Manual Therapy Journal - Volume 14, Issue 6, Pages 585-716 (December 2009)

VOLUME 14 NUMBER 6 PAGES 585–716 December 2009 Editors Ann Moore PhD, GradDipPhys, FCSP, CertEd, FMACP Clinical Researc...

Author: Editors: Ann Moore and Gwen Jull

383 downloads 1256 Views 7MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

VOLUME 14 NUMBER 6 PAGES 585–716 December 2009

Editors 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 Gwendolen Jull PhD, MPhty, Grad Dip ManTher, FACP Department of Physiotherapy University of Queensland Brisbane QLD 4072, Australia

Associate Editor’s Deborah Falla PhD, BPhty(Hons) Department of Health Science and Technology Aalborg University, Fredrik BajersVej 7, D-3, DK-9220 Aalborg Denmark Email:[email protected] Tim McClune D.O. Spinal Research Unit. University of Huddersfield 30 Queen Street Huddersfield 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: [email protected] Masterclass Editor Karen Beeton PhD, MPhty, BSc(Hons), MCSP MACP ex officio member Associate Head of School (Professional Development) School of Health and Emergency Professions University of Hertfordshire College Lane Hatfield AL10 9AB, UK E-mail: [email protected] Case Reports & Professional Issues Editor Jeffrey D. Boyling MSc, BPhty, GradDipAdvManTher, MCSP, MErgS Jeffrey Boyling Associates Broadway Chambers Hammersmith Broadway London W6 7AF, UK E-mail:[email protected] Book Review Editor Raymond Swinkels PhD, PT, MT Ulenpas 80 5655 JD Eindoven The Netherlands E-mail: [email protected]

International Advisory Board K. Bennell (Melbourne, Australia) K. Burton (Huddersfield, UK) B. Carstensen (Frederiksberg, Denmark) T. Chiu (Kowloon, Hong Kong) J. Cleland (Concord, NH, USA) M. Coppieters (Brisbane, Australia) E. Cruz (Setubal, Portugal) L. Danneels (Maríakerke, Belgium) I. Diener (Stellenbosch, South Africa) S. Durrell (London, UK) S. Edmondston (Perth, Australia) L. Exelby (Biggleswade, UK) J. Greening (London, UK) A. Gross (Hamilton, Canada) T. Hall (Perth, Australia) W. Hing (Auckland, New Zealand) M. Jones (Adelaide, Australia) 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. Maffey-Ward (Calgary, Canada) E. Maheu (Quebec, Canada) C. McCarthy (Coventry, UK) J. McConnell (Northbridge, Australia) S. Mercer (Brisbane, Australia) P. Michaelson (Luleå, Sweden) D. Newham (London, UK) J. Ng (Hung Hom, Hong Kong) S. O’Leary (Brisbane, Australia) 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) E. Rasmussen Barr (Stockholm, Sweden) D. Reid (Auckland, New Zealand) D.A. Rivett (Callaghan, Australia) A. Rushton (Birmingham, UK) M. A. Schmitt (Amersfoort, The Netherlands) M. Shacklock (Adelaide, Australia) D. Shirley (Sydney, Australia) C. Snijders (Rotterdam, The Netherlands) P. Spencer (Barnstaple, UK) M. Sterling (Brisbane, Australia) M. Stokes (Southampton, UK) P. Tehan (Melbourne, Australia) M. Testa (Alassio, Italy) P. van der Wurff (Doorn, The Netherlands) P. van Roy (Brussels, Belgium) O. Vasseljen (Trondheim, Norway) B.Vicenzino (Brisbane, Australia) M. Wessely (Paris, France) A. Wright (Perth, Australia) M. Zusman (Perth, Australia)

Visit the journal website at http://www.elsevier.com/math doi:10.1016/S1356-689X(09)00161-1

Manual Therapy 14 (2009) 585

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Editorial

World Congress of Physical Therapy 2011 and IFOMT 2012

As January 2010 approaches the build-up to two signiﬁcant conference events is in ﬂow. These events are the World Congress of Physical Therapy to be held June 20th to June 23rd 2011 in Amsterdam in Holland, and IFOMT 2012 which will be held between 30th September and 5th October in Quebec, Canada. Both these events will showcase research and developments in practice, but tend to attract different audiences who attend for different reasons and purposes. Whilst IFOMT attracts physical therapists, osteopaths and chiropractors specialising or wishing to specialise in musculoskeletal therapy and sets increasingly high standards in terms of research rigour and endeavour, the World Confederation of Physical Therapy as an organisation, is committed to taking forward physical therapy as a profession and its contribution to global health. The Confederation promotes and encourages high standards of physical therapy research, education and practice. Globally, physical therapy is in very different stages of development, for example, in several countries, physical therapists act as ﬁrst contact practitioners, some are holding consultant posts in hospital settings and working almost autonomously, whilst other physiotherapists in some parts of the world are struggling to achieve degree status and ﬁrst contact rights. Wherever physical therapists are based we all face issues and challenges in day to day practice and in professional status and development. Some physical therapists face incredible personal and professional challenges when working, for example, in areas of conﬂict and severe hardship, dealing with patients who are mal-nourished and who are also suffering other consequences of living in poverty-stricken areas. Many of us working in western societies are perhaps lulled into our own senses of security and are not exposed to the difﬁculties that some of our colleagues face in different parts of the world. Research in physical therapy and standards of education and practice are at different levels of maturity in different parts of the world. Anything that can be done as a profession to reduce this diversity can only help strengthen the profession globally. I was very honoured to be elected to be Chair of the International Scientiﬁc Committee for the World Congress of Physical Therapy 2011. I am highly committed, as is the rest of the International Scientiﬁc Committee, to provide a conference in 2011 which celebrates physical therapy as a whole global community which, whilst recognising specialism and distinct areas of practice, also recognises the importance of sharing ideas, new knowledge, new approaches to practice, education and research. The committee is also aiming to construct a conference which works towards enabling all those attending to cross specialisation boundaries, cross international divides and enables the

1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.10.004

whole of the physical therapy worldwide community to grow intellectually. For some this will be via the sheer exposure to new research ﬁndings and for some it will be by gaining insights into other specialities, other ways of working, but also by recognising what obstacles physical therapists in some countries face and how they are dealing with these obstacles. The conference should be a learning experience for everyone, whether you come as a top international researcher, or a new graduate physical therapist from a country where physiotherapy is still in its early developmental stages. There is something at WCPT for everyone. The World Confederation of Physical Therapy represents over 300,000 physical therapists worldwide and has 101 member organisations. The World Congress is expected to attract around 3500 delegates, providing a rich and vibrant forum for networking and debate and the opportunity to contribute to other physical therapists’ knowledge base development. WCPT 2011 will be held at the Rai Exhibition and Conference Centre in Amsterdam and the Royal Dutch Society for Physical Therapy (KNGF) is hosting the conference. Please visit the WCPT website for ongoing information. Importantly, the call for abstract submissions for the 2011 conference will be open in January 2010, so think now about submitting an abstract for a platform, poster or interactive poster presentation. Additionally, other sessions will be available at WCPT 2011, for example focused symposia and satellite programmes which will have a strong international focus as will debating sessions, workshops and of course a range of networking and social activities. Put WCPT 2011 in your calendar now! If you have not attended a WCPT congress before, it will be a new experience and one hopefully not to forget. The International Scientiﬁc Committee members will be doing its very best to provide a memorable conference for everyone and we will look forward to seeing you there. And whilst you are getting your diaries out to put in the WCPT 2011 dates, please also put the dates in your diary for IFOMT 2012. In the meantime, Seasons Greetings to All. Ann Moore, Executive Editor, Chair of the International Scientiﬁc Committee, WCPT 2011* Director Clinical Research Centre for Health Professions, Aldro Building, 49, Darley Road, Eastbourne BN20 7UR, United Kingdom Tel.: þ44 1273 643 766; fax: þ44 1273 643 944. E-mail address: [email protected]

Manual Therapy 14 (2009) 586–595

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Masterclass

Thoracic outlet syndrome part 1: Clinical manifestations, differentiation and treatment pathways L.A. Watson a, b, T. Pizzari b, *, S. Balster a a b

LifeCare Prahran Sports Medicine Centre, 316 Malvern Road, Prahran, VIC 3181, Australia Musculoskeletal Research Centre, La Trobe University, Bundoora VIC 3086, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 12 January 2009 Received in revised form 7 July 2009 Accepted 10 August 2009

Thoracic outlet syndrome (TOS) is a challenging condition to diagnose correctly and manage appropriately. This is the result of a number of factors including the multifaceted contribution to the syndrome, the limitations of current clinical diagnostic tests, the insufﬁcient recognition of the sub-types of TOS and the dearth of research into the optimal treatment approach. This masterclass identiﬁes the subtypes of TOS, highlights the possible factors that contribute to the condition and outlines the clinical examination required to diagnose the presence of TOS. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Thoracic outlet syndrome Entrapment neuropathy Classiﬁcation Diagnosis

1. Introduction

2. Deﬁnition

Opinions in the literature about thoracic outlet syndrome (TOS) vary in the extreme, swaying from the belief that it is the most underrated, overlooked and misdiagnosed peripheral nerve compression in the upper extremity (Shukla and Carlton, 1996; Sheth and Belzberg, 2001) to questioning whether it exists (Wilbourn, 1990). These varying beliefs highlight the need for the clinician to be rigorous in their clinical assessment so that patients are not misdiagnosed and are appropriately managed. Unfortunately the diagnosis of TOS remains essentially clinical and is often one of exclusion with no investigation being a speciﬁc predictor. This may be attributed, in part, to the fact that TOS is considered to be a collection of quite diverse syndromes rather than a single entity (Yanaka et al., 2004). Consequently, this also results in TOS being one of the most difﬁcult upper limb conditions to manage. The aim of this paper (Part 1) is to clarify the nomenclature, classiﬁcation, varying clinical presentations and assessment techniques so that the reader may attempt to assess and differentially diagnose patients presenting with TOS. The second paper (Part 2) will outline speciﬁc rehabilitation approaches used by the authors to treat one sub-type of TOS.

A broad deﬁnition of TOS is a symptom complex characterized by pain, paresthesia, weakness and discomfort in the upper limb which is aggravated by elevation of the arms or by exaggerated movements of the head and neck (Lindgren and Oksala, 1995).

* Corresponding author. Tel.: þ61 3 94795872. E-mail address: [email protected] (T. Pizzari). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.08.007

3. Anatomical considerations The pain and discomfort of TOS are generally attributed to the compression of the subclavian vein, subclavian artery and the lower trunk of the brachial plexus as they pass through the thoracic outlet (Cooke, 2003; Samarasam et al., 2004; Barkhordarian, 2007). Three sites of compression of the vessels and nerves are possible (Fig. 1). The lower roots of the brachial plexus may be compressed as they exit from the thoracic cavity and rise up over the ﬁrst rib (or a cervical rib or band when present) and pass between the anterior and middle scalene muscles (or even sometimes through the anterior scalene muscle). The upper roots of the brachial plexus can also be compressed between the scalene muscles but actually exit the cervical spine not the thorax, and should technically be referred to as cervical outlet syndrome (Ranney, 1996). The second potential site of entrapment is beneath the clavicle in the costoclavicular space, where the neural elements are already outside the thorax. The third potential site is more distal in the sub-coracoid tunnel (beneath the tendon of the pectoralis minor) where the plexus may be stretched by shoulder abduction (Ranney, 1996; Rayan, 1998; Demondion et al., 2003; Wright and Jennings, 2005).

L.A. Watson et al. / Manual Therapy 14 (2009) 586–595

587

The classiﬁcations for vTOS, tnTOS, and sTOS can be seen in Table 1. 5. Incidence

Fig. 1. Thoracic outlet anatomy. Three possible site of compression and structures compressed; A: Subclavian artery and lower roots of the brachial plexus may be compressed as they exit from the thoracic cavity and rise up over the ﬁrst rib and pass between the anterior and middle scalene muscles. B: Subclavian artery and vein and/or lower trunk of the brachial plexus beneath the clavicle in the costoclavicular space. C: The axillary artery and/or vein and/or one of the cords of the brachial plexus in the sub-coracoid tunnel.

Very rarely is this clariﬁed in the literature and the reader should be aware that many authors utilize the global term of TOS with little attempt to differentiate which sub-type of TOS they are treating. This may well account for the enormous variation in treatment outcomes described. We believe it is essential that the clinician carefully consider and at least attempt to clinically differentiate, where possible, exactly which component of the neurovascular complex is being affected and precisely where it is being compressed. This will direct not only what further investigations are required, but may well impact on what is the most appropriate treatment strategy. In reality this is often easier said than done. 4. Classiﬁcation and pathophysiology TOS is often categorized into two speciﬁc clinical entities: Vascular TOS (vTOS) and Neurological TOS (nTOS) (Atasoy, 1996; Rayan, 1998; Sharp et al., 2001). vTOS can be divided into arterial and venous TOS syndromes due to compression or angulation of either the subclavian or axillary artery or vein (Rayan, 1998; Davidovic et al., 2003). Usually it is caused by a structural lesion, either a cervical rib or another bony anomaly (Rayan, 1998). Arterial involvement is more common than venous involvement (Davidovic et al., 2003; Singh, 2006) and vTOS is generally easier to deﬁne, diagnose and treat than nTOS (Sharp et al., 2001). The subset of patients with bony abnormalities such as cervical ribs, are generally accepted as ‘‘true cases’’ of TOS and this commonly occurs in vTOS and true neurological TOS (tnTOS) (Roos, 1982; Samarasam et al., 2004). tnTOS is caused by irritation, compression or traction of the brachial plexus. The remaining larger group of patients with no radiological or electro-physical abnormalities are usually labeled as ‘‘disputed TOS’’ (Cherington, 1989) or ‘‘non-speciﬁc nTOS’’ (Sobey et al., 1993) or ‘‘symptomatic TOS’’ sTOS (Rayan, 1998). sTOS remains the most controversial form of TOS. There has been some suggestion that this may be an early or mild form of vTOS or nTOS and hence may mimic the symptoms with no deﬁnitive evidence for either (Rayan, 1998; Seror, 2005; Lee et al., 2006). In some cases patients may present with ‘‘combined TOS,’’ the simultaneous compression of vascular and neurological structures. This may be mixture of arterial and venous or arterial/venous and neurological or all three.

The incidence of TOS is reported to be approximately 8% of the population (Davidovic et al., 2003), is extremely rare in children (Cagli et al., 2006), and affects females more than males (between 4:1 and 2:1 ratios) (Gockel et al., 1994; Davidovic et al., 2003; Demondion et al., 2003; Degeorges et al., 2004). In particular, tnTOS is typically found in young women (van Es, 2001). According to Davidovic et al. (2003), 98% of all patients with TOS fall into the nTOS category and only 2% have vTOS. However this ﬁgure is clouded by the fact there is no distinction between tnTOS and sTOS (Urschel et al., 1994; Urschel and Razzuk, 1997; Goff et al., 1998). While neurological symptoms appear more prominently, the majority of these will fall into the sTOS classiﬁcation (Wilbourn, 1990; Rayan, 1998; Davidovic et al., 2003). 6. Etiology Bony pathology or soft tissue alterations are commonly attributed to the etiology of TOS. Numerous causes have been cited in the literature ranging from congenital abnormalities (anomalies of the transverse process of seventh cervical vertebra, cervical rib, ﬁrst rib, enlarged scalene tubercle, scalene muscles, costoclavicular ligaments, subclavius or pectoralis minor) to traumatic in origin (such as a motor vehicle accident or sporting incident) (Gruber, 1952; Makhoul and Machleder, 1992; Rockwood et al., 1997; Athanassiadi et al., 2001; Jain et al., 2002; Barkhordarian, 2007). Cervical ribs and other anatomic variations are not prerequisites for the diagnosis of TOS but may be implicated in some cases. Traumatic bony lesions include bone remodeling after fractures of the clavicle or ﬁrst rib or posterior subluxation of the acromioclavicular joint. Soft tissue pathologies such as anterior scalene muscle hypertrophy, muscle ﬁbre type adaptive transformation, spasm and excessive contraction particularly post cervical trauma have all been implicated in TOS (Roos, 1982; Machleder et al., 1986; Mackinnon, 1994; Schwartzman and Maleki, 1999; Kai et al., 2001; Pascarelli and Hsu, 2001; Davidovic et al., 2003). Less commonly, upper lung tumors have been implicated in the etiology (Machleder et al., 1986; Makhoul and Machleder, 1992; Barkhordarian, 2007). Postural or occupational stressors with repetitive overuse and associated soft tissue adaptations such as hypertrophy in some muscles and atrophy in others, have been implicated in all forms of TOS. Poor posture, especially in patients with large amounts of breast tissue or swelling due to trauma in the area, may predispose to TOS. Compression occurs when the size and the shape of the thoracic outlet is altered. This is commonly caused by poor posture, such as lowering the anterior chest wall with drooping shoulders and holding the head in a forward position (Aligne and Barral, 1992; Novak et al., 1995; Ranney, 1996; Rayan, 1998; Skandalakis and Mirilas, 2001; Barkhordarian, 2007). 7. Diagnosis Diagnosis of TOS is clinical and based on a detailed history, subjective and objective examination of neurovascular and musculoskeletal systems of the neck, shoulder, arm and hands (Roos, 1982; Novak et al., 1995). Frequently a multitude of further investigations are required, many of which in the case of sTOS may indeed prove to be negative (Barkhordarian, 2007). The literature laments that there is no one test or investigation that consistently proves the diagnosis of TOS. Given that TOS really is a ‘‘collection’’ of symptom complexes, often multifaceted, it is unreasonable to

588

L.A. Watson et al. / Manual Therapy 14 (2009) 586–595

Table 1 Classiﬁcations, pathophysiology and investigations. Classiﬁcation

Sub-type

Pathology

Signs & Symptoms

Vascular TOS (vTOS)

1. Arterial TOS (aTOS)

Compression of the subclavian artery that produces any combination of stenosis, poststenotic dilatation, intimal injury, formation of aneurysms and mural thrombosis.

2. Venous TOS (venTOS)

Unilateral arm swelling without thrombosis, when not caused by lymphatic obstruction may be due to subclavian vein compression.

1. True Neurological TOS (tnTOS)

Irritation, compression or traction of the brachial plexus creating compromised nerve function. Compression usually occurs via a bony or soft tissue anomaly present congenitally, created by either repetitive or signiﬁcant trauma and often inﬂuenced by postural, occupational or sporting factors.

2. Symptomatic TOS (sTOS)

Usually no bony or soft tissue anomaly can be demonstrated. Intermittent compression of the neurovascular complex due to repetitive postural, occupational or sporting forces that create temporary compression at varying sites in the cervical or thoracic outlet

Upper limb ischaemia Multiple upper limb arterial embolization Acute hand ischaemia Claudication Vasomotor phenomena Digital gangrene Absent or decreased arterial pulse Swelling, feeling of stiffness/heaviness, fatigability, coldness, pain of muscle cramp in the upper limb or hand Paresthesia (due to ischaemia) Asymmetrical upper extremity oedema (bilateral oedema can occur) Pain, cyanosis, fatigability and a feeling of stiffness or heaviness of the upper extremity Venous engorgement with collateralization of peripheral vessels Axillary or subclavian vein thrombosis Pulmonary embolism Paresthesia Upper plexus syndrome – C5/6/7 pattern: Sensory changes in the ﬁrst three ﬁngers þ/ numbness in cheek, earlobe, back of shoulder, or lateral arm Weakness in deltoid, biceps, triceps, scapula muscles and forearm extensors Pain in anterior neck, chest, supraclavicular region, triceps, deltoids, parascapular muscle, outer arm to the extensor muscles of the forearm þ/ pain in the neck, pectoral region (pseudoangina), face, mandible, temple and ear with occipital headaches þ/ dizziness, vertigo and blurred vision Lower plexus syndrome – C7/8/T1 pattern: Sensory changes in the fourth and ﬁfth ﬁngers, sensory loss above medial elbow Pain and paresthesia over the medial aspect of the arm, forearm, ulnar 1½ digits Hand weakness, loss of dexterity and wasting (lateral thenar muscles, profundi of the little and ring ﬁngers, the ulna intrinsics and the hypothenar muscles and extend into the forearm) Predominantly neurological, intermittent and transient in nature Paresthesia in digits (variable distribution) on awakening Distal symptoms range from pain, aching ‘‘spasm’’ to tingling, numbness and tightness Feelings of weakness and fatigue either in the hand or entire upper limb (especially when it is used overhead) Feeling of tenderness, swelling or loss of motor control Pain in forearm, hands and wrist þ/ Pain in lower neck and shoulder, elbow and upper back, especially over pectoralis minor, lateral humerus, suprascapular and medial scapula regions þ/ Concurrent cervical pain and headache Pain aggravated by repetitive, suspensory, or sustained overhead forward elevation of the shoulder and activities that depress the shoulder girdle Pain at rest and night pain

Neurological TOS (nTOS)

assume that any one test or one investigation can always accurately examine the whole spectrum of pathology. Diagnosis of sTOS is dependent on a systematic, comprehensive upper-body examination and several authors highlight that postural exacerbation of symptoms is an essential component of the diagnosis (Roos and Owens, 1966; Novak et al., 1995). Lindgren (1997) ﬁrst tried to systemize the diagnosis of sTOS by describing a clinical index (Fig. 2). While this index is a good initial guideline, there are other criteria that need to be added to ensure that the sTOS diagnosis is not missed in patients.

monitored as well as any changes in skin temperature, color, texture, blotching, hair growth, swelling, stiffness or loss of motor control. Less commonly seen are symptoms of tachycardia or pseudoangina, occipital headache, vertigo, dizziness, and tinnitus (Malas and Ozcakar, 2006). Behaviour of the symptoms should be noted including, morning and/or night pain and any speciﬁc

Patients should have at least three of the following four symptoms or signs. 1. a history of aggravation of symptoms with the arm in an elevated position

7.1. Subjective examination A detailed global body chart must be completed looking for total distribution of pain, neurological and vascular symptoms not only in the upper limb but in the head, neck, chest and the other side. In particular the type, nature and intensity of symptoms should be

2. a history of paraesthesia originating from the spinal segments C8/T1. 3. supraclavicular tenderness over the brachial plexus 4. a positive hands up abduction/external rotation or stress test. Fig. 2. Clinical index for diagnosis of sTOS (Lindgren, 1997).

L.A. Watson et al. / Manual Therapy 14 (2009) 586–595

589

Table 2 Differential diagnosis. Differential

Signs in common with TOS

Differing signs

Investigations/Tests

Carpal tunnel syndrome

Paresthesia of the hand (can be the entire hand) Proximal pain Night pain Hand pain aggravated by use Pain over lateral wrist and thumb

Loss of wrist range of motion (predominantly extension)

Wrist range of motion Tinel’s sign Phalen’s and reverse Phalen’s Tethered median nerve stress test (Pascarlei) EMG and nerve conduction

Local tenderness and swelling Pain – resisted thumb extension Pain – passive thumb ﬂexion Pain and point tenderness lateral epicondyle Pain – resisted wrist extension, gripping and morning stiffness Pain and point tenderness medial epicondyle Pain – resisted wrist ﬂexion and wringing activities Changes in the color and temperature of the skin over the affected limb, skin sensitivity, sweating, swelling and changes in nail and hair growth. Ptosis of the eye and a constricted pupil

Finkelstein’s test

deQuervain’s tenosynovitis Lateral epicondylitis

Pain in lateral forearm

Medial epicondylitis

Pain in medial forearm

Complex regional pain syndrome (CRPS I or II).

‘‘Burning’’ pain in the upper limb, Motor disability

Horner’s Syndrome

Can co-exist with TOS due to compression affecting nerves as well as stellate ganglion Vasospastic disorder mimic TOS Discolouration ﬁngers and cold sensitivity

Raynaud’s disease

Cervical disease (especially disc)

May present with pain in cervical spine, radiating in to the upper limb and medial scapula

Brachial plexus trauma Systemic disorders: inﬂammatory disease, esophageal or cardiac disease Upper extremity deep venous thrombosis (UEDVT), Paget– Schroetter syndrome

Varying from a neuropraxia to a neurotmesis Upper limb pain þ/ chest pain

Tightness or ‘‘heaviness’’ in affected biceps muscle, shoulder, neck, upper back and axilla Provocation tests are positive

Rotator cuff pathology

Restricted and painful shoulder range of motion Weakness in shoulder muscles

Glenohumeral joint instability

History of repeated overuse in the overhead position or trauma ‘‘Dead arm’’ symptoms or transient neurological symptoms

Discolouration also of toes (occasionally other extremities) in a characteristic pattern in time: white, blue and red Symptoms aggravated by cervical movements rather than arm motion. Ease factor may be elevation of the arm whilst this is an aggravating position in TOS

Hand, upper arm, posterolateral shoulder can be swollen and red with increased tissue temperature over the shoulder Painful limitation of internal and external rotation active motion may be present as well as positive rotator cuff tests Ecchymosis and non-edematous swelling of the shoulder, arm and hand, functional impairment, discolouration and mottled skin and distention of the cutaneous veins of the involved upper extremity Positive rotator cuff testing

Positive glenohumeral instability testing

Ultrasound scan

Ultrasound scan

Investigation autonomic nervous system

Radiological, autonomic and neurological investigation to differentiate. May need to be excluded from vTOS by an angiogram Allen’s test Cervical range of motion Neurological examination (decreased reﬂex in severe disc pathology) Cervical compression and distraction tests Spurling’s maneuver MRI Brachial Plexus Traction test Nerve conduction tests Blood tests (inﬂammation) Stress electrocardiography (cardiac disease)

This condition can cause a potentially dangerous or even fatal complication.

Clinical tests: - Neer and Hawkins impingement tests - Jobe (supraspinatus) test - Speed’s test (biceps) - External rotation test (infraspinatus) - Lift off & press belly test (subscapularis). Clinical tests: - Apprehension test, - Anterior and posterior draw tests in the adducted & abducted shoulder - The sulcus test - Dynamic anterior & posterior stability tests

590

L.A. Watson et al. / Manual Therapy 14 (2009) 586–595

aggravating factors especially; sustained shoulder elevation, suspensory holding activities, lying on the arm, carrying a back pack, carrying articles by the side, prolonged postures (especially sitting), repetitive use of the upper limb and hand dexterity. A detailed history should include the past history of prior traumatic insult to the surrounding neck, shoulder and arm areas that may indicate co-existence of cervical, glenohumeral (especially instability), acromioclavicular or sternoclavicular joint pathology that may confound, confuse or contribute to the clinical presentation (Barkhordarian, 2007). Loss or gain of weight or muscle mass (especially scalenes or pectoralis minor region) should be noted as should a history of use of growth hormone or steroids (Simovitch et al., 2006). 7.2. Physical examination The physical examination of TOS is frequently long and complex as the clinician needs to examine the entire upper limb and cervical spine. Not only is a neurological examination required, but frequently peripheral nerve entrapment tests also need to be performed. 7.2.1. Postural alignment Postural malalignment should be examined. The physique of classic TOS patients is that of a long neck with sloping shoulders (Kai et al., 2001). Many other variations of scapula malpositioning or ‘‘poor posture’’ may also occur in TOS (Pascarelli and Hsu, 2001). If sTOS is suspected then speciﬁc attention should be made to scapula position both at rest, motion and on loading (Refer to Part 2). 7.2.1.1. Palpation. Upper limb pain or symptom reproduction after digital palpation and palpation tenderness (mechanical alodynia) (Schwartzman and Maleki, 1999), especially in the supra and infraclavicular fossae, are considered to be useful in the diagnosis of nTOS. Morley test or the brachial plexus compression test (compression of the brachial plexus in the supraclavicular region) is considered ‘‘positive’’ if there is reproduction of an aching

sensation and typical localized paresthesia and not just mere tenderness of the area (Hasan and Romeo, 2001). This test is reported to be positive in up to 68% of patients with nTOS (Seror, 2005). In some cases fullness or even a palpable hard mass may be present in the supraclavicular region (Cagli et al., 2006). This may be an indicator of a true structural lesion potentially creating either vTOS or tnTOS but the mass itself must also be examined (chest xray and ultrasound) to make sure it is not of a more signiﬁcant nature (Ozguclu and Ozcakar, 2006). Palpation distally may also be required if local joint pathology or peripheral nerve entrapment needs to be excluded. 7.2.1.2. Active/passive motion. Active and passive motions of the cervical spine, cervicothoracic junction, shoulder, elbow, wrist and hands should be performed looking for; joint hyperlaxity, limitation of motion, dyskinesia or abnormal compensatory motions or symptom reproduction (Pascarelli and Hsu, 2001). At a minimum, cervical and shoulder range of motion should be objectively documented using a goniometer or inclinometer at initial assessment. Restriction of glenohumeral joint range of motion has been noted by several authors in sTOS (Sucher, 1990; Aligne and Barral, 1992; Rayan, 1998). This restriction may be due to the increased anterior tilt of the scapula. Any shoulder, scapula, elbow, wrist or hand muscle weakness should also be objectively assessed preferably using an objective assessment device (such as a dynamometer) or at the very least by using the standard 0–5 classiﬁcation (Kendall et al., 1971). 7.2.1.3. Rotator cuff tests and glenohumeral joint instability tests. Rotator cuff tests are examined for pain, weakness, and symptom reproduction to assess rotator cuff pathology (Table 2). If there is a history of repeated overuse in the overhead position (throwing athlete) or trauma, then the glenohumeral joint should be examined for instability (Table 2). 7.2.1.4. Neurological examination. A thorough neurological examination of the upper extremities, including motor, sensory and deep tendon reﬂexes is essential. Sensation can be measured to light

Fig. 3. Adson’s maneuver. A. Patient seated upright. Arms remain supported in patient’s lap and the patient performs cervical spine rotation and extension to the tested side. This is followed by a deep inspirational breath, which is held for up to 30 s, as the examiner palpates for any changes in the radial pulse. B. Modiﬁcation – perform in 15 shoulder abduction and maintain the head in the tested position for 1 min while the subject breathes normally.

L.A. Watson et al. / Manual Therapy 14 (2009) 586–595

touch and pin prick at a minimum but if possible by Semmes– Weinstein monoﬁlament testing (Gillenson et al., 1998). Attention should be given to; skin temperature, presence of tremor, atrophy and swelling (Pascarelli and Hsu, 2001). A decrease in sensation and strength in the absence of any aberrance of the deep tendon reﬂexes may indicate tnTOS. Any change in deep tendon reﬂexes (regardless of any change in sensation and motor control) may indicate a more proximal or central neurological pathology and needs to be referred on for further investigation. Muscle weakness will be manifested in either C5,6 muscle groups (upper plexus) or C8, T1 muscle groups (lower plexus) reﬂected by poor grip strength. This is not common and largely presents in tnTOS (Athanassiadi et al., 2001). 7.2.1.5. Peripheral nerve tests. There are many peripheral examination tests that can be performed. Carpal tunnel syndrome (CTS) is the most commonly cited peripheral nerve entrapment that may be

591

confused with TOS and therefore should be assessed as part of the standard physical examination (Seror, 2004, 2005). Further tests may be required as part of differential diagnosis (Table 2). 7.2.1.6. Cervical spine. In addition to active/passive range of motion, scalene muscle tightness should be examined. Restriction of cervical range and scalene muscle tightness is more likely associated with upper plexus entrapment (Skandalakis and Mirilas, 2001). Cervical nerve root compression caused by cervical disc disease should be excluded (Table 2) using a test such as Spurling’s test (Spurling and Bradford, 1939; Bradford and Spurling, 1942). Thoracic spine kyphosis or scoliosis should be noted and any compensatory lordosis (Sucher, 1990). 7.2.1.7. Provocation testing. The provocation tests most commonly described in the literature to diagnose TOS are presented in Figs. 3–6. These tests are purported to help delineate the possible level of compression of the neurovascular structures in either the

Fig. 4. a – Costoclavicular maneuver. Patient sitting, therapist assists the patient in performing scapula retraction (A), depression (B), elevation (C) and protraction (D), holding each position for up to 30 s. Subject rests his or her forearms on thighs while the examiner simultaneously monitors a change in pulse and symptom onset, note which positions exacerbates/eases symptoms. Recommended modiﬁcation: performed with both arms by the side, holding each position for 1 min. b – Modiﬁed to the military brace position – exaggerates backward and downward bracing of the shoulders. This movement obliterates the pulses most readily.

592

L.A. Watson et al. / Manual Therapy 14 (2009) 586–595

Fig. 5. Wright’s test – Hyperabduction maneuver. The test is performed in two steps. The patient sits in a comfortable position, head forward, while the arm is passively brought into abduction and external rotation to 90 without tilting the head. The elbow is ﬂexed no more than 45 . The arm is then held for 1 min (Rayan and Jensen, 1995). The tester monitors the patient’s symptom onset and the quality of the radial pulse. The test is repeated with extremity in hyperabduction (end range of abduction).

scalene, costoclavicular or axillary (sub-coracoid) intervals. Given the numerous possible causes and symptoms associated with TOS, no single test can unequivocally establish the presence or absence of the condition, particularly where sTOS is concerned (Roos, 1982; Lindgren, 1997). The classic provocation tests have been reported to be unreliable and frequently positive (up to 90%) for pulse obliteration in healthy patients (Hachulla et al., 1990; Urschel et al., 1994; Rayan and Jensen, 1995; Nannapaneni and Marks, 2003). No study to date has analyzed the speciﬁcity, sensitivity and predictability of the provocation tests in relationship to the separate categories of TOS. 7.2.1.7.1. Adson’s maneuver. This test is considered positive if there is an obliteration or diminution of the radial pulse and/or a precipitation of patient’s symptoms (Fig. 3) (Adson and Coffey,

Fig. 6. Roos stress test – EAST test (elevated arm stress test) The patient sits with the head in the neutral position, the arms abducted and externally rotated to 90 and the elbows ﬂexed to 90 . The patient is then requested to ﬂex and extend the ﬁngers for up to 3 min. The examiner watches for any dropping of the extremity during this time, which could indicate fatigue or arterial compromise. The therapist should also observe the color of the distal extremity, comparing left with right and monitor symptoms onset.

1927; Leffert and Perlmutter, 1999). Distribution of pain þ/ paresthesia should be noted and graded as mild, moderate, severe (Rayan and Jensen, 1995). The importance of the obliteration of the pulse has been questioned, especially in nTOS (Gergoudis and Barnes, 1980). This test is thought to stress the scalene triangle but may also stress the contralateral scalene triangle, indirectly bringing on symptoms (Walsh, 1994). 7.2.1.7.2. Costoclavicular maneuver. This maneuver (Fig. 4) is thought to stress the costoclavicular interval where either the subclavian artery, vein or brachial plexus may be entrapped by structures such as subclavius or costocoracoid ligament (Falconer and Weddell, 1943). The test is positive when radial pulse changes and/or patient’s symptoms are provoked. 7.2.1.7.3. Wright’s test – hyperabduction maneuver. The stress hyperabduction test (Fig. 5) is thought to implicate the axillary interval (space posterior to pectoralis minor) in the etiology of TOS (Wright, 1945). The test has two components and a positive result is a decrease in the radial pulse and/or reproduction of the patient’s symptoms. Distribution and severity of symptoms should be recorded (Rayan and Jensen, 1995). The ﬁrst part of the maneuver could implicate the subclavian vessels and plexus as they are stretched around the coracoid process (pectoralis minor impingement). The second part places the extremity in hyperabduction. A positive test is said to implicate the costoclavicular interval (Walsh, 1994). Other authors have described adding on the effect of cervical spine motion (ﬂexion, extension, left and right rotation) (Seror, 2005). 7.2.1.7.4. Roos stress test – EAST test (elevated arm stress test). Originally described by Roos and Owens (1966) and purported by the developers to be the most sensitive and speciﬁc test to detect nTOS (Fig. 6). This test is believed to stress all three intervals (scalene, costoclavicular, axillary) since this position places the arterial, venous and nervous systems in tension. The test is positive when the patient is unable to maintain elevation for the 3-min period or when symptoms are induced (Roos and Owens, 1966; Walsh, 1994). Accuracy of the test is reported to be best at angles of 90 or less (Hachulla et al., 1990). Despite the test being deemed the most sensitive and speciﬁc of the provocation tests (Roos and Owens, 1966), a study by Seror (2005) showed that 14% of patients could not complete the 3-min test with their symptomatic upper limb and 58% of patients with conﬁrmed CTS also had positive stress tests. In the same study only 5% of patients with CTS had a positive Adson’s test (Seror, 2005). 7.2.1.7.5. Other considerations. When using provocation tests for cases of vTOS then obliteration of the radial artery pulse, looking for distal ischaemic signs, oedema, and cyanosis of the upper extremity, measuring blood pressure and auscultation for a bruit in both upper extremities with the arms by the side and in provocation positions (Athanassiadi et al., 2001) is required and likely to be signiﬁcant if found positive (Singh, 2006). In cases of nTOS it would be logical that provocation tests should not only be performed to obliterate the radial artery pulse but also to recreate the patient’s discomfort and symptoms (Konin et al., 1997). Due to the low speciﬁcity of these various tests, some authors argue that if these tests are being utilized for a diagnosis of sTOS then two or three tests should be positive in a given patient (Hachulla et al., 1990). 7.2.1.8. Postural and scapula correction. Exacerbation of symptoms during testing by alterations in posture is considered a strong argument for the diagnosis of sTOS (Hachulla et al., 1990; Rayan and Jensen, 1995). Poor posture or malalignment of the shoulder girdle (such as drooping or rounded shoulders) has been cited by many authors as being a potential cause of sTOS due to the alteration of the anatomical position of the shoulder girdle potentially

L.A. Watson et al. / Manual Therapy 14 (2009) 586–595

593

Fig. 7. Correction of scapula position – Elevation/Upward Rotation. Performed in standing with the examiner behind the patient. Patient’s symptoms are provoked either by performing an active motion (such as abduction) or one of the Provocation tests. The point in range or time to onset of symptoms is objectively noted. Examiner then places his/her hand in the patient’s axilla (A). The examiner’s thumb in the posterior axilla, ﬁngers anterior. The scapula is then passively elevated until it is level with the other side (if the patient has unilateral symptoms) or to a level that approximates the normal resting position of the scapula. A small component (10 –15 ) of scapula upward rotation may also be added. Correction is maintained for 1 min as this will allow tractional forces to be relieved off the brachial plexus. Any alterations in the patient’s resting symptoms are noted. The patient then performs the active movement or provocation test whilst the examiner maintains the passive correction force. Any alteration in patient’s symptoms – distribution, intensity, location or type (worsening or improvement) should be noted. A positive correction response is a signiﬁcant reduction or absence of pain or an increased time duration in the ability to hold provocation positions or perform the stress test before symptom onset.

decreasing the space available in either the scalene triangle, the thoracic outlet, costoclavicular space or sub-coracoid tunnel. Clinically we have found that manual correction of the scapula position is an extremely useful clinical sign to help establish the diagnosis of sTOS and to determine if rehabilitation strategies that focus on strengthening of the scapula stabilizers and altering the scapula position at rest and in motion are likely to be successful (see Figs. 7 and 8). Attempts should be made, where possible, to objectively document the scapula asymmetries observed (Watson et al., 2005). As a general rule, if correction of the scapula improves the patient’s symptoms then the test is positive. If the patient’s symptoms are aggravated or not changed by repositioning then the test is negative. This would indicate that either the wrong correction position for the scapula has been chosen or the patient may not have a form of TOS that will be assisted by rehabilitation strategies for the scapula. This may help establish the diagnosis of either vTOS or nTOS and may indicate greater likelihood that surgical intervention is required. 8. Differential diagnosis The ﬁrst step in the differential diagnosis of TOS is to separate it from other painful conditions of the upper extremity and neck. Other pathologies may mimic TOS or have some clinical overlap (Table 2). It should be taken into account that co-existence of pathology can occur. Upton and McComas (1973) introduced the ‘double crush’ hypothesis, stating that a proximal level of compression could cause more distal sites along the nerve to be more susceptible to compression (Mackinnon, 1994). This hypothesis is extremely pertinent for the patient with nTOS who may be symptomatic from a combination of multiple levels of nerve compression. Each site in and of itself may not be signiﬁcant to produce symptoms but the cumulative effect of minor

compression at several sites along the nerve will result in signiﬁcant symptoms. The most commonly seen clinical picture is the association of carpal and ulnar nerve compression with TOS (Nannapaneni and Marks, 2003) but a similar phenomenon has been reported with cervical spine pathology and TOS (Kai et al., 2001). 9. Treatment Treatment strategies for TOS, particularly with regard to surgical intervention, remain highly controversial. The available literature does not provide strong support either for or against surgery or conservative management (Degeorges et al., 2004). The sub-type of TOS to some extent determines the appropriate treatment pathway. Part 2 of this article will comprehensively outline conservative management. vTOS generally requires surgical treatment and surgery usually involves decompression of the thoracic outlet with removal of the cervical rib (if present) and/or ﬁrst rib excision together with associated muscles and other soft tissue structures as indicated (Gergoudis and Barnes, 1980; Urschel and Razzuk, 1991; Bondarev et al., 1992; Atasoy, 1996; Pupka et al., 2004). The general consensus is that surgery is usually required for aTOS since there is often a structural lesion demonstrated. The decision to operate for venous symptoms is often more difﬁcult as many patients have no bony structures that can be proven responsible for compression. Initially conservative management may be trialed (thromboembolytic therapy and monitoring) but if symptoms persist or progress then decompression may be required (Jamieson and Chinnick, 1996; Azakie et al., 1998; Sultan et al., 2001). In tnTOS there is a high association with structural anomalies (such as a cervical rib) and objective conﬁrmatory tests are positive, potentially justifying surgical intervention. Despite this, many authors still recommend a trial of conservative treatment and only

594

L.A. Watson et al. / Manual Therapy 14 (2009) 586–595

References

Fig. 8. Scapula correction for posterior tilt. Performed when there is an increased anterior tilt or winging of the scapula. Examiner then places one hand anteriorly over the coracoid process and the other hand posteriorly over the blade of the scapula. The scapula is then passively tilted posteriorly and pressure maintained for 1 min. Any alterations in the patients resting symptoms are noted. The patient then reperforms the active movement or provocation test whilst the examiner maintains the passive correction force. If symptoms are worsened with posterior tilt then the test should be repeated applying correction of elevation as well as posterior tilt.

perform surgery if neurological symptoms such as muscle wasting progress (Dale, 1982; Mingoli et al., 1995; Sanders and Hammond, 2002; Degeorges et al., 2004). In sTOS there is often no obvious structural cause and objective conﬁrmatory testing may be lacking. The optimal approach for both surgical and conservative treatment remains controversial and variable (Lindgren, 1997; Sharp et al., 2001). Conservative management is almost universally accepted as the ﬁrst step (Sharp et al., 2001). 10. Conclusion To diagnose TOS is a difﬁcult process that requires time and effort. Given that the etiology of TOS is multifactorial and the signs and symptoms so varied, it would appear logical that physical therapy can successfully be employed in the optimal management of TOS patients (both conservative and surgical). There is a need for the development of a systemized approach to conservative management for TOS (refer to Part 2). If a better objective framework can be established this could facilitate communication between the disciplines to improve patient selection for both surgical and conservative treatment and help develop treatment algorithms that do reliably achieve consistently good or excellent objective treatment outcomes that are sustainable.

Adson AW, Coffey JR. Cervical rib: a method of anterior approach for relief of symptoms by division of the scalenus anticus. Annals of Surgery 1927; 85(6):839–57. Aligne C, Barral X. Rehabilitation of patients with thoracic outlet syndrome. Annals of Vascular Surgery 1992;6(4):381–9. Atasoy E. Thoracic outlet compression syndrome. Orthopedic Clinics of North America 1996;27(2):265–303. Athanassiadi K, Kalavrouziotis G, Karydakis K, Bellenis I. Treatment of thoracic outlet syndrome: long-term results. World Journal of Surgery 2001;25(5):553–7. Azakie A, McElhinney DB, Thompson RW, Raven RB, Messina LM, Stoney RJ. Surgical management of subclavian-vein effort thrombosis as a result of thoracic outlet compression. Journal of Vascular Surgery 1998;28(5):777–86. Barkhordarian S. First rib resection in thoracic outlet syndrome. Journal of Hand Surgery 2007;32(4):565–70. Bondarev VI, Kiandarian AK, Ablitsov NP, Baziak AP. New approaches in diagnosis and therapy of the thoracic outlet syndrome. Klinichna Khirurhia 1992;11:43–5. Bradford RK, Spurling RG. The intervertebral disc with special reference to the rupture of the annulous ﬁbrosus with herniation. Journal of Nervous and Mental Disease 1942;95(6):768. Cagli K, Ozcakar L, Beyazit M, Sirmali M. Thoracic outlet syndrome in an adolescent with bilateral biﬁd ribs. Clinical Anatomy 2006;19(6):558–60. Cherington M. A conservative point of view of the thoracic outlet syndrome. American Journal of Surgery 1989;158(5):394–5. Cooke RA. Thoracic outlet syndrome – aspects of diagnosis in the differential diagnosis of hand-arm vibration syndrome. Occupational Medicine 2003; 53(5):331–6. Dale WA. Thoracic outlet compression syndrome. Critique in 1982. Archives of Surgery 1982;117(11):1437–45. Davidovic LB, Kostic DM, Jakovljevic NS, Kuzmanovic IL, Simic TM. Vascular thoracic outlet syndrome. World Journal of Surgery 2003;27(5):545–50. Degeorges R, Reynaud C, Becquemin JP. Thoracic outlet syndrome surgery: longterm functional results. Annals of Vascular Surgery 2004;18(5):558–65. Demondion X, Bacqueville E, Paul C, Duquesnoy B, Hachulla E, Cotten A. Thoracic outlet: assessment with MR imaging in asymptomatic and symptomatic populations. Radiology 2003;227(2):461–8. Falconer MA, Weddell G. Costoclavicular compression of the subclavian artery and vein. Costoclavicular compression of the subclavian artery and vein. Lancet 1943;2:539–43. Gergoudis R, Barnes RW. Thoracic outlet arterial compression: prevalence in normal persons. Angiology 1980;31(8):538–41. Gillenson SP, Parets N, Bear-Lehman J, Stanton DB. The effect of wrist position on testing light touch sensation using the Semmes–Weinstein pressure aesthesiometer: a preliminary study. Journal of Hand Therapy 1998;11(1):27–31. Gockel M, Vastamaki M, Alaranta H. Long-term results of primary scalenotomy in the treatment of thoracic outlet syndrome. Journal of Hand Surgery [Br] 1994;19(2):229–33. Goff CD, Parent FN, Sato DT, Robinson KD, Gregory RT, Gayle RG, et al. A comparison of surgery for neurogenic thoracic outlet syndrome between laborers and nonlaborers. American Journal of Surgery 1998;176(2):215–8. Gruber E. Cervical rib; report of a case. South African Medical Journal 1952;26(27):551–2. Hachulla E, Camilleri G, Fournier C, Vinckier L. Clinical, ﬂowmetric and radiologic study of the thoracic outlet in 95 healthy subjects: physiologic limitations and practical impact. Revue De Medecine Interne 1990;11(1):19–24. Hasan SS, Romeo AA. Thoracic outlet syndrome secondary to an anomalous subclavius muscle. Orthopedics 2001;24(8):793–4. Jain S, Monbaliu D, Thompson JF. Thoracic outlet syndrome caused by chronic retrosternal dislocation of the clavicle. Successful treatment by transaxillary resection of the ﬁrst rib. Journal of Bone and Joint Surgery 2002;84B(1):116–8. Jamieson WG, Chinnick B. Thoracic outlet syndrome: fact or fancy? A review of 409 consecutive patients who underwent operation. Canadian Journal of Surgery 1996;39(4):321–6. Kai Y, Oyama M, Kurose S, Inadome T, Oketani Y, Masuda Y. Neurogenic thoracic outlet syndrome in whiplash injury. Journal of Spinal Disorders 2001;14(6):487–93. Kendall HO, Kendall FP, Wadsworth GE. Muscles, testing and function. 2nd ed. Baltimore: Williams and Wilkins; 1971. Konin J, Wiksten D, Isear J. Special tests for orthopedic examination. New Jersey: Slack Incorporated; 1997. Lee AD, Agarwal S, Sadhu D. Doppler Adson’s test: predictor of outcome of surgery in non-speciﬁc thoracic outlet syndrome. World Journal of Surgery 2006;30(3):291–2. Leffert RD, Perlmutter GS. Thoracic outlet syndrome. Results of 282 transaxillary ﬁrst rib resections. Clinical Orthopedics and Related Research 1999; (368): 66–79. Lindgren KA. Conservative treatment of thoracic outlet syndrome: a 2-year followup. Archives of Physical Medicine and Rehabilitation 1997;78(4):373–8. Lindgren KA, Oksala I. Long-term outcome of surgery for thoracic outlet syndrome. American Journal of Surgery 1995;169(3):358–60. Machleder HI, Moll F, Verity MA. The anterior scalene muscle in thoracic outlet compression syndrome. Histochemical and morphometric studies. Archives of Surgery 1986;121(10):1141–4.

L.A. Watson et al. / Manual Therapy 14 (2009) 586–595 Mackinnon SE. Thoracic outlet syndrome. Annals of Thoracic Surgery 1994;58(2):287–9. Makhoul RG, Machleder HI. Developmental anomalies at the thoracic outlet: an analysis of 200 consecutive cases. Journal of Vascular Surgery 1992;16(4):534–42. Malas FU, Ozcakar L. Legends of thoracic outlet syndrome. Rheumatology International 2006;27(1):109–10. Mingoli A, Feldhaus RJ, Farina C, Cavallari N, Sapienza P, di Marzo L, et al. Long-term outcome after transaxillary approach for thoracic outlet syndrome. Surgery 1995;118(5):840–4. Nannapaneni R, Marks SM. Neurogenic thoracic outlet syndrome. British Journal of Neurosurgery 2003;17(2):144–8. Novak CB, Collins ED, Mackinnon SE. Outcome following conservative management of thoracic outlet syndrome. Journal of Hand Surgery 1995;20(4):542–8. Ozguclu E, Ozcakar L. Supraclavicular mass disguising as thoracic outlet syndrome. Rheumatology International 2006;26(8):777–8. 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. Pupka A, Rucinski A, Skora J, Janczak D, Pawlowski S, Kaluza G, et al. The treatment of subclavian artery compression with the use of ringed polytetraﬂuoroethylene vascular prostheses. Polimery w Medycynie 2004;34(4):53–61. Ranney D. Thoracic outlet: an anatomical redeﬁnition that makes clinical sense. Clinical Anatomy 1996;9(1):50–2. Rayan GM. Thoracic outlet syndrome. Journal of Shoulder and Elbow Surgery 1998;7(4):440–51. Rayan GM, Jensen C. Thoracic outlet syndrome: provocative examination maneuvers in a typical population. Journal of Shoulder and Elbow Surgery 1995;4(2):113–7. Rockwood Jr CA, Groh GI, Wirth MA, Grassi FA. Resection arthroplasty of the sternoclavicular joint. Journal of Bone and Joint Surgery 1997;79A(3):387–93. Roos DB. The place for scalenectomy and ﬁrst-rib resection in thoracic outlet syndrome. Surgery 1982;92(6):1077–85. Roos DB, Owens JC. Thoracic outlet syndrome. Archives of Surgery 1966;93(1):71–4. Samarasam I, Sadhu D, Agarwal S, Nayak S. Surgical management of thoracic outlet syndrome: a 10-year experience. ANZ Journal of Surgery 2004;74(6):450–4. Sanders RJ, Hammond SL. Management of cervical ribs and anomalous ﬁrst ribs causing neurogenic thoracic outlet syndrome. Journal of Vascular Surgery 2002;36(1):51–6. Schwartzman RJ, Maleki J. Postinjury neuropathic pain syndromes. Medical Clinics of North America 1999;83(3):597–626. Seror P. Medial antebrachial cutaneous nerve conduction study, a new tool to demonstrate mild lower brachial plexus lesions. A report of 16 cases. Clinical Neurophysiology 2004;115(10):2316–22. Seror P. Frequency of neurogenic thoracic outlet syndrome in patients with deﬁnite carpal tunnel syndrome: an electrophysiological evaluation in 100 women. Clinical Neurophysiology 2005;116(2):259–63. Sharp WJ, Nowak LR, Zamani T, Kresowik TF, Hoballah JJ, Ballinger BA, et al. Longterm follow-up and patient satisfaction after surgery for thoracic outlet syndrome. Annals of Vascular Surgery 2001;15(1):32–6.

595

Sheth RN, Belzberg AJ. Diagnosis and treatment of thoracic outlet syndrome. Neurosurgery Clinics of North America 2001;12(2):295–309. Shukla PC, Carlton Jr FB. Diagnosis of thoracic outlet syndrome in the emergency department. Southern Medical Journal 1996;89(2):212–7. Simovitch RW, Bal GK, Basamania CJ. Thoracic outlet syndrome in a competitive baseball player secondary to the anomalous insertion of an atrophic pectoralis minor muscle: a case report. American Journal of Sports Medicine 2006;34(6): 1016–9. Singh D. Arterial complications of thoracic outlet syndrome. Surgical Practice 2006;10:52–6. Skandalakis JE, Mirilas P. Benign anatomical mistakes: the thoracic outlet syndrome. American Surgeon 2001;67(10):1007–10. Sobey AV, Grewal RP, Hutchison KJ, Urschel JD. Investigation of nonspeciﬁc neurogenic thoracic outlet syndrome. Journal of Cardiovascular Surgery 1993;34(4):343–5. Spurling RG, Bradford RK. Neurologic signs in trauma of the brain and spinal cord. Southern Medical Journal 1939;32(1):59–65. Sucher BM. Thoracic outlet syndrome – a myofascial variant: part 2. Treatment. Journal of the American Osteopathic Association 1990;90(9):810–2. 817–823. Sultan S, Evoy D, Eldin AS, Eldeeb M, Elmehairy N. Atraumatic acute upper limb ischemia: a series of 64 patients in a Middle East tertiary vascular center and literature review. Vascular Surgery 2001;35(3):181–97. Upton AR, McComas AJ. The double crush in nerve entrapment syndromes. Lancet 1973;2(7825):359–62. Urschel Jr HC, Razzuk MA. Improved management of the Paget–Schroetter syndrome secondary to thoracic outlet compression. Annals of Thoracic Surgery 1991;52(6):1217–21. Urschel Jr HC, Razzuk MA. Upper plexus thoracic outlet syndrome: optimal therapy. Annals of Thoracic Surgery 1997;63(4):935–9. Urschel JD, Hameed SM, Grewal RP. Neurogenic thoracic outlet syndromes. Postgraduate Medical Journal 1994;70(829):785–9. van Es HW. MRI of the brachial plexus. European Radiology 2001;11(2):325–36. Walsh MT. Therapist management of thoracic outlet syndrome. Journal of Hand Therapy 1994;7(2):131–44. Watson L, Balster SM, Finch C, Dalziel R. Measurement of scapula upward rotation: a reliable clinical procedure. British Journal of Sports Medicine 2005;39(9): 599–603. Wilbourn AJ. The thoracic outlet syndrome is overdiagnosed. Archives of Neurology 1990;47(3):328–30. Wright IS. The neurovascular syndrome produced by hyperabduction of the arms. The immediate change produced in 150 normals and the effects on some persons of prolonged hyperabduction of the arms as in sleeping and in certain occupations. American Heart Journal 1945;29:1–19. Wright D, Jennings PR. Thoracic outlet syndrome. Journal of the American Academy of Physician Assistants 2005;18(5):57–8. Yanaka K, Asakawa H, Matsumaru Y, Kujiraoka Y, Nose T. Diagnosis of vascular compression at the thoracic outlet using magnetic resonance angiography. European Neurology 2004;51(2):122–3.

Manual Therapy 14 (2009) 596–604

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Systematic Review

The effect of age on lumbar range of motion: A systematic review Pattariya Intolo a, Stephan Milosavljevic a, *, David G. Baxter a, Allan B. Carman a, Poonam Pal a, Joanne Munn b a b

Centre for Physiotherapy Research, School of Physiotherapy, University of Otago, Dunedin 9015, New Zealand Faculty of Health Sciences, University of Sydney, Cumberland Campus, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 2 September 2008 Received in revised form 22 July 2009 Accepted 6 August 2009

A systematic review and meta-analysis to determine the effect of age on lumbar range of motion (ROM). Assessment of lumbar ROM is commonly used in spinal clinical examination. Although known to reduce with advancing age, it is unclear how this occurs across different age bands; how this compares between movement planes; and what differences exist between males and females. Ten electronic databases were searched to ﬁnd studies matching predetermined inclusion criteria. Methodological quality was assessed with a quality assessment tool for quantitative studies. Evidence for effect of age on ROM in all planes was investigated with meta-analyses. Sixteen studies met inclusion criteria with results showing age-related reductions in ﬂexion, extension and lateral ﬂexion particularly from 40 to 50 and after 60 years of age. There was very little age effect on lumbar rotation. There is strong evidence for a non-linear age-related reduction in lumbar sagittal and coronal ROM after 40 years of age that also appears to be asymmetric in the coronal plane. These factors should be considered during the evaluation of spinal ROM in patients who present with lumbar disorders. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Age Lumbar Range of motion Mobility

1. Introduction Low back pain (LBP) is a common and costly health problem (Dagenais et al., 2008). Back injury claims cost the Accident Compensation Corporation (ACC) of New Zealand NZ$233 million in compensation in 2002/2003 (Pal et al., 2006) while they are estimated to be £1632 million per year (Maniadakis and Gray, 2000) in the UK. Back problems have a lifetime prevalence ranging from 52% to 91% across all age groups (Jin et al., 2004; Raspe et al., 2004; Walker et al., 2004), and prevalence increases as age progresses. There is debate regarding lumbar range of motion (ROM) as a predictor of successful rehabilitation in LBP with recent evidence suggesting angular velocity and acceleration may be more sensitive indicators (Marras, 2005). Although such dynamic measures are plausible directions for future research, technological and time restraints limit their use in clinical examinations. Clinical assessment of movement impairment in LBP is still most commonly quantiﬁed by ROM, being used to guide treatment and assess the patient’s response. Although lumbar ROM reduces with advancing

age it is still unclear how this reduction occurs across different age categories and clinicians may be uncertain of normative expectations when considering age and sex of a given patient. Thus it may be important to know whether movement reduces with age and whether it does this consistently across different age strata. Assessment of spinal ROM is done with a variety of equipment, procedures and analyses. A number of these studies use variable age categories (Moll and Wright, 1971; Einkauf et al., 1987; Burton and Tillotson, 1988; McGill et al., 1999; Tully et al., 2002), do not address sex differences in ROM (Fitzgerald et al., 1983; Einkauf et al., 1987; Milosavljevic et al., 2005), and do not investigate all planar movements (Einkauf et al., 1987; Russell et al., 1993; McGill et al., 1999; Tully et al., 2002; Milosavljevic et al., 2005). Uncertainty regarding how lumbar ROM reduces across age categories, and how sex may affect such movement is the prime driver for the current review. Our aim is to systematically examine the evidence for effect of age on lumbar ROM in healthy male and female participants. The criteria focused on non-invasive procedures for measuring ROM. 2. Methods

* Corresponding author. Centre for Physiotherapy Research, School of Physiotherapy, University of Otago, PO Box 56, Dunedin 9015, New Zealand. Tel.: þ64 3 479 7193; fax: þ64 3 479 8414. E-mail address: [email protected] (S. Milosavljevic). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.08.006

2.1. Literature search Electronic searches included Ovid; Medline; CINAHL; PEDro; ScienceDirect; Scopus; PubMed; ProQuest; EMBASE; and Web of

P. Intolo et al. / Manual Therapy 14 (2009) 596–604

Science. The search strategy used combinations of the terms ‘age’, ‘healthy’, ‘lumbar’, and ‘ROM’. Manual searches of relevant review bibliographies and reference lists of primary studies were undertaken to look for possible studies not captured by the electronic search.

597

2.4. Participants Studies that included healthy male and/or female participants in any age ranges who free of LBP were considered for review. In addition, participants were required to have had no history of serious spinal or hip joint trauma, including surgery; and/or no local or systematic disease likely to affect the spine.

2.2. Experimental design In order to determine initial relevance for inclusion, citation postings were independently screened by two reviewers (PI and PP). Discussion between the reviewers led to a consensus for articles that met the criteria for abstract review. If abstract review indicated that inclusion criteria were met the full article was extracted. Following independent review of these manuscripts the reviewers discussed whether consensus had been met for inclusion criteria (Fig. 1), study design and participant inclusion. Consensus allowed the given study to be included in this review. A lack of consensus would lead to a third reviewer (JM) independently examining the manuscript and a decision made on majority opinion.

2.3. Study design Cross-sectional investigation of studies measuring lumbar ROM in healthy participants.

2.5. Outcomes Studies reporting lumbar ROM in different ages, including female and male participants. Studies reporting non-invasive procedures measuring lumbar ROM (e.g. 3 Space Isotrak) were included.

2.6. Methodological assessment Study quality was assessed with Quality Assessment Tool for Quantitative Studies, developed by the Effective Public Health Practice Project 2003, Canada (Jackson et al., 2005). Two reviewers (PI and PP) separately evaluated all included studies for quality rating. When reviewers could not agree on a quality rating a third reviewer (JM) was asked to review and make judgment on study quality. All studies were rated according to the components in Table 2 including selection and allocation bias, confounders, blinding, data collection methods,

Fig. 1. Selection of studies for inclusion in systematic review.

598

P. Intolo et al. / Manual Therapy 14 (2009) 596–604

withdrawals and drop-outs, and statistical analysis. These ratings were judged as strong, moderate or weak. 2.7. Data extraction Data for means and standard deviations of lumbar ROM for the different age categories and directions of movement in both females and males (from reported values or extrapolated from ﬁgures) and study sample size were extracted by PI. Where relevant data were unavailable, additional data were requested from authors of these studies. Extracted data were then checked for accuracy by a second reviewer (PP). Disagreement was resolved by discussion and consensus. 2.8. Data synthesis The mean in degrees and estimated standard deviation for ﬂexion, extension, left and right lateral ﬂexion, and left and right rotation ROM (in degrees) as well as sample sizes for each age group were extracted for all pooled data. Results of individual studies were pooled across age group categories for similar measurement protocols using Meta-analysis Interactive eXplanations-version 1.54 (MIXTM) software employing a random effects model and reported as mean difference (X2–X1, 95%CI) (Egger et al., 1997, 1998). Statistical signiﬁcance was accepted at p < 0.05. Data were calculated for ﬂexion, extension, right and left lateral ﬂexion, right and left rotation within the relevant age and sex groups. This software requires the use of the mean and standard deviation of ROM, as well as the sample size for each chosen age strata in each sex group, allowing calculation of 95% conﬁdence intervals (CIs) for age band within each sex category in order to estimate the difference between two population means in different age groups. Statistical tests were considered signiﬁcant at a level of a ¼ 0.05. 3. Results A title review was conducted on 511 extracted articles with 19 relevant manuscripts identiﬁed and full papers obtained. Seven were excluded, where ﬁve used invasive radiological measures (Pearcy and Tibrewel, 1984; Pearcy et al., 1984; Hayes et al., 1989; Harada et al., 2000; Wong et al., 2004), one did not meet participant selection criteria (Twomey, 1979) and one duplicated previously reported results (Troke et al., 2001, 2005). Four further studies were identiﬁed from the reference lists of the remaining articles (Fitzgerald et al., 1983; Einkauf et al., 1987; Hindle et al., 1990; Dvorak et al., 1995). A total of 16 studies were therefore included in this systematic review (Fig. 1, Table 1). Total participants available for meta-analysis included 1323 female and 1001 male participants aged between 8 and 90 years who were described as being in good health. 3.1. Methodological quality Study ratings varied (Table 2) with four not reporting data for height, weight, body mass index (BMI) and recreational activity (Batti’e et al., 1987). All articles described either reliability or validity of measurement devices, three indicated the percentage of participants completing the study. Although the majority provided sample size power calculations (n ¼ 15), substantial differences were noted among all manuscripts for age group descriptions. 3.2. Participants Two studies included participants who reported no back pain within the past year; ﬁve included participants who were free of LBP for at least six months; two included those who were free of

LBP within the last three months; and seven described participants with no current LBP and no previous pathology affecting the spine (Table 1). Burton and Tillotson (1988) divided subjects into 20 year age spans (16–34, 35–54 years old) while Vachalathiti et al. (1995) divided subjects into 15 year age ranges. A further 10 studies divided the age of subjects into decades as follows; 20–29 years and 30–39 years. Two studies categorized age of subjects into either 15–24 years or 16–24 years of age. Tully et al. (2002) divided participants into two age groups 8–10 and 18–23 years old. McGill et al. (1999) studied two distinct mean age groups; 21 years (3.4) and 64 years (3.5) (Table 1). Ultimately 7 of the 16 reviewed articles had sufﬁciently common criteria for inclusion in at least one of the planar meta-analyses. Thus data for 109 females and 154 males were available for meta-analysis. 3.3. Instrumentation Four studies used the 3 Space Isotrak tracking system (McDonnel Douglas Electronics Company, VT, USA); three used the CA-6000 spinal motion analyzer (Orthopedic Systems, Incorporate, Hayward, CA, USA); and three used tape measures or goniometry. The remaining six studies either used a motion analysis system (Motion Analysis CorporationÔ, CA, USA), a ﬂuid-ﬁlled inclinometer (MEDesign Ltd., Southport, UK), Flexicurve, a B-200 Lumbar dynamometer (Isotechnologies, North Carolina, USA), a geometric CAD process (Northern Digital Inc, Ontario, Canada), or videorecording (Table 1). Studies with similar methodology and similar age categories included the angular measures of Fitzgerald et al. (1983), Einkauf et al. (1987), Hindle et al. (1990), Russell et al. (1993), McGregor et al. (1995), Herp et al. (2000), Troke et al. (2005), and Milosavljevic et al. (2005) and were included in the data pool for meta-analysis. Flexion was linearly measured by distraction by Fitzgerald et al. (1983) and Einkauf et al. (1987) and thus the ﬂexion results were not included in the meta-analysis. However these authors used goniometry to measure lumbar extension and lateral ﬂexion thus allowing these angular measures to be included in the data pool. Although Dvorak et al. (1995) used a similar angular measurement methodology for all planes of movement the results were not included in the data pool as examiner overpressure in passive end range stance was applied prior to angular measurement. This procedure was not used by the other included authors. 4. Effect of age on lumbar ROM These results use the 20–29 year age band as the primary benchmark comparison for all other age bands in the expectation that this age band will likely demonstrate the greatest ROM against which reductions in motion within other age bands can be compared. 4.1. Females 4.1.1. Sagittal plane Data pooling (n ¼ 88) were used for four ﬂexion (Russell et al., 1993; McGregor et al., 1995; Herp et al., 2000; Troke et al., 2005) and ﬁve extension (n ¼ 109) studies (Einkauf et al., 1987; Russell et al., 1993; McGregor et al., 1995; Herp et al., 2000; Troke et al., 2005). Although non-signiﬁcant there was a trend for a small increase (2.4 ) in ﬂexion for the 30–39 year group in comparison to the 20–29 year age group. There was a signiﬁcant reduction in ﬂexion (Fig. 2, Table 3) after 40 years of age with a mean difference of 3.5 when comparing the 20–29 to the 40–49 year age groups. A further signiﬁcant reduction of 9.2 was observed between the

Table 1 Details of included studies. Author, date, country

Participants criteria

Burton and Tillotson, 1988, UK

Instruments/company, city

Position/Motion

Spinal level

Results

268F, 242 M no notable low 10 to >54 years old: 10–12, back trouble 16–34, 35–54, >54

Flexicurve

Sitting, prone lying/static

T12–S2

Dvorak et al., 1995, Switzerland

42F, 62M no history of LBP in 20–70 years old: 20–29, 30– the past year 39, 40–49, >50

Einkauf et al., 1987, USAa

109F no history of LBP in past 3 months

CA-6000 Spinal Motion Analyzer/ Standing/dynamic (Passive ROM) Orthopedic System, Incorporate, Hayward, CA, USA Distraction method, Goniometer Standing/static

Spinal mobility decreases with. advancing age. Males showed a reduction in ﬂexion range in middle year whereas females showed reduced ﬂexion younger, maintained that level through middle-age and have a further decline over 65 years of age. A normative database was established showing signiﬁcantly decreased motion as age increased.

Fitzgerald et al., 1983, USAa

Gomez et al., 1991, Canada

Hindle et al., 1990, UK

McGill et al., 1999, Canada

McGregor et al., 1995, UKa

Milosavljevic et al., 2005, New Zealand Moll and Wright, 1971, UK

Russell et al., 1993, Australiaa

Sullivan et al., 1994, USA

Troke et al., 2005, UKa

Tully et al., 2002, Australia

Vachalathiti et al., 1995, Australia

20–84 years old: 20–29, 30– 39, 40–49, 50–59, 60–69, 70– 84 4F, 168M no LBP currently 20–82 years old: 20–29, 30– Distraction method Goniometer no history of LBP in past 3 39, 40–49, 50–59, 60–69, 70– 82 months 83F, 85M no history of LBP in 18–68 years old: <30,30–39, B-200 Lumbar Dynamometer/ Isotechnologies, Carrboro, North past 6 months, no history of 40–49,>50 Carolina, USA back surgery 50F, 50M no history of 20–77 years old: 20–29, 30– 3 Space Isotrak, Polhemus recent LBP 39, 40–49, 50–59, 60þ Navigator Sciences, McDonnel Douglas Electronic Company Colchester, VT, USA 3 Space Isotrak, Polhemus 40F, 40M no history of LBP in 20 to >50 years old: 20–29, Navigation Science, McDonell past 6 months, no history of 30–39, 40–49, >50 Douglas Electronics Company, back surgery Chlchester, VT, USA 21 3.4 years old, 64 35 3 Space Isotrak, Polhemus 7F, 5M no history of low years old Navigation Science, McDonell back injury, or recent Douglas Electronics Company, recurrent pain Chlchester, VT, USA 20–70 years old: 20–29, 30– CA-6000 Spinal Motion Analyzer/ 100F, 103M no LBP Orthopedic System, Incorporate, currently, no history of LBP 39, 40–49, 50–59, 60–70 Hayward, CA, USA in past 6 months 128M no LBP 19–59 years old 20–29, 30–39, Geometric CAD/Northern Digital 40–49, 50–59 Inc, Waterloo, Ontario, Canada Distraction method 118F, 119M no LBP 15 to >75 years old: 15– 24,25–34, 35–44,45–54, 55– 64, 65–74, >75 78 F, 103M no history of LBP 20–69 years old: 20–29, 30– 3 Space Isotrak, Polhemus Navigation Science, McDonnell and any pathology affected 39, 40–49, 50–59, 60–69 Douglas Electronics Company, the spine Colchester, VT, USA 15–65 years old: 16–24, 25– Fluid-ﬁlled inclinometer/ 686F, 440M no previous 34, 35–44, 45–65 MEDesign Ltd., Southport, UK experience of LBP during lifetimes 196F, 209M no LBP currently 16–90 years old: 10–20, 21– CA-6000 Spinal Motion Analyzer/ no history of LBP in past 12 30, 31–40,41–50, 51–60, 61– Orthopedic System, Incorporate, Hayward, CA, USA months or pain in previous 6 70, 71–80, 81–90 months 8–10, 18–23 years old Videotape 22 Adults, 22 Children no movement dysfunction, no history of pathology or pain in the hip 20 to >60 years old: 20–35, Motion Analysis System/Motion 54 Female, 46 Male no Analysis CorporationTM Santa history or back or lower limb 36–59, >60 pain for at least 6 months Rosa, California, USA

TL junction–S

L-S

Spinal mobility decreases with advancing age.

Standing/static

L-S

Spinal mobility decreases with advancing age.

Standing/dynamic

L

ROM reduced with advancing age.

Standing/dynamic

T12–S1

A clear trend of reducing motion with age in both males and females

Standing/dynamic

L1–S

A general trend for decreasing mobility with age.

Standing/dynamic

T12–S

The elderly had a reduced ROM in full ﬂexion and lateral bend but not in axial rotation.

Standing/dynamic

TL junction– PSIS/dynamic

Age appeared to have an inﬂuence on motion, and a gradual reduction was seen with each decade.

Standing/static

T12–PSIS

Lumbar mobility had decreased in advancing age.

Standing/dynamic

L–S

Standing/dynamic

L1–S1

An initiation increase in mean mobility from 15 to 24 decade to the 25–34 decade was followed by a progressive decrease with advancing age. Lumbar ROM was seen to be affected by age.

Prone lying/static

T12–S2

Total sagittal ROM, ﬂexion angle, and extension angle declined as age increased.

Standing/dynamic

T12–S2

Normative ﬂexion and lateral ﬂexion range declined across age spectrum. Extension declined the greatest. In axial rotation no age-related decline was observed.

Standing/static

T10, L1–PSIS

Lumbar movement in adults group were more than that of in children

Sitting/dynamic

T12–L5

Advancing age, signiﬁcant reductions in the ranges of forward and side ﬂexion, but not axial rotation were found

599

(F) Female; (M) Male; (LBP) Low Back Pain; (ROM) Range of motion; (TL) Thoracolumbar; (L) Lumbar; (S) Sacrum; (PSIS) Posterior Superior Iliac Spine. a Results of these studies were eligible and pooled for meta-analysis.

P. Intolo et al. / Manual Therapy 14 (2009) 596–604

Herp et al., 2000, UKa

Age: categories

600

P. Intolo et al. / Manual Therapy 14 (2009) 596–604

Table 2 Quality of studies (n ¼ 16) based on the quality assessment tool for quantitative studies. Study

Burton and Tillotson, 1988 Dvorak et al., 1995 Einkauf et al., 1987 Fitzgerald et al., 1983 Gomez et al., 1991 Herp et al., 2000 Hindle et al., 1990 McGill et al., 1999 McGregor et al., 1995 Milosavljevic et al., 2005 Moll and Wright, 1971 Russell et al., 1993 Sullivan et al., 1994 Troke et al., 2005 Tully et al., 2002 Vachalathiti et al., 1995

Summary of component ratings Selection bias

Allocation bias

Confounders

S W M M W M W W W W W M M M W W

W W W W W W W W W M W W W W W W

M M M W M W W W M M M M M M M M

Blinding

NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA

Data collection methods

Withdrawals and drop-outs

S S S S S S S S S S S S S S S S

S W W W S W W W W S W W W W W W

Statistical analysis Sample size calculation

Signiﬁcant difference

Statistic methods appropriate

Y Y Y Y Y Y Y N Y Y Y Y Y Y Y Y

Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y

Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y

Integrity intervention

NR NR NR NR NR NR NR NR NR NR NR NR NR NR NR NR

Key: Selection bias –Does the sample represent the target population. Allocation bias –The extent that assessments of exposure and outcome are likely to be independent. Confounders –Is a risk factor associated statistically associated with exposure to the putative cause. Blinding – The purpose of blinding the assessor (who might also be the care providers) is to protect against detection bias. Data collection methods – Reliability and validity can be reported in the study or in a separate study. For example, some standard assessment tolls have known reliability and validity. Withdrawals and drop-outs – It indicates the percentage of participants completing the study. Statistical Analysis – Appropriate statistic signiﬁcant needs to be determined between groups. Intervention integrity – What percentage of participants received the allocated intervention or exposure of interest? (S) Strong; (M) Moderate; (W) Weak; (NA) Not Applicable; (Y) Yes; (N) No; (NR) Not relevant.

20–29 and the 60–69 year age groups. No differences were observed between the 20–29 and 50–59 year age groups. In comparison to the 30–39 year age group there are signiﬁcant mean reductions of 5.9 in the 40–49; 6.2 in the 50–59; and 11.7 for the 60–69 year age groups (Fig. 2, Table 3). Compared to the 40–49 year age group ﬂexion did not signiﬁcantly reduce for the 50–59 age group however there was a comparative reduction of 5.9 for the 60–69

year age group. There was also a signiﬁcant 5.6 reduction when comparing ﬂexion between the 50–59 and 60–69 year age groups. Extension ROM signiﬁcantly reduced each decade (Fig. 2, Table 3) for those subjects aged beyond 40 years. There was a mean extension reduction of 7.7 between the 20–29 and 40–49 year age groups; a mean reduction of 10.5 when comparing the 20–29 to 50–59 year age groups, and a 13.9 mean extension reduction when comparing

Fig. 2. The inﬂuence of age category on sagittal ROM in females and males. Symbol (A) represents mean pooled effect of age on ROM in degrees; whiskers represent 95% CIs.

P. Intolo et al. / Manual Therapy 14 (2009) 596–604

601

Table 3 Mean and 95%CI difference (degrees) and p-value between two age categories. Age category

Flexion

Left lateral ﬂexion

Female

Male

Female

Left rotation Male

Female

Male

Group1

Group2

X 95%CI

p-Value

X 95%CI

p-Value

X 95%CI

p-Value

X 95%CI

p-Value

X 95%CI

p-Value

X 95%CI

p-Value

20–29

30–39 40–49 50–59 60–69

2.4 3.2 3.5 3.2 2.8 6.3 9.2 3.6

0.13 0.03a 0.39 0.0001a

2.1 10.7 4.0 8.6 7.1 5.4 16.3 4.4

0.7 0.37 0.01a 0.001a

1.8 1.7 4.2 1.7 5.4 2.0 7.4 2.3

0.04a 0.00001a 0.001a 0.0001a

1.7 1.4 5.6 7.1 6.7 7.6 10.8 7.0

0.04a 0.00001a 0.001a 0.0001a

0.7 1.3 0.3 2.2 1.4 1.4 1.8 3.6

0.26 0.78 0.04a 0.31

2.17 1.6 2.0 1.6 1.7 1.6 2.4 1.9

0.01a 0.015a 0.04a 0.01a

30–39

40–49 50–59 60–69

5.9 3.3 6.2 3.4 11.7 2.8

0.0005a 0.0004a 0.00001a

3.1 2.5 5.1 6.0 10.5 5.0

0.02a 0.01a 0.001a

2.4 1.9 2.4 1.9 5.6 1.9

0.01a 0.01a 0.0001a

4.0 6.8 5.2 7.0 9.3 6.3

0.01a 0.01a 0.001a

0.1 2.7 0.7 1.3 1.8 1.6

0.94 0.26 0.5

0.1 1.1 0.4 1.7 0.2 1.4

0.91 0.63 0.74

40–49

50–59 60–69

0.3 5.1 5.9 3.3

0.9 0.004a

3.0 4.7 1.9 5.5

0.21 0.5

0.3 2.1 3.3 2.1

0.82 0.002a

1.4 1.8 5.0 2.0

0.82 0.002a

1.8 1.7 0.8 1.7

0.03 0.33

0.4 1.6 0.2 1.6

0.65 0.76

50–59

60–69

5.6 3.3

0.00001a

7.0 5.1

0.007a

3.6 2.3

0.0018a

3.6 2.2

0.0018a

1.1 1.4

0.11

0.6 1.7

0.47

Age category

Extension

Right lateral ﬂexion

Female

Male

Female

Right rotation Male

Female

Group2

X 95%CI

p-Value

X 95%CI

p-Value

X 95%CI

p-Value

X 95%CI

p-Value

20–29

30–39 40–49 50–59 60–69

2.9 3.9 7.7 3.5 10.5 2.7 13.9 4.0

0.15 0.00001a 0.00001a 0.00001a

7.8 1.7 5.4 1.9 8.5 2.8 8.5 7.6

0.03 0.001 0.001 0.03

2.8 1.8 4.4 1.8 4.8 1.9 7.9 2.3

0.002a 0.0001a 0.001a 0.0001a

1.2 1.2 5.3 4.2 5.7 6.0 9.4 10.4

0.05a 0.01a 0.06 0.08

0.1 1.5 1.9 4.4 1.9 4.1 2.8 6.9

0.96 0.41 0.36 0.42

2.7 3.6 1.8 2.3 1.5 1.6 0.6 3.0

0.15 0.12 0.06 0.68

30–39

40–49 50–59 60–69

4.9 2.7 7.9 3.8 10.8 5.4

0.00003a 0.00001a 0.00001a

3.7 1.7 6.6 3.3 6.5 7.1

0.001 0.001 0.07

1.8 1.9 2.1 2.1 5.3 2.0

0.07 0.05a 0.0001a

3.6 3.4 4.1 5.1 7.8 9.3

0.04a 0.16 0.1

0.2 1.7 0.1 2.2 1.3 5.2

0.78 0.99 0.63

0.5 2.5 0.8 2.8 1.7 3.1

0.69 0.56 0.28

40–49

50–59 60–69

3.0 3.1 6.0 4.6

0.06 0.01a

2.6 2.1 3.6 6.8

0.01 0.3

0.3 2.0 3.6 1.9

0.77 0.0002a

0.2 1.6 3.8 5.1

0.8 0.15

0.1 1.7 0.4 2.6

0.94 0.76

1.2 2.7 1.0 2.0

0.78 0.34

50–59

60–69

3.7 1.8

0.01a

1.6 6.7

0.6

3.4 3.0

0.03a

3.5 4.3

0.11

0.5 2.2

0.63

1.0 2.3

0.4

a

X 95%CI

Male

Group1

p-Value

X 95%CI

p-Value

Signiﬁcant difference at level 0.05.

the 20–29 and 60–69 year age groups. There were signiﬁcant reductions in mean extension of 4.9 , 7.9 and 10.8 when comparing the 30–39 to the 40–49, 50–59 and 60–69 age groups, respectively. For the 40–49 age group there was a non-signiﬁcant 3.0 mean reduction compared to the 50–59 and a signiﬁcant 6.0 mean reduction compared to the 60–69 year age group. Extension also signiﬁcantly reduced by 3.7 from 50–59 to 60–69 year age groups. 4.1.2. Coronal plane Data could be pooled for four lateral ﬂexion (n ¼ 93) studies (Einkauf et al., 1987; McGregor et al., 1995; Herp et al., 2000; Troke et al., 2005). Left lateral ﬂexion was found to reduce (Fig. 3, Table 3) by a mean difference of 4.2 when comparing the 20–29 to the 40–49 year age groups; and by 4.5 mean when comparing the 20–29 to the 50–59 year age groups. There were also signiﬁcant reductions in mean left lateral ﬂexion of 2.4 , 2.4 and 5.6 when comparing the 30–39 to the 40–49, 50–59 and 60–69 year age groups. Although there were no signiﬁcant comparative differences between the 40–49 and the 50–59 age groups, signiﬁcant reductions of 3.3 and 3.6 were observed when comparing the 40–49 to the 60–69 as well as the 50–59 to the 60–69 year age groups. There were signiﬁcant mean reductions in right lateral ﬂexion (Fig. 3, Table 3) of 2.8 between the 20–29 and 30–39 year age groups; 4.4 when comparing the 20–29 to 40–49 year age groups; 4.8 reductions when comparing the 20–29 and 50–59 year age groups, and 7.9 when comparing the 20–29 and 60–69 year age groups. There were signiﬁcant reductions of 2.1 and 5.3 when comparing the 30–39 to the 50–59 and 60–69 year age groups, respectively. Although there were signiﬁcant reductions of 3.6 and

3.4 (respectively) when comparing the 40–49 to the 60–69 and the 50–59 to the 60–69 years age groups the difference between the 40–49 and the 50–59 year age groups was not signiﬁcant. 4.1.3. Transverse plane Data were pooled (n ¼ 72) for three studies (McGregor et al., 1995; Herp et al., 2000; Troke et al., 2005). Although there was a small yet statistically signiﬁcant difference of 1.4 when comparing rotation between the 20–29 and the 50–59 years age groups (Fig. 4, Table 3), there were no further statistically signiﬁcant differences in left and right rotations between 20 and 70 years age bands. 4.2. Males 4.2.1. Sagittal plane Data were pooled (n ¼ 133) from ﬁve ﬂexion (Russell et al., 1993; McGregor et al., 1995; Herp et al., 2000; Milosavljevic et al., 2005; Troke et al., 2005) and six extension (n ¼ 164) studies (Fitzgerald et al., 1983; Hindle et al., 1990; Russell et al., 1993; Herp et al., 2000; Milosavljevic et al., 2005; Troke et al., 2005). There were statistically signiﬁcant reductions in ﬂexion with mean differences of 7.0 and 16.3 when comparing the 20–29 to the 50–59 and 60–69 year age groups (Fig. 2, Table 3). In comparison to the 30–39 year age group there are signiﬁcant mean reductions of 3.1 in the 40–49; 5.1 in the 50–59; and 10.5 for the 60–69 year age groups. There was also a statistically signiﬁcant difference of 7.0 when comparing ﬂexion between the 50–59 and the 60–69 years age groups. There were signiﬁcant mean extension reductions of 5.4 , 8.5 , and 8.5 between the 20–29 and the 40–49, 50–59 and 60–69 year

602

P. Intolo et al. / Manual Therapy 14 (2009) 596–604

Fig. 3. The inﬂuence of age category on coronal ROM in females and males. Symbol (A) represents mean pooled effect of age on ROM in degrees; whiskers represent 95% CIs.

age groups, respectively (Fig. 2, Table 3). ROM signiﬁcantly reduced by a mean 3.7 and 6.6 between the 30–39 and the 40–49 year, and the 30–39 and 50–59 age groups respectively. There was also a signiﬁcant reduction of 2.6 when comparing the 40–49 and 50–59 year age group. There was no signiﬁcant reduction in extension when comparing the 60–69 age group to the 30–39, 40–49 and 50–59 age groups.

