered by the anterior deltoid and then the middle deltoid during the manoeuvre. The thickness of these muscles may however differ between individuals, and may vary as the tissues stretch during the movement. Further error may be attributable to the difficulty in picking out one co-ordinate of the humeral movement

ARTICLE IN PRESS D. Morrissey et al. / Manual Therapy 13 (2008) 529–535

to track, with the other two co-ordinates being dictated by the scapular movement. The humeral translation values derived from the reported formula will need to be interpreted with some caution in further studies, as the confidence intervals for the formula were large. The possibility of modelling the kinematic centre of the movement was explored. Numerous methods including regression analysis, spherical fitting of marker trajectories and helical axes have been compared in elevation and shown to generate significantly different estimates of the instantaneous axis of rotation (IAR) (Stokdijk et al., 2000). These methods assume negligible translation if a single IAR is to be defined, clearly not the case in this study. This problem has been surmounted by using the IAR for sequential portions of a total range of movement in studies of elevation (Ludewig and Cook, 2002; Ludewig et al., 2002). These methods were not possible in this study due to methodological problems associated with using humeral wands attached to the distal humerus to provide attachment sites for additional markers, as validated previously for elevation (Ludewig et al., 2002). In our study, it was noted that the wand trajectory did not closely follow that of the humerus leading to under- or over-estimation of the total movement. This appeared to be particularly true in subjects with looser soft tissues, as subjectively detected by the examiner. Future studies assessing the kinematic IAR would require further validation of the use of humeral wands during rotational movements. It could be argued that the accuracy and development work presented in this paper should have been done in subjects with pathology as they might, potentially, move differently from normal subjects, with different variability as has been shown in measuring laxity at the knee (Anderson et al., 1992). This argument would be based on structural derangement and pain affecting movement. The argument was judged to be minimally relevant in this particular study for two reasons. Firstly, the aim of this study was specifically to measure the accuracy of palpation in tracking movement, rather than to assess test–retest or inter-rater reliability— repeated movements being more affected by intermovement variability. Secondly, scapular and humeral movements are inherently complex and non-uniform between subjects, even in a normal shoulder, so the accuracy of motion tracking demonstrated in this study is likely to be generalisable to patients, although this would be a reasonable avenue for further research.

5. Conclusion This study has shown that the coracoid can be accurately palpated and tracked using palpation during the MRT manoeuvre (by comparison with 3DUS as a

535

gold standard), but that the movement is slightly underestimated. The palpation is accurate both in terms of Euclidian distance and in three dimensions. Anterior translation of the humerus was significantly underestimated in a systematic fashion by palpation, with accurate information about the degree of movement occurring therefore obtainable by application of a conversion formula derived from linear regression.

References Abboud JA, Soslowsky LJ. Interplay of the static and dynamic restraints in glenohumeral instability. Clinical Orthopaedics and Related Research 2002;400:48–57. Anderson AF, Snyder RB, Federspiel CF, Lipscomb AB, Grana WA. Instrumented evaluation of knee laxity—a comparison of 5 arthrometers. American Journal of Sports Medicine 1992; 20(2):135–40. Barnett ND, Duncan RDD, Johnson GR. The measurement of threedimensional scapulohumeral kinematics—a study of reliability. Clinical Biomechanics 1999;14(4):287–90. Comerford MJ, Mottram SL. Movement and stability dysfunctioncontemporary developments. Manual Therapy 2001;6(1):15–26. De Wilde LF, Berghs BM, Audenaert E, Sys G, Van Maele GO, Barbaix E. About the variability of the shape of the glenoid cavity. Surgical and Radiologic Anatomy 2004;26(1):54–9. Harryman DT. Translation of the humeral head on the glenoid with passive gleno-humeral motion. The Journal of Bone and Joint Surgery 1990;72A(9):1334–43. Itoi E, Morrey BF, An KN. Biomechanics of the Shoulder. In: Rockwood AC, Matsen AF, editors. The shoulder. 3rd ed., vol. 1. WB Saunders; 2004. Lee SB, Kim KJ, O’Driscoll SW, Morrey BF, An KN. Dynamic glenohumeral stability provided by the rotator cuff muscles in the mid-range and end-range of motion—a study in Cadavera. Journal of Bone and Joint Surgery—American Volume 2000;82A(6): 849–57. Ludewig PM, Cook TA. Translations of the humerus in persons with shoulder impingement symptoms. Journal of Orthopaedic and Sports Physical Therapy 2002;32(6):248–59. Ludewig PA, Cook TM, Shields RK. Comparison of surface sensor and bone-fixed measurement of humeral motion. Journal of Applied Biomechanics 2002;18(2):163–70. Novotny JE, Woolley CT, Nichols CE, Beynnon BD. In vivo technique to quantify the internal-external rotation kinematics of the human glenohumeral joint. Journal of Orthopaedic Research 2000;18(2):190–4. Sincich P. Simple Linear Regression and Correlation. In: Statistics by example, 5th ed., 1993. Stokdijk M, Nagels J, Rozing PM. The glenohumeral joint rotation centre in vivo. Journal of Biomechanics 2000;33(12):1629–36. van der Helm FCT. A standardized protocol for the description of shoulder motions. International Shoulder Group. Ref Type: Electronic Citation, 2002. Veeger HEJ. The position of the rotation center of the glenohumeral joint. Journal of Biomechanics 2000;33(12):1711–5. Werner CML, Nyffeler RW, Jacob HAC, Gerber C. The effect of capsular tightening on humeral head translations. Journal of Orthopaedic Research 2004;22(1):194–201. Wu G, van der Helm FCT, Veeger HEJ, Makhsous M, Van Roy P, Anglin C, et al. ISB recommendation on definitions of joint coordinate systems of various joints for the reporting of human joint motion – Part II: Shoulder, elbow, wrist and hand. Journal of Biomechanics 2005;38(5):981–92.

ARTICLE IN PRESS

Manual Therapy 13 (2008) 536–543 www.elsevier.com/math

Original article

Longitudinal sliding of the median nerve in patients with non-specific arm pain Andrew Dilley, Samuel Odeyinde, Jane Greening, Bruce Lynn Department of Physiology, University College London, Gower Street, London WC1E 6BT, UK Received 13 September 2006; received in revised form 8 June 2007; accepted 15 July 2007

Abstract In patients with non-specific arm pain (NSAP; also known as repetitive strain injury), there are clinical signs of altered median nerve sliding. It is possible that a restriction along the nerve course will lead to abnormal increases in local strain during limb movements, possibly contributing to symptoms. The present study uses ultrasound imaging to examine median nerve sliding through the proximal and distal nerve segments in 18 NSAP patients. Longitudinal nerve sliding was measured during metacarpophalangeal, wrist and elbow movements. During elbow movements, the angle of elbow extension at which the nerve begins to move was determined, since this was expected to decrease with a restriction through the shoulder. The results from this study were compared with previously reported data. Nerve movements ranged from 1.26 to 4.73 mm in patients compared with 1.43–5.57 mm in controls. There was no significant difference in nerve sliding (p40.05) or in the angle of elbow extension at which the nerve began to move (mean ¼ 53.41 in patients, 52.01 in controls; p40.05). In summary, restriction of median nerve sliding is unlikely to play a major role in NSAP. Therefore, painful responses during limb movements which tension the nerve are unlikely to result from abnormal increases in nerve strain. r 2007 Elsevier Ltd. All rights reserved. Keywords: Non-specific arm pain; Repetitive strain injury; Median nerve; Ultrasound

1. Introduction The diffuse painful upper limb condition termed nonspecific arm pain (NSAP; also known as repetitive strain injury) is a significant occupational problem for those people who perform repetitive tasks, for example keyboard operators, musicians and production line workers (Bernard et al., 1994; Macfarlane et al., 2000; Greening et al., 2001). NSAP is a diagnosis made following the exclusion of other specific conditions, for example carpal tunnel syndrome and tenosynovitis (Harrington et al., 1998). Many of the symptoms Corresponding author. Anesthesia, Critical Care, and Pain Management, Beth Israel Deaconess Medical Center, Harvard Medical School, 330 Brookline Ave., Dana 717, Boston, MA 02215, USA. Tel.: +1 617 667 4965; fax: +1 617 667 1500. E-mail address: [email protected] (A. Dilley).

1356-689X/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2007.07.004

experienced by NSAP patients can be attributed to altered function of the median nerve. For example, changes in both small and large fibre function, particularly affecting the median and ulnar nerve, have been previously reported (Greening et al., 2003). Limb movements are often painful, especially when the affected limb is positioned in postures that tension the brachial plexus and the median nerve (for example, the upper limb tension test 1 (ULTT1); Byng 1997; Greening et al., 2001). Such painful responses might indicate altered longitudinal sliding of the median nerve due to a restriction along its course (Lynn et al., 2002; Greening et al., 2005). If a nerve is prevented from sliding freely during joint movements, then the section of nerve closest to the moving joint will have to stretch more in order to accommodate the change in bed length. It is anticipated that these increases in stretch may be sufficient to cause altered nerve function, contributing

ARTICLE IN PRESS A. Dilley et al. / Manual Therapy 13 (2008) 536–543

towards painful symptoms (McLellan and Swash, 1976; Wilgis and Murphy, 1986). Painful responses to tests that stretch the median nerve, and to direct digital pressure over the affected nerve trunk (Greening et al., 2005), also suggest that neural mechanosensitivity may play a role in symptom production. It has been suggested that some patients with NSAP have a minor nerve injury and that there may be multiple sites of median nerve pathology (Greening and Lynn, 1998). The most likely sites are at the wrist where the median nerve passes through the carpal tunnel, in the forearm under the fibrous arch of flexor digitorum superficialis muscle, between the two heads of the pronator teres muscle, and where the medial cord of the plexus passes over the first rib at the thoracic outlet (Johnson et al., 1979; Wertsch and Melvin, 1982; Bilecenoglu et al., 2005). These sites are where the nerve passes through narrow spaces, and therefore could become compromised by the structures surrounding it. Magnetic resonance imaging (MRI) and ultrasound studies on NSAP patients have previously demonstrated reduced transverse median nerve sliding at the wrist during wrist movements, possibly indicating distal changes to nerve biomechanics (Greening et al., 1999, 2001). Shoulder protraction has also been shown to alter longitudinal median nerve sliding through the shoulder region (Julius et al., 2004), which is consistent with the high incidence of poor shoulder posture in NSAP patients (Pascarelli and Hsu, 2001). The present study is a preliminary examination of longitudinal median nerve sliding in patients with NSAP. Using methods previously described (Dilley et al., 2001, 2003), longitudinal median nerve sliding was examined using ultrasound imaging in the forearm and upper arm in response to metacarpophalangeal (MCP) joint and wrist movements, to determine whether the nerve is restricted from sliding freely through the elbow region, forearm and carpal tunnel. Nerve sliding was also examined in the forearm during elbow extension, to assess nerve movement through the shoulder region. It has previously been reported that the median nerve is slack when the elbow is flexed to 901, and it is not until the elbow is extended to approximately 451 flexion that the nerve straightens, and therefore begins to slide in the forearm (see Dilley et al., 2003). It was suggested in this report that when the elbow is flexed, most of the slack lies within the proximal nerve segment. It is anticipated that a restriction within the shoulder region is therefore likely to affect the angle of elbow flexion at which the nerve begins to slide in the forearm. In the present study, we have looked specifically at the angle of elbow flexion at which the nerve begins to move within the forearm in NSAP patients and controls. In summary, the aim of this study is to determine whether longitudinal median nerve sliding is altered through the wrist, forearm and shoulder region in patients with NSAP.

537

2. Methods 2.1. Patient criteria Eighteen NSAP patients (13 female and 5 male, mean age ¼ 36.9 years, S.D. 9.5 years) and 39 healthy controls (23 female and 16 male, mean age ¼ 34.1 years, S.D. 11.7 years) were included in the study. All the patients were office workers who attributed their symptoms to prolonged and intensive keyboard use. Patients were recruited through physiotherapy clinics and the London RSI Support Group. Each patient was screened to ensure they conformed to the diagnostic criteria of NSAP (Harrington et al., 1998) and were not suffering from any other upper limb pathology. Symptoms were recorded on a standardized body chart, which was used to determine the most affected side. A test designed to tension the brachial plexus (the upper limb tension test (ULTT1) Butler 1991), was also used to assess median nerve involvement. The ULTT1 consists of 901 shoulder abduction, lateral rotation of the glenohumeral joint, forearm supination, wrist extension and finally extension of the elbow from 901 flexion. This test was positive (restricted joint range and reproduction of the patient’s symptoms) on the most affected side in all patients. Control subjects were screened to exclude any cervical spine or upper limb pathology, or a previous history of NSAP or cervical whiplash injury. Controls were also excluded if they used display equipment for more than 40% of their working week and had a positive ULTT1. 2.2. Ultrasound imaging Ultrasound imaging was carried out as previously described by Dilley et al. (2001, 2003). A Diasuss Ultrasound system (Dynamic Imaging, Livingston, Scotland, UK) was employed with a 10–22 MHz, 26 mm linear array transducer. Sequences of images were obtained at 10 frames/s, converted to digital format, and analysed offline using software developed in Matlabs. The image resolution was 0.093 mm/pixel with an image size of 280  440 pixels. Offline analysis employed a cross-correlation algorithm to determine relative movement between adjacent frames in sequences of the US images (see Dilley et al., 2001). This method is reliable to 0.1 mm (Dilley et al., 2001). To control for probe movement, deep static structures, e.g., bony features or the interosseous membrane, were also tracked using the same methods. Any movement of these structures was subtracted from the nerve excursion values to give the best estimate of the nerve movement in millimetres. The arm length was measured in each subject to normalise the position of the transducer location. With the arm by the side and the forearm supinated, the arm was measured from the C6 spinous process to the tip of

ARTICLE IN PRESS 538

A. Dilley et al. / Manual Therapy 13 (2008) 536–543

the index finger following the approximate course of the median nerve. 2.3. Examination of distal longitudinal nerve sliding In 11 NSAP patients, either wrist (n ¼ 5) or MCP (n ¼ 6) extension were used to determine whether median nerve sliding was altered through the wrist and forearm. Nerve excursion values in NSAP patients were compared with control data (also published in Dilley et al., 2003; Erel et al., 2003). 2.3.1. Wrist extension For wrist extension, imaging was performed on the anterior surface of the distal third of the forearm and in the upper arm, proximal to the elbow. Imaging locations were similar between patients and controls (forearm location ¼ 64–71% and upper arm location ¼ 45–53% of total arm length from C6 to the digits). All of the NSAP patients (n ¼ 5) and 6 of 9 control subjects were imaged in both the upper arm and forearm. Subjects were positioned supine with their shoulder abducted to 451 (Fig. 1a), an angle previously shown to tension the median nerve (Dilley et al., 2003). The forearm was supported on a Perspex plate with the elbow straight. The digits and MCP joints were maintained in extension using a plate attached to the palmar aspect of the hand. Velcro strapping was used to attach supporting plates to the limb. All joint angles were checked using a goniometer. The wrist was passively moved from neutral to 401 extension during imaging. Each movement lasted 2–3 s and was repeated 3 times. The results from 2–3 trials were averaged for each subject.

