Manual Therapy Journal - Volume 10, Issue 4, Pages 239-304 (November 2005)

VOLUME 10 NUMBER 4 PAGES 239– 304 NOVEMBER 2005

Editors

International Advisory Board

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

K. Bennell (Victoria, Australia) K. Burton (Hudders¢eld, UK) B. Carstensen (Frederiksberg, Denmark) E. Cruz (Setubal Portugal) L. Danneels (Mar|¤ akerke, Belgium) S. Durrell (London, UK) S. Edmondston (Perth, Australia) J. Endresen (Flaktvei, Norway) L. Exelby (Biggleswade, UK) J. Greening (London, UK) C. J. Groen (Utrecht,The Netherlands) A. Gross (Hamilton, Canada) T. Hall (West Leederville, Australia) W. Hing (Auckland, New Zealand) M. Jones (Adelaide, Australia) S. King (Glamorgan, UK) B.W. Koes (Amsterdam,The Netherlands) D. Lawrence (Davenport, IA, USA) D. Lee (Delta, Canada) R. Lee (Hung Hom, Hong Kong) C. Liebenson (Los Angeles, CA, USA) L. Ma¡ey-Ward (Calgary, Canada) C. McCarthy (Manchester, UK) J. McConnell (Northbridge, Australia) S. Mercer (Queensland, Australia) E. Maheu (Quebec, Canada) D. Newham (London, UK) J. Ng (Hung Hom, Hong Kong) L. Ombregt (Kanegem-Tielt, Belgium) N. Osbourne (Bournemouth, UK) M. Paatelma (Jyvaskyla, Finland) N. Petty (Eastbourne, UK) A. Pool-Goudzwaard (The Netherlands) M. Pope (Aberdeen, UK) G. Rankin (London, UK) D. Reid (Auckland, New Zealand) M. Rocabado (Santiago, Chile) C. Shacklady (Manchester, UK) M. Shacklock (Adelaide, Australia) D. Shirley (Lidcombe, Australia) V. Smedmark (Stenhamra, Sweden) W. Smeets (Tongeren, Belgium) C. Snijders (Rotterdam,The Netherlands) M. Sterling (St Lucia, Australia) R. Soames (Leeds, UK) P. Spencer (Barnstaple, UK) P. Tehan (Victoria, Australia) M. Testa (Alassio, Italy) M. Uys (Tygerberg, South Africa) P. van Roy (Brussels, Belgium) B.Vicenzino (St Lucia, Australia) H.J.M.Von Piekartz (Wierden,The Netherlands) M.Wallin (Spanga, Sweden) M.Wessely(Paris, France) A.Wright (Perth, Australia) M. Zusman (Mount Lawley, Australia)

Gwendolen Jull PhD, MPhty, Grad Dip ManTher, FACP Department of Physiotherapy University of Queensland Brisbane QLD 4072, Australia Editorial Committee Karen Beeton MPhty, BSc(Hons), MCSP (Masterclass Editor) MACP ex o⁄cio member Department of Allied Health Professions—Physiotherapy University of Hertfordshire College Lane Hat¢eld AL10 9AB, UK Je¡rey D. Boyling MSc, BPhty, GradDipAdvManTher, MAPA, MCSP, MErgS (Case reports & Professional Issues Editor) Je¡rey Boyling Associates Broadway Chambers Hammersmith Broadway LondonW6 7AF, UK Tim McClune D.O. Spinal Research Unit. University of Hudders¢eld 30 Queen Street Hudders¢eld HD12SP, UK Darren A. Rivett PhD, MAppSc, MPhty, GradDip ManTher, BAppSc (Phty) (Case reports & Professional Issues Editor) Discipline of Physiotherapy Faculty of Health The University of Newcastle Callaghan, NSW 2308, Australia Kevin P. Singer PhD Centre for Musculoskeletal Studies Department of Surgery The University of Western Australia, Royal Perth Hospital Level 2, MRF Building, 50 rear, Murray Street Perth,WA 6000, Australia Raymond Swinkels MSc, PT, MT (Book Review editor) Ulenpas 80 5655 JD Eindoven The Netherlands

Visit the journal website at http://www.intl.elsevierhealth.com/journals/math doi:10.1016/S1356-689X(05)00132-3

ARTICLE IN PRESS

Manual Therapy 10 (2005) 239–241 www.elsevier.com/locate/math

Editorial

Why is the recent research regarding non-specific pain so non-specific? In line with recent guidelines, advocating the use of ‘‘non-specific’’ labels for various musculoskeletal conditions, such as CSAG (1994) and RCGP (1996) low back pain guidelines, researchers and funders have increasingly adopted these labels within research, specifically within randomised controlled trials (RCTs). The search for effective treatment for non-specific pain syndromes such as ‘‘non-specific low back pain (NSLBP)’’, ‘‘nonspecific neck pain’’ and ‘‘anterior knee pain’’ is ongoing and a number of RCTs have recently been published that have taken a pragmatic approach to evaluating physiotherapy practice in this area (Frost et al., 2004; Hay et al., 2005; Dziedzic et al., 2005). Doubtless, these trials will be included in future meta-analyses and hence incorporated within future clinical guidelines. These trials were undertaken with rigour and answered the questions they were set. However their clinical implications and conclusions need to be interpreted in light of the limitations in our understanding of this area. The RCT methodology is certainly a methodologically rigorous method of enquiry, in terms of reducing experimental bias, and is quite rightly the cornerstone of evidence based practice. However, it is the Manipulation Association of Chartered Physiotherapists’ (MACP) contention that the pragmatic RCT, evaluating the relative effectiveness of one treatment package against another, in non-specific pain syndromes, is fraught with underlying difficulties. It is unlikely that such a simple approach will advance our understanding regarding our effectiveness until some of these difficulties are addressed. Reassuringly, a number of innovative RCTs have been completed that have begun to address these issues (Fritz et al., 2003; Wand et al., 2004; UK BEAM Trial Team, 2004; Klaber Moffett et al., 2005). Whilst the RCT is an important tool in assessing effectiveness in the management of pain syndromes the process of scientific enquiry starts much earlier. Long before a RCT can be used a host of underlying factors must be considered. In this regard, it may be helpful to think of the management of pain syndromes in terms of a simple model. The model highlights the stages in a cyclical process and the underlying knowledge that links the stages. The model adopts the biopsychosocial 1356-689X/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2005.10.001

management paradigm and encourages biopsychosocial principles to be applied to diagnosis, outcome measurement and treatment. An appreciation of these concepts will facilitate an understanding of the knowledge that needs to be gained in order to ensure that further pragmatic RCTs will be of maximum clinical value. Fig. 1 shows a schematic of a cyclical model of management for pain syndromes. Starting at the bottom left of the triangle is the concept of ‘‘Measurement’’. This refers to the data we collect during our patient interactions, including social, psychological, clinical and laboratory data. Above this is ‘‘Profile’’. This is the process of assimilation of data and the formation of hypotheses or diagnoses. If we accept that this process will be multifaceted and include information pertaining to the biomedical, psychological and social profile of the patient we must accept that within this process will be a system of prioritisation or weighting. Thus, a patient is classified according to their predominant barrier to recovery but will have secondary and tertiary barriers to be considered. Thus, a ‘‘one label diagnosis’’ approach for most pain patients will not reflect the multiple factors in their ‘‘profile’’. To the right of the figure is ‘‘Intervention’’, reflecting the multiple factors included in our interaction with patients. If we are to accurately evaluate our practice we must consider that our intervention cannot be quantified solely by the treatment modality being provided. It is in the links between these processes where much work is needed to ensure that scientific evaluation of our management truly reflects the individuality of clinical practice and it is in these areas that we can see the strengths and weaknesses of recent publications. Currently, the link between our measurements and the diagnostic profile of our patients is unclear. The validity of our hypotheses and diagnostic labels are confounded by the lack of gold standards in the field. As yet, a patient’s biopsychosocial profile cannot be accurately classified into valid diagnoses in the majority of patients and consequently the diagnostic process has entailed excluding rare identifiable pathologies and labelling the resultant group as ‘‘non-specific’’. Thus, whilst there is a strong recognition that the non-specific group contain

ARTICLE IN PRESS Editorial / Manual Therapy 10 (2005) 239–241

240

BIOPSYCHOSOCIAL PROFILE LINKS

LINKS

VALID MEASURES

MATCHING PATIENT PROFILES WITH INTERVENTION

VALID DIAGNOSES

BIOPSYCHOSOCIAL MEASUREMENT

BIOPSYCHOSOCIAL INTERVENTION LINKS VALID OUTCOME MEASURES

Fig. 1. A model of management for pain syndromes.