4.2.2. Coronal plane Data pooling from four (n ¼ 90) studies (Fitzgerald et al., 1983; McGregor et al., 1995; Herp et al., 2000; Troke et al., 2005) showed statistically signiﬁcant 1.7 and 10.8 reductions in left lateral ﬂexion when comparing the 20–29 to the 30–39 and 60–69 year age groups, respectively (Fig. 3, Table 3). There were also signiﬁcant reductions of 4.0 , 5.2 and 9.3 comparing the 30–39 to 40–49,

Fig. 4. The inﬂuence of age category on horizontal ROM in females and males. Symbol (A) represents mean pooled effect of age on ROM in degrees; whiskers represent 95% CIs.

P. Intolo et al. / Manual Therapy 14 (2009) 596–604

50–59 and 60–69 year age groups, respectively. Further signiﬁcant reductions of 5.0 and 3.6 (respectively) were observed when comparing the 40–49 to the 60–69 year age groups and the 50–59 to the 60–69 year age groups. Right lateral ﬂexion signiﬁcantly reduced by a mean 1.2 and 5.3 between the 20–29 and 30–39 and the 40–49 year age groups, respectively (Fig. 3, Table 3). While there was a signiﬁcant mean reduction of 3.6 when comparing the 30–39 to 40–49 year age group there were no differences noted for respective between group comparisons for the 40–49, 50–59 and 60–69 year age groups. 4.2.3. Transverse plane Three studies were eligible (n ¼ 59) for data pooling (McGregor et al., 1995; Herp et al., 2000; Troke et al., 2001) showing small but signiﬁcant reductions in mean left rotation of 2.1, 2.0 , 1.7 and 2.4 when comparing the 20–29 to the 30–39, 40–49 50–59 and 60–69 year age groups (Fig. 4, Table 3). However there were no statistically signiﬁcant differences in right rotation from 20 years of age to 70 years old. In summary lumbar ROM appears to reduce most noticeably after 40 years of age for the sagittal and coronal planes of movement in both sexes with only a minimal effect for age on transverse plane movement. 5. Discussion 5.1. Effect of age on lumbar ROM While signiﬁcant age-related reductions in lumbar ﬂexion, extension, and lateral ﬂexion were generally observed for both females and males the reductions in rotation were minor (Table 3). For females extension reduced by a mean 13.9 from 20 to 70 years whereas ﬂexion reduced by a mean 9.0 . Comparatively for males extension reduced by a mean 8.0 in the same age span whereas ﬂexion reduced by a mean 16.3 . Flexion reduction was more pronounced after 40 and 50 years for females and males, respectively, while extension reduced in each decade after 40 and 30 years of age, respectively. All studies showed a greater reduction in ﬂexion for males to some extent. The reasons for such sex differences are unclear and might be due to a variety of factors that include occupational demands, previous pregnancies and sex differences in traditional daily functional activities (Hagstromer et al., 2007). It will take future research to verify and elucidate the reasons for this ﬁnding. Similarly the loss of extension ROM appears to be associated with advancing age and was more obvious in females. As many daily living and work-related activities involve ﬂexed postures it is likely that this helps to maintain ROM compared to extension and perhaps this effect is more pronounced in females for reasons that are yet unknown. For females, lateral ﬂexion reduced in an incremental manner by about 2.0 for each decade beyond 30 years of age. For males the pattern of reduction was less clear, with left and right lateral ﬂexion reducing by less than 2.0 from 20 to 30 years of age and then an asymmetric reduction where left lateral ﬂexion reduced by about 11.0 by 60 years of age, whereas the reduction in right lateral ﬂexion was approximately 5.0 at 40 years with no signiﬁcant reduction beyond this (Table 3 and Fig. 4). The reason for this occurring more predominantly in males is unknown. Investigations for factors such as hand dominance, structural asymmetry and/or lifestyle activities in sport and occupation will be required. The approximate 5.0 and 3.0 reduction in left rotation (less for right rotation) between 20 and 70 years of age for both females and males, respectively, is a 15 and 11% loss of range respectively over

603

a 50 year age span and it is thus apparent that rotation is not strongly inﬂuenced by age – at least when quantifying this movement with a non-invasive measure. 5.2. Limitations A potential weakness of this review and of most systematic reviews is the risk of an incomplete literature search. Although an extensive literature search was conducted it is possible that some relevant published studies were not identiﬁed due to either alternative key words or a poorly worded abstract. Although both sexes and similar age categories were included in the meta-analysis the studies used a variety of deﬁnitions for an absence of a history of LBP prior to testing including no LBP within the past 3 months; 6 months; the past year; and during participants lifetime. Such variation in symptom-related inclusion/exclusion criteria could lead to measurable differences in ROM. There was also considerable variation in the tools used to measure angular ROM in the pooled manuscripts, including use of the 3 Space Isotrak (Russell et al., 1993; Herp et al., 2000; Troke et al., 2005), CA-6000 Spinal Motion Analyzer (McGregor et al., 1995), Geometric CAD analysis (Milosavljevic et al., 2005), and goniometry (Fitzgerald et al., 1983; Einkauf et al., 1987). Good to excellent levels of either inter-rater and/or intra-observer reliability were reported by all studies while an acceptable (low) system error was only described by Russell et al. (1993) and Herp et al. (2000) for the 3 Space Isotrak – and was based on the work of Pearcy and Hindle (1989). ROM was either measured in static end range positions or alternatively during continual dynamic movement. The use of either static or dynamic tests will likely have varying relationships to limitations of functional movement, and further studies are required to truly test the sensitivity effects of static versus dynamic comparative movement measures. It is also possible that, given the small number of pooled manuscripts, mean values from each study may have a strong inﬂuence on the outcome of the meta-analysis, particularly those that are at the lower and higher ends of pooled angular data (e.g. Fitzgerald et al., 1983 and Milosavljevic et al., 2005 respectively). However when the meta-analysis was repeated without the data from these two papers there was very little change (<1.0 ) in ROM measures indicating stability in the mean data and little likelihood of skewing of mean pooled data from any one study. As a variety of different spinal levels were used for measuring lumbar ROM (T12 to L5 or alternatively L1 to S1) variation can also occur in the calculation of the magnitude of ROM. Despite such variations in landmark identiﬁcation the studies were still included if they examined for similar age categories and with similar protocols. Variations on testing position were also evident and included measurements from a sitting posture, from standing, and from prone lying. Some of these variations in test position are likely to inﬂuence spinal and trunk ROM by limiting pelvic movement during the measurement of spinal mobility. 5.3. Clinical relevance This review may help inform clinicians by providing normative data on the expected loss of lumbar ROM in healthy individuals with aging. Evidence from this review demonstrates age-related reductions in lumbar ﬂexion, extension, and lateral ﬂexion albeit in a non-linear and often asymmetric (lateral ﬂexion) manner, most evident after approximately 40 years of age. This review thus provides a guide in predicting normal reduction of ROM as age advances. Therefore ROM measures taken for a given clinical presentation, at a given point in time, can now be compared to

604

P. Intolo et al. / Manual Therapy 14 (2009) 596–604

expected ROM means and CIs for 10 year increments of age strata within each sex category. Although these patterns of ROM reduction by age strata may be comparatively useful for evaluating level of impairment; for evaluating treatment effectiveness; and for determining intervention outcome, further research is required to conﬁrm the clinical utility of such information. Furthermore it is unknown whether such decreases in ROM are inevitable or irreversible. A focus on maintaining or enhancing lumbar mobility, as well as enhanced muscle recruitment and muscle strength, through appropriate exercise may also have clinical utility in patients with lumbar disorders. 5.4. Implications for future research It is still unclear what effect age has on other parameters such as the dynamic patterns of angular displacement, velocity and acceleration, or alternatively the sequences of movement initiation or completion. Evaluating pattern of spinal movement as a more sensitive indicator of spinal motion variation in the different age categories has been initiated by our research team. In addition, investigation of coupled movement will be an important aspect for future investigation – perhaps indicating multi-structural and multi-planar movement components to function. Such research will investigate variations in right and left movement strategies for both females and males and will further explore the effect of age on these movement dynamics. Exploration of these effects is warranted in order to better understand how dynamic functional movement occurs and interacts, and in particular to clarify the place of movement examination in clinical disorders of the low back. The quality assessment used in this review should also be considered for use in any future research in order to increase the evidence base for variations in spinal movement in both asymptomatic healthy participants and LBP patients. 6. Conclusion This review shows spinal ROM in the sagittal and coronal planes reduces with advancing age. It is clear that the reduction occurs most markedly in both sexes after 40 years of age and subsequent reductions in range are consistent in each following decade for extension but somewhat inconsistent for further reductions in ﬂexion. For lateral ﬂexion, the initial reduction occurs after 30 years of age and continues to reduce in each decade for females but this effect is not observed in males. We recommend consideration of such age effects particularly in the assessment and re-assessment of lumbar ROM in low back patients by manual therapists. Acknowledgments This systematic review was undertaken by the ﬁrst author in partial fulﬁllment of the Doctor of Philosophy, School of Physiotherapy, University of Otago. Support was provided by the University of Otago (PhD Scholarship and Establishment Grant for GDB). References Batti’e MC, Bigos SJ, Sheehy A, et al. Spinal ﬂexibility and individual factors that inﬂuence it. Physical Therapy 1987;67(5):653–8. Burton AK, Tillotson KM. Reference values for ‘normal’ regional lumbar sagittal mobility. Clinical Biomechanics 1988;3(2):106–13.

Dagenais S, Caro J, Haldeman S. A systematic review of low back pain cost of illness studies in the United States and internationally. Spine Journal 2008;8(1):8–20. Dvorak J, Vajda EG, Grob D, et al. Normal motion of the lumbar spine as related to age and sex. European Spine Journal 1995;4(1):18–23. Egger M, Smith GD, Phillips AN. Meta-analysis: principles and procedures. BMJ 1997;315(7121):1533–7. Egger M, Schneider M, Smith GD. Spurious precision? Meta-analysis of observational studies. BMJ 1998;316(7125):140–4. Einkauf DK, Gohdes ML, Jensen GM, et al. Changes in spinal ROM with increasing age in women. Physical Therapy 1987;67(3):371–5. Fitzgerald GK, Wynveen KJ, Rheault W, et al. Objective assessment with establishment of normal values for lumbar spinal range of motion. Physical Therapy 1983;63(11):1776–81. Gomez T, Beach G, Cooke C, et al. Normative database for trunk range of motion, strength, velocity, and endurance with the isostation B-200 lumbar dynamometer. Spine 1991;16(1):15–21. Hagstromer M, Oja P, Sjostrom M. Physical activity and inactivity in an adult population assessed by accelerometry. Medicine & Science in Sports & Exercise 2007;39(9):1502–8. Harada M, Abumi K, Ito M, et al. Cineradiographic motion analysis of normal lumbar spine during forward and backward ﬂexion. Spine 2000;25(15):1932–7. Hayes MA, Howard TC, Shepherd J. Roentgenographic evaluation of lumbar spine ﬂexion–extension in a symptomatic individuals. Spine 1989;14(3):327–31. Herp GV, Rowe P, Salter P, et al. Three-dimensional lumbar spinal kinematics: a study of range of movement in 100 healthy subjects aged 20 to 60þ years. Rheumatology 2000;39(12):1337–40. Hindle RJ, Pearcy MJ, Cross AT, et al. Three-dimensional kinematics of the human back. Clinical Biomechanics 1990;5(4):218–28. Jackson N, Water E, Guidelines for Systematic Reviews in Health Promotion and Public Health Taskforce. Criteria for the systematic review of health promotion and public health interventions. Health Promotion International 2005;20(4):367–74. Jin K, Sorok GX, Courtney TK. Prevalence of low back pain in three occupational groups in Shanghai, People’s Republic of China. Journal of Safety Research 2004;35(1):23–8. Maniadakis N, Gray A. The economic burden of back pain in the UK. Pain 2000;84(1):95–103. Marras WS. The future of research in understanding and controlling work – related low back disorders. Ergonomics 2005;48(5):464–77. McGill SM, Yingling VR, Peach JP. Three-dimensional kinematics and trunk muscle myoelectric activity in the elderly spine – a database compared to young people. Clinical Biomechanics 1999;14(6):389–95. McGregor AH, McCarthy ID, Hughes SP. Motion characteristics of the lumbar spine in the normal population. Spine 1995;20(22):2421–8. Milosavljevic S, Milburn PD, Knox BW. The inﬂuence of occupation on lumbar sagittal motion and posture. Ergonomics 2005;48(6):657–67. Moll JM, Wright V. Normal range of spinal mobility. An objective clinical study. Annals of the Rheumatic Diseases 1971;30(4):381–6. Pal P, Milosavljevic S, Sole G, et al. Hip and lumbar continuous motion characteristics during ﬂexion and return in young healthy males. European Spine Journal 2006;16(6):741–7. Pearcy MJ, Hindle RJ. New method for the non-invasive three-dimensional measurement of human back movement. Clinical Biomechanics 1989;4(1):73–9. Pearcy MJ, Tibrewal SB. Axial rotation and lateral bending in the normal lumbar spine measured by three-dimension radiography. Spine 1984;9(6):582–7. Pearcy M, Protek I, Shepherd J. Three-dimensional X-ray analysis of normal movement in the lumbar spine. Spine 1984;9(3):294–7. Raspe H, Matthis C, Croft P, et al. Variation in back pain between countries the example of Britain and Germany. Spine 2004;29(9):1017–21. Russell P, Pearcy MJ, Unsworth A. Measurement of the range and coupled movements observed in the lumbar spine. British Journal of Rheumatology 1993;32(6):490–7. Sullivan MS, Dickinson CE, Troup JD. The inﬂuence of age and sex on lumbar spine sagittal plane range of motion. A study of 1126 healthy subjects. Spine 1994;19(6):682–6. Troke M, Moore AP, Maillardet FJ, et al. A new, comprehensive normative database of lumbar spine ranges of motion. Clinical Rehabilitation 2001;15(4):371–9. Troke M, Moore AP, Maillardet FJ, et al. A normative database of lumbar spine range of motion. Manual Therapy 2005;10(3):198–206. Tully EA, Wagh P, Galea MP. Lumbofemoral rhythm during hip ﬂexion in young adults and children. Spine 2002;27(20):E432–40. Twomey L. The effects of age on the range of motions of the lumbar region. Australian Journal of Physiotherapy 1979;25(6):257–63. Vachalathiti R, Crosbie J, Smith R. Effects of age, sex and speed on three dimensional lumbar spine kinematics. Australian Journal of Physiotherapy 1995;41(4): 245–53. Walker BF, Muller R, Grant WD. Low back pain in Australian adults: prevalence and associated disability. Journal of Manipulative and Physiological Therapeutics 2004;27(4):238–44. Wong KW, Leong JCY, Chan M, et al. The ﬂexion–extension proﬁle of lumbar spine in 100 healthy volunteers. Spine 2004;29(15):1636–41.

Manual Therapy 14 (2009) 605–610

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

The association between degenerative hip joint pathology and size of the gluteus medius, gluteus minimus and piriformis muscles Alison Grimaldi a, *, Carolyn Richardson a, Warren Stanton b, Gail Durbridge c, William Donnelly d, Julie Hides a, b a

Division of Physiotherapy, School of Health and Rehabilitation Sciences, The University of Queensland, Brisbane 4072, Australia The UQ/Mater Back Stability Clinic, Mater Health Services, Raymond Terrace, South Brisbane, Queensland 4101, Australia Centre for Magnetic Resonance Imaging, Brisbane, Australia d Brisbane Orthopaedic Specialist Services, Brisbane, Australia b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 13 November 2008 Received in revised form 19 June 2009 Accepted 8 July 2009

This study aimed to investigate changes in the deep abductor muscles, gluteus medius (GMED), piriformis (PIRI), and gluteus minimus (GMIN), occurring in association with differing stages of unilateral degenerative hip joint pathology (mild: n ¼ 6, and advanced: n ¼ 6). Muscle volume assessed via magnetic resonance imaging was compared for each muscle between sides, and between groups (mild, advanced, control (n ¼ 12)). GMED and PIRI muscle volume was smaller around the affected hip in subjects with advanced pathology (p < 0.01, p < 0.05) while no signiﬁcant asymmetry was present in the mild and control groups. GMIN showed a trend towards asymmetry in the advanced group (p ¼ 0.1) and the control group (p ¼ 0.076) which appears to have been associated with leg dominance. Between group differences revealed a signiﬁcant difference for the GMED muscle reﬂecting larger muscle volumes on the affected side in subjects with mild pathology, compared to matched control hips. This information suggests that while GMED appears to atrophy in subjects with advanced hip joint pathology, it may be predisposed to hypertrophy in early stages of pathology. Assessment and exercise prescription methods should consider that the response of muscles of the abductor synergy to joint pathology is not homogenous between muscles or across stages of pathology. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: OA Gluteus medius Gluteus minimus Piriformis

1. Introduction Osteoarthritis (OA) of the hip poses a considerable problem for modern society. As the incidence of OA of the hip increases with the aging population it has been declared by March and Bagga (2004) that ‘primary and secondary programs aimed at improving rehabilitation and physical activity are urgently required’ in the management of OA. Therapeutic exercise programmes designed to improve muscle function around the affected hip will only be maximally effective when we have further information available on both normal muscle function, and changes occurring in association with joint disease. Hip abductor muscle function has been a primary focus of research due to the importance of these muscles in performing single leg function, the basis of human locomotion. Patients with

* Correspondence to: Alison Grimaldi, PhysioTec Physiotherapy, 23 Weller Rd, Tarragindi, Brisbane, Queensland 4121, Australia. Tel./fax: þ61 7 3342 4284. E-mail address: [email protected] (A. Grimaldi). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.07.004

OA of the hip have demonstrated a change in pelvic-femur alignment during gait depending on stage of pathology. Those with mild OA demonstrate increased hip adduction during stance (Watelain et al., 2001), while those with more advanced changes reduce adduction by increasing frontal plane trunk movement (Krebs et al., 1998). The speciﬁc changes in abductor muscle function occurring in association with OA are however unclear at this point. While some authors have demonstrated reduced electromyographic (EMG) activity in the gluteus medius (GMED) muscle in subjects with OA of the hip (Long et al., 1993), others have shown increased EMG activity during dynamic function (Angielczyk and Bronarski, 1982; Sims et al., 2002). EMG testing of the tensor fascia lata (TFL) muscle has shown similar inconsistency (Long et al., 1993; Sims et al., 2002). No EMG investigations of the other hip abductor muscles, upper gluteus maximus (UGM), gluteus minimus (GMIN) or piriformis (PIRI) muscles, in patients with OA of the hip, have been reported in the literature. Studies that have involved strength testing as a measure of hip abductor muscle function in subjects with OA of the hip, have used dynamometry to measure open chain isometric or isokinetic abduction strength, providing a global

606

A. Grimaldi et al. / Manual Therapy 14 (2009) 605–610

assessment of the abductor synergy (UGM, TFL, GMED, GMIN, PIRI). These studies have, like EMG studies, displayed considerable variability (Murray and Sepic, 1968; Teshima, 1994; Jandric, 1997; Arokoski et al., 2002; Sims et al., 2002). The body of literature to date thus provides an incomplete and unclear picture of hip abductor muscle dysfunction. More speciﬁc information on patterns of change within the abductor synergy is required. The use of magnetic resonance imaging (MRI) provides an opportunity to assess each individual member of the abductor synergy simultaneously. One previous MRI study assessed cross sectional area (CSA) of the abductor muscles in subjects with OA of the hip, however most of the muscles were grouped together providing a global measure of abductor muscle size (Arokoski et al., 2002). In addition, single CSA measurements are unlikely to be as reﬂective of a muscle’s morphology as a measurement of muscle volume. The research undertaken by the current authors used MRI to assess muscle volume of each individual member of the abductor synergy in subjects with OA of the hip. This has been presented as two papers with muscles divided on an anatomical basis. An initial study (Grimaldi et al., in press) investigated changes present in the superﬁcial lateral musculature (UGM and TFL) that insert into the iliotibial band (Williams et al., 1989). The TFL was unaffected by the presence of joint pathology, while the UGM demonstrated asymmetry in subjects with advanced unilateral OA that appeared to be more closely related to hypertrophy of the unaffected side, than atrophy around the affected hip (Grimaldi et al., in press). The main aim of the current study was to investigate in these same subjects, size of the muscles of the deep lateral stability mechanism of the hip, the GMED, GMIN, and PIRI muscles, that assert their effect via direct insertion into the greater trochanter. Subjects with either mild or advanced unilateral degenerative pathology of the hip were chosen for maximum clarity of effect. The speciﬁc aims were to examine i) if there was signiﬁcant asymmetry in the deep abductor muscles across 3 groups (mild degenerative change, advanced degenerative change, control) and ii) if there were signiﬁcant differences in actual muscle size among the pathology and control groups. This study also examined the association of both stage of pathology, and muscle size, with the factors of age, height, weight, pain, function and activity levels. Leg dominance was also tested as all of these factors were considered to have the potential to impact upon muscle size and symmetry. The hypotheses of the study were that ia) there would be signiﬁcant asymmetry in size of the GMED, GMIN, and PIRI in subjects with hip joint pathology, but not in controls, ib) asymmetry would be greater in subjects with advanced pathology, and ii) the GMED, GMIN and PIRI muscles would be smaller around the affected hip in those with advanced pathology compared to the matched hip of control subjects. 2. Methods 2.1. Subjects Twelve subjects with degenerative hip joint pathology, and twelve age and sex matched control subjects were recruited for this study via medical practitioners and community advertisement. Control group subjects were required to be within 5 years of the age of their matched subject with joint pathology. Each group had equal numbers of males and females and all participants gave their informed consent to participate in this study after receiving detailed information on the study. Ethical approval was provided by the institutional review boards. Inclusion criteria required subjects with pathology to have a medical diagnosis of unilateral degenerative joint pathology, and radiographic evidence of their pathology. Subjects with OA were

Table 1 Subject characteristics for each group. Group

Mild Adv Con

No

6 6 12

Sex

Age

Weight(kg)

Height(cm)

M:F

Mean(SD)

Mean(SD)

Mean(SD)

BMI Mean(SD)

3:3 3:3 6:6

46.5 (9.5) 57.7 (6.7) 51.8 (9.7)

80.4 (15.1) 78.3 (8.5) 73.5 (13.3)

171.3 (9.7) 172.0 (7.4) 168.2 (10.2)

27.3 (3.5) 26.6 (4.4) 25.9 (3.3)

Number (No); Body Mass Index (BMI); Male:Female (M:F). Standard deviation (SD); Advanced Pathology (Adv); Control (Con).

recruited for either a ‘Mild’ or an ‘Advanced’ group. Those determined by an experienced radiologist to have early joint space narrowing and osteophytes (Kellgren/Lawrence (K/L) global scoring system grades 1–2 (Kellgren and Lawrence, 1957; Hirsch et al., 1998) were included in the mild group. Subjects with moderate to severe joint space narrowing and osteophytes (K/L grades 3–4) were recruited for the advanced group. Pathology was right sided for 5 subjects and left for 7 subjects. An analysis of variance (ANOVA) reported previously for these subjects determined that there was comparability between the mild, advanced and control group subjects for the factors of age, height and weight (Grimaldi et al., in press). Details of subject characteristics are listed in Table 1. Exclusion criteria included any factors that may represent confounding variables for muscle size or asymmetry such as systemic diseases of the muscular of nervous systems, congenital or childhood hip disease, any history of hip trauma, surgery, inﬂammatory joint disease, tumours, or lower limb or lower back injury within 2 years of testing. In addition subjects were excluded if they reported any signiﬁcant lifetime history of lower back pain that resulted in a period of immobility, investigation, or treatment. Subjects were also excluded in both groups if they reported participation in unilateral sports, use of a walking aid, or any problems that would preclude them from MRI scanning procedures (e.g. pacemaker, metal implants, pregnancy, claustrophobia). Control group subjects must have had no history of pain in the hip region.

2.2. Procedure Self-Report Questionnaires. Subjects activity levels were rated using a 12 month Leisure Time Physical Activity questionnaire providing an activity metabolic index (AMI) calculated with the formula: AMI ¼ Intensity code (mean metabolic units) average number of times the activity is performed per month the number of months per year (frequency) the time the activity was performed per occasion (duration). AMI for each activity is added so total AMI is compared across individuals (Taylor et al., 1978; Arokoski et al., 2002). A previously reported ANOVA for these subjects found no signiﬁcant differences between groups for metabolic activity (Grimaldi et al., in press). Pain and function were also assessed for pathology groups using the Modiﬁed Harris Hip Score (MHHS) (Byrd and Jones, 2000). This analysis has been reported in a prior paper revealing a signiﬁcantly lower score for the advanced group (p < 0.05), reﬂecting higher pain and lower function (Grimaldi et al., in press). The relationship between pain, function, and radiographic change has been discussed in detail in the same paper. Testing of Leg Dominance. Leg dominance during kicking function was tested with the weight-bearing leg recorded as ‘‘stance dominant’’ and the kicking leg as the ‘‘skill dominant’’ leg (Herneth et al., 2004). All subjects in this study were left stance dominant. MRI Assessment. After medical screening for suitability for MRI procedures subjects were positioned in supine with their legs extended to a neutral position. A 1.5 Tesla Siemens Sonata MR system was employed to collect a T2 True FISP sequence using 2

A. Grimaldi et al. / Manual Therapy 14 (2009) 605–610 Table 2 Intra-rater reliability across repeated measurement for the same image sequence for gluteus medius (GMED), gluteus minimus (GMIN) and piriformis (PIRI) muscles. Muscle

ICC2,1 (95% CI)

SEM cm2

SDD cm2

GMED GMIN PIRI

0.998 (0.997–0.999) 0.997 (0.994–0.998) 0.985 (0.955–0.995)

0.506 0.379 0.675

7.86 3.72 6.74

Intraclass correlation coefﬁcient (ICC), (95% conﬁdence interval at p < 0.05); Standard error of measurement (SEM); Standard deviation of the difference (SDD).

series of 28 6 mm contiguous slices throughout the pelvis (TR: 3.78 ms/TE:1.89 ms/FOV:390 mm). Measurement Procedure. CSA (cm2) of GMED, GMIN and PIRI muscles was measured by tracing each muscle on each slice using an MRI measurement software package (Osiris Version 4.19, University Hospitals of Geneva, Switzerland). Muscle volume (cm3) was determined as the sum of the muscles CSA on each slice in which the muscle appeared, multiplied by the slice width (Fukunaga et al., 1992; Alkner and Tesch, 2004). Reliability of the assessor’s measurement technique was tested by retracing all slices of one subject with an interim period of 6 weeks. Intra-tester reliability was tested for each separate measurement on each slice using a two sided bootstrapped interval of intraclass correlation coefﬁcient (ICC2,1). Intra-rater reliability was found to be very good, with correlation coefﬁcients ranging from 0.985 to 0.989. Standard error of measurement (SEM) was calculated using the formula SEM ¼ pooled SD (1 ICC)½ (Wallwork et al., 2007). Standard deviation of the difference (SDD) was also calculated as the standard deviation of the differences between measurement 1 and 2. ICC, SDD and SEM values are presented in Table 2. 2.3. Statistical analysis Analysis was performed using the Statistical Package for the Social Sciences (version 14; www.spss.com). The ﬁrst analysis addressed the issue of symmetry in muscle size between sides across the 3 groups. A comparison of muscle volumes among groups and between sides was performed using a mixed linear model describing muscle volume with group as a between-subjects factor (control, mild, advanced), and side (affected/unaffected for the pathology groups; left/right for the control group) as a withinsubjects factor (Dependant variable: muscle volume; Independent variables: side, group). Each muscle (GMED, GMIN, PIRI) was analysed separately. Contrasts of means were performed to compare sides within groups. Further analysis was conducted to assess whether control group subjects had larger hip abductor muscles than subjects with hip pathology. Separate ANOVAs were conducted for each side to compare muscle volumes across groups. Side comparisons were determined via the following method: if the pathological side was left, the left side muscle volume of the matched control subject was used for comparison, and the right compared with the unaffected side value of the pathology group counterpart. The dependant variable was muscle volume and the independent variable was group. Each muscle (GMED, GMIN, PIRI) was analysed separately, and contrasts of means were performed to compare size across groups. For ease of presentation of results, percent differences were calculated using the formula: % Difference ¼ [(larger value smaller value)/larger value] 100 (Hides et al., 1996). Analyses were also conducted to assess participant characteristics in relation to the extent of association with muscle size. The association between the patient characteristics of age, height,

607

weight, pain, function, and AMI and GMED, GMIN, or PIRI muscle size was assessed using analysis of covariance.

3. Results 3.1. Side to side differences in muscle volumes within groups There was no signiﬁcant asymmetry in the control group for GMED, GMIN or PIRI muscle volume, although there was a trend for the GMIN muscle to be larger on the left side (p ¼ 0.076, 9 of 12 control subjects larger on the left). No signiﬁcant differences were observed for any of the muscles studied for the mild group. GMED and PIRI were both signiﬁcantly smaller on the affected side for subjects with advanced pathology (t ¼ 2.951, p ¼ 0.008; t ¼ 2.195, p ¼ 0.03 respectively). Although comparisons of GMIN muscle volume did not reach statistical signiﬁcance, there was a trend for asymmetry in the advanced group (p ¼ 0.1) with smaller GMIN size around the affected hip (mean 8.3% smaller). Five of the 6 subjects in this group were on average 21.5% smaller on the affected side, with one subject demonstrating a 48% larger GMIN muscle volume on the affected side. Means, standard deviations, and percentage difference in muscle volumes are reported for each group in Table 3. Examples of side to side differences are illustrated for each group in Fig. 1.

3.2. Differences in muscle volumes between groups Comparisons between groups revealed that the GMED muscle was signiﬁcantly larger (mean 15%) around the affected hip in the mild group, compared with the same hip of the matched control subjects (p ¼ 0.026). No differences were evident between groups for the GMIN or PIRI muscles.

3.3. Effect of subject characteristics on muscle size There was no signiﬁcant relationship between the patient characteristics of age, height, weight, and metabolic activity, or pain and function, and GMED, GMIN or PIRI muscle volume (p > 0.05).

4. Discussion This study investigated the inﬂuence of degenerative hip joint pathology on size of the deep abductor muscles, GMED, GMIN and PIRI.

Table 3 Muscle volumes (cm3) for gluteus medius, gluteus minimus, and piriformis muscles, and percentage difference between sides. Group

Side

Mild (n ¼ 6)

Affected Unaffected % Difference Affected Unaffected % Difference Left Right % Difference

Advanced (n ¼ 6) Control (n ¼ 12)

GMED

GMIN

PIRI

Mean (SD)

Mean (SD)

Mean (SD)

369 (63) 367 (62) 0.4% 317 (94) 361 (71) 12%** 317 (75) 305 (88) 3.7%

87 (23) 95 (32) 7.9% 84 (34) 91 (33) 8.2% 86 (21) 79 (21) 8.3%

28 (10) 29 (14) 2.6% 28 (8) 33 (8) 14.4%* 28 (8) 28 (8) 0.4%

Gluteus medius muscle (GMED); Gluteus minimus muscle (GMIN); Piriformis muscle (PIRI); Standard deviation (SD); * p < 0.05, **p < 0.01.

608

A. Grimaldi et al. / Manual Therapy 14 (2009) 605–610

these muscles on the unaffected side as this side becomes favoured for weight-bearing function. Despite a lack of statistically signiﬁcant asymmetry in the deepest abductor muscle, GMIN, there was a trend towards asymmetry in the advanced group (mean 8.3% smaller on affected side, p ¼ 0.1). The importance of this trend is further highlighted when the removal of a single subject results in an asymmetry reﬂecting an average 21.5% smaller GMIN muscle volume on the affected side. The reason for the lack of atrophy around the affected hip in the remaining subject is unclear. This subject did remain very active with an AMI just below the average for normal control subjects, which may provide some explanation for this variation. Without this subject there is a clear pattern of asymmetry, smaller on the affected side, in the majority of the advanced pathology group. Atrophy in this deepest hip abductor muscle would be consistent with atrophy evident in other local muscles involved in joint protection, such as the multiﬁdus muscle in the lumbar spine (Hides et al., 1994), although some concurrent hypertrophy on the unaffected side cannot be excluded. The other consideration in the interpretation of results for the GMIN muscle is the trend towards GMIN asymmetry, larger on the left side, in control subjects (p ¼ 0.076). This asymmetry may be related to leg dominance as all subjects were left stance dominant. The GMIN muscle may be particularly important in weight-bearing function to assist in joint protection and stabilisation of the femoral head in the acetabulum (Beck et al., 2000; Walters et al., 2001). The relevance of this trend towards asymmetry in control group subjects is that for subjects with left sided hip joint pathology, the loss of muscle size may be underestimated. The only other study to date to investigate symmetry of hip abductor muscle size in subjects with OA of the hip showed a 6% smaller CSA of the ‘gluteal muscles’ around the most affected hip in those with unilateral or bilateral OA (Arokoski et al., 2002). Although the general picture is consistent with our ﬁndings the combined measure of all hip abductor muscles is difﬁcult to directly compare to that of the present study. 4.2. Differences in muscle volumes between groups

Fig. 1. The gluteus medius muscle ( in web version), gluteus minimus muscle ( in web version), and piriformis muscle ( in web version) in axial images above the hip joint in control group subject (A), and subjects with mild left osteoarthritis (B), and advanced left osteoarthritis (C). White dot indicates left ilium.

4.1. Side to side differences in muscle volumes within groups Although subjects with mild degenerative hip joint pathology were not signiﬁcantly asymmetrical, those with advanced pathology demonstrated signiﬁcant asymmetry for the GMED and PIRI muscles with smaller muscle volumes around the affected hip (mean 12%, p < 0.01 and mean 14.4%, p < 0.05 respectively). This is consistent with the changes in gait pattern at this stage of pathology (Krebs et al., 1998). Peak acetabular pressures have been shown to coincide with peak GMED activity rather than peak ground reaction forces (Krebs et al., 1998). The associated increases in lateral trunk ﬂexion over the weight-bearing leg during stance phase of gait was proposed to be a strategy to reduce abductor muscle activity, thereby reducing compressive forces across painful degenerated joint surfaces. This functional disuse would be in line with the muscle atrophy illustrated in the current study. Part of the asymmetry revealed may also be accounted for by hypertrophy of

Differences in muscle volumes between groups were not signiﬁcant for PIRI and GMIN muscles, consistent with the lack of between group difference (OA and control) reported by Arokoski et al. (2002). A signiﬁcant difference between control and mild pathology groups for the GMED muscle however, provides some important information for understanding changes occurring in this muscle, and inconsistencies in previous EMG research. For subjects with mild joint pathology, GMED muscle volume of the affected side was on average 16% larger than those of normal control subjects (p < 0.05). This information may indicate that the GMED muscle could be more predisposed to hypertrophy rather than atrophy in the early stages of joint pathology. This could help explain why subjects with early OA of the hip exhibit higher levels of EMG for this muscle (Sims et al., 2002), while patients just prior to arthroplasty exhibit reduced GMED EMG activity (Long et al., 1993). Differing gait patterns may provide some further explanation for the apparent disparity in GMED response across stages of joint pathology. As GMED muscle atrophy appears inherently linked to ofﬂoading strategies used in gait during late stage joint pathology (Krebs et al., 1998), GMED muscle hypertrophy may occur in early joint pathology associated with increases in relative hip adduction (Watelain et al., 2001). Kumagai et al. (1997) determined that the GMED muscle provides maximal contribution to abduction force from a position of 20 hip adduction and more speciﬁcally, the most superﬁcial, ‘middle’ portion of the GMED muscle is more active in a position of hip adduction than the deeper anterior and posterior

A. Grimaldi et al. / Manual Therapy 14 (2009) 605–610

609

Furthermore in the early stages of pathology motor control changes are likely to preempt changes in muscle size. Future research aimed at quantifying not only size, but ideally concurrent dynamic EMG activity of each member of the abductor synergy, including the functionally separate portions within the GMED muscle, may be able to elucidate the speciﬁc functions and exercise requirements for muscles of the abductor synergy. 5. Conclusion

Fig. 2. The three separate portions of the gluteus medius muscle. Anterior (A), Middle (M), Posterior (P).

portions (Fig. 2), and the GMIN muscle, which are favoured in a more neutral hip position. Increasing pelvic tilt or lateral shift to a position of increased adduction may be an inherent compensatory strategy to increase the contribution from the more superﬁcial abductors to lateral pelvic support. This alignment not only creates preferential recruitment in the superﬁcial portion of the GMED muscle, but also pretensions the iliotibial band potentially increasing the effect of the TFL and UGM muscles. As the GMED muscle is composed of 3 fascially distinct portions, anterior and posterior portions sitting deep to the middle portion (Jaegers et al., 1992) (Fig. 2), all with independent nerve supply (Gottschalk et al., 1989), it is possible that while the overall volume of the GMED muscle increased, the deeper anterior and posterior portions may be responding differently to their superﬁcial counterpart. 4.3. Possible clinical implications Information from this and our previous study (Grimaldi et al., in press) together demonstrate that the abductor synergy does not respond homogenously to joint pathology. While the deeper abductor muscles GMED, PIRI and GMIN demonstrate atrophy in subjects with advanced OA, superﬁcial abductor muscles UGM and TFL appear less affected by underlying pathology. Another ﬁnding of important clinical signiﬁcance is that the GMED muscle may hypertrophy in patients with mild joint pathology. In light of the fact that peak acetabular pressures during gait are associated with peaks in GMED muscle activity (Krebs et al., 1998), non speciﬁc exercise programmes focusing on generalised abductor strengthening may need to be reassessed. Programmes assessing and retraining speciﬁc portions of the abductor synergy, with particular attention to pelvic-femur alignment, may be most effective in both rehabilitation and prevention strategies. Real time ultrasound has been used successfully for assessment and speciﬁc rehabilitation of deep trunk musculature (Stokes et al., 1997; Painter et al., 2007). This tool also holds great potential for use in assessment and retraining of deeper members of the hip abductor synergy. 4.4. Limitations and future directions This study provides information from only a small subject population. This may have impacted on our ability to demonstrate signiﬁcant differences in muscle size in subjects with mild pathology. The other factor that may have resulted in underestimation of muscle loss is the technique of measuring around the circumference of a muscle. This technique does not account for replacement of viable muscle tissue with intramuscular fatty or connective tissue. As fatty atrophy has been shown to be unevenly distributed within the GMED and GMIN muscles (Pﬁrrmann et al., 2005) however, the use of a volume measurement should provide the most valid estimation of muscle size in comparison to a single CSA.

This study has shown that the deeper members of the hip abductor synergy, the GMED, GMIN, and PIRI muscles are smaller around the affected hip in subjects with advanced unilateral hip joint pathology. This atrophy was not measurable in subjects with mild pathology, however differing processes are likely in place associated with differing functional weight-bearing patterns. In subjects with mild pathology GMED muscle size was signiﬁcantly larger on the affected side than control group subjects suggesting the GMED muscle may hypertrophy at this stage of pathology. Assessment and rehabilitation strategies should carefully consider stage of pathology and speciﬁc changes occurring within the abductor synergy. This more speciﬁc approach may improve long term outcomes of conservative intervention in the management of OA of the hip, and may provide a direction for future prevention programmes. References Alkner BA, Tesch PA. Knee extensor and plantar ﬂexor muscle size and function following 90 days of bed rest with or without resistance exercise. European Journal of Applied Physiology 2004;93:294–305. Angielczyk A, Bronarski J. Electromyographic analysis of the gluteus medius muscle in osteoarthritis of the hip. Chirurgia Narzadow Ruchu I Ortopedica Polska 1982;47:201–4. Arokoski MH, Arokoski JPA, Haara M, Kankaanpaa M, Vesterinen M, Niemitukia LH, et al. Hip muscle strength and muscle cross sectional area in men with and without hip osteoarthritis. Journal of Rheumatology 2002;29:2185–95. Beck M, Sledge J, Gautier E, Dora C, Ganz R. The anatomy and function of the gluteus minimus muscle. Journal of Bone and Joint Surgery British 2000;82B(2): 358–63. Byrd JWT, Jones KS. Prospective analysis of hip arthroscopy with 2-year follow up. Arthroscopy 2000;16(6):578–87. Fukunaga T, Roy RR, Shellock FG, Day MK, Lee PL, Kwong-Fu H, et al. Physiological cross-sectional area of human leg muscles based on magnetic resonance imaging. Journal of Orthopedic Research 1992;10(6):926–34. Grimaldi AM, Richardson CA, Hides JA, Donnelly W, Durbridge G. The association between degenerative hip joint pathology and size of the gluteus maximus and tensor fascia lata muscles. Manual Therapy, in press. Gottschalk F, Kourosh S, Leveau B. The functional anatomy of tensor fascia latae and gluteus medius and minimus. Journal of Anatomy 1989;166:179–89. Herneth A, Philip M, Pretterklieber M, Balassy C, Winkelbauer F, Beaulieu C. Asymmetric closure of ischiopubic synchondrosis in pediatric patients: correlation with foot dominance. American Journal of Radiology 2004;182(2):361–5. Hides JA, Richardson CA, Jull GA. Multiﬁdus muscle recovery is not automatic after resolution of acute, ﬁrst-episode low back pain. Spine 1996;21(23):2763–9. Hides J, Stokes M, Saide M, Jull G, Cooper D. Evidence of lumbar multiﬁdus muscle wasting ipsilateral to symptoms in patients with acute/subacute low back pain. Spine 1994;19:165–72. Hirsch R, Fernandes RJ, Pillemer SR, Hochberg MC, Lane NE, Altman RD, et al. Hip osteoarthritis prevalence estimates by three radiographic scoring systems. Arthritis & Rheumatism 1998;41(2):361–8. Jaegers S, Dantuma R, deJongh H. Three dimensional reconstruction of the hip on the basis of magnetic resonance images. Surgical Radiologic Anatomy 1992;14:241–9. Jandric S. Muscle parameters in coxarthrosis. Medicinski Pregled 1997;50 (7–8):301–4. Kellgren J, Lawrence J. Radiological assessment of osteoarthritis. Annals of the Rheumatic Diseases 1957;16:494–502. Kumagai M, Shiba N, Higuchi F, Nishimura H, Inoue A. Functional evaluation of hip abductor muscles with use of magnetic resonance imaging. Journal of Orthopaedic Research 1997;15:888–93. Krebs DE, Robbins CE, Lavine L, Mann RW. Hip biomechanics during gait. Journal of Orthopedic and Sports Physical Therapy 1998;28(1):51–9. Long W, Dorr L, Healy B, Perry J. Functional recovery of noncemented total hip arthroplasty. Clinical Orthopaedics and Related Research 1993;288:73–7.

610

A. Grimaldi et al. / Manual Therapy 14 (2009) 605–610

March LM, Bagga H. Epidemiology of osteoarthritis in Australia. Medical Journal of Australia 2004;180(Supplement):S6–17. Murray MP, Sepic SB. Maximum isometric torque of hip abductor and adductor muscle. Physical Therapy 1968;48:1327–35. Painter E, Ogle M, Tehyen D. Lumbopelvic dysfunction and stress urinary incontinence: a case report applying rehabilitative ultrasound imaging. Journal of Sport and Physical Therapy 2007;37(8):499–504. Pﬁrrmann CWA, Notzli HP, Dora C, Hodler J, Zanetti. Abductor tendons and muscle assessed at MR imaging after total hip arthroplasty in asymptomatic and symptomatic patients. Radiology 2005;235:969–76. Sims K, Richardson CA, Brauer SG. Investigation of hip abductor activation in subjects with clinical unilateral osteoarthritis. Annals of the Rheumatic Diseases 2002;61:687–92. Stokes M, Hides J, Nassiri D. Musculoskeletal ultrasound imaging: diagnostic and treatment aid in rehabilitation. Physical Therapy Reviews 1997;2(2): 73–92.

Taylor HL, Jacobs DR, Schucker B, Knudsen J, Leon AS, Debacker G. A questionnaire for the assessment of leisure time activities. Journal of Chronic Diseases 1978;31:741–55. Teshima K. Hip abduction force in osteoarthritis of the hip. Acta Medica Nagasakiensia 1994;39(3):21–30. Watelain E, Dujardin F, Babier F, Dubois D, Allard P. Pelvic and lower limb compensatory actions of subjects in an early stage of hip osteoarthritis. Archives of Physical Medicine and Rehabilitation 2001;82:1705–11. Wallwork TL, Hides JA, Stanton WR. Intrarater and interrater reliability of assessment of lumbar multiﬁdus muscle thickness using rehabilitative ultrasound imaging. Journal of Orthopedic and Sports Physical Therapy 2007;37(10): 608–12. Walters J, Solomons M, Davies J. Gluteus minimus: observations on its insertion. Journal of Anatomy 2001;198:239–42. Williams P, Warwick R, Dyson M, Bannister L. Grays anatomy. 37th ed. Edinburgh: Churchill Livingstone; 1989.

Manual Therapy 14 (2009) 611–617

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

The association between degenerative hip joint pathology and size of the gluteus maximus and tensor fascia lata muscles Alison Grimaldi a, *, Carolyn Richardson a, Gail Durbridge b, William Donnelly c, Ross Darnell a, Julie Hides a, d a

Division of Physiotherapy, School of Health and Rehabilitation Sciences, The University of Queensland, Brisbane 4072, Australia Centre for Magnetic Resonance Imaging, Brisbane, Australia Brisbane Orthopaedic Specialist Services, Brisbane, Australia d The UQ/Mater Back Stability Clinic, Mater Health Services, Raymond Terrace, South Brisbane, Queensland 4101, Australia b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 10 August 2007 Received in revised form 28 October 2008 Accepted 8 November 2008

The aim of this study was to obtain, using Magnetic Resonance Imaging (MRI), muscle volume measurements for the gluteus maximus (upper: UGM and lower: LGM portions) and tensor fascia lata (TFL) muscles in both healthy subjects (n ¼ 12) and those with unilateral osteoarthritis (OA) of the hip (mild: n ¼ 6, and advanced: n ¼ 6). While control group subjects were symmetrical between sides for the muscles measured, subjects with hip joint pathology showed asymmetry in GM muscle volume dependent on stage of pathology. The LGM demonstrated atrophy around the affected hip in subjects with advanced pathology (p < 0.05), however asymmetry of the UGM (p < 0.01) could be attributed largely to hypertrophy on the unaffected side, based on between group comparisons of muscle volume. TFL showed no signiﬁcant asymmetry, or difference compared to the normal control group. This study highlights the functional separation of UGM and LGM, and the similarities of the UGM and TFL, both superﬁcial abductors appearing to maintain their size around the affected hip. Further research is required to determine the speciﬁc changes occurring in the deeper abductor muscles. This information may assist in the development of more targeted and effective exercise programmes in the management of OA of the hip. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Hip osteoarthritis Gluteus maximus Tensor fascia lata Magnetic resonance imaging

1. Introduction Therapeutic exercise has been cited as an important approach used in management of osteoarthritis (OA) of the hip (Hochberg et al., 1995; Altman et al., 2000; Smidt et al., 2005; National Collaborating Centre for Chronic Conditions, 2008; Zhang et al., 2008). There is however, a distinct scarcity of literature investigating the effectiveness of therapeutic exercise of the hip. Programmes have often been quite generalised with small to moderate short term effects and poorer long term effects (van Baar et al., 2001; Tak et al., 2005). Outcomes may be improved through the development of more speciﬁc programmes based on a greater understanding of muscle function and dysfunction around the hip joint. One of the most consistent ﬁndings in subjects with hip dysfunction is an inability to maintain adequate lateral control of the hip and pelvis in single leg stance (Hardcastle and Nade, 1985). Studies assessing hip abductor muscle strength in subjects with OA

* Correspondence to: Alison Grimaldi, PhysioTec Physiotherapy, 23 Weller Road, Tarragindi, Brisbane, Queensland 4121, Australia. Tel./fax: þ61 7 3342 4284. E-mail address: [email protected] (A. Grimaldi). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.11.002

of the hip have found deﬁcits of up to 31% (Murray and Sepic, 1968; Jandric, 1997; Arokoski et al., 2002), while others have found no signiﬁcant losses in abductor strength (Teshima, 1994; Sims et al., 2002). These apparent inconsistencies may be associated with speciﬁc changes occurring within muscles of the abductor synergy, and the association of these changes with stage of pathology. While strength testing provides information on global abductor muscle function, a resultant effect of all synergists, speciﬁc changes within the synergy will only become evident by addressing each muscle individually. Muscles of the abductor synergy providing lateral stability of the hip and pelvis could be divided into superﬁcial muscles that provide their effect via insertion into the iliotibial band (ITB), and deeper muscles that act via insertion into the greater trochanter. Muscles of the superﬁcial system include the tensor fascia lata (TFL) muscle and the gluteus maximus (GM) muscle. The deep system would include the gluteus medius (GMED), piriformis (PIRI) and gluteus minimus (GMIN) muscles. This paper will focus on the study of muscles of the superﬁcial system, while the deep muscle system will be addressed in a further publication (Grimaldi et al., unpublished). In clinical rehabilitation settings, the GM muscle has been targeted for strengthening exercises, due to its reported tendency to

612

A. Grimaldi et al. / Manual Therapy 14 (2009) 611–617

weaken and atrophy (Janda, 1983; Sims, 1999; Sahrmann, 2002). In contrast, the TFL muscle has been targeted for lengthening techniques, due to its reported tendency to become excessively active (Janda, 1983; Sims, 1999; Sahrmann, 2002). There has been little attention paid in either research or clinical settings, to the impact of the functional differentiation of the GM muscle on joint mechanics and the prescription of therapeutic exercise. The upper portion of the GM muscle (UGM) arises from the posterior iliac crest, while the lower portion of the GM muscle (LGM) arises from the inferior sacrum and upper lateral coccyx (Williams et al., 1989). Despite a lack of fascial separation in adult humans, studies on morphogenesis of the GM muscle have revealed that it arises from two muscle primordia with a loose connective tissue separation between cranial and caudal portions in the foetus followed by fusion in the prenatal period (Tichy and Grim, 1985). The UGM, acting above the centre of rotation of the hip, has a primary function of hip abduction, and does not have a role in hip extension. While both portions may externally rotate the femur, the lower portion of the GM muscle (LGM), acting below the centre of rotation of the hip, is the primary hip extensor (Stern, 1972; Stern et al., 1980; Lyons et al., 1983; Jaegers et al., 1992) playing an important protective role in absorbing ground reaction forces at heel strike during gait. The role of the hip abductor synergy in joint protection is less clear. While hip abductor strengthening is generally considered as a priority in patients with hip pain, an in vivo study on joint loads during gait revealed that peak joint loads were associated with peaks in hip abductor muscle activity during stance phase rather than solely loads applied from body weight (Krebs et al., 1998). Contrary to common clinical belief, the authors from this study recommended that clinicians aiming to reduce joint load should reduce hip abductor activity. Another important aspect that should be considered in the prescription of therapeutic exercise for patients with OA of the hip is the stage of pathology. While global atrophy of hip muscles may be present in end stage pathology, in the earlier stages of the condition, more speciﬁc changes in the muscles of the hip abductor synergy may occur. It has been proposed that these changes can result in alteration of the orientation of the resultant hip joint vector, and ultimately result in joint damage over time (Kummer, 1993; Sims, 1999). Further information pertaining to hip muscle dysfunction at different stages of pathology would be useful as it could be used in the development of more speciﬁc and possibly more effective conservative intervention or prevention programmes for those with degenerative hip joint pathology. Imaging studies provide an excellent opportunity to analyse individual muscles of the hip. Only one study has measured muscle size in subjects with OA of the hip. Arokoski et al. (2002) used magnetic resonance imaging (MRI) to measure hip muscle cross sectional area (CSA) in men with and without hip OA. Two axial slices through the pelvis provided a single CSA for LGM and a combined CSA of all hip abductors, including the UGM. This measure unfortunately failed to provide speciﬁc information of individual muscles of the abductor synergy. Furthermore, volume measurements rather than single slice CSA measurements, may be more representative of the complex pelvic musculature. One study has reported muscle volume measurements of the hip muscles for

three healthy subjects (Jaegers et al., 1992), but no volume measurements have been reported in subjects with hip OA. The main aim of this study was to investigate size of the muscles of the superﬁcial lateral stability mechanism of the hip, TFL and GM muscles, in subjects with either mild or advanced degenerative pathology of the hip. Subjects with unilateral pathology were selected in order to provide both within and between subject comparisons. The speciﬁc aims were to examine i) if there was signiﬁcant side asymmetry in the superﬁcial muscles across 3 groups (mild degenerative change, advanced degenerative change, matched controls), ii) if there were signiﬁcant differences in actual muscle size among the pathology and control groups, and iii) if the functionally separate portions of the GM muscle, UGM, and LGM, display similar patterns of change in subjects with hip pathology. This study also examined the association of both stage of pathology, and muscle size, with the factors of age, height, weight, pain, function and activity levels. The hypotheses of the study were that ia) there would be significant asymmetry in size of the UGM, LGM, and TFL in subjects with hip joint pathology, but not in controls, ib) asymmetry would be greater in subjects with advanced pathology, ii) the affected side LGM muscle would be smaller that the comparable side in control subjects, based on clinical expectation (Sims, 1999; Sahrmann, 2002), and iii) changes in the UGM would more closely reﬂect changes in the TFL muscle based on their close functional relationship. 2. Methods 2.1. Subjects Twenty-four subjects (12 subjects with hip joint pathology and 12 control subjects) were recruited for this study via community advertisement and via contact with medical practitioners. Control subjects were recruited to match each subject with pathology by sex and age. The age of the control subject was required to be within 5 years of the age of the matched subject with hip pathology. There was an equal distribution of males and females in each group. Subject details are listed in Table 1. Subjects with hip joint pathology were included in the study if they had both a medical diagnosis and radiographic evidence of unilateral degenerative hip joint pathology. Radiographic evidence included X-Ray or MRI demonstrating OA or atraumatic, degenerative labral pathology. OA of the hip joint was classiﬁed by an experienced radiologist using the Kellgren/Lawrence (K/L) global scoring system (Kellgren and Lawrence, 1957; Hirsch et al., 1998). Six subjects with early joint space narrowing and osteophytes (K/L grades 1–2) were recruited for the ‘Mild Group’ and 6 subjects with moderate to severe joint space narrowing and osteophytes (K/L grades 3–4) were recruited for the ‘Advanced Group’. Seven subjects had left sided pathology and ﬁve subjects had right sided pathology. Exclusion criteria for all subjects included any systemic disease affecting the muscular or nervous system, history of congenital or adolescent hip disease, hip trauma or previous surgery, inﬂammatory joint disease, presence of tumour, any lower limb injuries in the previous 2 years, participation in unilateral sports, use of a walking aid, and factors that would preclude them from MRI

Table 1 Subject characteristics for each group. Group No Sex M:F Age Mean(SD) Weight(kg) Mean(SD) Height(cm) Mean(SD) AMI Mean(SD) Mild Adv Con

6 6 12

3:3 3:3 6:6

46.5(9.5) 57.7 (6.7) 51.8 (9.7)

80.4 (15.1) 78.3 (8.5) 73.5 (13.3)

171.3 (9.7) 172.0 (7.4) 168.2 (10.2)

MHHS(P) Mean(SD) MHHS(F) Mean(SD) MHHS(Total) Mean(SD)

63 667 (23 884) 25 (10.5) 82 890 (75 410) 16.7 (5.2) 123 175 (68 766) –

41.5 (3.0) 36.2 (5.5) –

73.2 *(11.3) 58.1 *(58.7) –

No ¼ Number. BMI ¼ Body Mass Index. AMI ¼ Activity Metabolic Index. MHHS ¼ Modiﬁed Harris Hip Score. P ¼ Pain. F ¼ Function. M:F ¼ Male:Female. SD ¼ Standard deviation. Adv ¼ Advanced Pathology. Con ¼ Control. *Signiﬁcant difference between pathology groups (p < 0.05).

A. Grimaldi et al. / Manual Therapy 14 (2009) 611–617

scanning procedures (eg. pacemaker, metal implants, pregnancy, claustrophobia). Subjects in both groups were also excluded if they had experienced any lower back pain in the previous 2 years or if there had been any signiﬁcant lifetime history of lower back pain that resulted in a period of immobility, or required further investigation or treatment. Subjects in the control group were excluded if they had any history of hip pain. Information on the study was sent to the subjects prior to admission to the study. Ethical approval was obtained from the institutional review boards and informed consent was obtained from all subjects.

2.2. Procedure 2.2.1. Self-report questionnaires Information on subject activity levels was gathered using a 12 month Leisure Time Physical Activity questionnaire providing an activity metabolic index (AMI) (Taylor et al., 1978; Arokoski et al., 2002). Activities were coded using the intensity code provided (Taylor et al., 1978).The AMI for each activity the subject participated in was calculated with the formula: AMI ¼ Intensity code (mean metabolic units) average number of times per month the number of months per year (frequency) the time the activity was performed per occasion (duration). Total AMI reﬂects the addition of AMI for all activities (Taylor et al., 1978) and provides a measure of metabolic units used per year. The Modiﬁed Harris Hip Score (MHHS) was used to assess pain and function in the subjects with OA of the hip (Byrd and Jones, 2000). The pain section consisted of 44 points, where a score of 44 represents a pain-free state. The function section consisted of 47 points, where a score of 47 points represents full, normal function. The multiplier 1.1 was used to achieve a total score out of a possible 100 (pain-free normal function).

613

2.2.2. Testing of leg dominance Subjects were also tested for leg dominance. Kicking was used as the test function (Herneth et al., 2004). The weight-bearing leg was recorded as ‘‘stance dominant’’ and the kicking leg as the ‘‘skill dominant’’ leg. 2.2.3. MRI assessment Subjects were ﬁrst screened for contraindications to MRI by a medical practitioner. Subjects were positioned in supine lying with legs extended to a neutral position. Leg rotation was controlled with the use of sand bags. A 1.5 Tesla Siemens Sonata MR system was used. A T2 True Fast Imaging with Steady State Progression (FISP) sequence using 2 series of 28 6 mm contiguous slices from the iliac crest to the most distal extent of the GM muscle was employed (Time to Repetition (TR): 3.78 ms/Echo Time (TE): 1.89 ms/Field of View (FOV): 390 mm). 2.2.4. Measurement procedure An MRI measurement software package (Osiris) was used to measure CSA (cm2) of UGM, LGM and TFL muscles on each image in which the muscle appeared. Muscle volume (cm3) was calculated by multiplying CSA by slice width and then adding the volumes from each slice to determine the total muscle volume (Fukunaga et al., 1992; Alkner and Tesch, 2004) (Fig. 2). The two functionally separate parts of GM were measured (UGM and LGM). The UGM includes that part of the muscle acting above the centre of rotation of the femoral head. These ﬁbres insert almost exclusively into the ITB via a thick laminar tendon (Lieberman et al., 2006). The LGM inserts below the centre of rotation, superﬁcial ﬁbres into the ITB, deep ﬁbres into the gluteal ridge of the femur (Lieberman et al., 2006). This anatomy is depicted in Fig. 1. In this study the largest CSA of the femoral head was used as an anatomical landmark to functionally separate the UGM from the LGM muscle, to approximate the centre of rotation of the femoral head (Stern, 1972). Reliability of the assessor’s measurement technique was tested by retracing all slices of one subject (44 slices) with an interim period of 6 weeks. Intra-tester reliability was tested for each separate measurement on each slice using a two sided bootstrapped interval of intraclass correlation coefﬁcient (ICC2,1). Intrarater reliability was found to be good, with correlation coefﬁcients ranging from 0.87 to 0.99. Standard error of measurement (SEM) was calculated using the formula SEM ¼ pooled SD (1-ICC)1/2 (Wallwork et al., 2007). Standard deviation of the difference (SDD) was also calculated as the standard deviation of the differences between measurement 1 and 2. SEM for the GM muscle was 0.495 cm2 and the SDD was 3.87 cm2, while for the TFL muscle the SEM was 0.536 cm2 and the SDD was 2.44 cm2. These values represent good measurement stability with low error. 2.3. Statistical analysis

Fig. 1. Diagramatic representation of the portions of the GM muscle. UGM ; ITB .

; LGM

The comparison of muscle volumes among groups and between sides was performed using a mixed linear model describing muscle volume with group as a between-subject factor, and side as a within-subject factor (Dependent variable ¼ muscle volume, Independent variables ¼ sides and groups). Each muscle was analysed separately. Contrasts of means were performed to compare sides within groups. Muscle volumes around the affected and unaffected hips of the subjects with hip joint pathology were compared with muscle volumes of the corresponding sides of their matched control subjects. That is, if the pathological side was left, the left side muscle volume of the matched control subject was used for comparison, and the right compared with the unaffected side value of the pathology group counterpart. Percent differences were calculated using the formula: % Difference ¼ [(larger value smaller value)/larger value] 100 (Hides et al., 1996).

614

A. Grimaldi et al. / Manual Therapy 14 (2009) 611–617

Fig. 2. Axial MRIs through the pelvis: images A,C & E through ilia showing UGM in the proximal pelvis; Images B, D & F showing LGM and TFL just below the hip joint. A & B: control TFL . subject; C & D: subject with mild left hip OA (right side as viewed in image); E & F: subject with advanced left hip OA. GM

Analyses were also conducted to assess participant characteristics in relation to a) the similarity of the groups and b) the extent of association with muscle size. One way analysis of variance was used to assess group equivalence across each of the dependent measures of age, height, weight, pain, function, and metabolic activity. The association between these patient characteristics and UGM, LGM, or TFL muscle size was assessed using analysis of covariance.

signiﬁcant. In the group with advanced pathology there were signiﬁcant between side differences in the GM but not the TFL muscle. The asymmetry was greater in the UGM muscle (mean difference 21%, p < 0.01) than the LGM muscle (mean difference 19.7%, p < 0.05). Means, standard deviations, and percentage difference in muscle volumes are reported for each group in Table 2. Examples of side to side differences are illustrated for each group in Fig. 2.

3. Results 3.2. Differences in muscle volumes between groups 3.1. Side to side differences in muscle volumes within groups There were no signiﬁcant side to side differences in the control or mild pathology groups. While LGM size was smaller on the affected side in all but one subject in the group with mild joint changes, the asymmetry was not great enough to be statistically

No signiﬁcant differences in muscle volumes were found between the mild and advanced pathology groups. The UGM muscles were signiﬁcantly larger on the unaffected side (Mean difference 30.5%) of the subjects in the advanced pathology group when compared with matched controls (p < 0.05, Table 3). No other

A. Grimaldi et al. / Manual Therapy 14 (2009) 611–617 Table 2 Side to side differences in muscle volume (cm3), and percentage differences within groups for UGM, LGM, and TFL muscles. GROUP

SIDE

UGM Mean (SD)

LGM Mean (SD)

TFL Mean (SD)

Mild

Affected Unaffected % Difference Affected Unaffected % Difference Left Right % Difference

405 (70) 421 (60) 3.8% 378 (96) 479 (118) 21.0% 352 (106) 359 (125) 2.0%

508 (118) 539 (120) 5.8% 457 (158) 569 (144) 19.7%* 453 (130) 495 (158) 8.6%

82.5 (20) 73.8 (19) 10.5% 86.2 (38) 89.5 (27) 3.8% 74.3 (24) 80.6 (29) 7.8%

Advanced

Control

615

4. Discussion This study investigated the inﬂuence of degenerative hip joint pathology on size of the GM and TFL muscles. 4.1. Side to side differences in muscle volumes within groups

SD ¼ Standard Deviation. UGM ¼ Upper gluteus maximus muscle. LGM ¼ Lower gluteus maximus muscle. TFL ¼ Tensor fascia lata muscle. p < 0.01 * p < 0.05.

comparisons reached statistical signiﬁcance although LGM volumes were on average 15.2% larger (p ¼ 0.12) on the unaffected side in the group with advanced pathology, compared with controls, statistical analysis did not reveal a signiﬁcant difference in this relatively small sample size. Means, standard deviations, and percentage difference in muscle volumes around matched hips of the pathology and control groups are reported in Table 3.

3.3. Self-report questionnaires Results of the AMI for all subjects and the MHHS for subjects with OA of the hip are shown in Table 1. Pain and function scores were lower for the group with more advanced radiological changes, reﬂecting higher pain levels and more functional disability, as measured by the MHHS. These scores considered alone were not signiﬁcantly different statistically, however when the total score was calculated there was a signiﬁcant difference between scores in the mild and advanced pathology groups (p < 0.05). There was no statistically signiﬁcant difference between groups for the AMI.

3.4. Leg dominance All subjects were left stance dominant/right skill dominant.

3.5. Effect of subject characteristics on muscle size Results of the analyses indicated the groups were comparable in terms of age, height, weight, and metabolic activity (all p > 0.05). In addition there was no signiﬁcant relationship between these patient characteristics, or pain and function, and UGM, LGM or TFL muscle volume (p > 0.05).

The results of this study showed that subjects with demonstrated unilateral hip joint pathology exhibited marked side to side differences in the size of the GM muscle, speciﬁc to stage of pathology. While asymmetry in LGM size in subjects with mild joint pathology was not great enough to be statistically signiﬁcant, in those with advanced joint changes the mean volume of the LGM muscle was on average 19.7% smaller on the affected side (p < 0.05). The only previous study to investigate muscle size in those with OA of the hip/s reported that the mean CSA of the LGM muscle was 9% smaller on the side of the worse hip in those with either unilateral or bilateral OA (Arokoski et al., 2002). The most likely explanations for the smaller percent difference are the inclusion of subjects with bilateral pathology in the latter study which would be expected to reduce the degree of side to side difference demonstrated, and the inclusion of subjects with both mild and advanced joint pathology in the analysis. Some explanation may also be provided by the different measurement techniques. A single CSA measurement may not provide a true reﬂection of change in total muscle volume. The UGM muscle similarly showed no signiﬁcant side to side difference in those with mild joint pathology. In the presence of advanced pathology, the UGM was on average 21% smaller on the affected side, representing a signiﬁcant side to side difference in muscle size (p < 0.01). The TFL muscle was not signiﬁcantly different between sides in either pathology group, although the mild group was on average 10.5% larger on the affected side. In contrast Arokoski et al. (2002) reported that the CSA of the TFL muscle was 13% smaller on the more affected side in men with OA. This difference is again most likely due to differences in subject selection and/or measurement technique. Another important consideration when interpreting side to side differences in muscle size is that in the absence of longitudinal data, the determination of side to side differences as atrophy or hypertrophy around weight-bearing joints must be approached with caution. Side to side differences could reﬂect either atrophy or hypertrophy. Decreases in muscle size on the affected side could occur in response to pain (Lund et al., 1991) or reﬂex inhibition (Stokes and Young, 1984). However, as pain causes an instinctive shift in weight-bearing towards the unaffected side, side to side volume differences may occur due to disuse atrophy around the affected hip and/or overuse hypertrophy of the unaffected side. For this reason, a control group was included for comparison of actual muscle volumes between groups, thereby assisting in the interpretation of side to side differences. 4.2. Differences in muscle volumes between groups

Table 3 Between group differences in muscle volume (cm3) for UGM, LGM, and TFL muscles. SIDE

GROUP

UGM Mean (SD)

LGM Mean (SD)

TFL

Affected

Mild Advanced Controla Mild Advanced Controla

405 378 354 421 479 361

508 457 460 539 569 489

82.5 86.2 74.9 73.8 89.5 75.4

Unaffected

(70) (96) (103) (60) (118)* (119)

(118) (158) (128) (120) (144) (150)

(20) (38) (24) (19) (27) (26)

UGM ¼ Upper gluteus maximus muscle. LGM ¼ Lower gluteus maximus muscle. TFL ¼ Tensor fascia lata muscle. SD ¼ Standard Deviation. Side refers to the named side in the pathology group, and for the control group side is aligned by matched pair dependent on side of pathology; * p < 0.05. a Reference group for signiﬁcance values.

As with Arokoski et al. (2002) study, the current study was unable to demonstrate any between group difference in LGM size. This may be simply due to the inherent variability within the population and the relatively small sample size. Another consideration is the fact that the measurement of muscle size by tracing around the perimeter of a muscle in the subjects with pathology of the hip joint may underestimate the loss of contractile muscle tissue. Replacement of normal viable muscle tissue with intramuscular fatty or connective tissue has been reported as ‘fatty atrophy’ at the hip in the GMED muscle (Pﬁrrmann et al., 2005). Differences in tissue quality of the LGM muscle are observable as increased black markings within the muscle on the side of the

616

A. Grimaldi et al. / Manual Therapy 14 (2009) 611–617

affected hip in Fig. 1D and F. This assists in the support of the assumption that side to side differences in the LGM muscle in those with hip pathology are at least in part due to atrophy around the affected hip. It is most likely however that together with atrophy around the affected hip, there may be concurrent hypertrophy of the unaffected side LGM secondary to patterns of antalgic weight shift. The ﬁnding that advanced group subjects LGM volumes were 15.2% larger on the unaffected side than matched control subjects (p ¼ 0.12), provides some support for this effect although not reaching statistical signiﬁcance. Between group differences for the UGM muscle showed that the mean muscle volume of the UGM muscle on the unaffected side in those with advanced pathology was signiﬁcantly (Mean difference 30.5%) larger than the corresponding muscles in the control group subjects. This ﬁnding suggests that the signiﬁcant asymmetry (Mean difference 21%) observed in subjects with advanced joint pathology may be largely attributable to hypertrophy on the unaffected side. Some degree of atrophy on the affected side however cannot be discounted although fatty atrophy was not commonly observed in the UGM muscle. Around the affected hip neither the UGM muscle, nor the other superﬁcial hip abductor, the TFL muscle, were signiﬁcantly different in size to a normal population. The other information that was assessed with regard to the subjects of this study was gathered through self-report questionnaires and leg dominance testing. While pain, function and leg dominance had no signiﬁcant effect on GM or TFL muscle size, the information collected provided 2 important pieces of information. 4.3. Pain, function and radiological change The ﬁrst of these relate to the association between pain, function, and radiological change. It has been previously noted that there is often no clear relationship between severity of radiological change in an osteoarthritic joint and severity of pain or degree of disability (Hurley, 1999). In studies of subjects with OA of the knee, advanced radiological change may in some people be accompanied by very little pain, while others with only mild degenerative change may experience severe disabling pain (Claessens et al., 1990; McAlindon et al., 1993). Arokoski et al. (2002) in their study of men with OA of the hip were unable to demonstrate a correlation between grade of severity of OA and pain measured on a visual analogue scale. There was however signiﬁcantly more pain within individuals on the side with the highest radiographic OA score. Similarly the ﬁndings of the current study reﬂect the difﬁculty in linking a pain score alone to degree of radiographic change. By combining measures of pain and function, the MHHS was able to demonstrate signiﬁcant differences between subjects with early radiographic change and those with advanced radiographic change. This may suggest that this particular combination of questions may be more sensitive to degree of radiographic change than those available for OA of the knee. 4.4. The inﬂuence of leg dominance The second ﬁnding of importance relates to the potentially confounding variable of leg dominance. Although there is evidence that dominance has an effect on muscle strength (Balogen and Onigbinde, 1992), particularly in upper limb strength in those involved in unilateral sports (Ducher et al., 2005; Ellenbecker et al., 2006), there is a much weaker link between leg dominance and muscle strength (Hunter et al., 2000; Zakas, 2006), and little evidence to link leg dominance to asymmetry in muscle size. Greater muscle strength of the dominant limb may be associated with improved neuromuscular functioning, rather than muscle size alone. In the current study the exclusion of all subjects involved in unilateral sports sought to avoid the effect of this potentially

confounding variable on muscle symmetry. The results of this study were able to demonstrate that for the normal control subjects tested there was no signiﬁcant asymmetry in muscle size for the muscles measured. All subjects were left stance dominant which, if this factor were imparting an effect, would favour a larger muscle volume on the left side particularly for the weight-bearing LGM muscle. This was not the case, allowing greater clarity in interpretation of results for the pathology groups. 4.5. Possible clinical implications The balance of muscle activity around a joint may either protect a joint from injury or accelerate destructive joint forces. Both the UGM and LGM muscles are known to be active at heel strike in gait to help absorb ground reaction forces causing lateral pelvic drop and ﬂexion moments at the hip and knee (Stern et al., 1980; Lyons et al., 1983). While reduced activation of the GM muscle may fail to absorb these ground reaction forces, excessive activation in the abductor muscles, may lead to an increase in joint loading (Krebs et al., 1998). So both atrophy of the LGM muscle around the affected hip, and hypertrophy of the UGM muscle around the unaffected hip may have negative effects on their respective underlying joints. Hurley (1999) has suggested that the presence of bilateral muscle dysfunction may help to explain why unilateral OA years later often becomes bilateral OA. The ﬁndings of this study imply that the LGM and UGM muscles should be assessed individually, and on both sides, with clinical management directed towards restoring normal symmetrical weight-bearing patterns and muscle bulk. Further, the ﬁnding that neither of the superﬁcial hip abductor muscles appear to be affected on the side of pathology, and recommendations to reduce recruitment of the hip abductor muscles in order to reduce peak acetabular pressures during gait (Krebs et al., 1998), the current clinical rationale for generalised hip abductor muscle strengthening could be questioned. While some authors have reported hip abductor muscle strength deﬁcits of up to 31% (Murray and Sepic, 1968; Jandric, 1997; Arokoski et al., 2002), others have reported no signiﬁcant difference (Teshima, 1994; Sims et al., 2002). These variable ﬁndings may be a reﬂection of the relative degrees of atrophy of individual muscles of the abductor synergy. If both superﬁcial abductor muscles are not signiﬁcantly affected by pathology, strength changes may possibly reﬂect weakness in the deeper abductor muscles. Together with the information provided by this study, further information on the response of the deep muscle system to degenerative change of the hip may provide further insight into speciﬁc changes within the abductor synergy. Greater speciﬁcity in exercise prescription around the hip may allow development of interventions that achieve more signiﬁcant and longer lasting changes in pain and function scores in patients with OA of the hip. 4.6. Limitations and future directions The main limitation of this study was low subject numbers. Valuable additional information may be gained by subsequent studies with larger subject numbers and the inclusion of a method to measure quality of muscle tissue. Furthermore, this study assessed only two of many hip muscles which may be associated with hip pathology. Further investigation of other muscles, such as the deeper abductor muscles, is required to provide a more complete picture of muscle dysfunction. 5. Conclusion This study has demonstrated that the GM muscle should be considered as 2 functionally separate entities, the UGM a hip abductor and the LGM, a hip extensor, these muscles having

A. Grimaldi et al. / Manual Therapy 14 (2009) 611–617

differing responses to the presence of joint pathology. The UGM muscle like its functional counterpart, the TFL, appears unaffected on the side of joint pathology, while the LGM muscle demonstrates local atrophy. The lack of affect on the superﬁcial hip abductors suggests that muscle weakness demonstrated in subjects with OA of the hip may be related to changes in the deeper hip abductors (GMED, GMIN and PIRI) and require more speciﬁc therapeutic exercise intervention. References Alkner BA, Tesch PA. Knee extensor and plantar ﬂexor muscle size and function following 90 days of bed rest with or without resistance exercise. European Journal of Applied Physiology 2004;93:294–305. Altman RD, Hochberg MC, Moskowitz RW, Schnitzer TJ. Recommendations for the medical management of osteoarthritis of the hip and knee:2000 update. Arthritis and Rheumatism 2000;43:1905–15. Arokoski MH, Arokoski JPA, Haara M, Kankaanpaa M, Vesterinen M, Niemitukia LH, et al. Hip muscle strength and muscle cross sectional area in men with and without hip osteoarthritis. The Journal of Rheumatology 2002;29:2185–95. Balogen JA, Onigbinde AT. Hand and leg dominance: do they really affect limb muscle strength? Physiotherapy Theory and Practice 1992;8(2):89–96. Byrd JWT, Jones KS. Prospective analysis of hip arthroscopy with 2-year follow up. Arthroscopy 2000;16(6):578–87. Claessens AA, Schouten JS, van den Ouweland FA, Valkenburg HA. Do clinical ﬁndings associate with radiographic osteoarthritis of the knee? Annals of the Rheumatic Diseases 1990;49:771–4. Ducher G, Courteix D, Me´me´ S, Magni C, Viala J, Benhamou C. Bone geometry in response to long-term tennis playing and its relationship with muscle volume: A quantitative magnetic resonance imaging study in tennis players. Bone 2005;37(4):457–66. Ellenbecker TS, Roetert EP, Riewald S. Isokinetic proﬁle of wrist and forearm strength in elite female junior tennis players. British Journal of Sports Medicine 2006;40:411–4. Fukunaga T, Roy RR, Shellock FG, Day MK, Lee PL, Kwong-Fu H, et al. Physiological cross-sectional area of human leg muscles based on magnetic resonance imaging. Journal of Orthopedic Research 1992;10(6):926–34. Grimaldi A, Richardson C, Stanton W, Durbridge G, Donnelly W, Hides J. The association between degenerative hip joint pathology and size of the gluteus medius, gluteus minimus, and piriformis muscles, unpublished. Hardcastle P, Nade S. The signiﬁcance of the trendelenberg test. The Journal of Bone and Joint Surgery British 1985;67B:741–6. Herneth A, Philip M, Pretterklieber M, Balassy C, Winkelbauer F, Beaulieu C. Asymmetric closure of ischiopubic synchondrosis in pediatric patients: correlation with foot dominance. American Journal of Radiology 2004;182(2):361–5. Hochberg MC, Altman RD, Brandt KD, Clark BM, Dieppe PA, Grifﬁn MR, et al. Guidelines for the medical management of osteoarthritis. Part 1. Osteoarthritis of the hip. Arthritis and Rheumatism 1995;38:1535–40. Hides JA, Richardson CA, Jull GA. Multiﬁdus muscle recovery is not automatic after resolution of acute, ﬁrst-episode low back pain. Spine 1996;21(23):2763–9. Hirsch R, Fernandes RJ, Pillemer SR, Hochberg MC, Lane NE, Altman RD, et al. Hip osteoarthritis prevalence estimates by three radiographic scoring systems. Arthritis and Rheumatism 1998;41(2):361–8. Hunter SK, Thompson MW, Adams RD. Relationships among age-associated strength changes and physical activity level, limb dominance, and muscle group in women. Journal of Gerentology 2000;55A(6):B246–72. Hurley MV. The role of muscle weakness in the pathogenesis of osteoarthritis. Rheumatic Disease Clinics of North America 1999;25(2):283–98. Jaegers S, Dantuma R, deJongh H. Three dimensional reconstruction of the hip on the basis of magnetic resonance images. Surgical Radiologic Anatomy 1992;14:241–9. Jandric S. Muscle parameters in coxarthrosis. Medicinski Pregled 1997;50(7–8): 301–4. Janda V. Muscle function testing. London, Boston: Butterworths; 1983. Kellgren J, Lawrence J. Radiological assessment of osteoarthritis. Annals of the Rheumatic Diseases 1957;16:494–502.