2.3.2. MCP extension For MCP extension, imaging was performed at similar locations on the anterior surface of the distal third of the forearm (5–15 cm proximal from the distal wrist crease). Subjects (n ¼ 6 NSAP patients and 19 controls) were positioned supine with their shoulder abducted to 901 (Fig. 1b). The forearm was supported on a Perspex plate with the elbow and wrist straight. The interphalangeal joints were maintained in extension using a plate strapped to the palmar aspect of the digits. The MCP joints were passively moved from 901 flexion to neutral during imaging. Each movement lasted 2–3 s and was repeated 3 times. 2.4. Examination of proximal longitudinal nerve sliding Elbow extension was used to determine whether nerve sliding was altered through the shoulder region in NSAP patients (n ¼ 8) compared with controls (n ¼ 13). Imaging of the median nerve was performed on the anterior surface of the mid-forearm. Subjects were positioned supine with the shoulder abducted to 901 (Fig. 1c). The upper arm was supported on a Perspex plate, and the forearm and hand were supported on a separate plate with wrist and digits straight. The elbow was passively extended from 901 flexion to neutral (01 of flexion) during imaging. A 10-turn potentiometer attached to the forearm plate by means of a pulley system allowed measurement of the angle of extension. Potentiometer resistance was logged continuously (10/s) to a computer using PowerLabs (ADInstruments, Colorado, US). The relation between resistance and elbow angle was determined in a separate calibration using a goniometer to measure elbow angle. A minimum

Fig. 1. Diagram illustrating the upper limb posture, transducer position and joint movement for the (a) wrist, (b) MCP and (c) elbow extension groups. The circles on each limb illustration represent the shoulder, elbow and wrist joints (labelled on (a)). The angle of shoulder abduction is also given.

ARTICLE IN PRESS A. Dilley et al. / Manual Therapy 13 (2008) 536–543

of three trials were performed on each subject. In some subjects, it was necessary to perform up to 5 trials in order to obtain enough (2–3) sequences of sufficient image quality for the cross-correlation analysis. At the end of each series of repeats, the potentiometer was calibrated to convert the voltage output into degrees. The elbow angle corresponding to the start of nerve movement was determined for each trial. The time point at the onset of nerve movement was obtained from either a velocity or cumulative nerve movement profile, and the elbow angle corresponding to that time was determined from the potentiometer data. As confirmation, the cumulative nerve movement data were plotted against the elbow angle, and the start angle was determined from the graph (Fig. 2). Due to the significant length of time required to perform each procedure (wrist, MCP and elbow extension) and the short time period that some patients could tolerate the test position, only one procedure was performed at any one session. Consequently, the sample size varied between each group. 2.5. Statistical methods Comparisons between mean values were carried out using unpaired t tests. Comparisons between nerve movements during wrist extension were also carried out using analysis of covariance, with the transducer position along the arm as the co-variate. Lower 5% confidence limits for the differences in nerve sliding were determined for pooled data in the wrist and MCP group and expressed as a percent of the control mean. These values were used to give an indication of the sensitivity of the method, i.e. the smallest percentage reduction that could be reliably detected with the present data. A regression analysis was performed between nerve excursion during elbow movements and subject age. Standard errors have been given.

539

3. Results 3.1. Examination of distal longitudinal nerve sliding Demographic data for each group are shown in Table 1. 3.1.1. Wrist extension In all subjects (n ¼ 5 patients and 9 controls), the median nerve moved distally during 401 wrist extension. There was no significant difference in the amount of excursion between the patient and control group (p ¼ 0.21, analysis of covariance) (Fig. 3a). In the upper arm, the mean nerve excursion was 2.00 mm (0.25 SEM; range 1.43–2.74) in patients and 2.31 mm (0.31 SEM; range 1.43–3.30) in controls (p ¼ 0.47). In the distal forearm, the nerve moved 3.97 mm (0.33 SEM; range 3.06–4.73) in patients and 4.40 mm (0.20 SEM; range 3.43–5.57) in controls (p ¼ 0.26) (Fig. 3b). The lower 5% confidence limits for the differences in sliding were 1.23 mm in the upper arm and 1.21 mm in the forearm, equivalent to 53% and 28% below the control group mean, respectively. Although there have not been any reported differences in nerve sliding between males and females, since all the patients were females, the patient group was also compared with the females in the control group. The mean nerve excursion in the female only control group was 1.69 mm (0.19 SEM; range 1.43–2.05) in the upper arm (n ¼ 3) and 4.18 mm (0.20 SEM; range 3.43–4.94) in the forearm (n ¼ 6), which were not significantly different compared with the patients (p ¼ 0.43 and 0.58 for the upper arm and forearm, respectively). 3.1.2. MCP movements In all subjects (n ¼ 6 patients and 19 controls) the nerve moved distally during movement of the MCP joints from 901 flexion to neutral. The mean nerve Table 1 Demographic data Movement

Group

Wrist extension

NSAP

MCP extension

% females

Mean age (SD) range

5

100%

Controls

9

67%

33.8 (8.8) 24–48 33.4 (11.1) 21–48

NSAP

6

100%

19

68%

8

38%

13

50%

Controls Elbow extension Fig. 2. A typical graph from a control subject showing the cumulative nerve movement plotted against the elbow angle. The nerve started to move at 641 flexion.

n

NSAP Controls

36 (10.9) 24–55 41.3 (9.9) 26–58 38.9 (9.1) 32–53 24.6 (6.8) 18–40

ARTICLE IN PRESS 540

A. Dilley et al. / Manual Therapy 13 (2008) 536–543

Fig. 4. (a) Longitudinal median nerve movement in the forearm during elbow extension from 901 flexion to straight. The graph shows the average of 2–3 repeat trials for each subject (+). The mean is represented by the horizontal line within each group. Note the single outlier in the patient group (nerve movement ¼ 1.26 mm). (b) The mean angle of extension at which the median nerve begins to slide in the forearm (n ¼ 13 patients and 8 controls). Note there is no clear difference between each group. Error bars ¼ SEM.

Fig. 3. Longitudinal median nerve movement in the forearm and upper arm of NSAP patients (n ¼ 5) and controls (n ¼ 9 in forearm and 6 in upper arm) during 401 wrist extension. (a) Nerve movement values for individual subjects showing the location along the arm. Each point is the average of 2–3 repeat trials. Distance along the arm has been expressed as percent of total distance from C6. (b) The same data showing the average nerve movement for the patient and control group in the forearm and upper arm. Although there is a small reduction in nerve sliding in the NSAP patient group compared with controls, this is not significant. Error bars ¼ SEM.

excursion was 2.68 mm (0.44 SEM; range 1.46–4.01) in patients compared with 2.62 mm (0.15 SEM; range 1.63–4.54) in control subjects, which was not significantly different (p ¼ 0.85). The lower 5% confidence limit for the differences in nerve sliding was 0.78 mm, equivalent to 30% below the control group mean. Since all of the patients were female, the patient group was again compared with the females in the control group. The mean nerve movement in the female control group (n ¼ 13) was 2.46 mm (0.13 SEM; range 1.63–3.53), which was not significantly different to the patient group (p ¼ 0.50). 3.2. Examination of proximal longitudinal nerve sliding In all subjects (n ¼ 8 patients and 13 controls) the nerve moved proximally in the forearm during elbow extension from 901 flexion to neutral. There was a

17.9% reduction in nerve sliding in the patient group (mean ¼ 3.34 mm (0.35 SEM; range 1.26–4.38; n ¼ 8)) compared with controls (mean ¼ 4.07 mm (0.20 SEM; range: 2.93–4.88; n ¼ 12)), which was close to significance (p ¼ 0.07). It would, however, seem that this trend was in part due to a single outlier in the patient group (see Fig. 4a). Despite the difference in ages between the patient and control group, there was only a weak correlation between nerve excursion and age (r ¼ 0.21 for the combined control and patient group without the outlier patient). Comparing patients and controls of a similar age range (controls: n ¼ 5; mean age ¼ 31.8 (5.4 SD) range 25–40; NSAP patients: n ¼ 6; mean age ¼ 34.5 (4.8 SD) range 28–41) resulted in less of a difference in nerve sliding between the groups (mean nerve sliding 4.07 mm (0.32 SEM) in controls compared with 3.56 mm (0.24 SEM) in patients; p ¼ 0.22). The angle at which the median nerve begins to slide in the forearm during elbow extension was determined in all subjects. There was no significant difference in the start angle between the patient group (n ¼ 13; mean angle ¼ 53.41 (2.2 SEM) range 39.4–67.91) and control groups (n ¼ 8; mean angle ¼ 52.01 (4.8 SEM) range 36.2–66.51) (p ¼ 0.77) (Fig. 4b).

4. Discussion There is much uncertainty as to whether restriction of longitudinal sliding of peripheral nerves contributes to symptoms in patients with diffuse upper limb pain. The most compelling evidence to suggest that restricted nerve sliding may play a role is that joint movements which tension peripheral nerves are often painful in

ARTICLE IN PRESS A. Dilley et al. / Manual Therapy 13 (2008) 536–543

these patients. For inclusion, all the patients in the present study had painful responses when their most affected limb was positioned in a posture which tensioned the brachial plexus/median nerve (the ULTT1). However, these preliminary results for direct imaging of nerve sliding indicate that during limb movements, median nerve sliding in NSAP patients was not significantly different from normal control subjects, i.e. there was no evidence for a restriction. Imaging in the forearm and upper arm during wrist movements, and in the forearm during MCP and elbow movements, allowed nerve sliding to be studied at the most likely sites of restriction, namely the carpal tunnel, as the nerve passes between the two heads of the pronator teres muscle and through the shoulder region. In 11 patients, either wrist or MCP movements were used to assess whether there was a restriction within the carpal tunnel. Nerve movements in response to MCP extension were expected to be reduced in the forearm in the presence of a restriction. The expected change in the forearm with wrist extension was, however, not so obvious. If there was a restriction over the centre of rotation of the wrist joint, then there would be a tendency for nerve sliding to be increased in the forearm with wrist extension, since the nerve would be stretching over a shorter length (from the wrist to the spine rather than the digits to spine). If the restriction was proximal to the centre of rotation, then nerve sliding would decrease in the forearm. The present findings showed no clear difference in nerve sliding in the forearm between patients and control subjects in either the MCP or wrist group, suggesting that the median nerve was probably not restricted at the wrist. It is, however, possible that variability between subjects may mask small trends. Therefore, the lower 5% confidence limits for the differences in sliding were calculated as an indicator of the sensitivity of the method (the smallest percent reduction in sliding that could be reliably detected). These values suggested that the variability of the data was sufficiently reliable to exclude changes in nerve sliding greater than 30%. In a small number of patients, the median nerve was also imaged in the upper arm during wrist extension to determine whether there was a restriction through the forearm. However, the lack of a significant movement reduction in the patient group did not indicate any such restriction. NSAP patients were also examined for restricted median nerve sliding through the shoulder region. Since it was not possible to identify the nerve clearly by ultrasound within the shoulder, an indirect approach was used. Based on previous findings (Dilley et al., 2003), it was anticipated that a restriction may compromise the available slack through the shoulder region during elbow extension from 901 flexion, resulting in an earlier nerve movement in the forearm.

541

Cumulative nerve movement would also be increased because the nerve would be moving over a longer part of the elbow movement. There was, however, no significant difference in the onset of nerve sliding or the total nerve movement in the patient group compared with controls, suggesting that the median nerve is not restricted through the shoulder in these patients. Nerve movement in fact decreased in the patient group (close to significance), mainly due to a single outlier. The reduced nerve excursion value of this outlier subject corresponded to their late onset of nerve movement (361 flexion). In this group there was a notable difference in ages between patients and controls. However, there was no clear correlation between age and nerve movement. Age matching of the patients to the controls did not reveal any trends. The present results indicate that longitudinal nerve movements in patients with NSAP are comparable with those in normal individuals. Although small differences in nerve sliding (e.g. o30% in the forearm with wrist/ MCP movements) cannot be excluded, such differences would have only a small effect on nerve strain. It would, therefore, seem unlikely that increases in nerve strain due to lengthening of the median nerve over a shorter segment will contribute to painful symptoms in the patients examined. Since the sample size was relatively small and the diagnosis of NSAP is made by exclusion (Harrington et al., 1998), restricted median nerve sliding cannot be ruled out as a cause of symptoms in other sub-groups of NSAP patients with distinct aetiologies. However, it should be noted that even in carpal tunnel syndrome, where entrapment of the median nerve is thought to be part of syndrome aetiology, changes in longitudinal sliding were not seen in vivo (Erel et al., 2003). The lack of difference in nerve sliding between patients and control subjects is consistent with our previous results in carpal tunnel syndrome patients (Erel et al., 2003). Inherent strains in the median nerve may reach 3–4% in the forearm during physiological movements that maximally extend the nerve bed length (Dilley et al., 2003). Even in this extended position, in the absence of a nerve restriction, forearm strains are still below the levels that can alter nerve function (46%; Grewal et al., 1996). Close to moving joints, where median nerve strain may be at its highest (Dilley et al., 2003), the inherent strain is still likely to be below the level that can affect nerve function. In fact, 4% represents the maximum strain in the forearm closest to the wrist joint, with the wrist extended (Dilley et al., 2003). Therefore, an alternative hypothesis must be considered. Increases in nerve trunk mechanical sensitivity could explain why limb movements that tension the median nerve are frequently painful in patients (Dilley et al., 2005). The induction of a local neuritis in the rat can cause nerve fibres in continuity to become