specific homogenous sub-groups of patient profiles who are likely to be suited to specific management approaches, they are yet to be identified (McCarthy et al., 2004). Three recent pragmatic RCTs comparing varying degrees of advice and education or brief pain management packages with ‘‘traditional’’ physiotherapy intervention have shown no difference in effectiveness (Frost et al., 2004; Hay et al., 2005; Dziedzic et al., 2005). In contrast, another recent study in neck pain showed a small benefit in favour of ‘‘traditional’’ physiotherapy (Klaber Moffett et al., 2005) with the UK BEAM study showing a similar sized benefit over GP care for manual therapy (UK BEAM Trial Team, 2004). Although all powered correctly to show a pre-determined difference, the natural variability of patient response in the clinical setting and the inevitable heterogeneity of ‘‘nonspecific’’ groups may well be ‘‘washing out’’ the larger, positive effects observed in some by the smaller, or negative effects seen in others. Although this is the nature of research, it is important that we do not discount techniques or treatments that may be beneficial to some patients purely because we have failed to identify them. What stifles the clinical impact of these trials is the size of the effects observed. With a host of contradictory evidence and a strong theme of small effect sizes we need to establish if these effects truly reflect the intervention or merely a flaw in our methodology that is reducing the chance of observing larger effects. The link between diagnostic profile and the type of intervention suited to that profile is not clear. It is clear that establishing a match between patient profile and optimum intervention cannot be feasibly established within a RCT methodology. It is likely that having established effective matches between patient profiles and interventions in large cohort studies, the RCT will be crucial to their validation. Encouragingly, early attempts to link specific patient profiles to specific

treatments have shown a moderate degree of benefit over a non-specific multimodal approach (Fritz et al., 2003) and it is clear that this is an area that needs much investigation. Finally, the link between intervention and measurement needs to be strengthened. How should we be measuring our effectiveness? Does a simple measure of self-reported disability have the degree of content validity we need to encapsulate our effect? We are rightly encouraged to use a barrage of outcome measures to increase our measurement validity (Deyo et al., 1998) however our trials are powered to describe a change in only one outcome. It is likely that in order to fully evaluate our intervention we will need to accept that biomedical, psychological and social outcome measures will need to be evaluated and the relative importance of these measures established. Innovative trial design may be the way forward in the continuing hunt for effective management strategies. Three recent studies have taken an innovative approach to assessing the links between intervention and patient preference (Klaber Moffett et al., 2005); early versus delayed treatment (Wand et al., 2004) and the effect of the setting in which the treatment takes place (UK BEAM Trial Team, 2004). It is these types of innovations that will expand our knowledge of the processes involved in our interventions and facilitate our understanding of the outcome measurements we should be utilising. In conclusion, we are faced with a body of recent evidence that is contradictory regarding the effectiveness of ‘‘traditional’’ physiotherapy and manual therapy. In light of the small effect sizes observed, and the contradictory nature of the conclusions drawn, we must face the possibility that the current approach to evaluating our practice lacks the discriminatory ability necessary to truly represent complex clinical interactions. Alternatively, we can accept that traditional musculoskeletal physiotherapy provides no additional benefit for patients’ non-specific spinal pain. The MACP is of the opinion that before we accept the latter, effort is made to inform and underpin sophisticated, pragmatic RCTs. Until we establish valid links between the measures we use and the diagnoses we produce; until we match patient profiles with optimal treatments and until we routinely evaluate our treatments with valid outcomes we will continue to produce under whelming evidence. Future, well-informed pragmatic RCTs may, by necessity be some way off, but a delay will be worthwhile if the questions being asked are both clinically relevant and truly evaluate our clinical practice. There is no doubt that the recent studies have been valuable in the evaluation of our practise but it is vital for the development of our profession that we accept the messages from this body of work and address the important questions it has raised.

ARTICLE IN PRESS Editorial / Manual Therapy 10 (2005) 239–241

References CSAG. Clinical Standards Advisory Group Back Pain. London: HMSO; 1994. Deyo R, Beurskens A, Bombardier C, Croft P, Koes B, Malmivaara A, et al. Outcome measures for low back pain research. A proposal for standardised use. Spine 1998;23(18):2003–13. Dziedzic K, Hill J, Lewis M, Sim J, Daniels J, Hay EM. Effectiveness of manual therapy or pulsed shortwave diathermy in addition to advice and exercise for neck disorders: a pragmatic randomized controlled trial in physical therapy clinics. Arthritis and Rheumatism 2005;53(2):214–22. Fritz JM, Delitto A, Erhard RE. Comparison of classification-based physical therapy with therapy based on clinical practice guidelines for patients with acute low back pain: a randomized clinical trial. Spine 2003;28(13):1363–71. Frost H, Lamb SE, Doll HA, Carver PT, Stewart-Brown S. Randomised controlled trial of physiotherapy compared with advice for low back pain. British Medical Journal 2004;329(7468):708. Hay EM, Mullis R, Lewis M, Vohora K, Main CJ, Watson P, et al. Comparison of physical treatments versus a brief pain-management programme for back pain in primary care: a randomised clinical trial in physiotherapy practice. The Lancet 2005;365(9476):2024–30. Klaber Moffett JA, Jackson DA, Richmond S, Hahn S, Coulton S, Farrin A, et al. Randomised trial of a brief physiotherapy intervention compared with usual physiotherapy for neck pain patients: outcomes and patients’ preference. British Medical Journal 2005;330(7482):75. McCarthy CJ, Arnall FA, Strimpakos N, Freemont AJ, Oldham JA. The bio-psycho-social classification of non-specific low back

241

pain: A systematic review. Physical Therapy Reviews 2004;9: 17–30. Royal College of General Practitioners. 1996. Clinical guidelines for the management of acute low back pain. RCGP. UK BEAM Trial Team. United Kingdom back pain exercise and manipulation (UK BEAM) randomised trial: effectiveness of physical treatments for back pain in primary care. British Medical Journal 2004;329(7479):1377. Wand BM, Bird C, McAuley JH, Dore CJ, MacDowell M, De Souza LH. Early intervention for the management of acute low back pain: a single-blind randomized controlled trial of biopsychosocial education, manual therapy, and exercise. Spine 2004;29(21): 2350–6.

Christopher J. McCarthy The Centre for Rehabilitation Science, University of Manchester, Manchester Royal Infirmary, Oxford Road, Manchester, M13 9WL, UK E-mail address: [email protected] Mindy C. Cairns School of Paramedic Sciences, Physiotherapy and Radiography, Faculty of Health and Human Sciences, University of Hertfordshire, College Lane, Hatfield, Hertfordshire, AL10 9AB E-mail address: [email protected]

ARTICLE IN PRESS

Manual Therapy 10 (2005) 242–255 www.elsevier.com/locate/math

Masterclass

Diagnosis and classification of chronic low back pain disorders: Maladaptive movement and motor control impairments as underlying mechanism Peter O’Sullivana,b, a

Body-logic Physiotherapy, 146 Salvado Rd, Wembley, WA 6014, Australia School of Physiotherapy, Curtin University of Technology, Perth, Western Australia

b

Received 3 April 2005; accepted 9 July 2005

Abstract Low back pain (LBP) is a very common but largely self-limiting condition. The problem arises however, when LBP disorders do not resolve beyond normal expected tissue healing time and become chronic. Eighty five percent of chronic low back pain (CLBP) disorders have no known diagnosis leading to a classification of ‘non-specific CLBP’ that leaves a diagnostic and management vacuum. Even when a specific radiological diagnosis is reached the underlying pain mechanism cannot always be assumed. It is now widely accepted that CLBP disorders are multi-factorial in nature. However the presence and dominance of the patho-anatomical, physical, neuro-physiological, psychological and social factors that can influence the disorder is different for each individual. Classification of CLBP pain disorders into sub-groups, based on the mechanism underlying the disorder, is considered critical to ensure appropriate management. It is proposed that three broad sub-groups of CLBP disorders exist. The first group of disorders present where underlying pathological processes drive the pain, and the patients’ motor responses in the disorder are adaptive. A second group of disorders present where psychological and/or social factors represent the primary mechanism underlying the disorder that centrally drives pain, and where the patient’s coping and motor control strategies are mal-adaptive in nature. Finally it is proposed that there is a large group of CLBP disorders where patients present with either movement impairments (characterized by pain avoidance behaviour) or control impairments (characterized by pain provocation behaviour). These pain disorders are predominantly mechanically induced and patients typically present with mal-adaptive primary physical and secondary cognitive compensations for their disorders that become a mechanism for ongoing pain. These subjects present either with an excess or deficit in spinal stability, which underlies their pain disorder. For this group, physiotherapy interventions that are specifically directed and classification based, have the potential to impact on both the physical and cognitive drivers of pain leading to resolution of the disorder. Two case studies highlight the different mechanisms involved in patients with movement and control impairment disorder outlining distinct treatment approaches involved for management. Although growing evidence exists to support this approach, further research is required to fully validate it. r 2005 Elsevier Ltd. All rights reserved.

1. The need to classify CLBP disorders Low back pain (LBP) is common with up to 80% of people reporting LBP over their life time (Dillingham, 1995). The majority of acute LBP disorders resolve Corresponding author at: Body-logic Physiotherapy, 146 Salvado Rd, Wembley, WA 6014, Australia. E-mail address: [email protected].

1356-689X/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2005.07.001

within a 4 week period although recurrence is common (Croft et al., 1998). A small number of disorders (10–40%) become chronic and represent a major cost burden for society (Dillingham, 1995; Croft et al., 1998). In spite of the small number of pathological conditions that can give rise to back pain, most cases (85%) are classified as ‘‘non-specific’’ because a definitive diagnosis cannot be achieved by current radiological methods (Dillingham, 1995). Even when a specific diagnosis is

ARTICLE IN PRESS P. O’Sullivan / Manual Therapy 10 (2005) 242–255

made, the validity of the diagnosis can often be questioned. This leaves a diagnostic and management vacuum (Leboeuf-Yde et al., 1997). This situation commonly results in the ‘‘signs and symptoms’’ of the disorder being treated without consideration for the underlying basis or mechanism for the pain disorder. It is well recognized that the classification of chronic low back pain (CLBP) disorders into homogenous groups, and the application of specific interventions tailored for these groups is likely to enhance treatment efficacy (Leboeuf-Yde et al., 1997). It is also well established that LBP is a multi-dimensional problem (Borkan et al., 2002; McCarthy et al., 2004). These dimensions consist of pathoanatomical, neurophysiological, physical and psychosocial factors (Waddell, 2004). To date, the majority of studies that relate to the classification of back pain have focused only on a single dimension of the problem, rather than consideration being given to all dimensions of LBP (Ford et al., 2003). For a classification system to be clinically useful it should be based on identifying the underlying mechanism(s) driving the disorder, in order to guide targeted interventions, which in turn should predict the outcome of the disorder.