617

Krebs DE, Robbins CE, Lavine L, Mann RW. Hip biomechanics during gait. Journal of Orthopedic and Sports Physical Therapy 1998;28(1):51–9. Kummer B. Is the Pauwels theory of hip biomechanics still valid? a critical analysis, based on modern methods. Annals of Anatomy 1993;175:203–10. Lieberman D, Raichlen D, Pontzer H, Bramble D, Cutright-Smith E. The human gluteus maximus and its role in running. The Journal of Experimental Biology 2006;209(11):2143–55. Lund JP, Donga R, Widmer CG, Stohler CS. The pain-adaptation model: a discussion of the relationship between chronic musculoskeletal pain and motor activity. Canadian Journal of Physiology and Pharmacology 1991;69(5): 683–94. Lyons K, Perry J, Gronley JK, Barnes L, Antonelli D. Timing and relative intensity of hip extensor and abductor muscle action during level and stair ambulation. Physical Therapy 1983;63(10):1597–605. McAlindon TE, Cooper C, Kirwan JR, Dieppe PA. Determinants of disability in osteoarthritis of the knee. Annals of the Rheumatic Diseases 1993;52: 258–62. Murray MP, Sepic SB. Maximum isometric torque of hip abductor and adductor muscle. Physical Therapy 1968;48:1327–35. National Collaborating Centre for Chronic Conditions. Osteoarthritis: national clinical guideline for care and management in adults. London: Royal College of Physicians; 2008. Pﬁrrmann CWA, Notzli HP, Dora C, Hodler J, Zanetti. Abductor tendons and muscle assessed at MR imaging after total hip arthroplasty in asymptomatic and symptomatic patients. Radiology 2005;235:969–76. Sahrmann S. Diagnosis and treatment of movement impairment syndromes. St. Louis: Mosby; 2002. Sims K, Richardson CA, Brauer SG. Investigation of hip abductor activation in subjects with clinical unilateral osteoarthritis. Annals of the Rheumatic Diseases 2002;61:687–92. Sims K. The development of hip osteoarthritis: implications for conservative management. Manual Therapy 1999;4:127–35. Smidt N, de Vet HCW, Bouter LM, Dekker J, for the Exercise Therapy Group. Effectiveness of exercise therapy: a best evidence summary of systematic reviews. Australian Journal of Physiotherapy 2005;51:71–83. Stern JT. Anatomical and functional specializations of the human gluteus maximus. American Journal of Physical Anthropology 1972;36:315–40. Stern JT, Pare EB, Schwartz JM. New perspectives on muscle during locomotion. Electromyographic studies of rapid and complex behaviours. Journal of the American Orthopedic Association 1980;80(4):287–91. Stokes M, Young A. The contribution of reﬂex inhibition to arthrogenous muscle weakness. Clinical Science 1984;67(1):7–14. Tak E, Staats P, van Hespen A, Hopman-Roc M. The effects of an exercise programme for older adults with osteoarthritis of the hip. The Journal of Rheumatology 2005;32:1106–13. Taylor HL, Jacobs DR, Schucker B, Knudsen J, Leon AS, Debacker G. A questionnaire for the assessment of leisure time activities. Journal of Chronic Diseases 1978;31:741–55. Teshima K. Hip abduction force in osteoarthritis of the hip. Acta Medica Nagasakiensia 1994;39(3):21–30. Tichy M, Grim M. Morphogenesis of the human gluteus maximus muscle arising from two muscle primordia. Anatomy and Embrology 1985;173(2): 275–7. van Baar M, Dekker J, Oostendorp R, Bijl D, Voorn T, Bijlsma J. Effectiveness of exercise in patients with osteoarthritis of hip or knee: nine months’ follow up. Annals of the Rheumatic Diseases 2001;60:1123–30. Wallwork TL, Hides JA, Stanton WR. Intrarater and interrater reliability of assessment of lumbar multiﬁdus muscle thickness using rehabilitative ultrasound imaging. Journal of Orthopedic and Sports Physical Therapy 2007;37(10): 608–12. Williams P, Warwick R, Dyson M, Bannister L. Grays anatomy. 37th ed. Edinburgh: Churchill Livingstone; 1989. Zakas A. Bilateral isokinetic peak torque of quadriceps and hamstring muscles in professional soccer players with dominance on one or both two sides. Journal of Sports Medicine and Physical Fitness 2006;46:28–35. Zhang W, Moskowitz RW, Nuki G, Abramson S, Altman RD, Arden N, et al. OARSI recommendations for the management of hip and knee osteoarthritis, Part II: OARSI evidence-based, expert consensus guidelines. Osteoarthritis and Cartilage 2008;16:137–62.

Manual Therapy 14 (2009) 618–622

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

Video based measurement of sagittal range of spinal motion in young and older adultsq Yi-Liang Kuo a, Elizabeth A. Tully b, *, Mary P. Galea b a b

Department of Physical Therapy, Shu Zen College of Medicine and Management, Kaohsiung County 821, Taiwan School of Physiotherapy, Faulty of Medicine, Dentistry and Health Sciences, The University of Melbourne, 200 Berkeley Street, Parkville, Victoria 3010, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 23 May 2008 Received in revised form 20 November 2008 Accepted 3 December 2008

A revised model of skin marker placement with the two-dimensional (2D) PEAK Motus system was used to investigate the effect of aging on sagittal range of spinal motion. Twenty-four healthy young adults and twenty-two healthy older adults were videotaped while performing the movements of ﬂexion and extension in each spinal region d cervical, thoracic and lumbar spine. Alternative movement tests that may allow a greater range of motion (ROM) for thoracic extension and lumbar ﬂexion were also investigated. Older adults demonstrated signiﬁcantly decreased ﬂexion/extension ranges in the cervical, thoracic and lumbar spine. The movement of cat-stretch in the all-fours position allowed greater thoracic extension, and the movement of toe-touch in standing permitted greater lumbar ﬂexion. This study provides reference data for sagittal ranges of spinal motion in healthy young and older adults as measured by a 2D imaged-based system. The sagittal model of skin marker placement used in this study can have a broader application for ROM measurement in the clinical setting using a digital camera and freely downloadable software. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Spine Measurement Range of motion Aging

1. Introduction Knowledge of the expected range of motion (ROM) in healthy subjects provides the basis for assessment and for establishing appropriate treatment goals in clinical practice. Radiographic methods are considered the ‘gold standard’ for ROM measurement (Portek et al., 1983), however the risk of radiation exposure limits its use. Previous studies agree that increasing age is associated with decreased spinal motion, however the descriptive information provided from simple clinical tools (Loebl, 1967; Moll and Wright, 1971; Kuhlman, 1993) has frequently been jeopardized by measurement issues. For example, the Schober tape measure method (Moll and Wright, 1971) provides an index of lumbar movement (in cm) reported as unreliable (Portek et al., 1983; Miller et al., 1992). Use of a single inclinometer positioned over a speciﬁc spinous process (e.g. T12/L1) is gravity referenced, and thus only indicates the orientation of the body segment in space, with the angle being dependent on the position of the more caudal body segments. The dual inclinometer method provides more valid and

q This research was carried out as part of a PhD by Yi-Liang Kuo at The University of Melbourne. * Corresponding author. Tel.: þ61 3 8344 4171; fax: þ61 3 8344 4188. E-mail address: [email protected] (E.A. Tully). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.12.006

accurate measurement for lumbar ﬂexion (Loebl, 1967; Saur et al., 1996), however the reliability of measurement for lumbar extension has shown to be low (Merritt et al., 1986; Dillard et al., 1991). In contrast, a motion analysis system that tracks the displacement of reference markers attached to the skin over relevant bony landmarks can provide more reliable and accurate measurement of human movement. However, previous ROM studies have been compromised by problematic models of marker placement. For example, Hu et al. (2006) placed markers over a swimming cap and a sleeveless shirt to measure ROM in the cervical spine. Possible stretch or slide of the cap and clothes as well as skin movement errors likely inﬂuenced the validity and reliability of the data. In addition, the cervical spine was measured as a whole without acknowledging the functional differences between the upper and lower cervical regions. Also, in most surface-based studies information regarding ROM has been limited to a single spinal region (Dvorak et al., 1995; McGill et al., 1999; Sforza et al., 2002; Wu et al., 2007), with no attempt to determine the mobility of adjacent regions in this functionally interdependent chain of joints. Another factor inﬂuencing the ROM is test movement. The clinical tests for lumbar ﬂexion and extension are often performed in standing, and patients with balance problems may have difﬁculty achieving full lumbar extension. Similarly patients attempting thoracic extension in sitting tend to lean backwards from the hips so that the full available thoracic movement is not achieved. An

Y.-L. Kuo et al. / Manual Therapy 14 (2009) 618–622

alternative test movement may be provided by the ‘cat-stretch’, a common spinal ROM exercise performed in the all-fours position (Fig. 1). Patients arch their back upward and downward accompanied by coordinated head movement while kneeling on the hands and knees. When patients hollow their back, the end position is achieved by simultaneous extension of the thoracic and lumbar spine. The opposite movement, arching the back toward the ceiling, is comprised of thoracic and lumbar ﬂexion. The movement of catstretch in the all-fours position is stable, allows the normal movement interaction between the spinal regions, may permit a greater range of thoracolumbar motion, and also minimises the trouble of changing test positions. Therefore, the aim of this study was to use our previously developed two-dimensional (2D) model of marker placement (Tully and Stillman, 1997) with the 2D PEAK Motus video analysis system (PEAK Performance Technologies Inc., Englewood, Colorado, USA) to establish reference values for sagittal ROM in all spinal regions, and to investigate the effect of aging on regional mobility. The second aim of this study was to investigate the feasibility of an alternative movement, cat-stretch, for testing ROM in the lumbar and thoracic spine. 2. Material and methods 2.1. Subjects Twenty-four healthy young adults (15 women, 9 men; age: 17– 27 years; weight: 62.6 8.9 kg; height: 170.2 9.1 cm; BMI: 21.5 1.9 kg/m2) and 22 healthy older adults (14 women, 8 men; age: 60–83 years; weight: 69.3 12.1 kg; height: 163.9 8.4 cm;

619

BMI: 25.8 4.0 kg/m2) volunteered for this study. Exclusion criteria were: 1) signiﬁcant spinal lateral deviation or lower limb deformity; 2) severe pain and/or injury/pathology in spine or lower extremities requiring treatment during the preceding 6 months. This study was approved by the Human Research Ethics Committee of the university. Prior to involvement in this study, subjects were informed of the details of this project and signed a written consent form. 2.2. Experimental protocol Nine spherical reﬂective markers (B&L Engineering, Tustin, CA, USA) were attached to speciﬁc anatomic landmarks of subjects in standing (Fig. 2). For ease of application, one face marker was attached to the centre of the ﬂexible ear hook of a headphone piece (SHS3201/97, Philips, Sydney, NSW, Australia), while the other face marker was attached to the mid point between the right corner of the mouth and the right nasal ala. Details for locating the other skin reference markers have been reported previously (Tully et al., 2002; Kuo et al., 2008). Good test-retest reliability for skin marker placement was established in standing (ICC1, 1 ¼ 0.85–0.92). Subjects were videotaped while performing various ROM tests. They were instructed to move at their own comfortable speed to the maximal available end position and then return to the starting position. Before performing three trials, subjects were given two practice trials to familiarize themselves with the speciﬁc movement. The investigator corrected any faulty movement, including out of plane movement, during practice. A brief rest period was given between each trial. Cervical ﬂexion and extension were tested in sitting by looking down toward the chest and up toward the ceiling while maintaining an upright posture. Thoracic and lumbar ﬂexion and extension were tested when subjects performed the cat-stretch in the all-fours position. Thoracic extension was also tested in sitting by instructing subjects to arch their thoracic spine while maintaining their lower back and hips steady. Lumbar ﬂexion was also tested by performing toe touching from upright standing with knees extended. Subjects who could touch their toes stood on a platform, and were instructed to bend downward until their ﬁngers reached the edge of the platform or beyond. 2.3. Data management Fig. 2 illustrates the angle deﬁnitions and calculation. Increasing upper and lower cervical spine angles indicate extension, and

Fig. 1. End positions of cat-stretch in the all-fours position.

Fig. 2. Marker placement and angle deﬁnition.

620

Y.-L. Kuo et al. / Manual Therapy 14 (2009) 618–622

decreasing thoracic ﬂexion angle indicates extension. Negative lumbar spine angles denote lumbar extension. The videotaped images of the three test movements were automatically digitised using the 2D PEAK software program. The video images were then converted to kinematic data and smoothed using a fourth order Butterworth (high cut-off) ﬁlter (Winter, 2005) at an optimum cutoff frequency determined by the software (Jackson, 1979). The average of the three trials was used for statistical analysis. ROM was expressed as the maximal available angle and total ﬂexion/extension range. The maximal available angle was obtained from the end position of the joint/segment during the test movement. The total ﬂexion/extension range was calculated by subtracting the maximal available extension angle from the maximal available ﬂexion angle. Prior to statistical analysis, normal distribution for each numerical variable was examined. Descriptive statistics were summarised for all data accordingly. To determine the group difference in physical characteristics and sagittal spinal ROM, unpaired t-tests (or appropriate non-parametric tests) were used. Two-way repeated measures ANOVA, with group as a betweensubject factor and method as a within-subject factor, were used to determine the statistical signiﬁcance for measurement of thoracic extension and lumbar ﬂexion. If there was a signiﬁcant interaction in the repeated measures ANOVA, the simple factor analysis was performed to assess age-related differences at each position and position-related differences within the age group. The statistical signiﬁcance level was set at P < 0.05. 3. Results The young group was signiﬁcantly taller (t ¼ 2.44, P ¼ 0.02) and lighter (t ¼ 2.15, P ¼ 0.04) than the older group, and had a lower body mass index (t ¼ 4.50, P < 0.001). 3.1. Upper and lower cervical spine ﬂexion/extension Table 1 summarizes the maximal available ﬂexion and extension angles and ﬂexion/extension range in the cervical spine for both groups. The older group commenced the movements from a more forward head position, i.e. upper cervical extension (old vs. young: 126.6 vs. 114.4 ) and lower cervical ﬂexion (old vs. young: 70.4 vs. 79.6 ). When subjects were instructed to look downward the older group achieved a smaller upper cervical ﬂexion angle (t ¼ 4.80, P < 0.001) and a larger lower cervical ﬂexion angle (t ¼ 2.35, P ¼ 0.02). Although the young group extended more in both regions of the cervical spine when looking upward, the signiﬁcant difference between groups was only found in the lower cervical spine (t ¼ 6.49, P < 0.001) not in the upper cervical spine (t ¼ 1.06, P ¼ 0.3). Overall, the older group had signiﬁcantly smaller total

ranges of ﬂexion/extension in the upper (t ¼ 5.36, P < 0.001) and lower (t ¼ 5.65, P < 0.001) cervical spine. 3.2. Thoracic ﬂexion/extension When arching the back toward the ceiling during the movement of cat-stretch, the older group achieved a larger thoracic ﬂexion angle (58.8 9.0 ) than the young group (52.4 11.4 ). The mean difference of 6.4 between age groups was signiﬁcant (t ¼ 2.09, P ¼ 0.04). The maximal available thoracic extension angle (expressed as the minimal thoracic angle) was inﬂuenced by the main effects of group (F1,40 ¼ 19.91, SS ¼ 4196.11, MS ¼ 4196.11, P < 0.001) and method (F1,40 ¼ 21.73, SS ¼ 1016.68, MS ¼ 1016.68, P < 0.001). There was also a signiﬁcant interaction effect of group and method (F1,40 ¼ 14.30, SS ¼ 668.69, MS ¼ 668.69, P ¼ 0.001). The older group achieved a signiﬁcantly smaller minimal thoracic angle in the all-fours position than in the sitting position (t ¼ 4.82, P < 0.001, 95% CI 7.1 to 18.1). In other words, the older group was able to extend their thoracic spine further in the allfours position. However, the young group obtained a similar minimal thoracic angle in both positions (t ¼ 0.84, P ¼ 0.4, 95% CI 2.0 to 4.6). In both position, the older group demonstrated a signiﬁcantly larger minimal thoracic ﬂexion angle than the young group (all-fours position: t ¼ 6.03, P < 0.001, 95% CI 28.2 to 14.0; sitting: z ¼ 2.24, P ¼ 0.03) (Fig. 3), which means that older subjects had decreased maximal available thoracic extension. Overall, the total range of thoracic ﬂexion/extension during catstretch for the older group (33.6 15.6 ) was signiﬁcantly smaller than that (48.5 12.4 ) for the young group by 14.9 (t ¼ 3.59, P ¼ 0.001, 95% CI 6.5–23.2). 3.3. Lumbar ﬂexion/extension Similar to results for thoracic extension, the maximal available lumbar ﬂexion angle was affected by the main effects of group (F1,43 ¼ 29.70, SS ¼ 2017.58, MS ¼ 2017.58, P < 0.001) and method (F1,43 ¼ 107.31, SS ¼ 5265.25, MS ¼ 5265.25, P < 0.001). There was also a signiﬁcant interaction effect of group and method (F1,43 ¼ 5.58, SS ¼ 453.05, MS ¼ 453.05, P ¼ 0.02). Both groups achieved a greater maximal available lumbar ﬂexion angle during toe-touch in standing than during cat-stretch in all-fours (older group: t ¼ 7.0, P < 0.001, 95% CI 14.0 to 7.6; young group: t ¼ 7.85, P < 0.001, 95% CI 21.7 to 12.7) (Fig. 4).

Table 1 Sagittal ranges of motion in the cervical spine for the young (n ¼ 24) and older (n ¼ 22) groups. Angles

Young

Old

(Mean SD)

(Mean SD)

93.1 7.9 152.4 8.8 60.0 8.0

103.9 6.9 150.0 6.4 46.5 8.7

54.6 4.3 90.9 8.4 36.9 8.2

51.3 4.9 74.7 8.6 23.5 7.4

Mean Difference (95% CI) Upper Cervical Flexion* Extension Total range* Lower Cervical Flexion* Extension* Total range*

10.7 (6.2–15.3) 2.4 (7.0 to 2.2) 13.5 (8.4–18.6) 3.3 (6.0 to 0.5) 16.2 (21.3 to 11.2) 13.3 (8.6–18.1)

Abbreviation: SD ¼ standard deviation, CI ¼ conﬁdence interval. * Denotes a statistical signiﬁcant difference, P < 0.05.

Fig. 3. Mean and 95% conﬁdence intervals for maximal available thoracic extension angle (expressed as minimal thoracic angle) achieved in the sitting and all-fours positions in the young and older groups.

Y.-L. Kuo et al. / Manual Therapy 14 (2009) 618–622

621

4.2. Thoracic spine

Fig. 4. Mean and 95% conﬁdence interval for maximal available lumbar ﬂexion angle achieved in the all-fours and standing positions in the young and older groups.

The older group also demonstrated decreased maximal available lumbar ﬂexion angle than the young group in both all-fours (t ¼ 2.91, P ¼ 0.006, 95% CI 2.3–12.8) and standing positions (t ¼ 6.62, P < 0.001, 95% CI 9.7–18.2). No signiﬁcant difference was found in the maximal lumbar extension angle between the young (18.9 13.6 ) and older (11.1 9.2 ) groups during the movement of cat-stretch (t ¼ 1.85, P ¼ 0.07, 95% CI 12.5 to 0.5). Overall, the total range of lumbar ﬂexion/extension during catstretch for the old group (28.9 11.5 ) was signiﬁcantly smaller than that (42.4 15.8 ) of the young group by 13.5 (t ¼ 3.30, P ¼ 0.002, 95% CI 5.3–21.8).

4. Discussion This study has established reference values for sagittal range of spinal motion in young and older adults using image-based measurement. The comparison of traditional and alternative test positions provides useful information for clinicians who are involved in designing an exercise program for older adults to improve spinal ROM and in measuring the effects of the intervention.

Although older adults achieved a slightly larger (6.2 ) thoracic ﬂexion angle, again possibly associated with their more kyphotic posture (old vs. young: 42.8 vs. 38.3 ) in standing, there was a large difference in the results for thoracic extension (21 ) indicating stiffness in the thoracic spine associated with age. Whether or not this amount of limited thoracic extension is sufﬁcient to interfere with daily tasks that older subjects would usually perform, including those requiring full arm elevation remains to be determined. Results suggest that the movement of cat-stretch is more suitable for testing thoracic extension because it allows greater ROM. Performing thoracic extension in sitting appeared unnatural to most subjects, because they habitually initiated the movement from the lumbar spine, or swayed backwards from the hips. As a result, many appeared to simultaneously restrict movement in the thoracic spine. Persson et al. (2007) observed a similar ﬁnding in the movement of head protraction and retraction. When subjects were instructed to restrain their back against the backrest of the chair, their sagittal protraction/retraction excursion of the head signiﬁcantly reduced. These ﬁndings suggest that there exists a close relationship between the adjacent segments in the kinematic chain and between posture and ROM. The advantage of the cat-stretch is that it allows simultaneous movement in the adjacent cervical and lumbar spine, and maximises the physiologic ROM in the thoracic spine. Therefore, both groups, especially young subjects, achieved a smaller thoracic ﬂexion angle (increased extension) when performing the catstretch. Although other positions could have been trialled to measure maximal thoracic ﬂexion, including sitting and standing, the focus of this study was on the more clinically relevant movement of thoracic extension. A limitation of testing thoracic extension in the all-fours position was that the cat-stretch was a novel task. However, subjects became familiarised with the movement sequence after a few practice trials. In addition, although stable, older adults who have limited wrist extension or painful knees may require modiﬁcation in positioning to minimise discomfort at these joints, and some older adults may require assistance to get down or up from the position. 4.3. Lumbar spine

4.1. Cervical spine The separate measurement of upper and lower cervical ROM provided valuable information about the contribution of these functionally different regions to overall head movement. When older adults ﬂexed their head the movement was primarily limited in the upper cervical spine. On the other hand, the lower cervical spine was more restricted during attempts to extend the head and neck. These age-related changes in upper and lower cervical ﬂexibility may possibly be explained by the forward head posture seen in the older group. As the upper cervical starting position was more extended (by 12.2 ) it is possible that adaptive shortening in suboccipital muscles and ligaments may have limited head nodding. In contrast, older adults’ lower cervical spine had the biomechanical advantage of moving into ﬂexion because it was already positioned in more ﬂexion (by 9.2 ). This mechanism involving habitual posture and structural adaptation can explain age-related changes in the maximal available upper and lower cervical extension angles. Similar ﬁndings were found in young adults where a forward head posture affected the total sagittal motion of the cervical spine (Fiebert et al., 1999).

Results conﬁrmed that the commonly used movement test, toetouch in standing, produced greater lumbar ﬂexion than the alternative movement test, cat-stretch in the all-fours position. At the end position of toe-touch, tension in hamstrings possibly limited movement of the pelvis, and the gravity acting on the upper body aided lumbar ﬂexion. In other words, the lumbar spine was ‘passively’ stretched from both ends. On the other hand, arching the lumbar spine toward the ceiling in the all-fours position involved active contraction of abdominal muscles to ﬂex the lumbar region against gravity. Therefore, it was not unexpected that the maximal available lumbar ﬂexion angle achieved by toe-touch was greater than that obtained during the cat-stretch. Although lumbar ﬂexion was decreased in the older subjects, lumbar extension was not signiﬁcantly altered, possibly because at best the lumbar spine only has a relatively small range of extension at each segmental level, as shown by the radiological studies of Pearcy et al. (1984). A direct comparison of lumbar extension values with those previously reported was limited due to different measurement instruments and test movements. An advantage of the cat-stretch for measurement of lumbar extension is the removal of potential balance problems associated with use of dual

622

Y.-L. Kuo et al. / Manual Therapy 14 (2009) 618–622

inclinometers (Gerhardt et al., 2002) in standing. However, a greater range of (passive) lumbar extension may be achieved by instructing subjects to push up on their hands in the prone position. These results showing decreased ranges of spinal motion in older adults are in agreement with previous studies, however few studies have measured the upper and lower cervical spine separately and no study has compared age-related differences using surface-based methods. Greater differences in ranges of spinal motion would be expected in a more sedentary group of older adults or older adults with pathology. As a result of these ﬁndings, future studies need to include age-matched controls when investigating ROM. As with all surface-based measurement techniques, the reliability of the angles measured in this study depended largely on the operator’s ability to locate body landmarks for marker placement. Although high reliability was established in a test-retest pilot study by the investigator on the same day, the reliability over a longer time frame, and between operators, needs to be established for clinicians wishing to adopt this method. 4.4. Clinical implication The measurement method used in this study has several advantages over use of simple clinical tools, such as a universal goniometer (Norkin and White, 2003) or dual inclinometers (Gerhardt et al., 2002). Video analysis removes the problem of trying to align the stationary arm of the goniometer to an estimated vertical line, or hold the ﬂat base of two inclinometers against a curved surface. Moreover, patients do not have to hold the end range position, which may be painful in many cases. They can move easily towards and away from the limit position without restraint from the investigator’s attempt to position and read the measurement tool. In addition, as the effect of hypomobility of one spinal region may be accompanied by compensatory hypermobility in an adjacent region, video analysis has the advantage of measuring more than one joint or region at the same time. Thus the interaction of the spinal regions within the kinematic chain can be evaluated. The model of skin marker placement has a broader implication in the clinical setting. Skin reference markers can be quickly attached to the subject, and a digital camera used to obtain a sagittal image of the patient in the end range position. The limit angle for the joint of interest can then be measured on the image at a later time, using easy-to-use and freely downloadable image analysis software, ImageJ (Rasband, 1997–2007). 5. Conclusion Using surface-based measurement, older adults demonstrated signiﬁcantly decreased ﬂexion/extension ranges in the cervical, thoracic and lumbar spine compared to young adults. The movement of ‘cat-stretch’ was a feasible alternative for ROM measurement in the thoracic and lumbar spine. Thoracic extension achieved in the all-fours position was greater than in sitting, however the

‘cat-stretch’ did not appear effective in showing possible group differences in lumbar extension, and the ‘toe-touch’ in standing permitted greater lumbar ﬂexion. The obtained images, with ROM values attached, can provide a useful record in the patient’s history. References Dillard J, Traﬁmow J, Andersson GBJ, Cronin K. Motion of the lumbar spine – reliability of 2 measurement techniques. Spine 1991;16:321–4. Dvorak J, Vajda EG, Grob D, Panjabi MM. Normal motion of the lumbar spine as related to age and gender. European Spine Journal 1995;4:18–23. Fiebert IM, Roach KE, Yang SS, Dierking LD, Hart FE. Cervical range of motion and strength during resting and neutral head postures in healthy young adults. Journal of Back and Musculoskeletal Rehabilitation 1999;12:165–78. Gerhardt J, Cocchiarella L, Lea R. Measuring joints in the spine. In: The practical guide to range of motion assessment. The American Medical Association; 2002. p. 25–45 [chapter 2]. Hu HT, Li ZZ, Yan JB, Wang XF, Xiao H, Duan JY, et al. Measurements of voluntary joint range of motion of the Chinese elderly living in Beijing area by a photographic method. International Journal of Industrial Ergonomics 2006;36:861–7. Jackson K. Fitting of mathematical functions to biomechanical data. IEEE Transactions on Biomedical Engineering 1979;26:122–4. Kuhlman KA. Cervical range of motion in the elderly. Archives of Physical Medicine and Rehabilitation 1993;74:1071–9. Kuo Y-L, Tully EA, Galea MP. Skin movement errors in measurement of sagittal lumbar and hip angles in young and elderly subjects. Gait & Posture 2008;27:264–70. Loebl WY. Measurement of spinal posture and range of spinal movement. Rheumatology 1967;9:103–10. McGill SM, Yingling VR, Peach JP. Three-dimensional kinematics and trunk muscle myoelectric activity in the elderly spine: a database compared to young people. Clinical Biomechanics 1999;14:389–95. Merritt JL, McLean TJ, Erickson RP, Offord KP. Measurement of trunk ﬂexibility in normal subjects: reproducibility of 3 clinical methods. Mayo Clinic Proceedings 1986;61:192–7. Miller SA, Mayer T, Cox R, Gatchel RJ. Reliability problems associated with the modiﬁed Schober technique for true lumbar ﬂexion measurement. Spine 1992;17:345–8. Moll JMH, Wright V. Normal range of spinal mobility: objective clinical study. Annals of the Rheumatic Diseases 1971;30:381–6. doi:10.1136/ard.30.4.381. Norkin CC, White DJ. The cervical spine. In: Measurement of joint motion: a guide to goniometry. 3rd spiral ed. F.A. Davis Company; 2003. p. 181–96 [chapter 10]. Pearcy M, Portek I, Shepherd J. Three-dimensional x-ray analysis of normal movement in the lumbar spine. Spine 1984;9(3):294–7. Persson PR, Hirschfeld H, Nilsson-Wikmar L. Associated sagittal spinal movements in performance of head pro- and retraction in healthy women: a kinematic analysis. Manual Therapy 2007;12:119–25. Portek I, Pearcy MJ, Reader GP, Mowat AG. Correlation between radiographic and clinical measurement of lumbar spine movement. British Journal of Rheumatology 1983;22:197–205. Rasband WS. ImageJ, http://rsb.info.nih.gov/ij/ 1997–2007 [accessed 20.02.07]. Saur PMM, Ensink F-BM, Frese K, Seeger D, Hildebrandt J. Lumbar range of motion: reliability and validity of the inclinometer technique in the clinical measurement of trunk ﬂexibility. Spine 1996;21:1332–8. Sforza C, Grassi G, Fragnito N, Turci M, Ferrario VF. Three-dimensional analysis of active head and cervical spine range of motion: effect of age in healthy male subjects. Clinical Biomechanics 2002;17:611–4. Tully EA, Stillman BC. Computer-aided video analysis of vertebrofemoral motion during toe touching in healthy subjects. Archives of Physical Medicine and Rehabilitation 1997;78:759–66. Tully EA, Wagh P, Galea MP. Lumbofemoral rhythm during hip ﬂexion in young adults and children. Spine 2002;27:E432–40. Winter DA. Kinematics, Biomechanics and motor control of human movement. 3rd ed. New Jersey: John Wiley & Sons; 2005 [chapter 2], pp 13–57. Wu SK, Lan HHC, Kuo LC, Tsai SW, Chen CL, Su FC. The feasibility of a video-based motion analysis system in measuring the segmental movements between upper and lower cervical spine. Gait & Posture 2007;26:161–6.

Manual Therapy 14 (2009) 623–629

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

Reliability, validity and diagnostic accuracy of palpation of the sciatic, tibial and common peroneal nerves in the examination of low back related leg painq Jeremy Walsh a, *, Toby Hall b, c a

School of Physiotherapy, Royal College of Surgeons in Ireland, 123 St. Stephen’s Green, Dublin 2, Ireland School of Physiotherapy, Curtin University, Perth, Australia c Manual Concepts, Perth, Australia b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 24 June 2008 Received in revised form 2 December 2008 Accepted 14 December 2008

This study investigated the reliability, validity and diagnostic accuracy of manual palpation of the sciatic, tibial and common peroneal nerves in the examination of 45 subjects with low-back related leg pain. The nerves were palpated manually and with an algometer, to determine pressure pain thresholds (PPTs). A second examiner performed the straight leg raise (SLR) and slump tests to determine nerve trunk mechanosensitivity. The procedure was repeated by another examiner to determine inter-rater reliability (n ¼ 20). Kappa scores for agreement between raters for manual palpation were 0.80, 0.70 and 0.79 for the sciatic, tibial and common peroneal nerves respectively, demonstrating excellent reliability. PPTs were signiﬁcantly lower on the symptomatic side, for each of the three nerves, in subjects who were positive on manual palpation. In subjects who were negative on manual palpation, PPTs were not signiﬁcantly different between sides, demonstrating criterion-based validity, using PPT as the criterion. Highest scores of diagnostic accuracy were obtained when two or more of the three nerves were positive on palpation (sensitivity ¼ 0.83; speciﬁcity ¼ 0.73). Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Sciatic Tibial Common peroneal Nerve Palpation Reliability Validity Sensitivity and speciﬁcity

1. Introduction Neural tissue mechanosensitivity (local tenderness over nerve trunks and pain in response to limb movements that elongate the nerve) is a recognised feature of pain of neural origin (Dilley et al., 2005; Hall and Elvey, 2005). The straight leg raise (SLR) (Hall et al., 1998) and slump (Maitland, 1979) tests are used to assess mechanosensitivity of the sciatic nerve tract. Reproduction of symptoms in response to the SLR or slump tests, which is intensiﬁed by ankle dorsiﬂexion, is considered as one factor in the determination of sciatic nerve mechanosensitivity (Hall and Elvey, 2005). Nerve palpation has been advocated as an additional assessment technique in the examination of neural tissue pain disorders (Butler, 1989; Elvey and Hall, 1997; Jepsen et al., 2006). Under normal circumstances, peripheral nerve trunks are usually painless to non-noxious stimuli (Hall and Quintner, 1996). However, if the nerve trunks are inﬂamed, even mild mechanical provocation, such as gentle palpation, can cause pain and protective

q This work is attributed to the institution: Discipline of Physiotherapy, Trinity College Dublin, Dublin 2, Ireland. * Corresponding author. Tel.: þ353 1 4022258; fax: þ353 1 4022471. E-mail address: [email protected] (J. Walsh). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.12.007

muscle responses (Hall and Quintner, 1996). Therefore, if the sciatic nerve tract is sensitized and pain is provoked by the SLR and slump tests, then a similar pain response should be elicited by gentle nerve palpation. Furthermore, the increased sensitivity of nerve trunks to palpation (Hall and Quintner, 1996; Dilley et al., 2005) should be manifested by reduced pressure pain thresholds (PPTs) (Sterling et al., 2000). However, no studies have evaluated nerve palpation in the lower limb. The aim of this study was to determine the reliability, validity and diagnostic accuracy of manual palpation of the sciatic, tibial and common peroneal nerves in the examination of low-back related leg pain. 2. Methods In studies of diagnostic accuracy, the index test under review is compared with a reference standard (Bossuyt et al., 2003). Manual palpation was the index test and in the absence of a gold standard for sciatic nerve mechanosensitivity, the SLR and slump tests were used as the reference standard. Ethical approval was granted by the St. James’s Hospital/Adelaide and Meath hospitals incorporating the National Children’s Hospital Joint Research Ethics Committee. Subjects were able to withdraw from the study at any time and gave written informed consent prior to the study commencement.

624

J. Walsh, T. Hall / Manual Therapy 14 (2009) 623–629

2.1. Subjects Subjects were recruited from consecutive patients attending the Back Pain Screening Clinic (BPSC) at the Adelaide and Meath hospitals incorporating the National Children’s Hospital (AMNCH), Dublin, in June/July 2007. All patients underwent screening examination by one of the two attending BPSC physiotherapists as routine. Recruitment was based on presenting symptoms as determined during this examination. Consecutive patients who satisﬁed the inclusion criteria (presence of unilateral low-back related leg pain, able to understand English, age 18–70) and were not disqualiﬁed by the exclusion criteria (absence of unilateral lowback related leg pain, signs of serious pathology, history of spinal surgery or neurological disease, unable to tolerate testing process), were invited to participate in the study. Data collection was planned before the index test and reference standard were performed. 2.2. Procedure 2.2.1. Manual palpation Data were collected in the physiotherapy department at the AMNCH by three examiners who were aware of the inclusion and exclusion criteria and the presenting symptoms. After routine examination by the BPSC physiotherapist, participating subjects were seen immediately by the ﬁrst examiner (Tester 1, who had been qualiﬁed as a physiotherapist for eleven years and had completed a Masters in Manipulative Therapy qualiﬁcation seven years previously). Manual palpation was performed according to a standardised procedure (Fig. 1) using gentle pressure at three locations: the sciatic nerve at the midway point of a line from ischial tuberosity to the greater trochanter of the femur; the tibial nerve where it bisects the popliteal fossa at the mid-point of the popliteal crease; and the common peroneal nerve where it passes behind the head of ﬁbula to wind around the neck of ﬁbula (Field and Hutchinson, 2006; Moore and Dalley, 2006). Subjects wore shorts and were positioned in prone lying for palpation of the sciatic and tibial nerves and in crook lying for the common peroneal nerve. Nerves were palpated bilaterally, simultaneously. Subjects were asked if there was any pain or discomfort and if so, on which side. If pain or discomfort was reported bilaterally, the subject was asked if it was worse on one side and if so which side. Reporting of pain or discomfort on the symptomatic side, or more pain or discomfort on the symptomatic side compared to the asymptomatic side was recorded as positive. Otherwise, a negative ﬁnding was recorded. Fig. 1. Manual palpation at the a) sciatic, b) tibial and c) common peroneal nerves.

2.2.2. Mechanical palpation (PPT) Following manual palpation, mechanical palpation was performed (Fig. 2). An electronic digital algometer (Somedic AB) was used to record PPT. This consists of a circular probe with a 2 cm2 round rubber tip connected to a pressure transducer within the handle of the unit. The tester held the handle in his right hand and brought the rubber tip into contact with the site to be tested so that the probe was held perpendicular to the limb, stabilised by the testers left thumb and index ﬁnger (Sterling et al., 2000). Pressure was then applied by the tester at a rate of 50 kPa/s. Subjects were instructed to press a switch when the sensation under the probe changed from one of pressure, to one of pressure and pain. At this time the application of pressure was terminated and the measurement was stored in the memory of the algometer. To familiarise subjects with the procedure, a preliminary trial was ﬁrst performed on the forearm. For each nerve, three measures were taken (at the same sites as used in the manual palpation component of the examination) on the asymptomatic side followed by the symptomatic side, with a 10 s rest period between each

measurement. The mean of these three measures was taken as the PPT for each site. The sciatic nerve was tested ﬁrst followed by the tibial and common peroneal nerves in order. 2.2.3. SLR and slump tests On completion of the nerve palpation examination, the SLR and slump tests were immediately performed on each side (asymptomatic limb followed by symptomatic limb) on all subjects by a second examiner (Tester 2, who had been qualiﬁed as a physiotherapist for one year and was blinded to the ﬁndings of the ﬁrst examiner). For the SLR test, the subject was positioned in supine lying with the head resting on a pillow (the same pillow was used for all subjects for standardisation). The examiner passively raised the limb into hip ﬂexion (maintaining neutral adduction/abduction and internal/external rotation) with knee extension until signiﬁcant resistance to the movement was detected by the examiner, or the subject reported pain, at which point the location of any

J. Walsh, T. Hall / Manual Therapy 14 (2009) 623–629

625

dorsiﬂexed and it was determined whether this had the effect of rendering the presenting symptoms the same, better or worse. For both tests, if the subject experienced pain in the starting position (supine lying for the SLR test, slumped sitting with cervical ﬂexion for the slump test), hip ﬂexion (SLR) or knee extension (slump) was performed to the onset of signiﬁcant resistance, or an increase in starting pain (rather than the onset of pain), after which the procedure used for subjects without pain in the starting position was followed. For each test, reproduction of presenting symptoms, which was made worse by dorsiﬂexion, was recorded as a positive ﬁnding – any other responses were recorded as negative. Subjects who were positive on both SLR and slump tests were recorded as positive for sciatic nerve mechanosensitivity, while those who were negative on one or both tests were recorded as negative for sciatic nerve mechanosensitivity. To establish the reliability of testing procedures, a third examiner (Tester 3, a qualiﬁed physiotherapist with three months clinical experience, blinded to the ﬁndings of the other examiners) immediately repeated the entire procedure in the ﬁrst 20 subjects. 2.3. Data analysis 2.3.1. Reliability Manual palpation ﬁndings of Tester 1 and SLR test and slump test ﬁndings of Tester 2 were cross-tabulated by those of Tester 3 and the Kappa (k) statistic was used to determine inter-rater reliability of manual palpation, the SLR and slump tests. To determine the inter-rater reliability of mechanical palpation, Intraclass Correlation Coefﬁcients (ICCs) were determined for each site, using a two way ANOVA (mixed effects model). Mean differences (and 95% conﬁdence intervals) in PPT measures at each site between the two raters were calculated (Rankin and Stokes, 1998). The standard error of measurement (SEM) was also calculated, using the formula (SEM ¼ S O(1 ICC)), where S is the pooled standard deviation and ICC is the reliability coefﬁcient (O’Sullivan et al., 2007). 2.3.2. Validity Subjects were categorised according to whether they were negative or positive on manual palpation. To determine validity of manual palpation, using PPT as the criterion, paired t-tests were used to compare PPTs between the symptomatic and the asymptomatic legs in subjects who were either negative or positive on manual palpation.

Fig. 2. Mechanical palpation, using an algometer at the a) sciatic, b) tibial and c) common peroneal nerves.

response was ascertained through verbal questioning. In the event that presenting symptoms were reproduced, the ankle was then passively dorsiﬂexed and it was ascertained whether this had the effect of rendering the presenting symptoms the same, better, or worse. For the slump test, the subject was placed in slumped sitting with cervical ﬂexion by the examiner and asked to maintain this position for the duration of the test. The knee was passively extended to the point of signiﬁcant resistance as detected by the examiner, or the onset of pain. The location of any response was determined through verbal questioning. In the event that presenting symptoms were reproduced, the ankle was then passively

2.3.3. Diagnostic accuracy Manual palpation ﬁndings were cross-tabulated by mechanosensitivity ﬁndings. Sensitivity, speciﬁcity, positive predictive values (PPVs) and negative predictive values (NPVs) were determined for manual palpation at the sciatic, tibial and common peroneal nerves and also for ‘one or more’, ‘two or more’ and ‘three out of three’ positive manual palpation sites. 95% conﬁdence intervals around the observed k values and around the estimates of diagnostic accuracy were determined by the general method described by Fleiss (1981) and implemented in an online web calculator (www.statpages.org). 3. Results 3.1. Participants Of 134 consecutive new patients attending the BPSC, 55 were excluded from the study for the following reasons: absence of leg pain (47), unable to understand English (3), history of spinal surgery (1), suspected serious pathology (3), unable to tolerate

626

J. Walsh, T. Hall / Manual Therapy 14 (2009) 623–629

testing position (1) (Fig. 3). Therefore, 79 patients were invited to take part, of whom 34 declined, and so 45 subjects participated. Characteristics of the 45 study patients are detailed in Table 1, all of whom underwent the index test and the reference standard. There were no inconclusive or indeterminate results and no adverse events from performing any of the tests. 3.2. Reliability 3.2.1. Manual palpation Manual palpation ﬁndings of Tester 1 and SLR test and slump test ﬁndings of Tester 2 were cross-tabulated by those of Tester 3 (Table 2). k scores and 95% conﬁdence intervals for agreement between the two raters for nerve palpation and mechanosensitivity tests are detailed in Table 3. Substantial agreement was found for all tests (Landis and Koch, 1977). 3.2.2. Mechanical palpation Mean PPTs for Tester 1 and Tester 3 at each of the three sites on both sides are presented in Table 4. ICCs ranged from 0.85 to 0.96, demonstrating excellent reliability. Mean differences between raters (ranging from 14 to 11) were all close to zero indicating excellent agreement (Rankin and Stokes, 1998), while the 95% conﬁdence interval around the mean differences contained zero in all cases, indicating a lack of bias between raters (Brennan and Silman, 1992). 3.3. Validity Twenty-seven, 20 and 24 subjects were positive on manual palpation at the sciatic, tibial and common peroneal nerves, respectively. There were no signiﬁcant differences in PPTs between sides at any of the nerves in subjects who were negative on manual palpation. In subjects who were positive on manual palpation, mean PPTs were signiﬁcantly lower on the symptomatic side compared to the asymptomatic side for each of the nerves (Fig. 4 and Table 5).

Table 1 Participant characteristics. Characteristic Gender Male Female Age Mean (SD) Range Mean (SD) duration of symptoms Mean (SD) pain intensity

Value 22 23 46 (11) years 26–70 years 5.6 (5.7) months 6.1 (2.6)/10

3.4. Diagnostic accuracy No subjects reported pain in the starting position for either test. Twenty-three subjects were positive on SLR testing while 22 were positive on slump testing. Twenty subjects were positive on both tests and so determined positive for sciatic nerve mechanosensitivity. Cross tabulations of the results of the nerve palpation tests by the results of the mechanosensitivity test are presented in Table 6. 3.4.1. Estimates Sensitivity, speciﬁcity, PPVs, and NPVs, with 95% conﬁdence intervals, for nerve palpation are detailed in Table 7.

4. Discussion Radiating leg pain is a common problem affecting up to 57% of patients with low-back pain (Selim et al., 1998). Twenty-ﬁve subjects were found to be negative on either or both the SLR and slump tests and so determined to be negative for sciatic nerve mechanosensitivity. These subjects were assumed to have some other source of low-back related leg pain, e.g. somatic referred pain, although this is only conjecture and goes beyond the scope of this study. The relatively high proportion of subjects with a positive SLR and slump test (20/45) reﬂects the importance of nerve trunk mechanosensitivity in low-back related leg pain. Classiﬁcation of patients with low-back related leg pain into various sub-categories,

Fig. 3. Procedural ﬂowchart for ‘2 or more’ positive manual palpation sites.

J. Walsh, T. Hall / Manual Therapy 14 (2009) 623–629 Table 2 Cross tabulations for manual palpation, SLR test and slump test ﬁndings of Tester 1 by those of Tester 3 (inter-rater reliability, n ¼ 20).

Manual palpation

Sciatic Tibial Common peroneal

Mechanosensitvity

SLR Slump

P N P N P N P N P N

a

Sciatic Nerve Symptomatic Leg

350

Tester 3 P

N

8 1 9 1 7 1 8 1 8 3

1 10 2 8 1 11 1 10 0 9

Asymptomatic Leg

300 250

PPT

Tester 1

627

**

200 150 100 50 0

P ¼ Positive, N ¼ Negative.

-ve (n=18)

+ve (n=27)

Manual Palpation

Manual palpation

Sciatic Tibial Peroneal

SLR Slump

k

(95% CI)

0.8 0.7 0.79 0.8 0.71

(0.39, (0.28, (0.38, (0.39, (0.33,

0.94) 0.86) 0.94) 0.94) 0.71)

b

Tibial Nerve

350

Symptomatic Leg

300

Asymptomatic Leg

250

PPT

Table 3 Inter-rater reliability of manual palpation. Kappa (k) scores for agreement for manual palpation, SLR and slump tests (n ¼ 20) with 95% conﬁdence intervals (CI).

200 150

*

100

Table 4 Inter-rater reliability and agreement for mechanical palpation (PPT) (n ¼ 20). Site

Side Mean PPT (kPa)

Diff. (95%CI)

ICC (95%CI)

11 1 4 14 3 2

0.96 0.92 0.87 0.96 0.91 0.85

SEM (kPa)

Tester 1 Tester 3 Sciatic

A S Tibial A S Common peroneal A S

310 260 170 149 167 148

299 259 174 163 170 146

(2, 34) (31, 33) (26, 18) (29, 1) (24, 18) (20, 24)

(0.91, (0.79, (0.66, (0.91, (0.78, (0.63,

0.99) 0.97) 0.95) 0.99) 0.97) 0.94)

25 35 25 16 23 25

A ¼ asymptomatic, S ¼ symptomatic, kPa ¼ kilopascals, Diff. ¼ mean difference in PPT between testers, CI ¼ conﬁdence interval, ICC ¼ intraclass correlation coefﬁcient, SEM ¼ standard error of the measurement.

50 0

-ve (n=25)

+ve (n=20)

Manual Palpation

c

350

Common Peroneal Nerve Symptomatic Leg

300

Asymptomatic Leg

250

PPT

with targeted intervention, has been suggested as a mean of improving treatment outcome (Schafer et al., 2009) and there is preliminary evidence to support this (Schafer et al., 2008). The presence of increased neural tissue mechanosensitivity identiﬁed in part by nerve palpation is one of the key factors in the classiﬁcation of low-back related leg pain (Hall and Elvey, 2005; Schafer et al., 2009). k coefﬁcients ranging from 0.61 to 0.80 represent substantial agreement between raters (Landis and Koch, 1977). In this study, k coefﬁcients for manual palpation of the sciatic and common peroneal nerves were at the top end of this range (0.80 and 0.79, respectively). Although agreement between raters of manual palpation at the tibial nerve was also substantial (k ¼ 0.70), this was slightly less than that of the sciatic or common peroneal nerves. Reliability of palpation of lower limb nerve trunks has not previously been investigated. However, Jepsen et al. (2006) studied the inter-rater reliability of manual palpation of upper limb nerves. Moderate reproducibility was determined (k ¼ 0.53). Schmid et al. (2008) also investigated the inter-rater reliability of manual palpation at a number of sites of the medial, radial and ulnar nerves. k coefﬁcients ranged from 0.36 to 0.79, while overall interrater reliability was moderate (k ¼ 0.59). Lower reliability for

200 150

**

100 50 0

-ve (n=21)

+ve (n=24)

Manual Palpation Fig. 4. Mean PPTs on symptomatic and asymptomatic sides in subjects who were either negative or positive on manual palpation at the a) sciatic, b) tibial and c) common peroneal nerves. ve ¼ negative, þve ¼ positive, * ¼ difference is signiﬁcant at the p < 0.05 level, ** ¼ signiﬁcant at the p < 0.01 level.

manual palpation of upper limb nerves may be due to greater difﬁculty in isolating the nerves in the upper limb. Perhaps nerve size may be inﬂuential. Excellent reliability and agreement was determined for PPTs at all tested sites. This is consistent with asymptomatic upper limb nerve PPTs which were also found to have excellent inter-rater reliability with ICCs ranging from 0.92 to 0.97 (Sterling et al., 2000). To our knowledge there have been no previous reports of reliability of PPT for lower limb nerve sites in symptomatic subjects. The ﬁnding that PPTs were signiﬁcantly lower on the symptomatic side in subjects who were positive on manual palpation, while there were no signiﬁcant differences between sides in subjects who were negative on manual palpation, demonstrates

628

J. Walsh, T. Hall / Manual Therapy 14 (2009) 623–629

Table 5 Validity of manual palpation. Difference in PPTs between sides in subjects who were either negative or positive on manual palpation. Site

Manual palpation

Side

Mean PPT (kPa)

Sciatic

ve

A S A S A S A S A S A S

298 292 284 208 168 159 141 115 149 164 144 116

þve Tibial

ve þve

Common peroneal

ve þve

Mean Diff. (95% CI)

p

6 (60, 49)

0.84

76 (98, 54)

<0.001

9 (22, 3)

0.126

26 (45, 7)

0.01

15 (6, 37) 28 (40, 16)

0.155 <0.001

ve ¼ negative, þve ¼ positive, A ¼ asymptomatic, S ¼ symptomatic, kPa ¼ kilopascals, Mean Diff. ¼ mean difference in PPTs between asymptomatic and symptomatic sides, CI ¼ conﬁdence interval, p ¼ p-value.

Table 6 Cross tabulations for manual palpation ﬁndings by mechanosensitivity ﬁndings. Mechanosensitivity P

N

Total

17 3 13 7 13 7 18 2 17 3 8 12 20

10 15 7 18 11 14 16 9 8 17 4 21 25

27 18 20 25 24 21 34 11 25 20 12 33 45

Palpation Sciatic Tibial Peroneal ‘1 or more’ positive ‘2 or more’ positive ‘3 out of 3’ positive Total

P N P N P N P N P N P N

P ¼ Positive, N ¼ Negative.

criterion-based validity of manual palpation at the three sites, using PPT as the criterion. Palpation of the sciatic nerve had greatest accuracy in the identiﬁcation of sciatic nerve mechanosensitivity, followed, in order, by the tibial and common peroneal nerves (Table 7). ‘One or more’ positive palpation sites had high sensitivity but low speciﬁcity. A negative ﬁnding in this case (i.e. no positive palpation sites) may provide an indication that the target condition (sciatic nerve mechanosensitivity) can be ruled out. Conversely, ‘three out of three’ positive sites had low sensitivity but high speciﬁcity. A positive ﬁnding in this case (three positive palpation sites) may provide an indication that the target condition is present. ‘Two or more’ positive palpation sites had the greatest overall diagnostic accuracy. Therefore, palpation of all three sites is recommended in the examination of low-back related leg pain. If ‘two or more’ of

these sites are positive, sciatic nerve mechanosensitivity may be present. A limitation of this study may be that, in the absence of a diagnostic gold standard for sciatic nerve mechanosensitivity, positive responses to both the SLR and slump tests were used as the reference standard. A positive ﬁnding on either the SLR or slump test alone may be a sign of sciatic nerve mechanosensitivity. The stipulation in this study that both tests had to be positive to determine the presence of sciatic nerve mechanosensitivity may have increased the accuracy of identifying this target condition. Although the SLR and slump tests may be seen as tests of ‘stretch’ mechanosensitivity, while palpation may be seen as a sign of ‘pressure’ mechanosensitivity, in animal models, both pressure and stretch mechanosensitivity developed in response to nerve trunk inﬂammation (Bove et al., 2003; Dilley et al., 2005). Therefore, the use of the SLR and slump tests as the reference standard may be justiﬁed. Coveney et al. (1997) used nerve conduction studies as the gold standard in a study to determine the validity of a provocative test of the brachial plexus, in subjects with carpal tunnel syndrome. However, studies have shown that nerves can be mechanosensitised in the absence of any axonal damage (Eliav et al., 1999, 2001). Other studies have shown that mechanosensitivity does not necessarily exist in the presence of nerve compression (Jonsson et al., 1997; Kaptan et al., 2007). Furthermore, there is controversy as to whether nerve conduction studies are useful in the identiﬁcation of neural pain syndromes, at least in the upper limb (Concannon et al., 1997; Rosenbaum, 1999). Therefore, the use of nerve conduction studies as a gold standard in studies of mechanosensitivity tests cannot be supported. Real-time ultrasound has been used in the investigation of neural tissue pain disorders in the upper limb (Dilley et al., 2001; Greening et al., 2001, 2005) and sciatic nerve movement in the lower limb (Ellis and Hing, 2008; Ellis et al., 2008). However, at this stage, the usefulness of this procedure may be restricted to measurement of nerve movement rather than mechanosensitivity. Recent advances in magnetic resonance (MR) imaging of nerves has ensured that MR neurography is capable of providing information about nerve compression, nerve inﬂammation, nerve trauma, systemic neuropathies, nerve tumours, and recovery of nerve from pathological states (Filler et al., 2004; Freund et al., 2007). However, MR neurography has yet to be used to correlate these ﬁndings with mechanosensitivity. Therefore, clinical tests such as the SLR, slump and nerve palpation tests may be the best current means of determining sciatic nerve mechanosensitivity. In this study, each nerve was only palpated at one site. Other sites have also been proposed (Butler, 2000). Furthermore, the femoral or other lower limb nerves were not tested. Therefore, palpation of the sciatic nerve tract at other sites and of other lower limb nerves should be investigated to provide a more complete picture of this examination procedure. In this study, the examiners were aware of the presenting symptoms. This may have affected the blinding of the study.

Table 7 Diagnostic accuracy of manual palpation.

Sciatic Tibial Peroneal ‘1 or more’ positive ‘2 or more’ positive ‘3 out of 3’ positive

Sens (95% CI)

Spec (95% CI)

PPV (95% CI)

NPV (95% CI)

0.85 0.65 0.65 0.90 0.83 0.40

0.60 0.72 0.56 0.36 0.73 0.84

0.63 0.65 0.54 0.53 0.76 0.67

0.83 0.72 0.60 0.82 0.80 0.64

(0.75, (0.51, (0.51, (0.81, (0.72, (0.26,

0.95) 0.79) 0.79) 0.99) 0.94) 0.54)

(0.46, 0.74) (0.59, 0.85) (0.41, 0.7) (0.22, 0.5) (0.6, 0.86) (0.73, 0.95)

(0.49, 0.77) (0.51, 0.79) (0.4, 0.69) (0.38, 0.68) (0.64, 0.88) (0.53, 0.8)

Sens ¼ sensitivity, Spec ¼ speciﬁcity, PPV ¼ positive predictive value, NPV ¼ negative predictive value, CI ¼ conﬁdence interval.

(0.72, 0.94) (0.59, 0.85) (0.46, 0.74) (0.71, 0.93) (0.68, 0.92) (0.5, 0.78)

J. Walsh, T. Hall / Manual Therapy 14 (2009) 623–629

However, when performing these tests clinically, the examiner would be aware of this information. Furthermore, it was necessary to know which side was symptomatic in order to be able to determine whether any manual palpation ﬁndings were positive or negative, but also so that the asymptomatic side could be tested ﬁrst in the case of mechanical palpation, SLR and slump tests. 5. Conclusion This study provides support for the use of nerve palpation in clinical examination, with evidence of excellent reliability and diagnostic accuracy as well as validity of manual palpation for three lower limb nerve sites. Acknowledgements The authors would like to thank Mark Kenneally, Joanne Hayes, Mary Cassells, Eimear Cassidy and the staff at the BPSC, AMNCH for facilitating the study; and Dr. Kathleen Bennett, Professor Ronan Conroy and Dr. John Pezullo for statistical support. References Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, Irwig LM, Moher D, Rennie D, de Vet HC, Lijmer JG. The STARD statement for reporting studies of diagnostic accuracy: explanation and elaboration. Clinical Chemistry 2003;49(1):7–18. 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. Brennan P, Silman A. Statistical methods for assessing observer variability in clinical measures. British Medical Journal 1992;304(6840):1491–4. Butler D. Adverse mechanical tension in the nervous system: a model for assessment and treatment. Australian Journal of Physiotherapy 1989;35:227–38. Butler DS. The sensitive nervous system. Adelaide: Noigroup; 2000 [chapter 8]176– 209. Concannon MJ, Gainor B, Petroski GF, Puckett CL. The predictive value of electrodiagnostic studies in carpal tunnel syndrome. Plastic and Reconstructive Surgery 1997;100(6):1452–8. Coveney B, Trott P, Grimmer KA. The upper limb tension test in a group of patients with a presentation of carpal tunnel syndrome. In: Proceedings: tenth biennial conference. Melbourne: Manipulation Physiotherapists Association of Australia; 1997. Dilley A, Greening J, Lynn B, Leary R, Morris V. The use of cross-correlation 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, 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. Eliav E, Herzberg U, Ruda MA, Bennett GJ. Neuropathic pain from an experimental neuritis of the rat sciatic nerve. Pain 1999;83(2):169–82. Ellis R, Hing W, Dilley A, McNair P. Reliability of measuring sciatic and tibial nerve movement with diagnostic ultrasound during a neural mobilisation technique. Ultrasound in Medicine and Biology 2008;34(8):1209–16. Ellis R, Hing WA. Ultrasound imaging of sciatic nerve movement in a neural preloaded compared to an unloaded position – quantitative assessment and reliability, IFOMT 2008. Rotterdam: Elsevier; 2008.

629

Elvey R, Hall T. Neural tissue evaluation and treatment. In: Donatelli R, editor. Physical therapy of the shoulder. 3rd ed. New York: Churchill Livingstone; 1997. p. 131–52 [chapter 5]. Field D, Hutchinson JO. Field’s anatomy palpation and surface markings. 4th ed. Edinburgh: Butterworth Heinemann; 2006. p. 220–1. Filler AG, Maravilla KR, Tsuruda JS. MR neurography and muscle MR imaging for image diagnosis of disorders affecting the peripheral nerves and musculature. Neurologic Clinics 2004;22(3):643–82. vi–vii. Fleiss JL. Statistical methods for rates and proportions. 2nd ed. New York: John Wiley & Sons; 1981 [section 5.6]. Freund W, Brinkmann A, Wagner F, Dinse A, Aschoff AJ, Stuber G, Schmitz B. MR neurography with multiplanar reconstruction of 3D MRI datasets: an anatomical study and clinical applications. Neuroradiology 2007;49(4):335–41. Greening J, Dilley A, Lynn B. In vivo study of nerve movement and mechanosensitivity of the median nerve in whiplash and non-speciﬁc arm pain patients. Pain 2005;115(3):248–53. 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 non-speciﬁc arm pain. Journal of Hand Surgery [Br] 2001;26(5):401–6. discussion 407–8. Hall T, Elvey RL. Management of mechanosensitivity of the nervous system in spinal pain syndromes. In: Boyling J, Jull G, editors. Modern manual therapy of the vertebral column. 3rd ed. Edinburgh: Churchill Livingstone; 2005. p. 413–31 [chapter 29]. Hall T, Quintner J. Responses to mechanical stimulation of the upper limb in painful cervical radiculopathy. Australian Journal of Physiotherapy 1996; 42(4):277–85. Hall T, Zusman M, Elvey R. Adverse mechanical tension in the nervous system? Analysis of straight leg raise. Manual Therapy 1998;3(3):140–6. Jepsen J, Laursen L, Hagert C-G, Kreiner S, Larsen A. Diagnostic accuracy of the neurological upper limb examination I: inter-rater reproducibility of selected ﬁndings and patterns. BMC Neurology 2006;6(1):8. Jonsson B, Annertz M, Sjoberg C, Stromqvist B. A prospective and consecutive study of surgically treated lumbar spinal stenosis. Part I: clinical features related to radiographic ﬁndings. Spine 1997;22(24):2932–7. Kaptan H, Kasimcan O, Cakiroglu K, Ilhan MN, Kilic C. Lumbar spinal stenosis in elderly patients. Annals of the New York Academy of Sciences 2007;1100:173–8. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33(1):159–74. Maitland G. Negative disc exploration: positive canal signs. Australian Journal of Physiotherapy 1979;25:129–34. Moore KL, Dalley AF. Clinically oriented anatomy. 5th ed. Philadelphia: Lippincott, Williams & Wilkins; 2006. p. 629. O’Sullivan C, Bentman S, Bennett K, Stokes M. Rehabilitative ultrasound imaging of the lower trapezius muscle: technical description and reliability. Journal of Orthopaedic and Sports Physical Therapy 2007;37(10):620–6. Rankin G, Stokes M. Reliability of assessment tools in rehabilitation: an illustration of appropriate statistical analyses. Clinical Rehabilitation 1998;12(3):187–99. Rosenbaum R. Carpal tunnel syndrome and the myth of El Dorado. Muscle Nerve 1999;22(9):1165–7. Schafer A, Hall T, Briffa K. Classiﬁcation of low back-related leg pain-A proposed patho-mechanism-based approach. Manual Therapy 2009;14(2):222–30. Schafer GM, Hall TM, Ludtke K, Muller G, Briffa NK. Classiﬁcation of low back related leg pain: differences in treatment response between diagnostic groups, IFOMT 2008. Rotterdam: Elsevier; 2008. Schmid A, Brunner F, Luomajoki H, Held U, Bachmann L, Kunzer S, Coppieters M. Reliability of clinical tests to evaluate conduction loss and increased mechanosensitivity of the peripheral nervous system, IFOMT 2008. Rotterdam: Elsevier; 2008. Selim AJ, Ren XS, Fincke G, Deyo RA, Rogers W, Miller D, Linzer M, Kazis L. The importance of radiating leg pain in assessing health outcomes among patients with low back pain. Results from the veterans health study. Spine 1998;23(4):470–4. Sterling M, Treleaven J, Edwards S, Jull G. Pressure pain thresholds of upper limb peripheral nerve trunks in asymptomatic subjects. Physiotherapy Research International 2000;5(4):220–9. Pezullo JC. Interactive statistical calculations, www.statpages.org; 2008, http:// StatPages.org/ctab2x2.html; 2008 [last accessed 16 May 2008].

Manual Therapy 14 (2009) 630–635

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

Altered lumbopelvic movement control but not generalized joint hypermobility is associated with increased injury in dancers. A prospective studyq Nathalie Anne Roussel a, c, e, *, Jo Nijs a, b, Sarah Mottram d, Annouk Van Moorsel c, Steven Truijen a, Gaetane Stassijns e a

Division of Musculoskeletal Physiotherapy, Department of Health Sciences, Artesis University College of Antwerp, Belgium Spinal Research Group, Faculty of Physical Education and Physiotherapy, Vrije Universiteit Brussel, Belgium Department of Dance, Artesis University College of Antwerp, Belgium d KC International, UK e Department of Physical Medicine and Rehabilitation, University Hospital Antwerp, University of Antwerp, Belgium b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 26 March 2008 Received in revised form 10 November 2008 Accepted 3 December 2008

Dancers experience signiﬁcant more low back pain (LBP) than non-dancers and are at increased risk of developing musculoskeletal injuries. Literature concerning the relationship between joint hypermobility and injury in dancers remains controversial. The purpose of this study was therefore to examine whether lumbopelvic movement control and/or generalized joint hypermobility would predict injuries in dancers. Four clinical tests examining the control of lumbopelvic movement during active hip movements were used in combination with joint hypermobility assessment in 32 dancers. Occurrence of musculoskeletal injuries, requiring time away from dancing, was recorded during a 6-month prospective study. Logistic regression analysis was used to predict the probability of developing lower limb and/or lumbar spine injuries. Twenty-six injuries were registered in 32 dancers. Forty-four percent of the dancers were hypermobile. A logistic regression model using two movement control tests, correctly allocated 78% of the dancers. The results suggest that the outcome of two lumbopelvic movement control tests is associated with an increased risk of developing lower extremities or lumbar spine injuries in dancers. Neither generalized joint hypermobility, evaluated with the Beigthon score, nor a history of LBP was predictive of injuries. Further study of these interactions is required. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Low back pain Movement control Dance Hypermobility

1. Introduction Dancers are at increased risk of developing Low Back Pain (LBP) (McMeeken et al., 2001), as they regularly perform repetitive extensions, high velocity twisting and bending movements. Given the exceptionally high ﬂexibility required for dance, it is not surprising that repetitive movements to extreme positions can contribute to pain. Several studies have revealed increased

q This study was ﬁnancially supported by a PhD grant (‘Motor control of the lumbopelvic region in dancers and patients with low back pain’ – G806) and by a research grant (‘Analysis of static and dynamic parameters to reduce injuries in Dancers’ – G801) supplied by the University College of Antwerp. No commercial party having a direct ﬁnancial interest in the results of the research supporting this article has or will confer a beneﬁt on the authors or on any organization in which the authors are associated. * Corresponding author. Campus HIKE, Departement G, Artesis Hogeschool Antwerpen, Van Aertselaerstraat 31, 2170 Merksem, Belgium. Tel.: þ32 3 641 82 65; fax: þ32 3 641 82 70. E-mail address: [email protected] (N.A. Roussel). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.12.004

ﬂexibility and hypermobility in dancers (Klemp et al., 1984; Gannon and Bird, 1999; McCormack et al., 2004). While hypermobile (nondancing) individuals may be asymptomatic, hypermobility is a predisposing factor of musculoskeletal pain/injury (Kirk et al., 1967; Simmonds and Keer, 2007). It has been suggested recently that evaluating the quality of movement could be more important than measuring the quantity of movement in hypermobile individuals (Simmonds and Keer, 2007). Impaired proprioception has been found in hypermobile individuals (Mallik et al., 1994; Hall et al., 1995). It has been suggested that this could lead to recurrent joint trauma and consequently musculoskeletal pain (Fitzcharles, 2000). Hence, proprioceptive and motor control training have been used in the treatment of hypermobile individuals (Russek, 2000; Ferrell et al., 2004). The literature concerning the relationship between joint hypermobility and injury in dancers remains controversial (Klemp and Learmonth, 1984; Klemp et al., 1984; McCormack et al., 2004). An extremely high prevalence of injuries has been described in dancers (Garrick and Requa, 1993). Of all professional dancers in Australia, 89% sustain injuries which affect their career and

N.A. Roussel et al. / Manual Therapy 14 (2009) 630–635

approximately 50% of professional dancers have persistent/recurrent injuries (Crookshanks, 1999; Negus et al., 2005). However, it is not clear whether this high prevalence of injuries is related to hypermobility. One hypothesis is that the high prevalence of injuries including LBP is due to repetitive movements in the hypermobile range of movement, which is typical for dancing (McCormack et al., 2004). Another hypothesis is that impaired motor control of the lumbopelvic region leads to compensatory movements of the spine and lower limbs, which results in injuries (Zazulak et al. 2007a, b). However, motor control has not been examined in dancers, despite the high prevalence of LBP in this young population. Therefore, we undertook a study to examine the relationship between motor control, hypermobility and injuries (including LBP) in dancers. Generalized joint hypermobility can be easily screened according to the Beighton modiﬁcation of the Carter and Wilkinson criteria. Hypermobility is generally deﬁned as a score higher or equal to 4/9 on the Beighton scale (Beighton et al., 1999). Less consensus exists for deﬁning motor control impairments in clinical settings. A typical feature of impaired motor control is a reduced control of active movements (Luomajoki et al., 2007). An important part of the rehabilitation process, therefore, consists of training of speciﬁc lumbopelvic stabilization, independent of any trunk, lower or upper limb movement (Richardson et al., 1992; Sahrmann, 2002). The ability to activate muscles to isometrically hold a position or prevent motion at one joint, while concurrently producing an active movement at another joint, is a movement control test (Mottram and Comerford, 2008). Several authors voice the need for a clinical assessment of active movement control in LBP-patients (Maluf et al., 2000; O’Sullivan, 2005; Luomajoki et al., 2007), but information regarding the clinimetric properties of simple clinical movement control tests is lacking. 1.1. Study aims The purpose of this study was to examine whether altered lumbopelvic movement control and/or generalized joint hypermobility would predict musculoskeletal injuries to the spine and lower extremities in dancers. In addition, the inter-observer reliability and internal consistency of the four clinical tests examining lumbopelvic movement control were evaluated in patients with chronic LBP and healthy subjects. 2. Methods 2.1. Subjects and research design All students following a full-time professional Dance Program in Belgium (n ¼ 32) were recruited for the prospective part of the study. Twenty-six female (81%) and 6 male (19%) students, aged 20 2 years (range[17–25]) participated in the study. Baseline assessment included medical history, examination of lumbopelvic movement control and generalized joint hypermobility. Movement control and hypermobility were examined by an assessor blinded to the medical history of the dancer. The occurrence of injuries of the dancers was recorded every 2 weeks during a 6-month follow-up period, by assessors blinded to the outcome of the baseline assessment. Injuries were deﬁned as any musculoskeletal condition requiring time away from dancing and were registered using a standardized questionnaire and subjective evaluation. Prior to participation, all subjects received verbal and written information addressing the nature of the study. Demographic information was recorded by the time of testing. The Human Research Ethics Committee of the University Hospital approved the

631

study and written informed consent was obtained from all participants prior to testing. 2.2. Instrumentation The Pressure Biofeedback (PBU) has been developed to monitor lumbopelvic movement by recording pressure changes during assessment and exercise (Richardson et al., 1992; Jull et al., 1993). Calibration studies demonstrated that pressure recordings resulted from lumbopelvic movement and positional changes and were independent of the individual body weight (Jull et al., 1993). The PBU is sensitive to small movements associated with deep muscle recruitment within 2 mm Hg of pressure change (Falla et al., 2003). A high level of agreement has been found between the results of the prone abdominal drawing-in test, recorded with the PBU and converted to categories and a delayed contraction of transversus abdominis (Hodges et al., 1996). Furthermore, a blinded observer was able to detect the presence or absence of LBP with the use of the prone abdominal drawing-in test, recorded with the PBU (Cairns et al., 2000). The visual analogue scale (VAS – 100 mm) was used for the assessment of lumbar pain severity. The VAS score is believed to be reliable, valid, and sensitive to change (Jensen et al., 1986; Ogon et al., 1996). An international long-arms goniometer1 was used for the evaluation of elbow and knee joint angles (assessment of generalized joint hypermobility). A standardized questionnaire was used to collect demographic information at baseline, and an injury registration form was used for the assessment of musculoskeletal symptoms and injuries (Cumps et al., 2007). With this injury registration form information about the symptoms and injury occurrence, the time loss and the medical diagnosis were gathered. This injury registration form has already been used in prospective epidemiology research in sportsmen (Cumps et al., 2007). 2.3. Procedure Generalized joint hypermobility was assessed according to the description provided by Beighton et al. (1999). The clinimetric properties of the Beighton score have been summarized elsewhere (Nijs, 2005). Three subgroups were deﬁned based on the individual Beighton scores: tight (0–3); hypermobile (4–6); extremely hypermobile (7–9) (Stewart and Burden, 2004). Lumbopelvic movement control was assessed by evaluating the subjects’ ability to control movement of lumbopelvic region while performing simple movements in the hips. Four commonly used clinical tests, i.e. the Active Straight Leg Raising (ASLR), Bent Knee Fall Out (BKFO), Knee Lift Abdominal Test (KLAT) and Standing Bow (SB), were used in the present study for the evaluation of lumbopelvic movement control. ASLR, BKFO and KLAT were performed in supine position and monitored with a PBU. The pressure was inﬂated to 40 mm Hg (baseline pressure) (Richardson et al., 1992). Prior to the test, the subjects performed two inspirations and expirations. The pressure was then readjusted to 40 mm Hg. The participants were instructed to maintain neutral spine position (i.e. preventing spinal movement) during lower extremity movement. The other leg was extended (ASLR, BKFO) or ﬂexed (KLAT), and rested on the table. A pre-testing trial was organized to familiarize the subjects with the PBU and the clinical tests. Maximal pressure deviation from baseline was recorded during each test and these scores were used for further analyses.

1

Gymna. Pasweg 6 C, 3740 Bilzen, Belgie¨.

632

N.A. Roussel et al. / Manual Therapy 14 (2009) 630–635

The aim of the test was to have as little deviation from the baseline pressure as possible. Although it is unnatural to keep the spine rigid during movements, previous research demonstrated that healthy subjects with good trunk stabilization are able to maintain neutral spine position while moving the legs (Jull et al., 1993). Jull et al. (1993) developed a method to measure active control of the trunk by loading in the sagittal plane (unilateral leg load). These unilateral leg movements were considered to be low load. In this study, subjects who automatically pre-activated the transversus abdominis and internal oblique controlled the lumbopelvic position during the application of low load, with low pressure changes as result. Jull et al. (1993) therefore suggested that ‘excessive’ pressure changes during low loaded exercise reﬂect an inability to maintain isometric contraction of the abdominal muscles, resulting in uncontrolled movement of the lumbar spine. They did, however, not deﬁned these ‘excessive’ pressure changes. ASLR was performed in the supine position, according to the procedure as described in Roussel et al. (2007). During the ﬁrst phase, the participant lifted the extended leg 20 cm above the table (phase 1). Next, the subject held this position for 20 s (phase 2). The PBU was placed horizontally under the spine of the participant, with the lower edge at the level of the posterior superior iliac spines. For BKFO (see Fig. 1), the subject was positioned supine in partial crook lying position, as described by Comerford and Mottram (2001). The participant then slowly lowered out the bent leg to approximately 45 of abduction/lateral rotation, while keeping the foot supported beside the straight leg, and then returned to the starting position. It has been suggested that abdominal muscles should activate to stabilize the trunk in coordination with the adductors, which eccentrically lower the leg (Comerford and Mottram, 2001). For BKFO, the PBU was positioned vertically under the lumbar spine on the side of the bent leg, with the lower edge 2 cm caudal of the posterior superior iliac spine. A folded towel was placed under the lumbar spine on the side of the extended leg, so as to keep both sides of the lumbar spine at the same height. Pressure changes were recorded during the outward movement only. In case of left rotation, the pressure of the left PBU increases (i.e. rotation towards that side), while the pressure in the right PBU decreases. KLAT (see Fig. 2) was based on the abdominal exercises described by Sahrmann (2002) and on the isometric stability test performed by Wohlfahrt et al. (1993). The subjects were positioned in crook lying and were asked to lift one foot off the table to 90 of hip ﬂexion with knee ﬂexion, keeping the lumbar spine stable. Differences in temporal patterns of activation between healthy subjects and LBP-patients have been found during this test, indicating a lack of synergistic co-activitation in LBP-patients (Hubley-

Fig. 1. Bent Knee Fall Out. The subject is instructed to lower out the bent leg to approximately 45 of abduction/lateral rotation, while keeping the foot supported beside the straight leg, and then to return to the starting position.

Fig. 2. Knee Lift Abdominal Test. The subject is instructed to lift one foot off the table to 90 of hip ﬂexion with knee ﬂexion, keeping the lumbar spine stable.

Kozey and Vezina, 2002). The PBU was placed horizontally under the spine of the participant, with the lower edge at the level of the posterior superior iliac spines. An increase in pressure during the test indicates lumbar ﬂexion or posterior pelvic tilt, while a pressure decrease suggests lumbar extension or anterior pelvic tilt (Richardson et al., 1992). The SB (see Fig. 3) was executed in standing. First, the examiner positioned the spine in a neutral position, i.e. in the midrange between anterior and posterior pelvic tilt. Next, the participant was instructed to keep the spine in a neutral position while moving forwards in the hips till approximately 50 , without ﬂexing/ extending the lumbar spine (Luomajoki et al., 2007). The test was scored with visual inspection and palpation (Comerford et al., 2007). Bending forwards from the hips, keeping the spine in neutral position, certainly does not reﬂect normal movement and therefore challenges lumbopelvic movement control. For a correct performance of SB (score 0), stabilizing muscles of the trunk activate isometrically to keep the spine in neutral position, while the hip and back muscles contract to bend forwards. A lack of ability to actively control or prevent a compensatory movement when required or instructed to do so is considered to be uncontrolled motion (Mottram and Comerford, 2008). As the subjects were unfamiliar with the test, only clear movement dysfunction (ﬂexion or extension during the test) was rated as a loss of lumbopelvic stability (score 1). If the movement control improved by instruction and correction, it was considered that it did not infer a relevant movement dysfunction (Luomajoki et al., 2007), and a score of 0 was given. 2.4. Analysis of the inter-observer reliability and internal consistency The inter-observer reliability and internal consistency of these 4 lumbopelvic movement control tests were assessed prior to the study. Two assessors examined the ASLR, BKFO, KLAT and SB in 27 patients with chronic (>3 months) non-speciﬁc LBP, diagnosed by a physician, and 25 healthy subjects. The ﬁrst test session was followed by 10-min rest. The subject was then examined by the second investigator. In three 1-h training sessions prior to data collection, the examiners were trained in performing the tests under supervision of two manual therapists. Intra Class Correlation-Coefﬁcients (based on the pressure recordings during ASLR, BKFO and KLAT) and k-coefﬁcients (for SB) were used for the reliability analysis. Moderate to high interobserver reliability was found for ASLR and BKFO with ICC-coefﬁcients varying between 0.61 and 0.91, except for the left second phase of the ASLR in healthy people (ICC ¼ 0.41). High reliability was found for KLAT (ICC > 0.85). The weighted kappa for the interobserver reliability of the SB in healthy subjects and in patients was

N.A. Roussel et al. / Manual Therapy 14 (2009) 630–635

633

Table 1 Localization of the injuries in dancers. Number of injuries (n ¼ 26) Hip Knee Muscles lower legs Ankle & Foot Spine Upper extremities

Fig. 3. Standing Bow. Reprinted with kind permission of Ó PhysioTools Ltd.2

0.80 and 0.78 respectively (p < 0.001). The Cronbach a-coefﬁcient for internal consistency for ASLR, BKFO and KLAT was 0.83 for the patients and 0.65 for the healthy subjects (p < 0.01).

3 4 4 8 5 2

(12%) (15%) (15%) (31%) (19%) (8%)

months prospective study (see Table 1). Table 2 display the results of the movement control tests ASLR, BKFO and KLAT. Four dancers (13%) could not maintain the neutral position of the lumbar spine during SB. The mean Beighton score for generalized joint hypermobility was 4.0 2.3 (range:[0–9]). Fourteen of the 32 dancers (44%) scored above the 4/9 criterion for hypermobility (see Fig. 4). Eight of these 14 dancers (25%) presented a score ranging from 4 to 6, and six dancers (19%) were excessively hypermobile (score 7–9 according to the classiﬁcation of Stewart and Burden (2004)). The movement control test battery was used in combination with the assessment of generalized joint hypermobility in order to analyze the predictive value of these tests, i.e. the prediction of the probability of developing injuries to the lower extremities or the lumbar spine. The dancers were divided into 2 groups based on the results of the prospective study (i.e. developing musculoskeletal injuries during the 6-months follow-up or not). A logistic regression model using KLAT and SB, correctly allocated 78% of the dancers in one of the 2 groups (p < 0.05). Data of the regression analysis is presented in Table 3. Generalized joint hypermobility did not correlate with the motor control tests (rho ranging between 0.03 and 0.33), and was neither associated with the development of musculoskeletal injuries (rho ¼ 0.03, p ¼ 0.89), nor with a history of LBP (rho ¼ 0.03, p ¼ 0.89). Dancers with a history of LBP did not develop more injuries than dancers without a history of LBP (p ¼ 0.93, t ¼ 0.90).

2.5. Statistical analysis All data were analyzed using SPSS 12.0Ó for Windows.3 A 1sample Kolmogorov–Smirnov goodness-of-ﬁt test was used to examine whether the variables were normally distributed. All variables were found to be normally distributed (p > 0.05). A stepwise conditional logistic regression analysis, an independent t-test and Spearman correlation-coefﬁcients were used, in addition to descriptive statistics. The development of injuries to the lower limbs and lumbar spine during the prospective part of the study was considered as the dependant variable. The signiﬁcance level was set at 0.05, except for the correlation analysis, where the signiﬁcance level was set at 0.01 to help protect against potential type-I errors. A power analysis (using SigmaStat4) determined that 25 subjects per group were necessary for the reliability analysis to establish statistical signiﬁcance at a power of 0.90. This power analysis was based on a presumed correlation of 0.60 between the observers. 3. Results At baseline assessment, 63% of the dancers reported a history of LBP. Twenty-six injuries were registered in 32 dancers during the 6-

2

PhysioTools UK, 8 Culverwell Cottages, Pilton BA4 4DG, United Kingdom. 3 SPSS Inc. Headquarters, 233s. Wacker Drive, 11th ﬂoor, Chicago, Illinois 60606, USA. 4 Systat Software, Inc. 1735, Technology Drive, Ste 430, San Jose, CA 95110, USA.

4. Discussion Results regarding the relationship between joint hypermobility and injury in dancers remain controversial (Klemp and Learmonth, 1984; McCormack et al., 2004). While hypermobile individuals may be asymptomatic, hypermobility may be a predisposing factor of musculoskeletal pain/injury (Kirk et al., 1967; Simmonds and Keer, 2007). However, it has been suggested recently that an evaluation of the quality of movement could be more important than measuring the quantity of movement in hypermobile individuals (Simmonds and Keer, 2007). For this reason, both movement control and generalized joint hypermobility were assessed in professional dancers in the present study. Our results show that two movement control tests are able to predict injuries in dancers. In contrast, generalized joint hypermobility is not associated with a higher prevalence of musculoskeletal injuries. Table 2 Mean (X) and Standard Deviations (SD) are given for Active Straight Leg Raise (ASLR), Bent Knee Fall Out (BKFO) and Knee Lift Abdominal Test (KLAT) in 32 dancers. Left side

ASLR Phase 1 (mm Hg) ASLR Phase 2 (mm Hg) BKFO (mm Hg) KLAT (mm Hg)

Left side

Right side

Right side

Mean (SD)

Range

Mean (SD)

Range

45.0 45.9 46.1 48.0

[34–52] [38–52] [40–54] [44–60]

43.9 45.1 45.3 47.2

[30–52] [38–54] [40–50] [44–54]

(3.93) (3.63) (2.73) (4.05)

(4.72) (3.56) (2.31) (2.68)

634

N.A. Roussel et al. / Manual Therapy 14 (2009) 630–635

Beigthon score 10

Count

8

6

4

2

0

0

1

2

3

4

5

6

7

8

9

B-Tot Fig. 4. : Spread of hypermobility ratings (Beigthon Total Score) in 32 dancers. B-Tot ¼ Beigthon Total Score (0–9); Count ¼ number of dancers with a particular Beigthon score.