ARTICLE IN PRESS 542

A. Dilley et al. / Manual Therapy 13 (2008) 536–543

mechanically sensitive, responding to pressure and small stretches (Eliav et al., 2001; Bove et al., 2003; Dilley et al., 2005). Therefore, in patients with diffuse limb pain, relatively minor inflammation of the nerve trunk or adjacent tissue may be sufficient to cause local increases in axonal mechanical sensitivity. Consistent with the high incidence of NSAP in workers who perform repetitive tasks (e.g. keyboard operators), low-force highly repetitive movements can trigger an immune response within the median nerves and tendons of rats (Barbe et al., 2003; Clark et al., 2003; Al-Shatti et al., 2005). Interestingly, in the animal model of axonal mechanical sensitivity, the most sensitive fibres (C and A) responded to approximately 3% stretch (Dilley et al., 2005), which is similar to the range of median nerve strain that occurs during normal physiological movements in the upper limb. In patients with neuropathic pain, central sensitization is considered to play a substantial role in symptom production (reviewed in Campbell and Meyer, 2006; see also Campbell et al., 1988). Central sensitization is, however, dependent upon the maintenance of ongoing input from the periphery (LaMotte et al., 1991; Gracely et al., 1992). Therefore, even in patients where central processes seem to be the major cause of symptoms, a peripheral component, however, minor, is still likely to persist. Therefore, painful symptoms associated with mechanical stimulation of affected nerves are probably due to the combined effects of central and peripheral mechanisms. Nerve fibres supplying the nerve sheath (nervi nervorum; Bove and Light, 1995) and which respond to mechanical stimulation may also contribute to symptoms in patients with NSAP. The majority of these fibres also have receptive fields in muscle (Bove and Light, 1995), which is consistent with the pattern of symptom production during nerve stretch tests (i.e. typically painful responses to nerve stretch tests are referred to deep tissues rather than localised over the nerve). However, contrary to this mechanism, nervi nervorum fibres seem to be a small minority, and mechanical stimulation of untreated rat sciatic nerve has failed to find such units (Dilley et al., 2005). Subtle changes to nerve sliding have previously been reported in NSAP patients. Our previous studies have shown a reduction in transverse nerve sliding at the wrist during wrist flexion (Greening et al., 1999, 2001). This lack of transverse sliding means that the median nerve is vulnerable to compression from the flexor tendons as they move anteriorly during wrist and finger flexion. We have also demonstrated a reduction in longitudinal sliding in the forearm during deep inspiration (Greening et al., 2005), a result which probably indicates a reduction in first rib excursion. These findings do not provide evidence of longitudinal nerve restriction, since these changes will have negligible effect on overall

strain. Rather, they represent multiple sites of possible change in the nerve environment, situations that may lead to localised nerve inflammation. The present findings are important for the clinician. In NSAP patients when links have been made between diagnosis and presumed changed neural mechanics, attempts to normalise nerve movement may be ineffective. Such techniques are likely to contribute to an increase in the patient’s symptoms in the presence of nerve trunk mechanosensitivity. From the results in this study, a more effective treatment strategy may be one that is focussed on reducing nerve irritation from adjacent tissue interfaces (e.g. tissues around the thoracic outlet). The restoration of correct postural alignment and movement control plus resolution of any local inflammation is likely to be beneficial in the ‘‘desensitisation’’ of hyperalgesic neural tissue. In summary, median nerve restriction is unlikely to play a major role in the NSAP patients examined. MCP, wrist and elbow extension were used to examine both proximal and distal nerve segments for signs of a restriction. There were no clear signs of reduced nerve sliding in these patients and the methods were sufficiently reliable to exclude changes of 30% or more. It would seem that painful responses during limb movements which tension the median nerve are more likely to result from an increase in local axonal mechanosensitivity and altered central processing. In this situation, nerve mechanosensitivity would result in upper limb pain during everyday upper limb activities.

Acknowledgments This work was supported in part by the Arthritis Research Campaign (Project Grant L0541). References Al-Shatti T, Barr AE, Safadi FF, Amin M, Barbe MF. Increase in inflammatory cytokines in median nerves in a rat model of repetitive motion injury. Journal of Neuroimmunology 2005;167(1–2):13–22. Barbe MF, Barr AE, Gorzelany I, Amin M, Gaughan JP, Safadi FF. Chronic repetitive reaching and grasping results in decreased motor performance and widespread tissue responses in a rat model of MSD. Journal of Orthopaedic Research 2003;21(1): 167–76. Bernard B, Sauter S, Fine L, Petersen M, Hales T. Job task and psychosocial risk factors for work-related musculoskeletal disorders among newspaper employees. Scandinavian Journal of Work, Environment and Health 1994;20(6):417–26. Bilecenoglu B, Uz A, Karalezli N. Possible anatomic structures causing entrapment neuropathies of the median nerve: an anatomic study. Acta Orthopaedica Belgica 2005;71(2):169–76. Bove GM, Light AR. Unmyelinated nociceptors of rat paraspinal tissues. Journal of Neurophysiology 1995;73(5):1752–62.

ARTICLE IN PRESS A. Dilley et al. / Manual Therapy 13 (2008) 536–543 Bove GM, Ransil BJ, Lin HC, Leem JG. Inflammation induces ectopic mechanical sensitivity in axons of nociceptors innervating deep tissues. Journal of Neurophysiology 2003;90(3):1949–55. Butler DS. Mobilisation of the nervous system, 1st ed. London: Churchill Livingstone; 1991. Byng J. Overuse syndromes of the upper limb and the upper limb tension test: a comparison between patients, asymptomatic keyboard workers and asymptomatic non-keyboard workers. Manual Therapy 1997;2(3):157–64. Campbell JN, Meyer RA. Mechanisms of neuropathic pain. Neuron 2006;52(1):77–92. Campbell JN, Raja SN, Meyer RA, Mackinnon SE. Myelinated afferents signal the hyperalgesia associated with nerve injury. Pain 1988;32(1):89–94. Clark BD, Barr AE, Safadi FF, Beitman L, Al-Shatti T, Amin M, et al. Median nerve trauma in a rat model of work-related musculoskeletal disorder. Journal of Neurotrauma 2003;20(7):681–95. Dilley A, Greening J, Lynn B, Leary R, Morris V. The use of crosscorrelation analysis between high-frequency ultrasound images to measure longitudinal median nerve movement. Ultrasound in Medicine and Biology 2001;27(9):1211–8. Dilley A, Lynn B, Greening J, DeLeon N. Quantitative in vivo studies of median nerve sliding in response to wrist, elbow, shoulder and neck movements. Clinical Biomechanics 2003;18(10):899–907. Dilley A, Lynn B, Pang SJ. Pressure and stretch mechanosensitivity of peripheral nerve fibres following local inflammation of the nerve trunk. Pain 2005;117(3):462–72. Eliav E, Benoliel R, Tal M. Inflammation with no axonal damage of the rat saphenous nerve trunk induces ectopic discharge and mechanosensitivity in myelinated axons. Neuroscience Letters 2001;311(1):49–52. Erel E, Dilley A, Greening J, Morris V, Cohen B, Lynn B. Longitudinal sliding of the median nerve in patients with carpal tunnel syndrome. Journal of Hand Surgery (British) 2003;28(5): 439–43. Gracely RH, Lynch SA, Bennett GJ. Painful neuropathy: altered central processing maintained dynamically by peripheral input. Pain 1992;51(2):175–94. Greening J, Lynn B. Vibration sense in the upper limb in-patients with repetitive strain injury and a group of at risk office workers. International Archives of Occupational and Environmental Health 1998;71(1):29–34. Greening J, Smart S, Leary R, Hall-Craggs M, O’Higgins P, Lynn B. Reduced movement of median nerve in carpal tunnel during wrist

543

flexion in patients with non-specific arm pain. Lancet 1999; 354(9174):217–8. Greening J, Lynn B, Leary R, Warren L, O’Higgins P, Hall-Craggs M. The use of ultrasound imaging to demonstrate reduced movement of the median nerve during wrist flexion in patients with nonspecific arm pain. Journal of Hand Surgery (British) 2001;26(5): 401–6. Greening J, Lynn B, Leary R. Sensory and autonomic function in the hands of patients with non-specific arm pain (NSAP) and asymptomatic office workers. Pain 2003;104(1–2):275–81. Greening J, Dilley A, Lynn B. In vivo study of nerve movement and mechanosensitivity of the median nerve in whiplash and nonspecific arm pain patients. Pain 2005;115(3):248–53. Grewal R, Xu J, Sotereanos DG, Woo SL. Biomechanical properties of peripheral nerves. Hand Clinics 1996;12(2):195–204. Harrington JM, Carter JT, Birrell L, Gompertz D. Surveillance case definitions for work related upper limb pain syndromes. Occupational Environmental Medicine 1998;55(4):264–71. Johnson RK, Spinner M, Shrewsbury MM. Median nerve entrapment syndrome in the proximal forearm. Journal of Hand Surgery (American) 1979;4(1):48–51. Julius A, Lees R, Dilley A, Lynn B. Shoulder posture and median nerve sliding. BMC Musculoskeletal Disorders 2004;5:23. LaMotte RH, Shain CN, Simone DA, Tsai EF. Neurogenic hyperalgesia: psychophysical studies of underlying mechanisms. Journal of Neurophysiology 1991;66(1):190–211. Lynn B, Greening J, Leary R. Sensory and autonomic function and ultrasound nerve imaging in RSI patients and keyboard workers. CRR 417/2002. Health and Safety Executive, 2002. Macfarlane GJ, Hunt IM, Silman AJ. Role of mechanical and psychosocial factors in the onset of forearm pain: prospective population based study. British Medical Journal 2000;321(7262): 676–9. McLellan DL, Swash M. Longitudinal sliding of the median nerve during movements of the upper limb. Journal of Neurology Neurosurgery and Psychiatry 1976;39(6):566–70. Pascarelli EF, Hsu YP. Understanding work-related upper extremity disorders: clinical findings in 485 computer users, musicians, and others. Journal of Occupational Rehabilitation 2001;11(1):1–21. Wertsch JJ, Melvin J. Median nerve anatomy and entrapment syndromes: a review. Archives of Physical Medicine and Rehabilitation 1982;63(12):623–7. Wilgis EF, Murphy R. The significance of longitudinal excursion in peripheral nerves. Hand Clinics 1986;2(4):761–6.

ARTICLE IN PRESS

Manual Therapy 13 (2008) 544–551 www.elsevier.com/math

Original article

Shoulder kinematic features using arm elevation and rotation tests for classifying patients with frozen shoulder syndrome who respond to physical therapy$ Jing-lan Yanga, Chein-wei Changa, Shiau-yee Chenb, Jiu-jenq Linc, a

Department of Physical Medicine and Rehabilitation, National Taiwan University Hospital, 7 Chun-Shan S Road, Taipei, Taiwan b Department of Internal Medicine, Taipei Medical University-Municipal Wan Fang Hospital, Taipei, Taiwan c College of Medicine, School and Graduate Institute of Physical Therapy, National Taiwan University, Floor 3, No.17, Xuzhou Road, Zhongzheng District, Taipei City 100, Taiwan Received 21 August 2006; received in revised form 8 June 2007; accepted 15 July 2007

Abstract Physical therapy is an intervention commonly used in the treatment of subjects with frozen shoulder symptoms, with limited proven effect. The purpose of this study was to identify the kinematic features of patients with frozen shoulder who are more likely to respond to physical therapy. Thirty-four subjects presenting frozen shoulder syndrome were studied to determine altered shoulder kinematics and functional disability. Subjects received the same standardized treatment with passive mobilization/stretching techniques, physical modalities (i.e. ultrasound, shortwave diathermy and/or electrotherapy) and active exercises twice a week for 3 months. Initially, subjects were asked to perform full active motion in 3 tests: abduction in the scapular plane, hand-to-neck and hand-to-scapula. During the test, shoulder kinematics were measured using a 3-D electromagnetic motion-capturing system. In the initial and follow-up sessions, the self-reported Flexilevel Scale of Shoulder Function (FLEX-SF) was used to determine functional disability from symptoms. Improvement with treatment was determined using percent change in FLEX-SF scores over three months of treatment [(final scoreinitial score)/initial score  100, 420% improvement and o ¼ 20% nonimprovement]. Shoulder kinematics were first analysed for univariate accuracy in predicting improvement and then combined into a multivariate prediction method. A prediction method with two variables (scapular tipping 48.41 during arm elevation, and external rotation 438.91 during hand to neck) were identified. The presence of these two variables (positive likelihood ratio ¼ 15.71) increased the probability of improvement with treatment from 41% to 92%. It appears that shoulder kinematics may predict improvement in subjects with frozen shoulder syndrome. Prospective validation of the proposed prediction method is warranted. r 2007 Elsevier Ltd. All rights reserved. Keywords: Frozen shoulder; Shoulder kinematics; Likelihood ratio; Prediction method

0. Introduction Patients exhibiting frozen shoulder symptoms typically suffer pain, a limited range of motion and muscle weakness from disuse for periods ranging from several $ This study was supported by a grant from the National Science Council, Taiwan (NSC 94-2314-B-002-088), awarded to Dr. Lin. Corresponding author. Tel.: +886 2 33228126; fax: +886 2 33228161. E-mail address: [email protected] (J.-j. Lin).

1356-689X/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2007.07.006

months to many years (Reeves, 1976; Shaffer et al., 1992). These symptoms usually respond to stretching/ mobilization (Griggs et al., 2000; Vermeulen et al., 2006), but many patients with frozen shoulder syndrome still have some degree of pain and stiffness several years after onset of the disease (Reeves, 1976; Shaffer et al., 1992). Indeed, for patients with persistent symptoms, more aggressive interventions such as hydrodilatation, arthroscopic release or manipulation under anesthesia have been advocated (Dias et al., 2005). A prospective study of 41 patients with 5–10 years’ follow-up indicated

ARTICLE IN PRESS J.-l. Yang et al. / Manual Therapy 13 (2008) 544–551

that 39% had full recovery, 54% had clinical limitation without functional disability and 7% had functional limitation (Reeves, 1976). Research regarding the efficacy of the early use of treatment strategies is warranted. Although physical therapy is an intervention commonly used in the treatment of subjects with frozen shoulder symptoms, the effectiveness of physical therapy intervention has been limited (Griggs et al., 2000; Diercks and Stevens, 2004). One possible explanation for the lack of positive effects is the inability to define subgroups of patients who are most likely to respond to physical therapy. Developing an effective method for classifying patients with frozen shoulder symptoms could improve decision making by determining the patients who are most likely to benefit from physical therapy. Thus, determining methods for classifying patients with frozen shoulder symptoms is an important priority in the clinic and research. Some investigators have suggested that frozen shoulder symptoms may be related to persistent synovitis, capsule contracture, contracted soft tissues and/or shoulder kinematic features (Loyd and Loyd, 1983; Neviaser, 1987; Parker et al., 1989; Mao et al., 1997; Griggs et al., 2000). Regardless of the potential factors related to frozen shoulder symptoms, altered shoulder kinematics is believed to exacerbate the condition and predispose patients to subacromial impingement, rotator cuff tendonitis, altered shoulder joint forces and possible degenerative changes (Ludewig and Cook, 2000; Lin et al., 2006). Thus, a more difficult and chronic course of frozen shoulder symptoms may develop. Additionally, previous research has indicated that the course of other types of shoulder dysfunction such as impingement (Ludewig and Cook, 2000; Lin et al., 2005) and shoulder tightness (Lin et al., 2006) may be associated with altered shoulder kinematics. It was the object of this study to identify the kinematic features of patients with frozen shoulder syndrome who are more likely to respond to physical therapy. Specifically, this study used a prediction method modified from a clinical prediction rule (McGinn et al., 2000) to determine whether impaired shoulder kinematics are associated with the degree of symptom-related functional disability in patients with frozen shoulder syndrome.