2. Models for the diagnosis and classification of CLBP Current approaches or models used for the diagnosis and classification of CLBP have tended to only focus on a single dimension of the disorder, limiting their validity (Ford et al., 2003). The following overview is not designed to be exhaustive, but highlights to the clinician the strengths and weaknesses of these different approaches. 2.1. Patho-anatomical model The traditional medical approach to diagnosis of CLBP has been from a pathoanatomical perspective (Nachemson, 1999). The findings of intervertebral disc (IVD) and facet joint degeneration, annular tears, IVD prolapse, spondylolisthesis, foraminal and spinal stenosis with associated nerve pain are commonly assumed to be related to back pain (and in some cases associated neurogenic pain), with interventions provided on the basis of this assumption (Nachemson, 1999). However, the problem with pathoanatomical diagnoses for CLBP is that many ‘abnormal’ findings are also commonly observed in the pain free population and pathoanatomical findings correlate poorly with levels of pain and disability (Nachemson, 1999). Frequently, little consideration is given to the confounding impact of psycho-social, neuro-physiological and physical factors that may co-exist and contribute to the underlying basis

243

of these disorders (Nachemson, 1999). Because of this, even when a specific pathoanatomical diagnosis can been made, there is still a need to classify the disorder based on the mechanism(s) that drive the pain disorder to ensure appropriate management. 2.2. Peripheral pain generator model More recently there has been a focus on the identification of the painful structure (peripheral pain generator) based on the patient’s history, area of pain, clinical examination findings and diagnostic blocks (Donatelli and Wooden, 1989; Laslett and Williams, 1994; Schwarzer et al., 1994; Bogduk, 1995; Bogduk, 2004). This has led to studies that have reported that the majority of chronic back pain originates in the IVD (45%), with a smaller number of subjects with facet joint (20%) and sacro-iliac joint (15%) pain (Bogduk, 1995). These studies have led to diagnostic and therapeutic procedures to identify, block or denervate the nociceptive source (Bogduk, 2004). The major limitation of this treatment model is that it treats the symptom of pain without consideration for the underlying mechanism or cause of the pain generation, and these approaches frequently only result in short term pain relief and lack broad therapeutic utility (Nachemson, 1999). 2.3. Neuro-physiological model An increased focus on the study of the nervous system and its involvement in pain disorders has documented complex biochemical and neuro-modulation changes at a peripheral, as well as at spinal cord and cortical levels (Flor and Turk, 1984; Flor et al., 1997; Moseley, 2003; Wright and Zusman, 2004). This has highlighted that pain can be generated and maintained at a peripheral level, as well as centrally at both spinal cord and cortical levels. Central sensitisation of pain which is manifest in most CLBP disorders (to varying degrees) can occur secondary to sustained peripheral noniceptive input resulting in changes at spinal cord and cortical levels (Zusman, 2002). This can be both amplified and inhibited by fore-brain descending input (see psychosocial section) (Zusman, 2002). As well as this there is growing evidence that the nervous system undergoes changes to its cortical mapping and possesses a pain ‘memory’ which may leave it pre-sensitized to the exacerbation and recurrence of pain (Zusman, 2002). This new knowledge has lead to an increased focus on medical interventions to inhibit both peripheral and central processing of pain (Bogduk, 2004), as well as psychological and cognitive interventions to reduce the forebrain facilitation of pain (Woby et al., 2004).

ARTICLE IN PRESS 244

P. O’Sullivan / Manual Therapy 10 (2005) 242–255

2.4. Psychosocial model The focus on the nervous systems’ role in pain modulation has coincided with increasing research investigating the impact of psychological and social factors on the modulation of pain and in particular, their capacity to increase the central nervous system mediated drive of pain via the forebrain (Linton, 2000; Zusman, 2002; Waddell, 2004). Mal-adaptive coping strategies such as negative thinking, pathological fear and abnormal anxiety regarding pain, avoidant behaviour, catastrophizing and hyper-vigilance have been shown to be associated with high levels of pain, disability and muscle guarding (Frymoyer et al., 1985; Main and Watson, 1996; Nachemson, 1999; Linton, 2000). Social factors such as the compensation system, work place disputes, work and family tensions and cultural issues affecting beliefs reinforce the psychological factors that can increase the central drive of pain (Nachemson, 1999). Despite this advanced knowledge there is debate regarding the relative contribution of these factors to pain disorders and whether these factors predispose, or are as a result of a pain disorder. In contrast positive factors such as adaptive coping strategies, appropriate pacing and distraction (reduced hypervigilance) can have a descending inhibitory effect on pain via the forebrain (Zusman, 2002). Certainly there is evidence that cognitive behavioural interventions are effective in reducing disability in specific groups with non-specific CLBP (Woby et al., 2004), however there appears to be a growing trend within physiotherapy to classify most patients with non-specific CLBP as primarily psychosocial driven due to a lack of an alternative diagnosis. Although all CLBP disorders have psychological and social impact with associated cognitive issues related to the disorder, it appears that only a small sub-group exist where these factors become the dominant or primary pathological basis for the disorder. 2.5. Mechanical loading model Both high and low levels of physical activity are reported to be risk factors for LBP while moderate levels of activity appear protective (Newcomer and Sinaki, 1996; Balague et al., 1999). Mechanical factors are usually reported to be associated with the initial development of LBP and are frequently reported to contribute to the recurrence of LBP and the exacerbation of CLBP. These factors include; sustained low load postures and movements (such as sitting, standing, bending and twisting), exposure to whole body vibration, high loading tasks (such as repeated lifting and bending), as well as sudden and repeated spinal loading in sports specific and manual work situations (Pope and Hansen, 1992; Adams et al., 1999; Nachemson, 1999; Abenhaim et al., 2000; McGill, 2004). These different

mechanical exposures are also influenced by ergonomic and environmental factors (McGill, 2004), such as seating design, lifting technique, work place design and sporting equipment. Individual physical factors such as where in its range a spinal articulation is loaded (neutral zone vs. elastic zone), reduced trunk muscle strength and endurance, impaired flexibility, ligamentous laxity and motor control dysfunction as well as anthropometric considerations have also been reported to be associated with LBP (Adams et al., 1999; Abenhaim et al., 2000; McGill, 2004; Dankaerts et al., 2005b; O’Sullivan et al., 2005). Although little direct evidence supports the efficacy of ergonomic interventions for the management of LBP, there is little doubt that physical factors such as sustained end range spinal loading, lifting with flexion and rotation, exposure to vibration and specific sporting activities involving cyclical end range loading of the spine (especially combined with rotation) do negatively impact on the musculo-skeletal system and have the potential to cause ongoing peripheral nociceptor sensitization (Adams et al. 1999; Nachemson, 1999; Abenhaim et al., 2000; Burnett et al., 2004; McGill, 2004). 2.6. Signs and symptoms model The area and nature of pain, impairments in spinal movement and function, changes in segmental spinal mobility (hyper and hypo), as well as pain responses to mechanical stress (provocation tests) and movement (peripheralisation and centralisation of pain with repeated movement) have formed the basis for classifying LBP disorders (McKenzie, 1981; Maitland, 1986; McKenzie, 2000). These approaches are based on biomechanical and pathoanatomical models and have lead to the assessment and treatment of signs and symptoms associated with CLBP (McKenzie, 1981; Maitland, 1986; McKenzie, 2000). Evidence for the efficacy of these approaches for the management of CLBP disorders remains limited (Maher et al., 1999; Abenhaim et al., 2000; Bogduk, 2004). This may in part be due to the limitations of the research design for some of these studies, as well as a neglect to account for the complex biopsychosocial nature of chronic pain disorders (Elvey and O’Sullivan, 2004). 2.7. Motor control model There has been an increased focus on the management of CLBP from a motor control perspective (Richardson and Jull, 1995; O’Sullivan, 1997, 2000; Sahrmann, 2001). While it is well recognized that movement and motor control impairments exist with CLBP disorders, they are highly variable and their presence does not establish cause and effect. Movement and motor control impairments are known to occur secondary to the presence of pain (Hodges and Moseley, 2003; Van-Dieen et al.,