Sixty-three percent of the dancers reported a history of low back pain. This is in line with the study by McMeeken et al. (2001), who found an incidence rate of 49% in female and 59% in male dancers. Twenty-six injuries occurred during the 6-months follow-up period: seventy-three percent to the lower extremities and nearly 20% to the spine. These results are in accordance with the literature about dance injuries (Garrick and Requa, 1993; Byhring and Bo, 2002). Dancers are therefore at increased risk for developing musculoskeletal complaints to the spine and lower extremities. The ability to stabilize the lumbopelvic region during limb movement has been studied in elite gymnasts (Mulhearn and George, 1999), but not in dancers. Mulhearn and George (1999) found an association between impaired postural muscle endurance, a lordotic posture and LBP. As they reported a cross-sectional study, no conclusions could be drawn about the causal relationship. In the present study, four simple tests were used to evaluate lumbopelvic movement control in a clinical setting. To our knowledge, ASLR, BKFO and KLAT have not been evaluated before with the PBU. Therefore the reliability and internal consistency were evaluated in LBP-patients and healthy subjects prior to the prospective study in dancers. Moderate to high ICC-values have been found for ASLR and BKFO, high ICC-coefﬁcients were recorded for KLAT, and excellent inter-observer reliability was found for SB. The results of the reliability study therefore suggest that these tests can be used with acceptable reliability in clinical practice. The Cronbach a-coefﬁcient for internal consistency for ASLR, BKFO and KLAT was high, suggesting that all these tests assess the same underlying dimension, i.e. impaired movement control. Interestingly, different compensatory strategies were observed during testing. While a posterior pelvic tilt was observed during the movement control tests in some subjects, anterior pelvic tilt was Table 3 Logistic Regression analysis in 32 dancers. KLAT ¼ Knee Lift Abdominal Test, Exp(B) ¼ Exponent of B-coefﬁcient, CI ¼ Conﬁdence Interval. B-coefﬁcient

KLAT right Standing bow

0.531 2.173

Wald Z score

Exp(B)

5.920 4.740

0.588 8.782

p-Value

0.015 0.029

95%CI for Exp(B) Lower

Upper

0.383 1.242

0.902 62.086

also seen in others. Posterior pelvic tilt leads to a pressure increase, whereas anterior pelvic tilt decreases the pressure (Richardson et al., 1992). Unfortunately, only the maximal pressure excursion was registered and not the variation in pressure within one test performance. Anterior pelvic tilt and variation in pressure were indeed observed during clinical evaluation in dancers, but were not registered in the present study. A (hyper)lordotic posture is common in dancers and gymnasts and has been associated with an increased injury risk in female gymnasts (Steele and White, 1986). The increase in hyperlordosis during dancing could be the result of a deﬁcit in abdominal control to counteract anterior pelvic tilt during hip extension. Further study is nevertheless required to verify this assumption. The hypermobility scores found in the dancers are comparable to other studies performed in dancers (Gannon and Bird, 1999). However, generalized joint hypermobility did not correlate with the motor control tests and was not associated with a history of LBP. Moreover, neither joint hypermobility or a history of low back pain, but instead the outcome of two lumbopelvic movement control tests at baseline measurement were able to predict the probability of developing injuries to the lower extremities or lumbar spine in the present study. These results can be explained by an optimal neuromuscular control which compensates a decrease of the passive stability system, due to joint hypermobility (Panjabi, 1992; Reeves et al., 2007). Altered lumbopelvic movement control may force the dancer to compensate in the lower limbs, leading to musculoskeletal injuries. A research of the existing literature revealed only one study examining the predictive value of a lumbopelvic motor control evaluation. Zazulak et al. (2007a, b) demonstrated that impaired core stability predicted the risk of knee injuries with high sensitivity and moderate speciﬁcity in female athletes. However, injuries to the lumbar spine, hips or ankles were not examined in their study. There is no other data available regarding the predictive value of lumbopelvic motor control tests. These preliminary results are exciting, and could have an important clinical consequence. Indeed, it is possible to improve lumbopelvic motor control during physiotherapy sessions. Our data suggest that especially dancers with a positive SB and low pressure increase during KLAT are at risk to develop injuries to the lower limbs. An uncontrolled anterior pelvic tilt may account for this negative relationship in the regression analysis between the pressure results during KLAT and the increased risk for developing injuries. As the pelvic movement was not directly measured in the present study, further study of these interactions is required. 4.1. Study limitations Firstly, only the maximal pressure deviation was monitored during the tests. Some subjects ﬁrst increased the pressure and afterwards decreased the pressure during the movement control tests. These variations in pressure were not registered. Secondly, the physical activity levels were not taken into account. All the students were equally active during the day (as they all have the same dance programme) but the amount of physical activity outside the dance lessons was not evaluated. This could have inﬂuenced the predictive analysis. Further research is therefore necessary. Finally, a standardized questionnaire was used for the registration of the injuries in combination with a subjective evaluation. There is no data available regarding the reliability and the validity of this questionnaire. 5. Conclusion The Knee Lift Abdominal Test and the Standing Bow can be used for the assessment of lumbopelvic movement control. Since these

N.A. Roussel et al. / Manual Therapy 14 (2009) 630–635

tests correctly allocated 78% of the dancers, they may be useful for the identiﬁcation of dancers at risk for developing musculoskeletal injuries to the lower extremities and lumbar spine. In contrast, neither generalized joint hypermobility, evaluated with the Beigthon score, or a history of LBP is predictive of musculoskeletal injuries. Further research regarding the interaction between altered movement control and the prevention of injuries in dancers is required. Acknowledgement The authors would like to thank all the participants, physicians and physical therapists for their kind cooperation. Special thanks to Hilke Heselmans, Lander Gils, Joke Moonen, Ines Smits and Davina Caproni for aiding in data collection, to Cathy Devel for editing the manuscript, and to Mark Comerford and Raymond Lee for critical review of the manuscript. References Beighton P, Grahame R, Bird H. Hypermobility of joints. London: Springer Verlag; 1999. Byhring S, Bo K. Musculoskeletal injuries in the Norweigan National Ballet: a prospective cohort study. Scandinavian Journal of Medicine & Science in Sports 2002;12(6):365–70. Cairns MC, Harrison K, Wright C. Pressure biofeedback: a useful tool in the quantiﬁcation of abdominal muscular dysfunction? Physiotherapy 2000;86(3):127–38. Comerford MJ, Mottram SL. Functional stability re-training: principles and strategies for managing mechanical dysfunction. Manual Therapy 2001;6(1):3–14. Comerford MJ, Mottram SL, Gibbons SGT. Diagnosis of mechanical back pain subgroups & stability retraining of the lumbar spine. UK: Kinetic Control, http:// www.kineticcontrol.com/clinicalNotes.asp; 2007. Crookshanks D. Safe dance report III: the occurrence of injury in the australian professional dance population. Caberra, Australia: Australia Dance Council; 1999. Cumps E, Verhagen E, Meeusen R. Prospective epidemiological study of basketball injuries during one competitive season: ankle sprains and overuse knee injuries. Journal of Sports Science and Medicine 2007;6:204–11. Falla DL, Campbell CD, Fagan AE, et al. Relationship between cranio-cervical ﬂexion range of motion and pressure change during the cranio-cervical ﬂexion test. Manual Therapy 2003;8(2):92–6. Ferrell WR, Tennant N, Sturrock RD, Ashton L, Creed G, Brydson G, et al. Amelioration of symptoms by enhancement of proprioception in patients with joint hypermobility syndrome. Arthritis & Rheumatism 2004;50(10):3323–8. Fitzcharles MA. Is hypermobility a factor in ﬁbromyalgia? Journal of Rheumatology 2000;27:1587–9. Gannon LM, Bird HA. The quantiﬁcation of joint laxity in dancers and gymnasts. Journal of Sports Sciences 1999;17:743–50. Garrick J, Requa R. Ballet injuries. An analysis of epidemiology and ﬁnancial outcome. The American Journal of Sports Medicine 1993;21(4):586–90. Hall MG, Ferrell WR, Sturrock RD, Hamblen DL, Baxendale RH. The effect of the hypermobility syndrome on knee proprioception. British Journal of Rheumatology 1995;34(2):121–5. Hodges P, Richardson C, Jull G. Evaluation of the relationship between laboratory and clinical tests of transversus abdominis function. Physiotherapy Research International 1996;1(1):30–40. Hubley-Kozey CL, Vezina MJ. Differentiating temporal electromyographic waveforms between those with chronic low back pain and healthy controls. Clinical Biomechanics 2002;17(9–10):621–9. Jensen MP, Karoly P, Braver S. The measurement of clinical pain intensity: a comparison of six methods. Pain 1986;27(1):117–26. Jull G, Richardson CA, Toppenberg R, Comerford M, Bui B. Towards a measurement of active muscle control for lumbar stabilisation. Australian Journal of Physiotherapy 1993;39:187–93.

635

Kirk JA, Ansell BM, Bywaters EG. The hypermobility syndrome. Musculoskeletal complaints associated with generalized joint hypermobility. Annals of the Rheumatic Diseases 1967;26:419–25. Klemp P, Learmonth ID. Hypermobility and injuries in a professional ballet company. British Journal of Sports Medicine 1984;18:143–8. Klemp P, Stevens JE, Isaacs S. A hypermobility study in ballet dancers. Journal of Rheumatology 1984;11:692–6. Luomajoki H, Kool J, de Bruin ED, Airaksinen O. Reliability of movement control tests in the lumbar spine. BMC Musculoskeletal Disorders 2007;8:90. Maluf KS, Sahrmann SA, van Dillen LR. Use of a classiﬁcation system to guide nonsurgical management of a patient with chronic low back pain. Physical Therapy 2000;80(11):1097–111. Mallik AK, Ferrell WR, McDonald AG, Sturrock RD. Impaired proprioceptive acuity at the proximal interphalangeal joint in patients with the hypermobility syndrome. British Journal of Rheumatology 1994;33(7):631–7. McMeeken JM, Tully E, Stillman B, Natrass C, Bygott I, Story I. The experience of back pain in young Australians. Manual Therapy 2001;6(4):213–20. McCormack M, Briggs J, Hakim A, Grahame R. Joint laxity and the benign joint hypermobility syndrome in student and professional ballet dancers. The Journal of Rheumatology 2004;31(1):173–8. Mottram S, Comerford M. A new perspective on risk assessment. Physical Therapy in Sport 2008;9:40–51. Mulhearn S, George K. Abdominal muscle endurance and its association with posture and low back pain. Physiotherapy 1999;85(5):210–6. Negus V, Hopper D, Briffa N. Associations between turnout and lower extremity injuries in classical ballet dancers. Journal of Orthopedic & Sports Physical Therapies 2005;35(5):307–18. Nijs J. Generalized joint hypermobility: an issue in ﬁbromyalgia and chronic fatigue syndrome? Journal of Bodywork and Movement Therapies 2005;9:310–7. O’Sullivan P. Diagnosis and classiﬁcation of chronic low back pain disorders: maladaptive movement and motor control impairments as underlying mechanism. Manual Therapy 2005;10(4):242–55. Ogon M, Krismer M, So¨llner W, Kantner-Rumplmair W, Lampe A. Chronic low back pain measurement with visual analogue scales in different settings. Pain 1996;64(3):425–8. Panjabi MM. The stabilizing system of the spine. Part I. Function, dysfunction, adaptation, and enhancement. Journal of Spinal Disorders 1992;5:383–9. PhysioTools. Lumbar spine dynamic stability. Mark Comerford [Computer software module]. Reeves NP, Narendra KS, Cholewicki J. Spine stability: the six blind men and the elephant. Clinical Biomechanics 2007;22(3):266–74. Richardson CA, Jull GA, Toppenberg R, Comerford MJ. Techniques for active lumbar stabilisation for spinal protection: a pilot study. Australian Journal of Physiotherapy 1992;38:105–12. Roussel NA, Nijs J, Truijen S, Smeuninx L, Stassijns G. Low back pain. Clinimetric properties of the Trendelenburg test, the Active Straight Leg Raise test and the breathing pattern during the Active Straight Leg Raise test. Journal of Manipulative and Physiological Therapeutics 2007;30(4):270–8. Russek LN. Examination and treatment of a patient with hypermobility syndrome. Physical Therapy 2000;80:386–98. Sahrmann SA. Diagnosis and treatment of movement impairment syndromes. St Louis: Mosby; 2002. Simmonds JV, Keer RJ. Hypermobility and the hypermobility syndrome. Manual Therapy 2007;12:298–309. Steele VA, White JA. Injury prediction in female gymnasts. British Journal Sports Medicine 1986;20:31–3. Stewart DR, Burden SB. Does generalised ligamentous laxity increase seasonal incidence of injuries in male ﬁrst division club rugby players? British Journal of Sports Medicine 2004;38:457–60. Wohlfahrt D, Jull GA, Richardson CA. The relationship between the dynamic and static function of abdominal muscles. Australian Journal of Physiotherapy 1993;39:9–13. Zazulak BT, Hewett TE, Reeves NP, Goldberg B, Cholewicki J. Deﬁcits in neuromuscular control of the trunk predict knee injury risk: a prospective biomechanical-epidemiologic study. The American Journal Of Sports Medicine 2007a;35(7):1123–30. Zazulak BT, Hewett TE, Reeves NP, Goldberg B, Cholewicki J. The effects of core proprioception on knee injury: a prospective biomechanical-epidemiological study. The American Journal Of Sports Medicine 2007b;35(3):368–73.

Manual Therapy 14 (2009) 636–641

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

Cervical musculoskeletal impairment is common in elders with headache Sureeporn Uthaikhup a, *, Michele Sterling a, b, Gwendolen Jull a a

Division of Physiotherapy and National Health and Medical Research Council, Centre for Clinical Research Excellence in Spinal Pain, Injury and Health (CCRE Spine), School of Health and Rehabilitation Sciences, The University of Queensland, Brisbane, Queensland 4072, Australia b Centre for National Research on Disability and Rehabilitation Medicine (CONROD), The University of Queensland, Queensland 4006, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 11 September 2008 Received in revised form 2 December 2008 Accepted 19 December 2008

There is an opinion that with increasing cervical degenerative joint disease with ageing, cervicogenic headaches become more frequent. This study aimed to determine if cervical musculoskeletal dysfunction was speciﬁc to headache classiﬁable as cervicogenic or was more generic to headache in elders. Subjects (n ¼ 118), aged 60–75 years with recurrent headache and 44 controls were recruited. Neck function measures included range of motion (ROM), cervical joint dysfunction, cranio-cervical ﬂexor muscle function, joint position sense (JPS) and cervical muscle strength. A questionnaire documented the characteristics of headaches for classiﬁcation. A cluster analysis based on three musculoskeletal variables aligned previously with cervicogenic headache, divided headache subjects into two groups; cluster 1 (n ¼ 57), cluster 2 (n ¼ 50). Dysfunctions were greater in cluster 1 than in 2 for extension range and C1–2 joint dysfunction (p < 0.05). Most cervicogenic headaches were grouped in cluster 1, but musculoskeletal dysfunction was also found in headaches classiﬁable as migraine or tension-type headache. Neck dysfunction is not uniquely conﬁned to cervicogenic headache in elders. Further research such as headache responsiveness to management of the neck disorder is required to better understand about the neck’s causative or contributing role to elders’ headache. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Elderly Headache Cervical spine Musculoskeletal dysfunction

1. Introduction Migraine and tension-type headaches are classiﬁed as primary headaches while cervicogenic headache is a secondary headache to a primary cervical musculoskeletal disorder (International Headache Society, 2004). A number of studies have documented cervical musculoskeletal dysfunction (CMD) in cervicogenic headache (Hall and Robinson, 2004; Zito et al., 2005). Not surprisingly, while CMD is the predominant feature of cervicogenic headache, it has been found to be generally absent in the primary migraine and tensiontype headaches (Zwart, 1997; Amiri et al., 2007; Jull et al., 2007). Headache continues to be a common complaint in the elderly, although the incidence declines with advancing age (Lyngberg et al., 2005). Classical features of migraine tend to change with age (Haan et al., 2007). Throbbing, nausea and vomiting are less common and neck pain triggers more frequent. It is possible that the incidence of secondary headache increases in the elderly. It is the opinion of some clinical researchers (Pearce, 1995; Biondi and Saper, 2000) that with increasing cervical degenerative joint disease with ageing, cervicogenic headache is common in the

* Corresponding author. Tel.: þ61 7 3365 2275; fax: þ61 7 3365 1622. E-mail address: [email protected] (S. Uthaikhup). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.12.008

otherwise healthy over 50 population. Little is known about the role of CMD in headache in elders although recent studies indicate that neck pain is a frequent feature of primary headaches in this age group (Kelman, 2006; Martins et al., 2006). Nevertheless age alone affects musculoskeletal function for example, both cervical range of motion (ROM) and muscle strength decrease with age (Dvorak et al., 1992; Jorden et al., 1999). Thus CMD may merely reﬂect agerelated changes rather than indicating a cervicogenic origin of headache. The aim of this study was to measure several aspects of cervical musculoskeletal function in elders with and without headache, to investigate whether CMD was more speciﬁc to headaches classiﬁable as cervicogenic or if it was also present in other headache types as a generic feature and an age-related factor. The measures included range of cervical movement, muscle function, JPS and manual examination of cervical joint dysfunction, features which have been found previously to be impaired in cervicogenic headache and cervical disorders (Zwart, 1997; Zito et al., 2005). The combination of reduced range of movement, symptomatic cervical joint dysfunction and impaired muscle function in the craniocervical ﬂexion test has been shown to have high sensitivity and speciﬁcity for cervicogenic headache and differentiates it from primary headaches, such as migraine and tension-type and control subjects (Amiri et al., 2007; Jull et al., 2007). Thus in this study,

S. Uthaikhup et al. / Manual Therapy 14 (2009) 636–641

participants were sub-grouped on the basis of this combination of signs of CMD and the relationship between this grouping and classiﬁcation of headache type was investigated. Muscle strength and JPS were added as secondary measures for a more complete musculoskeletal assessment. It was hypothesized that CMD would be greater in cervicogenic headache when compared to other headache and non-headache groups of elders.

2. Materials and methods 2.1. Subjects Volunteer healthy elders (age range 60–75 years) were sought for the study from the general community through a University Centre of Ageing and advertising in the local press. Headache group inclusion criteria were headaches at least once per month for the past year. A structured questionnaire was used to gain a medical history from all volunteers. Volunteers were excluded if headaches had been diagnosed as secondary to neurological or systemic disorders. Control subjects were eligible if they did not suffer from neck pain and were either headache free, or had no more than occasional mild headache (<5/year) for which they had not sought medical treatment. Volunteers were excluded if medically unﬁt for age (cardiopulmonary disorders, inﬂammatory arthritis, history of bone fragility fracture, cognitive disturbance). The study was approved by the Institutional Medical Research Ethics Committee and all subjects provided written informed consent. 2.2. Questionnaires Demographic data and the presence of co-existing musculoskeletal pain were collected from all subjects. Headache subjects completed a headache questionnaire for later headache classiﬁcation and the Neck Disability Index (NDI) (Vernon and Mior, 1991) (see Appendix online). The headache questionnaire included the characteristics for migraine, tension-type headache (International Headache Society, 2004) and cervicogenic headache (Sjaastad et al., 1998) as well as questions on headache frequency and intensity. Medication use was also recorded.

2.3. Tests of the cervical musculoskeletal function 2.3.1. Cervical ROM Range of movement was measured with a 3-space Fastrak system (Polhenus, USA) and calculated with a customised software program (MATlab (7.0), Mathworks Inc, USA). Movement was measured in the directions of ﬂexion, extension, lateral ﬂexion and axial rotation using methodology as described by Dall’Alba et al. (2001). Three trials were performed in each movement plane and the mean of the primary plane values was used in the analysis.

637

2.3.3. Cervical JPS JPS was measured according to the method described by Revel et al. (1991). The Fastrak system was used to measure the subjects’ ability to relocate the natural starting head position following active movements from rotation to the left and right and extension. Absolute error was calculated with a customised software program (Matlab (7.0), Mathworks Inc, USA). Three repetitions of each movement direction were undertaken and the mean value of the error was used for the analysis. 2.3.4. Cranio-cervical ﬂexion muscle test (CCFT) The CCFT was performed in ﬁve incremental stages of increasing range in the supine position as described by Jull et al. (2004). An inﬂatable air-ﬁlled pressure sensor (Stabilizer, Chattanooga Group Inc, USA), placed behind the neck and inﬂated to a baseline pressure of 20 mmHg, guided the subject through the ﬁve stages (2 mmHg increments) of the test. Subjects were familiarized with the test prior to formal testing. Pairs of Ag/AgCl surface electrodes (11 mm-disc, 3 mm-diameter) (Grass Telefactor, Astro-Med Inc, USA) were positioned over the lower one-third of the sternocleidomatoid (SCM) bellies bilaterally (Falla et al., 2002). The ground electrode (universal electrosurgical pad-3 M HealthCare, USA) was placed on the upper part of the thoracic spine. Presumedly, electromyographic (EMG) signals were ampliﬁed (gain, 1 mV), band pass ﬁltered between 20 Hz and 1 kHz and sampled at 2 kHz using an Octal Bio Amp and a Power Lab/8SP unit (ADInstruments Pty Ltd. Australia). The maximum root mean square (RMS) of each stage of the CCFT was calculated using a software program (Matlab (7.0), Mathworks Inc, USA). The RMS values obtained during each stage of the CCFT were normalized against RMS values obtained during a reference contraction – a head lift task. 2.3.5. Cranio-cervical ﬂexor and extensor strength Strength of both the cranio-cervical ﬂexors and extensors was measured in sitting using the NeckMetrixÔ dynamometer (O’Leary et al., 2005). The dynamometer measures cranio-cervical ﬂexion and extension muscle torque about its axis that is aligned to axis of rotation of the C0/1 motion segment (concha of the ear). Maximal ﬂexion and extension effort is resisted by the pad of the dynamometer placed at the inferior and superior surfaces of the mandible, respectively, generating torque at the dynamometer axis that is recorded in Newton-meters (Nm) via a computer equipped with a custom-written data acquisition program (Labview 6i Virtual Instruments, Austin, USA). All dynamometry tests were performed with the head in a neutral position as determined by an anthropometric neutral head ﬂexion/extension position (Frankfort Plane). A warm-up of three sub-maximal repetitions was performed for each direction. This was followed by three trials of maximal contractions with 60 s rest between each trial. The maximal torque value of cranio-cervical ﬂexor and extensor strength was used for analysis. 2.4. Procedure

2.3.2. Symptomatic cervical joint dysfunction A manual examination of the cervical segments was conducted by a musculoskeletal physiotherapist to detect the presence or absence of symptomatic cervical joint dysfunction (Jull et al., 1988). Subjects rated any pain provoked on palpation on a Numeric Rating Scale (NRS) (Williamson and Hoggart, 2005) (where 0 ¼ no pain and 10 ¼ the worst pain imaginable) and the examiner rated the perceived tissue resistance to the manual palpation as normal, slight, moderate, marked resistance. A joint was classiﬁed as symptomatic if any pain provoked by manual palpation was 2/10 in combination with the examiner’s rating of moderately or markedly abnormal tissue compliance.

Initial screening was conducted by telephone or electronic mail for inclusion/exclusion purposes and an appointment made to attend the study. On the testing day, all eligible subjects completed the questionnaires. Any headache subject reporting more than one headache type completed separate questionnaires for each headache. The physical measures were conducted in a standard order for efﬁcient use of equipment. The examiner performing the cervical manual examination was blinded to the subject’s headache or control status. To further reduce examiner bias, headache classiﬁcation from the questionnaires was not undertaken until all subjects completed testing. Classiﬁcation was initially undertaken

638

S. Uthaikhup et al. / Manual Therapy 14 (2009) 636–641

independently by two researchers and then by consensus. Headaches were classiﬁed as migraine, tension-type according to the IHS criteria (International Headache Society, 2004), cervicogenic headache according to the Cervicogenic Headache International Study Group criteria (Sjaastad et al., 1998) or, in cases where symptoms were mixed or there was no consensus, headaches were designated as unclassiﬁable. No comment could be made on medication overuse headache due to a lack of data on frequency of medication use. 2.5. Data management and statistical analysis K-means cluster analysis was used to extract groups from the headache population. Cervical extension range, symptomatic C0–3 joints and cervical ﬂexor muscle function (normalized RMS EMG activity of SCM in CCFT) were chosen as the candidate variables to deﬁne clusters which did or did not align with musculoskeletal dysfunction previously found in cervicogenic headache (Amiri et al., 2007; Jull et al., 2007). A two-cluster solution was set for the analysis as it was expected to have two groups from the headache sample, i.e. cluster 1 with and cluster 2 without CMD. Outliers are usually selected as initial cluster centers in the K-means clustering. Thus the sample of 118 headache subjects was screened for outliers and ﬁve were removed from the initial analysis. A further six subjects were removed because of missing data in the EMG recordings. The cluster analysis was performed on the remaining sample of 107 headache subjects. The observed signiﬁcance levels were used to indicate how much the CMD contributed to the separation of the clusters. To test for differences in the three cervical musculoskeletal variables between the two clusters and the control group, a series of one-way ANOVAs and Bonferroni post hoc tests were performed for analysis of continuous variables and Chi-square was used for categorical variables. Signiﬁcance was set at p < 0.05. One-way ANOVAs and Bonferroni post hoc tests and chi-square analyses were used to test for any differences in other musculoskeletal variables tested in the study between the groups (clusters 1, 2 and control). Gender was entered as a covariate in the model (ANCOVA) for the analysis of the muscle strength measurement. The imputation technique to replace the missing data with the mean was carried out for the data missing at random in the study (range 1–5% of values). To minimize the possibility of a type I error, signiﬁcance was set at p 0.01. Headache classiﬁcations and cluster membership were crosstabulated to determine the distribution of headache type between clusters. Headache variables were also analyzed between the clusters using independent t-test or chi-square. Signiﬁcance was set as p < 0.05. All data were analyzed for a normal distribution using the Kolmogorov–Smirnov test and transformation was used for data with a skewed distribution. Statistical analyses were conducted using SPSS statistical package (version 15.0). 3. Results 3.1. Participants A total of 270 volunteers responded to advertisements. Fiftythree elders failed to meet the inclusion criteria and 55 declined to participate because of travelling distance to the testing venue. One hundred and eighteen headache and 44 non-headache control subjects entered the study. Table 1 presents subject demographics and general headache features. Forty-two percent of subjects reported a headache on the testing day but it did not preclude testing in any subject.

Table 1 Subject demographic and general headache characteristics. Variables

Headache (n ¼ 118)

Control (n ¼ 44)

Gender (female, %) Age (years, mean SD) BMI (kg/m2, mean SD) MS pain (yes, %) Headache intensity (1–10 VAS, mean SD) Headache frequency (15 days/month, %) Headache history (years, mean SD) NDI score, % score (mean SD) Neck pain with headache (yes, %)

56.8 65.9 4.6 26.6 4.4 61.0* 6.1 2.0 41.5 26.4 18.1 24.8 13.3 75

68.2 66.4 4.1 25.4 3.7 22.7 – – – – –

MS pain: co-existing musculoskeletal pain in other body regions. VAS: visual analogue scale. *p < 0.01.

3.2. Group clusters The cluster analysis grouped 57 of the 107 headache subjects into cluster 1 and 50 into cluster 2 on the basis of the three candidate musculoskeletal variables. Cluster 1 subjects had higher values of CMD and were separated from cluster 2 on the basis of less extension ROM and a higher incidence of symptomatic joint dysfunction at C1–2. The RMS EMG amplitudes of the CCFT did not contribute to the separation of the headache clusters. The three musculoskeletal variables were compared between clusters 1, 2 and the control group (Table 2). The ANOVA revealed signiﬁcant differences between the groups in range of extension and the incidence of symptomatic joint dysfunction (C0–3) but not in RMS EMG amplitudes in the CCFT. Post hoc analysis determined that cluster 1 subjects had signiﬁcantly reduced range of cervical extension compared to cluster 2 and control subjects, between whom there was no signiﬁcant difference. At C0–1, there were signiﬁcant differences in the frequency of joint dysfunction between cluster 1 and control subjects; at C1–2, cluster 1 subjects had a signiﬁcantly higher frequency of dysfunction than cluster 2 subjects which in turn was signiﬁcantly higher than control subjects. At C2–3, there were no signiﬁcant differences between clusters 1 and 2, but both clusters were signiﬁcantly different to control subjects. 3.3. Other cervical musculoskeletal measures The analyses for other CMD variables revealed greater dysfunctions in the headache clusters than controls in lateral Table 2 Results of the between group (clusters 1, 2 and the control group) analysis for the three candidate cervical musculoskeletal variables. Means with the same letter (a or b) were not signiﬁcantly different in post hoc between group analysis (p > 0.05). Variables

Cluster 1 (n ¼ 57)

Cluster 2 (n ¼ 50)

Control (n ¼ 44)

F

p Value

Extension (degrees)

31.0 1.06

48.6 1.00a

47.3 1.68a

66.8

0.00

14.0a,b (4.4–23.6) 46.0 (32.2–59.8) 58.0a (44.3–71.7) 0.45 0.03a

6.8b (0.6 to 14.2) 11.4 (1.8–20.2) 20.5 (9.0–33.0) 0.42 0.03a

Joint dysfunction (mean% (CI)) C0–1 28.1a (16.3–39.7) C1–2 64.9 (52.6–77.4) C2–3 68.4a (55.9–80.1) CCFT 22–26 0.40 0.02a mmHg (normalized RMS)c

CI: 95% conﬁdence intervals. Values are mean SE or as otherwise indicated. c The values are transformed data. d Chi-square value.

8.4d

0.02

29.3d

0.00

24.3d

0.00

0.9

0.41

S. Uthaikhup et al. / Manual Therapy 14 (2009) 636–641

639

ﬂexion and rotation range and in the frequency of symptomatic joint dysfunction (C3–4 to C7–T1) (all p 0.01) (Table 3). There were no signiﬁcant between group differences in ﬂexion ROM, cranio-cervical muscle strength and cervical JPS (p > 0.01). There were trends for higher levels of dysfunction in cluster 1 than in cluster 2 in the majority of measures, but these only reached signiﬁcance for cervical rotation range.

Table 4 Distribution of headache types in clusters.

Cluster 1 (n ¼ 57)

Cluster 2 (n ¼ 50)

Single headaches Migraine Tension-type Cervicogenic

11 (45.8) 7 (77.8) 16 (72.7)

13 (54.2) 2 (22.2) 6 (27.3)

24 9 22

3.4. Headache distribution between clusters

Non-classiﬁable headaches

13 (39.4)

20 (60.6)

33

5 (55.6) 5 (50)

4 (44.4) 5 (50)

9 10

In the classiﬁcation of the 107 headache subjects, 24 were classiﬁed as migraine, 9 as tension-type, 22 as cervicogenic headaches, 33 headaches were unclassiﬁable and 19 subjects presented with two or more headaches. Cross-tabulations between headache classiﬁcation and assigned cluster are presented in Table 4. As can be observed, while cluster 1 contained the majority of headaches classiﬁed as cervicogenic, it nevertheless contained all headache types. 3.5. Characteristics of headaches in clusters Headache characteristics were compared between clusters to determine whether or not there were distinguishing features of elders’ headache in cluster 1 with higher levels of associated CMD (Table 5). This analysis indicated that cluster 1 subjects had higher headache frequency (headaches 15 per month) and more frequently had a previous history of head or neck trauma (p < 0.05). 4. Discussion This study revealed the presence of CMD in elders with headache compared to healthy controls and CMD appears to be a generic feature of headache in this age group. Some distinction was made between headache in elders on the basis of the magnitude of CMD.

Table 3 Results of the differences in other cervical musculoskeletal variables. Means with the same letter (a or b) were not signiﬁcantly different in post hoc between group analysis (p > 0.01). Variables ROM (degrees) Flexion Lateral ﬂexion Axial rotation

Cluster 1 (n ¼ 57)

Cluster 2 (n ¼ 50)

Control (n ¼ 44)

F

p Value

40.9 1.38a 20.6 0.80a 46.0 1.20

42.4 1.29a 23.5 0.94a,b 52.6 1.34a

45.5 1.50a 25.0 0.85b 53.2 1.14a

2.6 6.5 10.9

0.07 0.00 0.00

Symptomatic cervical joint dysfunction (mean% (CI)) 48.0a 9.1 C3–4 54.4a (41.1–67.0) (34.2–61.8) (0.5–17.5) a a 42.0 6.8 C4–5 47.4 (34.0–60.0) (28.3–55.7) (0.6 to 14.2) a a 34.0 4.5 C5–6 43.9 (31.1–56.9) (20.9–47.1) (1.6 to 10.6) 22.0a 0 C6–7 24.0a (12.9–35.1) (10.5–33.5) (0–0) 0a C7–T1 15.8 4.0a (1.4 to 9.4) (0–0) (6.5–25.5)

23.9d

0.00

20.5

d

0.00

19.4

d

0.00

12.4d

0.00

10.4d

0.01

Muscle strength (Nm) 2.7 0.08a Cranio-ﬂexionc Cranio-extensionc 2.7 0.10

2.7 0.09a 2.8 0.10a

2.6 0.09c 3.0 0.11a

0.2 3.6

0.80 0.03

Joint position error (degrees) Extension 5.1 0.36a Rotation (left) 3.8 0.28a Rotation (right) 5.5 0.53a

5.1 0.45a 3.3 0.22a 5.2 0.59a

4.3 0.37a 3.9 0.36a 5.2 0.55a

1.2 1.3 0.1

0.30 0.29 0.93

CI: 95% conﬁdence interval. Values are mean SE or as otherwise indicated. c The values are transformed data after controlling for gender. d Chi-square values.

Classiﬁcation

Two or more headaches Cervicogenic headachea No cervicogenic headacheb Total

No. (%) of subjects

57

Total

50

107

a

Cervicogenic headache was classiﬁed as one of the subjects’ two or more headaches. b None of the headaches were classiﬁed as cervicogenic.

Subjects in cluster 1 (53% of the sample) displayed higher levels of CMD than those in cluster 2 and both clusters displayed more CMD than control subjects. The values for cervical musculoskeletal features measured across all groups are reduced when compared to values from younger populations (Dall’Alba et al., 2001; Prushansky et al., 2006) and would reﬂect age changes (ten Have and Eulderink, 1981; Dvorak et al., 1992). The changes in these elder headache groups cannot be attributed to ageing alone as they were signiﬁcantly greater than those measured in the control group. Pearce (1995) and Biondi and Saper (2000) reasoned that with increasing cervical degenerative joint disease with age, cervicogenic headache becomes more common. Our results indicate a less distinct and perhaps more complex picture of headache in the elderly regarding the potential role of CMD. The majority of subjects with cervicogenic headache as a single headache (16/22) were grouped into cluster 1, the group with greater levels of CMD. Nevertheless elders with other headache types (11/24 migraine, 7/9 tension-type) were also grouped in this cluster. All headache types were represented in cluster 2 although only 20% were classiﬁed as cervicogenic headache (as a single or one of multiple headache). This cluster had less measured CMD than cluster 1, but still presented with more dysfunction than control subjects. Thus all headache subjects had CMD to a greater or lesser degree, regardless of headache classiﬁcation or length of headache history, although a higher percentage of cluster 1 subjects (53% versus 32% in cluster 2) had frequent headaches (15 days per month). Our hypothesis of the greater relationship of CMD with cervicogenic headache than other headache types was rejected for our elders with headache. Hagen et al. (2002) also observed this interaction between musculoskeletal symptoms and both migrainous and nonmigrainous headache. Table 5 Summary of categorical variables in clusters 1 and 2 and signiﬁcance values.

Demographics Gender, female (%) Age, years (mean SD) BMI, kg/m2 (mean SD) Headache history, years (mean SD) Headache frequency, 15 days/month (%) Headache intensity, 1–10 VAS (mean SD) NDI score, % score (mean SD) Associated neck pain with headache (%) History of head-neck trauma (%) Co-existing musculoskeletal pain (%) VAS ¼ visual analogue scale.

Cluster 1 (n ¼ 57)

Cluster 2 (n ¼ 50)

p Value

52.6 65.8 4.3 27.0 4.0 26.3 19.0 52.6 6.4 2.0 26.2 13.4 77.2 43.9 66.7

64.0 65.7 4.7 25.7 4.4 25.7 18.6 32.0 6.3 2.3 21.8 12.3 67.3 22.0 56.0

0.24 0.91 0.11 0.87 0.03 0.87 0.08 0.26 0.02 0.26

640

S. Uthaikhup et al. / Manual Therapy 14 (2009) 636–641

The CMD in the headache of elders reﬂects that documented in other studies of cervicogenic headache. Reduced range of cervical extension and rotation has been reported previously (Zwart, 1997; Jull et al., 2007), as has palpable tenderness over the upper cervical joints (Jull et al., 2007; Sjaastad and Bakketeig, 2008). The results of the CCFT did not differ between the headache clusters when compared to controls as previously found in cervicogenic headache and neck pain disorders (Jull et al., 2004, 2007). The CCFT is a cognitive test which relies on ﬁne motor skills. The performance of the test may reﬂect age effects on cognition, learning and motor skill acquisition (Wishart and Lee, 1997). The lack of differences between headache and control subjects in cranio-cervical extensor muscle strength may also reﬂect age-related changes in the muscle system (Jorden et al., 1999). There were no between group differences in cervical JPS as also noted by others (Zito et al., 2005; Jull et al., 2007). Headache classiﬁcation was challenging. Approximately 30% could not be classiﬁed as the descriptions presented features that crossed the characteristics of tension-type, migraine and cervicogenic headaches. There was an unusually low incidence of tensiontype headache. It is probable, based on general headache epidemiology, that many of the unclassiﬁable headaches were tension-type. In addition, 19 subjects reported concurrent headaches. Difﬁculty in classiﬁcation is not novel (Fishbain et al., 2001) and the overlap in headache symptoms has often been encountered (D’Amico et al., 1994; van Suijlekom et al., 1999). Indeed Srikiatkhachorn (1991) found that a large number of headaches in elders (61%) could not be classiﬁed using the IHS criteria, having characteristics of a combination of vascular and tension-type. This probably reﬂects the age-related changes in headache features (Haan et al., 2007). Even though an exclusive link between CMD and cervicogenic headache was not made in this study, the changing nature of headache together with the widespread presence of CMD in all headache types could support the possibility of a primary headache evolving to a secondary cervicogenic or mixed headache in the elderly (Kelman, 2006). Our results might suggest a possible cervical musculoskeletal contribution to a variety of headaches in the elderly. In line with this contention, the presence of neck pain with headache in our elders was common (75%), supporting Kelman’s (2006) ﬁndings in elders with primary headaches. Neck pain accompanies up to 60–80% of headaches in the general population (Hagen et al., 2002; Fishbain et al., 2003) but this does not necessarily infer a cervicogenic origin of pain (Jull et al., 2007). Neck pain may be a referred pain of a primary headache, reﬂecting the bidirectional pathway in the trigeminocervical nucleus (Bartsch and Goadsby, 2003). Conversely its origin may be in the periphery (cervical structures) consistent with the CMD measured in this study. Elders with headache also reported more frequent coexisting musculoskeletal pain than controls. The relative high incidence of co-morbid shoulder, back and knee pains in both headache clusters is consistent with other studies of chronic headache (Terwindt et al., 2000; Wiendels et al., 2006). Wiendels et al. (2006) also found that subjects with chronic headache had more co-morbid musculoskeletal pain than those with infrequent headache, which was supported in this study. Additionally, a history of neck trauma was more prevalent in cluster 1 subjects but trauma can precipitate either primary or secondary headaches (Radanov et al., 2001). This study demonstrated that symptomatic CMD beyond that which could be attributed to ageing is present in elders with headache which challenges diagnosis. While its role cannot be explained by this study, the results suggest that CMD might, in some cases, be the origin of headache (cervicogenic headache), in others it could be a trigger of headache (Cook et al., 1989; Solomon

et al., 1990), or its presence may indicate a transition from a primary to a secondary headache as an age-related process. Alternately the CMD may be merely a co-morbid feature which may compound the pain of an elder’s headache syndrome. Further research is required to better understand the role or impact of CMD on headache in elders and this might be achieved by examining the relative outcomes in a clinical trial of treatment to the neck. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.math.2008.12.008. References Amiri M, Jull G, Bullock-Saxton J, Darnell R, Lander C. Cervical musculoskeletal impairment in frequent intermittent headache, part 2: subjects with concurrent headache types. Cephalalgia 2007;27(8):891–8. Bartsch T, Goadsby PJ. Increased responses in trigeminocervical nociceptive neurons to cervical input after stimulation of the dura mater. Brain 2003;126(8): 1801–13. Biondi DM, Saper JR. Geriatric headache: how to make the diagnosis and manage the pain. Geriatric 2000;55(12):40–50. Cook N, Evans DA, Funkenstein HH, Scherr PA, Ostfeld AM, Taylor JO, et al. Correlates of headache in a population-based cohort of elderly. Archives of Neurology 1989;46:1338–44. Dall’Alba PT, Sterling MM, Treleaven JM, Edwards SL, Jull GA. Cervical range of motion discriminates between asymptomatic persons and those with whiplash. Spine 2001;26(19):2090–4. D’Amico D, Leone M, Bussone G. Side-locked unilaterality and pain localization in long-lasting headaches – migraine, tension-type headache, and cervicogenic headache. Headache 1994;34(9):526–30. 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. Falla D, Dall’Alba P, Rainoldi A, Merletti R, Jull G. Location of innervation zones of sternocleidomastoid and scalene muscles: a basis for clinical and research electromyography applications. Clinical Neurophysiology 2002;113(1):57–63. Fishbain DA, Cutler R, Cole B, Rosomoff HL, Rosomoff RS. International Headache Society: headache diagnostic patterns in pain facility patients. Clinical Journal of Pain 2001;17(1):78–93. Fishbain DA, Lewis J, Cole B, Cutler RB, Rosomoff RS, Rosomoff HL. Do the proposed cervicogenic headache diagnostic criteria demonstrate speciﬁcity in terms of separating cervicogenic headache from migraine? Current Pain and Headache Reports 2003;7(5):387–94. Haan J, Hollander J, Ferrari M. Migraine in the elderly: a review. Cephalalgia 2007;27(2):97–106. Hagen K, Einarsen C, Zwart JA, Svebak S, Bovim G. The co-occurrence of headache and musculoskeletal symptoms amongst 51,050 adults in Norway. European Journal of Neurology 2002;9(5):527–33. 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. ten Have HA, Eulderink F. Mobility and degenerative changes of the ageing cervical spine: a macroscopic and statistical study. Gerontology 1981;27(1–2):42–50. International Headache Society. The international classiﬁcation of headache disorders: 2nd edition. Cephalalgia 2004;24(Suppl. 1):S1–151. Jorden A, Mehlsen J, Bulow PM, Ostergaard K, Danneskiold SB. Maximal isometric strength of the cervical musculature in 100 healthy volunteers. Spine 1999;24(13):1343–8. Jull G, Bogduk N, Marsland A. The accuracy of manual diagnosis for cervical zygapophysial joint pain syndromes. Medical Journal of Australia 1988;148(5):233–6. Jull G, Kristjansson E, Dall’Alba P. Impairment in the cervical ﬂexors: a comparison of whiplash and insidious onset neck pain patients. Manual Therapy 2004;9(2):89–94. Jull G, Amiri M, Bullock-Saxton J, Darnell R, Lander C. Cervical musculoskeletal impairment in frequent intermittent headache, part 1: subjects with single headaches. Cephalalgia 2007;27(7):793–802. Kelman L. Migraine changes with age: IMPACT on migraine classiﬁcation. Headache 2006;46(7):1161–71. Lyngberg AC, Rasmussen BK, Jorgensen T, Jensen R. Incidence of primary headache: a Danish epidemiologic follow-up study. American Journal of Epidemiology 2005;161(11):1066–73. Martins KM, Bordini CA, Bigal ME, Speciali JG. Migraine in the elderly: a comparison with migraine in young adults. Headache 2006;46(2):312–6. O’Leary SP, Vicenzino BT, Jull GA. A new method of isometric dynamometry for the craniocervical ﬂexor muscles. Physical Therapy 2005;85(6):556–64. Pearce JMS. The importance of cervicogenic headache in the over-ﬁfties. Headache Quarterly-Current Treatment and Research 1995;6(4):293–6.

S. Uthaikhup et al. / Manual Therapy 14 (2009) 636–641 Prushansky T, Pevzner E, Gordon C, Dvir Z. Performance of cervical motion in chronic whiplash patients and healthy subjects: the case of atypical patients. Spine 2006;31(1):37–43. Radanov BP, Stefano GD, Augustiny KF. Symptomatic approach to posttraumatic headache and its possible implications for treatment. European Spine Journal 2001;10:403–7. Revel M, Andre-Deshays C, Minguet M. Cervicocephalic kinesthetic sensibility in patients with cervical pain. Archives of Physical Medicine and Rehabilitation 1991;72(5):288–91. Sjaastad O, Fredriksen TA, Pfaffenrath V. Cervicogenic headache: diagnostic criteria. Headache 1998;38(6):442–5. Sjaastad O, Bakketeig LS. Migraine without aura: comparison with cervicogenic headache: vaga study of headache epidemiology. Acta Neurologica Scandinavia 2008;117(6):377–83. Solomon GD, Kunkel RS, Frame J. Demographics of headache in elderly patients. Headache 1990;30(5):273–6. Srikiatkhachorn A. Epidemiology of headache in the Thai elderly: a study in the Bangkae home for the aged. Headache 1991;31(10):677–81.

641

van Suijlekom JA, de Vet HCW, van den Berg SGM, Weber WEJ. Interobserver reliability of diagnostic criteria for cervicogenic headache. Cephalalgia 1999;19(9):817–23. Terwindt GM, Ferrari MD, Tijhuis M, Groenen SM, Picavet HS, Launer LJ. The impact of migraine on quality of life in the general population: the GEM study. Neurology 2000;55(5):624–9. Vernon H, Mior S. The Neck Disability Index: a study of reliability and validity. Journal of Manipulative and Physiological Therapeutics 1991;14(7):409–15. Wiendels NJ, van Haestregt A, Knuistingh Neven A, Spinhoven P, Zitman FG, Assendelft WJ, et al. Chronic frequent headache in the general population: comorbidity and quality of life. Cephalalgia 2006;26(12):1443–50. Williamson A, Hoggart B. Pain: a review of three commonly used pain rating scales. Journal of Clinical Nursing 2005;14(7):798–804. Wishart LR, Lee TD. Effects of aging and reduced relative frequency of knowledge of results on learning a motor skill. Perceptual and Motor Skills 1997;84(3 Pt 1):1107–22. Zito G, Jull G, Story I. Clinical tests of musculoskeletal dysfunction in the diagnosis of cervicogenic headache. Manual Therapy 2005;11(2):118–29. Zwart JA. Neck mobility in different headache disorders. Headache 1997;37(1):6–11.

Manual Therapy 14 (2009) 642–646

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

Abdominal muscle activity during abdominal hollowing in four starting positions Pakkanaporn Chanthapetch*, Rotsalai Kanlayanaphotporn 1, Chitanongk Gaogasigam, Adit Chiradejnant Department of Physical Therapy, Faculty of Allied Health Sciences, Chulalongkorn University, Soi Chula 12 Pathumwan, Bangkok 10330, Thailand

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 January 2008 Received in revised form 3 December 2008 Accepted 19 December 2008

The aim of this study was to investigate the activity of the rectus abdominis (RA), external abdominal oblique (EO), and transversus abdominis/internal abdominal oblique (TrA/IO) muscles during abdominal hollowing (AH) in four positions: crook lying, prone lying, four-point kneeling, and wall support standing. Thirty-two healthy participants, aged 21.3 0.8 years were recruited. They were instructed to perform maximal voluntary contraction (MVC) and AH. The electromyography (EMG) data of each muscle during AH were normalized as a percentage of MVC. During AH in all four starting positions, signiﬁcant differences were found in the EMG activity of RA, EO, and TrA/IO (p < 0.001). The TrA/IO exhibited the highest while the RA exhibited the lowest EMG activity. Among the four different starting positions, only the TrA/IO showed signiﬁcant difference in mean EMG activity (p < 0.001). The results suggest that all four starting positions can facilitate TrA/IO activity with minimal activity from RA and EO. Crown Copyright Ó 2009 Published by Elsevier Ltd. All rights reserved.

Keywords: Exercises Abdominal muscles Transversus abdominis Low back pain

1. Introduction Low back pain (LBP) is a common problem that occurs in the general population. One year prevalence of LBP has been reported to range from 40.5 percent to 64 percent (Barrero et al., 2006; Ihlebaek et al., 2006). It has been found that approximately 60–80 percent of the population report LBP once in their life time (Manchikanti, 2000; Ihlebaek et al., 2006). One hypothesis for the development of LBP is that there is a dysfunction in the control of the abdominal and back muscles (Panjabi, 1992; Richardson and Jull, 1995; O’Sullivan et al., 1997; Hides et al., 2001). Speciﬁc exercises that aim to train these trunk muscles to function properly are, thus, a contemporary approach for treating LBP. Abdominal hollowing (AH) is one of these exercises that is widely used in patients suffering from LBP (O’Sullivan et al., 1997; Hides et al., 2001; Rasmussen-Barr et al., 2003; Shaughnessy and Caulﬁeld, 2004; Goldby et al., 2006). To learn how to perform AH, it is recommended that a patient with LBP should start practising AH in a position that facilitates the co-contraction of the deep abdominal and back muscles. When the patient can master AH, the starting position can be changed. The muscles that should be activated during AH are the transversus abdominis (TrA), the internal abdominal oblique (IO) (lower ﬁbres), and the lumbar multiﬁdus (deep ﬁbres) which have * Corresponding author. E-mail addresses: [email protected] (P. Chanthapetch), rotsalai.k@ chula.ac.th, [email protected] (R. Kanlayanaphotporn). 1 Tel.: þ66 2 218 3765; fax: þ66 2 218 3766.

been proposed to function synergistically (Richardson et al., 2004). To be effective, co-contraction of these deep trunk muscles should occur in isolation from the rectus abdominis (RA) and the external abdominal oblique (EO) which lie superﬁcially. Empirically, four positions have been suggested by clinicians as the starting positions for performing AH. These positions are crook lying (O’Sullivan, 2000), prone lying (Richardson and Jull, 1995; O’Sullivan, 2000), four-point kneeling (Richardson and Jull, 1995; Norris, 1999; O’Sullivan, 2000), and wall support standing (Norris, 1999). To date, there have been only two studies that have compared the effectiveness of the starting positions for performing AH (Beith et al., 2001; Urquhart et al., 2005). Beith et al. (2001) compared the prone lying to the four-point kneeling position and found no statistical difference in the activity of IO between positions. However, an isolated activation of IO tended to occur more easily in the four-point kneeling position than in prone lying position. Urquhart et al. (2005) compared crook lying with prone lying positions. They found that crook lying could encourage the TrA to work in isolation better than the prone lying position. The aims of this study were to determine 1) whether there was any signiﬁcant difference in electromyography (EMG) activity among the three abdominal muscles in each of the four starting positions (crook lying, prone lying, four-point kneeling, and wall support standing); 2) whether there was any signiﬁcant difference in the EMG activity of each muscle among the four different starting positions; and 3) whether there was any difference in the frequency of non-activation and isolation of three abdominal muscles among the four starting positions.

1356-689X/$ – see front matter Crown Copyright Ó 2009 Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.12.009

P. Chanthapetch et al. / Manual Therapy 14 (2009) 642–646

2. Methods 2.1. Participants A convenience sample of 32 asymptomatic LBP (14 male, 18 female) took part in this study. Their mean standard deviation of age was 21.3 0.8 years, their weight 50.2 8.2 kg, their height 1.64 0.08 m, and their body mass index was 18.6 1.8 kg/m2. The participants were recruited from the students and staff of the Faculty of Allied Health Sciences at the Chulalongkorn University. Participant recruitment commenced in September 2005 and continued until January 2006. They were excluded if they had a history of LBP, practised AH, or any abnormalities of the spinal column or abdominal region such as fractures, surgery, burns, or cancer. Moreover, participants with a skinfold thickness in the abdominal and supra-iliac area greater than 20 mm were also excluded. This aimed to decrease the EMG artifact due to adipose tissue lying between the surface electrodes and the tested muscles (Neumann and Gill, 2002). All the participants had thin skinfold (abdominal skinfold thickness was 16 4 mm and supra-iliac skinfold thickness was 9 3 mm). Ethical approval for the study was granted by the Research Ethics Committee, Chulalongkorn University, Thailand. 2.2. Procedure Participants gave written informed consent prior to participation in the study. Initially, all participants were trained to perform AH in four positions. The standard protocol suggested by Richardson and Jull (1995) as described below was practised until they were able to correctly perform the AH. After the training session, only the participants who could perform AH correctly continued with the study. Then, they were instructed to perform maximal voluntary contraction (MVC) and AH. The order of the position was randomly assigned using a 4 4 balanced Latin square (Portney and Watkins, 2000). All participants were tested in the afternoon.

643

extended. The distance from the wall to their heels was 6 inches (Norris, 1995). Brieﬂy, the AH was performed by gently pulling the navel in and up while not allowing any movement at the spine, rib, or pelvis (Norris, 1995; Richardson and Jull, 1995; O’Sullivan, 2000). After the navel has been drawn close to the spine, the participants were instructed to hold the abdominal contraction for 10 s while continually breathing normally. This aimed to activate the TrA at a low level of muscle contraction which should be approximately 25 percent of its MVC (Richardson and Jull, 1995). The lumbar spine was kept in a neutral position such that the anterior superior iliac spine and posterior superior iliac spine were aligned vertically (Richardson et al., 2004). The duration of training for each participant varied from 10 to 40 min. 2.3. Measurement EMG recordings were made using silver/silver chloride surface electrodes of 1 cm in diameter which were placed with a centre-tocentre spacing of 2.2 cm (Ng et al., 1998). The three channels method was used in which a reference electrode for each muscle was placed adjacent to the paired electrodes of that muscle. All abdominal muscles were recorded on the right side by positioning the surface electrodes in parallel to the muscle ﬁbres (Fig. 1). All EMG placements followed those recommended by Ng et al. (1998). For the RA, the electrodes were placed in a cephalad/caudad orientation at 2 cm inferior to the navel and 1 cm lateral to the midline. For the EO, the electrodes were placed diagonally on the inferior edge along a line connecting the most inferior point of the costal margin and the contralateral pubic tubercle. For the TrA and IO, the electrodes were placed in the area where the TrA and IO fuse together and this was called TrA/IO. The TrA/IO electrodes were placed horizontally at 2 cm inferior and medial to the anterior superior iliac spine (Marshall and Murphy, 2003). EMG were sampled at 1000 Hz over a bandwidth of 8–500 Hz using the ME3000P8 EMG systemÒ (Mega Electronics Ltd, Kuopio, Finland):

2.2.1. Maximal voluntary contraction All participants were asked to perform three manoeuvres which were expected to generate maximal EMG activity for each of the three abdominal muscles. These manoeuvres were trunk ﬂexion, trunk ﬂexion with rotation to the left, and trunk ﬂexion with rotation to the right. Each manoeuvre was performed against manual resistance once in crook lying and then in sitting (Beith et al., 2001). During the performances, the participants were instructed to avoid any jerky contractions in order to decrease the chance of injury. Each manoeuvre was held for 5 s with a 2-min rest between trials to prevent muscle fatigue (Ng et al., 2002). For each muscle, the greater EMG that was produced either in the crook lying or sitting position was chosen as a reference value for normalization. 2.2.2. Abdominal hollowing All participants were required to perform AH for 10 s, three times in each position, with a 1-min rest between each time. For the crook lying position, the knees were ﬂexed at 90 (Drysdale et al., 2004). For the prone lying position, a small pillow was placed under the ankles (Richardson and Jull, 1995). For the four-point kneeling position, the participants were asked to look at the ﬂoor with their ears in horizontal line to the shoulder joints, their knees directly below their hips and their wrists directly below the shoulders (Norris, 1999). A small pillow was placed under their ankles (Richardson and Jull, 1995). For the wall support standing position, participants were asked to stand with their backs against the wall while their hips were slightly ﬂexed and their knees were

Fig. 1. Location for attaching surface electrodes to the abdominal wall. (A1 ¼ Paired electrode of the rectus abdominis muscle, A2 ¼ Reference electrode of the rectus abdominis muscle, B1 ¼ Paired electrode of the external abdominal oblique, B2 ¼ Reference electrode of the external abdominal oblique, C1 ¼ Paired electrode of the transversus abdominis/internal abdominal oblique, and C2 ¼ Reference electrode of the transversus abdominis/internal abdominal oblique).

644

P. Chanthapetch et al. / Manual Therapy 14 (2009) 642–646

with a differential ampliﬁcation gain of 375 and a common mode rejection ratio >110 dB. The data were converted from analogue to digital and simultaneously displayed on a personal computer monitor which was recording for subsequent analysis. 2.4. Data processing

3. Results

The raw surface EMG signals generated during both MVC and AH were analyzed using the root mean squared technique. The mean EMG activity over a 1-s period was calculated for both MVC and AH. The average was performed at the highest area for MVC while it was collected at a 2-s period prior to the end of the EMG activity for AH (Beith et al., 2001). The EMG activity during AH was normalized by the MVC for each muscle. 2.5. Statistical analysis Statistical analysis was performed using the SPSS version 13.0 for Windows. Based on analysis by one-sample Kolmogorov– Smirnov test which found that the EMG data were not normally distributed, the non-parametric statistical analyses were therefore used. Initially, the Friedman two-way analysis of variance (ANOVA) was performed to test for the differences in EMG activity due to the muscles or the starting positions. The signiﬁcant difference level was set at p < 0.05. Where a signiﬁcant difference emerged, a multiple comparison procedure with the Wilcoxon signed-ranks test was used to test which pairwise differences were signiﬁcant. A Bonferroni adjusted alpha level was used to safeguard for the overall Type I error to be accepted as signiﬁcant (Portney and Watkins, 2000). The percentage of participants who showed non-activation of the global muscles (RA or EO) and isolation of the local muscles (contraction of TrA/IO with non-activation of RA and EO) was calculated. Participants who could keep their RA, EO, or both unchanged from the baseline over the three trials of AH were categorized as the ‘always’ non-activation group. The participants who could keep neither their RA, EO, nor both unchanged at all three trials were categorized as the ‘never’ non-activation group. The participants who could sometimes keep their RA, EO, or both unchanged over the three trials were categorized as the ‘sometimes’ non-activation group. For the isolated activation of the local muscles, the participants were also classiﬁed into three categories. They were always, never, and sometimes activation groups. The same percentages of participants in each category of the nonactivation groups were also applied to each category of the isolated Table 1 Descriptive statistics of electromyographic activity of three abdominal muscles [percentage of maximal voluntary contraction (MVC)] during abdominal hollowing in four positions (N ¼ 32).

Crook lying

Prone lying

Four-point kneeling

Wall support standing

Muscles

RA EO TrA/IO RA EO TrA/IO RA EO TrA/IO RA EO TrA/IO

Table 1 presents the mean, minimum, maximum, and standard deviation of the EMG activity of three abdominal muscles during AH in four positions. 3.1. Comparison of EMG activity within starting positions The highest EMG activity was always found in TrA/IO with minimal EMG activity in EO and RA. Approximately 20–30 percent of MVC was demonstrated in TrA/IO while the associated EMG activity of EO and RA was less than 6.5 percent of MVC (Table 1, Fig. 2). The Friedman two-way ANOVA showed signiﬁcant differences in the EMG activity of three abdominal muscles in all four starting positions (p < 0.001) (Table 2). To determine in each starting position which pairwise comparisons of the EMG activity of three abdominal muscles were signiﬁcantly different, post hoc analysis was performed. The results showed that the EMG activity of all three abdominal muscles was signiﬁcantly different from each other in all four starting positions (p < 0.001). 3.2. Comparison of EMG activity within muscles The TrA/IO EMG activity was highest in the prone lying position (Fig. 2). Only the TrA/IO showed signiﬁcant difference in mean EMG activity among four different starting positions (p < 0.001) (Table 3). The EO EMG activity, however, approached signiﬁcance (p ¼ 0.053). For TrA/IO, post hoc analysis showed signiﬁcant difference only between the prone lying and four-point kneeling positions (p < 0.001). 3.3. Non-activation and isolation of abdominal muscle in four starting positions In all four positions, more participants could reduce RA EMG activity than EO EMG activity (Fig. 3A and B). Forty percent or more of participants could sometimes perform AH with no contribution from RA. By contrast, 75 percent or more of participants could never perform AH without contribution from EO (Fig. 3B). During AH, the number of participants who could activate TrA/IO in isolation from RA and EO was similar across all four positions

60 RA

Abdominal muscle activity (% MVC) Minimum

Mean (SD)

Maximum

0.00 0.00 4.96 0.00 0.00 4.60 0.00 0.00 3.94 0.00 0.11 5.91

1.84 4.82 26.21 1.35 6.09 27.59 1.35 4.52 18.75 2.09 6.28 20.89

16.67 34.57 84.00 4.67 35.65 69.60 6.67 29.17 77.36 13.88 24.54 82.60

(3.16) (6.78) (22.86) (1.49) (6.87) (19.22) (1.93) (6.01) (16.68) (3.40) (5.58) (16.01)

RA ¼ Rectus abdominis, EO ¼ External abdominal oblique, TrA/IO ¼ Transversus abdominis/internal abdominal oblique.

EO

TrA/IO

50

MVC

Position

activation groups. This was because all participants were required to activate their local muscles while trying to keep their global muscles non-activated during AH. The non-activation of the global muscles would in turn reﬂect the isolated activation of the local muscles.

40 30 20 10 0

CR

PR

FO

WA

Fig. 2. Mean and standard deviation of electromyographic activity of three abdominal muscles during abdominal hollowing in four positions. RA, EO, and TrA/IO represented rectus abdominis, external abdominal oblique, and transversus abdominis/internal abdominal oblique, respectively. (CR, PR, FO, and WA represented crook lying, prone lying, four-point kneeling, and wall support standing positions, respectively).

P. Chanthapetch et al. / Manual Therapy 14 (2009) 642–646

Testing conditions

p-value

Crook lying Prone lying Four-point kneeling Wall support standing

<0.001 <0.001 <0.001 <0.001

A of participants

Table 2 Results of Friedman two-way analysis of variance for assessing the differences in electromyographic activity due to position (N ¼ 32).

RA ¼ Rectus abdominis, EO ¼ External abdominal oblique, and TrA/IO ¼ Transversus abdominis/internal abdominal oblique muscles.

90 80 70 60 50 40 30 20 10 0

CR

Always

4. Discussion The results of this study suggest that the abdominal muscle activity during AH varies with the starting position. However, all four starting positions can facilitate TrA/IO activity with minimal activity from RA and EO.

Table 3 Results of Friedman two-way analysis of variance for assessing the differences in electromyographic activity due to muscle (N ¼ 32). Testing conditions Four positions

p-value RA EO TrA/IO

0.746 0.053 <0.001

RA ¼ Rectus abdominis, EO ¼ External abdominal oblique, and TrA/IO ¼ Transversus abdominis/internal abdominal oblique muscles.

PR

FO

WA

Sometimes

Never

Fig. 3. Frequency of inhibited activity of (A) rectus abdominis (RA) and (B) external abdominal oblique (EO) in four positions during abdominal hollowing. (CR ¼ Crook lying, PR ¼ Prone lying, FO ¼ Four-point kneeling, and WA ¼ Wall support standing positions).

was not presented in the normal prone lying position may have occurred in the previous study. In general, variation in the TrA/IO EMG activity among the starting positions might be explained by the differences in the amount of support provided in each position. In the prone lying and crook lying positions, the whole bodies of the participants were fully supported on the plinth so that all of their postural muscles could be relaxed. More concentration could therefore be placed on the contraction of TrA/IO which resulted in a higher TrA/IO EMG activity in these positions. On the other hand, the spine was not supported in the four-point kneeling position and the body was partially supported in the wall support standing position. The trunk and limb muscles would be activated to control these postures and this might cause the TrA/IO to contract with some difﬁculty. 4.3. Non-activation and isolation of abdominal muscles in four starting positions Apart from the high activity of TrA/IO, one essential criterion that determines the effectiveness of AH is the isolation of TrA/IO

of participants

The EMG activity of all three abdominal muscles varied with starting position. This variation is obvious in TrA/IO. In this study, the crook lying and prone lying positions tended to facilitate TrA/IO activity better than the four-point kneeling and wall support standing positions. The ﬁnding is consistent with a previous study that compared between prone lying (34.7% MVC) and four-point kneeling positions (18.5% MVC) (Beith et al., 2001). However, the slightly higher TrA/IO EMG activity in the prone lying position than in the crook lying position found in this study conﬂicts with the results of the previous study (Urquhart et al., 2005). This might be due to the differences in the settings for the prone lying position between the studies. The prone lying position in the previous study was modiﬁed from normal and differed from the one that is commonly recommended to patients. The researchers had their participants lie prone with two small boxes supported at the xiphisternum and pubic symphysis, respectively. The abdomen did not contact with the plinth. The stretch stimulus on the TrA that

WA

Never

CR

Always

4.2. Comparison of EMG activity within muscles

FO

80 70 60 50 40 30 20 10 0

4.1. Comparison of EMG activity within starting positions All four starting positions produced signiﬁcantly higher EMG activity of TrA/IO than EO and RA. The ﬁndings suggest that these four starting positions are appropriate for performing AH. They encourage contraction of the deep muscles with minimal contribution from the superﬁcial muscles. In addition, the amplitude of TrA/IO activity which was approximately 25 percent of MVC is also in line with the recommendation for performing AH (Richardson and Jull, 1995). These results support the use of these four positions for practising AH in the early stage of speciﬁc exercise training for treating LBP (Richardson and Jull, 1995; Norris, 1999; O’Sullivan, 2000).

PR

Sometimes

B 100 90 of participants

(Fig. 4). Eighty percent or more of participants could never activate TrA/IO in isolation from RA and EO while performing AH.

645

100 90 80 70 60 50 40 30 20 10 0

CR

Always

Sometimes

PR

FO

WA

Never

Fig. 4. Frequency of isolated activity of transversus abdominis/internal abdominal oblique (TrA/IO) in four positions during abdominal hollowing. (CR ¼ Crook lying, PR ¼ Prone lying, FO ¼ Four-point kneeling, and WA ¼ Wall support standing positions).

646

P. Chanthapetch et al. / Manual Therapy 14 (2009) 642–646

(Richardson and Jull, 1995). In other words, the non-activation of RA and EO should also be considered. In this study, the majority of participants could reduce RA EMG activity better than EO EMG activity. This difference in the non-activation of RA and EO might relate to the differences in the anatomical attachment between these two muscles. RA and TrA/IO share minimal attachments at their origins and insertions. The independent activation of RA to TrA/IO might therefore be easier. This was demonstrated as the minimal mean EMG activity of RA (ranging from 1.35 to 2.09 percent of MVC) recorded in all four starting positions. In contrast, EO and TrA/IO share the same ﬁbro-osseous attachments in that they attach to the costal cartilages, the thoracolumbar fascia, the iliac crest, and the pubis (Moore and Dalley, 1999). As a result, these two muscles inevitably work together to ﬂatten the abdomen during AH. Due to the characteristics of the non-activation of RA and EO found in this study, these ﬁndings support the previous study’s conclusion that the elimination of activity in EO is more difﬁcult to achieve (Beith et al., 2001). Nevertheless, the EO EMG activity in this present study was minimal in all four starting positions (ranging from 4.52 to 6.28 percent of MVC). The present ﬁndings should be considered in light of a few limitations. First, this study used surface EMG electrodes to record EMG activity from the deep abdominal muscles. Although the TrA/ IO electrodes were aligned with TrA muscle ﬁbres in this region (Marshall and Murphy, 2003), the recorded EMG signal probably represented activity from both TrA and IO. Since IO overlies the deeper muscle TrA, the EMG signals recorded from the surface electrode position for TrA/IO would be largely from IO rather than TrA. However, this limitation should not affect the results of this study which aimed to investigate the EMG activity of three abdominal muscles during AH in four starting positions. This is because both TrA and IO (lower ﬁbres) have been proven to function as local muscles that should be facilitated during AH (Marshall and Murphy, 2003). Secondly, the crosstalk phenomenon in which the electrodes pick up activity from the adjacent muscles might have occurred in this study. Anatomically, it is known that EO, IO, and TrA form layers in front of the trunk. Any surface electrodes that are attached on the anterolateral aspect of the abdomen would be able to pick up activity from these three muscles. However, the careful electrode placement in this study seems to have reduced this crosstalk phenomenon satisfactorily. Previous work suggests that crosstalk between EO and IO is minimal at the EO recording site (Beith and Harrison, 2004). In addition, the crosstalk between the RA and other abdominal muscles is unlikely (Fuglevand et al., 1992). This was shown as signiﬁcantly different in the EMG activity recorded from EO and TrA/IO. Lastly, all participants in this study were asymptomatic LBP. The presence of pain might have altered the ability of participants to contract their abdominal muscles. Consequently, differences in the starting positions may cause differences in the activation of the abdominal muscles. Nevertheless, the results were deemed to demonstrate the plausible facilitating potential of each starting position for performing AH in well-controlled conditions. Similar results would therefore be expected in the symptomatic population in which the controlling of any confounding factors is relatively more difﬁcult. However, further study in symptomatic LBP is needed.

5. Conclusion This study has implications for early AH training in clinical practice. The results provide support for the use of AH as a speciﬁc stabilization exercise. The ﬁndings of both EMG amplitude as well as the frequency of non-activation and isolation of the abdominal muscles suggest that all four positions can facilitate EMG activity in the local muscles with minimal EMG activity in the global muscles. References Barrero LH, Hsu YH, Terwedow H, Perry MJ, Dennerlein JT, Brain JD, et al. Prevalence and physical determinants of low back pain in a rural Chinese population. Spine 2006;31(23):2728–34. Beith ID, Harrison PJ. Stretch reﬂexes in human abdominal muscle. Experimental Brain Research 2004;159(2):206–13. Beith ID, Synnott RE, Newman SA. Abdominal muscle activity during the abdominal hollowing manoeuvre in the four point kneeling and prone positions. Manual Therapy 2001;6(2):82–7. Drysdale CL, Earl JE, Hertel J. Surface electromyographic activity of the abdominal muscles during pelvic-tilt and abdominal-hollowing exercises. Journal of Athletic Training 2004;39(1):32–6. Fuglevand AJ, Winter DA, Patla AE, Stashuk D. Detection of motor unit action potentials with surface electrodes: inﬂuence of electrode size and spacing. Biological Cybernetics 1992;67(2):143–53. Goldby LJ, Moore AP, Doust J, Trew ME. A randomized controlled trial investigating the efﬁciency of musculoskeletal physiotherapy on chronic low back disorder. Spine 2006;31(10):1083–93. Hides JA, Jull GA, Richardson CA. Long-term effects of speciﬁc stabilizing exercises for ﬁrst-episode low back pain. Spine 2001;26(11):E243–8. Ihlebaek C, Hansson TH, Laerum E, Brage S, Eriksen H, Holm SH, et al. Prevalence of low back pain and sickness absence: a ‘‘borderline’’ study in Norway and Sweden. Scandinavian Journal of Public Health 2006;34(5):555–8. Manchikanti L. Epidemiology of low back pain. Pain Physician 2000;3(2):167–92. Marshall PW, Murphy BA. The validity and reliability of surface EMG to assess the neuromuscular response of the abdominal muscles to rapid limb movement. Journal of Electromyography and Kinesiology 2003;13(5):477–89. Moore KL, Dalley AF. Clinically oriented anatomy. 4th ed. Philadelphia: Lippincott Williams & Wilkins; 1999 [chapter 2]. p. 184–85. Neumann P, Gill V. Pelvic ﬂoor and abdominal muscle interaction: EMG activity and intra-abdominal pressure. International Urogynecology Journal 2002;13(2):125–32. Ng JK, Kippers V, Richardson CA. Muscle ﬁbre orientation of abdominal muscles and suggested surface EMG electrode positions. Electromyography and Clinical Neurophysiology 1998;38(1):51–8. Ng JK, Richardson CA, Parnianpour M, Kippers V. EMG activity of trunk muscles and torque output during isometric axial rotation exertion: a comparison between back pain patients and matched controls. Journal of Orthopaedic Research 2002;20(1):112–21. Norris CM. Spinal stabilisation: 5. An exercise programme to enhance lumbar stabilisation. Physiotherapy 1995;81(3):138–46. Norris CM. Functional load abdominal training: part 2. Journal of Bodywork and Movement Therapies 1999;3(4):208–14. O’Sullivan PB, Phyty GD, Twomey LT, Allison GT. Evaluation of speciﬁc stabilizing exercise in the treatment of chronic low back pain with radiologic diagnosis of spondylolysis or spondylolisthesis. Spine 1997;22(24):2959–67. O’Sullivan PB. Lumbar segmental ‘instability’: clinical presentation and speciﬁc stabilising exercise management. Manual Therapy 2000;5(1):2–12. Panjabi MM. The stabilizing system of the spine. Part I. Function, dysfunction, adaptation, and enhancement. Journal of Spinal Disorders 1992;5(4):383–9. Portney LG, Watkins MP. Foundations of clinical research: applications to practice. 2nd ed. Upper Saddle River: Prentice Hall Health; 2000 [chapter 10]. p. 191. Rasmussen-Barr E, Nilsson-Wikmar L, Arvidsson I. Stabilizing training compared with manual treatment in sub-acute and chronic low-back pain. Manual Therapy 2003;8(4):233–41. Richardson CA, Jull GA. Muscle control-pain control. What exercises would you prescribe? Manual Therapy 1995;1(1):2–10. Richardson CA, Hodges PW, Hides JA. Therapeutic exercise for lumbopelvic stabilization. In: A motor control approach for the treatment and prevention of low back pain. 2nd ed. Edinburgh: Churchill Livingstone; 2004. p. 31–57 [chapter 3]. Shaughnessy M, Caulﬁeld B. A pilot study to investigate the effect of lumbar stabilisation exercise training on functional ability and quality of life in patients with chronic low back pain. International Journal of Rehabilitation Research 2004;27(4):297–301. Urquhart DM, Hodges PW, Allen TJ, Story IH. Abdominal muscle recruitment during a range of voluntary exercises. Manual Therapy 2005;10(2):144–53.

Manual Therapy 14 (2009) 647–653

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

The association between postural alignment and psychosocial factors to upper quadrant pain in high school students: A prospective study Yolandi Brink a, *,1, Lynette Christine Crous b, Quinette Abigail Louw b, Karen Grimmer-Somers c, Kristiaan Schreve d a

Department of Physiotherapy, Stellenbosch University, South Africa, PO Box 2101, Windmeul 7630, Paarl, South Africa Division of Physiotherapy, Department of Interdisciplinary Health Sciences, Stellenbosch University, South Africa Division of Health Sciences, University of South Australia, Australia d Department of Mechanical and Mechatronic Engineering, Stellenbosch University, South Africa b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 15 July 2008 Received in revised form 12 November 2008 Accepted 17 February 2009

Prolonged sitting and psychosocial factors have been associated with musculoskeletal symptoms among adolescents. However, the impact of prolonged static sitting on musculoskeletal pain among South African high school students is uncertain. A prospective observational study was performed to determine whether sitting postural alignment and psychosocial factors contribute to the development of upper quadrant musculoskeletal pain (UQMP) in grade ten high school students working on desktop computers. The sitting postural alignment, depression, anxiety and computer use of 104 asymptomatic students were measured at baseline. At three and six months post baseline, the prevalence of UQMP was determined. Twenty-seven students developed UQMP due to seated or computer-related activities. An extreme cervical angle (<34.75 or >43.95 ; OR 2.8; 95% CI: 1.1–7.3) and a combination of extreme cervical and thoracic angles (<63.1 or >71.1 ; OR 2.2; 95% CI: 1.1–5.6) were signiﬁcant postural risk factors for the development of UQMP. Boys with any extreme angle were more likely to suffer pain compared with boys with all middle range angles (OR 4.9; 95% CI: 1.0–24.5). No similar effect was found for girls. There was no strong relationship between depression, anxiety, computer exposure and UQMP among South African high school students. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Sitting posture Psychosocial factors Computer Musculoskeletal pain

1. Introduction The prevalence of self-reported upper quadrant musculoskeletal pain (UQMP) among adolescents is high (Niemi et al., 1996; Cho et al., 2003; Diepenmaat et al., 2006). Musculoskeletal pain experienced during adolescence may develop into chronic musculoskeletal pain syndromes that persist into adulthood (Brattberg, 2004; Stahl et al., 2004). The monthly prevalence rates for adolescent neck and shoulder pain are increasing and range between 11.5% and 29% (Diepenmaat et al., 2006; Straker et al., 2008a; Smith et al., 2008). Risk factors associated with adolescent UQMP are physical factors, psychosocial factors, gender and age (Prins et al., 2008). Almost 70% of high school students in the Western Cape metropole of South Africa, who are regularly using computers, suffer

* Corresponding author. Tel.: þ27 2 1872 8695; fax: þ27 2 1872 8694. E-mail address: [email protected] (Y. Brink). 1 Division of Physiotherapy, Department of Interdisciplinary Health Sciences, Stellenbosch University, South Africa, PO Box 19063, Tygerberg 7505, South Africa. 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.02.005

from musculoskeletal pain (Smith, 2007). Computers are widely available in developing countries and are used frequently for long time periods in schools and other settings (Straker and Pollock, 2005). Prolonged sitting while using the computer is associated with UQMP (Ramos et al., 2005; Auvinen et al., 2007; Smith et al., 2008). South African high school students who are using computers for about 9 h per week are signiﬁcantly at risk of developing neck pain (Smith et al., 2008). There is little knowledge of the relationship between UQMP and the sitting posture of computing adolescents (Straker et al., 2008b). Studies that investigated sitting postural alignment and muscle activation patterns of computing adolescents did not establish a relationship between musculoskeletal pain and sitting posture (Grieg et al., 2005; Straker et al., 2008c). However it is generally assumed that sitting with a neutral spinal posture will be beneﬁcial to the musculoskeletal structures and reduce musculoskeletal pain symptoms (Straker et al., 2008b). A recent review of the literature indicates that there is limited evidence about speciﬁc postural angles that could lead to musculoskeletal discomfort (Prins et al., 2008). The measurement tools used in the reviewed literature for assessing static sitting posture varied from

648

Y. Brink et al. / Manual Therapy 14 (2009) 647–653

direct observation and direct measurement of postural angles (Harris and Straker, 2000; Murphy et al., 2004) to self-reported questionnaires (Niemi et al., 1996; Cho et al., 2003; Ramos et al., 2005). Only Murphy et al. (2004) measured the angles of sitting posture and pain simultaneously. Although biomechanical measures are the preferred manner to report postural alignment, the results of the self-reported questionnaires showed signiﬁcant associations between prolonged static sitting posture and UQMP among high school students, irrespective of computer use or speciﬁc postural angles (Niemi et al., 1996; Cho et al., 2003; Ramos et al., 2005). The lack of conclusive evidence to support sitting posture as a risk factor for adolescent neck, shoulder and arm pain may be due to the limited research conducted in this ﬁeld of research and in particular also because of the poor methods used to assess sitting posture. It is prominent in the current literature that psychosocial factors especially depression, anxiety and psychosomatic symptoms are common associative factors that inﬂuence the prevalence of musculoskeletal pain in adolescents (Feldman et al., 2002; Brattberg, 2004; Diepenmaat et al., 2006). Egger et al. (1999) found that depressed girls and depressed boys had a 13 times and a 10 times greater chance of complaining of musculoskeletal pain respectively. Due to cross-sectional study designs, it is as yet unknown whether adolescents suffer musculoskeletal pain due to adverse psychosocial factors or the adverse psychosocial factors cause musculoskeletal pain. It is also unknown as to what extent psychosocial factors inﬂuence South African adolescents’ experience of musculoskeletal pain. There is a dearth of research that longitudinally establishes the causation of UQMP in adolescents. Cross-sectional study designs are not conducive to identify the physical and psychosocial risk factors that are predictive of musculoskeletal pain. The aim of this prospective study was to measure sitting postural alignment, anxiety and depression as the primary exposures and computer use as a secondary exposure in order to determine if these exposures are risk factors for the development of UQMP in computing high school students. 2. Methods 2.1. Ethical considerations Approval for the study was obtained from the Committee for Human Research of Stellenbosch University. Written informed consent was obtained from the Western Cape Education Department as well as from the child and his/her parents.