1. Methods 1.1. Subject recruitment This was a predictive validity/diagnostic test study. It was conducted at the outpatient clinic of the Department of Physical Medicine and Rehabilitation at National Taiwan University Hospital. All subjects gave written informed consent. Subjects were recruited if they

545

fulfilled the following inclusion criteria: 50% loss of passive movement of the shoulder joint relative to the nonaffected side, in 1 or more of 3 movement directions (i.e. forward flexion, abduction in the frontal plane, or external rotation in 01 of abduction) (Lundberg, 1969; Rizk et al., 1994; Diercks and Stevens, 2004); and duration of complaints of at least 3 months. Exclusion criteria were a history of stroke with residual upperextremity involvement, diabetes mellitus, rheumatoid arthritis, rotator cuff tear, surgical stabilization of the shoulder, osteoporosis or malignancies in the shoulder region. Subjects who had pain or disorders of the cervical spine, elbow, wrist, or hand, or who had pain radiating from the shoulder to the arm were also excluded. 1.2. Subjects A sample of 40 subjects was selected, based on availability for interview at the time of initial presentation to the clinic between August 2004 and May 2006 (Table 1). Three subjects did not return after the first session and were not included in the analysis. Two subjects left the study because of personal or workrelated circumstances. In addition, 1 subject was excluded from further participation in the study during the followup interview because she revealed the existence of bilateral frozen shoulder syndrome with severe and progressive symptoms, precluding the possibility of resolving the symptom course of a unilateral frozen shoulder. The final data analysis was therefore conducted on 34 patients. Two of these patients had had previous steroid injections at least 2 months before and none of them had had previous physical therapy treatments. 1.3. Shoulder kinematics assessments The FASTRAK 3-D electromagnetic motion-capturing system (Polhemus Inc., Colchester, VT, USA) was Table 1 Baseline (n ¼ 40) and follow up (n ¼ 34) conditions of patients with frozen shoulder syndrome receiving 3-month treatment Affected shoulder

FLEX-SFa score Duration (months)b Flexion Abduction External rotation Internal rotation Painc a

Unaffected shoulder

Baseline

Follow up

– – 1727151 1657181 857161 747191 –

27.576.3 6.478.3 122781 1057131 327161 227131 473

34.275.8 – 1457101 1237181 437221 347241 273

FLEX-SF ¼ Flexilevel Scale of Shoulder Function. Duration of symptom (pain or limited range of motion). c Pain intensity at the time of evaluation as determined with a visual analog scale (0–10). b

ARTICLE IN PRESS 546

J.-l. Yang et al. / Manual Therapy 13 (2008) 544–551

used to detect shoulder complex movements. Three sensors for the system were attached to the bony landmarks with adhesive tape. Each sensor was 2.3 cm in length, 2.8 cm in width, 1.5 cm in height and weighed 17 g. One sensor was attached to the sternum, and one sensor was attached to the flat bony surface of the scapular acromion with adhesive tape. The third sensor was attached to the distal humerus with Velcro straps. Data collection was performed as outlined in previous investigations (Lin et al., 2005). In general, we followed the ISB guidelines for constructing a shoulder joint coordinate system (Wu et al., 2005). Recordings started with the subject in a sitting position, the arms relaxed at the sides. Kinematics was collected for 5 s in this resting seated posture. Subjects were then asked to perform full active ROM in 3 tests: abduction in the scapular plane, hand-to-neck and hand-to-scapula. The hand-to-neck and hand-to-scapula tests represented function-related tests (Yang and Lin, 2006). For abduction in the scapular plane, subjects were guided to maintain in the scapular plane oriented 401 anterior to the coronal plane (Ludewig and Cook, 2000). Three replicated movements were performed in each test to the maximum possible active motions of the arms. The order of tests was randomized. To quantitatively characterize shoulder and scapular kinematics, the peak humeral elevation angle, the scapulo-humeral rhythm (slope of scapular upward rotation to glenohumeral elevation) and the peak scapular tilt were used as dependent variables in the abduction in the scapular plane test (Fig. 1). For the hand-to-neck and hand-to-scapula tests, the peak external rotation ROM and peak internal rotation ROM were used as dependent variables (Fig. 1). 1.4. Reliability and accuracy of kinematic variables For the system accuracy, within a 76-cm source-tosensor separation, the RMS system accuracy is 0.151 for orientation and 0.3–0.8 mm for position (Ludewig and Cook, 2000). In our study, the within-session reliability ICC (2, k) values ranged from .91 to .99 for subjects with frozen shoulder symptoms. These ICC values indicated good within-session reliability for the measured variables across the 3 tests. Furthermore, the standard error of our measurement p was 0.21, ffias calculated by the ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi equation SEM ¼ SD ð1  ICCÞ, where SD is the standard deviation. However, due to the progress of frozen shoulder symptoms in these patients, we did not test between-session reliability. 1.5. Functional evaluation The self-reported Flexilevel Scale of Shoulder Function (FLEX-SF) was used to determine functional disability from symptoms (Cook et al., 2003). In this scale, respondents answer a single question that grossly

anterior/posterior tipping

upward/downward rotation

humerus external/internal rotation Fig. 1. Axes and rotations used to describe scapular and humeral orientation and position.

classifies their level of function as low, medium or high. They then respond to only the items that target their level of function. Scores were recorded from 1, representing the most limited function, to 50, representing full function. Each patient was asked to indicate functional disability at the baseline and at a 3-month follow up. The percentage change in FLEX-SF was calculated (final scoreinitial score)/initial score  100). To develop a prediction method, we need to justify that the two subgroups show improvement and nonimprovement. If the change was 420%, the patient was categorized in the improvement group. If change was p20%, the patient was categorized in the nonimprovement group. We chose 20% change in FLEX-SF as the improvement criterion because the patients generally felt satisfied with 20% improvement from our investigation in the clinic.

2. Treatments Subjects received the same standardized treatment approach. The therapies included passive mobilization, stretching techniques, physical modalities (i.e. ultrasound, shortwave diathermy and/or electrotherapy) and active exercises. Each subject was treated by physical therapists with at least 3 years of clinical experience with

ARTICLE IN PRESS J.-l. Yang et al. / Manual Therapy 13 (2008) 544–551

2.1. Statistical analysis Improvement or nonimprovement was then used as the reference outcome. To address potential confounding variables, we compared the duration of symptoms, initial FLEX-SF scores and compliance with physical therapy treatment between the improvement and nonimprovement groups. Additionally, the planes of arm elevation between the two groups were also examined. Individual variables from the shoulder kinematics were tested for their relationship with the reference outcome using independent sample t tests. Variables with a significance level of po0.10 were retained as potential prediction variables; a more liberal significance level was chosen at this stage to avoid excluding potential predictive variables. For a significant relationship, sensitivity and specificity values were calculated for all possible cut-off points and then plotted as a receiver operator characteristic (ROC) curve (Hagen, 1995). The point on the curve nearest the upper left-hand corner represents the value with the best diagnostic accuracy, and this point was selected as the cut-off defining a positive test (Deyo and Centor, 1986). Sensitivity, specificity and positive likelihood ratios (PLR) were calculated for all potential prediction variables (Sackett, 1992). The PLR is calculated as sensitivity/(1specificity) and indicates the increase in the probability of improvement given a significant altered kinematic result. A PLR of 1 indicates that the kinematics does nothing to alter the probability of improvement, whereas PLR values 41 increase the probability of improvement given a significant altered kinematic result. PLR values between 2.0 and 5.0 generate small shifts in probability, values between 5.0 and 10.0 generate moderate shifts and values 410.0 generate large and often conclusive shifts in probability (Jaeschke et al., 1994). Potential prediction variables were entered into a stepwise logistic regression equation to determine the kinematics predictors for improvement using a multivariate model. A significance of 0.05 was required to enter a variable into the model, and a significance of 0.10 was required to remove it. Variables retained in the regression model were used to develop a multivariate prediction method for determining shoulder kinematics in the prediction of the progress of frozen shoulder syndrome.

41–65 years). The involved shoulder distribution in the subjects with unilateral involvement was 18 right dominant (53%), 6 right nondominant (18%), and 10 left nondominant (29%). Overall, the mean improvement in FLEX-SF scores over the 3-month period was 4.873.4, with a mean percentage improvement of 15.275.3%. Fourteen subjects (41%) were classified as showing improvement and 20 (59%) as showing nonimprovement. The mean improvement in FLEXSF scores in the improvement group over the 3-month period was 6.574.3, with a mean percentage improvement of 24.674.6%. In the nonimprovement group, the mean FLEX-SF score change was 2.471.8, with a mean percentage change of 8.876.5% (Fig. 2). There were no significant differences between the improvement and nonimprovement groups in duration of symptoms (5.977.1 versus 6.679.5 months), initial FLEX-SF scores (29.577.3 versus 26.575.6), or compliance with physical therapy treatment (20.670.9 versus 19.870.8 visits). Additionally, there was also no significant difference in received number of treatments regarding mobilization, stretching and/or physical modalities between the two groups. There was no difference between the two groups regarding the plane of arm elevation (40.670.51 versus 40.570.41). Representative kinematic data from a subject during arm elevation in the scapular plane are presented in Fig. 3. Although there was substantial variability among subjects, the general pattern was for the scapula to upwardly rotate and move posteriorly toward less anteriorly tipped positions as the arm elevated. Four prediction variables were retained from shoulder kinematic variables (Table 2): scapular tipping, humeral elevation, scapulohumeral rhythm and external rotation (hand to neck). Cut-off values and diagnostic statistics for retained variables were obtained from ROC curve 40 Flexilevel Scale of Shoulder Function

the application of mobilization/stretching techniques for patients with frozen shoulder syndrome. Subjects were treated twice a week for 3 months. Subjects did not receive a home exercise programme but were advised to use the affected shoulder in daily activities whenever possible.

547

Initial Three-month Follow-up

35 30 25 20 15 10 5 0 Improvement

3. Results Of the 34 subjects completing the study, 26 (77%) were female. The mean age was 54.176.1 years (range

Nonimprovement

Fig. 2. Initial and 3-month follow-up Flexilevel scale of shoulder function for the improvement and nonimprovement groups. The mean percent change in the improvement group was 24.674.6%. For the nonimprovement group, the mean percent change was 8.876.5%.

ARTICLE IN PRESS J.-l. Yang et al. / Manual Therapy 13 (2008) 544–551

548

100

80

Elevation Upward rotation Posterior tipping

analyses (Table 3). Among the kinematic variables, scapular tipping was the most predictive of improvement (PLR ¼ 3.32). The 4 potential prediction variables were entered into the logistic regression. Two were retained in the final model: scapular tipping and external rotation (hand to neck, model w2 ¼ 14.03, df ¼ 2, P ¼ 0.001, Nagelkerke R2 ¼ 0.46); 11 of the 14 subjects were in the improvement group for two retained prediction variables at baseline. Three of the 14 subjects, with 1 of 2 variables present, were in the improvement group. Accuracy statistics were calculated for each level of the prediction method (Table 4). Based on the probability of improvement found in this study (41%) and the PLR values calculated, a subject with 2 variables present at baseline has an increased probability of improvement from 41% to 92%. If the criteria were changed to 1 variable present, the probability of improvement would increase to only 56%. Additionally, we did not find significant differences in the other kinematic variables.

Humeral Elevation Scapular Upward Rotation Scapular Posterior Tipping

Angle (degree)

60

40

20

0

-20 Time Fig. 3. Data for a representative subject with frozen shoulder. Scapular posterior tipping, scapular upward rotation and humeral elevation during arm elevation in the scapular plane.

4. Discussion Although course of frozen shoulder syndrome is long term and the etiology is unclear, our results indicate that shoulder kinematics may be used as predictors of the clinical course of patients with frozen shoulder syndrome. Similar to other studies (Shaffer et al., 1992; Griggs et al., 2000; Diercks and Stevens, 2004), we were able to show only adequate effects of 3 months of treatment for some of our subjects’ symptoms. Additionally, we found that symptoms worsened in some subjects despite treatment for 3 months. As noted by previous studies, symptoms of frozen shoulder may develop over 6 months and may be mainfested as longlasting pain and restricted motion (Reeves, 1976; Shaffer et al., 1992; Dias et al., 2005). Nevertheless, by considering shoulder kinematics, especially scapula and humeral motions together, we were able to develop a prediction method that may be useful for assisting clinicians in identifying important shoulder kinematics that are likely to predict improvement with physiotherapy in patients with frozen shoulder syndrome. The developed prediction method contains 2 variables: scapular tipping 4 8.41 and external rotation (hand to neck) 438.91. These findings are generally consistent with previous theories and research (Ludewig and Cook, 2000; Rundquist and Ludewig, 2004; Lin et al. 2006; Vermeulen et al., 2006). Adequate posterior tipping of the scapula elevates the anterior acromion and may be critical in obtaining adequate clearance of subacromial tissues, which may exclude impingement and further frozen shoulder syndrome (Ludewig and Cook, 2000; Lin et al., 2006). Our results support this hypothesis. Additionally, impaired external rotation has been reported in subjects with frozen shoulder syndrome (Rundquist and Ludewig, 2004; Vermeulen et al., 2006). Limited external rotation is related to tightened capsules and/or ligaments (Rundquist and Ludewig, 2004; Vermeulen et al., 2006). Cyriax proposed that tightness in a joint capsule would restrict motion in a predictable pattern, a capsular pattern (Cyriax, 1978). In the case of the frozen shoulder (adhesive capsulitis), a capsular

Table 2 Shoulder kinematic variables used in this study (N ¼ 34)

Scapular posterior tipping Scapular upward rotation Humeral elevation Scapulohumeral rhythm External rotation (Hand to neck) Internal rotation (Hand to back)

Improvement (420% FLEX-SF scores, N ¼ 14/34)

Nonimprovement (p20% FLEX-SF scores, N ¼ 20/34)

p-value

18.075.81* 31.8712.61 102.5719.91* 0.7070.13* 52.3726.61* 15.6711.11

10.977.31 35.175.91 89.3718.01 0.9070.23 39.9716.21 11.075.51

0.004 0.304 0.057 0.008 0.097 0.121

*Variables with a significance level of po0.10 based on independent sample t tests.