ARTICLE IN PRESS P. O’Sullivan / Manual Therapy 10 (2005) 242–255

2003). Pathological processes such as neurogenic and radicular pain, neuropathic and centrally mediated pain and inflammatory disorders result in adaptive or protective altered motor behaviour in response to pain (Hall and Elvey, 1999; Elvey and O’Sullivan, 2004). Psychological processes such as stress, fear, anxiety, depression, hysteria, and somatisation are also known to disrupt motor behaviour (Frymoyer et al., 1985; Hodges and Moseley, 2003). Attempts to ‘‘normalize’’ movement or motor control impairments or treat dysfunction in the spinal muscles in many of these disorders would be inappropriate and ineffective due to the nonmechanical basis of these disorders. There is however growing evidence that CLBP disorders do exist where mal-adaptive movement and motor control impairments appear to result in ongoing abnormal tissue loading and mechanically provoked pain (Burnett et al., 2004; Dankaerts et al., 2005b; O’Sullivan et al., 2005). Following an acute episode of low back pain (when tissue healing would have normally occurred), ongoing mal-adaptive motor control behaviour provides a basis for ongoing peripherally driven nociceptor sensitisation leading to a chronic pain state. These disorders are amenable to tailored physiotherapy interventions directed at their specific physical and cognitive impairments (O’Sullivan et al., 1997a–c; Stuge et al., 2004). 2.8. Biopsychosocial model What is clear from the scientific literature and clinical practice, is that a multi-dimensional approach to dealing with CLBP based on a biopsychosocial model is required (Elvey and O’Sullivan, 2004; McCarthy et al., 2004; Waddell, 2004). The relative contribution of the different dimensions and their dominance associated with a CLBP disorder will differ for each patient. The role of the treating clinician is to consider all dimensions of the disorder based on an interview, thorough physical examination (assessing all aspects of the neuromusculosketetal system) combined with review of radiological imaging, medical tests and screening questionnaires (Elvey and O’Sullivan, 2004; O’Sullivan, 2004; Waddell, 2004) (Fig. 1). A clinical reasoning process allows determination of which factors are dominant in the disorder and whether the patient has adapted to the disorder in a positive or negative manner. Consideration of all the factors outlined allows for a diagnosis and mechanism based classification guiding management of the disorder (Elvey and O’Sullivan, 2004) (Fig. 1).

3. Diagnosis and classification of back pain The Quebec task force classification system provides a logical approach for the diagnosis and classification of

245

LBP disorders within a biopsychosocial framework (Spitzer, 1987; Abenhaim et al., 2000; Waddell, 2004). Under this framework red flags are considered in a diagnostic triage. The patient is screened for yellow flags or non-organic features suggestive of psychological and/ or social factors dominating in the disorder. Under this classification system, disorders can be diagnosed as specific (especially nerve root pain) or non-specific, and staged (acute, sub-acute and chronic). 3.1. Diagnosis: specific and non-specific CLBP disorders Specific pathoanatomical diagnoses, although critical for the understanding of many disorders, require further classification. For example, a diagnosis of lumbar spine stenosis (central or foraminal/lateral—chronic stage) may be associated with an adaptive (protective) motor response associated with a functional reduction of the lumbar lordosis with associated lumbar multifidus inhibition, to unload sensitized neural tissue. In this case attempts to normalize the motor control impairments would result in exacerbation and deterioration of the disorder. On the other hand the same diagnosis may be associated with a mal-adaptive motor response, represented by a functional increase in lumbar lordosis with associated back muscle guarding, resulting in further neural compromise and direct aggravation of the disorder. In this case normalising the motor control impairments (to functionally reduce the lumbar lordosis) would be indicated and effective. This proposed classification (into adaptive/mal-adaptive motor control responses) directly influences whether the patients’ specific disorder is amenable for physiotherapy management that is aimed at normalising the motor control impairments or not. Alternatively, this diagnosis may be associated with a dominance of psychosocial factors and associated dominant central nervous system sensitisation, compromising the potential success of both conservative physiotherapy and surgical interventions. In this case the same specific diagnosis may present with a different classification, reflecting a different underlying pain mechanism and therefore indicating a different intervention (Elvey and O’Sullivan, 2004). Eighty-five percent of CLBP disorders do not have a specific diagnosis (Dillingham, 1995). These disorders are labelled ‘non-specific CLBP’ disorders and represent a large group of ‘tissue strains’ and ‘sprains’ that have not resolved beyond normal tissue healing time (Abenhaim et al., 2000). This group has been broadly classified based on the area of pain and defined as somatic referred or radicular in nature (Abenhaim et al., 2000). However this diagnostic/classification system is of limited clinical value as it does not identify the underlying mechanism driving the pain disorder, and consequently there is no clear direction for specific management (Padfield and Butler, 2002).

ARTICLE IN PRESS P. O’Sullivan / Manual Therapy 10 (2005) 242–255

246

Social factors - relationships – family, friends, work - work structure - medical advice - support structures - compensation – emotional, financial - cultural factors - socio-economic factors

Genetic factors

Patho-anatomical factors

- potentially influencing all other domains

- structural pathology - identify peripheral pain generator (IVD / Zt joint / SI Jt / neural tissue / myo-fascial / connective tissue)

Physical factors

Psychological factors - personality type - beliefs & attitudes - hypervigilance - coping strategies – confronter vs avoider - pacing - emotions - fear / anxiety / depression / anger - iIlness behaviour

Pain Neuro-physiological factors - peripheral sensitisation - central sensitisation - sympathetic nervous system activity - somatic complaints

- ‘passive’ structure competence (hypermobility) - developmental factors - mechanism of injury - disorder history and stage - area of pain – local / generalised / referred - pain behaviour – directional / centralisation - mechanical vs non-mechanical provocation - articular mobility - neural tissue provocation testing - neurological examination - motor control / myo-fascial considerations - adaptive vs mal-adaptive motor response - movement impairments (directional) - motor control impairments (directional) - activity levels / conditioning / strength / muscle endurance - work / home environment / lifestyle - ergonomic factors

Fig. 1. Factors that need consideration within a biopsychosocial framework, for the diagnosis and classification of CLBP disorders.

3.2. Classification of CLBP Due to the shortcomings of the current models, it is clear that both specific and non-specific CLBP disorders require further classification based on a biopsychosocial construct. There are a number of key clinical indicators regarding pain area and behaviour, which provide an important insight into the different mechanisms underlying and driving a pain disorder, allowing classification to be made. Considered simplistically, the presence of localized and anatomically defined pain associated with specific and consistent mechanical aggravating and easing factors, suggest that physical/mechanical factors are likely to dominate the disorder resulting in a primary peripheral nociceptive drive. Correlation between clinical examination and pathoanatomical findings is critical to determine their significance and relationship to the disorder. If pain is constant, non-remitting, widespread and is not greatly influenced by mechanical factors (or minor mechanical factors result in an exaggerated and disproportionate pain response), then inflammatory or centrally driven neurophysiological factors (such as altered central pain processing) are likely to dominate the disorder. High levels of anxiety, hypervigilance, fear and emotional stress presenting as primary aggravating or precipitating factors in the disorder, highlight the influence of psychological and in some cases social factors indicating the dominant forebrain drive of pain in a disorder (Linton, 2000). Understanding a patient’s social circumstances, work environment, lifestyle factors and beliefs regarding their disorder is also critical (Waddell, 2004). Whether the patient has active or

passive coping strategies in managing their disorder, and whether they pace themselves is important in understanding their capacity to actively manage their pain (Bergstrom et al., 2001). In reality most disorders will be associated with a combination of these factors, and the role of the clinician is to consider the balance and dominance of them in the disorder (Fig. 1). It is proposed that there are three broad sub-groups of patients that present with disabling CLBP associated with movement and control impairments (Fig. 4). (1) The first sub-group is represented by disorders where high levels of pain and disability, as well as movement and/or control impairments are secondary and adaptive to an underlying pathological process. These include red flag disorders, specific pathoanatomical disorders in some circumstances (such as IVD prolapse, spinal and foraminal stenosis with associated radicular pain 7 neurological deficits, internal disc disruption with associated inflammatory pain, ‘unstable’ grade 2–4 spondylolisthesis), inflammatory pain disorders, neuropathic and centrally or sympathetically mediated pain disorders. These patients present with antalgic movement patterns and altered motor control that is driven directly by the pain disorder. The therapist will quickly determine this as attempts to ‘normalize’ these motor control and movement impairments results in exacerbation or non-resolution of the disorder, as these impairments are adaptive and driven by pathological processes. If the pathological process resolves with time or secondary to specifically targeted interventions (i.e. appropriate medical and/or surgical management when indicated), the signs and symptoms (e.g. motor

ARTICLE IN PRESS P. O’Sullivan / Manual Therapy 10 (2005) 242–255

control and movement impairments) related to the disorder resolve. Specifically targeted therapy management may be indicated for some of these disorders in conjunction with other primary medical interventions with full knowledge of the non-mechanical underlying basis of the disorder (Elvey and O’Sullivan, 2004). These disorders represent a small but severely disabled group within the CLBP population. (2) A second small sub-group exists where the drive of the pain disorder is from the forebrain, secondary to a dominance of psychological and/or social (non-organic) factors. Although psychological and social impact occurs with all chronic disabling pain disorders, it appears that for a small group of patients it represents the dominant central drive of their disorder. This results in high levels of disability, altered central pain processing, amplified non-remitting pain, and resultant disordered movement and motor control impairments. These disorders commonly present with dominant psycho-social features, including pathological anxiety, fear, anger, depression, negative beliefs, un-resolved emotional issues, poor coping strategies (lack of pacing resulting in pain provocation or excessive avoidance of activity as means of controlling pain) as well as negative social and inter-personal circumstances (Linton, 2000; Bergstrom et al., 2001; Waddell, 2004). These psychological and social stresses present as dominant coexisting, precipitating and primary aggravating factors for the disorder (Linton, 2000). The key feature of these disorders is the absence of an organic basis to the disorder, and lack of clear and consistent mechanical provocation or relieving patterns (absence of peripheral nociceptor drive). When mechanical factors are provocative they are inconsistent and tend to result in abnormal and disproportionate pain, disability and emotional responses. These patients commonly present with high levels of dependence on strong analgesic medication and passive forms of health care provision by multiple practitioners, even though they report a poor response to these interventions (Waddell, 2004). It is important to note that a therapist should not arrive at this classification without consultation and confirmation by either a treating clinical psychologist or psychiatrist. In this sub-group, attempts to simply treat the ‘signs and symptoms’ of the disorder directly (e.g. movement and control impairments) does not result in their resolution, as the underlying mechanism driving the pain is not addressed. Management of these disorders requires multi-disciplinary management with a primary focus on cognitive behavioural therapy (Bergstrom et al., 2001) and psychiatric management. Physiotherapy management can play a specialized role in reinforcing graded functional recovery while reducing the focus on pain, however it cannot be seen as the primary treatment for these disorders (Elvey and O’Sullivan, 2004).