2.3. Measurement of postural and psychosocial exposure Sitting postural alignment was measured at baseline by using the Photographic Posture Analysis Method (PPAM). This measurement tool measures the sitting postural alignment of school students while they sit in front of desktop computers (Van Niekerk et al., 2008). This tool measures the following four postural angles in the sagittal plane and the angles are illustrated in Figs. 1 and 2. Head tilt – the angle between a line drawn from the canthus to the midpoint of the tragus and the horizontal line through the tragus (Raine and Twomey, 1997; Straker and Mekhora, 2000) Cervical angle – the angle between a line drawn from the midpoint of the tragus to the C7 spinous process (SP) and the horizontal line through the C7 SP (Raine and Twomey, 1997) Shoulder pro- and retraction angle – the angle between a line drawn from the midpoint of the humeral head to the C7 SP and the horizontal line through the midpoint of the humeral head (Raine and Twomey, 1997) Thoracic angle – the angle between a line drawn from the C7 SP to the midpoint of the superior border of the manubrium and a line drawn from the T8 SP to the midpoint of the superior border of the manubrium (Szeto et al., 2005). Reﬂective markers were placed on the student’s canthus, the midpoint of the tragus, the SP of C7 and T8, the superior border of the manubrium and the midpoint of the humeral head. Fig. 3 shows a student with the reﬂective markers in position. The validity and reliability of the PPAM were tested and illustrated excellent validity and reliability (Van Niekerk et al., 2008). Intellect Software was used to calculate the postural angles. Depression and anxiety were evaluated at baseline by using the standardised 21-item Beck Depression Inventory (BDI) (Beck and Steer, 1987) and the 39-item Multidimensional Anxiety Scale for Children (MASC) (March et al., 1997). The BDI has excellent screening measure properties among an adolescent school sample (Barrera and Garrison-Jones, 1988). The MASC demonstrates satisfactory to excellent reliability across age and gender and is sufﬁciently stable to use in a research setting (March and Sullivan, 1999). Both the BDI and the MASC have been used in a South African adolescent population (Seedat et al., 2004; Fincham et al., 2007). The BDI and MASC were used in a non-psychiatric population therefore only one score for each student for the respective BDI and MASC was used. 2.4. Computer usage and UQMP The CUQ was administered at baseline to evaluate computer exposure at school and elsewhere (Smith et al., submitted for

2.2. Participants A prospective cohort study was conducted over a six month period. Grade ten male and female (high school) students aged between 15 and 17 years, with no musculoskeletal pain at the onset of the study, were eligible to participate. Students who commenced with Computer Studies or Compu-typing as school subjects at the beginning of the 2007 academic school year were eligible as this was the ﬁrst time that these students were exposed to curriculum delivery via computers. Six high schools were randomly selected from 124 high schools with computer laboratories. Grade ten students were screened for musculoskeletal symptoms or discomfort by administering a component of the Computer Usage Questionnaire (CUQ) that dealt with the onset, area and intensity of musculoskeletal symptoms (Smith et al., submitted for publication). This instrument’s reliability and validity were previously tested in this population (Smith et al., submitted for publication).

Fig. 1. Head tilt and the thoracic angle.

Y. Brink et al. / Manual Therapy 14 (2009) 647–653

649

For correlation with other explanatory variables, each was considered in its continuous form: 2) To test an association with pain, each explanatory variable was divided into binary categories. Age, gender, depression and anxiety scores, and time spent at the computer were divided at the median to ensure equally balanced categories. The four postural angles were categorized differently however, into middle range, or extreme, angles. As it could be argued that both very small and very large angles could place an increased biomechanical load on surrounding tissues (Straker et al., 2008b), these were grouped into the one ‘at-risk’ category, and for analysis purposes, were compared with middle range angles. The extremely small or large angles were identiﬁed as data falling below the 25th or above the 75th percentiles for each angle. 2.7. Statistical analysis Fig. 2. Cervical angle and the shoulder pro- and retraction angle.

publication). Computer use was described by the duration per sitting session, the frequency of weekly usage and the amount of hours per week. UQMP during the preceding month was measured at three and six months, using the pain component of the CUQ. 2.5. Data collection procedures Students were instructed to sit in front of their computers as they do during computer classes. Two students were measured simultaneously and were instructed by the class teacher to perform a 10-min typing task (Szeto et al., 2002) while the postural measurements were taken. Three photographs were taken of which the ﬁrst photograph was taken after the student was settled behind the computer and the second photograph at 5 min. The ﬁrst two photographs were taken to accustom the students to the data collection procedure. A third photograph was taken at 10 min and the data of only the third photographs were analysed. The BDI, the MASC and the CUQ were then completed by the participants. Follow-up data collection took place at three and six months to measure the outcome of UQMP by administering the pain component of the CUQ. 2.6. Data organization UQMP was recorded in binary form (Yes, No). For analysis purposes, the explanatory variables were treated in two ways: 1)

Means, SD’s and percentiles were reported for each explanatory variable. Correlations between the explanatory variables were calculated using univariate regression models, and reported as Pearson r correlation coefﬁcients. Statistical signiﬁcance of these associations was reported as p < 0.05. Associations between the binary forms of the explanatory variables (postural angles, anxiety and depression), secondary exposure (computer use) and the outcome (UQMP) were calculated using univariate logistic regression models (SAS Version 9.1). The output was expressed as odds ratios (OR) and 95% conﬁdence intervals (CI). Although the sample was small, potential gender effects were considered for the significant pain-explanatory variable associations by using subgroup analysis. The signiﬁcance of these models was identiﬁed when the 95% CI did not encompass the value 1. 3. Results 3.1. Sample description at baseline, three and six months Fig. 4 illustrates the change in the sample composition from baseline until six months. The mean age for the sample at baseline was 16.0 (0.70) years. 3.2. Measurements at baseline 3.2.1. Baseline sitting postural alignment and psychosocial factors Table 1 presents the descriptive data for sitting posture and psychosocial factors measured at baseline. 3.2.2. Computer use Ninety-two of the 104 students also utilized computers outside of the school environment. Seventy-four percent and 46% of the students reported less than one year of computer use at school and elsewhere respectively. Seventy-four percent of the students reported 45 min of computer sessions at school and 72% reported 60 min per session of computer use elsewhere. Sixty-six percent and 33% of the students used the computer ﬁve times or more per week at school and elsewhere respectively. The boys reported 134 min (73.3) and the girls 99 min (49.3) computer use per day. The mean weekly computer usage was 7.6 h (5.06). The boys reported an average of 8.8 h (6.0) whereas the girls reported 6.2 h (3.2) per week computer usage. 3.3. Measurement of UQMP at three and six months

Fig. 3. The placement of the reﬂective markers on a student.

Fig. 5 presents 1) the numbers of students who developed UQMP that was related to seated activities such as sitting behind a computer or school desk; 2) the new cases at six months; 3) the

650

Y. Brink et al. / Manual Therapy 14 (2009) 647–653

Students screened for musculoskeletal pain

N=322

N=187

Students excluded after applying the inclusion and exclusion criteria

N=135

Eligible asymptomatic students invited to participate

N=109

Students returned informed consent

Two students were repeating grade ten, 162 students suffered from musculoskeletal pain and 23 students fell outside the age restriction.

N=5 Students absent from school

N=104 Students measured at baseline

N=6 Students absent from school

N=98 Students measured at three months

55 boys

N=5 Students absent from school

49 girls N=93 Students measured at six months

51 boys

47 girls 48 boys

45 girls Fig. 4. Flow chart to illustrate the sample composition.

students who suffered continuing pain and 4) the students who became pain free.

3.4. Identiﬁcation of the pain sufferers for analysis The students who developed UQMP at three months were grouped with the 13 new cases at six months to form a pain group of 27 students. The data analysis group consisted of the pain group

(n ¼ 27) and the no-pain group (n ¼ 67). Subsequent data analysis was performed on a group of 94 high school students. There was a cumulative pain incidence of 26% after six months, and the incidence rate was ﬁve new cases of UQMP per six months per 100 students using computers. There was no signiﬁcant difference in age between students with and without pain (mean age for the pain group 15.9 years (0.65), mean age for the no-pain group 16.0 years (0.73)). 3.5. Correlation between explanatory variables

Table 1 Descriptive values for the postural angles, anxiety and depression (n ¼ 104). Head tilt angle

Minimum 13.1 Maximum 34.7 Mean 13.78 SD 9.66 Median 14.65 75% Quartile 18.60 25% Quartile 6.90

Cervical angle

Shoulder pro- and retraction angle

Thoracic angle

Anxiety Depression

18.8 59.1 39.27 7.99 39.60 43.95 34.75

81.4 173.5 128.65 17.18 127.60 141.40 117.40

35.3 89.6 67.10 8.27 67.70 71.10 63.10

29 71 51 9.92 51.50 59.0 43.5

0 63 15 13.27 11.0 20.5 6.0

A number of signiﬁcant and strong correlations were identiﬁed between the explanatory variables. The ﬁndings are provided in Fig. 6. Weight was strongly and positively correlated with the shoulder pro- and retraction angle (r ¼ 22), suggesting that the heavier the child the more forward the shoulder posture. Weight on the other hand was strongly negatively associated with the other three angles (head tilt angle r ¼ 26, cervical angle r ¼ 24 and thoracic angle r ¼ 22%). This suggests that the heavier the child the more ﬂexed spinal postures are maintained. BMI showed similarly strong associations as weight, and in the same directions. Depression was negatively and strongly associated with shoulder pro- and

Y. Brink et al. / Manual Therapy 14 (2009) 647–653

14.0

3 months 6 months

Number of students

12.0 10.0 8.0 6.0 4.0

651

0.9–4.9), although this was not statistically signiﬁcant at p < 0.05. Individually, none of the extreme angles was signiﬁcantly associated with pain, for either boys or girls. However, comparing students who had any of the at-risk angles with students who had all middle range angles, boys had signiﬁcantly elevated risk that any extreme angle would be associated with pain (OR 4.9; 95% CI: 1.0– 24.5), compared with girls (OR 2.0; 95% CI: 0.5–8.9). There was no gender difference in the probability of suffering pain, when considering only those boys and girls with any extreme postural angle (OR 0.5; 95% CI: 0.2–1.5). 3.7. Other explanatory variables associated with UQMP after a six month period

2.0 0.0 girls

boys

continuing cases

new cases

painfree

Fig. 5. Survival analysis showing the number of students who developed UQMP.

retraction angle (r ¼ 22), suggesting that greater depression scores were related to more upright postures. Head tilt angle and cervical angle were strongly positively correlated (r ¼ 46), a ﬁnding that was not surprising as these have similar anatomical measurement points. The shoulder pro- and retraction angle was negatively correlated with the thoracic angle (r ¼ 21), suggesting that a more forward shoulder posture correlates with a more ﬂexed thoracic spine.

There was no signiﬁcant overall association between pain and anxiety, depression or computer use, or for gender subgroups. The results are shown in Table 2. 4. Discussion There is sparse literature providing evidence that postural alignment deviation leads to the development of musculoskeletal pain (Straker et al., 2008a). This prospective design contributes to the current knowledge by providing information about the temporality of musculoskeletal pain and upper quadrant postural alignment. The ﬁndings of the study indicate that there may be a causal link between postural alignment and UQMP. 4.1. Sitting postural alignment

3.6. Extreme posture associated with UQMP after a six month period The associations between UQMP with posture measures are presented in Table 2. As a group, the students with extreme (at-risk) cervical angles had signiﬁcantly elevated risk (OR 2.8) of developing UQMP (95% CI: 1.1–7.3). Moreover students with both extreme cervical and thoracic angles were at signiﬁcantly increased risk of developing UQMP (OR 2.2; 95% CI: 1.1–5.6). Neither shoulder pro- and retraction nor head tilt angle was associated with pain, either on its own, or in conjunction with other angles. Considering gender subgroups, there were indications that boys were at greater risk than girls for developing UQMP (OR 1.9; 95% CI:

Height Weight BMI Anxiety Depression Head tilt Cervical Shoulder proand retraction Thoracic

Height

Weight

BMI

Anxiety Depression

1

0.25

-0.18

-0.13

1

0.9 1

The study found that the only postural angles to be predictive of UQMP for computing high school students were the extreme cervical angle and the combination of the extreme cervical and thoracic angles. The postural angles were measured using the horizontal line as a reference. Therefore, as the cervical angle increases, the cervical spine is in a position of less ﬂexion. As the cervical angle decreases, the extensor moment around C7 increases and more isometric muscle activity is needed from the superﬁcial paraspinal extensors to counterbalance the gravitational moment (Briggs et al., 2004). The increased muscle activity could potentially lead to musculoskeletal pain or discomfort (Briggs et al., 2004; Grieg et al., 2005). When the cervical angle increases, the extensor moment is shortened, but muscle activity is still needed to stabilise

Cervical

-0.12

Head tilt

Shoulder pro- Thoracic Computer Computer and retraction minutes hours

0.04

-0.04

0.03

-0.08

-0.26

-0.24

0.22

-0.22

0.04

0.11

0.09

-0.04

-0.26

-0.22

0.22

-0.25

-0.03

0.006

1

0.11

0.13

0.03

0.03

-0.11

-0.06

0.11

1

0.005

0.006

-0.22

-0.11

-0.15

-0.06

1

0.46

0.14

-0.17

0.0009

0.1

1

0.02

-0.08

-0.04

-0.03

1

-0.21

-0.02

0.02

1

0.05

0.05

1

0.87

-0.02

0.11

Computer minutes Computer hours

0.19

0.26

1

Fig. 6. Graph to demonstrate the correlations between explanatory variables.

652

Y. Brink et al. / Manual Therapy 14 (2009) 647–653

Table 2 The association of UQMP by gender, postural angles, anxiety, depression and computer use.

Gender Height Weight BMI Head tilt Cervical Shoulder pro- and retraction Thoracic Anxiety Depression Computer use (minutes per day) Computer use (hours per week) Extreme cervical and thoracic

Boys Girls 1.63 54.2 20.4 6.9 or 18.6 34.75 or 43.95 117.4 or 141.4 63.1 or 71.1 51.5 11 105 min per day 6 h per week

Group OR

Group CI

Boys OR

Boys CI

Girls OR

Girls CI

1.9* 0.5 1.3 0.6 0.5 0.9 2.8** 1.1 2.3* 1.4 0.8 1.7 1.6 2.2**

0.9–4.9 0.2–1.3 0.5–3.1 0.3–1.5 0.2–1.3 0.3–2.5 1.1–7.3 0.4–2.9 0.8–6.3 0.6–3.3 0.4–1.9 0.7–4.2 0.7–3.8 1.1–5.6

0.9 2.6* 2.2 1.5 1.41 0.65 1.0 0.9 2.56*

0.3–3.8 0.8–9.1 0.6–8.7 0.5–5.2 0.4–4.7 0.2–2.5 0.3–3.4 0.3–2.8 0.7–8.9

0.7 3.1 0.4 2.4 1.87 1.0 0.5 1.8 1.5

0.1–4.0 0.7–13.7 0.1–3.3 0.5–11.8 0.4–8.4 0.2–4.0 0.1–4.3 0.4–8.7 0.3–7.2

The OR in bold print marked with a (**) shows signiﬁcance at the 95% conﬁdence interval. The OR in bold print marked with a (*) shows a tendency to reach signiﬁcance at the 95% conﬁdence interval.

the cervical spine in order to maintain an erect posture. With a greater cervical angle, the length-tension relationship of the deep suboccipital muscles is not optimal and this places more stress on the upper cervical zygapophyseal and intervertebral joints and can lead to musculoskeletal pain or discomfort (Briggs et al., 2004). It appears that postures with prolonged extreme cervical angles place repetitive or prolonged mechanical stress on neuromusculoskeletal structures and therefore these structures can become symptomatic. The ﬁndings from this study could be compared with studies by Straker et al. (2002) and Briggs et al. (2004) because these studies were consistent in the deﬁnition for head tilt and cervical angle and similar to the deﬁnition used in this study for the same age group. We found no difference between the head tilt angles of the high school students experiencing pain and those experiencing no pain. Szeto et al. (2002) reported similar results among the adult population and found that the head tilt angle did not differ between asymptomatic and symptomatic adult ofﬁce workers. This could possibly be because the head tilt angle reﬂects the position of the head segment when viewing the computer display and the height of the computer display did not differ in this study (Szeto et al., 2002). An above average weight was also associated with a speciﬁc postural alignment. This study found that overweight students sit with a pronounced ﬂexed posture (more upper and lower cervical and thoracic ﬂexion and a more protracted shoulder position). No other study could be found to support or contradict this result as little is known about the inﬂuence of weight on the sitting postural alignment of adolescents. 4.2. Depression and anxiety No causal relationship was found between anxiety or depression and the development of UQMP among high school students who work on desktop computers. The ﬁndings of this study contradicts the ﬁndings of Egger et al. (1999), Feldman et al. (2002) and Diepenmaat et al. (2006), who found a positive association between neck pain and high levels of depression or anxiety/stress. A study conducted among grade ten students in the Western Cape of South Africa found that adolescents report on average 3.5 childhood traumatic events and, as a result have difﬁculty expressing their emotions compared to adolescents from developed countries (Suliman et al., 2005). It could be that high levels of depression affected the South African adolescents differently than the adolescents from other developed countries and consequently we speculate that more depressed feelings might have suppressed the student’s expression of pain symptoms. According to the pain coping proﬁles for adolescents with chronic pain by Claar et al.

(2008), this study’s symptomatic high school students fall into the ‘‘infrequent copers’’ classiﬁcation. This classiﬁcation type do not use any other coping strategy assessed by the Pain Response Inventory and consequently these adolescents’ low levels of pain correlates with low levels of depression or anxiety. 4.3. Computer use In this study the pain group and the no-pain group had similar years of exposure to computers, with both groups containing more than 60% of students with less than one year’s exposure to computers. Research has shown that, as the years of computer use increases, the prevalence of musculoskeletal discomfort among adolescents also increases (Harris and Straker, 2000; Ramos et al., 2005). It is possible that there was no association found between computer usage and UQMP because the students in this study reported limited years (average 1.8 years) of exposure to computers. 4.4. Gender In this study, twice as many boys than girls reported pain even though there was almost an equal distribution of girls and boys at inception of the study. This contradicts previous research that found adolescent girls to have a signiﬁcantly greater risk for neck/ shoulder pain (Stahl et al., 2004; Diepenmaat et al., 2006). This greater risk might be because boys with any extreme angle had a signiﬁcantly elevated risk to develop pain (OR 4.9; 95% CI: 1.0– 24.5) compared to girls. The boys also demonstrated more daily and weekly computer use than their female counterparts. A study by Auvinen et al. (2007) reported that prolonged sitting due to computer work is associated with neck and occipital pain only among boys and not girls. 5. Conclusion Posture may be a cause of UQMP amid high school students. Having an extreme cervical angle or a combination of extreme cervical and thoracic angles may be associated with UQMP among high school students using desktop computers. Further research is needed to produce stronger evidence that shows whether postural alignment and psychosocial factors are risk factors for QUMP. The outcome of this research will enable clinicians to appropriately manage adolescent UQMP. Conﬂict of interest None.

Y. Brink et al. / Manual Therapy 14 (2009) 647–653

Acknowledgements This research was ﬁnancially supported by The South African Medical Research Council and Stellenbosch University Research Fund. We also acknowledge support from the WCED and the Khanya school computer project.

References Auvinen BM, Tammelin T, Taimela S, Zitting P, Karppinen J. Neck and shoulder pains in relation to physical activity and sedentary activities in adolescence. Spine 2007;32:1038–44. Barrera MJ, Garrison-Jones CY. Properties of the beck depression inventory as a screening instrument for adolescent depression. Journal of Abnormal Psychology 1988;16:263–73. Beck AT, Steer RA. Beck depression inventory manual. San Antonio, TX: Psychological Corp; 1987. Brattberg G. Do pain problems in young school children persist into early adulthood? A 13-year follow-up. European Journal of Pain 2004;8:187–99. Briggs A, Straker L, Grieg A. Upper quadrant postural changes of school children in response to interaction with different information technologies. Ergonomics 2004;47:790–819. Cho CY, Hwang IS, Chen CC. The association between psychosocial distress and musculoskeletal symptoms experienced by Chinese high school students. The Journal of Orthopaedic and Sports Physical Therapy 2003;33:344–53. Claar RL, Baber KF, Simons LE, Logan DE, Walker LS. Pain coping proﬁles in adolescents with chronic pain. Pain 2008;. doi:10.1016/j.pain.2008.09.007. Diepenmaat ACM, Van der Wal MFV, De Vet HCW. Neck/shoulder, low back, and arm pain in relation to computer use, physical activity, stress, and depression among Dutch adolescents. Pediatrics 2006;117:412–6. Egger HL, Costello EJ, Erkanli A, Angold A. Somatic complaints and psychopathology in children and adolescents: stomach aches, musculoskeletal pains and headaches. The Journal of American Academy of Child and Adolescent Psychiatry 1999;38:852–60. Feldman DE, Shrier I, Rossignol M, Abenhaim L. Risk factors for the development of neck and upper limb pain in adolescents. Spine 2002;27:523–8. Fincham D, Schickerling J, Temane M, Nel D, De Roover W, Seedat S. Exploratory and conﬁrmatory factor analysis of the Multidimensional Anxiety Scale for Children (MASC) among adolescents in the Cape Town metropole of South Africa. Depression and Anxiety Published online 14 November 2007. Grieg AM, Straker LM, Briggs AM. Cervical erector spinae and upper trapezius muscle activity in children using different information technologies. Physiotherapy 2005;91:119–26. Harris C, Straker L. Survey of physical ergonomics issues associated with school children’s use of laptop computers. Industrial Ergonomics 2000;26:337–46. March JS, Sullivan K. Test-retest reliability of the multidimensional anxiety scale for children. Journal of Anxiety Disorders 1999;13:349–58. March J, Parker J, Sullivan K, Stallings P, Conners C. The Multidimensional Anxiety Scale for Children (MASC): factor structure, reliability and validity. Journal of the American Academy of Child and Adolescent Psychiatry 1997;36:554–65. Murphy S, Buckle P, Stubbs D. Classroom posture and self-reported back and neck pain in school children. Applied Ergonomics 2004;35:113–20.

653

Niemi SM, Levoska S, Kemila J, Rekola K, Keinanen-Kiukaanniemi SM. Neck and shoulder symptoms and leisure time activities in high school students. The Journal of Orthopaedic Sports and Physical Therapy 1996;24:25–9. Prins Y, Crous L, Louw QA. A systematic review of posture and psychosocial factors as contributors to upper quadrant musculoskeletal pain in children and adolescents. Physiotherapy Theory and Practice 2008;24(4):221–42. Raine S, Twomey LT. Head and shoulder posture variations in 160 asymptomatic women and men. Archives of Physical Medicine and Rehabilitation 1997;78:1215–23. Ramos EMA, James CA, Bear-Lehman J. Children’s computer usage: are they at risk of developing repetitive strain injury? Work 2005;25:143–54. Seedat S, Nyamai N, Njenga F, Vythilingum B, Stein D. Trauma exposure and posttraumatic stress symptoms in urban African schools. British Journal of Psychiatry 2004;184:169–75. Smith L. Computer-related musculoskeletal dysfunction among adolescent school learners in the Cape Metropolitan region. Unpublished M thesis. Tygerberg: Stellenbosch University; 2007. Smith L, Louw QA, Crous LC, Grimmer-Somers K. Prevalence of neck pain and headaches: impact of computer use and other associative factors. Cephalalgia 2008; 29:250–7. Smith L, Louw QA, Crous L, Grimmer-Somers K. Assessing adolescent musculoskeletal dysfunction associated with computer use: development and testing of the computer usage questionnaire. Work Submitted for publication. Stahl M, Mikkelsson M, Kautiainen H, Hakkinen A, Ylinen J, Salminen JJ. Neck pain in adolescence. A 4-year follow up of pain-free preadolescents. Pain 2004;110:427–31. Straker L, Mekhora K. An evaluation of visual display unit placement by electromyography, posture, discomfort and preference. Industrial Ergonomics 2000;26:389–98. Straker L, Pollock C. Optimizing the interaction of children with information and communication technologies. Ergonomics 2005;48:506–21. Straker L, Briggs A, Grieg A. The effect of individually adjusted workstations on upper quadrant muscle activity in school children. Work 2002;18:239–48. Straker LM, O’Sullivan PB, Smith AJ, Perry MC, Coleman J. Sitting spinal posture in adolescents differs between genders, but is not clearly related to neck/shoulder pain: an observational study. Australian Journal of Physiotherapy 2008a;54:127–33. Straker L, Burgess-Limerick R, Pollock C, Murray K, Netto K, Coleman J, Skoss R. The impact of computer display height and desk design on 3D posture during information technology work by young adults. Journal of Electromyography and Kinesiology 2008b;18:336–49. Straker LM, Burgess-Limerick R, Pollock C, Maslen B. The effect of forearm support on children’s head, neck and upper limb posture and muscle activity during computer use. Journal of Electromyography and Kinesiology 2008c;. doi:10.1016/j.jelekin.2008.02.003. Suliman S, Kaminer D, Seedat S, Stein DJ. Assessing post-traumatic stress disorder in South African adolescents: using the child and adolescent trauma survey (CATS) as a screen tool. Annals of General Psychiatry 2005;4:2. Szeto GPY, Straker L, Raine S. A ﬁeld comparison of neck and shoulder postures in symptomatic and asymptomatic ofﬁce workers. Applied Ergonomics 2002;33:75–84. Szeto GPY, Straker L, O’Sullivan PB. A comparison of symptomatic and asymptomatic ofﬁce workers performing monotonous keyboard work-2: neck and shoulder kinematics. Manual Therapy 2005;10:281–91. Van Niekerk S, Louw QA, Vaughn C, Grimmer-Somers K, Schreve K. Photographic measurement of upper-body sitting posture of high school students: a reliability and validity study. BMC Musculoskeletal Disorders 2008;9:113.

Manual Therapy 14 (2009) 654–660

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

Informed consent practices of physiotherapists in the treatment of low back pain Anne Fenety a, *, Katherine Harman a, d, Alison Hoens b, c, Raewyn Bassett d, e a

School of Physiotherapy, Dalhousie University, 5869 University Avenue, Halifax, Nova Scotia, B3H 3J5, Canada Department of Physical Therapy, University of British Columbia, 212 - 2177 Wesbrook Mall, Vancouver, British Columbia V6T 1Z3, Canada c Providence Health Care, 3080 Prince Edward St, Vancouver, British Columbia V5T 3N4, Canada d Faculty of Health Professions, Dalhousie University, 5968 College Street, Halifax, Nova Scotia B3H 3J5, Canada e Faculty of Graduate Studies: 6299 South Street, Halifax, Nova Scotia B3H 4H6, Canada b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 1 May 2008 Received in revised form 5 February 2009 Accepted 17 February 2009

Our purpose was to explore and describe physiotherapists’ informed consent practices in the treatment of clients with low back pain. Forty-four physiotherapists were assigned to six focus group interviews. Focus group interaction elicits insights that are less accessible in individual interviews and which can be corroborated immediately through inbuilt checks and balances. Participating physiotherapists described not only fulﬁlling but also exceeding their regulatory and ethical duty to obtain explicit and implicit informed consent from clients according to professional guidelines. Client autonomy could not always take precedence in the fast-paced and seamless therapy session. A shared decision-making process of embodied, implicit consent or refusal was embedded in a reciprocal client–therapist care relationship of trust and rapport. A typology of modes of consent is provided. A process for obtaining a more explicit consent alongside implicit consent that does not interrupt the continuity of physiotherapy treatment is outlined. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Informed consent Embodied consent Qualitative methods Low back pain

1. Introduction Health professionals have a legal and moral duty to obtain informed consent from their clients, while health care clients have the right and responsibility to give or refuse fully informed, coercion-free consent. To make an informed decision about treatment, clients trust health professionals to understand, to fully disclose, and to help them meaningfully understand the facts necessary to the decision (Strom-Gottfried, 1998; O’Neill, 2003). In Canada, the Health Care Consent Act (1996) mandated the legal obligation of health practitioners to obtain consent from clients prior to treatment. This legal obligation is incorporated in most provincial physiotherapy licensing boards’ codes of ethics and in the Essential Competency Proﬁle for Physiotherapists in Canada (2004), establishing clients’ rights to be involved in decision-making about proposed treatments. While these regulatory/academic standards are region speciﬁc, the broader concepts of informed consent cross professional and geographical boundaries. Informed consent is deﬁned as a process of client decisionmaking based on information delivered by the treating physiotherapist, adapted to the client’s needs for understanding,

* Corresponding author. Tel.: þ1 902 494 2634; fax: þ1 902 494 1941. E-mail address: [email protected] (A. Fenety). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.02.007

including a description of treatment, risks and alternatives, and the consequences of refusing treatment (College of Physiotherapists of Ontario, 2000). The College goes on to suggest that ﬁrst, client’s questions are to be answered by the physiotherapist and their decision respected, second, consent, refusal, or revoking of consent during treatment must be voluntary, and treatment must not proceed before obtaining consent orally, in writing or by implication from the client’s actions, words or circumstances, and ﬁnally, the informed consent is presumed to cover continuation, variations, and adjustments to treatment in the same or another setting (College of Physiotherapists of Ontario, 2000). Much has been written about the need to obtain consent, but very little describes how physiotherapists actually do so. We invited physiotherapists to focus group interviews to discuss their treatment and informed consent practices with clients with low back pain. Informed consent was deﬁned according to local regulatory practice guidelines (Fig. 1). This paper focuses on how participants informed clients of proposed treatment plans and how consent was gauged. 2. Literature review In the post-war decades, informed consent to treatment was enshrined legally and ethically, transforming a health care system from paternalistic to client-centred decision-making (Purtilo,

A. Fenety et al. / Manual Therapy 14 (2009) 654–660

655

Nova Scotia College of Physiotherapists Practice Guidelines (1999)

The client or their proxy must be mentally competent have the legal capacity to give consent The information about the client’s proposed treatment must include a description of the treatment, why it is proposed & how it will be delivered the risks, treatment alternatives & consequences of refusal any personal commitment/financial costs to the client Consent will be informed voluntarily given by the client respected by the therapist expressed in writing, orally, or by implication from actions, words and circumstances presumed to apply to continuation & variation of treatment Refusal will be informed voluntarily given by the client respected by the therapist documented, including that the consequences have been explained Exceptions to client’s informed consent includes emergency care

Fig. 1. Key elements of local regulatory practice guidelines.

2000). In the past, health professionals made health care decisions for clients without the client’s consent (Elkin, 2001). Medical paternalism was accepted by both health professionals and clients, based on the expertise, knowledge and authority invested in the health professional. Mid-last century, collective actions such as the women’s movement, the civil rights movement, and the patients’ rights movement (Elkin, 2001; Gabard and Martin, 2003), and the discovery of unethical research studies such as the Tuskagee syphilis study in the United States (Jones, 1993) and ‘‘The unfortunate experiment’’ in cervical screening in New Zealand (Coney and Bunkle, 1987), prompted change from medical paternalism to client autonomy in health care decision-making. Client autonomy, too, has increasingly moved away from merely gaining consent from the client at one point in time to a focus on the quality and comprehensiveness of information disclosed in an ongoing dialogue between health professional and client (Sim, 1997). 2.1. The client consent process Client autonomy in decision-making as the antidote to autocratic paternalism, ignores the psycho-social context within which decision-making occurs (Corrigan, 2003). It has been pointed out that clients’ emotional and cognitive states, physical pain, anxiety levels and educational background affect the absorption, selection, recall and integration of information (Delaney, 1996, 2002; Doyal, 2001; Everett et al., 2005). Clients’ decisions are also inﬂuenced by health care routines, situational exigencies such as pain levels and drugs, and perceived sanctions for questioning or responding negatively to requests to consent (Dawes et al., 1993; Corrigan,

2003; Dixon-Woods et al., 2006). Further, clients want to be wellinformed but may not wish to be involved in decisions about treatment, preferring to take advice on trust (Waterworth and Luker, 1990; Dawes et al., 1993; Charles et al., 1997). Clients trust health professionals to act in their best interests (Corrigan, 2003); it is often trust rather than the information disclosed that determines their consent (Kent, 1996). 2.2. The health professional consent process Assent or dissent in the informed consent process rests upon the health professional’s disclosure of information, tailored to the client’s needs for understanding (Delaney, 1996). Disclosure includes information other professionals in the same circumstance would disclose (professional standard), information a reasonable person would want to know (prudent patient standard), as well as what the client would consider necessary to informed decision-making (subjective/dialogic standards) (Sim, 1997; Swisher and KruegerBrophy, 1998; Gabard and Martin, 2003). Beyond content, information must be comprehensively disclosed; the client must understand it in order to use it in decision-making (Sim, 1997). There is uncertainty surrounding how to obtain informed consent in clinical practices such as physiotherapy where assessment and treatment are seamless (Delaney, 2002) and fast-paced (Carlesso et al., 2007), and consent is readily given. Given this uncertainty, a lapse to medical paternalism (Gabard and Martin, 2003) may be practised by some health professionals ‘for the client’s own good’ (Delaney, 1996, 2007; Aveyard, 2005). Informed consent is implicitly entwined in intuitive and routine clinical practices

656

A. Fenety et al. / Manual Therapy 14 (2009) 654–660

characterised by a tension between giving clients choices and providing an efﬁcient, professional service (Aveyard, 2005; Delaney, 2007). Where client compliance or active participation is required, additional information may be provided, often overwhelming clients, contributing to anxiety about procedures, and limiting genuinely informed consent (Kent, 1996; O’Neill, 2003). Aveyard (2005) found that even with a client’s refusal, an intervention might be continued if it was clinically indicated; Delaney (2007) found that therapists sometimes provided information about the presumed beneﬁcial therapeutic outcomes they determined as in the client’s best interests, overshadowing client autonomy in decision-making. 2.3. Shared decision-making More recently, the concept of shared, informed decisionmaking has been suggested, yet what constitutes shared decisionmaking remains unclear. Proposed as a mechanism to decrease asymmetrical knowledge and power between health professional and client, the mere transfer of information to clients to correct a knowledge deﬁcit will not lead to shared, informed decisionmaking (Charles et al., 1997; Elkin, 2001; Swisher and KruegerBrophy, 1998). Treatment decisions are often fraught with uncertainty (e.g. lack of clear evidence or several competing treatment options), and providing that information to clients may evoke unnecessary anxiety (Elwyn et al., 1999). Shared decisionmaking includes therapist and client together sharing treatment alternatives and client preferences to reach a decision both agree is the best option, made with the information before them (Charles et al., 1997). Of note is that the Latin ‘‘consentire’’ from which the word ‘‘consent’’ is derived, means ‘‘to feel or think together’’ (Elkin, 2001, p. 103). Physiotherapists receive pre-licensure training on how to obtain consent in busy, dynamic clinical practices, yet professional guidelines have remained relatively silent on this topic. Practice guidelines indicate who should provide information to whom, but not how to do so. Where previously, consent was readily given and therefore assumed by physiotherapists, inserting meaningful informed consent into fast-paced sessions can be problematic. Physiotherapy literature on informed consent is replete with the importance of clear, unbiased information delivery, tailored to clients’ needs for meaningful, informed decision-making, emphasising a duty by physiotherapists to ensure clients understand the information. Written or verbal information must include proposed treatments and risks, including risks of refusal to proceed with treatment (Delaney, 1996; Sim, 1996; Gafni et al., 1998; Nova Scotia College of Physiotherapists, 1999). The literature is explicit about the importance of the ethical and legal duty to obtain consent that is informed, and various informed consent models are described. Sim (1996, 1997) deﬁnes three forms of consent: express consent is given orally or in writing, as assent or dissent; implicit consent is not speciﬁcally indicated by the client, but is implied in a performed action; and tacit consent is presumed from a failure to dissent. However, Sim (1996, 1997), as with much of the literature, does not speciﬁcally address how physiotherapists actually obtain and reafﬁrm informed consent in the continuous and seamless process of assessment and treatment in physiotherapy practice. Within a single session and over the course of treatment, physiotherapists make multiple treatment decisions. Perhaps cognisant of this dilemma, in a study almost two decades ago, physiotherapists indicated their concern that obtaining informed consent would interrupt the treatment ﬂow (Grant and Trott, 1991). Mere compliance with the legal duty to obtain consent has been cautioned as a mechanistically clinical approach (Delaney, 2005), non-collaborative, and event- rather than process-based (Sin, 2005).

Recognising the client’s autonomy and right to make decisions throughout the clinical encounter is a requisite of informed consent (Delaney, 2005). Cognisant of the complexities of dynamic clinical practices and procedures, this research addressed the following question: ‘‘How do physiotherapists obtain informed consent in the fast-paced, seamless assessment and treatment of low back pain?’’ 3. Methods Physiotherapists’ treatment practices with clients with low back pain were explored in a focus group study by a team of researchers consisting of three physiotherapists (AF, KH, AH) and one sociologist (RB). Following ethics reviews by the authors’ institutions, focus group interviews were held in the Fall of 2006. The interaction in focus group interviews elicits insights that are less accessible in individual interviews (Ulin et al., 2005), and the variety of views and insights can be veriﬁed immediately through inbuilt checks and balances such as probing of responses and dialogue between participants (Grbich, 1999). Notices about the two-hour focus group interviews were sent to private clinics treating clients primarily with musculoskeletal complaints, with and without insurance coverage, inviting interested physiotherapists to a discussion of their treatment practices. This was followed by a presentation on current evidence for physiotherapeutic intervention of low back pain, an incentive for therapists to volunteer to participate. The interviews were conducted over one month in urban and rural settings. The participants were 44 physiotherapists (36F; 8M), with a mean of 17.5 years of experience (range 0.5–38 years). Each participant was assigned to one of six focus groups conducted over a one month period in geographically dispersed areas in the region. Informed consent was provided by participants at the beginning of the audio-taped interview. The three lead investigators (KH, AF, AH) conducted the focus groups using standardised, structured questions. After the local regulatory guidelines on Informed Consent were provided (Fig. 1), participants were asked if they regularly obtained and recorded informed consent and how and when. Transcribed focus group data were imported into the qualitative software, ATLAS/ti. Initial code development was undertaken by RB. In an inductive process, themes that described, organised and interpreted the participants’ responses were identiﬁed by examining their words and phrases. In an iterative process, a label or code to assign meaning, drawing from participants’ words, was applied to selected text, and codes were clustered, compared and sorted into sufﬁciently distinct and comprehensive themes. This subjective, interpretive process involved the repetitive execution of the same sequence of instructions (code / cluster / compare / sort) multiple times (Boyatzis, 1998; Bassett et al., 2008). For example, focus group participants often mentioned the importance of rapport with the client in the informed consent process. Sections of text where this aspect was mentioned were assigned to a code labelled ‘rapport’. When participants’ responses had been assigned to sufﬁcient codes, each code was clustered according to themes. The code, ‘rapport’, for example, and the code, ‘trust’, were clustered as characteristics of the theme, ‘implied consent’. Themes were compared and sorted in a process of constant comparison to ensure a clear distinction between them (Corbin and Strauss, 2008). Code and theme development were discussed, adjusted where necessary, and consensus gained during frequent research team meetings (Bassett and Graham, 2007). We describe below a typology of modes of consent developed from the analysis. 4. Results In this section, the word, ‘participant’ is used to refer speciﬁcally to the physiotherapists participating in the focus groups, and

A. Fenety et al. / Manual Therapy 14 (2009) 654–660

‘physiotherapist’ to refer to physiotherapists more generally. A brief overview of the focus group ﬁndings on informed consent is provided, followed by the modes of consent used by participants. 4.1. Overview Participants noted the importance of client consent to physiotherapy procedures and how consent was recognised. In keeping with the local regulatory guidelines, participants informed the client of the diagnosis, prognosis, potential treatments, therapeutic plan, and risks and precautions, and obtained a general, blanket consent to any further treatment. Following this initial consent, participants sometimes documented informed consent for speciﬁc treatment but it was most often obtained orally or by implication. Participants described various modes of consent, ranging from blanket consent, documented, verbal, implied as well as embodied consent, and compliance. 4.2. Modes of consent Blanket consent. Consent was sought at the beginning of treatment, and in accordance with provincial guidelines was considered by many physiotherapists as the only consent required. Blanket consent was understood to cover all treatments and evaluations for the client’s treatment period in relation to their current diagnosis, as the following participant states. Can I just mention that the procedure that I use with informed consent is done right when the client comes initially? They are consenting for evaluation and treatment. And then from there, it’s just an individual discussion about treatment with them. Documented consent. At their ﬁrst visit, clients sign a standard form giving consent to a general range of procedures and techniques that could be provided to treat their low back pain. Although local regulatory guidelines did not require further consent from the client for treatment, some participants said they obtained consent as treatment proceeded and documented their consent discussions with clients using acronyms. For example, some used PCIC for high-risk procedures, which indicated that Precautions-and-Contraindications-andInformed-Consent were discussed and obtained. Consent for low-risk treatment was seldom documented. Many participants stated that a client’s consent to treatment was written up most often only for treatment considered risky, such as cervical manipulation. I think the riskier the intervention, for example cervical manipulation, that is one place where you want to cross your ‘t’s and dot your ‘i’s and maybe re-do a neuro check and do the special testing to be done. Then document that, absolutely. And the response to it. Verbal consent. Consent was obtained orally by many participants and, as with documented consent, pertained largely to the blanket consent given at the ﬁrst visit. Diagnosis, treatment proposed, and risks involved were explained, after which the participant conﬁrmed that the client was ready to proceed. Oh, I seek consent for any treatment basically but it is. I explain what the treatment is going to be, and, ‘‘Are you okay with this?’’ ‘‘Yes.’’ ‘‘Okay,’’ and then I proceed. Not ‘‘do I have your consent for the treatment’’ speciﬁcally. However, some participants said they may not inform the client of risks nor explicitly ask for permission to proceed with

657

treatment because to do so might confuse or frighten the client into leaving the clinic. When the focus group facilitator queried whether participants asked the client, ‘‘do I have your verbal consent to proceed?’’ they responded: Participant 1. No, if we did that, we would have very confused patients. They would be thinking ‘‘how bad is this going to be?’’ and it would scare them away. Participant 2. Like giving the risks and all that stuff? No, not necessarily. It was at the participant’s discretion that information was disclosed, based on evaluation of clients’ needs which might include their need for treatment. Implied consent. Implied consent consisted of two components. First, consent to further treatment was implied when the client signed or agreed orally to the general, blanket terms presented to them in the initial informed consent process upon their arrival at the clinic. I thought when they were coming to physio and signing like the . the one document that it is implied when we are treating them. Second, after explaining the treatment to be provided, participants considered it the responsibility of the client to explicitly refuse if they did not agree with the treatment option. I would probably say like, ‘‘I am going to just push on this joint,’’ or whatever but I wouldn’t ask the question, ‘‘is it alright?’’ I would probably wait for them to say no, to stop, or whatever. The responsibility was considered that of the client to opt-out of treatment. If they did not, consent was implied. This form of consent occurred more often with low-risk treatment techniques such as a joint mobilization or a muscle stretch. Rapport developed between client and physiotherapist was the basis of implied consent. Participants described the importance of the connection they develop with clients in the therapeutic relationship. Trust was important to this relationship. As one participant said, ‘‘You build rapport with them after understanding their situation and knowing where to go with it’’. It was taken-forgranted that with such rapport clients would stop the procedure if they were uncomfortable with it. A standing order that you can stop me at any time. Sort of that rapport, that anything that I do that you are not comfortable with, stop me at any time. I think that is an understanding that you have with your clients. Resistance or protest signiﬁed the active revoking of consent. Its absence was taken as active consent to treatment. Embodied consent. The client’s body language was assessed by the participant for consent to treatment, both prior to and while treatment was in progress. Trust and rapport between therapist and client extended to the body’s response to treatment. This, too, was considered an active, implied consent by focus group participants. And not just absolutely verbal but our rapport with our patients, and watching their body language. The embodied nature of consent is central to the treatment process and to physiotherapy practice as a whole. Compliance. Where rapport and trust are readily given, information-giving by the therapist may be absent. Participants described not always providing sufﬁcient information to the client. Without

658

A. Fenety et al. / Manual Therapy 14 (2009) 654–660

information from the physiotherapist about procedures and likely consequences, consent, informed or implied, cannot be given or rescinded by the client. Compliance replaces consent. A lot of times I think what they are going to get but I don’t tell them. And maybe I don’t take enough time, I think, to involve the patient in terms of what we expect to see over a course of time. I have my impressions but maybe I don’t take the time to explain to them what we are trying to achieve. 5. Discussion This study highlights two aspects of physiotherapy practice not previously described in the literature: (1) how informed consent is obtained in a speciﬁc practice context; and (2) a typology of the different modes of consent given and obtained in a collaborative client–physiotherapist relationship. The strength of this qualitative approach was the discovery of the implicit aspects of the informed consent process (Grant, 2005) described by physiotherapists. Sim (1996, 1997) has previously described three types of consent: express, implied, and tacit consent. In our study, the modes ‘documented’ and ‘verbal’ consent are equivalent to Sim’s ‘express’ consent. Participants in our study, too, discussed implied consent, however, our departure from Sim’s (1997, p. 68) typology of implied consent occurs in relation to his conceptualisation of tacit consent. Sim deﬁnes tacit consent as ‘‘the individual’s failure to dissent’’. Yet, participants in our study did not ﬁnd that clients failed to dissent. On the contrary, their non-verbal body language provided active cues of assent or dissent which were interpreted by physiotherapists as implying consent or refusal. The body language of embodied consent was central to physiotherapy practice. This study found that participants described consent as a routine, implicit, and intertwined component of clinical practice. Physiotherapists in focus group interviews were asked about informed consent in the context of their treatment practices. This allowed the development of a typology of explicit and implicit consent, and recognition of the role of embodied consent/refusal in client–physiotherapist interactions. We conclude that collaborative decision-making processes occur, albeit processes that could be improved. Following the discussion of results and study limitations, we outline an ongoing, collaborative client–therapist informed consent process. When compared against professional guidelines, study participants met their legal duty to obtain informed consent from clients. Most participants obtained written or verbal, blanket consent when clients arrived at the clinic, and assessed implied consent predominantly through the absence of resistance or protest as treatment progressed. Consent was not a one-time event but a process interwoven with multiple procedures and decisions throughout treatment. Participants described obtaining implicit consent using visual and kinaesthetic cues between themselves and the client. The client’s posture, movements and facial expression implied consent or its withdrawal. A ﬁne line exists between implied consent and compliance. Indeed, there is an element of complying, a ‘giving up of oneself’ (Waterworth and Luker, 1990) required from the client for many physiotherapy treatments. Trust between client and physiotherapist is essential to this process, described by participants as rapport, whereby a sign from the client, embodied as a tightening of a muscle, a bodily refusal to comply, or a grimace, was recognised and respected by the therapist, and treatment paused or ceased. Clients trust both physiotherapists’ skills and knowledge, and that treatment decisions are made in their best interests. A reciprocal trust is embedded in the care taken by the therapist not to overwhelm the client with too much information which might frighten them

and interfere with the ‘giving up of one’s self’ necessary for treatment. Caution is necessary to avoid erring towards too little information, with beneﬁcial therapeutic outcomes prioritised over client autonomy (Delaney, 2007). Shared decision-making in the informed consent process involves collaboration whereby clients have time to think about information and make reasoned decisions (Charles et al., 1997). As participants pointed out, in physiotherapy practice the immediacy of treatment leaves little or no time for such client decisions. Most procedures are fast-paced and continuous, and treatment could not occur were the therapist to stop and seek verbal consent for every detail. Time constraints in private practice also inﬂuence the mode of consent obtained. Implied consent in which the response of the client’s body is observed by the therapist while simultaneous ongoing dialogue informs the client of treatment procedures as they take place provide opportunities for consent or refusal. The ﬁndings of this study suggest that consent registered at the level of the body is an active form of consent. It is the presence (not absence) of information and active (not passive) authorization of consent that distinguishes implied consent from compliance. Asking clients for verbal afﬁrmation to proceed rather than relying solely on embodied consent would protect both client and therapist from potential allegations of deception or coercion. Knowing that consent is rescindable at any point in a procedure, stating aloud the standing order, telling the client they may refuse treatment, and that the therapist will observe bodily signs in the absence of voiced refusal or consent can lead to a more collaborative consent process. Evident in the ﬁndings is a tendency to prioritise the provision of beneﬁcial treatment outcomes over client choices. Yet one need not preclude the other. In shared decision-making where treatment information is disclosed comprehensively and client preferences are included, treatment outcomes that are beneﬁcial may be consented to by the client and endorsed by the therapist. On the other hand, beneﬁcial treatment outcomes may be embedded in medical paternalism in the current client autonomy model where information is provided to the clients who then, must alone, evaluate it with their preferences to come to a decision. The framework physiotherapists currently use has not kept up with the recent literature on shared decision-making. The local regulatory guidelines attend to the importance of imparting information to the client who then consents or refuses treatment as a one-time event. While transfer of information to clients by physiotherapists meets the criteria for decreasing therapist authority and increasing client autonomy, it also transfers responsibility from experienced and knowledgeable physiotherapists to far less clinically experienced and knowledgeable clients. Far from reducing power and information inequities, it exacerbates them. Information about proposed treatment should be accompanied by client preferences for treatment with both physiotherapist and client coming to a shared agreement on treatment options, as Charles et al. (1997) have suggested. Opportunity for this exists with goal-setting and development of treatment plans with clients. A more explicit consent process can be intertwined with implicit and embodied consent in a continuous treatment process. From the results, we developed a framework for a goal-focused, physiotherapist-client collaborative and explicit informed consent process that verbalises the currently tactile components of implied and embodied consent (Fig. 2). It involves the therapist eliciting the thought process by thinking aloud (Elkin, 2001). The model assumes competency on the part of the client throughout the treatment process. 5.1. Study limitations The presence of others in a focus group inﬂuences what is said by participants, with ‘public’ rather than ‘private’ views generally

A. Fenety et al. / Manual Therapy 14 (2009) 654–660

659

A collaborative, interwoven explicit and implicit consent process Step one:

Upon entry to a clinic, a client is assigned a therapist and signs a document providing blanket consent to treatment for their injury.

Step two:

Prior to assessment of their injury, it is suggested that clients be informed that examination may cause pain or injury flare-up. Verbal consent should be obtained before beginning the assessment.

Step three:

Following assessment, functional, client-identified goals and therapist-identified objectives to achieve those goals can be established collaboratively. The result is client-therapist sharing of decision-making with consent to treatment interwoven throughout. The therapist informs clients in the consent process by i) obtaining explicit written consent for high risk (e.g. manipulation), challenging or painful (e.g. aggressive treatment) procedures, ii) explicitly stating aloud to the client the standing order, iii) inviting questions, and iv) explaining intentions to observe for bodily signs of consent or refusal.

Step four:

As treatment proceeds, the client’s implied, embodied consent or refusal is observed by the therapist and made explicit by verbalising it to the client as part of the dialogue that occurs simultaneously with treatment procedures.

Step five:

Re-assessment of the client’s progress on the next visit begins the process again, with oral consent before proceeding.

Fig. 2. A collaborative, interwoven explicit and implicit consent process.

offered (Ulin et al., 2005). In this study, the presence of participant’s colleagues, employer or supervisor may have limited some conversations. Focus group interviews were two hours in length, however, the number of questions related to physiotherapists’ treatment practices constrained depth of response in favour of breadth of discussion. 6. Conclusion Described here is a model of practice regarding the informed consent process in physiotherapy that is implicit, embodied and continuous with the treatment process. We suggest this process can be improved to encompass a more collaborative explicit consent intertwined with an embodied, implicit consent process between therapist and client. Informed consent cannot occur separately from assessment and treatment processes. It is an integral aspect of physiotherapy practice and has a role at every stage of treatment. Acknowledgements We would like to thank the focus group participants for their time and thoughtful responses. We also thank the anonymous reviewers for their comments and suggestions. References Aveyard H. Informed consent prior to nursing care procedures. Nursing Ethics 2005;12:19–29. Bassett R, Graham JE. Memorabilities: enduring relationships, memories and abilities in dementia. Ageing & Society 2007;27:533–54. Bassett R, Chapman GE, Beagan BL. Autonomy and control: the co-construction of adolescent food choice. Appetite 2008;50:325–32. Boyatzis RE. Transforming qualitative information: thematic analysis and code development. Thousand Oaks, CA: SAGE; 1998.

Carlesso L, Bartlett DJ, Padﬁeld B, Chesworth BM. Cervical manipulation and informed consent: Canadian manipulative physiotherapists’ opinions on communicating risk. Physiotherapy Canada 2007;59:89–96. Charles C, Gafni A, Whelan T. Shared decision-making in the medical encounter: what does it mean? (or it takes two to tango). Social Science & Medicine 1997;44:681–92. Code of ethics. College of Physiotherapists of Ontario; 2000. Coney S, Bunkle P. The unfortunate experiment. Metro June 1987:46–65. Corbin J, Strauss A. In: Basics of qualitative research: techniques and procedures for developing grounded theory. 3rd ed. Newbury Park: SAGE; 2008. Corrigan O. Empty ethics: the problem with informed consent. Sociology of Health and Illness 2003;25:768–92. Dawes PJ, O’Keefe L, Adcock S. Informed consent: using a structured interview changes patients’ attitudes towards informed consent. Journal of Laryngology and Otology 1993;107:775–9. Delaney C. Should I warn the patient ﬁrst? Australian Journal of Physiotherapy 1996;42:249–55. Delaney C. Cervical manipulation – how might informed consent be obtained before treatment? Journal of Law & Medicine 2002;10:174–86. Delaney C. Respecting patient autonomy and obtaining their informed consent: ethical theory – missing in action. Physiotherapy 2005;91:197–203. Delaney C. In private practice, informed consent is interpreted as providing explanations rather than offering choices: a qualitative study. Australian Journal of Physiotherapy 2007;53:171–7. Dixon-Woods M, Williams SJ, Jackson CJ, Akkad A, Kenyon S, Habiba M. Why do women consent to surgery, even when they do not want to? An interactionist and Bourdieusian analysis. Social Science & Medicine 2006;62:2742–53. Doyal L. Informed consent: moral necessity or illusion? Quality in Health Care 2001;10:129–33. Elkin S. Informed consent: requirements for legal and ethical practice. Physiotherapy Theory and Practice 2001;17:97–105. Elwyn G, Edwards A, Gwyn R, Grol R. Towards a feasible model of shared decisionmaking: focus group study with general practice registrars. British Medical Journal 1999;319:753–6. Everett CR, Novoseletsky D, Cole S, Frank J, Remillard C, Patel RK. Informed consent in interventional spine procedures: how much do patients understand? Pain Physician 2005;8:251–5. Gabard DL, Martin MW. Physical therapy ethics. Philadelphia, PA: F.A. Davis Company; 2003. pp. 49–68 (chapter 3). Gafni A, Charles C, Whelan T. The physician-patient encounter: the physician as a perfect agent for the patient versus the informed treatment decision-making model. Social Science and Medicine 1998;47:347–54.

660

A. Fenety et al. / Manual Therapy 14 (2009) 654–660

Grant A. The use of qualitative research methodologies within musculoskeletal physiotherapy practice. Manual Therapy 2005;10:1–3. Grant, R., Trott, P., 1991. Pre-manipulative testing of the cervical spine – the APA protocol three years on. In: Manipulative Physiotherapists Association of Australia 7th Biennial Conference, Leura, NSW, Proceedings, pp 40–45. Grbich C. Qualitative research in health: an introduction. Thousand Oaks: Sage; 1999. Jones JH. Bad blood: the Tuskagee syphilis experiment. New York: Simon & Shuster; 1993. Kent G. Shared understandings for informed consent: the relevance of psychological research on the provision of information. Social Science & Medicine 1996;43:1517–23. Nova Scotia College of Physiotherapists practice guidelines. Nova Scotia College of Physiotherapists; 1999. O’Neill O. Some limits of informed consent. Journal of Medical Ethics 2003;29:4–7.

Purtilo RB. A time to harvest, a time to sow: ethics for a shifting landscape. Physical Therapy 2000;80:1112–9. Sim J. Ethical decision-making in therapy practice. Oxford: Reed; 1997. pp. 59–75 (chapter 4). Sim J. Informed consent and manual therapy. Manual Therapy 1996;2:104–6. Sin CH. Seeking informed consent: reﬂections on research practice. Sociology 2005;39:277–94. Strom-Gottfried K. Informed consent meets managed care. Health and Social Work 1998;23:1–11. Swisher LL, Krueger-Brophy C. Legal and ethical issues in physical therapy. Boston: Butterworth-Heinemann; 1998. pp. 173–199 (chapter 6). Ulin PR, Robinson ET, Tolley EE. Qualitative methods in public health: a ﬁeld guide for applied research. San Francisco: Jossey-Bass; 2005. p. 71 (chapter 4). Waterworth S, Luker KA. Reluctant collaborators: do patients want to be involved in decisions concerning care? Journal of Advanced Nursing 1990;15:971–6.

Manual Therapy 14 (2009) 661–664

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

Massage and mobilization of the feet and ankles in elderly adults: Effect on clinical balance performance Jacques Vaillant a, b, e, *, Audrey Rouland b, Pascale Martigne´ b, Renaud Braujou b, c, Michael J. Nissen d, Jean-Louis Caillat-Miousse b, Nicolas Vuillerme a, Vincent Nougier a, Robert Juvin e a

Laboratoire Sante´ Plasticite´ Motricite´, Universite´ Joseph Fourier-Grenoble 1, Grenoble, France Ecole de Kine´sithe´rapie du Centre Hospitalier Universitaire de Grenoble, France ˆ pital de Saint-Laurent-du-Pont, Saint-Laurent-du-Pont, France Ho d Service de Rhumatologie du Centre Hospitalier Universitaire de Gene`ve, Switzerland e Service de Rhumatologie du Centre Hospitalier Universitaire de Grenoble, France b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 25 March 2008 Received in revised form 19 February 2009 Accepted 2 March 2009

The aim of this study was to evaluate the effects of a session of plantar massage and joint mobilization of the feet and ankles on clinical balance performance in elderly people. A randomized, placebo-controlled, cross-over trial was used to examine the immediate effects of manual massage and mobilization of the feet and ankles. Twenty-eight subjects, aged from 65 to 95 years (78.8 8.5 years – mean SD) were recruited from community nursing homes. Main outcome measures were the performances in three tests: One Leg Balance (OLB) test, Timed Up and Go (TUG) test and Lateral Reach (LR) test. Results demonstrated a signiﬁcant improvement after massage and mobilization compared with placebo for the OLB test (1.1 1.7 s versus 0.4 1.2 s, p < 0.01) and the TUG test (0.9 2.6 s versus 0.2 1.2 s, p < 0.05). Conversely, performances in the LR test did not improve signiﬁcantly. These results emphasise the positive impact of a single session of manual therapy applied to the feet and ankles on balance in elderly subjects. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Balance Elderly Massage Mobilization Physical therapy Ankle Foot

1. Introduction The cause of falls in the elderly is multifactorial (Tinetti, 2003). Amongst the many factors described, the fact that the aging process results in reduced joint ﬂexibility and reduced afferent sensory information is well established. All joints show a signiﬁcant reduction in range of motion (ROM) with age. Ankle dorsiﬂexion (knee extended) shows the greatest age-related decline (James and Parker, 1989). Decrease in dorsiﬂexion ROM is associated with normal aging in both men and women (Vandervoort et al., 1992; Gadjosik et al., 1999). Fallers show a reduced ankle ROM (Kemoun et al., 2002). Despite the atrophy of the ankle musculature which occurs with aging, passive resistive torque of stretched connective tissue shows an increasing trend in older subjects (Vandervoort et al., 1992).

* Correspondence to: Jacques Vaillant, Ecole de kine´sithe´rapie du Centre Hospitalier Universitaire de Grenoble, BP 217, F 38041 Grenoble cedex 09, France. Tel.: þ33 4 76 76 52 56; fax: þ33 4 76 76 59 18. E-mail address: [email protected] (J. Vaillant). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.03.004

Adequate ROM of the ankle and metatarsophalangeal (MTP) joints is an important prerequisite to enable the performance of essential daily activities such as locomotion and balance (Lung et al., 1996). Walker et al. (1984) found an age-related decrease in ﬂexion of the ﬁrst MTP joint. Moreover, Mecagni et al. (2000) showed a correlation between ankle ROM and balance in community-dwelling elderly women. In addition to the musculoskeletal aspects, control of balance requires coordinated activity of the neuromuscular system. Accurate sensory inputs are necessary to organize motor programs and to generate effective motor output responses (Vandervoort, 1999). Sensations from the bottom of the feet play an important role during dynamic postural responses (Perry et al., 2000; Perry, 2006). Therefore, two important sources of information, cutaneous input from the feet and joint input from the feet and ankles, could be manipulated. In previous studies, the effects of (1) mechanical stimulation of the feet (Bernard-Demanze et al., 2004) and (2) massage and manipulation of the feet (Vaillant et al., 2008) on postural control during quiet standing have been shown. However, effects on functional balance performances were unknown.

662

J. Vaillant et al. / Manual Therapy 14 (2009) 661–664

Few functional balance tests have been validated, although they are commonly used. The One Leg Balance test (OLB) is one of the most common tests used to measure balance in older people (Jarnlo and Nordell, 2003) and is a simple predictive test for injury-related falls (Vellas et al., 1997). The Timed Up and Go (TUG) test is a convenient and reliable test for estimating physical mobility (Podsiadlo and Richardson, 1991). The importance of mediolateral postural control has been demonstrated to be a signiﬁcant factor in the identiﬁcation of elderly fallers (Maki et al., 1994; Maki and McIlroy, 1998; Brauer et al., 1999; Lord et al., 1999). The Lateral Reach (LR) test was found to be a valid indicator of lateral stability limits (Brauer et al., 1999). The aims of this study were ﬁrstly to compare the impact of massage and mobilization of the feet and ankles versus placebo on functional balance performance, and secondly to improve the understanding of the role played by distal lower limb inputs in balance control during functional activities. 2. Methods 2.1. Population Healthy volunteers were recruited from three community nursing homes, with help from managers. The criteria for inclusion were: age over 65 years and the ability to walk 10 m. Subjects were excluded if they had severe cognitive impairment, rapidly progressive or terminal illness, acute illness or unstable chronic illness, myocardial infarction or a fracture of the lower limb, within the six months prior to inclusion. Twenty-eight subjects with a mean age of 78.8 years (SD ¼ 8.5 years, range 65–95 years, 15/28 women) completed written consent to the study as required by the local Ethics Committee. One subject dropped out of the study because of lack of interest. Each of the remaining 27 subjects participated in two experimental sessions. Fourteen subjects received the massage and mobilization in the ﬁrst experimental session and the placebo condition during the second session, while the remaining 13 subjects received the same interventions in the reverse order. 2.2. Experimental design For this cross-over study, two sessions were organized: one involved massage and mobilization and the other, application of placebo. In order to avoid carry-over effects, at least one week separated the two sessions which were performed in random order. The measurements were obtained immediately before and after each of these protocols. 2.3. Measurement procedure 2.3.1. Protocol The functional tests, all performed in the following order, included the OLB test (Vellas et al., 1997), the TUG test (Podsiadlo and Richardson, 1991) and the LR test (Brauer et al., 1999). The OLB and TUG tests were timed with a digital stopwatch by an assessor blinded to the randomization protocol. Three trials were performed and the mean time was calculated. For the LR test, three trials on each side were carried out and the mean distance was calculated.

were allowed. During the test, the subject was not allowed to move the foot from the ﬂoor (Vellas et al., 1997). 2.3.2.2. TUG test. A high straight-backed ofﬁce chair with arm rests was placed 3 m from a wall. Subjects sat comfortably in the chair and were asked to rise and stand still momentarily, walk towards the wall, turn without touching the wall, walk back to the chair, turn around and then sit down again. The score given was the time in seconds required to complete the test (Podsiadlo and Richardson, 1991). 2.3.2.3. LR test. The subjects stood with their back against (but not in contact with) a wall. The feet were placed in a standardized position with 10 cm between the most medial aspects of the heels and at an outwards angle of 30 degrees. To ensure accurate recording of the initial hand position, subjects stood for 10 s with both arms abducted to 90 degrees and maintained equal weight bearing. Subjects were given standardized instructions and encouragements to reach directly sidewards as far as possible without overbalancing, taking a step or touching the wall. The contralateral arm remained by their side during the reach. Both feet had to remain fully in contact with the support surface throughout the task, no knee ﬂexion was permitted and no trunk rotation or ﬂexion was tolerated (Brauer et al., 1999). 2.4. Therapeutic protocol The massage and mobilization protocol (MMP) included a therapeutic manipulation of the feet and ankles. This intervention, widely used by physical therapists (Dufour, 1996) was designed to target the somatosensory system of the feet and ankles. Given that the somatosensory system includes multiple receptors that provide information about pressure distribution (cutaneous), muscle tension (Golgi tendon organs), joint angle changes (joint receptors) and muscle length changes (spindles), the intervention involved manual massage of the feet and mobilization of the feet and ankle joints. Intervention methods were standardized. Half of the allocated time was for massage and the other half for joint mobilization. The aim of the massage was to enhance local blood circulation and to stimulate cutaneous receptors. The massage technique (Clay and Pounds, 2006) involved the application of friction, static and glide pressure focus on the sole of the foot. Multidirectional, systematic tractions on the sole of the foot were performed particularly in the heel region and over the metatarsal heads. Mobilization involved dorsiﬂexion and plantar ﬂexion of the talocrural joints, eversion and inversion of the subtalar joints, anteroposterior glide, torsion, ﬂexion and extension of the midtarsal joints, anteroposterior glide and rotation of the tarsometatarsal joints, anteroposterior glide of the intermetatarsal joints, and plantar ﬂexion and extension of the MTP and interphalangeal joints. Each manipulation was performed three times per foot. Manual massage and joint mobilization were applied to the feet and ankles for a total of 20 min. 2.5. Placebo protocol (PP) The PP consisted of the application of three demagnetized magnets in the region of the ﬁfth metatarsals for 20 min. 2.6. Statistical analysis

2.3.2. Functional test description 2.3.2.1. OLB test. The subjects stood on one leg with the other slightly ﬂexed, ﬁrst the right and then the left, for as long as possible without shoes, looking at a target. Three trials on each leg

The data was analysed with ‘‘R’’ statistical software (Version 2.4.0) and as it was not normally distributed (Shapiro–Wilk normality test), the Wilcoxon rank test was used to test statistical

J. Vaillant et al. / Manual Therapy 14 (2009) 661–664

3. Results

4 3.5

Improvement (sec.)

differences between sessions. Test–retest reliability was performed on pre-intervention data by using Intraclass Correlation Coefﬁcients (ICCs). In addition, the Standard Error of Measurement (SEM) was determined for each of the continuous variables according to the following equation: SEM ¼ SD O (1 ICC) and the Smallest Detectable Difference (SDD) was calculated as SDD ¼ SEM 1.96 (z score for 95% conﬁdence) O2. A two-sided p value less than or equal to 0.05 was considered to indicate statistical signiﬁcance.

3.1. Reliability of the tests

663

*

*

3 2.5 2 1.5 1 0.5

The results demonstrated a good reliability between the three tests. Despite of the frailty of many of the patients, the ICCs of the observations for the same observer between pre-tests were: 0.98 for the OLB test, 0.98 for TUG test and 0.92 for LR test. 95% CI, SEM and SDD were also calculated (Table 1).

0

One Leg Balance

Timed Up & Go

placebo

treatment

*p value <0.05 Fig. 1. Changes (in s) in the TUG test and the OLB test between the value obtained after the treatment protocol versus placebo and the baseline value.

3.2. Treatment versus placebo (Table 2)

4. Discussion Results showed a statistically signiﬁcant improvement in performance for both the OLB test and the TUG test, whereas the LR test did not improve signiﬁcantly. Several other studies have reported ﬁndings of a similar nature. Andre´-Deshays and Revel (1988) demonstrated the sensory role of the plantar sole, particularly with regards to sense of movement in the mediolateral direction. Kavounoudias et al. (2001) established that the anterior region of the sole (heads of the ﬁrst metatarsal and of the ﬁfth metatarsal) demonstrate high discrimination capacities. Perry et al. (2000) reported the importance of cutaneous receptors from the plantar sole in controlling speciﬁc aspects of rapid compensatory stepping reactions caused by postural perturbations. Since then, Bernard-Demanze et al. (2004), Perry (2006) and Vaillant et al. (2008) have all demonstrated that the quality of somatosensory information plays a major role in postural control. Improvement of somatosensory information could potentially explain the beneﬁt observed in our study. On the other hand, the improvement may also be related to mechanical effects. The importance of ROM of the feet and ankle joints with regards to balance and locomotion performance is also known. Mecagni et al. (2000) showed a correlation between ankle ROM and the functional reach test. In our study, changes of joint ROM were not measured. This hypothesis needs to be explored further in future studies. Even though the tests demonstrated good reliability in our prestudy, the results have to be put into perspective, in that the improvements after the MMP were less than or equal to the SEM Table 1 Reliability of pre-test.

OLB (s) TUG (s) LR (cm)

and SDD (Tables 1 and 2). Moreover, the 1.1 and 0.9 s improvements for the OLB and TUG tests respectively observed in our study, were similar to the minimal detectable change usually considered to be between 1 and 2 s (Piva et al., 2004; Lim et al., 2005). Nevertheless, the use of demagnetized magnets as placebo, as previously described by Martel et al. (2002), would allow us to consider that these results are robust. In addition, the cross-over with the placebo session allowed us to neutralize the warm-up and apprenticeship effects. With regards to the LR test, the cross-over with the placebo session enabled us to neutralize the improvement in performance linked to the repetition of the movement (Brauer et al., 1999). This learning phenomenon could explain the 0.8 cm improvement in the placebo group, which was not statically different from the 1.3 cm improvement in the MMP. In the context of a progressively ageing population and increasing falls, these results after only a single 20 min intervention are very promising. There were however several limitations to the study protocol. Firstly, the brief period during which the subjects were treated is unusual. It is generally acknowledged that a standard protocol of therapeutic intervention consists of sessions 4 3.5 3

Improvement (cm)

After the MMP (Figs. 1 and 2), the improvement of performance in the OLB test and the TUG test was greater with respect to the PP. Mean (SD) changes in the OLB test were 1.1 (1.7) s for MMP and 0.4 (1.2) s for PP (p < 0.01). Mean (SD) changes in the TUG test were 0.9 (2.6) s for MMP and 0.2 (1.2) s for PP (p < 0.05). Conversely, the improvement in the LR test was not signiﬁcantly different between MMP (1.3 2.3 cm) and PP (0.8 1.3 cm).

2.5 2 1.5 1 0.5 0

ICC

95% CI

SEM

SDD

0.98 0.98 0.92

0.73–0.99 0.96–0.99 0.88–0.98

0.98 0.99 0.96

2.73 2.74 2.65

Lateral Reach placebo

treatment

Fig. 2. Changes (in cm) in the LR test between the value obtained after the treatment protocol versus placebo and the baseline value.

664

J. Vaillant et al. / Manual Therapy 14 (2009) 661–664

Table 2 Pre-tests, Post-tests and improvements (6) during MMP and PP (n ¼ 27). MMP

OLBa (s) TUGa (s) LRa (cm) a

PP

Pre-test

Post-test

6

Pre-test

Post-test

6

5.17 7.97 21.55 10.88 11.59 5.90

6.22 7.73 20.63 11.76 12.86 6.59

1.05 1.65 0.92 2.63 1.28 2.27

5.36 8.01 20.82 9.92 11.47 5.82

5.77 7.94 20.65 9.91 12.30 5.77

0.41 1.19 0.17 1.24 0.83 1.25

Mean (SD).

of at least 20 min that are repeated several times per week over a period of approximately 10 weeks (American Geriatrics Society et al., 2001). Secondly, in the present experiment, the effect of the therapeutic intervention on clinical performance was assessed immediately after the therapeutic manipulation. It is therefore impossible to judge the potential durability of the reported improvement. Moreover, the impact of manipulative techniques and joint mobilization versus massage of the plantar sole is unknown. Further research is required to clarify these outstanding questions.

References American Geriatrics Society, British Geriatrics Society, American Academy of Orthopaedic Surgeons Panel on Falls Prevention. Guideline for the prevention of falls in older persons. Journal of American Geriatrics Society 2001;49:664–72. Andre´-Deshays C, Revel M. Evaluation de la sensibilite´ kinesthe´sique de la cheville. Application a` la re´e´ducation. Annales de Re´adaptation et Me´decine Physique 1988;31:367–76. Bernard-Demanze L, Burdet C, Berger L, Rougier P. Recalibration of somesthetic plantar information in the control of undisturbed upright stance maintenance. Journal of Integrative Neurosciences 2004;3:433–51. Brauer S, Burns Y, Galley P. Lateral reach: a clinical measure of medio-lateral postural stability. Physiotherapy Research International 1999;2:81–8. Clay JH, Pounds DM. Basic clinical massage therapy: integrating anatomy and treatment. New-York: Lippincott Williams and Wilkins; 2006. Dufour M. Massages. Encyclope´die Me´dico-Chirurgicale, Kine´sithe´rapie-Me´decine physique-Re´adaptation. Paris: Elsevier; 1996. Gadjosik RL, Vander Linden DW, Williams AK. Inﬂuence of age on length and passive elastic stiffness characteristics of the calf muscle-tendon unit of women. Physical Therapy 1999;79:827–38. James B, Parker AW. Active and passive mobility of lower limb joints in elderly men and women. American Journal of Physical Medicine 1989;68(4):162–7. Jarnlo GB, Nordell E. Reliability of the modiﬁed ﬁgure of eight – a balance performance test for elderly women. Physiotherapy Theory Practice 2003;19:35–43. Kavounoudias A, Roll R, Roll JP. Foot sole and ankle muscle inputs contribute jointly to human erect posture regulation. Journal of Physiology 2001;532:869–78. Kemoun G, Thoumie P, Boisson D, Guieu JD. Ankle dorsiﬂexion delay can predict falls in elderly. Journal of Rehabilitation Medicine 2002;34:278–83. Lim L, van Wegen E, de Goede C, Jones D, Rochester L, Hetherington V, et al. Measuring gait and gait-related activities in Parkinson’s patients’ own home

environment: a reliability, responsiveness, and feasibility study. Parkinsonism & Related Disorders 2005;11:19–24. Lord SR, Rogers MW, Howland A, Fitzpatrick R. Lateral stability, sensorimotor function and falls in older people. Journal of American Geriatrics Society 1999;47(9):1077–81. Lung MW, Hartsell HD, Vandervoort AA. Effects of aging on joint stiffness: implications for exercise. Physiotherapy Canada 1996;48(2):96–106. Maki BE, McIlroy WE. Control of compensatory stepping reactions: age-related impairment and potential for remedial intervention. Physiotherapy Theory and Practice 1998;15:69–90. Maki BE, Holiday PJ, Topper AK. A prospective study of postural balance and risk of falling in an ambulatory and independent elderly population. Journal of Gerontology 1994;49:M72–84. Martel GF, Andrews SC, Roseboom CG. Comparison of static and placebo magnets on resting forearm blood ﬂow in young, healthy men. The Journal of Orthopaedic and Sports Physical Therapy 2002;32(10):518–24. Mecagni C, Smith JP, Roberts KE, O’Sullivan SB. Balance and ankle range of motion in community-dwelling women aged 64 to 87 years: a correlational study. Physical Therapy 2000;80:1004–11. Perry SD. Evaluation of age-related plantar-surface insensitivity and onset age of advanced insensitivity in older adults using vibratory and touch sensation tests. Neuroscience Letters 2006;392:62–7. Perry SD, McIlroy WE, Maki BE. The role of plantar cutaneous mechanoreceptors in the control of compensatory stepping reactions evoked by unpredictable, multidirectional perturbation. Brain Research 2000;877:401–6. Piva SR, Fitzgerald GK, Irrgang JJ, Bouzubar F, Starz TW. Get up and go test in patients with knee osteoarthritis. Archives of Physical Medicine and Rehabilitation 2004;5(2):284–9. Podsiadlo D, Richardson S. The timed « Up and Go »: a test of basic functional mobility for frail elderly persons. Journal of American Geriatrics Society 1991;39:142–8. Tinetti ME. Preventing falls in elderly persons. The New England Journal of Medicine 2003;348:42–9. Vaillant J, Vuillerme N, Janvy A, Braujou R, Louis F, Juvin R, et al. Effect of manipulation of the feet and ankles on postural control in elderly adults. Brain Research Bulletin 2008;75:18–22. Vandervoort AA. Ankle mobility and postural stability. Physiotherapy Theory Practice 1999;15:91–103. Vandervoort AA, Chesworth BM, Cunningham DA, Paterson DA, Rechnitzer PA, Koval JJ. Age and sex effects on mobility of the human ankle. Journal of Gerontology 1992;47:M17–21. Vellas BJ, Wayne SJ, Romero L, Baumgartner RN, Rubenstein LZ, Garry PJ. One-leg balance is an important predictor of injurious falls in older persons. Journal of American Geriatrics Society 1997;45:735–8. Walker JM, Sue D, Miles-Elkousy N, Ford G, Trevelyan H. Active mobility of the extremities in older subjects. Physical Therapy 1984;64:919–23.

Manual Therapy 14 (2009) 665–670

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

‘‘The goal is to be more ﬂexible’’ – Detailed analysis of goal setting in physiotherapy using a conversation analytic approach Veronika Schoeb* Department of physiotherapy, HECVSante´, HES-SO University of Applied Sciences Western Switzerland, Avenue Beaumont 21, CH-1011 Lausanne, Switzerland

a r t i c l e i n f o

a b s t r a c t

Article history: Received 9 October 2008 Received in revised form 16 February 2009 Accepted 24 February 2009

Standards of practice in physiotherapy require professionals to set goals in collaboration with patients. Some evidence exists regarding outcome, but no studies have investigated the process of goal setting in orthopaedic physiotherapy. The objective of this pilot study is to describe the interaction between professionals and patients in the activity of goal setting. Three physiotherapists in an orthopaedic outpatient department audio-recorded initial patient consultations. Conversation analysis (hereafter CA), a qualitative research method, is used to illustrate participants’ interaction in regards to turn taking, how talk is organised, what vocabulary they use and how they respond to each other’s utterances. Four distinct phases are found: eliciting patients’ preferences; introduction of goal setting activity; formulating goals; and closing the activity of goal setting. Formulating questions to elicit patients’ preferences requires considerable effort. Constant adjustment is needed in order to achieve goals acceptable to both participants. Goal setting is time-consuming if the patient is actively involved. Closing of goal setting activity in physiotherapy is comparable to this activity in doctor–patient interaction. In conclusion it is not straightforward to formulate and negotiate treatment goals collaboratively with patients. A balance has to be found between the input of physiotherapists and patients during the process. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Goal setting Conversation analysis Physiotherapy Orthopaedics

1. Introduction Physiotherapy is governed by Standards of Practice, the purpose of which are to enhance quality and effectiveness in the provision of health services. Standards of Physiotherapy Practice emphasize the importance of collaborative goal setting (WCPT, 1999; APTA, 2004; Physioswiss, 2007). The World Health Organisation deﬁnes a goal as ‘‘a general or speciﬁc objective towards which to strive; an ultimate desired state towards which actions and resources are directed’’ (WHO, 2004, p. 27). WHO recommendations include that ‘‘goal setting should be one strategy to encourage active rather than passive decision-making’’ (Peri et al., 2006, p. 1). Professional conduct means adhering to professional regulations. As such, one of the purposes of goal setting identiﬁed in published studies is to meet contractual, legislative or professional requirements (Levack et al., 2006a). Another purpose of goal setting relates to outcomes, either to improve outcomes or to evaluate them (Levack et al., 2006a).

* Tel.: þ41 21 314 69 16; fax: þ41 21 314 69 00. E-mail address: [email protected] 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.02.004

Even though goal setting is frequently used, the methods are hardly ever standardised (Holliday et al., 2005). There is inconsistent evidence regarding effectiveness of goal planning (Levack et al., 2006b). It is argued that goal setting is complex and potential disagreements regarding goals seem substantial (Bradley et al., 1999). However, there are results showing that patients would like to participate (Payton et al., 1998), and that physiotherapists are convinced about the importance of goal setting (Baker et al., 2001). Moreover, the types of treatment objectives can be identiﬁed: decrease of pain, increase of range of motion, increase of muscle strength, decrease of muscle tension, and worker education (Poitras et al., 2005). But there is uncertainty about the nature of the goal-setting process. Studies investigating barriers to goal setting found that the process had not been put into practice due to interactional (Wressle et al., 1999; Playford et al., 2000; SchulmanGreen et al., 2006) as well as organisational reasons (Playford et al., 2000). None of the previously cited studies used observational methods in order to investigate how the process of goal setting takes place. One study using a conversation analytic approach provides some insight into this aspect (Parry, 2004). The study in a neurological setting revealed that physiotherapists make goals

666

V. Schoeb / Manual Therapy 14 (2009) 665–670

explicit in only 8 out of 74 videotaped treatment sessions. Several hypotheses explaining this practice were generated through observations: goal setting is time-consuming, interactional difﬁculties can arise when discussing patients’ problems, and goal setting assumes that progress is possible which might not always be the case in stroke physiotherapy. We can see then, that although policy documentation emphasizes on collaborative goal setting, with some more or less robust evidence on outcome, there is very little evidence on the process in physiotherapy practice. The analysis of the interaction between patients and physiotherapists is fundamental to understanding goal setting in physiotherapy. The objective of this study is to describe the structure of goal-setting interactions in orthopaedic physiotherapy practice. 2. Methods 2.1. Setting and participants The setting for this study was an outpatient physiotherapy department in a Swiss University Hospital. The audio-recorded interactions were part of a research project investigating patient outcomes of physiotherapy services. Three physiotherapists each provided audio-tapes from one initial treatment session. Physiotherapists’ and patients’ characteristics are presented in Table 1. 2.2. Data collection Patients referred to physiotherapy and scheduled with one of the three physiotherapists were asked to participate. Ethics committee approval for the whole study was granted by the local commission. Participants signed a consent form and anonymity was guaranteed. The treating physiotherapist put a small digital audio-recorder on a chair in a closed treatment room. Apart from the presence of the tape-recorder, none of the procedures commonly used in the setting were changed. At the end of the treatment session, the physiotherapist provided the researcher with the tape for transcription and analysis. 2.3. Data analysis CA was the method chosen for this study. CA was developed by Harvey Sacks in collaboration with Emanuel Schegloff and Gail

Table 1 Physiotherapists’ and patients’ characteristics. Extracts Physiotherapist’s position

Physiotherapist’s characteristics

Patient’s medical diagnosis

Patient’s characteristics

Extract Junior physiother1.1 apist (male) Extract Code: PT1 2.2 Extract 3.1 Extract 3.2 Extract 4.2

3 years of experience Postgraduate training

Ankylosing spondylitis

Nurse employed in the same hospital (male) Code: Pat A

Extract Junior physiother- 5 years of experi- Acute ankle 1.2 apist (female) ence, sprain Code: PT2 new to the service

Work in IT (male) Code: Pat B

Extract Senior physiother- 15 years of 2.1 apist (male) experience Extract Code: PT3 4.1

Engineer (male) Code: Pat C

Knee pain after meniscectomy

Jefferson (Ten Have, 2004). It can be said that ‘‘it is both an interpretive enterprise seeking to capture the understandings and orientations displayed by the participants themselves, and at the same time it enforces rigorous standards of evidence made possible by the use of recorded data’’ (Clayman and Gill, 2004, p. 590). The theory underlying CA is that ‘‘previous actions are a primary aspect of the context of action, that meaning of an action is heavily shaped by the sequence of the previous action, and that social context itself is a dynamically created thing that is expressed in and through the sequential organisation interaction’’ (Heritage, 2005, p. 104). CA has been widely applied to the analysis of medical interaction (Heritage and Maynard, 2006). In this study, sequences of goal setting were selected from each taped initial examination and transcribed by the researcher using Jefferson’s transcription conventions (Jefferson, 1984; Appendix 1) which are the most commonly used conventions for CA data (Ten Have, 1999). As the data was recorded in French, a line-by-line translation is provided. The selected extracts were analysed in order to establish patterns of interaction (Drew et al., 2001). The focus was on the following aspects (Heritage, 2004, 2005): Turn-taking organisation: participants use special turn-taking procedures (e.g. question–answer during examination activity) and are expected to follow certain rules (e.g. patients answer questions or tell their story as to how an accident happened). Structural organisation: analysis of different sections of interaction (e.g. opening – greeting, problem initiation, disposal, closing). Sequence organisation: analysis of how actions are initiated, followed through and closed (e.g. how does goal setting activity start, continue and end) Turn-construction/design: analysis of alternative ways of saying (or doing in case of non-verbal communication) and how aspects of turn are articulated within the performance of organisational tasks (e.g. how goal setting is initiated and included in the physiotherapy evaluation and treatment process). Lexical choice: analysis of vocabulary used Forms of asymmetries: analysis of participation, knowledge and rights of access to knowledge (e.g. expert physiotherapist vs. expert patient). Using this analytical framework, conclusions (claims) can be drawn regarding the patient–physiotherapist interaction. In order for those conclusions to be valid, Arminen (2005) suggests applying the following rules: 1. Transparency: details of talk or action must be demonstrable 2. Validation by next turn: what comes after conﬁrms previous actions 3. Participant’s validation: attention is paid to the features of talkin-interaction oriented to by the participants themselves (p. 70) The presentation of the ﬁndings below includes the actual spoken interaction in order for the reader to test the claims for themselves. CA is an inductive, qualitative research method that does not set categories in advance. The elaboration of ﬁndings was rooted in the collected data. 3. Findings On the basis of the analysis of the audio-taped interactions, four phases can be described as part of the goal setting activity in orthopaedic physiotherapy practice:

V. Schoeb / Manual Therapy 14 (2009) 665–670

1. 2. 3. 4.