ARTICLE IN PRESS J.-l. Yang et al. / Manual Therapy 13 (2008) 544–551

549

Table 3 Sensitivity and specificity statistics (with 95% confidence intervals) for kinematic variables for predicting improvement Kinematic variables associated with improvement

Sensitivity

Specificity

Positive likelihood ratio

Scapular posterior tipping 4 8.41 Humeral elevation 497.01 Scapulohumeral Rhythm o0.88 External rotation (Hand to neck) 4 38.91

92.9 71.4 100.0 71.4

60.0 70.0 50.0 60.0

3.32 2.08 2.10 2.21

(66.1–98.8) (41.9–91.4) (76.7–100.0) (41.9–91.4)

(36.1–80.8) (45.7–88.0) (27.2–72.8) (36.1–80.8)

Table 4 A prediction method No. of Predictor Variables Present

Sensitivity

Specificity

Positive likelihood ratio

Probability of improvement* (%)

2 1+

78.6 (49.2–95.1) 100.0 (76.7–100.0)

95.0 (75.1–99.2) 45.0 (23.1–68.4)

15.71 1.82

92 56

*The probability of improvement is calculated using the positive likelihood ratio and assumes a pretest probability of improvement of 41%.

pattern is one in which external rotation is more limited than abduction, which in turn is more limited than internal rotation. In our study, impaired external rotation is likely to predict improvement of frozen shoulder syndrome with physiotherapy. An interesting finding was that humeral elevation and scapulohumeral rhythm were not associated with chronic frozen shoulder disability. Although they appeared to be significant when analyzed at baseline between groups, the apparent significance was lost when it was entered into a prediction method with a multifactorial model. Theoretically, an adequate amount of humeral external rotation is required for humeral elevation (Rundquist and Ludewig, 2004). Therefore, our results suggest that subjects with adequate scapular tipping and humeral external motion are likely to show a treatment improvement effect. The importance of a prediction method for determining shoulder kinematics in the prediction of the progress of frozen shoulder syndrome is best expressed using likelihood ratio statistics. When the subject meets the prediction rule’s criteria, PLR expresses the change in odds favouring the improvement (Sackett, 1992). In our sample, treatment of subjects with frozen shoulder syndrome may result in about a 41% probability of the improvement without any attempt at prediction. Using 2 criteria variables present at baseline (PLR ¼ 15.71), the probability of improvement is raised to 92%; therefore, these individuals respond to treatments. If only one variable is present, the probability of improvement increased only to 56%, which suggests that scapular tipping and external rotation of the shoulder, where they are judged to be less than the thresholds identified in this study, should be considered as important treatment goal/areas when

considering mobilization/stretching treatments in these subjects. Consideration in the assessment of the progress of frozen shoulder syndrome is important regarding the outcomes which are judged. Previous studies of assessment used impairment measures such as ROM and strength as primary means of evaluating the effectiveness of intervention (Reichmister and Friedman, 1999; Arslan and Celiker, 2001). In our opinion, traditional impairment measures may have insufficient reliability and validity. Therefore, we chose to reference functional FLEX-SF scores with good reliability and validity (Cook et al., 2003) that are representative of the desired outcome measures. The use of 20% improvement on the FLEX-SF scores as the reference standard was based on previous research involving intervention in subjects with frozen shoulder syndrome. Symptoms of frozen shoulder develop over 6 months, may last 2 years and may then gradually disappear (‘‘self-limiting character,’’ Lundberg, 1969; Reeves, 1976; Grey 1978). Sometimes, there may be long-lasting pain and restricted motion. Reeves (1976) described the natural history of frozen shoulder and found a mean duration of the symptoms of 30 months (range ¼ 12–42). As our subjects’ symptoms has been present for at least 3 months (range ¼ 3–9), we therefore thought that 20% improvement in the FLEX-SF scores over a 3-month period would provide adequate distinction between subjects responding to the intervention and those simply benefiting from the natural history of adhesive capsulitis. Additionally, we found that patients generally felt satisfied with 20% improvement from our investigation in the clinic. On the other hand, different results may be expected from other criteria. This needs to be further validated.

ARTICLE IN PRESS 550

J.-l. Yang et al. / Manual Therapy 13 (2008) 544–551

Methodological standards for developing and validating a prediction method, modified from a clinical prediction rule, include three steps (McGinn et al., 2000). The initial step is the development of the method, the second step is validation of the method and the third step is an analysis of the impact of the method on clinical behaviour. The purpose of the current study was to develop a prediction method that would identify impaired shoulder kinematics associated with the degree of symptom-related functional disability in patients with frozen shoulder syndrome. Validation of the proposed prediction method is the purpose of an ongoing intervention trial to determine whether altered shoulder kinematics subjected to intervention meet the prediction criteria and demonstrate improvement in their follow-up. Eventually, the prediction method should be shown to improve decision making in clinical practice. Limitations of the study should be noted. Interpretation of our results should be cautious because there was a lack of control over intervention in our study. We examined the treatments between improvement and nonimprovement groups and found that similar treatment protocols were applied to two groups. It is possible, however, that a specific modality tended to improve the condition over another. Although good within-session reliability was demonstrated in this study, the between-session reliability of these kinematic variables should be further investigated. In the present study, only shoulder kinematics and 3-month improvement were considered, and it is unknown whether other factors or long-term follow-up would provide similar results. Furthermore, the participant population was mainly subjects with idiopathic frozen shoulder. The generalizability of the study results to frozen shoulders with other pathologies is uncertain. The mean FLEX-SF score for the subjects with frozen shoulder was 27.5. Subjects with more severe impairments may be expected to show different alterations in kinematics. In addition, scapular kinematics were measured with sophisticated laboratory measures in our study: appropriate clinical measures of scapular motion/position corresponding to laboratory measures should be further investigated. This investigation supports the assertion that shoulder kinematics are associated with the progress of disability in subjects with frozen shoulder syndrome. Based on the prediction method we found, a subject with frozen shoulder syndrome who meets 2 criteria (scapular tipping 48.41 and humeral external rotation (hand to neck) 438.91) at baseline has a probability of 92% of demonstrating improvement at 3-month followup. Although humeral external rotation is generally advocated as a treatment focus, our results further suggest that scapular tipping and humeral external rotation should be managed together. Additionally, specific least angles are recommended. However, before

this prediction method can be considered ready for use in clinical practice, it should be validated in a prospective study.

References Arslan S, Celiker R. Comparison of the efficacy of local corticosteroid injection and physical therapy for the treatment of adhesive capsulitis. Rheumatology International 2001;21:20–3. Cook KF, Roddey TS, Gartsman GM, Olson SL. Development and psychometric evaluation of the Flexilevel Scale of Shoulder Function. Medical Care 2003;41:823–35. Cyriax J. Textbook of orthopedic medicine. Diagnosis of soft tissue lesions, vol. 1. New York: Macmillan; 1978. Deyo RA, Centor RM. Assessing the responsiveness of functional scales to clinical change: an analogy to diagnostic test performance. Journal of Chronic Diseases 1986;11:897–906. Dias R, Cutts S, Massoud S. Frozen shoulder. British Medical Journal 2005;331:1453–6. Diercks RL, Stevens M. Gentle thawing of the frozen shoulder: a prospective study of supervised neglect versus intensive physical therapy in seventy-seven patients with frozen shoulder syndrome followed up for two years. Journal of Shoulder and Elbow Surgery 2004;13:499–502. Grey RG. The natural history of ‘‘idiopathic’’ frozen shoulder. Journal of Bone and Joint Surgery (American volume) 1978;60:564. Griggs SM, Ahn A, Green A. Idiopathic adhesive capsulitis. A prospective functional outcome study of nonoperative treatment. Journal of Bone and Joint Surgery (American volume) 2000;82: 1398–407. Hagen MD. Test characteristics: how good is that test? Medical Decision Making 1995;22:213–33. Jaeschke R, Guyatt G, Sackett DL. Users’ guides to the medical literature: III. How to use an article about a diagnostic test. A. Are the results of the study valid? Journal of the American Medical Association 1994;271:389–91. Lin JJ, Hanten WP, Olson SL, et al. Functional activity characteristics of individuals with shoulder dysfunctions. Journal of Electromyography and Kinesiology 2005;15:576–86. Lin JJ, Lim HK, Yang JL. Effect of shoulder tightness on glenohumeral translation, scapular kinematics, and scapulohumeral rhythm in subjects with stiff shoulders. Journal of Orthopaedic Research 2006;24:1044–51. Loyd JA, Loyd HM. Adhesion capsulitis of the shoulder: arthrographic diagnosis and treatment. Southern Medical Journal 1983;76:879–83. Ludewig PM, Cook TM. Alterations in shoulder kinematics and associated muscle activity in people with symptoms of shoulder impingement. Physical Therapy 2000;80:276–91. Lundberg BJ. The frozen shoulder. Acta Orthopaedica Scandinavica 1969;119:1–59. Mao C, Jaw W, Cheng H. Frozen shoulder: correlation between the response to physical therapy and follow-up shoulder arthrography. Archives of Physical Medicine and Rehabilitation 1997;78: 857–9. McGinn TG, Guyatt GH, Wyer PC, et al. Users’ guide to the medical literature: XXII. How to use articles about clinical decision rules. Journal of the American Medical Association 2000;284:79–84. Neviaser TJ. Adhesive capsulitis. Orthopedic Clinics of North America 1987;18:439–43. Parker RD, Froimson AI, Winsberg DD, Arsham NZ. Frozen shoulder. Part I: chronology, pathogenesis, clinical picture, and treatment. Orthopedics 1989;12:869–73. Reeves B. The natural history of the frozen shoulder syndrome. Scandinavian Journal of Rheumatology 1976;4:193–6.

ARTICLE IN PRESS J.-l. Yang et al. / Manual Therapy 13 (2008) 544–551 Reichmister JP, Friedman SL. Long-term functional results after manipulation of the frozen shoulder. Maryland Medical Journal 1999;48:7–11. Rizk TE, Gavant ML, Pinals RS. Treatment of adhesive capsulitis (frozen shoulder) with arthrographic capsular distension and rupture. Archives of Physical Medicine and Rehabilitation 1994;75:803–7. Rundquist PJ, Ludewig PM. Patterns of motion loss in subjects with idiopathic loss of shoulder range of motion. Clinical biomechanics 2004;19:810–8. Sackett DL. A primer on the precision and accuracy of the clinical examination. Journal of the American Medical Association 1992;267:2638–44. Shaffer B, Tibone JE, Kerlan RK. Frozen shoulder. A long term follow up. Journal of Bone and Joint Surgery (American volume) 1992;74:738–46.

551

Vermeulen HM, Rozing PM, Obermann WR, le Cessie S, Vliet Vlieland TP. Comparison of high-grade and low-grade mobilization techniques in the management of adhesive capsulitis of the shoulder: randomized controlled trial. Physical Therapy 2006;86: 355–68. Wu G, van der Helm FC, Veeger HE, Makhsous M, Van Roy P, Anglin C, et al. International Society of Biomechanics. ISB recommendation on definitions of joint coordinate systems of various joints for the reporting of human joint motion—Part II: shoulder, elbow, wrist and hand. Journal of Biomechanics 2005; 38:981–92. Yang JL, Lin JJ. Reliability of function-related tests in patients with shoulder pathologies. Journal of Orthopaedic and Sports Physical Therapy 2006;36:572–6.

ARTICLE IN PRESS

Manual Therapy 13 (2008) 552–559 www.elsevier.com/math

Original article

The influence of age, gender, lifestyle factors and sub-clinical neck pain on the cervical flexion–rotation test and cervical range of motion Kenric Smith, Toby Hall, Kim Robinson School of Physiotherapy, Curtin University of Technology, GPO Box U1987, Perth WA 6845, Australia Received 6 July 2006; received in revised form 9 June 2007; accepted 16 July 2007

Abstract The flexion–rotation test (FRT) is commonly used when assessing cervicogenic headache. Additionally, active range of motion (AROM) is frequently used to evaluate impairment in neck pain. No studies have investigated the interaction of the FRT and AROM with age, gender, pain and lifestyle factors. The purpose of this study was to determine the influence of these factors on the FRT and cervical AROM. A group of 66 participants (aged 20–78) were studied, 28 experienced sub-clinical neck pain (recurrent neck pain or discomfort which has not received treatment from a healthcare professional) while 38 did not. Age, gender, lifestyle factors and sub-clinical neck pain were assessed using a questionnaire. Measurement of AROM was performed by two examiners blind to the results of the questionnaire. Multiple linear regression analysis found that 59% of the variance in the FRT was explained by the presence of sub-clinical pain and cervical lateral flexion measures. Secondly, 58–72% of the variance in active cervical ROM measures was influenced by factors including the FRT, gender and movements of the neck in other planes. This study found that lifestyle factors do not influence the cervical FRT and AROM. r 2007 Elsevier Ltd. All rights reserved. Keywords: Neck pain; Flexion–rotation test; Atlantoaxial joint; Range of motion

1. Introduction The prevalence of neck pain in the general population has been estimated to be as high as 34% over 1 year and 67% over an entire lifetime (Bovim et al., 1994; Cote et al., 1998), and it is recognized that many people with neck pain fail to seek formal treatment (Grant et al., 1995). Sub-clinical neck pain refers to recurrent neck pain or discomfort which has not received treatment from a healthcare professional (Lee et al., 2004). This untreated population group is of particular interest to health professionals including physiotherapists as they may represent an intermediary between individuals with no pain and those currently seeking treatment (Lee et al., 2005a). Before early intervention programs can be implemented, clinical features of sub-clinical pain Corresponding author. Tel.: +61 8 93811863.