247

(3) It is proposed that a large third sub-group exists where mal-adaptive movement or control impairments and associated faulty coping strategies result in chronic abnormal tissue loading (associated with either excessive or reduced spinal stability), pain, disability and distress. This group is classified on the basis that the ‘movement’ impairments (characterized by pain avoidance behaviour) or ‘control’ impairments (characterized by pain provocation behaviour) act as the underlying mechanism that drives the CLBP state. Normalisation of the movement or control impairments based on a cognitive behavioural approach results in resolution and/or control of these disorders. Disorders with a ‘movement’ and ‘control’ impairment classification present commonly in clinical practice, and they appear to have different underlying pain mechanisms from each other and therefore their management is distinctly different (Figs. 2 and 3). These disorders may present as specific (associated with a pathoanatomical diagnosis) or nonspecific CLBP disorders, and are commonly associated with psychological, social, neurophysiological (central sensitisation) factors, that may contribute to but do not dominate or drive the disorder. The classification of these disorders leaves them amenable to therapy intervention directed at the primary physical (movement and control) impairments while addressing the secondary cognitive aspects of the disorder (see Fig. 4). 3.2.1. Movement impairment classification CLBP disorders classified as ‘movement impairment’ present with a painful loss or impairment of normal (active and passive) physiological movement in one or more directions (Figs. 2, 3 and 5a). These disorders are associated with abnormally high levels of muscle guarding and co-contraction of lumbo-pelvic muscles when moving into the painful and impaired range. This appears to be driven by an exaggerated withdrawal motor response to pain. This leads to high levels of compressive loading across articulations, movement restriction and rigidity (excessive stability), resulting in a mechanism for tissue strain and ongoing peripheral nociceptor sensitisation. These patients are usually acutely aware of their pain and are fearful of moving into the painful movement direction as they perceive that pain provocation is damaging. The fear of movement appears to develop from the patients’ initial experience of severe acute pain, as well as their beliefs (reinforced by sympathetic family members and treatment providers) that pain is harmful. Movement related fear, hyper-vigilance and anxiety associated with the pain reinforces the faulty cognitive coping strategies and beliefs, further amplifying the pain centrally and reinforcing their muscle guarding. This represents a mal-adaptive response to the pain disorder, as the compensations for the pain in turn becomes the mechanism that drives the disorder. These disorders

ARTICLE IN PRESS 248

P. O’Sullivan / Manual Therapy 10 (2005) 242–255

(A) Movement impairment classification Nature and mechanism of pain: Localised pain +/- referral Severe pain of rapid onset Movement impairment in direction of pain Hyper-awareness of pain Exaggerated reflex withdrawal motor response Muscle guarding and abnormal tissue loading (↑spinal stability) Avoidance of movement into painful range Disability Directional (flexion, extension, rotation, lateral shift, loading) Multi-directional Result: Peripheral pain sensitisation

(B) Control impairment classification Nature and mechanism of pain: Localised pain +/- referral Gradual onset of pain from repeated or sustained strain No impaired movement in direction of pain Lack of awareness of pain triggers Poor lumbo-pelvic position sense Absence of reflex withdrawal motor response Ongoing tissue strain (↑or↓ spinal stability) Provocation into painful range Avoidance of painful activity Disability Directional (flexion, extension, rotation, lateral shift, loading) Multi-directional Result: Peripheral pain sensitisation

Anxiety related to movement pain Fear avoidance when moving in direction of pain (pathological) Hyper-vigilence Belief that pain is damaging (pathological)

Anxiety related to chronic disabling pain Fear of activity (non-pathological) Lack of control and awareness of disorder Belief that activity is damaging (non-pathological)

Result: Central pain sensitisation

Result: Central pain sensitisation

Normalisation of movement impairment leads to resolution / control of disorder

Normalisation of control impairment leads to resolution / control of disorder

Fig. 2. The nature and mechanism associated with mal-adaptive motor control disorders with: (A) Movement impairment classification and (B) control impairment classification (italics represent common features of the disorders / normal text highlights differences between the disorders).

may present in a directional manner (flexion, extension, side bending and rotational impairments) as well as combinations of these movements (multi-directional movement impairments). Management of this patient sub-group is directed at both the dominant physical and associated cognitive factors that underlie the disorder. The aim is first to educate the patient that their pain is not damaging and they have developed faulty compensations to their pain, which now act to maintain their disorder. Restoration of the painful impaired movement is critical for the resolution of the disorder. The aim of the intervention is to desensitize the nervous system by restoring normal movement, reducing the fear of movement into pain and associated muscle guarding. This is facilitated by graded movement exposure into the painful range in a relaxed and normal manner based on the individual patient presentation. The cognitive strategies of reducing fear and changing beliefs regarding pain is augmented by manual therapy ‘treatment’ to restore the movement impairment (articular mobilisation/manipulation and soft tissue techniques). This is combined with active ‘management’ approaches directed to restore the movement impairment (muscle relaxation, breathing control, postural adjustments, graded movement exposure ex-

ercises, cardio-vascular exercise and most importantly graded functional restoration to normalize motor control). As the movement impairment and associated movement-based fear reduces, so too does the disability and pain related to the disorder. Stabilising exercise programs and treatment approaches that focus on pain and reinforce the avoidance behaviour usually exacerbate these disorders and are contra-indicated. 3.2.1.1. Case study 1. A 28-year-old woman reported a 3 year history of disabling non-specific CLBP (central lower lumbar) that had developed following a lifting injury while working as a nurse. She was placed off work for three weeks and was told by her physiotherapist that she had injured her disc, should do ‘McKenzie extension exercises’, avoid flexion and maintain her lumbar lordosis at all times. She reported becoming disabled with pain and very fearful of bending her back which she avoided doing from that time. Her treatment history consisted of McKenzie extension exercises, Pilates, stabilisation training (with a focus on pelvic floor, transverse abdominal wall and lumbar multifidus co-activation) and swimming. She had seen an orthopaedic surgeon, pain specialist, clinical psychologist, a number of physiotherapists and was taking

ARTICLE IN PRESS P. O’Sullivan / Manual Therapy 10 (2005) 242–255

Mal-adaptive CLBP disorders -where ‘movement’ and ‘control’ impairments …. dominate and represent underlying mechanism for pain

Tissue injury / localised pain

Motor response

Movement impairment classification Factors that may influence pain and motor response

physical patho-anatomical genetic neuro-physiological motor control psycho-social coping strategies beliefs fear avoidance compensation

- segmental spinal - directional / multi-directional

Management Non resolution mal-adaptive patterns adopted poor coping strategies NMS response prolonged excessive↔reduced spinal stability abnormal tissue loading peripheral / central sensitisation

Resolution of the disorder

- education – regarding pain mechanism - reduce fear - cognitive behavioural approach - restore movement impairment - graded movement restoration - graded pain exposure - functional restoration - normalise movement behaviour

Control impairment classification - segmental spinal - directional / multi-directional

Management - education – regarding pain mechanism - cognitive behavioural motor control intervention - pain control (avoid provocation) - retrain faulty postures and movements - self control of pain - functional restoration - normalise movement behaviour

Fig. 3. Mal-adaptive motor control impairment CLBP disorders.

CLBP disorders associated with altered motor control

Adaptive / protective altered motor response to an underlying disorder - inflammatory disorders - centrally mediated pain - sympathetically maintained pain - neurogenic pain - neuropathic pain

Altered motor response and centrally mediated pain secondary to dominant psychosocial factors

Mal-adaptive motor control patterns that drive the pain disorder - movement impairments - control impairments (may result in an excess or loss of spinal stability)

Fig. 4. Altered motor responses in the presence of CLBP (3 groups).