Eliciting patients’ expectations of physiotherapy treatment Introduction of goal setting activity Formulating goals Closing of goal setting activity

Extract 1.2 PT2 – Pat B. 1 Physio 2

3.1. Eliciting patient’s expectation of physiotherapy treatment To initiate the goal setting activity physiotherapists ask patients about their expectation regarding physiotherapy treatment. This seemingly simple task is not straightforward. In Extract 1.1, the physiotherapist makes a considerable effort asking the question using rephrasing and small pauses without being able to ﬁnish the sentence. The response of the patient is hypothetical (‘‘if’’) and is followed by laughter. Extract 1.2 shows a sequence where the physiotherapist does not pursue the answer to her question. The patient was treated for the same problem (acute ankle sprain) a few months earlier. The physiotherapist’s question is ambiguous as to whether it refers to what happened now or then. The question–answer sequence is interrupted when the physiotherapist answers the ﬁrst question herself (line 2). She abandons her initial question about the patient’s expectation for this treatment (line 1) and continues with the next question regarding the patient’s last experience with physiotherapy (line 2). The patient responds to the second question, and the physiotherapist does not come back to her initial question. In conclusion, formulating a question in order to elicit a patient’s expectation about physiotherapy is not straightforward. It can be noted that considerable effort is required. Goal setting activities are part of the ﬁrst encounter between patients and physiotherapists. Higgs and Jones (2000) describe physiotherapy encounters as a cyclical process of a ‘‘growing understanding of the client and the clinical problem’’ (p. 11). A complete understanding cannot happen in the ﬁrst 30 min (length of the initial evaluation) but continued engagement is required. 3.2. Introduction of goal setting activity Introducing the goal setting activity is a practice commonly found in our data. Extracts 2.1 and 2.2 show the same pattern: the physiotherapist reformulates the problem for which the patient is coming to see the therapist. In Extract 2.1 (lines 1–3) the physiotherapist summarises the problem evoking the reason why the patient is seeking his help, while the physiotherapist in Extract 2.2 (lines 1–3) presents a summary of the patient’s medical diagnosis and his understanding of how the patient deals with it. The patients’ responses to those summaries are different. Whereas in Extract 2.1 the patient only acknowledges the accuracy of the physiotherapist’s statement, in Extract 2.2, the patient is

Extract 1.1 PT1 – Pat A. 1 Physio ok donc la` (.) ton objectif pour ehh- apre`s ces se´ances (.) ce serait [lequel Ok well (.) your objective for ehh – after those sessions (.) that would be which [one 2 tu aimerais arriver a` quel stade apre`s ces (0.2) apre`s ces neuf se´ances (.) ces neuf se´ances [pre´scrites (.) You would like to reach what stage after those (0.2) after the nine sessions (.) the nine sessions [prescribed (.) 3 maisBut4 Patient Si je te dis d’eˆtre souple ((rires)) If I tell you to be ﬂexible ((laughter))

667

3 Patient

Et puis vous attendez quoi de (0.2) du traitement cette [fois And then what do you expect from (0.2) the treatment this [time en fait vous connaissez de´ja` – en fait mais (0.3) vous faites (.) vous avez une ide´e de ce que va eˆtre le [traitement Actually you already know- actually but (0.3) you make (.) you have an idea what the treatment will be [like Une ide´e a` peu pre`s oui A broad idea yes

actively engaged and helping the physiotherapist to gain a deeper understanding of his clinical problem. The active participation allows for a more complete picture of the patient’s problem, but requires more time and attention from the physiotherapist to keep the conversation on track. The ﬁnal sequence of this phase (Extract 2.2) is initiated rather hesitantly by the physiotherapist, marked with longer pauses (1.5 s in line 1) and hesitations (‘‘euh’’ and ‘‘donc’’, lines 1 and 2). The patient is actively seeking to take the ﬂoor. The interruptions are either of afﬁrmative nature (lines 4 and 8) or are designed to complete sentences initiated by the physiotherapist (lines 3 and 5). 3.3. Formulating goals In two of the three audio recordings (physiotherapist PT2 and PT3) only a small amount of exchange takes place on goal setting. The most detailed exchange can be seen in Extract 3.1 when both the physiotherapist and the patient make goals explicit. However, difﬁculties in the formulation of goals are also found in this extract. The physiotherapist (lines 16–18) is unsure about the way how to formulate treatment objectives. The difﬁculty is related to the fact that the physiotherapist is using the patient’s vocabulary (‘‘I formulated’’ – my objective), however, it is the patient’s goal (‘‘it is you who formulated’’ – your objective). The patient does not interrupt the turn and agrees with a response token (line 19). The reason for this confusion of ‘‘my objective’’ versus ‘‘your objective’’ is unclear. In terms of setting a realistic time-frame for goals, we see in Extract 3.2 how participants negotiate this activity. The physiotherapist introduces this topic as a statement (‘‘9 sessions’’ – line 20) rather than as a question. Without waiting for an answer, the physiotherapist continues with a more speciﬁc question related to the race in which the patient would like to participate. This question is answered by the patient (line 21). It is one of the few question–answer formats during the goal-setting process here. Nevertheless, there is some uncertainty (line 24) as the physiotherapist does not know exactly how to respond to the patient’s comment about the timeline. An interesting feature can be found in line 25 when the patient repeats

Extract 2.1 PT3 – Pat C. 1 Physio

donc si j ¼ vous comprends bien (.) l-le proble`me qui vous ame`ne¼ So if I understand you correctly (.) the problem that brings you here¼

2

¼c’est pas un proble`me qui fait b-beaucoup [souffri:r (.) ¼is not a problem that makes you suffer a [lo:t (.)

3

euh c’est plutoˆt euh le b ¼ soin d’avoir quelques outi:ls pour pouvoir prendre s’en ehh it is more ehh the need to get some to:ols in order to be able to take

4

charge pour qu ¼ ça [n ¼ pose pas d ¼ proble`mes [plus tard] care to [prevent] further problems [later on]

5 Patient

[tout-a`-fait] [exactly]

[tout-a`-fait] [exactly]

668

V. Schoeb / Manual Therapy 14 (2009) 665–670

Extract 2.2 PT1 – Pat A. 1 Physio

donc si je re´sume e:n deux mots euh (1.5) euh donc c’est une maladie qu ¼ tu Alright if I summarize w:ith two words eh (1.5) eh well it is a disease that you

2

connais bien Bechterew eu:h tu fais re´gulie`rement la physi:o une a` deux fois par know well Bechterew e:h you are doing regularly physiotherapy once or twice a

3

anne´e[par] pe´riode [comme] ça de deyear [for] periods [like] that that-

4 Patient

[oui] [yes]

[oui] [yes]

5

en fait de manie`re [aigue voila`] j’essaie j’ai dit bien j’essaie de me stretcher¼ actually in an [acute way well] I try I say I try to stretch ¼

6 Physio

[des se´ances re´gulie`res] [regular sessions] ¼voila` (.) tu fais [des exercices a` domic[ile tu fais des trucs diso[ns ¼alright (.) you are [doing exercises at ho[me you are doing things [well [voila` [voila` [voila` [that’s right [that’s right [that’s right

7

8 Patient

‘‘yes’’ three times. It is argued that multiple sayings are a ‘‘resource speakers have to display that their turn is addressing an in progress course of action rather than only the just prior utterance’’ (Stivers, 2004, p. 260). In our data it could be interpreted that the physiotherapist unnecessarily comments on their agreement regarding the nine sessions (line 24). It takes quite a long time (lines 25–35) for the physiotherapist to reformulate the patient’s comment into an objective. The difﬁculty lies in the different formulation of the problem. The patient is explaining how he would like to have the burning at the upper part of the back (lines 26 and 27) decreased. The physiotherapist on the other hand is speaking of decreasing the tensions (line 36). The patient does not react to this adjustment of the goal. Parry (2005) suggests that reaching alignment on proposed goals can present interactional difﬁculties. In our data, alignment is not reached due to the various, and not negotiated, descriptions of symptoms. 3.4. Closing of the goal setting activity Collaboratively closing down the sequence (Schegloff, 2007) is found in our data. This activity is described as a two-turn closing (Extract 4.1; lines 8 and 9) or a four-turn design (Extract 4.2; lines 57–61) with reference to the arrangement made (Goldberg, 2004). It is interesting to note that in Extract 4.2, there is an additional reference to the next meeting scheduled the following day (lines 58–60). This echoes evidence from research on closings in medical visits. West (2006) notes that arrangements for the future are often used to initiate this phase. The reference to closing the goal-setting

Extract 3.1 PT1 – Pat A. 16 Physio donc ça c’est ton objectif (1.0) euh moi je l’ai formul:e´ enﬁn c’est toi qui Well this is your goal (1.0) eh I formula:ted it well it is you who formulated it 17 l’a formule´ c’est eˆtre plus souple et puis pour pouvoir faire euh ton sport- enﬁn mon It is to be more ﬂexible and then to be able to do your sport- well my 18 sport tranquillement quoi sport quietly 19 Patient ouais Yeah

process can be found in Extract 4.2 (line 60) when the physiotherapist is actively looking for agreement. In conclusion, it can be argued that goal setting is a challenging component of physiotherapy. It is a highly complex interaction and is time-consuming. The initiation of the process as well as the formulation of treatment goals requires considerable effort from both patients and physiotherapists. It is not such a straightforward activity as Standards of Practice suggest. 4. Discussion The objective of this study was to describe patterns of patient– physiotherapist interaction around goal setting in an orthopaedic outpatient setting. Existing literature suggests that one reason for infrequent goal setting in neurological physiotherapy lies in the time-consuming nature of the activity (Parry, 2004). This fact is evident in our data when we compare the amount of effort put in by the two participants in one example (PT1 – Pat A) compared to the other (PT3 – Pat C). Both physiotherapists are working in the same orthopaedic outpatient setting with tight scheduling practice (30 min per patient) and where hospital standards require speciﬁc documentation. The physiotherapist PT3 is more efﬁcient in that he has goals written down more quickly than the physiotherapist PT1. However, due to the short exchange during goal setting, physiotherapist PT3 might not get a complete picture of the patient. It would be interesting to follow up on those initial encounters to see how interactions develop throughout the treatment sessions. Two types of interaction are found in our data: a therapist-led goal-setting process and a more collaborative process. The therapist-led goal-setting process (Extract 2.1) is based on the physiotherapist’s assumption of what might be suitable for this patient. This approach is described as paternalistic decision-making (Charles et al., 1999). Even though shared decision-making is promoted in the medical literature (Charles et al., 1997, 1999), research evidence from various health professions (nursing, homeopathy) detected both unilateral and bilateral (collaborative) approaches of practitioners to making decisions about treatment procedures (Collins et al., 2005). It can be argued that the achievement of decision-making in health care is more complex than the shared decision-making model suggests. It has also been questioned, whether the shared decision-making model in clinical practice actually improves patient satisfaction, treatment adherence and health (De Haes, 2006). A more patient-centred approach can be found in the interaction between the physiotherapist PT1 and the patient Pat A and corresponds to the patient-centred medicine model (Stewart et al., 1995). Literature offers different deﬁnitions of ‘‘patient-centred care’’, yet without reaching consensus. One group of researchers deﬁnes it as ‘‘sharing control’’ and ‘‘focus on the whole person’’ (Lewin et al., 2001), whereas others suggest ﬁve conceptual dimensions: bio-psycho-social perspective; the patient as person; sharing of power and responsibility; therapeutic alliance; and the therapist as person (Mead and Bower, 2000). A third group identiﬁes two different conceptions of patient-centeredness: ﬁrstly, the inclusion of the patient’s perspective and, secondly, the stimulation of the patient to participate actively in the treatment (Michie et al., 2003). One theme common to the shared decision-making model and the patient-centred medicine is the concept of patient participation. Even though it is proposed that ‘‘patient participation in decision-making is justiﬁed on humane grounds alone and is in line with a patient’s right to self-determination’’ (Guadagnoli and Ward, 1998; p. 337), there is evidence that variations in preferences for patients’ participation in decision-making exist (Levinson et al., 2004). Reasons for low level participation in clinical practice are

V. Schoeb / Manual Therapy 14 (2009) 665–670 Extract 3.2 PT1 – Pat A. 20 Physio

21 Patient 22 Physio

23 Patient

24 Physio 25 Patient 26

27 28 Physio 29 Patient 30 Physio 31 Patient 32 Physio 33 Patient 34 Physio 35

Extract 4.2 PT1 – Pat A. Hein (.) donc ça ça serait l’objectif a` (.) a` neuf se´an- c’est quand cet:te [compe´tition la` Ok (.) well this this would be the goal for (.) for nine sessi- it is when your [race Euh 15 jours apre`s Paˆques (.) dans 3 semaines Eh two weeks after Easter (.) in three weeks oui Alright (1.0) Non mais meˆme si c’est pas- euh meˆme si j ¼ ai pas ﬁni mes 9 se´ances pour mon truc No but even if it is not- eh even if I have not ﬁnished my nine sessions for my thing oui on est d’accord (.) mhm Alright we agree (.) mhm oui ¼ oui ¼ oui disons que je me sente un peu mieux ﬁt Yes ¼ yes ¼ yes well that I feel myself a little bit ﬁtter que vraiment ces douleurs la` dans le dos la` (.) qui me- que je disaisau niveau bruˆlures That really this pain there in the back there (.) that I- that I say- the level of burning ça ça soit atte´nue´ quoi¼ It it should be decreased¼ Ok Ok ¼parc ¼ que quand j’ai un sac sur le dos ¼because if I have a backpack oui Yes et bein vraiment j’ai ces- (.) j’ai les skis qui partent quoi Well really I have those - (.) I have the skis parting Mhm mhm au bout d’une heure de temps ça me tire bien comme il fa:ut After an hour it starts pulling the way it’s suppo:sed to Ok oui (1.0) Ok yes (1.0) Donc ça c ¼ serait aussi un autre objectif c’est qu ¼ a` a` moyen terme disons Well this would be another goal as a mid-term goal let’s say

diverse: patient’s psychology, characteristics of illness, patient’s resources for communication and physician communication, as well as socio-demographic characteristics, and variables related to the visit characteristics (Robinson, 2003; Thompson et al., 2007). The strength of this detailed analysis is that it is based on actual patient–physiotherapist interaction and not on physiotherapists’ accounts of goal setting. Focus groups and interviews might help detect representations and conceptualisations of this process, but are never able to provide information about the interaction. The analysis of exchanges between patients and physiotherapists illustrates patterns of goal setting activities in an orthopaedic outpatient setting. Parry (2005) argues that a conversation analytic approach might ‘‘help therapists analyse and develop their own and others’ conduct in the area of interaction’’ (p. 211). Another researcher notes that single extracts of interaction can offer an indepth analysis of ‘‘several different domains of phenomena’’ (Schegloff, 1984; p. 111). Extract 4.1 PT3 – Pat C. 6 Physio 7 Patient 8 Physio 9 Patient

669

¼on est plutoˆt sur [l’apprentissage () ¼it is rather for [teaching () [c’est juste c’est juste] [that’s right that’s right] j’ai bien compris d’acc[ord I understood alright Ok[ay. [c’est juste [that’s correct

57 Physio 58 59 Patient 60 Physio 61 Patient

Ok (2.0) donc euh je crois qu’on a tout pour l’instant Ok (2.0) well I think it is all for now On se revoit de´ja` [demain We see each other already tomorrow [again oui 8 heures Yes at 8 o’clock et puis euh on partira la` d ¼ ssus (.) Ça [joue And well eh we will be starting with that (.) Is that [ok c’est bien (.) nickel That’s ok (.) super

4.1. Limitations of the study Limitations exist in regards to the context of the study. The participating physiotherapists were about to be trained in negotiation skills, and they were aware of the fact that the audio-tapes would be subsequently used for discussion during training. Therefore, they were potentially more sensitive to the topic. It could be argued that similar practices might not be found in another sample of physiotherapists. However, even though the participating physiotherapists were aware of the purpose of the audiotape, they continued to follow their understanding of professional practice. Another important limitation can be found in using audio-taped interactions. Physiotherapists very often use non-verbal communication and body language. The body matters in physiotherapy and should therefore always be taken into consideration. Videorecording of treatment sessions would have provided insight into those aspects of the interaction, but this option was considered too intrusive. Another limitation is related to the quality of the audiorecording. The small digital voice-recorder does not provide high quality recordings. Additional wireless microphones could enhance reliability (Pera¨kyla¨, 2004). This pilot study has a very small sample size. Future studies should include a wider data set using video-recordings in order to understand all aspects of communication.

5. Conclusion This study sheds light on the goal-setting process in orthopaedic physiotherapy. In order to formulate and negotiate treatment goals collaboratively, considerable effort is required by both participants. Evidence from this small pilot study shows that in order to gain a deeper understanding of the patients’ clinical problems physiotherapists are required to use a more participative approach. Through this process treatment goals can be made explicit. On the other hand, the results also suggest that if this effort is reduced, patient involvement is less likely to happen. This generates an imbalance of inputs which undermines the concept of shared decision-making. How to achieve this balance, and to what extent this is desirable, are questions to be investigated in the future. Qualitative methods in general and CA in particular provide useful tools to describe actual practices. Manual therapists may learn from the evidence of this study and reﬂect on their own interaction with patients.

Acknowledgment The study was ﬁnanced by DO-RE Funds of the Swiss National Science Foundation (13DPD3-108451) and the Swiss Physiotherapy Association’s Research Funds. Thanks to all participants and to Alison Pilnick, Robert Dingwall and Ruth Parry for their feedback.

670

V. Schoeb / Manual Therapy 14 (2009) 665–670

Appendix 1. Transcription convention (Jefferson, 1984).

Symbols

Explanation

[]

Indicates the point where overlap begins and ends Indicates elapsed time in silence in tenths of a second either within or between utterances Indicates a gap of less than 0.1 s Arrow up indicates a rising shift in intonation prior to the word Horizontal dash indicates that the word sounds abruptly ‘‘cut off’’ Indicates quieter passage of talk compared to the surrounding talk Indicates an extension of the syllable it follows Indicates that there is no interval between two utterances Indicates that the transcriptionist is not able to hear the utterance Indicates a description of a phenomenon (e.g. laughter, noise.)

(0.0)

(.) [ –

: ¼ () (())

References APTA American Physical Therapy Association. A normative model of physical therapy professional education: version 2004. Alexandria, VA: Department of Education APTA; 2004. Arminen I. Institutional interaction: studies of talk at work, vol. 2. Hants, UK: Ashgate; 2005. Baker SM, Marshak HH, Rice GT, Zimmerman GJ. Patient participation in physical therapy goal setting. Physical Therapy 2001;81(5):1118–26. Bradley EH, Bogardus ST, Tinetti ME, Inouye S. Goal-setting in clinical medicine. Social Science & Medicine 1999;49:267–78. Charles C, Gafni A, Whelan T. Shared decision-making in the medical encounter: what does it mean? (or it takes two to tango). Social Science & Medicine 1997;44(5):681–92. Charles C, Gafni A, Whelan T. Decision-making in the physician–patient encounter: revisiting the shared treatment decision-making model. Social Science & Medicine 1999;49:651–61. Clayman SE, Gill VT. Conversation analysis. In: Bryman HM, editor. Handbook of data analysis. London: Sage; 2004. p. 589–606. Collins S, Drew P, Watt I, Entwistle V. ‘Unilateral’ and ‘bilateral’ practitioner approaches in decision-making about treatment. Social Science & Medicine 2005;61:2611–27. De Haes H. Dilemmas in patient centeredness and shared decision making: a case for vulnerability. Patient Education and Counseling 2006;62(3):291–8. Drew P, Chatwin J, Collins S. Conversation analysis: a method for research into interactions between patients and health-care professionals. Health Expectations 2001;4:58–70. Goldberg JA. The amplitude shift mechanism in conversational closing sequences. In: Lerner GH, editor. Conversation analysis: studies from the ﬁrst generation. Amsterdam/Philadelphia: John Benjamins; 2004. p. 257–97. Guadagnoli E, Ward P. Patient participation in decision-making. Social Science & Medicine 1998;47(3):329–39. Heritage J. Conversation analysis and institutional talk: analysing data. In: Silverman D, editor. Qualitative research: theory, method and practice. London: Sage; 2004. p. 222–45 [chapter 12]. Heritage J. Conversation analysis and institutional talk. In: Fitch KL, Sanders RE, editors. Handbook of language and social interaction. Mahwan, NJ: Lawrence Erlbaum; 2005. p. 103–47 [chapter 5]. Heritage J, Maynard DW, editors. Communication in medical care: interaction between primary care physicians and patients. Cambridge: University Press; 2006 [chapter 1]pp. 1–22. Higgs J, Jones M. Clinical reasoning in the health professions. 2nd ed. Oxford: Butterworth-Heinemann; 2000. Holliday RC, Antoun M, Playford DE. A survey of goal-setting methods used in rehabilitation. Neuro Rehabilitation and Neural Repair 2005;19:227–31.

Jefferson G. Transcript notation. In: Atkinson JM, Heritage J, editors. Structures of social action: studies in conversation analysis. Cambridge: University Press; 1984. p. ix–xvi. Levack WM, Dean SG, Siegert RJ, McPherson KM. Purposes and mechanisms of goal planning in rehabilitation: the need for a critical distinction. Disability and Rehabilitation 2006a;28(12):741–9. Levack WM, Taylor K, Siegert RJ, Dean SG. Is goal planning in rehabilitation effective? A systematic review. Clinical Rehabilitation 2006b;20:739–55. Levinson W, Kao A, Kuby A, Thisted RA. Not all patients want to participate in decision-making: a national study of public preferences. Journal of General Internal Medicine 2004;20:531–5. Lewin SA, Skea ZC, Entwistle V, Zwarenstein M, Dick J. Interventions for providers to promote a patient-centered approach in clinical consultations. The Cochrane Database of Systematic Reviews 2001;4. Art. No.: CD003267. doi10.1002/ 14651858.CD003267. Mead N, Bower P. Patient-centeredness: a conceptual framework and review of the empirical literature. Social Science & Medicine 2000;51:1087–110. Michie S, Miles J, Weinmann J. Patient-centredness in chronic illness: what is it and does it matter? Patient Education and Counseling 2003;51(3):197–206. Parry RH. Communication during goal-setting in physiotherapy treatment sessions. Clinical Rehabilitation 2004;18:668–82. Parry R. A video analysis of how physiotherapists communicate with patients about errors of performance: insights for practice and policy. Physiotherapy 2005;91:204–14. Payton OD, Nelson CE, St. Clair Hobbs M. Physical therapy patients’ perception of their relationship with health care professionals. Physiotherapy Theory and Practice 1998;14:211–21. Pera¨kyla¨ A. Reliability and validity in research based on naturally occurring social interaction. In: Silverman D, editor. Qualitative research: theory, method and practice. London: Sage; 2004. p. 283–304 [chapter 15]. Peri K, Kerse N, Parsons J, McPherson KM, Latham N, Parsson N. Physical rehabilitation involving goal setting for older people living in the community (Protocol). Cochrane Database of Systematic Reviews 2006;1. No. CD005572. Physioswiss. Das Wesen der Physiotherapy. Swiss Physiotherapy Association, http:// www.ﬁsio.org/de/index.cfm%3FtreeID ¼ 532; 2007 [accessed 27.02.07]. Playford ED, Dawson L, Limbert V, Smith M, Ward CD, Wells R. Goal-setting in rehabilitation: report of a workshop to explore professionals’ perceptions of goal-setting. Clinical Rehabilitation 2000;14(5):491–6. Poitras S, Blais R, Swaine B, Rossignol M. Management of work-related low back pain: a population-based survey of physical therapists. Physical Therapy 2005;85(11):1168–81. Robinson JD. An interactional structure of medical activities during acute visits and its implications for patient’s participation. Health Communication 2003;15(1):27–59. Schegloff EA. Analyzing single extracts of interaction: an exercise in conversation analysis. Social Psychology Quarterly 1984;50(2):101–14. Schegloff EA. Sequence organization in interaction: A primer in conversation analysis 1. Cambridge: University Press; 2007. Schulman-Green DJ, Naik AD, Bradley EH, McCorkle R, Bogardus ST. Goal setting as a shared decision making strategy among clinicians and their older patients. Patient Education and Counseling 2006;63:145–51. Stewart M, Brown JB, Weston WW, McWhinney IR, McWilliam CL, Freeman TR. Patient-centered medicine. Transforming the clinical method. Thousand Oaks: Sage; 1995. Stivers T. ‘‘No no no’’ and other types of multiple sayings in social interaction. Human Communication Research 2004;30(2):260–93. Ten Have P. Doing conversation analysis: a practical guide. London: Sage; 1999. Ten Have P. Understanding qualitative research and ethnomethodolgy. London: Sage, http://site.ebrary.com/lib/uon/Doc%3Fid ¼ 1008097&ppg ¼ 24; 2004 [accessed 24.01.07]. Thompson A, Ruusuvuori J, Britten N, Collins S. An integrative approach to patient participation in consultations. In: Collins S, Britten N, Ruusuvuori J, Thompson A, editors. Patient participation in health care consultations: qualitative perspectives. New York: Open University Press – McGraw Hill; 2007. p. 176–93 [chapter 10]. WCPT. Declaration of principles and position statements, http://www.wcpt.org/ common/docs/WCPTPolicies.pdf; 1999 [accessed 28.02.07]. West C. Coordinating closings. In: Heritage J, Maynard DW, editors. Communication in medical care: interaction between primary care physicians and patients. Cambridge: University Press; 2006. p. 379–415 [chapter 13]. WHO World Health Organisation. A glossary of terms for community health care and services for older persons. Aging and health technical report, http://whqlibdoc. who.int/wkc/2004/WHO_WKC_Tech.Ser._04.2.pdf; 2004 [accessed 26.02.07]. Wressle E, Oberg B, Henriksson C. The rehabilitation process for the geriatric stroke patient – an exploratory study of goal setting and interventions. Disability and Rehabilitation 1999;21(2):80–7.

Manual Therapy 14 (2009) 671–678

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

Pillow use: The behaviour of cervical pain, sleep quality and pillow comfort in side sleepers Susan J. Gordon a, *, Karen Grimmer-Somers b, Patricia Trott b a b

School of Public Health, Tropical Medicine and Rehabilitation Sciences, James Cook University, Australia School of Health Sciences, University of South Australia, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 24 September 2008 Received in revised form 6 February 2009 Accepted 24 February 2009

A random allocation single blind block design pillow ﬁeld study was undertaken to investigate the behaviour of cervico-thoracic spine pain in relation to pillow use. Participants (N ¼ 106) who reported preference for side sleep position with one pillow were recruited via a telephone survey and newspaper advertisement. They recorded sleep quality and pillow comfort ratings, frequency of retiring and waking cervical pain and duration of waking cervical pain while sleeping for a week on their usual pillow, polyester, foam, feather and rubber pillows of regular shape and a foam contour pillow. Analysis was undertaken comparing sleep quality, pillow comfort, waking and temporal cervical pain reports, between the usual pillow and the trial pillows, between pillows of differing content and foam pillows of differing shape. This study provides evidence to support recommendation of rubber pillows in the management of waking cervical pain, and to improve sleep quality and pillow comfort. The rubber pillow performed better than subjects’ own pillow in most instances. Subjects’ own pillow performed similarly to foam and polyester pillows, and there is no evidence that the use of a foam contour pillow has advantages over the regular shaped pillows. Feather pillows should not be recommended. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Pillow Cervical pain Sleep quality Pillow comfort

1. Introduction Pillow performance research has largely involved testing pillows of novel shape and design (Erfanian et al., 1998; Persson and Moritz, 1998; Erfanian et al., 2004), comparing contour and noncontour shaped pillows (Lavin et al., 1997; Persson, 2006) and comparing contour pillows with the participants’ usual pillow (Hagino et al., 1998). Shields et al. (2006), who undertook a systematic review regarding the effect of contour or cervical pillow use on neck pain, highlighted the methodological ﬂaws in these studies and concluded that there was insufﬁcient evidence to support the use of contour pillows in the management of chronic neck pain. Helewa et al. (2007) reported that contour pillows were ineffective in the management of chronic neck pain unless combined with active neck exercises. The paucity of research has caused health professionals to provide patient advice based on the anecdotal suggestions of expert colleagues and professional associations. This advice has included the use of malleable pillows (Maitland, 1986; McKenzie and May,

2006), a cervical roll (Elkind, 1987; Kramer, 1990; McKenzie and May, 2006), a contour pillow (Jackson, 1976; Emberson, 1985) or a down or urethane pillow (Australian Physiotherapy Association, 2008). Furthermore the range of marketing advice provided by pillow manufacturers is confusing for consumers. This paper reports the performance of commonly used pillows and their association with cervical pain behaviour. A pillow ﬁeld trial was undertaken to: Compare the frequency of waking cervical pain reported when subjects slept on their own pillow and on ﬁve trial pillows; Examine temporal symptom reports, to determine if pillow content or shape was related to overnight abolition of retiring symptoms or overnight production of waking symptoms; and Compare pillow comfort and sleep quality ratings for participants’ usual pillow and the trial pillows.

1.1. Preliminary research * Corresponding author. Tel.: þ61 7 4781 6734; fax: þ61 7 4781 6868. E-mail address: [email protected] (S.J. Gordon). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.02.006

A telephone survey of over 800 individuals randomly selected from a known population established that:

672

S.J. Gordon et al. / Manual Therapy 14 (2009) 671–678

Side sleeping position was most prevalent (71.9%) and this position protected individuals against waking cervical pain (odds ratio (OR) 0.6; conﬁdence intervals (CIs) 0.4–0.9) (Gordon et al., 2007). Further it has been reported that adults spend 59–73% of their sleep in the side lying position (DeKoninck et al., 1992). Side sleepers, who slept on one pillow, used a variety of pillows including a polyester regular pillow (44.9%), rubber regular pillow (14.4%), foam contour pillow (12%), feather regular pillow (8.8%) and foam regular pillow (7.7%) (Gordon et al., 2007). 17.6% of subjects reported waking, at least once in a usual week, with cervical pain (Gordon et al., 2002, 2007).

People who reported an accident or injury affecting the cervicothoracic spine in the preceding year. 2.5. Sample size No studies were available on which to base sample size calculations. We considered that 100 participants was the largest feasible sample size given study constraints of time and funding. 2.6. Study management

2. Method

An independent trial ofﬁce was established to co-ordinate subject allocation to pillows, pillow delivery and data collection. The ﬁve trial pillows were randomly allocated into a ﬁve-block administration order design. Subjects were randomly allocated into receiving one of the blocks. The trial pillows were de-identiﬁed by removing identifying labels and covers, numbering them and placing them in a plain pillow case. Pillows were re-used and not cleaned or laundered during the trial. For hygiene reasons participants were encouraged to use a second pillowcase on the pillow.

2.1. Ethics approval

2.7. Data collection

Ethics approval was provided by the University of South Australia Human Research Ethics Committee.

Subjects initially provided their age and gender. No information was captured about subjects’ own pillow type. Outcome data was recorded on a seven day-night diary for each pillow. Subjects reported retiring and waking cervical pain, the duration of waking cervical pain, pillow comfort and sleep quality ratings. Subjects also provided free-text comments regarding their perceptions of the trial pillows on the last day of the pillow trial. The diary is provided in Appendix 1.

A sleep laboratory study established the validity and reliability of self reports of sleep position (Gordon et al., 2004). Therefore, it seemed appropriate to test the effect of pillow use on cervical pain and sleep quality in side sleepers.

2.2. Pillows tested Tontine, an Australian pillow manufacturer (Nicholson Street, East Brunswick, Victoria) provided polyester, synthetic ﬁbre ﬁll, pillows for the trial. Dentons, another Australian pillow manufacturer (Lewis Road, Wantirna South, Victoria) provided foam regular shape (Comfort Classic) and foam contour shape pillows (Medirest). Both foam pillows were molded from high density foam. Dunlopillo latex rubber pillows were provided by the University of South Australia and feather pillows were purchased by the principal author from Target, Australia. The pillows varied in length from 70 to 73 cm and in width from 45 to 46 cm. The depth of the foam regular pillow was 120 mm, the foam contour pillow varied between 120 and 142 mm across the contour, the latex pillow was 115 mm, the feather pillow was 120 mm and the polyester pillow was 118 mm. The study was conducted independently of any additional involvement of pillow manufacturers. Further information regarding the latex pillow can be found at http://www. dunlopillo.com.au/Products/Pillow-Range/Latex-Range.asp, the foam regular pillow at http://www.dentons.com.au/comfort.htm and the foam contour pillow at http://www.dentons.com.au/ therap.htm. 2.3. Study design A random allocation block design ﬁeld study was undertaken in subjects’ homes, in a large rural town in South Australia, Australia. Subject blinding to pillow type was attempted, although the shape and feel of some of the pillows may have constrained this. 2.4. Participants A sample of convenience was recruited from the respondents to the preliminary studies (Gordon et al., 2002, 2007) and additional subjects were recruited via local newspaper advertisements. Inclusion criteria: Over 18 years, reported usually sleeping on their side, on one pillow, and were not actively seeking treatment for their neck during the period of the trial. Exclusion criteria:

2.8. Pillow intervention The ﬁeld trial took 10 weeks. Data was initially captured on subject’s own pillow for a week, to establish baseline ‘usual’ symptoms. The subjects’ own pillow was chosen as the comparison ‘gold standard’ pillow on the assumption that it was likely to be the most comfortable pillow subjects had encountered. Over the next nine weeks, subjects tested each of the randomly allocated trial pillows for seven consecutive nights, interspersed by seven night’s sleep on their own pillow. Returning to their own pillow after each pillow trial provided a ‘washout period’ (Yin, 2003) which allowed subjects to return to their ‘normal’ sleeping state. Subjects were encouraged to test each trial pillow for the whole seven nights’ sleep unless they believed that symptom production or lack of sleep necessitated cessation of the trial of that pillow. 2.9. Data management Each subject could potentially provide seven nights’ data from each six pillows (42 observations each). 2.9.1. Invalid data Throughout the trial, the occasions on which subjects reported the presence of waking symptoms associated with identiﬁable causes (other than the pillow) were excluded from analyses, in order to retain a homogenous set of subjects for whom there was no other identiﬁable reason for waking pain except the pillow. 2.9.2. Drop-outs Subjects who dropped out of any week’s trial of a particular pillow were identiﬁed, and their ‘missing’ data removed from analysis once they had ceased providing data for that pillow. This left only valid waking pain scores for analysis. It was important to

S.J. Gordon et al. / Manual Therapy 14 (2009) 671–678

count those subjects who did not complete the week’s trial, as this was believed to reﬂect dislike of that particular pillow. Thus the number of subjects who completed the trial was reported, as well as the number of ‘valid’ days of data collection. 2.9.3. Temporal pain To analyse temporal pain patterns and pillow performance subjects who provided valid data were sub-classiﬁed into four groups: (1) going to bed and waking with no pain (pillow did not produce symptoms) (2) going to bed with pain and waking with no pain (a positive effect of the pillow) (3) going to bed with no pain and waking with pain (a negative effect of the pillow), and (4) going to bed with pain and waking with pain (pillow had no effect on symptoms).

2.9.4. Cumulative waking pain score A cumulative numeric pain frequency and duration score were developed to provide a per-week waking pain frequency–duration score per pillow using the seven day sum of pain frequency and duration values as indicated: subjects who woke without any pain on any day were assigned a score of 0 subjects who woke with pain lasting up to 30 min on any day were assigned a score of 0.5 (half an hour) subjects whose waking pain on any day lasted half a day were assigned a score of 12 (12 h) and subjects whose waking pain on any day lasted all day were assigned a score of 18 (18 h).

673

were described and tested in this manner. Signiﬁcance for tests of proportion was set at p < 0.05. The cumulative walking pain scores over the trial pillow weeks were examined for pillow type, subject age and gender effects using ANOVA models. Tests for differences were signiﬁcant at p < 0.05. Descriptive analysis of the cumulative scores was by means and standard deviations. ORs (95% CIs) were calculated to detect differences between pairs of pillows for waking pain (compared with no waking pain). The comparator in each case was subjects’ own pillow. 2 2 tables (StatCalc in EpiInfo Version 10) were used to calculate ORs and 95% CIs. When 95% CIs did not encompass 1, the ﬁnding was deemed to be signiﬁcant. Where required, multivariate logistic regression analysis (SAS Version 8.02) was used to test associations between pillows and pain categories, and to test the confounding effects of age group and gender. As for the 2 2 table comparisons, 95% CIs which did not encompass 1 indicated a signiﬁcant ﬁnding. Significant confounding effects were indicated by p values < 0.05 for each independent variable in the model. 3. Results 3.1. Demographics Of the 106 participants who commenced the own pillow trial 33 were male (average age 49.0 years (SD 14.3 years, range 23–76 years)); and 73 were female, (average age 49.9 years (SD 13.9 years, range 20–81 years)). Fifty-eight participants had participated in the previous telephone survey and a further 48 participants were recruited via newspaper advertisement. 3.2. Pillow trial completion

2.10. Analysis

3.2.1. Study withdrawals Seven participants withdrew from the study at various points throughout the ten weeks; three due to production of cervicothoracic symptoms while trialling the feather pillow, and single participants due to emergency lumbar spine surgery following a fall, transfer from town for work purposes, a loss of interest in participating while trialling their usual pillow, and death of a spouse. Given the random allocation block design, this inﬂuenced the number of subjects who commenced each trial pillow week. Considering the total number of starters (male, female), 105 subjects commenced the polyester pillow trial (72, 33), 101 commenced the foam regular pillow trial (69, 32), 103 commenced the foam contour pillow trial (71, 32), 101 commenced the feather pillow trial (71, 30) and 100 commenced the rubber pillow trial (70, 30).

The challenge in this analysis was to make sense of multiple categories of information on temporal pain change and waking pain, and to determine per pillow changes between pain categories for each day that subjects remained in the study. Descriptive analysis was used to report frequency distributions in terms of the percentage of subjects in each category of waking pain, temporal pain change, quality of sleep and pillow comfort for each pillow. A survival analysis approach was used to describe drop-outs throughout each of the trial pillow weeks, as subjects who refused to continue in the trial were considered to be those who sustained adverse effects of sleeping on the pillow in question. This was important to ﬂag as evidence of ‘treatment harm’. Differences in percentage of subjects in each category of waking pain, and the categories of temporal pain change over a night of sleeping, were tested using Chi Square test of proportions. The entire dataset, and gender and age subgroups where appropriate,

3.2.2. Invalid data Subjects whose data was excluded for any one day of the trial because of a known cause for waking pain are outlined in Table 1, with the number of observations excluded. Reasons for waking pain included the effects of alcohol, prescription medication, illness, wakeful spouse, children or pets, or external noises. There was no signiﬁcant difference between pillows, gender or age groups regarding the excluded observations. The daily reports of known reasons for waking pain were greatest for subject’s own pillow, followed by the feather pillow (Fig. 1). As the own pillow comparison data was provided ﬁrst, this suggests that subjects were perhaps most aware of classifying reasons for waking pain on their own pillow, than on any others. The number of subjects who completed each pillow trial varied, with the feather pillow having the lowest percentage of completers and rubber pillow having the highest. All subjects completed the

These arbitrary scores were assigned to investigate the extent of waking pain. 2.9.5. Clusters of waking pain cumulative scores It was hypothesized that there would be clusters of cumulative waking pain scores for each pillow. These would be classiﬁed as No pain (0 score), Occasional short term pain (1–3 days of pain lasting 30 min), Regular short term pain (4–7 days of pain lasting 30 min), Occasional half day pain (1–3 days of pain lasting half a day), Regular half day pain (4–7 days of pain lasting half a day) and Longer term pain (regular pain lasting for longer).

674

S.J. Gordon et al. / Manual Therapy 14 (2009) 671–678

Table 1 Data excluded per pillow. Pillow

Total possible obs

Total completed obs

N subjects providing reasons for exclusion

Total excluded obs (invalid)

Total valid obs

Dropouts Total obs lost through dropouts

Own Polyester Foam regular Foam contour Feather Rubber

742 732 736

742 713 706

50 [12 men, 38 women] 33 [11 men, 22 women] 19 [4 men, 15 women]

107 66 37

635 647 669

0 5 6

0 19 30

730

682

15 [3 men, 12 women]

45

637

12

48

707 700

563 691

14 [3 men, 11 women] 21 [5 men, 16 women]

18 40

545 651

33 3

144 9

Obs ¼ observations.

‘own pillow’ trial week. Overall, 100 subjects completed the polyester pillow trial week (95.4% of commencers), 95 completed the foam regular pillow trial week (94.1% of commencers), 91 completed the foam contour pillow trial week (88.3% of commencers), 68 completed the feather pillow trial week (67.3% of commencers) and 97 completed the rubber pillow trial week (97.0% of commencers). 3.3. Symptom behaviour 3.3.1. Waking pain The percentage of subjects with valid observations who reported any waking pain on any pillow is reported in Fig. 2. The feather pillow was by far the most problematic pillow. Considering the overall data, no signiﬁcant differences were found between proportions in waking pain categories, despite apparent differences from visual inspection. The lack of signiﬁcance is possibly attributed to the variability within any one category of waking pain across the pillows. Chi square tests of proportions identiﬁed a signiﬁcant effect for young women (aged under 40 years) (p < 0.05). Further exploration using ANOVA models showed that women aged under 40 years reported on average signiﬁcantly more events of waking pain than any other group of subjects (average 0.6 (SD 0.5) (women 40–59 years average 0.2 (SD 0.4)), women over 59 years (average 0.3 (SD 0.5)), compared with men

Fig. 1. Day-by-day percentage of subjects with known reasons for waking pain.

under 40 years average 0.3 (SD 0.4), men 40–59 years (0.2 (SD 0.4)) and men over 59 years (0.1 (SD 0.4))). 3.3.2. Duration of waking pain At least 30% subjects recorded no waking pain on any pillow throughout the trial. Using the pain frequency–duration classiﬁcations, different waking pain category proﬁles were identiﬁed for each pillow. Table 2 outlines the overall percentage of subjects with the different pain proﬁles. Examining the data of only those subjects who reported no waking pain on their own pillow, we considered their responses to the trial pillows (see Fig. 3). The polyester, foam contour and rubber pillows were most likely to continue the pattern of no waking pain reports. The regular foam and feather pillows produced more frequent events of waking pain. 3.3.3. Comparing the percentage change in pain symptoms between pillows Considering all subjects with valid data patterns of waking pain change between own pillow and trial pillows were compared. Table 3 outlines the percentage of subjects who remained the same, improved or worsened. There was no signiﬁcant gender or age effect on change in symptom distribution between own pillow and the trial pillows. Fig. 4 reports a survival analysis of subjects who remained free of waking pain on their own and each of the trial pillows throughout the seven day trial. The subjects’ own pillow and the feather pillow produced the greatest loss in the sample to cervical waking pain, whilst the rubber pillow was clearly superior. There was variable probability of subjects’ waking with pain on their own pillow, compared with each of the trial pillows. Accumulated valid data from all subjects who commenced each trial, found no difference in odds (95% CI) when comparing own

Fig. 2. The percentage of subjects who reported any waking pain.

S.J. Gordon et al. / Manual Therapy 14 (2009) 671–678

675

Table 2 Percentage of valid subject-observations classiﬁed into waking pain duration subsets overall% (male %, female %). Pillow

No pain

Occasional short term pain Regular short term pain Occasional half day pain Regular half day pain Longer term pain Excluded obs (%)

Own pillow Polyester Foam regular Foam contour Feather Rubber

42.5% (37.8%, 62.2%) 51% (37.3%, 62.8%) 54.2% (36.5%, 63.5%) 59.4% (40.4%, 59.7%) 48.2% (35.9%, 64.1%) 66.3% (33.9%, 6.2%)

3.8% (75%, 25%) 7% (14.3%, 85.7%) 11.5% (27.3%, 72.7%) 7.3% (0%, 100%) 12.4% (20%, 80%) 7.1% (42.9%, 57.1%)

1.9% (0%, 100%) 1% (0%, 100%) 0% (0%, 0%) 1.0% (0%, 100%) 1.2% (0%, 100%) 0% (0%, 0%)

3.8% (25%, 75%) 5% (0%, 100%) 11.5% (27.3%, 72.7%) 12.5% (25%, 75%) 17.3% (42.9%, 57.1%) 4.1% (0%, 100%)

0% (0%, 0%) 1% (0%, 100%) 2.1% (0%, 100%) 3.1% (33.3%, 66.7%) 3.7% (33.3%, 66.7%) 1% (0%, 100%)

0.94% (0%, 100%) 2% (0%, 100%) 1% (0%, 100%) 1.0% (0%, 100%) 0% (0%, 0%) 0% (0%, 0%)

47.1 33.0 19.7 15.7 17.2 21.5

Fig. 3. The responses of subjects who reported no waking pain on their own pillow, regarding waking pain patterns on the trial pillows.

pillow with polyester (1.1 (0.7–1.6)), foam regular (1.1 (0.8–1.6)) or foam contour (1.4 (0.9–2.0)). The odds were signiﬁcantly doubled for feather (1.9 (1.3–2.7)) and signiﬁcantly halved for rubber (0.6 (0.4–0.9)). Considering those subjects who reported occasional pain on their own pillow, there was no difference in probability of waking pain when comparing own pillow with polyester (0.9 (0.4–2.0)), foam regular (0.9 (0.4–2.0)) or foam contour (1.0 (0.5–2.1)). Again the odds were signiﬁcantly doubled for feather (2.4 (1.2–4.9)) and trended towards signiﬁcantly halved for rubber (0.4 (0.1–1.0)). Considering those subjects who reported frequent pain on their own pillow, there was no difference in probability when comparing own pillow with polyester (0.8 (0.5–1.4)), foam regular (1.3 (0.8– 2.2)) or foam contour (1.2 (0.7–2.0)). The odds were almost signiﬁcantly doubled for the feather pillows (1.7 (1.0–2.9)). The effect of the rubber pillow was not signiﬁcant (0.8 (0.5–1.4)).

Fig. 4. Survival analysis of the percentage of subjects who commenced each pillow trial who had valid observations and remained free of cervical waking pain on each consecutive day of the pillow trial.

3.3.4. Temporal symptoms and pillow use The rubber pillow was the best of all pillows at ensuring that subjects who went to bed without cervical pain woke without any, whilst the feather pillow was the worst. The pillow shape (contour compared with regular) appeared to make little difference to temporal patterns. Fig. 5 outlines the performance of the trial

Table 3 Percentage subjects with changed pain patterns when comparing own pillow with each trial pillow. Own pillow compared with

Worse (%)

Remained the same (%)

Improved (%)

Polyester Foam regular Foam contour Feather Rubber

10.4 16.9 9.3 16.0 6.0

77.1 60.4 79.6 62.0 82.0

12.5 22.7 11.1 22.0 12.0 Fig. 5. Performance of subjects with temporal symptoms on trial pillows.

676

S.J. Gordon et al. / Manual Therapy 14 (2009) 671–678

Table 4 Percentage of subjects reporting high pillow comfort and sleep quality ratings.

High comfort rating High sleep quality rating

Own (%)

Polyester (%)

Foam regular (%)

Foam contour (%)

Feather (%)

Rubber (%)

85 71

81 71

80 72

71 63

41 40

91.5 81

pillows in terms of changing temporal patterns. The odds (OR, 95% CI) of subjects waking worse than they retired in terms of cervical pain, when compared with their own pillow, were not signiﬁcant for the polyester (1.3 (0.6–2.6)), foam regular (1.3 (0.6–2.7)) and foam contour (1.4 (0.7–2.9)), and rubber pillows (0.7 (0.3–1.7)). The odds of this occurring were signiﬁcantly elevated with the feather pillow (2.4 (1.2–2.7)). There was a signiﬁcant difference in the proportion of subjects across the temporal pain categories when comparing the feather pillow and own pillow. There was no difference in proportions when comparing the other pillows. 3.4. Sleep quality and pillow comfort The percentages of participants’ valid observations who reported high sleep quality and high pillow comfort ratings are presented in Table 4. In Table 5, in comparison to their usual pillow, subjects were signiﬁcantly more likely to report low pillow comfort on all trial pillows except the rubber pillow, which signiﬁcantly protected subjects from a low pillow comfort rating. The foam contour and the feather pillows were signiﬁcantly more likely to produce low sleep quality, whilst the rubber pillow signiﬁcantly protected subjects. The comfort and quality ratings for the feather pillow were signiﬁcantly lower when compared with own pillow. The remaining pillows did not differ signiﬁcantly from the ratings provided for subjects’ own pillow. 4. Discussion This study provides the ﬁrst reports that pillow type can be recommended by health practitioners to alter the behaviour of cervical pain in side sleepers. It provides the basis for further investigation of pillow performance with speciﬁc groups of people (with known musculoskeletal problems or other sleeping positions) and a basis for evidence-based prescription of pillows for individuals suffering from regular waking pain, reduced pillow comfort or sleep quality. The study ﬁndings are in direct conﬂict with historically held anecdotal advice regarding pillow selection in the management of cervico-thoracic symptoms . 4.1. Pillow performance Subjects’ own pillow performed similarly to the polyester and foam pillows in terms of production of waking symptoms and maintenance of retiring pain. The shape of the foam pillow appeared to make no difference to waking pain or abolition of night pain. However the contour pillow was less comfortable and provided poorer quality sleep. Thus the contour pillow is less efﬁcacious for these reasons. Contour shaped foam pillows were initially developed to support the cervical lordosis in the supine

sleep position. Hence further investigation of the association between contour pillow shape and symptom behaviour in supine sleepers is indicated. Moreover, contour pillows of different heights require examination with respect to subject anthropometry, symptom behaviour, sleep quality and pillow comfort ratings. The feather pillow was a consistent poor performer in all outcome measures and therefore cannot be recommended as an alternative should subjects request a pillow which is better than their own. However the rubber pillow performed consistently well, and was a better performer than subjects’ own pillow in all outcome measures and should be recommended as an alternative should subjects seek a better performing pillow than their own. 4.2. Pillow trial period Seven days appears to be a suitable period for a pillow trial as all drop-outs occurred before the ﬁfth trial day. The ‘washout period’ of own pillow use for seven days between pillow trials also appeared to be appropriate to reduce trial pillow symptoms, to catch up on sleep from poor quality sleep from the trial pillows and to retain the interest of the study sample. 4.3. Anthropometry The ﬁt of pillow-to-human form has not been reported in the literature, and was not investigated in this study. Anthropometric studies may thus provide useful information regarding if, and what, anatomical parameters will ensure a comfortable, symptom free union between person, mattress and pillow. 4.4. Study limitations There were several limitations in this study including an inability to completely blind subjects for pillow type, reliance on daily self-report measures of pain occurrence and duration, sleep quality and pillow comfort, and a lack of information on subjects’ own pillows. There was a surprisingly high number of waking pain reports on subjects’ own pillows, questioning our assumption that these were the most comfortable pillows ever used by subjects. Although there was no difference in subject reports of known reasons for waking pain between the trial pillows, there was a noticeably high percentage on subjects’ own pillows in the ﬁrst trial week. This could lead to questions such as ‘Did this occur because subjects were anxious about the trial, or perhaps more aware of waking pain?’ If this is so, it must be considered that reports of cervical waking pain on the trial pillows may well have been related to known reasons other than the pillow, but were perhaps ascribed to the pillow itself, in error (Sackett, 1979). The potential for over- and under-reporting of symptoms during the trial pillow weeks must be considered.

Table 5 Odds ratios (95% conﬁdence intervals) comparing low sleep quality and pillow comfort ratings between the subjects’ usual pillow and each trial pillow. Signiﬁcant ﬁndings are in bold. Compared to usual pillow OR (95% CI)

Polyester

Foam regular

Foam contour

Feather

Rubber

Low comfort rating Low sleep quality

1.4 (1.1–1.9) 1.0 (0.8–1.3)

1.5 (1.1–2.0) 1.0 (0.8–1.3)

2.4 (1.8–3.1) 1.5 (1.2–1.8)

8.4 (6.4–11.0) 3.7 (3.0–4.8)

0.5 (0.4–0.8) 0.6 (0.5–0.7)

S.J. Gordon et al. / Manual Therapy 14 (2009) 671–678

2. Please choose one of the categories below to rate the comfort of your pillow last night.

5. Conclusion This study provides evidence to support recommendation of rubber pillows in the management of waking cervical pain and to improve sleep quality and pillow comfort. The rubber pillow performed better than subjects’ own pillow in most instances. Subjects’ own pillow performed similarly to foam and polyester pillows and there is no evidence that the use of a foam contour pillow has advantages over the regular shaped pillows. Feather pillows should not be recommended. Appendix 1. Diary As you were preparing to go to bed tonight did you have:Please tick Neck pain Neck stiffness Headache Aching between your shoulder blades

Morning 1 (after Night One) 1. Please choose one of the categories below to describe the quality of your sleep last night. Please tick one box only Excellent Very good Good Fair Poor

677

Please tick one Perfectly comfortable Very comfortable Quite comfortable Barely comfortable Uncomfortable

3. This morning did you wake with any of the following symptoms? How long did the symptom/s last? Please rate how bad the symptom was on a scale from 1 to 10 with 1 being slight pain and 10 severe pain Please

rate

1–10

1 h or less

Half a day

All day

Neck pain Stiff neck Headache Aching between the shoulder blades

4. Cause of symptoms If you woke, this morning, with any of the symptoms listed above can you think of anything, apart from your pillow that may have caused the symptoms? Please tick

YES

NO

If yes please describe what caused your symptoms.

678

S.J. Gordon et al. / Manual Therapy 14 (2009) 671–678

References Australian Physiotherapy Association, 2008. Neck pain poster. Marketing Solutions. DeKoninck J, Lorrain D, Gagnon P. Sleep positions and position shifts in ﬁve age groups: an ontogenetic picture. Sleep 1992;15:143–9. Elkind AH. Muscle contraction headache. Postgraduate Medicine 1987;83:203–18. Emberson MW. Support pillow. Physiotherapy 1985;71:498. Erfanian P, Hagino C, Guerriero R. A preliminary study assessing adverse effects of a semi-customized cervical pillow on asymptomatic adults. Journal of the Canadian Chiropractic Association 1998;42:156–62. Erfanian P, Tenzif S, Guerriero RC. Assessing effects of a semi-customized experimental cervical pillow on symptomatic adults with chronic neck pain with and without headache. Journal of the Canadian Chiropractic Association 2004;48:20–9. 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;48:9–15. Gordon SJ, Grimmer K, Trott P. Self-reported versus recorded sleep position: an observational study. Internet Journal of Allied Health Sciences and Practice(1), http://ijahsp.nova.edu, 2004;2. Gordon S, Grimmer K, Trott P. Sleep position, age, gender, sleep quality and waking cervico-thoracic symptoms. Internet Journal of Allied Health Sciences and Practice(1), http://ijahsp.nova.edu, 2007;5.

Hagino C, Boscariol J, Dover L, Letendre R, Wicks M. Before/after study to determine the effectiveness of the align-right cylindrical cervical pillow in reducing chronic neck pain severity. Journal of Manipulative and Physiological Therapeutics 1998;21:89–93. Helewa A, Goldsmith C, Smythe H, Lee P, Obright K, Stitt L. Effect of therapeutic exercise and sleeping neck support on patients with chronic neck pain: a randomized control trial. Journal of Rheumatology 2007;34(1):151–8. Jackson R. The cervical spine. 3rd ed. Springﬁeld, IL: Charles C. Thomas; 1976. Kramer J. Intervertebral disk diseases: causes, diagnosis, treatment and phophylaxis. 2nd ed. New York: Thieme Medical Publishers Inc; 1990. Lavin RA, Pappagallo M, Kuhlemeier KV. Cervical pain: a comparison of three pillows. Archives of Physical Medicine and Rehabilitation 1997;78:193–8. Maitland DG. Vertebral manipulation. London: Butterworths; 1986. McKenzie R, May SW. The cervical and thoracic spine. Mechanical diagnosis and therapy, vol. 2. New Zealand: Spinal Publications; 2006. Persson L, Moritz U. Neck support pillows: a comparative study. Journal of Manipulative and Physiological Therapeutics 1998;21:237–40. Persson L. Neck pain and pillows – a blinded study of the effect of pillows on nonspeciﬁc neck pain, headache and sleep. Advances in Physiotherapy 2006;8:122–7. Sackett DL. Bias in analytic research. Journal of Chronic Diseases 1979;32:51–63. Shields N, Capper J, Polak T, Taylor N. Are cervical pillows effective in reducing neck pain? New Zealand Journal of Physiotherapy 2006;34:3–9. Yin RK. Case study research: design and methods. In: Applied social research methods. 3rd ed., vol. 5. New York: Sage Press; 2003.

Manual Therapy 14 (2009) 679–684

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

Effects of posture on the thickness of transversus abdominis in pain-free subjects Angelica Reeve a, *, Andrew Dilley b, c a

19th Street Physiotherapy Clinic, Vancouver, V7M 1X5 Canada Division of Clinical and Laboratory Investigation, Brighton and Sussex Medical School, University of Sussex, Brighton BN1 9PS, UK c Department of Physiology, University College London, Gower Street, London, UK b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 30 April 2008 Received in revised form 1 February 2009 Accepted 24 February 2009

The role of transversus abdominis (TrA) on spinal stability may be important in low back pain (LBP). To date, there have not been any investigations into the inﬂuence of lumbo-pelvic neutral posture on TrA activity. The present study therefore examines whether posture inﬂuences TrA thickness. A normative within-subjects single-group study was carried out. Twenty healthy adults were recruited and taught ﬁve postures: (1) supine lying; (2) erect sitting (lumbo-pelvic neutral); (3) slouched sitting; (4) erect standing (lumbo-pelvic neutral); (5) sway-back standing. In each position, TrA thickness was measured (as an indirect measure of muscle activity) using ultrasound. In erect standing, TrA (mean TrA thickness: 4.63 1.35 mm) was signiﬁcantly thicker than in sway-back standing (mean TrA thickness: 3.32 0.95 mm) (p ¼ 00001). Similarly, in erect sitting TrA (mean thickness ¼ 4.30 mm 1.58 mm) was found to be signiﬁcantly thicker than in slouched sitting (mean thickness ¼ 3.46 mm 1.13 mm) (p ¼ 0002). In conclusion, lumbo-pelvic neutral postures may have a positive inﬂuence on spinal stability compared to equivalent poor postures (slouched sitting and sway-back standing) through the recruitment of TrA. Therefore, posture may be important for rehabilitation in patients with LBP. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Posture Transversus abdominis Spinal stability

1. Introduction There has been much recent interest in the stability of the lumbar spine and its relation to low back pain (LBP; van Dieen et al., 2003; Hodges, 2003; Silﬁes et al., 2005). Transversus abdominis (TrA) has been of particular interest to many physiotherapists as a core stability muscle due to its anatomy (O’Sullivan, 2000; Hodges, 2003; Golby et al., 2006). A delayed muscular response of TrA has been found in patients with a history of LBP (Hodges and Richardson, 1996, 1998; Ferreira et al., 2004), but it is still not clear if this delayed response is a predisposing factor to LBP or is a consequence of LBP (McGill et al., 2003; Hodges and Moseley, 2003). In a review on posture by Raine and Twomey (1994), it was concluded that there are still controversies and little evidence supporting claims on the beneﬁts of ideal posture or the suggestion that poor posture will lead to musculoskeletal pain. However, in a more recent systematic review, Prins et al. (2008) concluded that musculoskeletal pain may be inﬂuenced by sitting posture in children and adolescents. Moreover, a correlation has been

* Corresponding author. Tel.: þ1 604 988 5221; fax: þ1 604 984 4339, þ1 604 785 9234 (mobile). E-mail address: [email protected] (A. Reeve). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.02.008

observed between spinal posture and LBP (O’Sullivan et al., 2006). There is also an increased risk of LBP in people who have sedentary jobs, with symptoms increasing when sitting for long periods of time (Pope et al., 2002). In a study by Yip et al. (2008), cervical posture was correlated to cervical disability and pain. The inﬂuence of lumbar stability on poor posture versus upright posture has also been studied. It has been reported that there is a signiﬁcant decrease in activity of the internal oblique (IO) and multiﬁdus muscles in poor sitting and standing postures (Snijders et al., 1998; O’Sullivan et al., 2002, 2007). In these studies however, the activity of TrA was not measured. The effect of sitting postures on TrA has been studied by Ainscough-Potts et al. (2006), but a lumbo-pelvic neutral spine appeared not to be controlled. Deep needle electromyography (EMG) is an invasive technique which has resulted in alternative methods of measuring muscle activity. There is now growing evidence of real-time ultrasound as a valid tool to measure muscle thickness, with changes in thickness shown to be correlated to changes in muscle activity at lower levels of maximal voluntary contraction (McMeeken et al., 2004; Hodges et al., 2003a). The purpose of this study is to use ultrasound to examine changes in thickness of TrA in slouched sitting and sway standing, which are commonly adopted poor standing and sitting postures (Arnold et al., 2000; O’Sullivan et al., 2006), and compare these to erect lumbo-pevic neutral standing and sitting positions.

680

A. Reeve, A. Dilley / Manual Therapy 14 (2009) 679–684

2. Method 2.1. Subjects Following ethical approval, by University College London’s Committee on the ethics of non-NHS human research, consent was obtained from 20 healthy subjects (10 male and 10 female) with a mean age of 29 years (range 20–51 years). Subjects were excluded if there was a history of abdominal surgery, a history of LBP within the last two years, or a spinal deformity such as scoliosis. Subjects were also excluded if they were pregnant (Snijders et al., 1995). 2.2. Procedure In each subject, the TrA muscle was imaged in a series of ﬁve postures: (1) supine lying (legs hip width apart and arms folded in front); (2) erect sitting (on a stool with hips angled at 90 , legs hip width apart and hands folded); (3) slouched sitting (as erect sitting except with a slumped lumbar spine); (4) erect standing (legs hip width apart and hands folded); (5) sway-backed standing (as erect standing but with the hips also in relative extension and the pelvis in posterior tilt). These postures are summarized in Fig. 1. Subjects were given standardised instructions on how to position themselves. For erect sitting and standing, a plumb line was used with the line passing through the external auditory meatus, the acromion and greater trochanter and just anterior to the lateral malleolus, as identiﬁed by skin markers (Kendall et al., 1993, Fig. 1). Supine was the starting position in each subject. This was to allow the examiner to familiarise themselves with the subject’s individual abdominal muscle topography. The order of the other four positions was randomly selected. Prior to testing, each subject was taught (with the aid of demonstration, visualisation and manual facilitation when necessary) how to achieve these ﬁve different positions correctly. In erect postures, subjects learnt to uncurl themselves segmentally from a ﬂexed spinal position to ensure they achieved lumbo-pelvic neutral and did not over extend their spines when standing/sitting straight.

Ultrasound (US) imaging was performed using a Diasus ultrasound system (Livingston, Scotland, UK) with a 26 mm linear array, 10–22 MHz transducer head. The right side was scanned in all subjects as no asymmetry in the thickness of the abdominals has been found between sides (Beith et al., 2001). The ultrasound head was positioned between the iliac crest and the lowest rib along the anterior axillary line (Strohl et al., 1981). This position is where visualisation of the tranversus abdominis is found to be at its thickest and where the aponeurotic attachment of transversus abdominis is visible (Misuri et al., 1997). A local knowledge of the anatomy of the iliac crest and lowest rib was used to return to the original transducer position. Skin markings were not used, since movements of the skin during changes in posture can render these inaccurate (Ainscough-Potts et al., 2006). The transducer head was positioned perpendicular to the abdominal wall for accurate readings and optimal clarity of the image. The recordings of muscle thickness were timed at the end of inspiration, as the thickness of TrA varies with the respiratory cycle (Misuri et al., 1997; Ainscough-Potts et al., 2006). Three recordings were taken in each position to improve the reliability of the readings. A change in the length of TrA during relaxation or contraction of the muscle would result in a shift in the position of the aponeurotic attachment. Therefore, as an estimate of the change in length of TrA between erect and slouched sitting (n ¼ 5) and erect and sway-back standing (n ¼ 6), the ultrasound transducer position relative to the umbilicus was determined. This was expressed as a percentage of the circumference of the trunk at the level of the umbilicus. Three repeated measurements of the aponeurotic attachment of TrA were taken from the right side of the transducer to the umbilicus in each position. All images were randomised prior to measurement of the TrA. Muscle thickness was measured using Paint Shop Pro (Jasc, USA). Care was taken to measure perpendicular to the muscle ﬁbres, 15 mm from the TrA aponeurosis (Fig. 2). 2.3. Statistical analysis Two way mixed intraclass correlation coefﬁcients were calculated for each position as a measure of repeatability. Repeat

Fig. 1. Example of each position. (a) Erect sitting. (b) Slouched sitting. (c) Erect Standing. (d) Sway Standing.

A. Reeve, A. Dilley / Manual Therapy 14 (2009) 679–684

681

Muscle thickness (mm)

7.8 6.8 5.8 4.8 3.8 2.8 1.8 Erect Standing

Sway-Back Standing

Fig. 4. Line graph showing the change in TrA thickness in erect standing and swayback standing. Nineteen out of 20 subjects showed an increase in erect sitting compared to slouched sitting.

(Table 1), indicating lengthening of the muscle in sway-back standing. 3.2. Transversus abdominis thickness in erect sitting compared to slouched sitting Fig. 2. Real-time ultrasound scan showing the different layers of the lateral abdominals. Horizontal line, aponeurotic attachment 15 mm to point of measurement. Vertical line, measurement of TrA thickness.

measurements of TrA thickness were averaged for each subject. Paired Student t-tests were used to compare muscle thickness between erect standing with sway standing and erect sitting with slouched sitting. Bonferroni corrections were performed thus the level of signiﬁcance was set at p < 0.017. 3. Results

3.3. Reliability

The mean TrA thickness for each of the ﬁve positions is shown in Fig. 3. 3.1. TrA thickness in erect standing compared to sway standing There was a 39.5% increase in TrA thickness in erect standing (mean TrA thickness ¼ 4.63 1.35 mm SEM) compared to swayback standing (mean ¼ 3.32 0.95 mm), an increase that was signiﬁcant (p ¼ 0.00001, paired t test; Fig. 3). This trend was observed in 19 out of the 20 subjects (Fig. 4). In six subjects tested where the change in position of the aponeurotic attachment was determined, it was found that the ultrasound head was positioned closer to the umbilicus in sway-back standing than erect standing

6

TrA thickness (mm)

There was a 24.3% increase in TrA thickness in erect sitting (mean thickness ¼ 4.30 mm 1.58 mm SEM) compared to slouched sitting (mean ¼ 3.46 mm 1.13 mm SEM), an increase that was also signiﬁcant (p ¼ 0.0002, paired t test). Again this trend was observed in 19 out of 20 subjects. In the ﬁve subjects where the change in position of the aponeurotic attachment was determined, the ultrasound head was positioned closer to the umbilicus in slouched sitting than in erect sitting (Table 1), indicating a lengthening of the muscle in slouched sitting.

5 4 3 2 1 0 Lying

Erect Sitting

Slouched Sitting

Erect Standing

SwayBack Standing

Fig. 3. Thickness of TrA in different positions. Error bars ¼ SEM.

Intraclass correlation coefﬁcients calculated for supine lying, erect sitting, slouched sitting, erect standing and sway-backed standing were 0.98 (95% CI ¼ 0.96–0.99), 0.98 (95% CI ¼ 0.95–0.99), 0.96 (95% CI ¼ 0.92–0.98), 0.98 (95% CI ¼ 0.95–0.99) and 0.97 (95% CI ¼ 0.93–0.99) respectively, indicating excellent reliability between the three scans taken for each posture. Bland–Altman plots (comparing the average between trials plotted against their difference) are shown in Fig. 5. 4. Discussion The present study was an investigation into the changes in TrA thickness in commonly adopted poor postures (sway-back standing and slouched sitting) compared to equivalent neutral spine postures. The results show a signiﬁcant thickening of TrA in both lumbo-pelvic neutral erect standing and sitting postures compared to sway-back standing and slouched sitting. TrA thickness has been shown to be correlated with muscular activity (Hodges et al., 2003a; McMeeken et al., 2004). Therefore, the observed increase in thickness of TrA in erect standing compared to the sway-backed position suggests that there is more TrA activity in erect standing. This increase in activity may help to stabilize the spine (Richardson and Jull, 1995; O’Sullivan et al., 1997). In accordance with the present results, O’Sullivan et al. (2002) found a signiﬁcant decrease in the EMG activity of IO in sway standing compared to erect standing. In a different study using the same methodology, EMG activity from IO and TrA could not be distinguished (Marshal and Murphy, 2003), and therefore it was

682

A. Reeve, A. Dilley / Manual Therapy 14 (2009) 679–684

Table 1 Mean location of the transducer from the umbilicus in erect standing and sway-back standing (n ¼ 6), and erect sitting and slouched sitting (n ¼ 5) (average of three trials), and the difference in location between postures. Mean (cm)

Percentage of circumference

11.1 0.4 9.6 0.4

13.7 0.4 11.7 0.4

Difference SEM

1.5 0.4

2.0 0.5

Erect sitting Slouched sitting

11.5 1.1 10.9 1.3

13.8 0.8 12.3 0.9

0.7 0.3

1.5 0.5

Erect standing Sway standing

Difference SEM

The mean location has also been normalised and expressed as the percentage distance around the circumference (the waist) from the umbilicus. SEM ¼ standard error of the mean.

possible that TrA activity was also recorded. O’Sullivan et al. (2002) also found rectus abdominis to be the only abdominal muscle with a signiﬁcant increase in activity in sway-back standing. Due to the independent anatomy of RA in comparison to the other abdominal muscles (Askar, 1977; Rizk, 1980), RA is likely to have a different role in the sway-back position. The signiﬁcant increase in thickness of TrA in upright sitting compared to the slouched sitting position was consistent with the results for erect and sway-back standing, which indicated a reduction in TrA activity in the slouched sitting position (Hodges et al., 2003a; McMeeken et al., 2004). Together, these results support the idea that TrA is more active in an aligned posture. Ainscough-Potts et al. (2006) used ultrasound to measure TrA thickness in relaxed sitting on a chair compared to sitting on a gym ball. No signiﬁcant difference in thickness of TrA was found; however it does not appear that lumbo-pelvic neutral posture was taught or controlled in this study. O’Sullivan et al. (2002, 2007) also reported a signiﬁcant increase in recruitment of spinal stabilizing muscles in lumbo-pelvic upright sitting compared to slouched or thoracic upright sitting, however TrA activity was not measured. Sapsford et al. (2006) observed an increase in pelvic ﬂoor activity in

upright sitting compared to slouched sitting. Since the pelvic ﬂoor muscles and TrA are likely to work in synergy, an increase in pelvic ﬂoor activity might also indicate an increase in TrA activity (Sapsford et al., 2001). 4.1. Limitations of the study Passive tissue changes during ﬂexion or extension of the trunk may contribute to the observed changes in muscle thickness. There are however, a number of arguments suggesting that this is not likely to be the case. Firstly, the anterior abdominal wall is formed by the common aponeurosis of TrA, IO and external oblique (EO) which weaves around the rectal sheaths (Askar, 1977; Rizk, 1980). This interwoven arrangement allows movement in both longitudinal and transverse directions, enabling the anterior abdominal wall to adapt to movements of the spine, thus maintaining a relatively straight (linear) proﬁle in both ﬂexed and extended trunk positions (Askar, 1977). Therefore, during trunk ﬂexion, there may be passive folding of the subcutaneous tissue but the muscles layers will remain relatively straight. Secondly, as TrA ﬁbres mostly run in a transverse orientation (Urquhart et al., 2005), TrA will be particularly non-receptive to passive tissue changes during ﬂexion or extension of the trunk. Lastly, in the present study, TrA showed a reduction in thickness in the slouched sitting position, which is the opposite of what would be expected if the change in thickness had been due to a passive tissue change. This was conﬁrmed by movement of the ultrasound head closer to the umbilicus in the slouched sitting position compared to erect sitting, indicating lengthening, and therefore relaxation of the muscle. Contrary to these arguments, a factor that may conceivably cause a change in the proﬁle of TrA during trunk ﬂexion or extension is displacement of the abdominal contents, especially since the TrA is the deepest of the three abdominal wall muscles. It has been suggested by Teyhen et al. (2007) that a 7.5–10 MHz ultrasound probe is more suitable for imaging the lateral abdominal muscles, since it has the advantage of better visualisation of the three muscles. In the present study, a higher frequency 10–22 MHz probe was used. The only limitation with using a higher frequency

Fig. 5. Bland–Altman graphs for all trials in (a) supine lying, (b) erect sitting, (c) slouched sitting, (d) erect standing and (e) sway-backed standing. Solid horizontal lines indicate the limits of agreement (95% CI).

A. Reeve, A. Dilley / Manual Therapy 14 (2009) 679–684

probe is that image quality is reduced at lower depths. However, in this study the depth of TrA in all of the subjects was within the limits of the probe. Researchers have proposed a variety of different imaging locations to measure the thickness of the lateral abdominals and as yet, no consensus as to the best location has been reached (Teyhen et al., 2007). In the present study the measurement of TrA thickness was taken 15 mm from the aponeurosis. Fifteen millimetres was considered suitable, since at this location the upper and lower borders of TrA were parallel in all subjects. Expiration has been suggested to be a more stable part of the respiratory cycle for taking images (Teyhen et al., 2007), although no study as yet has conﬁrmed this. The excellent repeatability between the three individual trials in each position suggests that the inspiration phase of the respiratory cycle that was used in the present study was also sufﬁciently stable to produce reliable results. Between day reliability was not tested. 4.2. Clinical implications Since the increase in muscle thickness in neutral spine postures during sitting and standing most likely represents an increase in the muscle activity in these postures, then this study supports the evidence that TrA functions as a postural muscle. The function of TrA may help stabilize the lumbar spine by controlling intersegmental motion either through increasing intra-abdominal pressure or its fascial attachments (Cresswell et al., 1992; Hodges and Richardson, 1997; Hodges et al., 2003b). Patients with LBP have been found to have delayed activity of TrA (Hodges, 2003), with treatment of LBP targeting the re-training of this muscle (O’Sullivan et al., 1997; Hides et al., 2001; Richardson et al., 2002; Stuge et al., 2004). The current research suggests that lumbo-pelvic neutral posture may inﬂuence the recruitment of TrA and therefore supports the idea that posture re-education could be an important part of the rehabilitation process. Further research using deep needle EMG of TrA may be justiﬁed to support the evidence found in this study. 5. Conclusion In both erect lumbo-pelvic neutral standing and sitting postures there was an increase in TrA thickness compared to sway-back standing or slouched sitting, respectively. Should it be accepted that the observed changes in thickness represent an increase in muscular activity, these results support the evidence of TrA as a postural muscle, possibly assisting in the provision of stability to the spine. The present research also suggests that ‘good’ posture may inﬂuence the recruitment of TrA. Therefore, posture re-education could be important in the rehabilitation of TrA. Acknowledgments The assistance of Professor Bruce Lynn is most gratefully acknowledged and also, many thanks to Dr Iain Beith for his advice. References Ainscough-Potts AM, Morrissey MC, Critchley D. The response of the transverse abdominis and internal oblique muscles to different postures. Manual Therapy 2006;11(1):54–60. Arnold CM, Beatty B, Harrison EL, Olszynski W. The reliability of ﬁve clinical postural alignment measures for women with osteoporosis. Physiotherapy Canada 2000;52(4):286–94.

683

Askar OM. Surgical anatomy of the aponeurotic expansions of the anterior abdominal wall. Annals of the Royal College of Surgeons of England 1977;59(4):313–21. Beith ID, Critchley DJ, Copeman E, Newman DJ. Changes in thickness of the left and right human abdominal muscles in standing and lying. Journal of Physiology Proceedings 2001;531:147P. Cresswell AG, Grundstrom H, Thostensson A. Observations on intra-abdominal pressure and patterns of abdominal intra muscular activity in man. Acta Physiologica Scandinavica 1992;144:409–18. van Dieen JH, Selen LPJ, Cholewicki J. Trunk muscle activation in low back pain patients, an analysis of the literature. Journal of Electromyography and Kinesiology 2003;13:333–51. Ferreira PH, Ferreira ML, Hodges PW. Changes in recruitment of the abdominal muscles in people with low back pain: ultrasound measurement of muscle activity. Spine 2004;29:2560–6. Golby LJ, Moore AP, Doust J, Trew MP. A randomized controlled trial investigating the efﬁciency of musculoskeletal physiotherapy on chronic low back pain disorder. Spine 2006;31(10):1083–93. Hides JA, Jull GA, Richardson CA. Long-term effects of speciﬁc stabilizing exercises for ﬁrst-episode low back pain. Spine 2001;26(11):E242–8. Hodges PW, Richardson CA. Inefﬁcient muscular stabilization of the lumbar spine associated with back pain: a motor control evaluation of transverus abdominis. Spine 1996;21(22):2640–50. Hodges PW, Richardson CA. Relationship between limb movement speed and associated contraction of the trunk muscles. Ergonomics 1997;40(11): 1220–30. Hodges PW, Richardson CA. Delayed postural contraction of transversus abdominis in low back pain associated with movement of the lower limb. Journal of Spinal Disorders 1998;11(1):46–56. Hodges P. Core stability exercise in chronic back pain. Orthopedic Clinics of North America 2003;34:245–54. Hodges P, Holm AK, Holm S, Ekstrom L, Cresswell A, Hansson T, et al. Intervertebral stiffness of the spine is increased by evoked contraction of the transversus abdominis and the diaphragm: in vivo porcine studies. Spine 2003b;28(23):2594–601. Hodges PW, Moseley GL. Pain and motor control of the lumbopelvic region: effect and possible mechanisms. Journal of Electromyography and Kinesiology 2003;13:361–70. Hodges PW, Pengel LHM, Herbert RD, Gandevia SC. Measurement of muscle contraction with ultrasound imaging. Muscle and Nerve 2003a;27:682–92. Kendall FP, Kendall McCreary E, Provance PG. In: Muscle testing and function: with posture and pain. 2nd ed. Baltimore: Williams and Wilkins; 1993. pp. 75–85. Marshal P, Murphy B. The validity and reliability of surface EMG to assess the neuromuscular response of the abdominal muscles to rapid limb movement. Journal of Electromyography and Kinesiology 2003;13:477–89. McGill SM, Grenier S, Kavcic N, Cholewicki J. Coordination of muscle activity to assure stability of the lumbar spine. Journal of Electromyography and Kinesiology 2003;13:353–9. McMeeken J, Beith I, Newham D, Milligan P, Critchley D. The relationship between EMG and change in thickness of transversus abdominis. Clinical Biomechanics 2004;19(4):337–42. Misuri G, Colagrande S, Gorini M, Iandelli I, Mancini M, Duranti R, et al. In vivo ultrasound assessment of respiratory function of abdominal muscles in normal subjects. European Respiratory Journal 1997;10:2861–7. O’Sullivan PB, Twomey LT, Garry T. Evaluation of speciﬁc stabilization exercise in the treatment of chronic LBP with radiologic diagnosis of spondylolysis or spondylolisthesis. Spine 1997;22(24):2959–67. O’Sullivan PB. Lumbar segmental ‘stability’: clinical presentation and speciﬁc stabilizing exercise management. Manual Therapy 2000;5(1):2–12. O’Sullivan PB, Kirsty M, Kendell M, Lapenskie SC, Moller N, Richards K. The effect of different standing and sitting postures on trunk muscle activity in a pain-free population. Spine 2002;27(11):1238–44. O’Sullivan PB, Mitchell T, Bulich P, Waller R, Holte J. The relationship between posture and back muscle endurance in industrial workers with ﬂexion-related low back pain. Manual Therapy 2006;11(4):264–71. O’Sullivan PB, Dankaerts W, Burnett A, Farrell G, Jefford E, Naylor CS, et al. The effect different ‘upright’ sitting postures have on trunk muscle activation in pain-free subjects. Physiotherapy 2007;93(S1):S96. Pope MH, Goh KL, Magnusson ML. Spine ergonomics. Annual Review of Biomedical Engineering 2002;4:49–68. Prins Y, Crous L, Louw OA. A systematic review of posture and psychosocial factors as contributors to upper quadrant musculoskeletal pain in children and adolescents. Physiotherapy Theory and Practice 2008;24(4):221–42. Raine S, Twomey L. Attributes and qualities of human posture and their relationship to dysfunction or musculoskeletal pain. Critical Reviews in Physical and Rehabilitation Medicine 1994;6(4):409–37. Richardson CA, Jull GA. Muscle control – pain control. What exercises would you prescribe? Manual Therapy 1995;1:2–10. Richardson CA, Snijders CJ, Hides JA, Damen L, Pas MS, Storm J. The relation between the transversus abdominis muscles, sacroiliac joint mechanics, and low back pain. Spine 2002;27(4):399–405. Rizk NN. A new description of the anterior abdominal wall in man and mammals. Journal of Anatomy 1980;131(3):373–85.