E-mail address: [email protected] (T. Hall). 1356-689X/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2007.07.005

need to be examined further (Lee et al., 2005a). Previous research on sub-clinical neck pain has used active cervical ROM (CROM) as an outcome measure (Lee et al., 2004, 2005a, b). Lee et al. (2004) found statistically significant differences in active CROM between sub-clinical neck pain subjects and controls, with a mean difference of 51 for left rotation and 31 for extension. Assessment of active cervical movements is a routine part of examination of cervical spine disorders (Maitland et al., 2001). Active cervical examination, however, incorporates movements of both upper and lower cervical segments simultaneously. According to Bland (1998), the C1/2 motion segment accounts for 50% of the rotation in the cervical spine. Thus, the examination of C1/2 impairment may be useful in assessment of cervical spine disorders, particularly those involving the high cervical spine, such as cervicogenic headache (Hall and Robinson, 2004). Indeed pain arising from the C1/2 segment is a frequent finding in

ARTICLE IN PRESS K. Smith et al. / Manual Therapy 13 (2008) 552–559

subjects with cervicogenic headache (Bogduk, 2001; Aprill et al., 2002; Hall and Robinson, 2004). The flexion–rotation test (FRT) is a clinically useful test purportedly biased to assess rotational dysfunction at the C1/2 motion segment, though there is no direct evidence to support this. The validity of the FRT as a measure of C1/2 impairment is based on studies that have identified a positive FRT in those subjects with C1/2 dominant segmental dysfunction determined by manual examination (Hall and Robinson, 2004; Ogince et al., 2006). Likewise, the FRT has been shown to be negative in subjects with cervicogenic headache where pain arises from cervical levels other than C1/2 (Hall et al., 2007). Literature regarding age effects on cervical range of motion (ROM) tends to support the notion that with increasing age there is a decrease in motion (Dvorak et al., 1992; Youdas et al., 1992; Walmsley et al., 1996; Chen et al., 1999; Feipel et al., 1999). In contrast, the influence of gender on CROM has not been clearly supported in the literature (Dvorak et al., 1992; Trott et al., 1996; McClure et al., 1998; Mannion et al., 2000). Lifestyle factors including sleeping position, side dominant exercise or occupation and hours spent in sustained positions such as sitting all appear to have some influence on ROM measures (Guth, 1995; Gordon et al., 2002; Lee et al., 2004). Age, gender and lifestyle factors are often assessed during examination of patients with neck pain and headaches; however, their relationship to tests used in the physical examination of these disorders has not been studied adequately. The primary aim of this study was to determine whether age, gender, time spent sitting, sleep position, side dominant lifestyle, presence of sub-clinical pain, and active CROM, measured in the cardinal plane, influence the range of the FRT. The secondary aim was to determine whether these variables influenced active CROM.

2. Method 2.1. Study design A single blind descriptive design was used to determine the influence of age, gender, lifestyle factors and sub-clinical neck pain on ROM recorded during the FRT and active cervical movements in the cardinal planes. A questionnaire sought information regarding lifestyle. Following evaluation of the physical measures, subjects were questioned regarding the presence of subclinical neck pain. This order of testing was important to prevent examiner bias during the FRT. 2.2. Subjects Advertisements placed at the local university campus sought volunteers aged between 20 and 80 years of age.

553

No other inclusion criteria were specified. Potential participants were excluded from the study if they had a headache more than once per month, neck pain that had required treatment from a health professional in the last year, any history of cervical surgery, seropositive arthritis, Down’s syndrome, or known congenital anomalies of the cervical spine. In addition, subjects were excluded if they were unable to communicate with the examiner. Withdrawal criteria included withdrawal of consent or inability to tolerate the examination process. Ethical approval for this study was granted by the Curtin University Human Research Ethics Committee. Each subject provided informed consent before testing and the rights of each participant were respected at all times. A total of 66 subjects aged between 20 and 78 years of age (mean age, 33 years; SD, 13.5) volunteered to participate in this study with no subsequent withdrawals. The subject population comprised an even distribution of gender (49% female, 51% male) and subclinical neck pain (42% present, 58% absent). 2.3. Materials Range of active cervical motion in the cardinal planes was measured using the CROM device (Performance Attainment Associates, 958 Lydia Drive, Roseville, MN 55113, USA). This device has been shown to have good intra-tester and inter-tester reliability for measures of active cervical motion in the cardinal planes (CapuanoPucci et al., 1991; Rheault et al., 1992; Hall and Robinson, 2004). In particular, Hall and Robinson (2004) found intra-class correlation coefficients ranging between 0.92 (flexion) and 0.99 (extension), indicating high reliability for measures of active cervical motion. Furthermore, the CROM device has very good validity for measurement of CROM (Ordway et al., 1997; Tousignant et al., 2000, 2002; Malmstrom et al., 2003). For measurement of the FRT (Fig. 1), the CROM device was modified and attached to the top of the head at the axis of rotation. Hall and Robinson (2004) and Ogince et al. (2006) found this modified CROM device to have excellent reliability for measures of range of rotation during the FRT with intra-class correlation coefficients above 0.9 and a kappa value of 0.81. 2.4. Procedures Activity and lifestyle information was collected using the questionnaire in Appendix A. This questionnaire provided data on the participant’s age, gender, lifestyle, time spent sitting, and sleeping position. Subjects were not permitted to discuss these answers, nor their history of previous neck pain, with the physiotherapist during the testing procedure in order to prevent examiner bias.

ARTICLE IN PRESS 554

K. Smith et al. / Manual Therapy 13 (2008) 552–559

was passively rotated to maximum pain-free left and right rotation and ROM recorded. Immediately following the physical measurements, information regarding recurrent neck pain was obtained. The examiner deemed the subject to have sub-clinical neck pain if they answered yes to a question ‘‘Do you get any recurring neck pain?’’. 2.5. Data analysis

Fig. 1. The cervical flexion–rotation test.

All physical measurements were performed by one of two manipulative physiotherapists with greater than 20 years of clinical experience. Previous research has found excellent inter-rater reliability between these two physiotherapists for assessment of active ROM and the FRT (Hall and Robinson, 2004; Ogince et al., 2006). Furthermore, Hall and Robinson (2004) found no significant change in ROM with repeated trials for any of the measurements used in this study. Thus, measurements were only taken once, in each direction, by one physiotherapist, for each subject. To test active CROM, the subject was seated with a neutral spine posture, with the spine supported against a high-back chair. Once in this position, the CROM was fitted and the purpose of the device was explained. The subjects were then instructed to move their head and neck in all cardinal planes to maximum pain-free range. The participant was first given a practice run in each of the described directions. After the practice run, ROM was recorded in each direction. For the FRT, the modified CROM was arranged in the following sequence. Two Velcro straps were positioned to allow the compass to sit horizontal at the vertex of the skull. A floating compass (Plastimo Airguide Inc. (Compasses), 1110 Lake Cook Road, Buffalo Groove, IL 60089, USA) with inbuilt spirit level was attached at the axis of rotation for measurement of the FRT. Before the FRT procedure, subjects were instructed to inform the examiner if they felt any pain during the test. The subjects were asked to lay supine with their cervical spine projecting over the edge of the plinth supported by the examiner. The examiner instructed the subject to relax with their hands on their abdomen. The physiotherapist then passively moved the neck into maximum pain-free cervical flexion. In this position, the neck

All data were analysed using Statistical Package for Social Sciences version 12.0 (SPSS, Inc., Chicago, IL). In all cases, alpha were set at the 0.05 level. Following data collection, ROM data was transformed to produce fewer variables. Total movement in the sagittal plane was produced by simple addition of active flexion and extension. Total rotation and total lateral flexion was produced by simple addition of range to the left and right. Finally, total rotation in the FRT was produced by simple addition of left and right rotation with the cervical spine maximally flexed. Following data reduction, there were four dependent variables: total rotation, total lateral flexion, total flexion–extension and total rotation in the FRT. The independent variables used were age, gender, presence of sub-clinical pain, presence of side dominant lifestyle, presence of prone sleeping pattern and average daily hours spent sitting. Multiple linear regression was performed four times to determine which independent variables best predicted the variability in each of the four dependent variables. When performing each regression analysis, one dependent variable was investigated with the remaining three dependent variables being treated as independent variables in the analysis. For example, when total rotation in the FRT was the dependent variable, total rotation, total lateral flexion, and total flexion–extension were treated as independent variables alongside the independent variables described above. Prior to analysis, all assumptions for the use of multiple linear regression were met. 3. Results All 66 subjects completed the study. The mean, standard deviation and range of the independent variables in the study sample are presented in Table 1. Table 2 represents the mean and standard deviation of the dependent variables (unilateral and combined measures). Dependent variables are grouped according to the presence or not of sub-clinical pain with the total sample also presented. Table 3 represents results of the multiple linear regression analysis with T, p, R2 and F values presented. With total rotation in the FRT as the dependent

ARTICLE IN PRESS K. Smith et al. / Manual Therapy 13 (2008) 552–559 Table 1 Characteristics of the independent variables (n ¼ 66) Variable

Number of subjects

Age (years) Mean (SD) Range

34 (13.5) 20–78

Gender Male Female

34 (51%) 32 (49%)

Sub-clinical pain Yes No

28 (42%) 38 (58%)

Side dominant lifestyle Yes No

25 (38%) 41 (62%)

Prone sleeping position Yes No

15 (23%) 51 (77%)

Hours spent sitting daily Mean (SD) Range

8 (3.6) 2–18

Table 2 Mean (SD) scores for the dependent variables (n ¼ 66) Variable

Total subjects (n ¼ 66)

Total flexion– rotation Flexion–rotation (L) Flexion–rotation (R)

84 (14.4)

Total lateral flexion Lateral flexion (L) Lateral flexion (R)

81 (17.1)

No pain (n ¼ 38)

Sub-clinical pain (n ¼ 28)

90.8 (10.2)

74.9 (14.3)

45.5 (5.6) 45.3 (5.2)

37.8 (9.0) 37.2 (7.3)

86.5 (16.6) 43.5 (8.3) 43.0 (8.7)

74.6 (15.5) 38.2 (7.6) 36.4 (8.4)

Total rotation Rotation (L) Rotation (R)

146 (19.1)

151 (16.9) 74.6 (10.0) 76.4 (8.1)

140.3 (20.5) 70.6 (10.6) 69.8 (10.7)

Flexion–extension Flexion Extension

125 (19.7)

129.8 (21.2) 53.6 (12.3) 76.2 (15.2)

118.7 (15.8) 47.1 (8.9) 71.6 (12.7)

variable, the R2 values suggest that 59% of the variance in this measure can be explained by the presence of subclinical pain and total lateral flexion. This was highly significant (F(2,63) ¼ 45.2, po0.001), with examination of T-values indicating that both sub-clinical pain (t ¼ 4.06, po0.001) and total lateral flexion (t ¼ 6.66, po0.001) contribute to the prediction of total rotation in the FRT. Multiple linear regression analysis with total rotation as the dependent variable found that both flexion–extension and total rotation in the FRT explain

555

58% (R2 ¼ 0.58) of the variance in total rotation, which was highly significant (F(2,63) ¼ 43.6, po0.001). The remaining independent variables, lifestyle factors, gender and age, were excluded from these two analyses, thus suggesting that their influence on total rotation in the FRT and total rotation were negligible. In a clinical context, these analyses suggest that ROM during the FRT is only partly explained by a combination of sub-clinical pain and range of lateral flexion. Likewise, cervical rotation ROM is only partly explained by ROM during the FRT and cervical flexion/ extension. Analysis of multiple linear regression with total lateral flexion as the dependent variable concluded that total rotation in the FRT, flexion–extension, and gender explain 72% (R2 ¼ 0.72) of the variance in total axial rotation, which was highly significant (F(3,62) ¼ 53.2, po0.001). The final analysis with flexion–extension as the dependent variable found total lateral flexion and total rotation to explain 68% (R2 ¼ 0.68) of the variance in flexion—extension, which was highly significant (F(2,63) ¼ 66.3, po0.001). The remaining independent variables were excluded from these two analyses, thus suggesting that these variables had no influence on total lateral flexion and flexion– extension. These two regression analyses suggest that we can predict with moderate confidence cervical rotation ROM based on a combination of FRT ROM, flexion–extension range and gender. Similarly, flexion– extension range can be predicted with moderate confidence by total cervical rotation and total lateral flexion. 4. Discussion This is the first study to examine the influence of age, gender, lifestyle factors and sub-clinical neck pain on the FRT and active CROM. Our results concluded that 59% of the variance in FRT mobility can be explained by the presence of sub-clinical pain and by measurement of active lateral flexion. Secondly, 58% of the variance in total active rotation is influenced by capital mobility and FRT mobility. Thirdly, we concluded that 72% of the variance in active lateral flexion is influenced by FRT mobility, capital mobility and gender. Lastly, it was concluded that 68% of the variance in capital mobility (active flexion–extension) is influenced by measurement of total active rotation and lateral flexion. It is important to recognize that these results show only moderate correlations; factors other than those investigated in this study may also be involved to explain the variance in the FRT and active CROM. Previous studies have shown an influence of age on some measures of CROM. According to a number of authors (Dvorak et al., 1992; Youdas et al., 1992;

ARTICLE IN PRESS 556

K. Smith et al. / Manual Therapy 13 (2008) 552–559

Table 3 T, p, R2 and F values from multiple linear regression analysis p

R2

F

6.66 4.06

o0.001 o0.001

0.59

45.2

Flexion–extension Total flexion–rotation

5.02 3.52

o0.001 0.001

0.58

43.6

Total lateral flexion

Total flexion–rotation Flexion–extension Gender

4.52 6.76 2.29

o0.001 o0.001 0.026

0.72

53.2

Flexion–extension

Total lateral flexion Total rotation

5.93 3.71

o0.001 o0.001

0.68

66.3

Dependent variable

Significant independent variable

Total flexion–rotation

Total lateral flexion Sub-clinical pain

Total rotation

Walmsley et al., 1996; Chen et al., 1999; Feipel et al., 1999), with increasing age there is a reduction in measures of active cervical mobility in the cardinal planes. In contrast, the present study concluded that age did not significantly influence mobility during the FRT. This finding is supported by previous studies (Dvorak et al., 1992; Walmsley et al., 1996; Castro et al., 2000). One explanation for this is that the upper cervical spine undergoes minimal age-related degenerative changes, in comparison to joints lower in the cervical spine (Dvorak et al., 1992; Dvorak, 1998). As the FRT is purported to measure C1/2 mobility, age-related degenerative changes in the lower cervical spine should not influence the mobility during this movement. Therefore, this and other studies’ finding that age does not influence the FRT indirectly adds validity to the assumption that the FRT measures high cervical mobility. The current study found that gender did not influence the variability in either active CROM or the FRT. This finding is supported by a number of studies (Dvorak et al., 1992; Trott et al., 1996; McClure et al., 1998; Mannion et al., 2000). Lifestyle factors including sleep position, time spent sitting and side dominant lifestyle did not appear to influence variability in cervical mobility. These findings are favourable for interpretation of the FRT and active cervical movements. With a non-significant influence of extraneous variables such as lifestyle factors, confidence in the use of the above tests for measurement of impairment in neck disorders can be increased. When interpreting these results, however, validity and reliability of the questionnaire used for measurement of lifestyle factors needs consideration. In the present study, there was no formal investigation of reliability and validity. Future investigations may thus be needed to determine the reliability and validity of such questionnaires for measuring lifestyle factors. To place the results of the present study into a clinical context, it is apparent that ROM during the FRT and