249

ARTICLE IN PRESS 250

P. O’Sullivan / Manual Therapy 10 (2005) 242–255

8/10, her disability index (Oswestry disability index) was 40% and she had high levels of kinesiophobia (Tampa scale of Kinesiaphobia). Investigations: Physical examination Observation

X-rays/MRI Lumbar spine— NAD

  

she sat and walked with a rigid erect thoraco-lumbar spine posture she sat forward on the chair with a lordotic spinal posture she maintained thoracolumbar lordosis and avoided flexion when moving from sitting to standing and while undressing

AROM

Fig. 5. (a) Patient with classification of movement impairment into flexion (note the pain provocation into flexion is associated with an impairment of lumbar spine flexion). (b) Patient with classification of control impairment into flexion (note the pain provocation into flexion is not associated with an impairment of lumbar spinal flexion).

anti-depressants, strong analgesic and muscle relaxant medication. She was only able to work 2 days per week doing light duties because of her CLBP disorder. She reported that her symptoms were exacerbated by all flexion postures and movements such as slump sitting, bending, dressing and lifting activities. Extension related spinal movements such as standing and walking were pain free. She gained relief from her pain with heat and rest. She reported high levels of anxiety relating to pain, disability and an inability to work full time. She constantly worried about her back pain and believed that she would not get better as she had a disc injury that had not resolved. She coped with her back pain by avoiding provoking it and restricting her activities involving spinal flexion. Her pain intensity level was

Flexion—hip flexion 501, no thoraco-lumbar flexion with use of hands to support her and assist her return to upright (Fig. 5a) Extension—301 no pain Side bending—full ROM and pain free Repeated flexion increased guarding and report of pain Motion palpation L5/S1—hypo-mobile in flexion Provocation palpation of L4 and L5 centrally— reproduced pain (highly sensitized) SIJ NAD Neural provocation NAD tests Motor control 1. Functional movement tests—stated under observation 2. Specific movement testing—attempts to posteriorly rotate pelvis in sitting, supine and four point kneeling were associated with pain and muscle guarding. 3. Specific muscle testing—able to isolate co-activation of the transverse abdominal wall and lower lumbar multifidus in neutral lordosis (difficulty observed relaxing them). Diagnosis Classification

non-specific CLBP Movement impairment disorder–flexion pattern L5/S1

The disorder diagnosis of non-specific CLBP was based upon the non-resolution of a flexion back sprain and the absence of a specific diagnosis. The disorder classification of this patient was a movement impairment disorder (into flexion with localized pain at L5/S1).

ARTICLE IN PRESS P. O’Sullivan / Manual Therapy 10 (2005) 242–255

The mechanism underlying the pain is a movement impairment with a loss of normal physiological movement into flexion, with associated muscle guarding and fear of forward bending. This movement impairment and associated fear was initiated in the acute phase and was reinforced by her beliefs that pain associated with flexion of her spine was damaging for her. This patient avoided bending due to the knowledge that flexion will provoke pain and the belief (reinforced by treatment providers) that this movement causes ‘further damage’ and that by not moving into this painful direction will prevent damage. The basis of this pain disorder is linked to both dominant peripheral and secondary central pain mechanisms. Management of this patient was directed at both the dominant peripheral and secondary central mechanisms of the pain disorder over a 12 week period. Management first focussed on educating the patient regarding the basis and mechanism of her disorder. It was critical to change the patient’s beliefs, so that she understood that to relax the spinal muscles and restore normal movement in the direction of her pain was essential for resolution of the pain disorder. The patient was assured that her movement-provoked pain into flexion was not dangerous or damaging. The restoration of normal tissue compliance and reduction of muscle guarding was facilitated by ‘passive’ treatment techniques directed to restore flexion mobility to the lower lumbar spine (L5/S1 flexion articular mobilisation techniques and soft tissue inhibitory techniques directed to her back extensor and psoas muscles). This was combined with graded active movement into the restored range. This involved the patient initially being taught to posteriorly tilt her pelvis in a relaxed manner without trunk muscle guarding and breath holding (initially in supine and four point kneeling progressed to sitting and standing). She was instructed to cease cognitively contracting her spinal ‘stabilising muscles’ but rather to relax her upright postures so to reduce her thoracolumbar hyper-lordosis to a neutral spine posture. Finally the patient was trained to flex her spine in upright postures (sitting and standing) in a normal physiological manner without guarding. As the movement impairment was restored, the pain, disability and fear of bending also reduced. At this stage the patient reported that she had the capacity to control her pain. This new control was then introduced into previously provocative functional tasks such as dressing and housework. She reported that she could work longer and increase her general activity levels. She was encouraged to carry out regular cardio-vascular exercise and join a yoga class to maintain her spine mobility in a relaxed manner. The resolution of her CLBP disorder supported the classification and management approach taken.

251

3.3.1. Control impairment classification CLBP disorders classified as ‘control impairment’ appear to be most common in clinical practice. These disorders are associated with impairment or deficits in the control of the symptomatic spinal segment in the primary direction of pain. In these disorders there is no movement impairment in the direction of pain (Figs. 3 and 5b). Pain in these disorders is associated with a loss of functional control around the neutral zone of the spinal motion segment due to specific motor control deficits (and muscle guarding in some situations) of the spinal stabilising muscles. This is manifest during dynamic and/or static tasks as 1. ‘through range movement pain’ due to non-physiological motion of the spinal segment observed during dynamic tasks, 2. ‘loading pain’ due to non-physiological loading of the spinal segment (not end range) observed during static loading tasks and 3. ‘end of range pain’ or ‘overstrain’ due to repetitive strain of the spinal motion segment at the end of range observed during static and dynamic functional tasks. The irony with these patients is that they adopt postures and movement patterns that maximally stress their pain sensitive tissue (Burnett et al., 2004; O’Sullivan et al., 2004; Dankaerts et al., 2005b), and yet they have no awareness that they do this. One reason for this may relate to the fact that their pain is often of a gradual onset and therefore they lack a withdrawal reflex motor response, coupled with a lack of proprioceptive awareness of the lumbo-pelvic region (Fig. 2) (O’Sullivan et al., 2003; Burnett et al., 2004). This control deficit is clearly mal-adaptive and represents a powerful mechanism for ongoing pain (which is both peripherally and centrally mediated) and disability. These patients present with movement based fear that is real, as their movement strategies are highly provocative of their pain disorder, resulting in failure to respond to general exercise and conditioning interventions. These disorders frequently present in a directional manner (flexion, extension (passive or active) and lateral shift control impairment) as well as combinations of these directions (multi-directional control impairment). These disorders may be associated with deficits in the spinal stabilising muscles (i.e. flexion pattern) or excessive muscle activity resulting in increased spinal loading (i.e. active extension pattern). These directional patterns are described in detail elsewhere (O’Sullivan, 2000, 2004). Clinical instability of the lumbar spine represents a sub-group of these disorders (O’Sullivan, 2000, 2004). Management of this sub-group is based on a cognitive behavioural motor learning intervention model. This intervention is based on the premise that mal-adaptive

ARTICLE IN PRESS 252

P. O’Sullivan / Manual Therapy 10 (2005) 242–255

motor control behaviour provides an ongoing mechanism for tissue strain and peripheral nociceptive drive. The aim of the intervention is to desensitize the nervous system by educating the patient to control their pain provocative postures and movement patterns so as to avoid repetitive strain on the painful tissue, reduce the peripheral nociceptive drive and in turn enhance function. This is not simply an exercise program rather it follows a motor learning intervention model with the aim of changing movement behaviour via physical as well as cognitive learning processes. As the motor control is enhanced, the repeated stress on the symptomatic tissue reduces, resulting in less peripheral nociceptive drive into the nervous system, allowing the pain disorder to resolve. This provides the patient with the capacity to manage their disorder in an effective manner, which reduces their fear of activity and increase their levels of function. This intervention directly impacts on both the dominant peripheral nociceptive as well as the secondary central drives for the pain disorder. The role of manual therapy treatment in control impairment disorders is limited only to the restoration of articular movement away from the direction of pain provocation and only if this movement is impaired and inhibiting the muscle synergies controlling this movement. These techniques are never used in isolation, but rather they facilitate movement so as to enhance the restoration of motor control to dynamically unload the pain sensitive tissue. For example in a flexion pattern control impairment disorder, if a loss of segmental spinal extension prohibits restoring control over the lower lumbar lordosis, then manual therapy treatment may be used to facilitate extension. This is immediately followed by training active control over this movement so as to reduce the flexion load of the motion segment. The specifics of this intervention have been reported in detail previously (O’Sullivan, 2000, 2004). 3.3.1.1. Case study 2. A 42-year-old male reports a 2 year history of non-specific CLBP. He first developed central LBP while lifting (with a flexed lumbar spine) a 30 kg bag of fertilizer while working as a labourer. His back pain disorder did not resolve and he had not been able to return to work. His previous treatment consisted of physiotherapy, Pilates, gym based exercise programs, psychological intervention and medication (strong analgesics and antidepressants). He reported that his back pain was provoked by static flexed spinal postures (sitting, driving, semi-inclined bending) and activities (such as lifting, sit—stand, dressing). He reported that he avoided all such activities as they exacerbated his pain and it took days then to settle. He reported relief with extension or lordotic postures.

He reported feeling depressed due to the nature of his disability, his loss of independence and his alienation with his health providers, work and family and was tearful when describing this. He was also limited in his ability to socialize with his friends. He had been told there was nothing structurally wrong with his back and that he would have to learn to live with his problem and he believed that his condition was unlikely to improve. His pain intensity level was 7/10, his disability index (Oswestry disability index) was 42% and he had high levels of kinesiophobia (Tampa scale). Physical examination Observation



he sat down to undress, and used his hands to assist transferring from sitting to standing

AROM

Flexion—no lower lumbar movement impairment (full low lumbar ROM) into flexion with report of LBP mid range (Fig.5b) Extension—301 no pain Right and left side bending—full ROM Repeated and sustained spinal flexion increased his LBP PPIVM L5/S1—hyper-mobile in flexion Provocation palpation of L5/S1 central—painful with reproduction of back pain Neural NAD provocation tests Motor control: 1. Functional movement tests—forward bending, reaching, lifting, sit to stand and squatting were associated with increased flexion at the lower lumbar spine, a loss of anterior pelvic rotation and lordosis in the upper lumbar and thoracic spine (Fig. 4b). The use of the arms was observed to support the trunk with these activities. 2. Specific movement tests—Attempts to initiate anterior pelvic tilt and extend the lower lumbar spine in standing, sitting and supine were associated with upper lumbar and thoracic spine extension 3. Specific muscle testing—Inability to isolate the activation of the pelvic floor, transverse abdominal muscles and lumbar multifidus with posterior pelvic rotation and flexion of the lower lumbar spine, with bracing of the upper abdominal wall.