684

A. Reeve, A. Dilley / Manual Therapy 14 (2009) 679–684

Sapsford RR, Hodges PW, Richardson CA, Cooper DH, Markwell SJ, Jull GA. Coactivation of the abdominal and pelvic ﬂoor muscles during voluntary exercises. Neurourology Urodynamics 2001;20(1):31–42. Sapsford RR, Richardson CA, Stanton WR. Sitting posture affects pelvic ﬂoor muscle activity in Parous women: an observational study. Australian Journal of Physiotherapy 2006;52:219–22. Silﬁes SP, Squillante D, Maurer P, Westcott S, Karduna AR. Trunk muscle recruitment patterns in speciﬁc chronic low back pain populations. Clinical Biomechanics 2005;20:465–73. Snijders CJ, Slagter AHE, van Strik R, Vlemming A, Stoeckart R, StamWhy leg crossing? HJ. The inﬂuence of common postures on abdominal muscle activity. Spine 1995;20(18):1989–93. Snijders CJ, Vleeming A, Stoekart R, Mens JMA, Kleinrensink GJ. Biomechanical modelling of sacroiliac joint stability in different postures. Spine: State of the Art Reviews 1998;9:419–32.

Strohl KP, Mead J, Banzett RB, Loring SH, Kosch PC. Regional differences in abdominal muscle activity during various manoeuvres in humans. Journal of Applied Sciences 1981;70:3044–54. Stuge B, Veierod MB, Laerum E, Vollestad N. The efﬁcacy of a treatment programme focusing on speciﬁc stabilizing exercise for pelvic girdle pain after pregnancy: a two-year follow-up of a randomised clinical trial. Spine 2004;29(10): E197–203. Teyhen DS, Gill NW, Whittaker JL, Henry SM, Hides JA, Hodges P. Rehabilitative ultrasound imaging of the abdominal muscles. Journal of Orthopaedic and Sports Physical Therapy 2007;37(8):450–66. Urquhart DM, Barker PJ, Hodges PW, Story IH, Biggs CA. Regional morphology of the transversus abdominis and obliquus internus and externus abdominis muscles. Clinical Biomechanics 2005;20:233–41. Yip HTY, Chui TTW, Poon ATK. The relationship between head posture and severity and disability of patients with neck pain. Manual Therapy 2008;13:148–54.

Manual Therapy 14 (2009) 685–689

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

Stature recovery after sitting on land and in water Ba´rbara Maria Camilotti a, *, Andre´ L.F. Rodacki b,1, Vera Lu´cia Israel c, 2, Neil E. Fowler d, 3 ´ gua Verde, Curitiba, Parana ´, Setor de Cieˆncias Biologias, Mestrado em Tecnologia em Sau ˆ ndido Xavier, 817, ap 703, A ´, Brazil ´lica do Parana ´ de, Rua Ca Pontifı´cia Universidade Cato ´, Setor de Cieˆncias Biolo ´gicas, Departamento de Educaça ˜o Fı´sica, Centro de Estudos do Comportamento Motor. R. Coraça ˜o de Maria, 92, Universidade Federal do Parana ˆnico, Curitiba, Parana ´, Brazil BR116, Km 95, Jardim Bota c ´, Campus Litoral, Setor de Cieˆncias Biologias, Departamento de Fisioterapia. Rua Jaguariaı´va, 512, Caioba ´-Matinhos, Parana ´, Brazil Universidade Federal do Parana d The Manchester Metropolitan University, Dept of Exercise and Sport Sciences. Hassal Road, Alsager, Stoke-on-Trend ST72 HL, United Kingdom a

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 10 September 2008 Received in revised form 14 March 2009 Accepted 28 March 2009

Back pain treatment in water has been commonly used although there is little evidence about its effects. One purported advantage for exercise is the reduced loading due to the buoyant force. The purpose of this study was to compare stature change, as a marker of spinal loading, after sitting in aquatic and dry land environments. Fourteen asymptomatic volunteers had their stature measured in a precision stadiometer, before and after a bout of physical activity and during a recovery period either sitting in water (head out of water immersion; HOWI) and sitting in a chair on land (SITT). Stature loss following exercise was as expected similar in both groups (SITT ¼ 89.2 5.4% and HOWI ¼ 86.5 8.1%; p ¼ 0.33). When stature recovery was compared between the water and land environments, HOWI (102.2 8.7%) showed greater recovery than SITT (86.5 6.3%) after 30 min (p < 0.05). These results suggest that HOWI facilitated more rapid stature recovery through lower spinal loading and supports use of this technique to reduce spinal loading during recovery. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Spine Intervertebral discs Spinal shrinkage Aquatic therapy

1. Introduction Low back pain is both very common and of high impact in society. The World Health Organization (WHO) report that more than 80% of the global population are affected by back pain at some point in life, the impact of which is seen in personal suffering, lost productivity and health care cost. (Svensson and Andersson, 1989; Graf et al., 1995; Wilke et al., 1999; Norcross et al., 2003). Treatment costs are high and represent a burden for the patients and/or State (Svensson and Andersson, 1989; Deyo et al., 1991; Wilke et al., 1999). Compressive loading of the spine is a key factor in determining the risk of developing back pain and guidelines recommend seeking methods to minimize such loads (van Deursen et al., 2005). Long periods of compressive loading are associated with a progressive loss of ﬂuid from the intervertebral discs and associated increases in loads of other spinal structures e.g. facet joints (Althoff et al., 1992; Graf et al., 1995; Leivsth and Drerup, 1997; Lengsfeld et al., 2000). Measurement of the rate and extent of disc

* Corresponding author. Tel.: þ55 41 3311 11525; fax: þ55 41 3244 5401. E-mail address: [email protected] (B.M. Camilotti). 1 Tel.: þ55 41 3360 4333; fax: þ55 41 3360 4336. 2 Tel.: þ55 41 3452 8300; fax: þ55 41 3452 2662. 3 Tel.: þ44 (0)161 247 5466; fax: þ44 (0) 161 247 6375. 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.03.007

height has been advocated as an effective marker of spinal loading (Althoff et al., 1992; Rodacki et al., 2003). A number of studies have indicated that low back pain is closely related to the inability to recover stature following a period of loading (Rodacki et al., 2003; Healey et al., 2005b). Several studies have investigated different recovery strategies aimed at promoting stature recovery, for instance, Rodacki et al. (2003) and Healey et al. (2005b, 2008) analyzed the effects of the side lying, whereas others have analyzed the partial gravitational inversion (Healey et al., 2005a) and the sitting position (Althoff et al., 1992; Beynon and Reilly, 2001; van Deursen et al., 2005). In a recent work, Healey et al. (2005a) showed that postures in which spinal loading is reduced either by altering gravitational loading or muscle activity are more effective for the restoration of intervertebral space. A frequent treatment applied to reduce the loading on the spine and to treat back pain and spinal pathologies is aquatic therapy. Such treatment is predicated on the assumption that the buoyant force will reduce the spinal loading and thus offer relief from back pain (Konlian, 1999; Israel and Pardo, 2000; Masumoto et al., 2004). Several studies (Guillemin et al., 1994; Ariyoshi et al., 1999) have reported pain reduction after performing aquatic therapy. Others (Konlian, 1999) have indicated that aquatic therapy reduces the risk of further insult to the damaged tissues. However, studies quantifying the behavior of the spinal column in response to aquatic therapy are scarce in literature. Thus, the present study aimed to

686

B.M. Camilotti et al. / Manual Therapy 14 (2009) 685–689

Table 1 Average physical characteristics (SD) for the participants. Gender

Age (years)

Weight (kg)

Stature (cm)

BMI (Kg/m2)

Female Male

22.5 2.9 21.2 2.4

59.7 5.4 67.8 6.9

168.1 0.04 174.2 0.05

21.1 1.3 22.3 2.6

compare stature recovery in aquatic (head out of water immersion; HOWI) and dry land (SITT) environments. 2. Methods 2.1. Participants Fourteen healthy, asymptomatic individuals (7 men and 7 women), with no history of back pain in the 6 months prior the study, volunteered to participate. Participants were able to swim (familiarized with the water) and had a body mass index smaller than 25 kg/m2. Table 1 shows the physical characteristics of all participants. The study was approved by the Ethics Committee of the Pontiﬁcal Catholic University of the State of Parana´ and the participants signed a free and informed consent form. 2.2. Experimental procedures Each participant attended the laboratory on two different occasions, each in the morning and within the ﬁrst three hours after waking. Changes in stature were assessed in two conditions: sitting immersed in water (HOWI) and sitting on dry land (SITT). The order of the conditions was randomly assigned in an attempt to reduce any order effects. A minimum period of 24 was imposed between experimental sessions. The assessment protocol included: familiarization of the participants with the measuring equipment (precision stadiometer) (Fig. 1); resting in the Fowler’s position; loading done with the use of a backpack and stature-recovery period in the sitting position, on the land and in the water. The purpose of the training and familiarization with the stadiometer procedures was to improve the repeatability of measurement and the participants were considered as trained when the

standard deviation was less than 0.5 mm for 10 consecutive measurements of the stature (Eklund and Corlett, 1984; Rodacki et al., 2001, 2003; Dezan et al., 2003; Fowler et al., 2005). Thus, measurements were deemed to provide reliable data after this short period of training (5–8 min). After the familiarization period, participants rested for 30 min in a supine position, with their hip and knees ﬂexed and ankles supported on a comfortable surface (Fowler’s position). Such a position has been proven effective to recover the vertebral column’s length and to eliminate the interference prior loading to the spine (Dezan et al., 2003; Rodacki et al., 2003, 2005). After the resting period, participants stood up for 90s before the ﬁrst measurement of the individuals’ stature was carried out and named PRE. Participants were then asked to carry a dorsal backpack, whose weight corresponded to 10% of the body weight, for 20 min, in order to induce a compressive spinal load and induce the loss of stature. After the 20-min period, a second measurement was taken and named LOS. Following loading the participants undertook one of the two stature-recovery strategies, which involved sitting for 30 min in a chair without a backrest, with their feet on the ground either on dry land (SITT) or immersed in water (HOWI) up-to the level of the seventh cervical vertebra (C7) (Fig. 2). Small postural adjustments were allowed. The water temperature was kept at 33.1 C 0.6. During the recovery period (HOWI, SITT) dry land measurements of stature variations were taken with the stadiometer every 10 min and called HOWI10, HOWI20, HOWI30 and SITT10, SITT20 and SITT30. Before each measurement, the participants stood up for 90 s in order to reduce the effect of soft tissue creep deformation of the lower limbs (Foreman and Linge, 1989). 2.3. Stature-changing measurements Changes in stature were quantiﬁed using a precision stadiometer as proposed by Eklund and Corlett (1984) and modiﬁed by Rodacki et al. (2001). The stadiometer consisted of a rigid metal frame, tilted backwards, 15 in relation to the vertical plane. Five horizontal beams (postural controls) were adjusted at speciﬁc anatomic points (Fig. 1): the largest head protuberance (occipital); the deepest point of the cervical lordosis curve (approximately at

Fig. 1. Precision stadiometer (adapted from Rodacki et al., 2001).

B.M. Camilotti et al. / Manual Therapy 14 (2009) 685–689

Stature Lossð%Þ ¼

687

LOS 100 PRE

(1)

The percentage of stature recovery was calculated by the equations (2) and (3), for HOWI and SITT respectively, in each measurement (10, 20 and 30 min of recovery):

Stature Recoveryð%Þ ¼

LOS 100 HOWI

(2)

Stature Recoveryð%Þ ¼

LOS 100 SITT

(3)

Stature changes (loss and recovery) were compared by a Two Way ANOVA for repeated measures. In order to establish the differences in terms of height changes, the Scheffe´ test was applied. Values of p < 0.05 have indicated statistical signiﬁcance, and 95% conﬁdence intervals calculated. The Statistica Software, version 7.0, was used for the statistical analyses.

3. Results Fig. 2. Sitting position inside the water.

C4); the most prominent point of thoracic kyphosis (approximately at T7); the deepest point of lumbar lordosis curve (approximately at L3) and the apex of the buttocks (approximately at the middle ridge of the sacrum). At the lumbar and cervical curves, a screw (perpendicular to the postural control) was designed to move forwards to touch the deepest point of the concavity (Rodacki et al., 2001). For the measurements, the participants were instructed to stand up, knees straight and to place their feet comfortably with weight evenly distributed and to lean back against the postural controls. The arms were left hanging, on the side of their thighs. Undesirable head movements were controlled by a pair of glasses with two removable laser-emitting devices (class 2, wave-length 630– 680 mm and maximum output < 1 mW), battery-operated and assembled on the frame’s side. An elastic band helped to keep the participant’s spectacles in a comfortable and appropriate position, with relatively constant pressure. Horizontal and vertical alignment of the head was achieved by keeping the laser beams (left and right) at the center of two marks (diameter: 20 mm) that were ﬁtted on two small magnets placed on the metallic surface of the projection screen, that is, a metallic frame place above the participant’s head. A mirror (200 200 mm) was placed w300 mm in front of the participants in such a way that they were able to see and thus control the position of the laser beams. The magnets were positioned with the participant’s head and neck in a normal and comfortable position. A high-resolution (0.005 mm) Linear Variable Displacement Transducer (LVDT) was used to determine stature changes. The LVDT was placed on a rigid but adjustable frame, at the top of the stadiometer, in the middle of the horizontal beam and was positioned to coincide with the line of the longitudinal axis of the spine. Such settings permitted the distal end of the LVDT to rest directly on the highest apex of the head (vertex).

2.4. Statistical approach

The repeatability of the measurements reported in this study for both experimental groups was consistent with that reported by several previous authors (Rodacki et al., 2003, 2005; Healey et al., 2005a,b). The mean error of assessments was deemed as acceptable for the purpose of this study (0.42 0.13 mm) with the mean SD below the values (0.5 mm) reported in other studies. Loss in stature after the physical activity (carrying of a backpack) was 89.24 5.45% for SITT and 86.58 8.17 for HOWI. The stature losses were similar regardless the assessment days (p ¼ 0.33). Stature was seen to recover in the sitting position in both HOWI and SITT at all measurement points (Table 2). Comparing HOWI and SITT, stature recovery was signiﬁcantly greater (p ¼ 0.01) at all three recovery moments for the water immersed condition. However, no signiﬁcant differences were seen (p ¼ 0.38) between the recovery moments on SITT (SITT10, SITT20, SITT30). In HOWI, signiﬁcant stature difference was only seen at moment HOWI30 comparing to HOWI20 (p ¼ 0.03), when a full stature recovery was observed (102.23 8.76%). Fig. 3 shows the stature recovery at HOWI and SITT.

4. Discussion Findings in this study show that the sitting position used during recovery did allow participants to recover the majority of their stature in both land or water environments. However, recovery following sitting whilst immersed in water was greater than that observed in normal (on land) sitting. An analysis of the stature loss, induced by the loading task, conﬁrms that there was a similar loading imposed in each condition. It is therefore possible to compare the recover rates as the discs would be returning from a similar state of compression. It is also worth to note that the

Table 2 Percentage of stature recovery (SD) in HOWI and SITT at moments 10, 20 and 30 min.

SITT (%) HOWI (%)

The analysis of stature loss/recovery was carried out similarly to the approach described by Rodacki et al. (2003). The percentage of stature loss was determined by the equation (1):

10 min

20 min

30 min

88.06 5.71 (84.76 91.36) 96.23 8.78 (91.16 101.31)

86.59 5.10 (83.64 89.54) 95.03 7.92 (90.45 99.60)

86.58 6.39 (82.89 90.28) 102.23 8.76 (97.17 107.29)

Note: Values between brackets refer to the lower and upper conﬁdence intervals, respectively.

688

B.M. Camilotti et al. / Manual Therapy 14 (2009) 685–689

Fig. 3. Average percentage for stature recovery in HOWI and SITT at moments 10, 20 and 30 min. The symbol * indicates differences (p < 0.05) between experimental conditions (SITT and HOWI), while # refers to differences (p < 0.05) between recovery moments.

degree of stature loss is comparable with that induced in other studies (Rodacki et al., 2003; Healey et al., 2005a,b, 2008). Whilst the induced shrinkage in stature was comparable with previous studies, the magnitude of the recovery in SITT (86.58 6.39%) was smaller than that reported for studies where participants recovered using recumbent postures (Eklund and Corlett, 1984; Beynon and Reilly, 2001; Dezan et al., 2003; van Deursen et al., 2005; Healey et al., 2005a). Thus, irrespective of the whether immersion is used, sitting does not seem to offer the same degree of ‘unloading’ and thus may not be the most suitable recovery posture. Where immersed seating is used longer recovery time is necessary to achieve a full recovery of stature loss which may have implications for the viability of this technique for clinical application. The seated posture, in both water and on land is likely to have induced greater muscle activation to sustain this posture compared to recumbent postures used in other studies. Healey et al. (2008) have shown a close negative relationship between muscle activation and both the rate and extent of stature recovery post-exercise. Muscle activation changes may also have inﬂuenced differences between water (HOWI) and land (SITT) conditions. The greater stature recovery observed in the immersed condition may be attributed to some physical properties of the water e.g. buoyancy and hydrostatic pressure. These properties will have reduced the apparent weight of the body and thus diminished the muscular demand to sustain the erect posture of the upper segments (Skinner and Thomson, 1985; Cole et al., 1996; Degani, 1998; Konlian, 1999; Po¨yho¨nen et al., 1999; Masumoto et al., 2004). Further to this, it is likely that maintenance of the up-right sitting posture will have been facilitated by the viscous a hydro-static forces acting upon the body, this would also help reduce the muscle activation and so spinal loading. As well as the buoyant and hydro-static forces there may also have been beneﬁts associated with the thermal effects of immersion in warm water. It is well known that muscles are able to relax in a greater extend when thermal effects are present (Konlian, 1999; Hinman et al., 2007). Thus, reducing the muscle activation may have reduced the compressive loading and so permitted greater ﬂuid inﬂux to the nucleus pulposus and greater stature gains. Other studies including electromyography assessment are necessary to further explore these arguments. Rodacki et al. (2003) and Healey et al. (2005b, 2008) proposed that full stature recovery is important to reverse the effects of spinal loading and intervertebral disc loading which are believed to be closely related to back pain. Thus, methods involving strategies that

provide the most effective ability to recover stature are important. Thus, the relatively long time required to provide signiﬁcant changes in stature is not an attractive aspect of HOWI in comparison to other strategies used to recover stature (e.g. series of abdominal exercises). When sitting postures were compared (HOWI and SITT) a clear advantage of using water was identiﬁed. The results of the present study must be viewed with some caution when extrapolating to clinical populations. It is possible that responses in healthy participants may not replicate the same behavior as participants with ongoing back pain, although studies by Healey et al. (2005a,b) offer some support to the assertion that, at least for patients with mild back pain, similar patterns of stature loss and recovery can be expected. To identify the likely mechanisms for the beneﬁcial effects of immersion, it remains necessary to assess electromyography to determine the degree that water effects are due to buoyant forces or reduced muscle activation. Other studies are still required to further examine these preliminary ﬁndings. 5. Conclusion In this study it was possible to observe that stature recovery occurs in all sitting postures, however, water immersion (HOWI) showed better results when compared with normal (dry land) sitting (SITT). Full stature recovery occurred only after 30 min in water, which was earlier than in land. A combination of the buoyancy and hydro-static forces and reduced muscle activation were effective to reduce the magnitude of the internal loads. Although immersed sitting was better than sitting on dry land, the amount of recovery was lower than that observed for recumbent recovery postures. Conﬂict of interest statement Authors have exclusive academic interest in this manuscript and there are no conﬂicts of interest in the present submission. References Althoff I, Brinckmann W, Frobin W, Sandover J, Burton K. An improved method of stature measurement for quantitative determination of spinal loading: application to sitting postures and whole body vibration. Spine 1992;17(6):683–93. Ariyoshi M, Sonoda K, Nagata K, Mashima T, Zenmyo M, Paku C, et al. Efﬁcacy oh aquatic exercises for patients with low back pain. Kurume Medical Journal 1999;46(2):91–6. Beynon C, Reilly T. Spinal shrinkage during a seated break and standing break during simulated nursing tasks. Applied Ergonomics 2001;32:617–22. Cole AJ, Eagleston RE, Moschetti M, Sinnett E. Aquatic rehabilitation of the spine: water-based programs con help patients with spine injuries. Rehab Management 1996;9(3):55–60. Degani A. Hidroterapia: os efeitos fı´sicos, ﬁsiolo´gicos e terapeˆuticos da a´gua. Fisioterapia em Movimento 1998;11(1):91–106. van Deursen DL, van Deursen LL, Sniders CJ. Relationship between everyday activities and spinal shrinkage. Clinical Biomechanics 2005;20(5):547–50. Deyo RA, Cherkin D, Conrad D, Volinn E. Cost, controversy, crisis: low back pain ant the health public. Annual Review of Public Health 1991;12:141–50. Dezan V, Rodacki AL, Rodacki CL, Santos AM, Okazaki VHA, Sarrat TA. Comparaça˜o dos efeitos compressivos do disco intervertebral nas condiço˜es de levantamento de peso nas posiço˜es sentada e em pe´. Brazilian Journal of Biomechanics 2003;4(7):41–9. Eklund JA, Corlett NE. Shrinkage as a measure of the effect of load on the spine. Spine 1984;9(2):189–94. Foreman T, Linge K. The importance of hell compression in the measurement of diurnal stature variation. Applied Ergonomics 1989;4:299–300. Fowler NE, Rodacki AL, Rodacki CL. Spinal shrinkage and recovery in women with and without low back pain. Archives of Physical Medicine and Rehabilitation 2005;86(3):505–11. Graf M, Guggenbu¨hl U, Krueger H. An assessment of seated activity and postures at ﬁve work places. International Journal of Industrial Ergonomics 1995;15:81–90. Guillemin F, Constant E, Collin JF, Boulange M. Short and long-term effect of spa therapy in chronic low back pain. British Journal of Rheumatology 1994;33:148–51. Healey EL, Fowler NE, Burden AM, McEwan IM. The inﬂuence of different unloading positions upon stature recovery and paraspinal muscle activity. Clinical Biomechanics 2005a;20:365–71.

B.M. Camilotti et al. / Manual Therapy 14 (2009) 685–689 Healey EL, Fowler NE, Burden AM, McEwan IM. Raised paraspinal muscle activity reduces rate of stature recovery after loaded exercise in individuals with chronic low back pain. Archives of Physical Medicine and Rehabilitation 2005b;86:710–5. Healey EL, Burden AM, McEwan IM, Fowler NE. The impact of increasing paraspinal muscle activity on stature recovery in asymptomatic people. Archives of Physical Medicine and Rehabilitation 2008;89:749–53. Hinman RS, Heywood SE, Day AR. Aquatic physical therapy for hip and knee osteoarthritis: results of a single-blind randomized controlled trial. Physical Therapy 2007;87(1):32–43. Israel VL, Pardo MBL. Hidroterapia: tratamento do lesado medular em piscina terapeˆutica. Fisioterapia em Movimento 2000;13(1):111–27. Konlian C. Aquatic therapy: making a wave in the treatment of low back injuries. Orthopaedic Nursing 1999;18(1):11–20. Leivsth G, Drerup B. Spinal shrinkage during work in a sitting posture compared to work in a standing posture. Clinical Biomechanics 1997;12(7/8):409–18. Lengsfeld M, Frank A, van Deu¨rsen DL, Griss P. Lumbar spine curvature during ofﬁce chair sitting. Medical Engineering & Physics 2000;22:665–9. Masumoto K, Takasugi S, Hota N, Fujishima K, Iwamoto Y. Eletromyographic analysis of walking in water in healthy humans. Journal of Physiological Anthropology 2004;23(4):119–27.

689

Norcross JP, Lester GE, Weinhold P, Dahners LE. An in vivo model of degenerative disc disease. Journal of Orthopaedic Research 2003;21(1):183–8. Po¨yho¨nen T, Keskinen KL, Hautala A, Savolainen J, Ma¨lkia¨ E. Human isometric force production and electromyogram activity of knee extensor muscles in water and on dry land. European Journal of Applied Physiology 1999;80:52–6. Rodacki CL, Fowler NE, Rodacki AL, Birch K. Repeatability of measurement in determining stature in sitting and standing postures. Ergonomics 2001;84(12):1076–85. Rodacki CL, Fowler NE, Rodacki AL, Birch K. Stature loss and recovery in pregnant women with and without low back pain. Archives of Physical Medicine and Rehabilitation 2003;84(4):507–12. Rodacki AL, Fowler NE, Provensi CLG, Rodacki CL, Dezan V. Body mass as a factor in stature change. Clinical Biomechanics 2005;20(8):779–805. Skinner AT, Thomson AM. Dufﬁeld: exercı´cios na a´gua. 3rd ed. Sa˜o Paulo: Manole; 1985. pp. 4–23 [chapter 1]. Svensson HO, Andersson GBJ. The relationship of low-back pain, work history, work environment, and stress: a retrospective cross-sectional study of 38 to 64 year old women. Spine 1989;14(5):517–21. Wilke H, Neff P, Caimi M, Hoogland T, Claes L. New in vivo measurements of pressures in the intervertebral disc in daily life. Spine 1999;24(8):755–62. World Health Organization, [accessed 25.08.08].

Manual Therapy 14 (2009) 690–695

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

Comparison of changes in abdominal muscle thickness between standing and crook lying during active abdominal hollowing using ultrasound imaging Rosie Mew* MSc School of Human Health and Performance, Department of Physiology, University College London, Gower Street, London, UK

a r t i c l e i n f o

a b s t r a c t

Article history: Received 25 July 2008 Received in revised form 9 May 2009 Accepted 19 May 2009

To determine if transversus abdominis (TrA) demonstrates a greater increase in thickness on lower abdominal hollowing (LAH) in standing compared to crook lying. Muscle thickness measurements of TrA, addition of internal obliques (IO) and external obliques (EO) were measured using ultrasound imaging at rest and during LAH on 28 healthy controls (14 female, 14 male) in crook lying and standing. TrA demonstrated greater thickness changes on LAH in standing (þ0.88 mm 0.12 mm). IO and EO demonstrated greater thickness changes on LAH in crook lying (þ0.59 mm 0.08 mm and 0.87 mm 0.12 mm, respectively). These differences were all signiﬁcant (p < 0.001). Increased resting thickness was noted in standing in TrA (20.7%), IO (10.3%) and EO (1.2%). This increase was only significantly different between TrA and EO (P ¼ 0.004). TrA showed signiﬁcantly greater increases in thickness on LAH in standing compared to crook lying, and with greater speciﬁcity in relation to IO and maybe EO. If muscle thickness can be an indicator of muscle function or activity, then this suggests that TrA rehabilitation should be facilitated in positions of greater function, such as standing. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Transversus abdominis Muscle thickness Ultrasound Posture

1. Introduction Low back pain (LBP) is the commonest orthopaedic complaint, with up to 50 million working days lost and estimated costs to the UK economy of £5 billion per year (Waddell et al., 2002). Whilst the exact underlying mechanisms remains unclear, there is increasing evidence to support Panjabi’s (1992) proposal to suggest LBP is a result of suboptimal lumbar segmental control. Richardson et al. (1999) suggest that this change in spinal control is due in part to dysfunction in local segmental muscles such as transversus abdominis (TrA), and has frequently been found in patients with LBP (Richardson and Jull, 1995; Hodges and Richardson, 1996, 1998; O’Sullivan et al., 1997b). The use of lower abdominal hollowing (LAH) has been shown to preferentially activate TrA relative to the more superﬁcial lateral abdominal muscles (Teyhen et al., 2005; Springer et al., 2006). LAH or spinal stabilising exercises has been shown to be effective in the treatment of LBP (O’Sullivan et al., 1997b; Richardson and Hodges, 2004; Ferreira et al., 2007), signiﬁcantly reduce LBP and disability (O’Sullivan et al., 1997a) and

* Tel.: þ44 7967 078 382. E-mail address: [email protected] 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.05.003

a particularly effective method of reducing the long term reoccurrence of LBP (Rackwitz et al., 2006). Much of the previous research involving LAH has been performed in positions such as supine, prone and four point kneeling (Beith et al., 2001a; Critchley and Coutts, 2002) and has often been directly imitated in clinical practice. This, and the recent popularity of Pilates, has lead to an increasing amount of core stability work being performed in lying and other postures of greater support, even though there is evidence to suggest that positions of greater function may provide more effective activation of the deep abdominal muscles (Bunce et al., 2004; Urquhart et al., 2005). Studies investigating external oblique (EO) have reported delays in activities in sitting (Moore et al., 1992) compared to a similar task in standing (Hodges and Richardson, 1997). Meanwhile, Urquhart et al. (2005) observed delayed abdominal muscle activity with arm movement in sitting compared to standing. However to date there has been little research directly comparing TrA function between different body positions, or investigating which posture is preferential for TrA activation. Accurate evaluation of TrA activity is challenging due to its depth within the abdominal wall. Fine wire electromyography (EMG) recordings are impractical in the clinical setting and TrA cannot be isolated with surface EMG (Critchley and Coutts, 2002).

R. Mew / Manual Therapy 14 (2009) 690–695

Muscle ﬁbre contraction results in increased cross sectional area (CSA) proportional to changes in force tension (Takemori, 1990). The low forces involved with LAH results in minimal tendon stiffness with large changes in muscle geometry (Hodges et al., 2003). As change in width is constrained by the ribs and iliac crest, increases in CSA will be expressed as increases in muscle thickness. Measuring change in muscle thickness using real time ultrasound (US) imaging has been shown to demonstrate consistent correlations with EMG activity in TrA (McMeeken et al., 2004). Hodges et al. (2003) found strong correlations between US muscle thickness changes in addition of TrA and internal obliques (IO) and EMG activity less than 20% maximum voluntary contractions (MVC), but showed no such correlation for EO. As a result many clinicians have started to use US as a tool to examine changes in muscle function in response to speciﬁc tasks such as LAH, however to date much of the available research has only been done in crook lying and other nonfunctional positions. The aim of this present investigation was to compare changes in TrA thickness in a functional position, such as standing, to a less functional but more supportive posture, crook lying. The aim was to determine which posture showed greater changes in TrA thickness between rest and during LAH, relative to IO and EO.

691

and the apex of the ilium. This has been shown to be the thickest point of TrA (Strohl et al., 1981) and demonstrates the clearest image of TrA, IO and EO simultaneously (Critchley and Coutts, 2002). Small multidirectional movements (i.e. 26 mm) from this location were permitted to optimise the scan image. 3 repeat US images were taken at rest and during LAH in both crook lying and standing, and the muscle thicknesses of TrA, IO and EO were then measured by a blinded reviewer. Imaging was collected at end of inspiration when TrA is at its thinnest (Misuri et al., 1997). LAH was performed following guidelines by Richardson and Jull (1995), by asking subjects to ‘gently draw in their lower abdomen and lift up their pelvic ﬂoor’, as this has been found to improve TrA activation (Critchley, 2002). 2.3. Data analysis The changes in muscle thickness of TrA, IO and EO, between rest and during LAH, were compared between crook lying and standing, and analysed using Paired student ‘t-test’. Changes in resting muscle thickness between the 2 postures were also analysed using the paired student ‘t-test’. Where changes in the 3 muscles were addition of compared and signiﬁcance levels were adjusted to P 0.017 (Bonferroni correction). Demographic variables within the population group were analysed using the unpaired student ‘t-test’. Sub-group analysis also compared gender, trained and untrained, and previous history of LBP and none.

2. Method 2.1. Subjects 28 healthy voluntary subjects (14 female, 14 male, aged 21–42 years) were recruited from staff and students of University College, London. An information sheet was provided and all subjects signed informed consent. Exclusion criteria were a history of pelvic/ abdominal surgery, current LBP or pregnancy. All subjects completed a questionnaire recording their gender, age, height, weight, level of physical activity, and a history of any previous LBP. 14 subjects had some previous experience of TrA training (trained subjects) whilst 14 had none (untrained), and 6 subjects had a previous history of LBP. This study was approved by the university research ethics committee.

2.4. Reliability testing Intra-rater reliability was examined using interclass correlation coefﬁcients (ICC) as described by Rankin and Stokes (1998), using model 3,k (Koppenhaver et al., 2009). Good intra-rater reliability was demonstrated for the 3 blinded repeat muscle thickness measurements for each muscle, whilst relaxed and during LAH, and in both postures in 6 sets of data, with ICC results between 0.98 and 0.99, with a standard error of measurement of 0.26 mm. Varying the order of the starting posture or muscle state demonstrated no signiﬁcant difference (range of P-values 0.98–0.99).

2.2. Measurements 3. Results The two postures compared were crook lying (supine lying with 60 hip ﬂexion) and standing (feet hip width apart). The starting posture and starting muscle state (i.e. resting or during LAH) were randomised each time. US imaging was performed using a high frequency (10–22 MHz) 26 mm linear array transducer head (Diasus, Livingston, Scotland). Transducer head location was marked on the right-hand side of each subject mid way between the lowest rib

3.1. Lower abdominal hollowing TrA demonstrated greater thickness changes from rest to LAH in standing compared to crook lying (þ0.88 mm 0.12 mm) (Table 1 and Fig.1). IO and EO demonstrated greater thickness changes on LAH in crook lying (þ0.59 mm 0.08 mm and 0.87 mm 0.12 mm,

Table 1 Mean muscle thicknesses (mm) at rest and during LAH in crook lying and standing (N ¼ 28), with change in muscle thickness, difference between standing and crook lying, standard error mean (SEM) and P value. Mean muscle thickness (mm) Difference

SEM

P value

At rest During LAH Change in thickness

3.71 5.34 1.63

Crook lying

4.48 7.00 2.51

þ0.88

0.12

0.0000002

IO

At rest During LAH Change in thickness

8.02 9.33 1.31

8.85 9.57 0.72

0.59

0.08

0.00101

EO

At rest During LAH Change in thickness

5.76 6.13 0.37

5.83 5.32 0.51

0.87

0.12

0.00003

TrA

Standing

692

R. Mew / Manual Therapy 14 (2009) 690–695

10

3.0

TrA

Lying Standing

IO

9

EO

2.5 8

7

Muscle Thickness (mm)

Change in muscle thickness (mm)

2.0

1.5

1.0

6

5

4

3

0.5

2 0.0 TrA

IO

EO

-0.5

1

0 Crook Lying

-1.0 Fig. 1. Change in muscle thickness (mm) on LAH in transversus abdominis (TrA), internal obliques (IO) and external obliques (EO) between crook lying and standing (N ¼ 28). Error bars – SEM.

Standing

Fig. 2. Muscle thickness (mm) of transversus abdominis (TrA), internal obliques (IO) and external obliques (EO) at rest in crook lying and standing (N ¼ 28). Error bars – SEM.

thickness than the female subjects (0.08 mm) but no signiﬁcant difference was seen on LAH (male 0.93 mm and female 0.85 mm, p ¼ 0.74).

respectively), with a thinning of EO noted on LAH in standing. The means of TrA, IO and EO were all signiﬁcantly different between crook lying and standing (p < 0.001).

4. Discussion

3.2. Resting thickness

4.1. Lower abdominal hollowing

All 3 muscles demonstrated greater resting thickness in standing over crook lying (Table 1 and Fig. 2). TrA was thicker by 20.7% (þ0.77 mm 0.10), IO by 10.3% (þ0.83 mm 0.11), and EO by 1.2% (þ0.07 mm 0.01) (Fig. 2, Table 1). Signiﬁcant difference was found between TrA and EO (P ¼ 0.004) but not between TrA and IO, or IO and EO.

Standing produced statistically signiﬁcant greater increases in TrA thickness on LAH with reduced thicknesses in both IO and EO compared to crook lying. If changes in muscle architecture or thickness can be an indicator of muscle function or activity, then this may suggest that LAH in standing produces greater TrA activity compared to crook lying. Possible explanations for this could be the greater gravitational pull upon the abdomen in standing, resulting in greater feedback from TrA muscle stretch receptors raising the excitability of its motor-neurone pool and increasing TrA recruitment on LAH (Beith et al., 2001a). This gravitational pull will be reduced in crook lying so making recruitment of TrA in isolation more demanding, and possibly increasing the need for oblique recruitment to assist in LAH. EO being more of a torque producing muscle will be less responsive to the low forces involved with LAH, and the passive stretch in standing could instead result in a lengthening and thinning of EO. Due to the constraints of the ribs and iliac crest, increased TrA and IO thickness on LAH could also create a relative crushing force upon the more passive EO, resulting in this apparent thinning. The reduction in EO thickness during LAH in standing could be speciﬁcally beneﬁcial in rehabilitating those LBP patients who have excessive EO substitution strategies during LAH (Richardson and

3.3. Demographic variables No signiﬁcant differences were noted with age, height, weight, or level of physical activity; or between the different sub-groups of gender, trained and untrained, and previous history of LBP and none. Removing those subjects with a history of LBP from the study completely made no signiﬁcant difference to the results obtained (p ¼ 0.71). Those with previous LBP had slightly reduced resting TrA thickness (0.22 mm) but there was no signiﬁcant difference during LAH (0.80 mm and 0.99 mm, p ¼ 0.53). On LAH, the trained subjects showed a greater mean increases in TrA thickness than the untrained subjects but not signiﬁcantly (1.10 mm and 0.66 mm, respectively, p ¼ 0.09), and no signiﬁcance was seen with IO (0.56 mm and 0.62 mm, p ¼ 0.85) and EO (0.81 mm and 0.94 mm, p ¼ 0.72). Male subjects had a slightly larger resting TrA

R. Mew / Manual Therapy 14 (2009) 690–695

Jull, 1995; Henry and Westervelt, 2005). However, John and Beith (2007) found the relationship between muscle thickness and activity for EO was only signiﬁcant in 50% of their subjects. Hodges et al. (2003) found that whilst TrA and IO showed consistent thickness changes with incremental changes in activity less than 20% MVC, EO showed no such correlation, addition of implying that it does not get thicker on contraction at low level MVC. They proposed that whilst transversely orientated abdominal muscles may experience greater shortening, EO’s more oblique orientation may result in increased breadth rather than thickness. However the study by Hodges et al. (2003) was only performed on 3 subjects with possible correlations noted in 1 of the 3. Although more research is therefore required in this area, this may suggest that muscle thickness changes on US may have less validity for indicating EO activity. LAH resulted in an almost doubling of TrA thickness (56% in standing and 43% in crook lying) with relatively smaller changes in IO thickness (8% and 16%, respectively) and EO (9% and 6%, respectively). This is consistent with other recent abdominal US studies (Teyhen et al., 2005; Springer et al., 2006) and further validates the use of LAH as a speciﬁc exercise preferential to TrA. 4.2. Resting thickness Changes in resting muscle thickness between the 2 postures could be assumed to be primarily due to changes in ‘involuntary postural tone’. The increased postural tone of TrA in standing supports its primary postural role increasing in activity when the challenge to stability was increased (Hodges and Richardson, 1998). The lack of increase in EO, and its signiﬁcant difference to TrA is similar to ﬁnding by Beith et al. (2001b). This reinforces theories that EO is a force or torque producing muscles less affected by postural changes and has less role in segmental spinal stability. IO meanwhile showed no signiﬁcant difference from either TrA or EO suggesting that it plays some role in both. 4.3. Demographic variables No signiﬁcant differences were noted with any of the variables within the population group, or between the different sub-groups of gender, trained and untrained, and previous history of LBP and none. Although the males demonstrated slightly larger resting muscle thicknesses than the females, similar to ﬁndings by Rankin et al. (2006) and Springer et al. (2006), difference due to gender was not signiﬁcant at rest or during LAH in this present investigation. The slightly reduced resting TrA muscle thickness in subjects with a history of LBP may have been predicted due to de-conditioning or pain inhibition as has been found in Lumbar Multiﬁdus (Hides et al., 1994; Hides et al., 2008). Cairns et al. (2000) also found that subjects that have previously had LBP can present with deep abdominal muscular dysfunctions even when asymptomatic. However, in this present investigation no signiﬁcant difference was found between those with or without previous history of LBP at rest or during LAH. Those with previous exposure to LAH (trained subjects) demonstrated greater increases in TrA on LAH in standing, but again this was not enough to be statistically signiﬁcant. 4.4. Methodological considerations Single side only studies do not allow for any left to right variance as has been found in Lumbar Multiﬁdus (Hides et al., 2008) and the abdominal muscles (Mannion et al., 2008). Although only imaging a small proportion of the muscle, Beith et al. (2001b) showed minimal differences at 3 points on TrA, suggesting activity is similar

693

throughout the muscle length. However, when activated excited muscle ﬁbres slide over each other creating shorter thicker muscles, by only imaging the middle we cannot comment on these more distal portions. Whilst the lower ﬁbres of IO run more parallel to TrA, suggesting that they function in a similar fashion, the upper ﬁbres run more obliquely, implying that they may have a very different role to both the lower ﬁbres and to TrA. As this present investigation only scanned the upper ﬁbres of IO any conclusion made in relation to IO can only refer to these upper ﬁbres. The high frequency transducer (10–22 MHz) is higher than many previous US studies of the abdominal muscles (Hodges et al., 2003; Teyhen et al., 2005; Ainscough-Potts et al., 2006) and than recommended by an international consensus paper (Teyhen et al., 2007). Although higher frequency gives better resolution at shallower depths, errors could arise in visualising muscle borders deeper than 30 mm. Since no subjects in this present investigation had muscle thickness measurements greater than 30 mm good resolution was maintained, and measuring errors were minimised by demonstrating good intra-rater reliability. To gain a true representation of a change in muscle activity many previous studies have normalised the data by expressing it as a percentage of its maximum (Cholewicke et al., 1997; Beith et al., 2001a; McMeeken et al., 2004). Therefore, additional to the previous data analysis done in this present investigation, attempts were also made to normalise the data by expressing the increase in thickness on LAH as a percentage of the increase during MVC, whilst subjects performed a maximal abdominal clench. Mean MVC thicknesses of TrA, IO and EO were 7.17 mm (1.95), 11.16 mm (2.93) and 5.53 mm (1.44), respectively. Analysis with the Paired student ‘t-test’ however now resulted in less signiﬁcant results (p < 0001, p ¼ 0.01 and p < 01, respectively). Cholewicke and McGill (1996) suggest that muscle forces as low as 1–3% MVC may be sufﬁcient for segmental stability, therefore comparing to maximum may be inappropriate. Normalising all 3 muscles against the same task is also not entirely appropriate, as whilst MVC in standing is optimal for TrA, MVC in rotation would have been more suitable for IO and EO. Furthermore without twitch interpolation the author cannot state that MVC was produced but rather ‘maximal effort’. Hodges et al. (2003) demonstrated that TrA and IO muscle thickness increased incrementally only with activity less than 20% MVC. At low activity levels small changes in muscle activity produce large changes in muscle architecture, whilst at higher activity levels muscle architecture changes relatively little. This suggests that US can only reliably measure low level activity with less discrimination at higher levels of contraction and could further account for less signiﬁcant results obtained with MVC. 4.5. Implications for future practice When rehabilitating patients with LBP the clinician needs to ensure that the best exercise in the best position is chosen to achieve optimum outcome. Much of the previous research into core abdominal exercises has frequently used positions of such as supine, prone and four point kneeling (Beith et al., 2001a; Critchley and Coutts, 2002). As therapists, we are at risk of imitating this directly into our clinical practice using positions of greater support and less function such as Pilates-style exercises. This present investigation provides much needed evidence directly comparing which position provides greater changes in TrA thickness. If change in muscle thickness correlates to activity then these ﬁndings suggest that our traditional LAH posture of crook lying provides a less effective activation of TrA than standing, and builds on growing evidence that positions of greater function are more preferential for TrA activation. Beith et al. (2001a) found that TrA activity can be isolated more often and more consistently in four

694

R. Mew / Manual Therapy 14 (2009) 690–695

point kneeling than prone. Using US, Ainscough-Potts et al. (2006) found that by lifting one foot off the ﬂoor was the only way to signiﬁcantly increases TrA and IO activity compared to lying, sitting and sitting on a gym ball. Bunce et al. (2004) found that TrA activation on US was increased in standing and walking compared to crook lying. Using EMG Urquhart et al. (2005) observed a delay in abdominal muscle activity during rapid arm movement in supported sitting relative to standing. The focus of core stability rehabilitation has previously often been on the segmental approach to re-learning, practising in a position of greater support and then progression into function. Problems often arise if subjects spend too long in non-functional positions and frequently struggle to incorporate what they have learnt into functional activity. Cortical mapping studies by Van Vilet and Heneghan (2006) demonstrated that neural pathways adapt according to what is practised. They recommend that to improve motor control it is essential to incorporate functionally orientated exercise early into management to re-educate this feed-forward mechanism and improve spinal stability. O’Sullivan (2005) highlights the signiﬁcance of fear avoidance and faulty coping strategies with LBP. He stresses the importance of treating subjects in their pain provocation positions, usually positions of greater function, and believes that postures of greater support such as prone or supine should only be used if activation cannot be facilitated in functional postures such as sitting or standing. Previous studies into the use of US as a clinical tool to examine changes in muscle activity have reviewed non-functional postures such as supine and crook lying. As clinicians we must be aware that these ﬁndings may not be transferable to functional postures such as standing. This investigation provides much needed valuable information into the architectural changes of TrA, IO and EO on LAH in healthy normals, and with direct correlation between our traditional non-function posture of crook lying and standing. As more physiotherapists start to incorporate functional work into their rehabilitative programs the ﬁndings of this investigation will provide some normative data for further research in this area.

5. Conclusion This present investigation demonstrates that performing LAH in a posture of greater function produces greater changes in TrA thickness when compared to a posture of greater support. If changes in muscle thickness are an indicator of changes in muscle activity these ﬁnding suggest that standing is a more preferential position for TrA training over our traditional LAH posture of crook lying, and in relation to IO and possibly EO. These ﬁndings have signiﬁcant importance for future abdominal core stability rehabilitation for patients with lumbo-pelvic disorders.

Acknowledgements With thanks for the time, patience and advice given by Professor Bruce Lynn and Dr Andrew Dilley.

References Ainscough-Potts A-M, Morrissey M, Critchely D. The response of the transversus abdominis and internal oblique muscles to different postures. Manual Therapy 2006;11:54–60. Beith I, Synnott E, Newman SA. Abdominal muscle activity during the abdominal hollowing manoeuvre in the four point kneeling and prone positions. Manual Therapy 2001a;6(2):82–7. Beith I, Critchley D, Copeman E, Newham D. Changes in thickness of the left and right human abdominal muscles in standing and lying. Journal of Physiology 2001b;531:147P.

Bunce S, Hough A, Moore A. Measurements of abdominal muscle thickness using M-mode ultrasound imaging during functional activities. Manual Therapy 2004;9(1):41–4. Cairns M, Harrison K, Wright C. Pressure biofeedback: a useful tool in the quantiﬁcation of abdominal muscle dysfunction? Physiotherapy 2000;86(3):127–38. Cholewicke J, McGill S. Mechanical stability of the in vivo lumbar spine: implications for injury and chronic low back pain. Clinical Biomechanics 1996;11(1): 1–15. Cholewicke J, Panjabi M, Khachatryan A. Stabilising function of trunk ﬂexorextensor muscles around a neutral spine posture. Spine 1997;22(19):2207–12. Critchley D. Instructing pelvic ﬂoor contraction facilitates transversus abdominis thickness increase during low abdominal hollowing. Physiotherapy Research International 2002;7(2):65–75. Critchley D, Coutts F. Abdominal muscle function in chronic low back pain: measurement with real-time ultrasound scanning. Physiotherapy 2002;88(6):322–32. Ferreira ML, Ferreira PH, Latimer J, Herbert R, Hodges P, Jennings MD, et al. Comparison of general exercise, motor control exercise and spinal manipulative therapy for chronic low back pain: a randomised control trail. Pain 2007;131:31–7. Henry S, Westervelt K. The use of real time ultrasound feedback in teaching abdominal hollowing exercises to healthy subjects. Journal of Orthopaedic and Sports Physical Therapy 2005;35:338–45. Hides J, Stokes M, Saide M, Jull G, Cooper D. Evidence of lumbar multiﬁdus muscle wasting ipsilateral to symptoms in patients with acute/subacute low back pain. Spine 1994;19:165–72. Hides J, Gilmore C, Stanton W, Bohlscheid E. Multiﬁdus size and symmetry among chronic LBP and healthy asymptomatic subjects. Manual Therapy 2008;13(1):43–9. Hodges P, Richardson C. Inefﬁcient muscular stabilization of the lumbar spine associated with low back pain, a motor control evaluation. Spine 1996;21(22):2640–50. Hodges P, Richardson C. Relationship between limb movement speed and associated contractions of the trunk muscles. Ergonomics 1997;40(11):1220–30. Hodges P, Richardson C. Delayed postural contraction of transversus abdominis in low back pain associated with movement of the lower limb. Journal of Spinal Disorders 1998;11(1):46–56. Hodges P, Pengal L, Herbert R, Gandevia S. Measurement of muscle contraction with ultrasound imaging. Muscle and Nerve 2003;27(6):682–92. John E, Beith I. Can activity within the external abdominal oblique be measured using real time ultrasound imaging? Clinical Biomechanics 2007;22(9): 972–9. Koppenhaver S, Hebert J, Fritz J, Parent E, Teyhen D, Magel J. Reliability of rehabilitative ultrasound imaging of the transversus abdominis and lumbar multiﬁdus muscles. Archives of Physical Medicine and Rehabilitation 2009;90(1): 87–94. Mannion A, Pulkovski N, Toma V, Sprott H. Abdominal muscle size and symmetry at rest and during abdominal hollowing exercises in healthy control subjects. Journal of Anatomy 2008;213(2):173–82. McMeeken J, Beith I, Newman D, Miligan P, Critchley D. The relationship between EMG and changes in thickness of transversus abdominis. Clinical Biomechanics 2004;19:337–42. Misuri G, Colagrande S, Gorini M, Landelli I, Mancini M, Duranti R, et al. In vivo ultrasound assessment of respiratory function of abdominal muscles in normal subjects. European Respiratory Journal 1997;10:2861–7. Moore S, Brunt D, Nesbitt M, Juarez T. Investigation of evidence for anticipatory postural adjustments in seated subjects who performed a reaching task. Physical Therapy 1992;72(5):172–80. O’Sullivan P. Diagnosis and classiﬁcation of chronic low back pain disorders: maladaptive movement and motor control impairments as underlying mechanisms. Manual Therapy 2005;10(4):242–55. O’Sullivan P, Twomey L, Allison G. Evaluation of speciﬁc stabilising exercise in the treatment of chronic low back pain with radiological diagnosis of spondylolysis and spondylolisthesis. Spine 1997a;22(29):59–67. O’Sullivan P, Twomey L, Allison G, Sinclair J, Miller K, Knox J. Altered pattern of abdominal muscle activation in patients with chronic low back pain. Australian Journal of Physiotherapy 1997b;43(2):91–8. Panjabi M. The stabilizing system of the spine, part I: function, dysfunction, adaptation and enhancement. Journal of Spinal Disorders 1992;5(4):383–9. Rackwitz B, de Bie R, Limm H, von Garnier K, Ewert T, Stucki G. Segmental stabilising exercises and low back pain. What is the evidence? A systematic review of randomised controlled trails. Clinical Rehabilitation 2006;20:553–67. Rankin G, Stokes M. Reliability of assessment tools in rehabilitation: an illustration of appropriate statistical analysis. Clinical Rehabilitation 1998;12: 187–99. Rankin G, Stokes M, Newham D. Abdominal muscle size and symmetry in normal subjects. Muscle and Nerve 2006;34(3):320–6. Richardson C, Hodges P. Local segmental control. In: Richardson C, Hodges P, Hides J, editors. Therapeutic exercises of lumbopelvic stabilisation. A motor control approach for the treatment and prevention of low back pain. 2nd ed. Edinburgh: Churchill Livingstone; 2004. p. 185–219. Richardson C, Jull G. Muscle control – pain control: what exercises would you prescribe? Manual Therapy 1995;1:2–10. Richardson C, Jull G, Hodges P, Hides J. Therapeutic exercise for spinal segmental stabilization in low back pain. Toronto: Churchill Livingstone; 1999.

R. Mew / Manual Therapy 14 (2009) 690–695 Springer B, Mielcarek B, Nesﬁeld T, Teyhan D. Relationships among lateral abdominal muscles, gender, body mass index and hand dominance. Journal of Orthopaedic and Sports Physical Therapy 2006;36(5):289–97. Strohl K, Mead J, Bankett R, Loring S, Kosch P. Regional differences in abdominal muscle activity during various manoeuvres in humans. Journal of Applied Sciences 1981;70:3044–54. Takemori S. Inﬂuence of osmotic swelling on cross section and resting tension in isolated muscle ﬁbres. Japanese Journal of Physiology 1990;40:595–611. Teyhen D, Gill N, Whittaker J, Henry S, Hides J, Hodges P. Rehabilitative ultrasound imaging of the abdominal muscles. Journal of Orthopaedic and Sports Physical Therapy 2007;37(8):450–66.

695

Teyhen D, Miltenberger C, Deiters H, Del Toro Y, Pulliam J, Childs J, et al. The use of ultrasound imaging of the abdominal drawing-in manoeuvre in subjects with low back pain. Journal of Orthopaedic and Sports Physical Therapy 2005;35(6):346–55. Urquhart D, Hodges P, Story I. Postural activity of the abdominal muscles varies between regions of these muscles and between body positions. Gait and Posture 2005;22:295–301. Van Vilet P, Heneghan N. Motor control and the management of musculoskeletal dysfunction. Manual Therapy 2006;11(3):208–13. Waddell G, Alyward M, Sawney P. Back pain, incapacity for work and social security beneﬁts: an international review and analysis. The Royal Society of Medicine Press; Nov 2002.

Manual Therapy 14 (2009) 696–701

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Original Article

The effect of therapeutic exercise on activation of the deep cervical ﬂexor muscles in people with chronic neck pain G.A. Jull a, *, D. Falla b, B. Vicenzino a, P.W. Hodges a a

Centre of Clinical Research Excellence in Spinal Pain, Injury and Health, Division of Physiotherapy, School of Health and Rehabilitation Sciences, The University of Queensland, Brisbane, QLD 4072, Australia b Centre for Sensory-Motor Interaction (SMI), Department of Health Science and Technology, Aalborg University, Denmark

a r t i c l e i n f o

a b s t r a c t

Article history: Received 8 September 2008 Received in revised form 6 May 2009 Accepted 19 May 2009

Deep cervical ﬂexor muscle (DCF) activation is impaired with neck pain. This study investigated the effects of low load cranio-cervical ﬂexion (C-CF) and neck ﬂexor strengthening exercises on spatial and temporal characteristics of DCF activation during a neck movement task and a task challenging the neck’s postural stability. Forty-six chronic neck pain subjects were randomly assigned to an exercise group and undertook a 6-week training program. Electromyographic (EMG) activity was recorded from the DCF, sternocleidomastoid (SCM) and anterior scalene (AS) muscles pre and post intervention during the cranio-cervical ﬂexion test (CCFT) and during perturbations induced by rapid, unilateral shoulder ﬂexion and extension. C-CF training increased DCF EMG amplitude and decreased SCM and AS EMG amplitude across all stages of the CCFT (all P < 0.05). No change occurred in DCF EMG amplitude following strength training. There was no signiﬁcant between group difference in pre-post intervention change in relative latency of DCF but a greater proportion of the C-CF group shortened the relative latency between the activation of the deltoid and the DCF during rapid arm movement compared to the strength group (P < 0.05). Speciﬁc low load C-CF exercise changes spatial and temporal characteristics of DCF activation which may partially explain its efﬁcacy in rehabilitation. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Neck pain Electromyography Motor training Feedforward control

1. Introduction There is increasing evidence of impaired cervical ﬂexor muscle function in neck pain disorders. Although earlier studies focused on, and demonstrated, a reduction in ﬂexor strength and endurance (Watson and Trott, 1993; Barton and Hayes, 1996), recent research has provided evidence of more speciﬁc deﬁcits. Studies of the coordination of the deep and superﬁcial cervical ﬂexor muscles in a low load cranio-cervical ﬂexion (C-CF) task have revealed increased electromyographic (EMG) amplitude of the large superﬁcial sternocleidomastoid (SCM) (Jull et al., 2004) and anterior scalene (AS) muscles in patients with neck pain (Falla et al., 2004b). This was associated with reduced activation of the deep cervical ﬂexors (DCFs), longus capitis and longus colli, and reduced range of C-CF motion to perform the task (Falla et al., 2004b). Furthermore a delay in activation of both the deep and superﬁcial cervical ﬂexor muscles has been demonstrated during rapid arm movements,

* Corresponding author. Tel.: þ61 7 3365 2275; fax: þ61 7 3365 1622. E-mail address: [email protected] (G.A. Jull). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.05.004

indicating a change in the automatic feedforward control of the cervical spine (Falla et al., 2004a). Two contrasting exercise programs have been used to address impaired cervical ﬂexor muscle function: general strengthening exercises (e.g. head lift exercise) (Berg et al., 1994; Bronfort et al., 2001); and a low load program designed to focus more speciﬁcally on motor control aspects to train the coordination between the layers of neck ﬂexor muscles and the quality of C-CF movement (Jull et al., 2008). Clinical trials of both exercise regimes have demonstrated outcomes of reduced neck pain and headache (Bronfort et al., 2001; Jull et al., 2002). Although the low load exercise regime improved performance in the cranio-cervical ﬂexion test (CCFT), this was judged clinically by the subject’s ability to successfully complete higher stages of the test (Jull et al., 2002). It is unknown whether the coordination of the deep and superﬁcial cervical ﬂexor muscles was modiﬁed or restored by the exercise. Nor is it known if such speciﬁc task retraining is necessary or whether a general exercise, such as conventional strengthening exercises, would achieve the same effect. Finally, it is unknown if improvements following exercise with either regime translate to improvements in automatic function of the cervical muscles. Changes in activation of deep trunk muscles in an untrained task, following motor training

G.A. Jull et al. / Manual Therapy 14 (2009) 696–701

in patients with low back pain, suggest that such transfer could be expected (Tsao and Hodges, 2007, 2008). This study compared the physiological effects of low load C-CF exercise and neck ﬂexor strengthening to evaluate effects on deep and superﬁcial cervical muscle activity during the CCFT and on their automatic activation during rapid, unilateral arm movements in patients with non-severe chronic neck pain. We hypothesised that speciﬁc training would be more efﬁcient than general strengthening in addressing deep and superﬁcial muscle control in the CCFT and in automatic function when the neck is perturbed during rapid arm movements. 2. Methods 2.1. Subjects Participants were 46 female subjects with chronic neck pain greater than 3 months duration. Sample size was based on the difference in EMG amplitude of the SCM between patients with neck pain and controls in the CCFT (Sterling et al., 2003). Forty-two patients (21 per group) were required to detect a 70% (0.195) difference in EMG amplitude between patients with neck pain and controls with a SD of 0.223 at 80% power, and 95% conﬁdence. The sample was increased to 46 to allow for a 10% dropout rate. Subjects were recruited by advertisements in the local press. Two particular inclusion criteria were; (i) non-severe neck symptoms (Neck Disability Index score <15/50) to avoid exacerbation of pain with the strengthening exercises; (ii) poor performance in the CCFT – unable to control more than the second stage of the test (Jull, 2000) to ensure that subjects had the muscle impairment for which the training was required. Subjects were excluded if they previously had cervical spine surgery, neurological signs in the upper limb or participated in a neck exercise program in the past 12 months. Ethical approval for the study was granted by the Institutional Medical Research Ethics Committee. Written consent was provided before participation. 2.2. Exercise interventions Subjects were randomized into two exercise groups, low load or higher load strength training, by drawing a number in a sealed envelope from a box. Exercise regimes were of 6-weeks duration and were commenced within one week of the initial assessment. All subjects received personal instruction and supervision by one of 10 experienced physiotherapists once per week. No exercise sessions were longer than 30 min. Subjects were asked not to seek other interventions for neck pain although usual medication was not withheld. Subjects received an exercise diary and were requested to practice their respective regime twice per day (10– 20 min) for the duration of the trial, without provoking neck pain and with attention to performance of smooth uniplanar movements. 2.2.1. C-CF training. The low load training of the cranio-cervical ﬂexor muscles followed an established protocol (Jull et al., 2002; Jull et al., 2008). This exercise targets the deep ﬂexor muscles of the upper cervical region (longus capitis and longus colli), rather than the superﬁcial ﬂexor muscles (SCM and AS). The SCM has a large ﬂexor moment for the cervical region but does not contribute to ﬂexor moments at the cranio-cervical region (Vasavada et al., 1998) and the AS muscles have no attachment to the cranium. In the ﬁrst phase of training the physiotherapist taught the subject to perform a slow and controlled C-CF action in the supine position. The subject concentrated on feeling the back of the head slide in cephalad and caudad directions on the supporting surface to ensure a sagittal rotation rather than a retraction movement. Once the

697

correct C-CF motion was achieved, subjects began the second phase of training in which they were trained to hold progressively increasing ranges of C-CF using feedback from an air-ﬁlled pressure sensor (StabilizerÔ, Chattanooga Group Inc. USA) placed behind the neck. The feedback dial displayed the amount of pressure change as the cervical lordosis progressively ﬂattened during C-CF. The subject initially performed C-CF to sequentially reach 5 pressure targets in 2 mmHg increments from a baseline of 20 mmHg to the ﬁnal level of 30 mmHg. The physiotherapist identiﬁed the target level that the subject could hold steadily for 5 s without resorting to retraction, without dominant use of the superﬁcial neck ﬂexor muscles, and without a quick, jerky C-CF movement. Contribution from the superﬁcial muscles was monitored by the physiotherapist using palpation. Training commenced at the target level that the subject could achieve with a correct C-CF movement and without dominant use of the superﬁcial muscles. They then trained to be able to sustain progressively greater ranges of C-CF using feedback from the pressure sensor. For each target level, the contraction duration was increased to 10 s, and the subject trained to perform 10 repetitions with brief rest periods between each contraction (w3–5 s). Once a stage was achieved, the exercise was progressed to train at the next target level up to the ﬁnal target of 30 mmHg. 2.2.2. Strength training. The strength training consisted of a progressive resistance exercise program in supine with the head supported. Subjects slowly lifted the head and neck through as full a range of motion (ROM) as possible without causing discomfort or reproducing symptoms. It was a two-stage program of two weeks and then four weeks duration as recommended by McArdle et al. (1996) for initiating a weight training program in untrained individuals. In stage one, subjects performed 12–15 repetitions with a weight that they could lift 12 times at the ﬁrst session and progressed to 15 repetitions. They maintained this stage for the remainder of the two week period. In stage two, subjects performed 3 sets of 10 repetitions, with the ﬁrst set using a 50% 10 repetition maximum (RM) load, the second set a 75% 10 RM load and the third set a full 10 RM load. All repetitions were performed over a 1 s period with no rest between repetitions and with a 1 min rest interval between sets. If head weight was insufﬁcient to provide a 10 RM load, weights were applied to the subject’s forehead in 0.5 kg increments. If the subject could not perform the head lift or the head lift caused discomfort, the load on the neck ﬂexors was reduced by decreasing the vertical component of the head weight vector (the upper body was inclined up from horizontal). 2.3. Outcome measures Primary outcome measures were EMG amplitude of the DCF, SCM and AS muscles and ROM during the ﬁve stages of the CCFT, and the relative latencies between onset of DCF, SCM and AS EMG and that of deltoid during rapid unilateral arm movements. The latter measure had added importance as unlike the CCFT, it was a measure and task that was unrelated to the training protocols of either group. Secondary outcome measures were patient self reports of pain and disability and perceived beneﬁt of exercise. All measures were taken at baseline and in week 7 immediately after treatment except for the perceived beneﬁt of exercise which was obtained only following the intervention period. The researcher was blinded to subject group for the outcome assessments. 2.3.1. Electromyography Myoelectric signals were recorded from the DCF muscles unilaterally on the side of greatest pain. The apparatus consisted of bipolar silver wire electrode contacts (2 mm 0.6 mm, inter-

698

G.A. Jull et al. / Manual Therapy 14 (2009) 696–701

electrode distance: 10 mm) attached to a suction catheter (size 10 FG), with a heat sealed distal end, which was inserted via the nose to the posterior oropharyngeal wall. The electrode location was conﬁrmed by inspection through the mouth w1 cm lateral to the midline at the level of the uvula (Falla et al., 2003a). The electrode contacts were ﬁxed to the mucosal wall with a suction pressure of 30 mmHg via a portal between the two contacts. Before insertion, the nose and pharynx were anaesthetised with three metered doses of 2% XylocaineÒ spray (lidocaine, Astra Pharmaceuticals, Sweden) administered via the nostril and to the posterior oropharyngeal wall, via the mouth. Surface EMG signals were recorded from the sternal head of SCM and the AS bilaterally and the anterior and posterior deltoid unilaterally (on the side that DCF EMG was acquired) using Ag/AgCl electrodes (Grass Telefactor, Astro-Med Inc.) following skin preparation and guidelines for electrode placement (Hermens et al., 1999; Falla et al., 2002). The ground electrode was placed on the upper thoracic spine. EMG data were ampliﬁed (Gain ¼ 1000), band-pass ﬁltered between 20 Hz–1 kHz and sampled at 2 kHz. Data were sampled with Spike software using a micro1401 data acquisition system (Cambridge Electronic Design, Cambridge, UK) and converted into a format suitable for signal processing with Matlab (MathWorks, Inc. MA, USA). 2.3.2. Measures of pain, disability and perceived beneﬁt At baseline and post intervention, subjects completed the Neck Disability Index (NDI) (Vernon and Silvano, 1991) (score/50). The average intensity of neck pain was measured on a 10 cm Numerical Rating Scale (NRS) anchored with ‘‘no pain’’ and ‘‘the worst possible pain imaginable’’. Subjects rated perceived beneﬁt of the exercise program on a NRS anchored with ‘‘0%’’ and ‘‘100%’’. 2.4. Experimental procedure 2.4.1. CCFT. Subjects were comfortably positioned in supine, crook lying with head position standardised in a mid-position (Falla et al., 2003a). The pressure sensor was placed sub-occipitally behind the subject’s neck and inﬂated to a 20 mmHg baseline pressure. Subjects received visual feedback of pressure. They were instructed by the researcher in the C-CF action and practiced targeting the ﬁve test levels (progressive increments: 2 mmHg) between 22 and 30 mmHg in two practice trials before the electrodes were applied. Before experimental trials, EMG data were collected for 10 s during a standardised manoeuvre for normalisation purposes. The task involved cervical and C-CF to lift and hold the head just clear of the bed (reference voluntary contraction). Subjects then performed the ﬁve incremental stages (22–30 mmHg) of the CCFT to the best of their abilities, maintaining the pressure steady on each target for 10 s. Data collection commenced when the subject reached the pressure target. A 30 s rest was allowed between stages. C-CF ROM was recorded for each test stage using a digital imaging method as previously described (Falla et al., 2003b). 2.4.2. Arm movement task. Subjects performed ﬁve repetitions of rapid unilateral shoulder ﬂexion and extension to approximately 45 in each direction, always starting with the arm resting beside the body (shoulder in neutral rotation and elbow in full extension) whilst standing with feet placed shoulder width apart (Hodges and Richardson, 1997). Visual commands to move were provided by light emitting diodes ﬁxed to an adjustable board positioned at eye level. The voltage drop produced by the onset of the stimulus to move was recorded with the EMG signals. Directions of arm movement were randomized between subjects. The time between stimuli varied and was controlled by the investigator. Subjects were instructed to move ‘‘as fast as possible’’ with the emphasis on speed not distance. Subjects performed 2–3 practice repetitions in both

directions to check signal quality and to ensure consistent speed and distance of arm movement between repetitions. 2.5. Data analysis 2.5.1. CCFT. To obtain a measure of EMG signal amplitude, maximum root mean square (RMS) was calculated using a 1 s sliding window. For normalisation, EMG amplitude for each CCFT stage was expressed as a percentage of the 1 s maximum RMS values obtained during the reference voluntary contraction. A mixed design analysis of variance (ANOVA) was used to evaluate changes in RMS values post intervention with group (C-CF training, strength training) as the between subjects variable and time (pre, post intervention), muscle (DCF, left SCM, right SCM, left AS, right AS) and stage of the test (ﬁve stages of 2 mmHg) as the within subject variables. The ROM obtained at each stage of the CCFT (i.e. change in angle from the start position) was expressed as a percentage of the full range of C-CF. Mixed design ANOVA was used to evaluate changes in ROM with group as the between subjects variable and time and stage of the test as the within subject variables. Any signiﬁcant differences were investigated with post-hoc Student-Newman-Keuls (SNK) pair-wise comparisons. 2.5.2. Arm movement task. EMG data were rectiﬁed and displayed using interactive software and inspected visually to identify the EMG onset for each trace (Hodges and Richardson, 1997; Falla et al., 2004a). The onset was deﬁned as the earliest increase in EMG activity above the baseline level of activity. Recordings were enlarged to a resolution of 0.5 ms and were displayed individually without reference to the muscle or other temporal landmarks to exclude observer bias. Neck muscle EMG onsets were expressed relative to the onset of deltoid EMG, i.e. the relative latency (onset of the deltoid EMG subtracted from the onset of neck muscle EMG, expressed in ms). Any neck muscle EMG onsets that were more than 150 ms before or 500 ms after the onset of deltoid EMG were discarded from analysis as it is unlikely to be related to the perturbation resulting from movement of the arm. A mixed design ANOVA was used to evaluate pre-post intervention change in the relative latencies with group (C-CF training, strength training) as the between subjects variable and muscle (DCF, left SCM, right SCM, left AS, right AS) and direction (ﬂexion, extension) as within subject variables. Signiﬁcant differences were investigated with post-hoc Student-Newman-Keuls (SNK) pair-wise comparisons. Chi-squared analyses were performed to compare the distribution of subjects showing earlier or later relative latencies of the DCF muscles compared to pre-intervention baseline values. Data were categorized based on a change in the timing of the DCF (<40 – >40 ms in 10 ms increments). Paired t-tests were conducted to determine if NDI and NRS were signiﬁcantly different pre to post intervention for both groups. Independent t-tests were conducted to compare for group differences. Statistical analyses were performed using SPSS 10.0 for Windows. A value of P < 0.05 was used to indicate statistical signiﬁcance. 3. Results No subjects were lost to follow up assessment. Table 1 presents subject descriptive data. Baseline characteristics, pain and disability levels, EMG amplitude, ROM for the CCFT and relative latencies during the arm movement task were not different between groups (all: P > 0.05). All participants in the strength group and all but three in the C-CF group received all treatments. Procedural difﬁculties with insertion of the nasopharyngeal electrode resulted in a reduced number of subjects for the DCF EMG data during the CCFT

Strength training (n ¼ 23)

P

39.6 12.2 10.1 10.6 91.3

37.1 10.3 9.2 6.6 91.3

0.45 0.73 0.99

4.5 1.6 11.0 2.7

4.2 2.1 9.6 3.1

0.61 0.10

(C-CF training: n ¼ 20; strength training: n ¼ 20) and arm movement task (C-CF training: n ¼ 18; strength training: n ¼ 20). 3.1. CCFT A signiﬁcant interaction was observed between group and time for values of EMG amplitude (F ¼ 13.8, P < 0.001). Post hoc analysis demonstrated that a signiﬁcant change in EMG amplitude was only identiﬁed for the C-CF group (SNK: P < 0.0001). Accordingly, a signiﬁcant interaction occurred between group, time, muscle and stage of the CCFT for the values of EMG amplitude (F ¼ 1.6; P < 0.05). Post intervention, the DCF EMG amplitude was increased for the C-CF training group across all stages of the CCFT (SNK: all P < 0.05; Fig. 1). In contrast, there was no difference in DCF EMG amplitude in the strength-training group (SNK: all P > 0.05; Fig. 1). In the C-CF group, the EMG amplitude for the left and right SCM and AS decreased across all CCFT stages, except for the lowest level (22 mmHg) (SNK all P < 0.05; Figs. 2 and 3 respectively). There was no signiﬁcant reduction in EMG amplitude of the superﬁcial ﬂexors for the strength-training group except for the left SCM at 28 mmHg test stage (SNK P < 0.05; Fig. 2), for the left and right AS at 30 mmHg (SNK P < 0.05; Fig. 3). No differences in the full range of C-CF were observed after the intervention for either group (change in full range of C-CF group: 1.2 1.0 ; strength group: 0.4 1.0 ). Range of C-CF motion used during the CCFT depended on the interaction between group, time and stage (F ¼ 2.6; P < 0.05). The relative range of C-CF was increased across all CCFT stages for the CCF group post intervention (SNK: all P < 0.00001; Fig. 4). In contrast, the strength-training group only demonstrated an increase in ROM at the 22 mmHg and 28 mmHg stages (SNK: P < 0.05; Fig. 4). 3.2. Arm movement task The analysis revealed that changes in the relative latencies of DCF, SCM and AS were not dependent on time or group. Nevertheless, visual inspection of the data suggested a tendency for

80

Strength training

Post Pre

60 40

*

20

*

0

22

24

*

26

*

28

Stage of the C-CFT (mmHg)

*

30

22

24

26

28

30

Stage of the C-CFT (mmHg)

Fig. 1. Normalised RMS values (mean and standard deviation) for the DCF muscles for each stage of the CCFT. Data are presented for the C-CF retraining group and strengthtraining group both pre and post intervention. *indicates signiﬁcant difference between pre and post intervention data (P < 0.05).

699

C-CF training

100

Strength training

Post Pre

80 60

∗

∗

40 20

∗

∗

∗

0 100 80 60 40 20 0

∗ 22

24

∗ 26

∗

∗

28

30

22

Stage of the C-CFT (mmHg)

24

26

28

30

Stage of the C-CFT (mmHg)

Fig. 2. Normalised RMS values (mean and standard deviation) for the left and right SCM muscles for each stage of the CCFT. Data are presented for the C-CF training group and strength-training group both pre and post intervention. *indicates signiﬁcant difference between pre and post intervention data (P < 0.05).

earlier onsets of the DCF muscles in both directions of arm movement for the C-CF training group post intervention (Figs. 5 and 6). Calculation of frequencies indicated an earlier onset of DCF EMG in 83.5% and 89% of subjects during arm ﬂexion and extension respectively for the C-CF group compared with 55% in each direction for the strength-training group. When the onset data were categorized based on the change in timing of the DCF (<40 – >40 ms in 10 ms increments) and the distribution of data directly compared across the two training groups, the subsequent analysis showed that the distribution of changes in relative latencies across

C-CF training

(R) AS normalised RMS (%)

DCF normalised RMS (%)

100

C-CF training

(R) SCM normalised RMS (%)

Age (years) Length of History (years) Onset (insidious, trauma) % insidious Neck Pain Intensity (NRS 0–10) Neck Disability Index (50)

C-CF training (n ¼ 23)

(L) AS normalised RMS (%)

Table 1 Baseline characteristics of the C-CF and strengthening exercise groups (Mean and standard deviation).

(L) SCM normalised RMS (%)

G.A. Jull et al. / Manual Therapy 14 (2009) 696–701

140

Strength training

Post Pre

120 100 80

∗

60 40 20

∗

0

∗

∗

∗

140 120 100 80

∗

60 40

∗

20

∗

∗

∗

0 22

24

26

28

Stage of the C-CFT (mmHg)

30

22

24

26

28

30

Stage of the C-CFT (mmHg)

Fig. 3. Normalised RMS values (mean and standard deviation) for the left and right AS muscles for each stage of the CCFT. Data are presented for the C-CF training group and strength-training group both pre and post intervention. *indicates signiﬁcant difference between pre and post intervention data (P < 0.05).

G.A. Jull et al. / Manual Therapy 14 (2009) 696–701

∗

40

0

∗

∗

∗ 22

24

26

28

30

22

Stage of the C-CFT (mmHg)

24

26

28

30

Stage of the C-CFT (mmHg)

Fig. 4. Percentage of full C-CF ROM (mean and standard deviation) for each stage of the CCFT are presented for the C-CF training group and strength-training group both pre and post intervention. *indicates signiﬁcant difference between pre and post intervention data (P < 0.05).

subjects was different between the two training groups (ﬂexion: c2 ¼ 22.55, P < 0.01; extension: c2 ¼ 37.45, P < 0.01; Fig. 7). 3.3. Measures of pain, disability and perceived beneﬁt Both exercise groups demonstrated a signiﬁcant reduction in average pain intensity (NRS) (C-CF training, P < 0.001; strength training P < 0.05), and NDI score (C-CF training, P < 0.001; strength training, P < 0.001) but there were no between group differences (both P > 0.05) (Table 2). Perceived beneﬁt of the exercises was also similar between groups. 4. Discussion Speciﬁc deﬁcits in DCF muscle activation have been identiﬁed in patients with neck pain compared to asymptomatic individuals (Jull et al., 2004; Falla et al., 2004b). This study showed that activation of the DCF was increased at each of the ﬁve levels of the CCFT and activity of the SCM and AS muscles reduced following C-CF training. The interaction between the deep and superﬁcial ﬂexors during the test changed so that it closely mirrored that measured previously in asymptomatic subjects (Falla et al., 2004b). There was an increase in the angle of C-CF used in each test stage, which

C-CF training

Strength training

Arm Flexion Relative latencies (ms)

250 Pre Post

200 150

50

-20

(R)AS

(L)AS

(R)SCM

(L)SCM

DCF

(R)AS

(L)AS

(R)SCM

(L)SCM

(R)AS

(L)AS

(R)SCM

(L)SCM

DCF

(R)AS

C-CF training Strength training

-25 -30

Arm Flexion

25 20 15 10 5

30

Arm Extension

C-CF training Strength training

25 20 15 10 5 0

Fig. 5. Mean and standard deviation of the relative latencies of the DCF, left (L) and right (R) SCM and AS are presented for the C-CF training group and strength-training group both pre and post intervention for both arm movement directions.

(L)AS

-15

30

Proportion of subjects (%)

100

DCF

Arm Extension Relative latencies (ms)

-10

0 35

150

0

-5

suggests a more accurate performance of C-CF rather than an aberrant pattern inclusive of a retraction action (Falla et al., 2004b) and parallels improvement in activation of the DCF. The strength training produced no substantive change in the activation of the deep and superﬁcial ﬂexors, thus did not address the altered neuromuscular strategy in the CCFT that has been measured regularly in patients with neck pain disorders (Jull et al., 2004; Falla et al., 2004b). This result may not be surprising as the C-CF training exercise and the outcome task were similar and there is considerable evidence that task speciﬁc improvements can be achieved with training (Weir et al., 1995; Young et al., 2001). Nevertheless of note clinically, improvement in DCF activation capacity was achieved with a recumbent, low load exercise. Cagnie et al. (2008) in a recent MRI study of three cervical ﬂexor exercise tasks showed that T2 changes in the longus capitis and the longus colli in the C-CF exercise were 42% and 19% respectively of that achieved in the high load head lift exercise of C-CF with cervical ﬂexion. Thus the evidence suggests that the low load C-CF exercise can train the DCF effectively, even in the early stages of rehabilitation when pain or pathology might preclude high load exercise.

50

200

Arm Extension

Fig. 6. Pre-post intervention change in relative latency of the DCF, left (L) and right (R) SCM and AS are shown for the C-CF training group and strength-training group for both arm movement directions.

100

0 250

(R)SCM

0

Change in relative latency (ms)

60

∗

∗

∗

Arm Flexion

Proportion of subjects (%)

% Full C-CF ROM

Post Pre

80

20

Strength training DCF

C-CF training 100

(L)SCM

700

<-40

-30 -40

-20 -30

-10 -20

0 -10

0 10

10 20

20 30

30 40

>40

Change in relative latency (ms) Fig. 7. Proportion of subjects showing a change in relative latency of the DCF muscles during unilateral arm movements for the C-CF training group and strength-training group.

G.A. Jull et al. / Manual Therapy 14 (2009) 696–701 Table 2 The change from baseline for the C-CF and strengthening exercise group in measures of pain and disability following exercise as well as the perceived beneﬁt of exercise.

Neck Pain Intensity (NRS 0–10) Neck Disability Index (50) Perceived beneﬁt of exercise (%) a

C-CF training (n ¼ 23)

Strength training (n ¼ 23)

Pa

1.7 2.0 5.0 4.2 60.4 24.7

1.0 3.3 3.5 2.3 54.6 27.3

0.35 0.15 0.44

Results of the between group analysis.