T

aspects of CROM are inherently linked, but only ROM of the FRT is influenced, in part, by the presence of sub-clinical pain. The influence of sub-clinical pain on FRT mobility has not been previously reported. However, it has been shown in subjects with C1/2 dominant cervicogenic headache, determined by manual diagnosis, that FRT mobility was significantly reduced when compared with healthy controls (Hall and Robinson, 2004; Ogince et al., 2006). In the present study, the cervical segmental source of the sub-clinical pain was not clearly identified. Future research should involve measurement of the FRT, in patients with specific lower cervical disorders, e.g., lower cervical zygapophyseal joint pain. This would further investigate the relationship between lower and upper cervical mobility in the presence of a clinical neck condition. In addition, we recognize as a limitation of this study that the FRT is unlikely to completely isolate rotation to C1/2. It would be probable that other cervical levels contribute to some degree to rotation during the test, but the extent of this is unknown. This fact may also be reflected by the moderate correlations found between the FRT, sub-clinical neck pain and cervical active ROM. The mean range rotation measured by the FRT was 841. These results are comparable to Hall and Robinson (2004), who found the mean range of rotation in flexion to be 881, and Amiri et al. (2003), who found a mean of 841. In the case of a group with cervicogenic headache, unilateral ROM during the FRT, on the side of the headache, was 281, 161 less than the control group (Hall and Robinson, 2004). In the presence of sub-clinical pain in our study, unilateral range of movement was 371, 81 less than the control group. In the study by Hall and Robinson (2004), ROM for the FRT was inversely correlated to headache severity. This finding may thus explain why subjects with subclinical pain, presumably less severe than pain for which patients seek treatment, did not have as much

ARTICLE IN PRESS K. Smith et al. / Manual Therapy 13 (2008) 552–559

ROM restriction as subjects with cervicogenic headache found in previous studies. Although this may have been the case, 81 is a significant amount of movement loss that we suggest can be detected in a clinical environment. Lee et al. (2005a, b) examined CROM in three population groups including a clinical neck pain group, sub-clinical pain group and control group. Their study found a significant difference in ROM between the control group and pain groups. Conversely, no difference was found in ROM between the clinical pain and sub-clinical pain groups. From these results, one may conclude that sub-clinical pain and clinical neck pain do not differ in terms of ROM values. Assuming the above conclusion, the effect of sub-clinical pain on FRT mobility may be applicable to a clinical neck pain population. Lee et al. (2004) found statistically significant differences in active CROM between sub-clinical neck pain subjects and controls, with a mean difference of 51 for left rotation and 31 for extension. Although the current study found similar differences, multiple linear regression did not conclude that sub-clinical pain significantly influenced the variability in active cervical mobility. In contrast, the current study found subclinical neck pain to influence the variability in the FRT. Such findings may support the use of the FRT as an outcome measure in the evaluation of sub-clinical populations. It was not surprising to find that FRT mobility, and active CROM, in part, influences one another. As previously stated, movement during the FRT is believed to be localized to the upper cervical region and possibly

557

targeted at C1/2. It is known that mobility at C1/2 significantly contributes to overall cervical motion, particularly rotation and lateral flexion (Dvorak and Panjabi, 1987; Mimura et al., 1989); hence, lack of C1/2 mobility is likely to have a substantial influence on overall cervical movement. It has been documented that rotation and lateral flexion of the neck are intimately coupled, but in opposing patterns, in the upper and lower cervical spine (Mercer and Bogduk, 2001). For example, lateral flexion of the whole cervical spine induces coupling of the high cervical spine into contralateral rotation to allow the head to remain facing forward during this movement (Mercer and Bogduk, 2001). Hence, range of lateral flexion is limited by how much the C1/2 segment can rotate to compensate for the coupled rotation in the sub-axial cervical spine.

5. Conclusion This study found that lifestyle factors do not influence active CROM and the FRT. When examining active cervical movements, the presence of pain, gender and FRT mobility should be considered as a contributing factor to ROM deficits.

Acknowledgements The authors wish to thank Dr. Ritu Gupta, Department of Mathematics, Curtin University of Technology, for statistical advice.

ARTICLE IN PRESS 558

K. Smith et al. / Manual Therapy 13 (2008) 552–559

Appendix A. Activity and lifestyle questionnaire

References Amiri M, Jull G, Bullock-Saxton J. Measuring range of active cervical rotation in a position of full head flexion using the 3D Fastrak measurement system: an intra-tester reliability study. Manual Therapy 2003;8(3):176–9. Aprill C, Axinn M, Bogduk N. Occipital headaches stemming from the lateral atlanto-axial (C1–2) joint. Cephalalgia 2002;22:15–22. Bland J. Anatomy and pathology of the cervical spine. In: Singer K, Giles LGF, editors. Clinical Anatomy and Management of Cervical Spine Pain. Oxford: Butterworth Heinemann; 1998.

Bogduk N. Cervicogenic headache: anatomical basis and pathophysiological mechanisms. Journal of Current Pain Headache Report 2001;5:382–6. Bovim G, Schrader H, Sand T. Neck pain in the general population. Spine 1994;19(12):1307–9. Capuano-Pucci D, Rheault W, Aukai J, Bracke M, Day R, Pastrick M. Intratester and intertester reliability of the cervical range of motion device. Archives of Physical Medicine and Rehabilitation 1991;72(5):338–40. Castro WH, Sautmann A, Schilgen M, Sautmann M. Noninvasive three-dimensional analysis of cervical spine motion in normal

ARTICLE IN PRESS K. Smith et al. / Manual Therapy 13 (2008) 552–559 subjects in relation to age and sex. An experimental examination. Spine 2000;25(4):443–9. Chen J, Solinger AB, Poncet JF, Lantz CA. Meta-analysis of normative cervical motion. Spine 1999;24(15):1571–8. Cote P, Cassidy JD, Carroll L. The Saskatchewan health and back pain survey. The prevalence of neck pain and related disability in Saskatchewan adults. Spine 1998;23(15):1689–98. Dvorak J. Epidemiology, physical examination, and neurodiagnostics. Spine 1998;23(24):2663–73. Dvorak J, Panjabi M. Functional anatomy of the alar ligaments. Spine 1987;12:183–9. Dvorak J, Antinnes JA, Panjabi M, Loustalot D, Bonomo M. Age and gender related normal motion of the cervical spine. Spine 1992; 17(Suppl. 10):S393–8. Feipel V, Rondelet B, Le Pallec J, Rooze M. Normal global motion of the cervical spine: an electrogoniometric study. Clinical Biomechanics (Bristol, Avon) 1999;14(7):462–70. Gordon SJ, Trott P, Grimmer KA. Waking cervical pain and stiffness, headache, scapular or arm pain: gender and age effects. Australian Journal of Physiotherapy 2002;48(1):9–15. Grant R, Forrester C, Hides J. Screen based keyboard operation: the adverse effects on the neural system. Australian Journal of Physiotherapy 1995;41(2):99–107. Guth E. A comparison of cervical rotation in age-matched adolescent competitive swimmers and healthy males. Journal of Orthopaedic and Sports Physical Therapy 1995;21:21–7. Hall T, Robinson K. The flexion–rotation test and active cervical mobility—a comparative measurement study in cervicogenic headache. Manual Therapy 2004;9(4):197–202. Hall T, Robinson K, Fujinawa O, Kiyokazu A. The influence of examiner experience on interpretation of the cervical flexion– rotation test. In: Proceedings of the 15th International World Confederation for Physical Therapy, Vancouver, 2007. Lee H, Nicholson LL, Adams RD. Cervical range of motion associations with subclinical neck pain. Spine 2004;29(1):33–40. Lee H, Nicholson LL, Adams RD. Neck muscle endurance, selfreport, and range of motion data from subjects with treated and untreated neck pain. Journal of Manipulative and Physiological Therapeutics 2005a;28(1):25–32. Lee H, Nicholson LL, Adams RD, Bae SS. Proprioception and rotation range sensitization associated with subclinical neck pain. Spine 2005b;30(3):E60–7. Maitland G, Hengeveld E, Banks K, English K. Maitland’s vertebral manipulation. London: Butterworth Heinemann; 2001. Malmstrom EM, Karlberg M, Melander A, Magnusson M. Zebris versus Myrin: a comparative study between a three-dimensional

559

ultrasound movement analysis and an inclinometer/compass method: intradevice reliability, concurrent validity, intertester comparison, intratester reliability, and intraindividual variability. Spine 2003;28(21):E433–40. Mannion AF, Klein GN, Dvorak J, Lanz C. Range of global motion of the cervical spine: intraindividual reliability and the influence of measurement device. European Spine Journal 2000; 9(5):379–85. McClure P, Siegler S, Nobilini R. Three-dimensional flexibility characteristics of the human cervical spine in vivo. Spine 1998; 23(2):216–23. Mercer S, Bogduk N. Joints of the cervical vertebral column. Journal of Orthopaedic and Sports Physical Therapy 2001;31(4): 174–82. Mimura M, Moriya H, Watanabe T. Three-dimensional motion analysis of the cervical spine with special reference to the axial rotation. Spine 1989;14:1135–9. Ogince M, Hall T, Robinson K. The diagnostic validity of the cervical flexion–rotation test in C1/2 related cervicogenic headache. Manual Therapy 2006. Ordway NR, Seymour R, Donelson RG, Hojnowski L, Lee E, Edwards WT. Cervical sagittal range-of-motion analysis using three methods. Cervical range-of-motion device, space, and radiography. Spine 1997;22(5):501–8. Rheault W, Albright B, Byers C, Franta M, Johnson A, Skowronek M, et al. Intertester reliability of the cervical range of motion device. Journal of Orthopaedic and Sports Physical Therapy 1992;15(3):147–50. Tousignant M, de Bellefeuille L, O’Donoughue S, Grahovac S. Criterion validity of the cervical range of motion (CROM) goniometer for cervical flexion and extension. Spine 2000;25(3): 324–30. Tousignant M, Duclos E, Lafleche S, Mayer A, Tousignant-Laflamme Y, Brosseau L, et al. Validity study for the cervical range of motion device used for lateral flexion in patients with neck pain. Spine 2002;27(8):812–7. Trott PH, Pearcy MJ, Ruston SA, Fulton I, Brien C. Threedimensional analysis of active cervical motion: the effect of age and gender. Clinical Biomechanics (Bristol, Avon) 1996;11(4): 201–6. Walmsley RP, Kimber P, Culham E. The effect of initial head position on active cervical axial rotation range of motion in two age populations. Spine 1996;21(21):2435–42. Youdas JW, Garrett TR, Suman VJ, Bogard CL, Hallman HO, Carey JR. Normal range of motion of the cervical spine: an initial goniometric study. Physical Therapy 1992;72(11):770–80.

ARTICLE IN PRESS

Manual Therapy 13 (2008) 560–563 www.elsevier.com/math

Case report

Concepts for assessment and treatment of anterior knee pain related to altered spinal and pelvic biomechanics: A case report Andrew Thomas Connell Queensland Health, Sunshine Coast and Cooloola Area Health Service, Australia Received 13 June 2007; received in revised form 18 December 2007; accepted 21 December 2007

1. Introduction

2.1. History

Anterior knee pain is a musculoskeletal condition that is associated with symptoms including pain around the region of the patellofemoral articulation and can be aggravated by activities that increase load through the patellofemoral joint (McConnell, 1986). Differentiation of the cause of anterior knee pain based on location of pain is difficult since many conditions are characterized by pain around the patella. Furthermore, aggravating activities including squatting, kneeling, sitting and ascending or descending stairs are similar across numerous conditions. Most clinical practice draws on evidencebased local interventions including quadriceps muscle re-education and patellar realignment procedures including taping and stretching in an attempt to optimize the biomechanics of the patellofemoral joint (Crossley et al., 2001; Cowan et al., 2001; Hinman et al., 2003). In the reported case of a patient with anterior knee pain, improvements in knee pain and range of motion during squatting were found immediately following mobilization of the thoracic spine and the lumbosacral junction. It is the purpose of this case report to explore reasons behind this improvement and to discuss the implications for assessment and management.

The patient recalled falling a few metres from a rope swing 5 weeks prior to the consultation. She described landing in a way that the right foot struck the ground creating a valgus force around the knee. She experienced immediate pain which ‘‘shot’’ up the lower and upper leg. A bruise appeared immediately on the medial side of the right tibial condyle. Standard radiographs of the right knee revealed no bony, articular or soft tissue damage. She had not experienced previous problems with the right knee. There was no previous history of spinal dysfunction or trauma.

2. Case 1

Due to the nature of the injury, priority was given to establish the integrity of knee ligamentous restraints. Palpation revealed tenderness of the medial collateral ligament (MCL) of the knee at the joint line. Superficial fibres of the MCL were also tender. There was some swelling over the medial joint line but no increase in temperature. Active and passive knee range of motion was full and pain-free to overpressure if applied slowly.

This patient was a healthy 37 year old female teacher’s aide who presented with right patellofemoral pain, right anterolateral thigh pain, anterior right shin pain and anterolateral talofibular pain. E-mail address: consie@flexinet.com.au 1356-689X/$ - see front matter r 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2007.12.003

2.2. Subjective examination The patient reported knee pain as illustrated on full and partial weight bearing. Negotiating stairs was no worse than walking. There was occasional sharp P2 causing a giving-way sensation at the right knee. There was less P1, and 2 upon waking but all were immediately noticeable upon full weight-bearing out of bed. Both pains were felt intermittently. 2.3. Objective examination

ARTICLE IN PRESS A.T. Connell / Manual Therapy 13 (2008) 560–563

There was mild laxity on testing of the superficial fibres of the MCL. Squat allowed only 90% range of motion compared to the left knee and a painful block was reported in the popliteal fossa. The discrepancy between the knees on squat could have occurred due to pressure changes in the knee joint associated with weight bearing. Slow squat to 151 knee flexion showed slight adduction of the right knee towards the midline suggesting potential muscle control issues around the hip and pelvis. Motion of the right sacroiliac was assessed using Gillet’s test. Gillet’s test is performed in single leg stance. The displacement of the posterior superior iliac spine caudally on the side of the flexing hip is palpated relative to the position of the S1 prominence on the sacrum (Lee, 2004). The test revealed range of movement of the right posterior superior iliac spine (PSIS) inferiorly limited to 90% when compared to the left side suggesting an articular restriction of the right sacroiliac joint. Apart from the vague anterolateral thigh pain and the injury being an impact with the ground, there was no other suggestion that the source of the pains P1 and P2 were from any proximal source such as the hip or spine. Therefore, lumbar spine active movements were not examined at this point in time. Examination of passive right hip motion revealed tighter end feel on combined flexion to 901, adduction and external rotation quadrant.