Investigations Diagnosis Classification

X-rays/MRI lumbar spine— degenerative disc disease L5/S1 (mild) non-specific CLBP control impairment disorder— flexion pattern at L5/S1

ARTICLE IN PRESS P. O’Sullivan / Manual Therapy 10 (2005) 242–255

The diagnosis of non-specific CLBP was based on the non-resolution of a flexion back sprain beyond normal healing time and the lack of a specific diagnosis. The classification of this patient as control impairment disorder (flexion pattern) is based on the underlying mechanism of this pain disorder being directly linked to an ongoing flexion strain of the L5/S1 motion segment secondary to a loss of functional control of the segment into flexion. The patients’ sense of alienation, frustration, anger and depression further confounds his situation resulting in increased central drive of his pain. Management of this patient was directed on a cognitive behavioural motor learning frame-work (O’Sullivan, 2004). The patient was first educated that subsequent to his initial back sprain he had adopted a mal-adaptive motor control pattern that exposed the symptomatic segment to abnormal and repetitive strain into flexion, which in turn maintained his pain. This was further reinforced by his anxiety levels related to work and home, lack of control over his pain disorder and inactivity. Management focused on a motor control intervention to reduce the flexion strain at L5/S1 in a functionally specific manner with relaxation of the thoraco-lumbar spine and enhancing control of segmental lordosis at L5/S1. Initally he was taught to dis-associate lumbo-pelvic lordosis from thoracic in supine, sitting and standing. This was in order to develop proprioceptive awareness and control of this region and so reduce the flexion strain at L5/S1. Once this was achieved he was then taught to coactivate his lower lumbar multifidus with his transverse abdominal wall (in a neutral lordosis), with relaxation of his thoracic erector spinae and upper abdominal muscles (with normal respiration) in these postures. At this stage previously aggravating postures and movements into forward bending were targeted and retrained so that the patient could perform them (controlling the L5/S1 within a neutral lordosis), in a pain-free manner thereby enhancing his functional capacity. This in turn reduced his fear of movement and activity. His exercise program was then progressed into a gym setting where he was taught to integrate his lumbo-pelvic control into a graded cardiovascular exercise program as well as training strength and endurance with loaded tasks such as squats, lunges and resistance lifting tasks. As the patient’s functional mobility increased and pain reduced his coping strategies improved and he was capable of a graduated return to work. The resolution of the disorder supports the classification that the control impairment into flexion represented the dominant underlying mechanism driving the disorder.

4. Validity of the classification system There is a growing concensus within the literature that current diagnostic and classification approaches for

253

CLBP are limited, and a mechanism based classification of CLBP disorders from a biopsychosocial perspective is required (McCarthy et al., 2004). Although considerable research has documented the biopsychosocial nature of CLBP, further research is required to test the validity of this approach in management of CLBP disorders to determine whether it predicts and indeed improves patient outcomes. There is growing evidence to support the validity of the ‘control impairment’ classification system as a subgroup with CLBP. Recent research has shown that physiotherapists trained in the classification system can reliably identify five different subgroups with a classification of control impairment (Dankaerts et al., 2005a, b). Laboratory evidence for the presence of specific motor control and postural deficits have been documented in a series of studies conducted on patients with CLBP with a classification of ‘control impairments’ (O’Sullivan et al., 1997a–c, 2003; Burnett et al., 2004; O’Sullivan et al., 2004; Dankaerts et al., 2005b). Motor learning interventions have been shown efficacious in patient groups with a classification of control impairment, with documented reductions in pain and disability (O’Sullivan et al., 1997a–c, 1998, 2001; Dankaerts et al., 2004).

5. Summary CLBP disorders must be considered within a biopsychosocial framework. The presence and dominance of the potential pathoanatomical, physical, neurophysiological, psychological and social factors that may impact on these disorders is different for each individual with CLBP. This highlights the enormous complexity and individual nature of the problem. It is critical that classification of CLBP pain disorders be based on the mechanism (s) underlying and driving the disorder. It is proposed that motor control impairments may be adaptive or mal-adaptive in nature. The treatment of the signs and symptoms of a pain disorder cannot be justified without an understanding of its underlying mechanism as there are sub-groups of patients for whom physiotherapy treatment is not indicated. It is proposed that there is a large sub-group of CLBP disorders where mal-adaptive movement and control impairments dominate the disorder, resulting in either excessive or impaired dynamic spinal stability and loading. This in turn becomes a mechanism for ongoing pain. Physiotherapy interventions that are classification based and specifically directed to the underlying driving mechanism, have the potential to alter these disorders and impact on both the primary physical and secondary cognitive drivers of pain. This approach is not limited only to the lumbo-pelvic region but can be applied to all regions of the musculoskeletal system. The evidence to

ARTICLE IN PRESS 254

P. O’Sullivan / Manual Therapy 10 (2005) 242–255

date supports these proposals although further research is required to further develop and validate this approach.

References Abenhaim L, Rossignol M, Valat J, Nordin M, Avouac B, Blotman F, Charlet J, Dreiser R, Legrand E, Rozenberg S, Vautravers P. The role of activity in the therapeutic management of back pain— report of the International Paris task force on back pain. Spine 2000;25(4):1S–33S. Adams M, Mannion A, Dolan P. Personal risk factors for low back pain. Spine 1999;24:2497–505. Balague F, Troussier B, Salimen J. Non-specific low back pain in children and adolescents: risk factors. European Spine Journal 1999;8:429–38. Bergstrom G, Bodin L, Jensen I, Linton S, Nygren A. Long term, nonspecific spinal pain: reliable and valid sub-groups of patients. Behaviour Research and Therapy 2001;39:75–8. Bogduk N. The anatomical basis for spinal pain syndromes. Journal of Manipulative and Physiological Therapeutics 1995;18(9):603–5. Bogduk N. Management of chronic low back pain. Australian Medical Journal 2004;180:79–83. Borkan JVTM, Reis S, Schoene ML, Croft P, Hermoni D. Advances in the field of low back pain in primary care: a report from the fourth international forum. Spine 2002;27(5):E128–32. Burnett A, Cornelius A, Dankaerts W, O’Sullivan P. Spinal kinematics and trunk muscle activity in cyclists: a comparison between healthy controls and non-specific chronic low back pain subjects. Manual Therapy 2004;9:211–9. Croft P, Macfarlane G, Papageorgiou A, Thomas E, Silman A, Thomas W, Silman A. British Medical Journal 1998;2(May): 1356–9. Dankaerts W, O’Sullivan P, Burnett A, Straker, L. What precedes RCT’s? A new model for clinical research into NS-CLBP. Eighth International Federation of Manipulative Therapists Conference, Cape Town, South Africa, 2004. Dankaerts W, O’Sullivan P, Burnett A, Straker L, Skouen J. Towards a clinical validation of a classification method for non specific chronic low back pain patients with motor control impairment. Manual Therapy 2005a; in press. Dankaerts W, O’Sullivan PB, Burnett AF, Straker LM. Differences in sitting postures are associated with non-specific chronic low back pain disorders when sub-classified. Spine 2005b; in press. Dillingham T. Evaluation and management of low back pain: and overview. State of the Art Reviews 1995;9(3):559–74. Donatelli R, Wooden MJ. Orthopaedic physical therapy. New York: Churchill Livingstone; 1989. Elvey R, O’Sullivan P. A contemporary approach to manual therapy, Modern Manual Therapy, Boyling and Jull. 3rd ed. Amsterdam: Elsevier; 2004. Flor H, Braun C, Elbert T, Birbaumer N. Extensive reorganisation of primary somatosensory cortex in chronic back pain patients. Neuroscience Letters 1997;224:5–8. Flor H, Turk D. Etiological theories and treatments for chronic back pain. 1. Somatic models and interventions 1984;19:105–21. Ford J, Story I, McMeeken J, O’Sullivan P. A systematic review on methodology of classification system research for low back pain. Proceedings of Musculoskeletal Physiotherapy Australia 13th Biennial Conference, Sydney Australia 2003. Frymoyer J, Rosen J, Clements J, Pope M. Psychological factors in low back pain disability. Clinical Orthopaedics and Related Research 1985;195(May):178–84. Hall T, Elvey R. Nerve trunk pain: physical diagnosis and treatment. Manual Therapy 1999;4(2):63–73.