701

showed improved temporal characteristics of DCF muscle activation following cranio-C-CF training compared to strength training. These observations may partially explain the efﬁcacy of this exercise in rehabilitation of individuals with chronic neck pain. Acknowledgement Financial support was provided by a grant from the National Health and Medical Research Council of Australia. References

Coordination between the deep and superﬁcial ﬂexors is considered necessary for safe progression of exercise in patients with neck pain. There is evidence that both the cranio-cervical (Watson and Trott, 1993; O’Leary et al., 2007) and cervical ﬂexors (Barton and Hayes, 1996) have reduced strength and endurance in neck pain, thus warranting rehabilitation. It is unknown if the degree of impaired strength differs between the deep and the superﬁcial ﬂexors. However, it is known that SCM and AS together provide 83% of the cervical ﬂexion capacity while the longus capitis and longus colli provide only 17% (Vasavada et al., 1998). Thus if the coordination between the superﬁcial and deep ﬂexors is not corrected in the ﬁrst instance, work of the superﬁcial muscles might mask or substitute for any impaired performance of the DCF muscles in any premature progression to higher load exercise. Previous research in patients with neck pain revealed delays in activation of both the DCF and superﬁcial SCM and AS in response to postural perturbations, indicating a defect in the automatic feedforward control of the cervical spine (Falla et al., 2004a). Similar delays were recorded in this neck pain group with greatest delays identiﬁed for DCF EMG onset (Fig. 5). We proposed that the speciﬁc C-CF training exercise which focused on repeated activation of the DCF in a motor relearning model might redress this delay and do so more efﬁciently than the strengthening exercise, akin to the deep abdominal muscles using a similar training protocol (Tsao and Hodges, 2007, 2008). Although there were no signiﬁcant differences in pre-post intervention change in relative latency of DCF muscle between the two training groups, consistent with our hypothesis, a greater proportion of subjects showed earlier onsets of their DCF post intervention following C-CF training compared to strength training. Although earlier onset of DCF activity was identiﬁed during rapid arm movement, the relative latencies did not reach values consistent with data from a pain-free population (Falla et al., 2004a) and cannot be termed feedforward postural adjustments as the EMG onsets did not occur earlier than 50 ms after the onset of deltoid EMG. Further research is necessary to investigate whether increased training duration would induce greater changes. Studies have shown that patients achieve a progressive gain in EMG onset with continued training (Tsao and Hodges, 2008). Both exercise groups gained similar pain relief with the exercise training and considered the exercises beneﬁcial. The subjects, by design, had relatively mild neck pain syndromes and the study was not designed to evaluate the difference in efﬁcacy of the two interventions. Nevertheless these results reﬂect the pain relieving effects gained in previous trials of these exercise techniques (Bronfort et al., 2001; Jull et al., 2002). 5. Conclusion As hypothesised, speciﬁc low load C-CF training but not strength training enhanced the pattern of deep and superﬁcial muscle activity in the CCFT. In addition, a greater proportion of patients

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 and Medical Rehabilitation 1996;77:680–7. Berg HE, Berggren G, Tesch PA. Dynamic neck strength training effect on pain and function. Archives of Physical and Medical Rehabilitation 1994;75:661–5. Bronfort G, Evans R, Nelson B, Aker PD, Goldsmith CH, Vernon H. A randomised clinical trial of exercise and spinal manipulation for patients with chronic neck pain. Spine 2001;26:788–97. Cagnie B, Dickx N, Peeters I, Tuytens J, Achten E, Cambier D, et al. The use of functional MRI to evaluate cervical ﬂexor activity during different cervical ﬂexion exercises. Journal of Applied Physiology 2008;104:230–5. Falla D, Dall’Alba P, Rainoldi A, Merletti R, Jull G. Location of innervation zones of sternocleidomastoid and scalene muscles - a basis for clinical and research electromyography applications. Clinical Neurophysiology 2002;113:57–63. Falla D, Jull G, Dall’Alba P, Rainoldi A, Merletti R. An electromyographic analysis of the deep cervical ﬂexor muscles in performance of craniocervical ﬂexion. Physical Therapy 2003a;83:899–906. Falla DL, Campbell CD, Fagan AE, Thompson DC, Jull GA. Relationship between cranio-cervical ﬂexion range of motion and pressure change during the craniocervical ﬂexion test. Manual Therapy 2003b;8:92–6. Falla DL, Jull GA, Hodges PW. Patients with neck pain demonstrate reduced electromyographic activity of the deep cervical ﬂexor muscles during performance of the craniocervical ﬂexion test. Spine 2004a;29:2108–14. 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 2004b;157:43–8. Hermens H, Freriks B, Merletti R, Ha¨gg G, Stegeman D, Blok J, et al. European recommendations for surface electromyography (SENIAM 8). The Netherlands: Roessingh Research and Development; 1999. Hodges PW, Richardson CA. Feedforward contraction of transversus abdominis in not inﬂuenced by the direction of arm movement. Experimental Brain Research 1997;114:362–70. Jull GA. Deep cervical neck ﬂexor dysfunction in whiplash. Journal of Musculoskeletal Pain 2000;8(1/2):143–54. 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;27:1835–43. Jull G, Kristjansson E, Dall’Alba P. Impairment in the cervical ﬂexors: a comparison of whiplash and insidious onset neck pain patients. Manual Therapy 2004;9:89–94. Jull G, Sterling M, Falla D, Treleaven J, O’Leary S. Whiplash, headache and neck pain: research based directions for physical therapies. Edinburgh: Elsevier UK; 2008. McArdle WD, Katch FI, Katch VL. Exercise physiology: energy, nutrition and human performance. 4th ed. Baltimore: Williams and Wilkins; 1996. O’Leary S, Jull G, Kim M, Vicenzino B. Cranio-cervical ﬂexor muscle impairment at maximal, moderate, and low loads is a feature of neck pain. Manual Therapy 2007;12:34–9. Sterling M, Jull G, Vicenzino B, Kenardy J, Darnell R. Development of motor system dysfunction following whiplash injury. Pain 2003;103:65–73. Tsao H, Hodges P. Immediate changes in feedforward postural adjustments following voluntary motor training. Experimental Brain Research 2007;181:537–46. Tsao H, Hodges P. Persistence of improvements in postural strategies following motor control training in people with recurrent low back pain. Journal of Electromyography and Kinesiology 2008;18:559–67. Vasavada AN, Li S, Delp SL. Inﬂuence of muscle morphology and moment arms on moment-generating capacity of human neck muscles. Spine 1998;23:412–22. Vernon H, Silvano M. The neck disability index: a study of reliability and validity. Journal of Manipulative and Physiological Therapeutics 1991;14:409–15. Watson DH, Trott PH. Cervical headache: an investigation of natural head posture and upper cervical ﬂexor muscle performance. Cephalalgia 1993;13:272–84. Weir J, Housh T, Weir L, Johnson G. Effects of unilateral isometric strength training on joint angle speciﬁcity and cross-training. European Journal of Applied Physiology and Occupational Physiology 1995;70:337–43. Young W, McDowell M, Scarlett B. Speciﬁcity of sprint and agility training methods. Journal of Strength and Conditioning Research 2001;15:315–9.

Manual Therapy 14 (2009) 702–705

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Case Report

Positive patient outcome after spinal manipulation in a case of cervical angina Steven R. Passmore a, b, c, *, Andrew S. Dunn a, c a

Veterans Affairs of Western New York Health Care System, Buffalo, New York, USA Department of Kinesiology, McMaster University, Hamilton, Ontario, Canada c New York Chiropractic College, Depew, New York, USA b

a r t i c l e i n f o Article history: Received 1 September 2008 Received in revised form 23 February 2009 Accepted 23 March 2009 Keywords: Manipulation Spinal Radiculopathy Chest pain Angina

1. Introduction When angina pectoris is suspected but adequately ruled out, upper anterior chest pain and related symptoms may be attributed to cervical angina, particularly in the presence of radiculopathy and myelopathy (Nakajima et al., 2006). Cervical angina is theorized to involve the C6, C7, or T1 nerve roots, and possibly the medial and lateral pectoral nerves (Jacobs, 1990; Freccero and Donovan, 2005). While the prevalence of cervical angina is not completely clear, it is described as a virtually unknown and neglected clinical syndrome that may not be uncommon but is under diagnosed (Nakajima et al., 2006; Christensen et al., 2005). Aside from cardiac enzyme and exercise tolerance testing, Christensen et al. (2005) suggest cervical angina is potentially recognized from true angina through manual palpation of the spine and thorax. Manual palpation for locating vertebral joint dysfunction is most reliably indicated through eliciting tenderness (Jull et al., 1988; Hubka and Phelan, 1994; HaavikTaylor and Murphy, 2007). Further diagnostic tactics that may discern cervical from true angina may include improvement after corticotrophin therapy but not glyceryl trinitrate, normal temperature, and lack of vasomotor changes (Grieve, 1988). However, upon recognition of the presentation, what is an appropriate course of

* Corresponding author. Department of Kinesiology, McMaster University, Hamilton, Ontario, Canada L8S 4K1. Tel.: þ1 905 525 9140x26086; fax: þ1 905 527 3071. E-mail address: [email protected] (S.R. Passmore). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.03.005

care for cervical angina? Nakajima et al. (2006) suggest that anterior cervical surgical procedures that alleviate compression on nerve roots or the spinal cord and may be palliative. The same authors also mention surgery should be avoided until conservative measures have been exhausted, but caution the use of ‘‘aimless conservative therapy’’ (Nakajima et al., 2006). So the question remains what is an appropriate, focused approach to the conservative treatment of cervical angina? Furthermore, what type of intervention and treatment frequency should be recommended and what is the sustainability of such an intervention? This paper follows the case of a patient with cervical angina in whom radiographic ﬁndings can be correlated to patient presentation. The symptoms of cervical angina responded to a brief course of spinal manipulative therapy (SMT) targeted at the cervicothoracic junction, a novel ﬁnding previously absent in the literature. 2. Case description 2.1. History A 42-year-old woman was referred for evaluation and management of neck pain with cervical radiculopathy and comorbid cervical spondylosis. Upon subjective evaluation, symptoms reportedly developed over the previous ﬁve years, for which the patient had not actively pursued management options. During her

S.R. Passmore, A.S. Dunn / Manual Therapy 14 (2009) 702–705

history, the patient described her neck pain as ‘‘numb, dull, and sometimes accompanied by sharpness across the chest’’. The patient was naı¨ve to SMT, and was referred to the clinic by her primary care physician. 2.2. Examination A cervical spine radiographic study performed the week prior to consultation was compared to a study performed 3 years earlier that revealed minor anterior wedging of the C3 vertebral body without instability. The remaining cervical vertebral heights and alignment were otherwise maintained. The prevertebral soft tissues were intact. Subtle narrowing of the C4–5 disc space was observed and minimal uncovertebral spurring was noted on the left at the C3–4 and C5–6 levels (Fig. 1). The right intervertebral foramina were patent. During the objective examination a Neck Disability Index (Cleland et al., 2006) (NDI) score of 58 was recorded as a baseline outcome measure. Valsalva’s manoeuvre/test (Rubinstein et al., 2007), the Cervical compression test (Rubinstein et al., 2007), and Cervical traction/neck distraction test (Rubinstein et al., 2007) were positive. Muscular reﬂexes were 1–2þ and bilaterally symmetric at levels C5–7. Upper extremity clonus was absent. Motor strength was 4–5þ throughout. Sensory examination was hypoesthetic over the C7 dermatome. Passive cervical range of motion as observed was pain producing into extension, lateral ﬂexion and rotation toward end range. Hypokyphosis was noted in the thoracic spine. Segmental palpation (Jull et al., 1988) was painful at C3–4 bilaterally, and T4–5 on the left. Manual palpation revealed hypertonicity with tenderness in the suboccipital and levator scapulae musculature bilaterally. Due to the pre-existing diagnoses of cervical radiculopathy and spondylosis, with no additional upper extremity complaint peripheral joint examination was deferred at this time.

Fig. 1. Oblique view of the left cervical spine: Arrow indicates unconvertebral osteophytes encroaching at the C6 nerve root level.

703

2.3. Interventions and outcomes Treatment was initiated and the patient received high-velocity, low-amplitude (HVLA) SMT to the regions indicated by segmental palpation (C3–4 bilaterally, and T4–5 on the left) in conjunction with passive stretches for the suboccipital and levator scapulae bilaterally. Stretches were held once muscle tension was attained for 10 s. In the cervical spine a supine digit pillar pull (Peterson and Bergmann, 2002) was employed for SMT while in the thoracic spine a bilateral hypothenar transverse push (Peterson and Bergmann, 2002) was utilized. Immediately following treatment the patient verbally self-reported localized ‘‘decreased stiffness and pain’’ in the cervical region. These procedures and outcome occurred on two instances (treatment visits 1 and 2). The present body of literature on spinal manipulation has not speciﬁcally identiﬁed an optimal dosage of care in regard to the number and frequency of visits (Jull and Moore, 2002; Haas et al., 2004). While laboratory studies have examined changes immediately following a single instance of spinal manipulation (MartinezSegure et al., 2006; Tseng et al., 2006; Haavik-Taylor and Murphy, 2007), this immediate re-evaluation with completed outcome measures and full physical assessment has not yet translated into standard clinical practice. Clinical papers cite delay of thorough reevaluation for periods up to 9 or 12 visits, to follow a course of care as opposed to a single instance of intervention (Haas et al., 2004). In the present case the plan included a treatment frequency of two visits per week with re-evaluation after the forth visit in an attempt to minimize the number of patient visits needed to bring about clinical change or lack thereof, and indentify an appropriate end point early on in care to minimize potential over treatment (Dunn and Passmore, 2008). Upon presentation for her third scheduled appointment, she reported that the prior treatment had minimally sustained palliative effects, and that sleep the night before was interrupted by sensations of a self perceived myocardial infarction, including chest pain with dizziness, and pain in the lateral aspect of the left upper extremity with paresthesia distribution into the ﬁrst through third digits (Fig. 2), in addition to her previous neck pain. Manual therapy was deferred and the patient, who had a previous borderline exercise tolerance test (ETT) to investigate a prior apparent cardiac episode, was sent to the emergency department (ED) for evaluation. Immediately prior to referral for manual therapy the referring physician performed cardiac auscultation and described the presence of regular rate and rhythm of all heart sounds with the absence of murmur. The attending ED physician requested an electrocardiogram, and chest radiographs, but found no signiﬁcant ﬁndings. Serial cardiac enzyme laboratory analysis was also negative. Upper extremity pain radiation was provoked positionally by cervical spine left lateral ﬂexion (Lindgren et al., 1992) and the diagnosis of cervical angina was communicated. Having ruled out cardiac etiology, a return to manual therapy was determined. The patient’s course of care for neck pain with cervical spine radiculopathy was resumed at levels indicated by palpation in the cervical and thoracic spine (C3–4 bilaterally, and T4–5 on the left) with the addition of HVLA techniques directed speciﬁcally to hypomobile and tender segments (C5–6 on the left, C6–7 on the right and the T2 costotransverse joint on the left) at the cervicothoracic junction as cervical angina is theorized to involve the C6, C7, or T1 nerve roots (Jacobs, 1990; Freccero and Donovan, 2005). Manipulation directed at the cervicothoracic junction was delivered utilizing the thumb spinous push technique (Peterson and Bergmann, 2002). The patient lay prone and the clinician made a ﬁrst distal phalange to spinous process contact while inducing lateral ﬂexion targeted at facet joint manipulation. Also, the prone

704

S.R. Passmore, A.S. Dunn / Manual Therapy 14 (2009) 702–705

Fig. 2. Body chart illustrating pain and paresthesia presentation.

hypothenar costal push, (Peterson and Bergmann, 2002) was utilized in which the clinician’s hypothenar region contacted the patient’s upper costal angles, and HVLA SMT was delivered to manipulate the costovertebral articulations. These techniques were integrated into the existing treatment plan with delivery at an intended frequency of twice per week at dysfunctional motion segments (Dunn and Passmore, 2008). This was performed on two occasions before re-evaluation (treatment visits 3 and 4). After her ﬁfth scheduled appointment (treatment visit 4), a reevaluation was performed. Segmental palpation was no longer painful in the cervical or thoracic spine regions. Valsalva’s manoeuvre/test, Cervical Compression Test, and Cervical traction/ neck distraction test were painless and a re-evaluation NDI score of 48 was recorded. Immediately following treatment she noted a sustained decrease in cervical spine pain and chest pain, and following the second treatment directed to the cervicothoracic region her pain had resolved. While spontaneous recovery can never be completely dismissed, based on the temporal correlation of the decrease in symptom presentation, clinical ﬁndings, and outcome measure scores while undergoing a course of SMT it is theoretically feasible to attribute this patient’s improvement to the intervention. She followed up with her primary care physician for her biannual physical examination 11 weeks later without report of chest pain at that time. 3. Discussion Atypical chest pain presentations related to cervical radiculopathy entered the literature over 70 years ago when Nachlas (1934) identiﬁed what he described as ‘‘pseudo angina pectoris’’. Later, Hanﬂig (1936) suggested pain in the shoulder girdle, arm and precordium can be attributed to cervical arthritis. The term ‘‘cervical angina’’ came into favour following a publication by Jacobs (1990). Recent cases in the literature reported chest pain with associated nerve root impingement ranging from C6–T2 (Freccero and Donovan, 2005; Ozgur and Marshall, 2003; Yeung and Hagen, 1993). A retrospective chart review of 241 cases of C6–7 anterior cervical discectomy with unilateral nerve root impingement revealed that

15% presented with breast/chest pain (Ozgur and Marshall, 2003). Ozgur and Marshall (2003) proceed to state that of those 15% of individuals, 90% experienced long-term relief following anterior cervical discectomy and fusion, which they feel, is a clear indication of nerve root involvement. However, Grant and Keegan (1968) and more recently Erwin et al. (2000) suggest that costovertebral joints might be an under recognized site of pain generation in atypical chest, upper back and arm pain. Grant and Keegan (1968) found that pressure applied over the ribs that reproduced chest pain could be used to identify ‘‘costal syndrome’’, and additional pressure over the dorsal spine could identify ‘‘vertebro-costal syndrome’’ which occurred most often at a single vertebral level. Although no treatment options, descriptions, or protocol were detailed in the primarily diagnostic clinical paper the authors mentioned pain was often relieved by spinal traction. They also state that in their 41 clinical patients there were ﬁve consistent symptoms reported that were congruent with presentation which included accurate localization, accentuation by thoracic spine movement, exacerbation by breathing/coughing/ straining, association with posture/position, and a history of mechanical stress (Grant and Keegan, 1968). Erwin et al. (2000) demonstrate the anatomic possibility of pain production via evidence for the existence of enervated synovial folds in the costovertebral joints. It remains unclear if chest pain generation is occurring at the site of the nerve root, the costotransverse joints, or potentially the zygapophyseal joints (Erwin et al., 2000). In this patient when cervicothoracic SMT was delivered, it was directed at both the upper thoracic costotransverse joints, and lower cervical zygapophyseal joints. With a theoretical etiology of cervical radiculopathy, a positive response to manual therapy in this case of cervical angina was congruent with predictions theorized in a recent radiculopathy study (Cleland et al., 2007). In agreement with a cervical radiculopathy and manual therapy outcome prediction paper this patient’s anticipated response was favourable as her age is less than 54 years, and radicular symptoms were in her non-dominant arm (Cleland et al., 2007). Supporting a recent cervical radiculopathy outcome measure study, this patient demonstrated an NDI improvement that exceeds a minimally detected change of 7, and meets a minimally clinically important difference of 10 points improvement (Cleland et al., 2006). In a study on the diagnostic accuracy of cervical spine palpation, while there were no false-positives (medically asymptomatic zygapophysial joints diagnosed as symptomatic manually), there was an instance of a false-negative (medically symptomatic zygapophysial joints diagnosed as asymptomatic manually) (Jull et al., 1988). This false-negative occurred at the level of the C6–7 zygapophysial joint. It was also the only instance in which a joint below C5–6 was indicated by either medical or manual diagnosis. This ﬁnding could be interpreted as indicative that manual diagnoses of symptomatic lower cervical, or cervicothoracic joints may be more difﬁcult to discern. Symptomatic joints in the cervicothoracic region could have falsely been declared clear until patient symptomatic presentation clearly warranted more thorough assessment of the joints in this region. This was the case in the present course of management. This patient presented with unilateral osteophytes of the uncovertebral joints projecting into the intervertebral foramen potentially serving as the etiology of her cervical radiculopathy and cervical angina, but they may also be a benign comorbid condition. Had she not experienced a sustained response to SMT, other treatment options include but are not limited to active exercise, cervical spine traction, grade I–IV mobilisation, pharmacological intervention, or surgery. Authors of a recent study suggest that there is controversy over the possibility of relieving radicular symptoms without removal of offending osteophytes via direct surgical decompression of the uncovertebral joints. Shen et al. (2004) postulate that anterior cervical

S.R. Passmore, A.S. Dunn / Manual Therapy 14 (2009) 702–705

discectomy and fusion (ACDF) can add the 2–3 mm of distraction prerequisite for a favourable clinical outcome without osteophyte removal. Finding that it is unnecessary to remove potentially offending osteophytes to see clinical improvement supports why radicular symptoms could be alleviated by SMT, which has been demonstrated to increase zygapophysial joint space on MRI post manipulation (Cramer et al., 2002). However, should future patients fail to respond to cervicothoracic SMTor other conservative measures, and where patients fail to respond to ACDF, uncoforaminotomy to remove osteophytes is the suggested approach to care (Pechlivanis et al., 2006). When assessing a patient presenting with chest pain, a distinction between cervical angina, and true angina must be made to diagnose and manage cardiac issues distinct from mechanical neck, costovertebral and shoulder pain. Patients should be screened to rule out cardiac etiology ﬁrst. A study examining co-morbidity for people in their seventies with mild, moderate and severe neck and shoulder pain revealed a signiﬁcant percentage of individuals reported a history of angina (Vogt et al., 2003). With regard to neck pain rated mild, moderate, or severe, the associated percentages of people who reported angina were, respectively, 13.1%, 19.1%, and 16.4%. With regard to shoulder pain rated mild, moderate, or severe, the associated percentages of people who reported angina were, respectively 13.3%, 14.5%, and 18.6%.

4. Conclusion This case identiﬁed an individual with the under diagnosed phenomena of cervical angina. This patient demonstrated a sustained improvement up to 11 weeks following a brief trial of SMT directed to the cervicothoracic region, suggesting a mechanically based, musculoskeletal etiology to her presentation. Future prospective studies are needed to assess the viability of a course of SMT management, and the consideration of related treatments such as grade I–IV joint mobilisation for patients who have tested negative for true angina, but continue to present with unrelenting atypical chest and upper extremity pain prior to directing them for surgical management. These additional studies also need to conﬁrm the appropriate dosage of SMT for these patients following clearance and referral from cardiology through carefully controlled clinical studies with re-evaluation following each intervention. This case also raises the issue of the need for careful palpation to ascertain dysfunction in tissues, and the possibility that such dysfunction in the cervicothoracic junction may be more difﬁcult to identify through manual palpation then other regions. Future work is needed to determine whether it is HVLA manipulation of the costotransverse, zygapophyseal or the combination of these joints in conjunction with passive stretches that may have contributed to the therapeutic beneﬁt of this course of management.

Acknowledgments The research endeavours of Steven Passmore are funded in part by Fellowships from the Foundation for Chiroparactic Education and Research, New York Chiropratic College and an Ontario Graduate Scholarship.

705

References Christensen HW, Vach W, Gichangi A, Manniche C, Haghfelt T, Hoilund-Carlsen PF. Cervicothoracic angina indentiﬁed by case history and palpation ﬁndings in patients with stable angina pectoris. J Manipulative Physiol Ther 2005;28: 303–11. Cleland J, Fritz J, Whitman J, Heath R. Predictors of short-term outcome in people with a clinical diagnosis of cervical radiculopathy. Phys Ther 2007;87:1619–32. Cleland J, Fritz J, Whitman J, Palmer J. The reliability and construct validity of the Neck Disability Index and patient speciﬁc functional scale in patients with cervical radiculopathy. Spine 2006;31:598–602. Cramer GD, Gregerson DM, Knudsen JT, Hubbard BB, Ustas LM, Cantu JA. The effects of side-posture positioning and spinal adjusting on the lumbar Z joints: a randomized controlled trial with sixty-four subjects. Spine 2002;27:2459–66. Dunn AS, Passmore SR. Consultation request patterns, patient characteristics, and utilization of services within a Veterans Affairs Medical Center chiropractic clinic. Milit Med 2008;173:599–603. Erwin WM, Jackson P, Homonko D. Innervation of the human costovertebral joint: implications for clinical back pain syndromes. J Manipulative Physiol Ther 2000;23:395–403. Freccero D, Donovan D. Adjacent segment degeneration at T1–T2 presenting as chest pain. Spine 2005;30:E655–7. Grant AP, Keegan DA. Rib pain – a neglected diagnosis. Ulster Med J 1968;37:162–9. Grieve GP. In: Common vertebral joint problems. 2nd ed. Edinburgh: Churchill Livingstone; 1988. p. 393–4. Haas M, Groupp E, Aickin M, Fairweather A, Ganger B, Attwood M, et al. Dose response for chiropractic care of chronic cervicogenic headache and associated neck pain: a randomized pilot study. J Manipulative Physiol Ther 2004;27: 547–53. Haavik-Taylor H, Murphy B. Cervical spine manipulation alters sensorimotor integration: a somatosensory evoked potential study. Clin Neurophysiol 2007;118:391–402. Hanﬂig S. Pain in the shoulder girdle, arm, and precordium due to cervical arthritis. JAMA 1936;106:523–6. Hubka M, Phelan S. Interexaminer reliability of palpation for cervical spine tenderness. J Manipulative Physiol Ther 1994;17:591–5. Jacobs B. Cervical angina. NY State J Med 1990;90:8–11. Jull G, Bogduk N, Marsland A. The accuracy of manual diagnosis for cervical spine zygapophysial joint pain syndromes. Med J Aust 1988;148:233–6. Jull G, Moore A. What is a suitable dosage of physical therapy treatment? Man Ther 2002;7:181–2. Lindgren KA, Leino E, Manninen H. Cervical rotation lateral ﬂexion test in brachialgia. Arch Phys Med Rehabil 1992;73:735–7. Martinez-Segure R, Fenandez-de-las-Penas C, Ruiz-Saez M, Lopez-Jimenez C, Rodriguez-Blanco C. Immediate effects of neck pain and active range of motion after a single cervical high-velocity low-amplitude manipulation in subjects presenting with mechanical neck pain: a randomized control trial. J Manipulative Physiol Ther 2006;29:511–7. Nachlas I. Pseudo-angina pectoris originating in the cervical spine. JAMA 1934;103:323–5. Nakajima H, Uchida K, Kobayashi S, Kokubo Y, Yayama T, Sato R, et al. Cervical angina: a seemingly still neglected symptom of cervical spine disorder? Spinal Cord 2006;44:509–13. Ozgur B, Marshall L. Atypical presentation of C-7 radiculopathy. J Neurosurg Spine 2003;99:169–71. Pechlivanis I, Brenke C, Scholz M, Engelhardt M, Harders A, Schmieder K. Anterior uncoforaminotomy in the treatment of recurrent radiculopathy after anterior cervical diskectomy with fusion. Minim Invasive Neuorsurg 2006;49:323–7. Peterson DH, Bergmann TF, editors. Chiropractic technique. 2nd ed. St. Louis: Mosby; 2002. p. 222–76. Rubinstein SM, Pool JJ, van Tulder MW, Riphagen H, de Vet HC. A systematic review of the diagnostic accuracy of provocative tests of the neck for diagnosing cervical radiculopathy. Eur Spine J 2007;16:307–19. Shen F, Samartzis D, Khanna N, Goldberg E, An H. Comparisson of clinical and radiographic outcome in instrumented anterior cervical diskectomy and fusion with or without direct uncovertebral joint decompression. Spine J 2004;4: 629–35. Tseng Y, Wang W, Chen W, Hou T, Chen T, Lieu F. Predictors for the immediate responders to cervical manipulation in patients with neck pain. Man Ther 2006;11:306–15. Vogt M, Simonsick E, Harris T, Nevitt M, Kang J, Rubin S, et al. Neck and shoulder pain in 70- to 79-year-old men and women: ﬁndings from the health, aging and body composition study. Spine J 2003;3:435–41. Yeung M, Hagen N. Cervical disk herniation presenting with chest wall pain. Can J Neurol Sci 1993;20:59–61.

Manual Therapy 14 (2009) 706–708

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Case Report

Internal jugular vein thrombosis following cervical manipulation Thomas Hoffelner, Werner Maurer-Ertl*, Gregor Kienbacher, Roman Radl, Andreas Leithner, Reinhard Windhager University Clinic of Orthopaedic Surgery, Medical University of Graz, Auenbruggerplatz 5-7, 8036 Graz, Austria

a r t i c l e i n f o Article history: Received 5 April 2008 Received in revised form 5 February 2009 Accepted 6 February 2009 Keywords: Manual therapy Venous thrombosis In vitro fertilization Activated protein C resistance

1. Introduction Thrombosis of the internal jugular vein (IJVT) is a rare entity with potential for serious consequences. The possibility that jugular thrombosis may cause neck pain or a mass in the cervical region is usually overlooked. Exacerbated coagulation occurs due to a lesion of the vessel’s endothelium, venous stasis or hypercoagulability. IJVT is rarely spontaneous, which means there is no apparent predisposing factor for thrombosis (Unsal et al., 2003). Hereditary and acquired conditions predispose to thrombotic phenomena. In hereditary thrombophilia (HT), thromboembolic phenomena usually occur among young adults at unusual sites. The major causes of HT are Activated Protein C Resistance (APCR, Factor V Leiden mutation), antithrombin III and protein C and S deﬁciency, prothrombin gene mutation and hyperhomocysteinaemia. Common acquired causes of hypercoagulability are hormone treatment (increased blood viscosity), pregnancy, antiphospholipid syndrome, malignancy, dehydration, polycythaemia and nephrotic syndrome. Endothelial lesions often occur through venous catheterization, intravenous drug abuse, local infection, trauma and last but not least manipulation of the cervical spine. Clinical manifestation of IJVT may include the presence of a cervical mass, pain in the neck or shoulder during movement, facial oedema, cough and hoarseness (due to recurrent nerve compression by the thrombotic mass or oedema). There are

* Corresponding author. Tel.: þ43 664 859 6059. E-mail address: [email protected] (W. Maurer-Ertl). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.02.002

multiple reports in the literature of serious and at times fatal complications after cervical spine manipulation therapy (CSMT), even though CSMT is considered by some health providers to be an effective and safe therapeutic procedure for head and neck pain syndromes (Leon-Sanchez et al., 2007). The risk of severe vascular or neurological complications through manual therapy of the spine is low to very low, but minor side effects like fainting, dizziness or light headedness may develop even after many uneventful manipulations (Thiel et al., 2007). The popularity of chiropractic manipulations is rising among the general population, therefore as a consequence more adverse effects because of this treatment are reported. Many studies consider a causality relation between the neurological adverse events and the chiropractic manipulations, therefore the implementation of a risk alert system is recommended (Gouveia et al., 2007). 2. History A 30-year-old female, primigravida following in vitro fertilization and with a known heterozygous APCR, presented in our orthopaedic outpatient department one day after chiropractic manipulation of the cervical spine. In general she was a very ﬁt person although she did not do any sports except nordic walking. She stopped smoking recently. At the time of her visit she was not taking any anticoagulant. Two years ago the patient did have a subdermal single rod etonogestrel implant (Implanon). Three months after becoming pregnant following in vitro fertilization she was doing ﬁne and did not have any complications caused by the intervention.

T. Hoffelner et al. / Manual Therapy 14 (2009) 706–708

Complaining of spontaneous pain in the neck as well as in her right shoulder she consulted her chiropractor because the pain, continuously radiating to her right shoulder, had remained the same for two days and it stopped her from working as a hairdresser. At the time of consultation with the chiropractor she was not taking any medication. After examining the lady the chiropractor made the diagnosis of general hypermobilisation, a painful restriction of rotation of the neck to the right hand side, problems with elevation of the right shoulder and no neurological deﬁcits. In the same session manual treatment of the neck was performed apparently without consideration of the risk factors for thrombosis. Taking into account the pregnancy no X-ray of the cervical spine was ordered before treatment. Manipulation was deﬁned as the application of rotation of the vertebrae C2–C3 and the cervico-thoracic intersection. After intervention, manual unlocking was achieved, following which rotation to the left hand side and elevation of the right shoulder were normal. Shoulder pain decreased but neck pain seemed to become worse. A few hours later a right sided cervical swelling occurred and next morning she had mild dysphagia. In view of this she decided to see the orthopaedic outpatient department at our clinic. 3. Examination Observation revealed a tender, poorly deﬁned mass in the area of the right sternocleidomastoid muscle (Fig. 1) with no apparent source of infection in the head and neck region. There was a signiﬁcant decrease in the range of motion of the neck, but no neurological deﬁcit. An ultrasound scan (Fig. 2) showed complete occlusion of the right internal jugular vein (RIJV) with surrounding increased soft tissue density but no evidence of abscess formation. After initial treatment with high dose lowmolecular-weight (LMW) heparin (Enoxaparin 1 mg/kg/day), the neck oedema and discomfort promptly resolved within two days. After two weeks of hospital treatment at the gynaecological ward with relative bed rest, the follow-up ultrasound scan showed complete recanalization of the RIJV and the patient was discharged. Further treatment included anticoagulation with LMW Heparin (Enoxaparin 60 mg/0/40 mg) and a follow up examination was scheduled. 3.1. Review appointment three months later Three months after IJVT, the patient complained of vaginal bleeding and was again admitted to the gynaecological ward.

Fig. 1. Frontal picture on the day of admission with right sided cervical swelling.

707

Fig. 2. Axial doppler ultra-sound of the neck; Typical appearance of a RIJV thrombosis.

Clinical examination and ultrasound scan of the RIJV were unremarkable and the patient could be discharged the next day. The anticoagulation with LMW Heparin (Enoxaparin 60 mg/0/40 mg) was continued and a check-up in one month’s time was scheduled. Ultrasound follow-up studies showed complete restoration of the RIJV with unremarkable blood ﬂow. Laboratory ﬁndings for clotting factors including the anti-factor Xa were within normal range. The patient was sufﬁciently anticoagulated for the duration of her pregnancy and 6 months after IJVT she gave birth to a healthy boy. 4. Discussion IJVT is comparatively rare but a potentially fatal condition. As mentioned above, IJVT is more often due to direct trauma of the vein (e.g. direct catheterization, repeated intravenous injections by drug users, cervical manipulation or direct extension of tumor) compared with a spontaneous IJVT. In the reported case the patient unfortunately presented with three potential risk factors for getting an IJVT. First, in the medical history a HT caused by APC-Resistance is present, second the woman has had IVF and third her cervical spine was manipulated. Many cases reporting on chiropractic manipulations of the neck suggest that rotation or traction of the cervical spine can lead to dissections of the cervical vessels (Reuter et al., 2006). It may thus be speculated that chiropractic manipulations of the neck can lead to small lesions in the wall of the vessels which can remain asymptomatic or cause transient neurological symptoms which can be more or less worrying (Khan et al., 2005). There is little doubt that IJVT is underreported, as most cases present with symptoms, who do not give any information about the disease, such as little headache, neck pain and other small physical ﬁndings (Arya et al., 2001). In general compared to other reported cases of IJVT it is conspicuous that most of the patients concerned have been female. Although chiropractic interventions are very convenient in daily routine to treat pregnant women presenting cervical or lumbar pain, therapists also have to keep in mind that pregnant women are more likely to develop hypercoagulation caused by their hormone status, because the physiological changes that occur during pregnancy create a hypercoagulable milieu (Patnaik et al., 2007). With regard to this fact, literature about thrombophilia in pregnancy has conﬁrmed that women with thrombophilia are at increased risk of developing complications during pregnancy (Mancuso et al., 2005). Furthermore, Robertson et al. (2007) reported, that woman with

708

T. Hoffelner et al. / Manual Therapy 14 (2009) 706–708

thrombophilia are at increased risk of developing complications during pregnancy. Despite the increase in relative risk, the absolute risk of venous thromboembolism and adverse outcomes in pregnancy remains low. The casual relationship between complication rate of chiropractic manipulations and education or the speciﬁc knowledge of the manipulator is put into perspective by Reuter et al. (2006) who reported that 50% of the accidents were at the hands of well trained orthopedists. Due to the lack of studies, gaps still exist in our knowledge especially of the risk of pregnancy-related internal jugular vein thrombosis associated with APC-R and manipulation of the cervical spine. For this reason the authors recommend that more emphasis be given to the medical history before starting manipulation and consideration be given to alternative manual treatment methods in case of high risk patients. 5. Conclusion It remains unclear whether the IJVT in this case was caused by endothelial damage through chiropractic intervention in addition to the pre-existing risk factors, or a pre-existing IJVT was aggravated by compression of the neck. However, as there are several case reports describing vascular accidents subsequent to manual therapy, a thorough medical history including risk factors for thrombosis should be taken and a clinical examination should be performed before force is applied to the neck. In the case described it would have been prudent for the chiropractor not to have

manipulated the patient’s cervical spine given the number of risk factors listed above.

References Arya R, Shehata HA, Patel RK, Sahu S, Rajasingam D, Harrington KF, et al. Internal jugular vein thrombosis after assisted conception therapy. British Journal of Haematology 2001;115(1):153–5. Gouveia LO, Castanho P, Ferreira JJ, Guedes MM, Falcao F, e Melo TP. Chiropractic manipulation: reason for concern? 2007;109(10):922–25. Khan AM, Ahmad N, Li X, Korsten MA, Rosman A. Chiropractic sympathectomie: carotid artery dissection oculosympathetic palsy after chiropractic manipulation of the neck. The Mount Sinai Journal of Medicine, New York 2005;72(3):207–10. Leon-Sanchez A, Cuetter A, Ferrer G. Cervical spine manipulation: an alternative medical procedure with potentially fatal complications. Southern Medical Journal 2007;100(2):201–3. Mancuso A, De Vivo A, Fanara G, Di Leo R, Toscano R. Upper body venous thrombosis associated with ovarian stimulation: case report and review of the literature. Clinical and Experimental Obstetrics and Gynecology 2005;32(3): 149–54. Patnaik MM, Haddad T, Morton CT. Pregnancy and thrombophilia. Expert Review of Cardiovascular Therapy 2007;5(4):753–65. Reuter U, Ha¨mling M, Kavuk I, Einha¨upl KM, Schielke E. Vertebral artery dissection after chiropractic neck manipulation in Germany over three years. Journal of Neurology 2006;253(6):724–30. Robertson L, Wu O, Langhorne P, Twaddle S, Clark P, Lowe GD, et al. Thrombophilia in pregnancy: a systematic review. British Journal of Haematology 2007;132(2):171–96. Thiel HW, Bolton JE, Docherty S, Portlock JG. Safety of chiropractic manipulation of the cervical spine. A prospective national survey. Spine 2007;32(21):2375–8. Unsal EE, Karaca C, Ensarı´ S. Spontaneous internal jugular vein thrombosis associated with distant malignancies. European Archives of Oto-Rhino-Laryngology 2003;260(1):39–41.

Manual Therapy 14 (2009) 709–711

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Professional Issue

Manual therapy education. Does e-learning have a place? Paul Bowley, Liz Holey* School of Health and Social Care, University of Teesside, Borough Road, Middlesbrough TS1 3BA, UK

a r t i c l e i n f o Article history: Received 10 October 2008 Received in revised form 29 January 2009 Accepted 3 February 2009

1. Introduction The practical, psychomotor skills integral to manual therapy require considerable development to ensure a practitioner is competent to practise safely. Traditionally, this has been learnt through a cycle of observed demonstration, practice and teacher feedback where the student’s attempts are observed and commented upon, followed by a reﬁnement of practice, of tasks designed with a gradual increase in complexity. This process is both effective and efﬁcient for the learner. To enable autonomous professional clinical practise these skills must be embedded within a framework of assessment, diagnosis, clinical decision-making, evaluation and reﬂection. This ensures that an individual needs-based assessment package is prescribed and delivered effectively over a course of time (Holey and Cook, 2003). The resulting reﬂective practitioner (Schon, 1987) that is able to be self-critical and maintain competence over a working life. Experience has shown that learning the psychomotor skills and intellectual framework in an integrated way is the most effective. This has led to an assumption that e-learning, therefore, is an inappropriate learning and teaching strategy for manual therapy, but this paper argues that it has a place in supporting and enhancing the learning of the manipulative therapies. 2. Intellectual skills The expert becomes distinct from the novice by his/her ability to make complex clinical decisions through a sophisticated process of clinical reasoning. Progression from novice to expert is developed through clinical experience. It is thought that by experiencing a wide range of patients with their varied clinical presentations and differing responses to treatment enables the expert to recognise * Corresponding author. Tel.: þ44 1642 384119. E-mail address: [email protected] (L. Holey). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.02.001

patterns and thus makes more accurate clinical decisions, more quickly (Jones and Rivett, 2004). The integration of the underpinning theory and its use to inform problem-solving, decision-making and clinical reasoning can be learned through the exploration of case studies. These options can be undertaken on-line, without direct contact with tutor or peers and lone learning can be supported by explanations, examples and anecdotes being provided in a written or diagrammatic form on an electronic learning platform. There is a danger, however, that independent learning from written material may encourage a passive learning style. The passivity involved in learning from written material must, therefore, be supplemented by an interaction with the material and an active engagement with decision-making. 2.1. Psychomotor skill acquisition A student of manual therapy needs to experience the feel of normal tissues and joints and the way they respond to manipulative techniques. They then need to experience the same in a variety of patients with different abnormalities. Only then will the student understand, in a ﬁeld suffering from a paucity of objective measures of tissue response, when these responses to treatment are ‘normal’, desirable and related to symptoms. It is accepted that there is no substitute for real palpation and handling of joints and tissues, and communicating and responding to human individuals. High level technological learning objects can, however, supplement this practice in actuality and prepare students to practise on humans. 2.2. The possibilities of e-learning A virtual learning environment can be created in which a total learning experience can be enjoyed. Activities within this environment can be used to enrich learning in novel ways. Sensitive selection of learning activities and experiences at appropriate points, and skilful linkage of activity to learning outcomes can take

710

P. Bowley, L. Holey / Manual Therapy 14 (2009) 709–711

the learner through a journey of discovery, achievement, feedback and fun. The journey can be taken in the virtual company of others, via electronic communication. Sophistication varies from the reading of material, watching video clips and PowerPoint presentations, to watching animations and engaging in tests, quizzes, interactive case-based exercises and electronic discussions. 2.3. Deﬁnitions of e-learning Deﬁnitions of e-learning proliferate, but for the purposes of this discussion e-learning can be deﬁned as technology-based instruction delivered by means of a computer network (Welsh et al., 2003; Zhang et al., 2004), most commonly nowadays via the World Wide Web. Two distinct types of e-learning have been identiﬁed (Welsh et al., 2003): synchronous, where all students access learning materials at the same time; and asynchronous, where pre-prepared materials are accessed by students at any time. Currently, the most popular forms of asynchronous e-learning are blogs, wikis and podcasts, which will be the initial subject of discussion. Technologies that are more obviously applicable to the ﬁeld of manual therapy, namely simulations and devices that provide haptic feedback will then be explored. Whilst adding fun can be an added beneﬁt, it is important that the on-line teacher is able to select the tools which are appropriate to the learning outcomes and that ‘gimmicks’ are avoided. 2.4. Really simple syndication (RSS) RSS is a means of disseminating content from frequently updated Web sites through documents known as ‘feeds’, and is connected with blogs, wikis and podcasts. RSS feeds (composed of XML – Extensible Markup Language) can be gathered from Web sites through the use of programs known as ‘readers’ or ‘aggregators’. Users direct such programs to visit Web sites of interest to pick up the latest feeds, which may contain listings of news headlines or podcasts, for example. In this way, content can be collected conveniently from a variety of sources without the user having to visit a number of Web sites. 2.5. Blogs The term ‘blog’ is a contraction of ‘Web log’, and is thought to have come into use in 1999 (Williams and Jacobs, 2004). The blog in its most basic form is an on-line journal with entries displayed in reverse chronological order. Entries are mostly textual, but can contain links to other Web sites and multimedia. Blogs also generate RSS feeds. An important feature of the blog is that readers can post comments to each entry, which lends the blog a collaborative, conversational style. As an educational tool, therefore, Blogs provide a means of sharing knowledge and a forum for debate on any conceivable topic. In Williams and Jacobs (2004) study of an MBA blog at Brisbane Graduate School of Business, most students felt that the blog ‘increased the level of meaningful intellectual exchange’ amongst peers. Glogoff (2005) sees the addition of comments to blog entries as a useful means of giving immediate feedback to students on their blog entries, which in turn can be an aid to directed learning and guided discovery.

pages, the content of which can be modiﬁed by anyone with relative ease through a Web browser. Most wikis maintain archived copies of their Web pages, allowing the evolution of each page to be monitored. Like blogs, wikis can also generate RSS feeds for the dissemination of the latest updates. The open editing feature of the wiki lends itself to a number of uses in an e-learning context: collaborative writing projects, group problem-solving projects, the creation of shared information sources and the provision of customised electronic portfolios have been identiﬁed (Ferris and Wilder, 2006). The same feature has led to the accuracy of the content of wikis being viewed with some suspicion. However, Boulos et al. (2006) posit that the collaborative editing process should lead to a ‘survival of the ﬁttest’ content through a kind of natural selection. Nevertheless, wikis require careful monitoring to ensure accuracy of content. 2.7. Podcasts Cebeci and Tekdal (2006) deﬁne podcasting as the Web syndication (e.g. through RSS) of audio content (e.g. voice or music recordings) to mobile devices (e.g. mobile phones or MP3 players). In fact, the audio content of a podcast can be played on virtually any personal computer. The ease with which the podcast can be created and distributed is one of its key features. An inexpensive microphone, a personal computer and some free, open-source recording software will sufﬁce (Tavales and Skevoulis, 2006). In an educational context, podcasts can be used to distribute recordings of lectures and supplementary materials, making them a convenient reference resource that can be accessed regardless of time and, given the appropriate player, location. They also offer students the option of learning through listening, which some students may prefer and others, such as the visually challenged, may require (Cebeci and Tekdal, 2006). 2.8. Simulations Simulations can range from Web-based applications containing text, graphics and multimedia (sound, video) to three-dimensional (3D) model environments. An example of the former type is Web-SP, a Web-based simulated patient system (Brutlag et al., 2006). Cases are customised and can include patient interviews, physical examinations and laboratory tests. Students can also make diagnoses and receive feedback on them. Of the latter type is the 3D virtual radiography application developed at the University of Teesside. This application allows students to position a virtual patient and X-ray tube independently in a number of scenarios and see the effects of those manipulations on the resulting X-ray exposure. The program has so far proved invaluable in training students of diagnostic radiography in a radiation-free environment. Also of the latter type, this time incorporating a haptic feedback device, is the Virtual Haptic Back project (Williams et al., 2003). This application features a virtual 3D model of the human back and the Personal Haptic Interface Mechanism (PHANToM). The PHANToM is a mechanical device through which the student can apply forces to the virtual back and receive a realistic sensation of the resistance to those forces, as if he or she was palpating a human back. 2.9. Evidence-based learning and teaching

2.6. Wikis The word ‘wiki’ comes from the Hawaiian meaning ‘swift’ or ‘to hurry’ (Boulos et al., 2006). The most well-known example of the wiki is Wikipedia – The Free Encyclopedia. The wiki is a collection of Web

Just as our clinical practise should be evidence-based, so should our teaching. A detailed critical review of the literature is beyond the scope of this paper but it is recognised that evidence for the effectiveness of e-learning remains scanty. It has, however, been

P. Bowley, L. Holey / Manual Therapy 14 (2009) 709–711

found that e-learning is suitable for use in Higher Education (Koskela et al., 2005). To reduce learner isolation, a buddy system can be beneﬁcial (Palloff and Pratt, 1999). Unsurprisingly, students with high computer conﬁdence appear best suited to e-learning (Lain and Aston, 2004). Graff (2006) found that student performance is improved by a sense of on-line community, and engagement with assessment. Lake (2001) has identiﬁed that student satisfaction may decrease with the need to take learner responsibility. 2.10. Blended learning Our experience of delivering e-learning at the University of Teesside, UK has shown that the blended learning approach is particularly appropriate and effective. This is where e-learning is supplemented with periods of face-to-face contact and this is essential where manual therapy skills must be learnt safely. This must be woven into a total, logical learning experience where the context is set at induction, information is given and shared and the learning experience is structured via the electronic platform. Learners are most likely to complete the course if an on-line community is built. Assessment and testing can be electronic and include a practical assessment of manual skills which is commonly combined with a viva element to test clinical reasoning and justiﬁcation of intervention. A component of assessment which is essential for learning is feedback on performance and achievement. To enhance the quality of the electronic learning platform, a managed learning environment can be used. This is where the electronic delivery of materials is staged; for example, access is not given to new material until a quiz has been completed. This provides structure to learning, avoiding overwhelming the student by the volume of material required for completion of a whole module. It also enables close monitoring of student engagement, and can provide feedback through on-line tests and quizzes, which the student is motivated to complete so they can access more material and obtain formative feedback. 3. Conclusion E-learning can be useful to students. It can be effective in the learning and teaching of manual therapy if a blended learning approach is used. Engagement is more effectively sustained

711

through the provision of a sense of student community via interstudent electronic interaction and practical workshops. The enjoyment and quality of e-learning are enhanced by a high level of student engagement and the use of a managed learning environment to structure learning. References Boulos MNK, Maramba I, Wheeler S. Wikis, blogs and podcasts: a new generation of web-based tools for virtual collaborative practice and education. BMC Medical Education 2006;6:41. Brutlag P, Youngblood P, Ekorn E, Zary N, Fors U, Dev P, et al. Case-Ex: examining the applicability of web-based simulated patients for assessment in medical education. In: Reeves T, Yamashita S, editors. Proceedings of World Conference on E-Learning in Corporate, Government, Healthcare, and Higher Education 2006. p. 1869–70. Cebeci Z, Tekdal M. Using podcasts as audio learning objects. Interdisciplinary Journal of Knowledge and Learning Objects 2006;2:47–57. Ferris S, Wilder H. Uses and potentials of wikis in the classroom. Innovate 2006;2:5. Glogoff S. Instructional blogging: promoting interactivity, student-centered learning, and peer input. Innovate 2005;1:5. Graff M. The importance of on-line community in student performance. Electronic Journal of e-Learning 2006;4(2):127–32. Holey EA, Cook EM. Evidence-based therapeutic massage. Edinburgh: Churchill Livingstone; 2003. p. 99 [chapter 6]. Jones MA, Rivett DA. Clinical reasoning for manual therapists. 1st ed. Butterworth Heinemann; 2004. Koskela M, Kitti P, Vilpola I, Tervenoven J. Suitability of a virtual learning environment for higher education. Electronic Journal of e-Learning 2005;3(1): 21–30. Lain D, Aston J. Literature review of evidence on e-learning in the workplace. UK: Institute of Employment Studies; 2004. Lake D. Student performance and perceptions of a lecture-based course compared with the same course utilizing group discussion. Physical Therapy 2001;81(3):896–902. Palloff R, Pratt K. Building learning communities in cyberspace. San Francisco: Jossey-Bass; 1999. Schon D. Educating the reﬂective practitioner. San Francisco: Jossey Bass; 1987. Tavales S, Skevoulis S. Podcasts: changing the face of e-learning, http://ww1.ucmss. com/books/LFS/CSREA2006/SER4351.pdf; 2006 [accessed 31.08.08]. Welsh ET, Wanberg CR, Brown KG, Simmering MJ. E-learning: emerging uses, empirical results and future directions. International Journal of Training and Development 2003;7(4):245–58. Williams JB, Jacobs J. Exploring the use of blogs as learning spaces in the higher education sector. Australasian Journal of Educational Technology 2004;20(2):232–47. Williams II RL, Srivastava M, Conatser Jr. RR, Howell JN. The virtual haptic back project. Proceedings of the 2003 Image Society Conference, Scottsdale, AZ; July 2003. p. 14–8. Zhang D, Zhao JL, Zhou L, Nunamaker Jr JF. Can e-learning replace classroom learning? Communications of the ACM 2004;47(5):74–9.

Manual Therapy 14 (2009) 712–715

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Technical and Measurement Report

Analysis of tibial rotation using magnetic resonance imaging Mina Samukawa a, *, Toru Yamamoto b, Shigenori Miyamoto c, Aogu Yamaguchi d, Masaki Katayose e a

Division of Physical Therapy, Department of Health Sciences, Hokkaido University School of Medicine, Kita 12jo, Nishi 5chome, Kita-ku, Sapporo, Hokkaido 060-0812, Japan Division of Radiological Technology, Department of Health Sciences, Hokkaido University School of Medicine, Kita 12jo, Nishi 5chome, Kita-ku, Sapporo, Hokkaido 060-0812, Japan c Department of Physical Therapy, Faculty of Human Science, Hokkaido Bunkyo University, 196-1, Koganechuo 5 chome, Eniwa, Hokkaido 061-1449, Japan d Department of Radiology, Hokkaido University Hospital, Kita 14jo, Nishi 5chome, Kita-ku, Sapporo, Hokkaido 060-8648, Japan e Department of Physical Therapy, Graduate School of Health Sciences, Sapporo Medical University, Minami 1jo, Nishi 17chome, Chuo-ku, Sapporo, Hokkaido 060-8556, Japan b

a r t i c l e i n f o Article history: Received 25 April 2008 Received in revised form 17 December 2008 Accepted 2 January 2009

1. Introduction The knee joint is the largest joint in the human body and is capable of ﬂexion-extension, varus-valgus, and internal-external rotation movements (Takeda et al., 1994). Together with the surrounding muscles, the ligaments of the knee play a key role in controlling the displacement of the tibia on the femur, and injury to these ligaments can cause joint laxity, or instability, leading to loss of joint control (Magee, 2002). The tibia rotates internally when the knee ﬂexes and externally when it extends, a phenomenon known as ‘‘screw-home motion’’ (Hallen and Lindahl, 1966; Piazza and Cavanagh, 2000). The articular surfaces of the tibia and the femur are incongruent, enabling the two bones to move by differing amounts, guided by muscles and ligaments (Magee, 2002). Two features common to previous studies that need to be addressed are that the observed angles of tibial rotation vary widely and the studies are difﬁcult to generalize. One possible explanation is that the studies did not share a common deﬁnition of tibial rotation. In addition, tibial rotation was usually examined at angles close to terminal knee extension so as to conﬁrm the existence of screw-home motion and to determine the effects of injuries (Karrholm et al., 1989; Matsumoto, 1990; Matsumoto et al., 2000). Lastly, side-to-side differences are as yet undetermined even though this is considered essential information for clinicians wanting to see the effects of injuries and to regain tibiofemoral joint mobility.

* Corresponding author. Tel./fax: þ81 11 706 3392. E-mail address: [email protected] (M. Samukawa). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.01.003

Magnetic Resonance Imaging (MRI hereafter) is becoming an increasingly valuable tool for in vivo studies of musculoskeletal biomechanics, as it is a non-invasive means of obtaining precise anatomic and geometric information. To the best of our knowledge, MRI has never been used to measure tibial rotation in normal knees although previous researchers have compared the tibial rotation of pathological knees with that of intact knees (Czerniecki et al., 1988; Nagao et al., 1998; Brandsson et al., 2002; Georgoulis et al., 2003). Furthermore, most of these studies examined tibial rotation when the knee position was close to fully extended. Although several studies have been conducted to determine tibial displacement on the femur at different knee angles of ﬂexion (Iwaki et al., 2000; Logan et al., 2004), the degree of in vivo tibial rotation at different angles remains relatively undetermined. As such, the purposes of this study were two-fold: to determine the angle of tibial rotation at different angles of knee ﬂexion using MRI, and to identify any side-to-side differences. 2. Methods 2.1. Subjects Ten normal subjects (5 male and 5 female) were recruited for this study. The mean age was 19.9 1.6 years old; mean weight was 55.9 10.7 kg; and mean height was 162.3 8.4 cm. The subjects had no previous histories of knee injuries or neurological disorders. Potential subjects were excluded if they had any general joint laxity or history of claustrophobia. Sarji et al. (1998) found up to 0.5% of MRI examinations fail due to claustrophobia. All subjects were required to read and sign an informed consent form.

M. Samukawa et al. / Manual Therapy 14 (2009) 712–715

713

Fig. 1. Position of subject when measuring tibial rotation.

2.2. Test procedure The Sapporo Medical University Institutional Review Board approved this study. Subjects lay in a supine position inside the MRI magnet on a specially-designed apparatus made of wood and plastic that allowed the examiner to set the knee ﬂexion angles and also stabilized the femur and tibia (Fig. 1). The hip angle was ﬁxed at 45 degrees, and knee ﬂexion angles of 30, 60, and 90 degrees were chosen for measurement. The position of each subject on the MRI apparatus is shown in Fig. 2. Subjects were asked to relax during the measurement, and the knee ﬂexion angles were selected arbitrarily. A clinical 1.5-T whole-body MR scanner (Magnetom Symphony, Siemens Medical Solutions, Erlangen, Germany) was used throughout this study. MR images were obtained using 3-dimensional true fast imaging with steady-state procession (FISP) with the following imaging parameters: TR 4.3 ms, TE 2.08 ms, matrix

Fig. 2. Position of subject when measuring tibial rotation in MRI.

size 256 180, ﬂip angle 55 degrees, slice thickness 3.0 mm, ﬁeld of view 300 mm 281 mm. Using a built-in body coil, 52 axial slices were obtained from each knee. The 60 s acquisition time included L-R directional phase oversampling to avoid wrap-around from the other knee. The tibial rotation angle was veriﬁed from MR Images, and the same radiologic technologist (A.Y.) took all the images. The anatomical reference points used in the present study (Fig. 1) had previously been described and shown to be reliable (Lerner et al., 2003). The sagittal slices were reformatted from the set of 52 slices and the sagittal slice in which the top of the intercondylar notch of the femur ﬁrst appeared was chosen. To monitor the rotation of the femur, the position of the slice was determined on the selected sagittal slice perpendicular to the longitudinal axis of the femur and crossing the distal femoral point. This slice was also reformatted from the set of 52 slices so as to deﬁne the line connecting the medial and lateral femoral points (c in Fig. 3) as an indicator of

Fig. 3. Establishment of anatomic reference points for this measurement. The angle of tibial was deﬁned as the angle formed by the line c and the line f.

714

M. Samukawa et al. / Manual Therapy 14 (2009) 712–715

femoral line. Similarly, to monitor the tibial line, a line connecting the medial and lateral points on the posterior tibia was determined on the tibial aspect slice (f in Fig. 3), which was then reformatted from the set of 52 slices in respect to the longitudinal axis of the tibia and the position of the posterior condyle. The vector angle between the lines (c & f in Fig. 3) representing the rotational angle was then measured. A positive value was interpreted as the degree of internal rotation, and a negative value as the degree of external rotation.

Table 1 Descriptive statistic for measurement of tibial rotation at 30, 60, and 90 degrees of knee ﬂexion. Knee ﬂexion angels

Mean (deg) S.D. (deg) Median Minimum (deg) Maximum (deg)

30 degrees

60 degrees

90 degrees

Right

Left

Right

Left

Right

Left

4.3 10.9 6.0 19.0 15.0

7.1 4.5 7.0 1.0 14.0

7.2 6.6 7.0 1.0 19.0

7.2 4.2 7.0 1.0 14.0

5.4 4.6 4.0 1.0 15.0

5.6 5.6 4.5 1.0 18.0

2.3. Data analysis

‘‘’’ values meant tibia was externally rotated. The tibial rotation angles were internally rotated throughout the three angles of ﬂexion.

A repeated-measures analysis of variance (ANOVA) was used to ascertain any differences in the amount of tibial rotation at 30, 60, and 90 degrees of ﬂexion. In the presence of a main effect, a post hoc test (Fisher’s PLSD test) was then used to determine differences in the angle of tibial rotation at the 3 angles of knee ﬂexion. A paired ttest was also conducted to compare sides within each individual and to test for any systematic side-to-side differences. Pearson’s correlation coefﬁcient was used to compare the side-to-side differences of the rotational angles. Analyses were performed using Statview 5.0 software (SAS institutes Inc, Berkley, CA). A p-value of less than 0.05 was set as the signiﬁcance level.

and Lindahl, 1966; Todo et al., 1999; Iwaki et al., 2000; Piazza and Cavanagh, 2000; Logan et al., 2004). Speciﬁcally, previous researchers did not sufﬁciently describe how the tibial rotation measurements were conducted, either by failing to adequately deﬁne tibial rotation or use consistent measurement points, or both. With this in mind, the present study was designed to examine the nature of the congruence between the tibia and the femur at a greater range of ﬂexed knee positions (30, 60, and 90 degrees) and to clearly describe how to measure tibial rotation. The results of this study show that the tibia was rotated internally and that there was little difference in rotation among the 3 angles. The present research conﬁrms that the tibia rotates internally when knee ﬂexes (screw-home motion). Other tissues around the knee did not signiﬁcantly affect tibial rotation at those 3 angles. The presence of pathologies can be identiﬁed in a variety of ways, but bilateral symmetry in the internal–external rotation of the tibia has not yet been reported. Previous studies that have used a clinical testing apparatus to look for side-to-side differences in the range of tibial rotation in the right and the left knees are nearly identical (Shoemaker and Markolf 1982; Lusin and Gajdosik, 1983; Samukawa et al., 2007). A signiﬁcant side-to-side correlation was found in the present research. The limited number of subjects in this study, and the low correlation coefﬁcient make it difﬁcult to determine whether side-to-side differences actually exist or not. Further investigations are needed before any conclusions can be made. Our purpose-built apparatus allowed both the knee and hip angles to be ﬁxed, something none of the previous studies mention. The length of the muscles around the knee, especially quadriceps, hamstrings, IT-band, and popliteus, have previously been found to change the knee kinematics depending on position, possibly affecting the angle of tibial rotation (Kwak et al., 2000; Ferrari et al., 2003). Therefore, we tried to limit these effects as much as possible by using a special apparatus to ﬁx the knee and hip joint ﬂexion angles.

3. Results The tibial rotation angles at 30, 60, and 90 degrees of knee ﬂexion are shown in Fig. 4. There was little difference in tibial rotation among the 3 angles of knee ﬂexion (p ¼ 0.400.92). The descriptive results of tibial rotation at 30, 60, and 90 degrees of ﬂexion are shown in Table 1. The tibia was rotated internally throughout the three angles of ﬂexion. The tibial rotation relationship between the right and the left legs is determined in Fig. 5 and there was a signiﬁcant correlation among the 3 angles between the right and the left sides (r ¼ 0.40, p < 0.03). 4. Discussion Although morphological studies on the shape and congruence of the tibia and femur have been conducted (Iwaki et al., 2000), the nature of tibial rotation in normal subjects is not well understood, with many researchers having used measurements obtained at close to terminal extension to see if the tibia rotates externally with knee extension, an effect known as screw-home motion (Hallen

Fig. 4. Tibial rotation at 30, 60, and 90 degrees of knee ﬂexion. Error bars indicate SD. No signiﬁcant differences of tibial rotation were found between the right and the left sides (p ¼ 0.40–0.92).

Fig. 5. The side-to-side comparisons of tibial rotation. A signiﬁcant correlation was found between the right and the left sides (r ¼ 0.40, p < 0.03).

M. Samukawa et al. / Manual Therapy 14 (2009) 712–715

There were several potential limitations to this study. First, the resolution of the MRI images was very limited. The imaging protocols adopted in this study (acquisition time was 60 s) were referred to from previous studies (Incavo et al., 2003; Lerner et al., 2003). Ideally, scan times should be as short as possible to minimize subject anxiety. As less complicated sequences require less scan time, they are considered to cause less discomfort. Second, bone shape and/or gender-speciﬁc differences may exist. This study did not examine either possibility, but the potential for such differences should be emphasized. Furthermore, as only Japanese subjects participated in the study, the possibility of ethnic differences cannot be disregarded. Third, a conventional MRI apparatus with limited chamber space was used in this study, and the extended position of the knee was not reported on because the hip angle was ﬁxed at 45 degrees of ﬂexion to minimize the effect of different muscle lengths. 5. Conclusion This study, in which new MRI techniques were developed to allow the measurement of tibial rotation angles in normal subjects at 30, 60, and 90 degrees of knee ﬂexion, contains several important new ﬁndings. Firstly, the tibia was shown to rotate internally relative to the femur in all 3 positions. The relative relationship between the tibia and the femur changes very little over the three ﬂexion angles. Secondly, tibial rotation angles between the right and the left sides were signiﬁcantly correlated (r ¼ 0.40, p < 0.03). References Brandsson S, Karlsson J, Sward L, Kartus J, Eriksson BT, Karrholm J. Kinematics and laxity of the knee joint after anterior cruciate ligament reconstruction. Pre-and postoperative radiostereometric studies. The American Journal of Sports Medicine 2002;30:361–7. Czerniecki JM, Lippert F, Olerud JE. A biomechanical evaluation of tibiofemoral rotation in anterior cruciate deﬁcient knees during walking and running. The American Journal of Sports Medicine 1988;16:327–31. Ferrari DA, Wilson DR, Hayes WC. The effect of the popliteus and quadriceps force on rotation of the knee. Clinical Orthopaedics and Related Research 2003;412:225–33. Georgoulis AD, Papadonikolakis A, Papageorgiou CD, Mitsou A, Stergiou N. Threedimensional tibiofemoral kinematics of the anterior cruciate ligament-deﬁcient

715

and reconstructed knee during walking. The American Journal of Sports Medicine 2003;31:75–9. Hallen LG, Lindahl O. ‘‘The screw-home’’ movement in the knee joint. Acta Orthopaedica Scandinavica 1966;37:97–106. Incavo SJ, Coughlin KM, Pappas C, Beynnon BD. Anatomic rotational relationships of the proximal tibia, distal femur, and patella. Implications for rotational alignment in total knee arthroplasty. The Journal of Arthroplasty 2003;18:643–8. Iwaki H, Pinskerova V, Freeman MA. Tibiofemoral movement 1: the shapes and relative movements of the femur and tibia in the unloaded cadaver knee. The Journal of Bone and Joint Surgery British Volume 2000;82:1189–95. Karrholm J, Elmqvist L, Selvik G, Hansson LI. Chronic anterolateral instability of the knee. The American Journal of Sports Medicine 1989;17:555–63. Kwak SD, Ahmad CS, Gardner TR, Grelsamer RP, Henry JH, Blankevoort L, et al. Hamstrings and iliotibial band forces affect knee kinematics and contact pattern. Journal of Orthopaedic Research 2000;18:101–8. Lerner AL, Tamez-Pena JG, Houck JR, Yao J, Harmon HL, Salo AD, et al. The use of sequential MR image sets for determining tibiofemoral motion: reliability of coordinate systems and accuracy of motion tracking algorithm. Journal of Biomechanical Engineering 2003;125:246–53. Logan MC, Williams A, Lavelle J, Gedroyc W, Freeman M. Tibiofemoral kinematics following successful anterior cruciate ligament reconstruction using dynamic multiple resonance imaging. The American Journal of Sports Medicine 2004;32:984–92. Lusin GF, Gajdosik RL. Reliability of instrumentation and measurement procedures for active internal and external rotation. The Journal of Orthopaedic and Sports Physical Therapy 1983;4:154–7. Magee DJ. Orthopedic physical assessment. 4th ed. Philadelphia: WB Saunders; 2002 [chapter 12], p 684. Matsumoto H. Mechanism of the pivot shift. The Journal of Bone and Joint Surgery British Volume 1990;72:816–21. Matsumoto H, Seedhom B, Suda Y, Otani T, Fujikawa K. Axis location of tibial rotation and its change with ﬂexion angle. Clinical Orthopaedics and Related Research 2000;371:178–82. Nagao N, Tachibana T, Mizuno K. The rotational angle in osteoarthritic knees. International Orthopaedics 1998;22:282–7. Piazza SJ, Cavanagh PR. Measurement of the screw-home motion of the knee is sensitive to errors in axial alignment. Journal of Biomechanics 2000;33: 1029–34. Samukawa M, Magee DJ, Katayose M. The effect of tibial rotation on the presence of instability in the anterior cruciate ligament deﬁcient knee. Journal of Sport Rehabilitation 2007;16:2–17. Sarji SA, Abdullah BJJ, Kumar G, Tan AH, Narayanan P. Failed magnetic resonance imaging examinations due to claustrophobia. Australasian Radiology 1998;42:293–5. Shoemaker SC, Markolf KL. In vivo rotatory knee stability. Ligamentous and muscular contributions. The Journal of Bone and Joint Surgery American Volume 1982;64:208–16. Takeda Y, Xerogeanes JW, Livesay GA, Fu FH, Woo SL. Biomechanical function of the human anterior cruciate ligament. Arthroscopy 1994;10:140–7. Todo S, Kadoya Y, Moilanen T, Kobayashi A, Yamano Y, Iwaki H, et al. Anteroposterior and rotational movement of femur during knee ﬂexion. Clinical Orthopaedics and Related Research 1999;362:162–70.

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 716 www.elsevier.com/math

Diary of events Congresso OMT e Does the eﬃcacy of Manual Therapy depend on its speciﬁcity? Parma, Italy, April 10the11th April 2010. For further information visit info@ﬁsiodynacom.it NOI International conference UK and Ireland Nottingham UK e April 15e17, 2010 Dublin IRELAND April 21e23, 2010 For further details www.noi2010.com Fax þ 3906 51882443 MUSCULOSKELETAL DISORDERS IN PRIMARY CARE RESEARCH CONGRESS 2010 Rotterdam 11e13 October 2010 For further information visit www.bsl.nl/primarycare

1356-689X/$ - see front matter doi:10.1016/S1356-689X(09)00164-7

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.: þ1 301 656 0220; Fax: þ1 301 654 0333; website: www.painpoints.com/seminars/index.html E-mail: [email protected] The Diary of events will soon be moving to http://www.elsevier. com/math (click on the journal news tab). 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.

Manual Therapy 14 (2009) e14

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Book Review Textbook of pediatric osteopathy edition: translation of handbuch der padiatrischen Osteopathie, Eva Moeckel, Mitha Noori (Eds.), 1st ed. . Churchill Livingstone Elsevier (2005). Published in English in 2008, Price £39.99, No of pages: 490, ISBN: 978-0443-06864-5 The publication of this book in English generated much anticipation and excitement in the osteopathic profession. This could be attributed to the growing interest and scope of pediatric osteopathic practice and the need for a comprehensive textbook in this ﬁeld. In producing this book, the editors, Eva Moeckel and Noori Mitha, have made a substantial contribution to osteopathy by drawing together a breadth and wealth of knowledge from twenty-eight highly qualiﬁed contributors with expertise in a variety of ﬁelds, all of which, impact on the growing baby and child, and are of interest to the osteopath. The book is divided into two sections. The ﬁrst section follows a chronological development from pregnancy, birth, to diagnosis and treatment of the newborn, baby, and child. Section two provides information on common conditions and related medical ﬁelds. Each of the twenty chapters provides a good balance of detailed information, photographs, drawings, references, and case studies as examples. The pediatric patient is always considered within the context of family, psychosocial and other health-related inﬂuences. The text avoids being prescriptive. It clearly aims to present knowledge and dynamic concepts that stimulate readers to think and consider the possibilities.

doi:10.1016/j.math.2009.07.010

From the outset, osteopathy in the cranial ﬁeld is the fundamental approach used. However, there is much of value to those practitioners who do not primarily work in this area. It draws on research published in German which would not be readily accessible to many and highlights the new and growing evidence base in pediatric health care. The key strength of this book lies in collaboration amongst colleagues, not limited to the osteopathic profession, throughout Europe and Australia. At times, individual contributions, which occur within a chapter, can present as somewhat disjointed and it can be a little difﬁcult to locate references. However, this is easily overcome once the reader becomes familiar with the format, and a summary of useful references, websites, and organizations is provided in an Appendices section. This is an important textbook for osteopaths, not only interested in pediatrics, as it explores human development based on osteopathic principles. It is a book that takes time to explore and will be a useful, stimulating, and ongoing resource for a variety of health practitioners who work with the very young.

Denise Cornall Osteopathic Medicine, Victoria University, PO Box 14428, Melbourne, Victoria 8001, Australia E-mail address: [email protected] 31 July 2009

Manual Therapy 14 (2009) e15

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Book Review The physiology of the joints: volume three: the cervical column, pelvic girdle and head, A.I. Kapandji (Ed.), 6th ed. . Churchill Livingstone (2008). Price £ 49,99, No. of pages 335, ISBN: 9780702029592 A.I. Kapandji has published three volumes on the Physiology of Joints. From the ﬁrst edition on they have been a great success and the work has since then been translated into six languages, including Japanese. A quick search through the new third volume on the spine indicates that it classically covers the pelvic girdle and lumbar, thoracic and cervical spine. However, this sixth edition has a completely new color lay-out and is printed on a larger paper size which makes the work more attractive and improves the clearness of the drawings and schemes. This new edition also includes new items and chapters. Besides some new drawings in most of the chapters the chapter on the pelvis includes a new section on the pelvic ﬂoor muscles and their functions. The chapter on the cervical spine is enriched with a section on the vertebral artery and an additional chapter on the head covers the temporomadibular joint and movements of the eye. Though the title may be somewhat misleading, this book clearly focuses on the biomechanics of body and joints thus introducing readers into the ﬁeld of arthrokinematics. The author starts from a human functional anatomical realism converting it into

doi:10.1016/j.math.2009.07.009

a mechanistic representation. Such simpliﬁcation may be didactically welcomed but risks oversimplifying things introducing inaccuracy: e.g. the representation of the cervical intervertebral disc by the general lumbar model ignores new insights in to the speciﬁc disc structure in the cervical spine. Such simpliﬁcation for the sake of didactics tends to take borderline aspects in the additional multi-link representation for the pelvic girdle and especially in the mechanical cut-out and assembly model of the spine. Although the author on other occasions tends to say that biomechanics is ‘‘fuzzy’’ – as there is no well-deﬁned geometry the book does not make explicit reference towards anatomical variability. A chapter or epilogue on aspects as asymmetry in morphology and function as well as on inter individual and time dependent aspects of anatomical variability might enrich the educational value of this work. However, if instructors take this in consideration these critical remarks do not defeat the tremendous value of this book in undergraduate and graduate physiotherapy and related education. Erik Cattrysse Vrije Universiteit Brussel, Brussels, Belgium E-mail address: [email protected] 30 July 2009

Manual Therapy 14 (2009) e16

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Book Review Neuropathic Physical Medicine: Therapy and Practice for Manual Therapists and Neuropaths, L. Chaitow. Elsevier-Churchill Livingstone (2008). Price £41,99, No. of pages 594, ISBN: 9780443103902 This multi-author book covers those aspects of neuropathic medicine which are not pharmacological or mind-body by nature. I did ask myself whether there is a physical medicine that is not neuropathic. The answer was, I thought, no. So perhaps this book is about physical medicine and, having been a professor of physical medicine once, I thought this is right up my street. As it turns out, I was wrong. The book embraces all sorts of concepts which a scientiﬁc mind is likely to think of as mumbo jumbo. Examples are ‘Constitutional considerations, (p. 89) and ‘detoxiﬁcation’ (p. 81). I know that most neuropaths would disagree with me, but I fail to see any good evidence for there validity. More disturbingly I found that the concept of scientiﬁc evidence was dealt with in a most peculiar and misleading way. This culminates in a graph which implies that data from clinical

doi:10.1016/j.math.2009.07.008

trials and meta-analyses are just as relevant as case reports and in vitro studies for evaluating the effectiveness of a therapy. I don’t think this is true and I am sure that most methodologist would agree with me. The last part of the book is devoted to treatment of speciﬁc conditions. I was unconvinced by much of this too. For example, the implication that one can effectively treat serious heart conditions with manipulation worries me. Similarly the notion that ‘‘irrespective of the named condition’’, patients can be helped by physical medicine simply does not tally with my scientiﬁc knowledge nor my clinical experience. I am sorry not to be able to strike a more positive note after reading this book. Edzard Ernst Complementary Medicine, Peninsula Medical School, Universities of Exeter and Plymouth, 25 Victoria Park Road, Exeter EX2 4NT, United Kingdom E-mail address: [email protected] 29 July 2009

Manual Therapy 14 (2009) e17

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Letter to the Editor

Comments on the article by Lucy C. Thomas et al. ‘‘Validity of the Doppler velocimeter in examination of vertebral artery blood ﬂow and its use in pre-manipulative screening of the neck’’, Manual Therapy 2009;14(5):544–9

We acknowledge the efforts made by the authors of this article in researching this important topic and would like to make the following comments: Thomas incorrectly used a criteria of a 50% increase or decrease in peak systolic velocities with the duplex scan (Thomas et al., 2008). Haynes data suggested that an 85% or more could indicate such, while there was no data to support any percentage less than this (Haynes, 2002). The duplex results in Haynes’s study for the peak systolic velocities of the 5 arteries that showed marked decreases at end rotation were: 10.1 cm/s, 22.3 cm/s, 21.6 cm/s, 20.2 cm/s, 22.9 cm/s, which indicates that when peak systolic velocities are in the low 20s or lower, velocimetry will likely indicate a positive ﬁnding (Haynes, 2002). It is a better way of deﬁning the criteria for positive duplex results. This is because a very large peak systolic velocity in the neutral position of say 70 cm/s could have more than a 50% reduction in velocities with rotation, yet still have velocities much higher than the low 20s. This means that the velocimeter examination could be normal, yet the 50% criteria of Thomas study set for the duplex results would actually result in a false positive for duplex. We think that while 2 h training may give adequate proﬁciency for testing in the neutral position, it will probably take several months or longer of routine use for many practitioners to achieve high skill levels examining during rotation. We ask the authors why they excluded one patient from the study, who was known by one of the physiotherapist examiners (Thomas et al., 2008). All of the patients had histories and presentations devoid of any clues as to the presence or absence of

DOI of original article: 10.1016/j.math.2008.08.007, 10.1016/j.math.2009.07.006. 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.07.007

rotational stenosis. So this blinding, presumably as part of what they called double blinding, would have failed to provide any beneﬁt to the study. Our last comment is concerning the gold standart used in this research: while validity trials exist for mid cervical VA duplex in the neutral position, there seem to be none supporting it during cervical rotation. This is likely why there is abnormal proportion of positive cases of vertebral artery stenosis in rotation in Thomas study. The authors needed to have used a duplex technique that had at least some evidence of validity.

References Haynes MJ. Vertebral arteries and cervical movement: Doppler ultrasound velocimetry for screening before manipulation. J Manipulative Physiol Ther 2002;25: 556–67. Thomas LC, Rivett DA, Bolton PS. Validity of the Doppler velocimeter in examination of vertebral artery blood ﬂow and its use in pre-manipulative screening of the neck. Man Ther 2009;14(5):544–9.

Vincent Karl* Fanco-European Chiropractic Society (SOFEC), 147 Avenue Louis Imbert, 83160 La Valette du Var, France Tel.: þ33 4 9461 3397. E-mail address: [email protected] 7 June 2009

Manual Therapy 14 (2009) e18

Contents lists available at ScienceDirect

Manual Therapy journal homepage: www.elsevier.com/math

Authors Reply to Letter to Editor

Comments in response to letter to the editor by Karl et al. Manual Therapy 2009;14(6):e17 We thank the authors of this letter for their comments. Our choice of the duplex criteria of a greater than 50% increase or decrease in peak systolic (PS) velocity between the neutral head position and contralateral rotation position was based on Haynes (2000) study (Haynes, 2000) in which he suggests marked changes in PS velocity could be considered abnormal (his 2000 study reported decreases between 35% and 68%) and Freed’s et al. (1998) suggestion that a greater than 50% change in PS velocity in the vertebral artery (VA) between neutral and contralateral head positions is indicative of stenosis. In our study the average PS values in the neutral head position for the left and right vertebral arteries were 33.6 and 35.16 cm/s respectively and we did not observe the high values which the authors suggested may have made a 50% change less signiﬁcant. Our study had one subject with a decrease in duplex determined PS velocity to 18.65 cm/s (52%) but all other subjects showed increases on contralateral rotation ranging from 65% to 176% of the neutral head position value (see Table 1). We agree that an initial high PS value in the neutral head position may make a 50% change (reduction) on rotation less likely to compromise brain perfusion sufﬁcient to impact on its function. This however, was not the case in our study as reported in Table 1. In contrast to the report in the letter, our practitioners in fact had 2 months of regular practice using the velocimeter in addition to the 2 hrs training. As we have already stated in our previous correspondence to the editor (Thomas et al., 2009) we agree that practitioners would require several months of practice to become proﬁcient and we suggest they may need many months or even years of practice to attain the level of proﬁciency achieved by Dr Haynes. This lengthy training may not be practical for many clinicians. The one patient’s data that was excluded was done so because he also served as one of the examiners. We used duplex as our determinate of VA blood ﬂow and discussed the reasons for doing this in our previous correspondence (Thomas et al., 2009). It was important in our view to examine the vertebral arteries in the same location with both duplex and velocimeter to obtain a true comparison. The gold standard for examination of VA blood ﬂow is magnetic resonance angiography (MRA) but this was not available for the current study. In conclusion we would like to suggest that ﬂuctuations in VA ﬂow on neck rotation, such as those we have measured, may in

DOI of original article: 10.1016/j.math.2008.08.007, 10.1016/j.math.2009.07.007. 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.07.006

Table 1 Peak systolic (PS) velocity values for the 5 patients with >50% change from the neutral to contralateral head rotation position. PS values neutral head position (cm/s)

PS values contralateral rotation head Percentage position (cm/s) change (%)

39 47.07 41.7 38.85 44.8 17.55

18.65 100.4 155.5 93.55 74.3 32.8

52 113 177 141 66 87

fact be a normal variation and have no relevance in terms of predicting risk of dissection with cervical manipulation. This would seem to be supported by the fact that none of the patients in either our or Dr Haynes’ (Haynes, 2000) studies showed any signs of vertebrobasilar insufﬁciency despite large alterations in ﬂow on neck rotation. We further suggest that the more important factors in risk prediction for cervical manipulation are the force of the manipulation and the underlying state of the arterial wall and further research should be pursued with this focus.

References Freed KS, Brown LB, Carroll MD. The extracranial cerebral vessels. In: Rumack CM, Wilson SR, Charboneau JW, editors. Diagnostic ultrasound, vol. 1. St. Louis, MO,: Mosby Year Book, Inc.; 1998. p. 885–919. Haynes M. Vertebral arteries and neck rotation: Doppler velocimeter and duplex results compared. Ultrasound in Medicine and Biology 2000;26(1):57–62. Thomas L, Rivett D, Bolton PS. Comments in response to letters to editor regarding article: validity of the Doppler velocimeter in examination of vertebral artery blood ﬂow and its use in pre-manipulative screening of the neck. Manual Therapy 2009;14(5):e6.

Lucy C. Thomas*, Darren A. Rivett, Philip S. Bolton The University of Newcastle, Faculty of Health, University Drive, Callaghan, Newcastle 2308 NSW, Australia Tel.: þ61 2 4921 8680; fax: þ61 2 4921 7902. E-mail address: [email protected] 20 July 2009