561

Examination of the position of the right patella suggested no difference in alignment at rest or during active knee extension when compared to the left patella. There was palpable muscle spasm in the right popliteal area. The difference seen on the right Gillet’s test and the internal rotation of the right femur during squat suggested potential dysfunction around the region of the right pelvis and hip. However, in the presence of localized knee swelling and tenderness over the medial joint line structures following a direct trauma to the area, it was decided that local disruption of knee joint structures was the probable source of pain and swelling and that local measures be undertaken to address these issues. 2.4. Treatment With a strong case for primary disruption of the medial collateral ligament of the right knee as the source of symptoms, treatment was instigated to alleviate swelling and pain over the medial joint line structures and relieve muscle spasm in the popliteal region. This included local massage, contract/relax techniques for both hamstring and quadriceps muscle groups and exercises in supine and standing to facilitate quadriceps muscle action with attention paid to the equal recruitment of all vasti muscles.

ARTICLE IN PRESS 562

A.T. Connell / Manual Therapy 13 (2008) 560–563

After each treatment session, there was a temporary slight increase in range of squat. However, after 2 sessions squat had returned to pre-treatment range, the subjective feeling of restriction in the popliteal region had worsened and there was the return of giving way in the knee as the patient had previously described. A decision was made to assess thoracic, lumbar and pelvic function. Examination revealed an exaggerated thoracic spine kyphosis above T10/11 with lack of contribution to active spinal extension from this region. Passive accessory examination revealed intersegmental muscle spasm at T11/12 and T12/L1. There was also palpable restriction of L5 and sacral motion in a poster anterior direction and a caudal direction. Caudal pressure on the right ilium revealed stiffness compared to the left. Over the following 3 sessions, mobilisation of the levels described above was pursued bilaterally (Grade IV) until spasm resolved at T11/12 and T12/L1 and until stiffness resolved around the right L5/S1 and sacroiliac joint. During the first treatment session, re-assessment of the squat, revealed an immediate increase in range of right knee flexion and lessening of the restriction in the popliteal region. Over the next 2 sessions, treatment consisted only of mobilization of the spinal joints and the sacroiliac joints as outlined. The squat cleared to full range with the patient able to bounce repetitively in full knee flexion. Swelling had disappeared and there was no further pain experienced in the leg. A follow-up phone call one month after the last treatment session revealed that the patient had experienced no further right knee pain or dysfunction. A home exercise programme was then encouraged over the following few weeks to promote optimal biomechanics and muscle control around the right hip and pelvis.

3. Discussion In this patient, the immediate improvement in range of motion and knee pain on squat due to manual therapy applied further away at the spine and pelvis suggests that the aetiology of the anterior knee pain in this case may have been due to complex interactions involving abnormal biomechanics and muscle control across the lumbar spine, pelvis, hip and knee (Tyler et al., 2006). History of onset of pain may give clues as to the aetiology of anterior knee pain, although in this patient the precise mechanism is not absolutely clear. Local tenderness and laxity over the medial knee joint line suggested primary trauma to the knee may have occurred during the fall. However, it is also possible that the fall caused local muscle spasm at the thoracic and lumbar levels described above and anterior knee pain developed secondary to that. Whatever the case, it is important to understand how manual therapy in this patient changed knee pain where 2 sessions of local treatment had previously not.

Evidence indicates that spinal manipulation may impact primary afferent neurons from paraspinal tissues, the motor control system and pain processing (Floman et al., 1997; Dishman and Bulbian, 2000; Keller and Colloca, 2000). Lehman et al. (2001) and Herzog et al. (1995) demonstrated a decrease in bilateral local electromyographic (EMG) activity in the painful motion segment during the application of spinal manipulation. Perhaps relaxation of tight spinal muscles in this patient altered positional alignment or control of the lumbar spine and the pelvis and subsequently affected both muscle control of the femur and range of flexion at the hip joint. In this patient there was an increase in active hip range during squat, passive hip range during hip quadrant and less internal rotation of the femur during the first 151 of squat. Gill et al. (2007) in a single case study documented a change in one patient’s ability to perform a preferential transverse abdominus contraction immediately following a regional lumbopelvic manipulation. Similarly, Marshall and Murphy (2006), demonstrated an improvement in the feed forward response in the transverse abdominus muscle following sacroiliac joint manipulation. Could it be that this patient experienced similar improvement in transversus abdominus activity which then improved the squat via improved lumbopelvic muscular stabilization of the pelvis on the moving femur (Hodges and Richardson, 1996; Richardson et al., 2002). Although not measured in this patient, transversus abdominus muscle activity could be measured in the clinic with the use of ultrasound imaging or pressure biofeedback. It is interesting to speculate on other muscles around the pelvis and hips which may experience facilitation following manual therapy to the spine and pelvis. Spinal nerves from T12/L1 to L5/S1 innervate muscles that control the position of the femur, including the quadriceps group, piriformis, gluteals and deeper gemelli muscles (Gray, 2004). There is evidence of change in peripheral muscle activation following spinal manual therapy. Suter et al. (1999) demonstrated a decrease in quadriceps inhibition in patients with anterior knee pain following manipulation of the sacroiliac joint. The exact mechanism by which this occurs is not fully understood. However, it seems probable that the effect of reducing quadriceps inhibition would probably assist in the performance of a squat. Perhaps spinal manual therapy had similar inhibitory and excitatory effects on the intrinsic hip rotator muscles to allow more efficient rotational control of the femur. Concordantly, in this patient, after mobilization of the spine and pelvis there was less palpable restriction of combined hip flexion/ adduction/external rotation quadrant and improved control of femoral rotation during the first 151 of squat. Lee et al. (2003) and Powers (2003) have demonstrated that altered femoral rotation resulted in abnormal

ARTICLE IN PRESS A.T. Connell / Manual Therapy 13 (2008) 560–563

patellofemoral contact areas. Therefore, it is possible that improved hip muscle control of the femur may have improved the squat in this patient via improved patellofemoral biomechanics. Haavik-Taylor and Murphy (2007) assert that the effects of spinal manipulation on the somatosensory reflexes may be quite complex producing both excitatory and inhibitory effects. Furthermore, findings that manual therapy produces concurrent hypoalgesic and sympathoexcitatory effects have lead to the proposal that spinal manual therapy may exert its initial effect by activating descending pain inhibitory pathways from the dorsal peri-aquaductal gray (PAG) area of the midbrain. Stimulation of the dorsal PAG in animals has also been shown to have a facilitatory effect on motor activity (Sterling et al., 2001). It is, therefore, conceivable that manual mobilization of both the thoracic spine and the lumbosacral junction regions in this patient may have had broader effects on knee pain and function via a combination of mechanisms including alteration to central pain processing, thereby allowing a neurophysiological window of opportunity for improved muscle activation.

4. Summary Since threats to internal validity exist, it is not possible to affirm a cause and effect relationship in a case report. However, it seems possible that the improved knee range and reduced knee pain on squat was caused via several different mechanisms including improved lumbopelvic biomechanics and muscle control and perhaps via central pain processing mechanisms. Rather than a simple localised cause for the anterior knee pain in this patient, the aetiological mechanism may have been due to concurrent trauma to the spine and the knee and the ensuing dysfunction an example of a complex relationship across an interactive neuromusculoskeletal system. Therefore, to pursue the most appropriate and comprehensive management of anterior knee pain, accurate assessment and measurement of these interactive systems may be required to ascertain the primary aetiological mechanism. The author is currently involved in a randomized controlled trial to evaluate the relative benefit of localized knee treatment compared to the same localised treatment together with thoracic spine mobilization in patients with anterior knee pain. References Cowan SM, Bennell KL, Hodges PW, Crossley KM, McConnell J. Delayed onset of electromyographic activity of vastus medialis obliquus relative to vastus lateralis in subjects with patellofemoral pain syndrome. Archives of Physical Medicine and Rehabilitation 2001;82:183–9.

563

Crossley K, Bennell K, Green S, Cowan S, McConnell J. A systematic review of physical interventions for patellofemoral pain syndrome. Clinical Journal Of Sports Medicine 2001;11(2):103–10. Dishman JD, Bulbian R. Spinal reflex attenuation associated with spinal manipulation. Spine 2000;25(19):2519–25. Floman Y, Liram N, Gilai AN. Spinal manipulation results in immediate H-reflex changes in patients with unilateral disc herniation. European Spine Journal 1997;6(6):398–401. Gill NW, Teyhan DS, Lee IE. Improved contraction of the transverse abdominus immediately following spinal manipulation: A case study using real-time ultrasound imaging. Manual Therapy 2007; 12(3):280–5. Gray H. Grays Anatomy. In: Susan Standring, editor. 39th Ed. Williams and Warwick; 2004. Haavik-Taylor H, Murphy B. Cervical spine manipulation alters sensorimotor integration: a somatosensory evoked potential study. Clin Neurophysiol 2007;118(2):391–402. Herzog W, Conway PJ, Zhang YT, Gal J, Guimaraes AC. Reflex responses associated with manipulative treatments on the thoracic spine: a pilot study. Journal of Manipulative and Physiological Therapeutics 1995;18(4):233–6. Hinman RS, Bennett KL, Crossley KM, McConnell J. Immediate effects of adhesive tape on pain and disability in individuals with knee osteoarthritis. Rheumatology 2003;42:865–9. Hodges PW, Richardson CA. Inefficient muscular stabilization of the lumbar spine associated with low back pain. A motor control evaluation of transversus abdominis. Spine 1996;21(22): 2640–50. Keller TS, Colloca CJ. Mechanical force spinal manipulation increases trunk muscle strength assessed by electromyography: a comparative clinical trial. Journal of Manipulative and Physiological Therapeutics 2000;23(9):585–95. Lee D. The Pelvic Girdle: An approach to the examination and treatment of the lumbopelvic- hip region. 3rd ed. Edinburgh: Churchill Livingston; 2004 [Chapter 8, p. 87–8]. Lee TQ, Morris G, Csintalan RP. The influence of tibial and femoral rotation on patellofemoral contact area and pressure. Journal of Orthopaedic and Sports Physical Therapy 2003;33:686–93. Lehman GJ, Vernon H, McGill SM. Effects of Mechanical Pian Stimulus on erector spinae activity before and after a spinal manipulation in patients with back pain: a preliminary investigation. J Manipulative Physiol Ther 2001;24(6):402–6. Marshall P, Murphy B. The effect of sacroiliac joint manipulation on feed-forward activation times of the deep abdominal musculature. Journal of Manipulative and Physiological Therapeutics 2006; 29(3):196–202. McConnell J. The management of chondromalacia patellae: a long term solution. Australian Journal of Physiotherapy 1986;32: 215–23. Powers CM. The influence of lower- extremity kinematics on patellofemoral joint dysfunction: A theoretical perspective. Journal of Orthopaedic and Sports Physical Therapy 2003;33:639–46. Richardson CA, Snijders CJ, Hides JA, Damen L, Pas MS, Storm J. The relationship between the transversus abdominus muscles , sacroiliac joint mechanics and low back pain. Journal Of manipulative and Physiological Therapeutics 2002;27(4):399–405. Sterling M, Jull G, Wright A. Cervical mobilization: concurrent effects on pain, sympathetic nervous system activity and motor activity. Manual Therapy 2001;6(2):72–81. Suter E, McMorland G, Herzog W, Bray R. Decrease in quadriceps inhibition after sacroiliac joint manipulation in patients with anterior knee pain. Journal Of manipulative and Physiological Therapeutics 1999;22(3):149–53. Tyler TF, Nicholas SJ, Mullaney MJ, McHugh MP. The role of hip muscle function in the treatment of patellofemoral pain syndrome. American Journal of Sports Medicine 2006;34(4):630–6.

Manual Therapy (2008) 13(6), 564

Available online at www.sciencedirect.com

Diary of events

Back and beyond Theme The lumbar spine and pelvis Dates Sat 28–29th March 2009 Venue East Midlands Conference Centre, Nottingham For more details visit www.physiofirst.org.uk

MSc Masterclass Hypermobility and the Hypermobility Syndrome Facilitators include: Professor Rodney Grahame CBE, MD, FRCP, FACP Rosemary Keer MSc, MMACP and Jane Simmonds BApp Sc, PGDip Man Ther, MA, MMACP Date: 11th and 12th December, 2008 Venue: University of Hertfordshire, Hatfield, Herts, UK Cost: 110 5 M level credits are available for those who undertake the assessment (subject to approval) Details and bookings contact: Jane Simmonds j.1.simmonds @herts.ac.uk or Karen Wells [email protected]

NOI International conference UK and Ireland Nottingham UK - April 15–17, 2010 Dublin IRELAND April 21–23, 2010 For further details www.noi2010.com Fax + 3906 51882443

2nd World Congress on Manual Therapy and Sport Rehabilitation, The Spine II, in Roma Italy 6th–8th of February 2009 www.newmaster.it Tel: +3906 51600107 Fax: +3906 51882443

Janet G. Travell, MD Seminar Series, Bethesda, USA For information, contact: Myopain Seminars, 7830 Old Georgetown Road, Suite C-15, Bethesda, MD 20814-2432, USA. Tel.: +1 301 656 0220; Fax: +1 301 654 0333; website: www.painpoints.com/seminars.htm E-mail: [email protected]

MSc Conference Day Lifestyle, Physical Activity and Health Learn more about contemporary health promotion issues by attending this topical interactive conference day. The day will include topical presentations delivered by public health professionals including; managers and strategists, clinicians and researchers. Date: 23rd February, 2009 Venue: University of Hertfordshire, Hatfield, Herts, UK or virtual attendance from your home or workplace via Elluminate. Cost: 120 or 60 for virtual attendance Details and bookings contact: Jane Simmonds j.1.simmonds @herts.ac.uk or Karen Wells [email protected]

If you wish to advertise a course/conference, please contact: Karen Beeton, Associate Head of School (Professional Development), School of Health and Emergency Professions, University of Hertfordshire, College Lane, Hatfield, Herts AL10 9AB, UK. E-mail: [email protected] There is no charge for this service.

564