Hodges P, Moseley G. Pain and motor control of the lumbo-pelvic region: effect and possible mechanisms. Journal of Electromyography and Kinesiology 2003;13:361–70. Laslett M, Williams M. The reliability of selected pain provocation tests for sacro-iliac joint pathology. Spine 1994;19:1243–9. Leboeuf-Yde C, Lauritsen J, Lauritsen T. Why has the search for causes of low back pain largely been nonconclusive? Spine 1997;22(8):877–81. Linton S. A review of psychological risk factors in back and neck pain. Spine 2000;25:1148–56. Maher C, Latimer J, Refshauge K. Prescription of activity for low back pain: what works? Australian Journal of Physiotherapy 1999; 45:121–32. Main C, Watson P. Guarded movements: development of chronicity. Journal of Musculoskeletal Pain 1996;4(4):163–70. Maitland J. Vertebral manipulation. London: Butterworths; 1986. McCarthy C, Arnall F, Strimpakos N, Freemont A, Oldham J. The biopsychosocial classification of non-specific low back pain: a systematic review. Physical Therapy Reviews 2004;9:17–30. McGill S. Linking latest knowledge of injury mechanisms and spine function to the prevention of low back disorders. Journal of Electromyography and Kinesiology 2004;14:43–7. McKenzie R. The lumbar spine, mechanical diagnosis and treatment. Waikanae, New Zealand: Spinal Publications Ltd.; 1981. McKenzie R. Mechanical diagnosis and therapy for disorders of the low back. Physical Therapy for the low back. Ta Taylor, New York: Churchill Livingstone; 2000. p. 141–66. Moseley L. A pain neuromatrix approach to patients with chronic pain. Manual Therapy 2003;8(3):130–40. Nachemson A. Back pain; delimiting the problem in the next millenium. International Journal of Law Psychiatry 1999;22(5–6):473–80. Newcomer K, Sinaki M. Low back pain and its relationship to back strength and physical activity in children. Acta Paediatrica 1996;85(12):1433–9. O’Sullivan P. The efficacy of specific stabilising exercises in the management of chronic low back pain with radiological diagnosis of lumbar segmental instability. Curtin University of Technology; 1997. O’Sullivan P. Lumbar segmental instability: clinical presentation and specific exercise management. Manual Therapy 2000;5(1):2–12. O’Sullivan, P. Clinical instability of the lumbar spine, Modern Manual Therapy, Boyling and Jull. 3rd ed. Amsterdam: Elsevier; 2004. O’Sullivan P, Beales D, Avery A. Normalisation of aberrant motor patterns in subjects with sacro-iliac joint pain following a motor learning intervention: a multiple subject case study investigating the ASLR test. In Proceedings of: Manipulative Physiotherapists Association of Australia Twelth Biennial Conference, Adelaide, Australia, 2001. O’Sullivan P, Burnett A, Floyd A, Gadsdon K, Logiudice J, Quirke H. Lumbar repositioning deficit in a specific low back pain population. Spine 2003;28(10):1074–9. O’Sullivan P, Myers T, Jensen L, Murray K. Characteristics of children and adolescents with chronic non-specific musculo-skeletal pain. Australian Physiotherapy Association 8th International Physiotherapy Congress, Adelaide, APA 2004. O’Sullivan P, Twomey L, Allison G. Dysfunction of the neuromuscular system in the presence of low back pain—implications for physical therapy management. Journal of Manual and Manipulative Therapy 1997a;5(1):20–6. 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 1997b;22(24):2959–67. O’Sullivan P, Twomey L, Allison G, Sinclair J, Miller K, Knox J. Altered patterns of abdominal muscle activation in patients with chronic back pain. Australian Journal of Physiotherapy 1997c; 43(2):91–8.

ARTICLE IN PRESS P. O’Sullivan / Manual Therapy 10 (2005) 242–255 O’Sullivan P, Twomey L, Allison G, Taylor J. Altered motor control following specific exercise intervention in subjects with chronic low back pain and a clinical and radiological diagnosis of lumbar segmental ‘‘instability’’. Proceedings of the World Congress of Low Back and Pelvic Pain Conference, Vienna, Austria 1998. O’Sullivan PB, Mitchell T, Bulich P, Waller R, Holte J. The relationship between posture, lumbar muscle endurance and low back pain in industrial workers. Manual Therapy 2005; in press. Padfield BCB, Butler R. Use of an outcome measurement system to answer a clinical question: Is the Quebec task force classification system useful in an outpatient setting? Physiotherapy Canada 2002;(Fall):254–60. Pope M, Hansen T. Vibration and low back pain. Clinical Orthopaedics and Related Research 1992;279(June):49–58. Richardson CA, Jull GA. Muscle control—pain control. What exercises would you prescribe? Manual Therapy 1995;1(1):2–10. Sahrmann S. Movement impairment syndromes of the lumbar spine. St Louis: Mosby; 2001. Schwarzer A, Aprill C, Derby R, Fortin J, Kine G, Bogduk N. The relative contributions of the disc and zygapophyseal joint in chronic low back pain. Spine 1994;19(7):801–6.

255

Spitzer W. Scientific approach to the assessment and management of activity related spinal disorders. Spine 1987;7:S1–S55. Stuge B, Lærum E, Kirkesola G, Vøllestad N. The efficacy of a treatment program focusing on specific stabilizing exercises for pelvic girdle pain after pregnancy. A randomized controlled trial. Spine 2004;29:351–9. Van-Dieen J, Selen L, Cholewicki J. Trunk muscle activation in lowback patients, an analysis of the literature. Journal of Electromyography and Kinesiology 2003;13:333–51. Waddell G. The back pain revolution. Edinburgh: Churchill Livingstone; 2004. Woby S, Watson P, Roach N, Urmston M. Are changes in fearavoidance beliefs, catastrophizing, and appraisals of control, predictive of changes in chronic low back pain and disability? European Journal of Pain 2004;8(3):201–10. Wright A, Zusman M. Neurophysiology of pain and pain modulation. In: Modern Manual Therapy, Boyling and Jull. 3rd ed. Amsterdam: Elsevier; 2004. Zusman M. Forebrain-mediated sensitization of central pain pathways: ‘non-specific’ pain and a new image for MT. Manual Therapy 2002;7(2):80–8.

ARTICLE IN PRESS

Manual Therapy 10 (2005) 256–269 www.elsevier.com/locate/math

Review article

Inter-examiner reliability of passive assessment of intervertebral motion in the cervical and lumbar spine: A systematic review E. van Trijffela,, Q. Anderegga,b, P.M.M. Bossuyta, C. Lucasa a

Department of Clinical Epidemiology and Biostatistics, Academic Medical Centre, University of Amsterdam, Amsterdam, The Netherlands b Vrije University Hospital, Amsterdam, The Netherlands Received 7 December 2004; received in revised form 28 March 2005; accepted 27 April 2005

Abstract A systematic review was conducted to determine inter-examiner reliability of passive assessment of segmental intervertebral motion in the cervical and lumbar spine as well as to explore sources of heterogeneity. Passive assessment of motion is used to decide on treatments for neck and low-back pain patients. Inter-examiner reliability has been a matter of debate, resulting in questions about professional credibility and accountability. A structured search for relevant studies in MEDLINE and CINAHL was followed by extensive reference tracing and hand searching. Studies presenting estimates of reliability for individual motion segments were included. No language restrictions were imposed. Study quality was assessed using criteria derived from the Standards for Reporting of Diagnostic Accuracy (STARD) statement and a quality assessment tool for studies of diagnostic accuracy included in systematic reviews (QUADAS). Study selection, quality assessment, and data extraction were performed by two reviewers independently. Qualitative analyses and additional subgroup analyses were conducted. Nineteen studies were included. Two studies satisfied criteria for external and internal validity, of which one found fair to moderate reliability. Assessment of motion segments C1–C2 and C2–C3 almost consistently reached at least fair reliability. Overall, inter-examiner reliability was poor to fair. However, most studies were found to be of poor methodological quality. We propose explicit recommendations for the conduct and reporting of future research. r 2005 Elsevier Ltd. All rights reserved. Keywords: Spine; Motion assessment; Reliability; Reproducibility of results; Systematic review

1. Introduction An overview of epidemiologic research has shown high prevalence rates of neck and low-back pain in developed countries (Nachemson et al., 2000). Generally, at some point during the clinical course, many patients suffering from these conditions are treated by manual practitioners, such as physiotherapists, manual therapists, chiropractors, osteopaths and physicians. In the Netherlands, according to guidelines, general practitioners may refer patients with low-back pain persisting longer than 6 weeks (Faas et al., 1996). FiftyCorresponding author. Plantageweg 30, 3061 PK, Rotterdam, The Netherlands. Tel.: +31 10 4527409. E-mail address: [email protected] (E. van Trijffel).

1356-689X/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2005.04.008

nine percent of patients with chronic neck pain are referred to a physiotherapist or manual therapist (Borghouts et al., 1999). Passive assessment of the quantity and quality of motion—also known as motion palpation—in individual vertebral motion segments guides decisions on treatment (Jull et al., 1994). Reliability reflects the extent to which practitioners are able to differentiate diagnostically among individuals who vary in characteristics (Streiner and Norman, 2003). Furthermore, an estimate of inter-examiner reliability can be used to quantify the extent to which practitioners show variability in diagnostic assessment (Brennan and Silman, 1992). A satisfactory level of inter-examiner reliability is a prerequisite for valid and uniform decisions about patients (Bartko and Carpenter, 1976). Variability

ARTICLE IN PRESS E. van Trijffel et al. / Manual Therapy 10 (2005) 256–269

among examiners has empirically been shown to affect diagnostic accuracy (Whiting et al., 2004). At this moment, it is unclear to what extent practitioners vary in their motion assessment of the spine. Inter-examiner reliability of passive intervertebral motion assessment has been a matter of debate, resulting in questions about professional credibility and accountability (Breen, 1992; Maher and Latimer, 1992; Troyanovich and Harrison, 1998). Four narrative reviews (Keating, 1989; Haas, 1991b; Panzer, 1992; Huijbregts, 2002) concerning reliability of spinal motion assessment have been published, of which two (Keating, 1989; Panzer, 1992) dealt with the lumbar spine only. None of these reviews formally assessed methodological quality of included studies. Two extensive systematic reviews have appeared covering reliability of chiropractic tests for the lumbo-pelvic spine (Hestbœk and Leboeuf-Yde, 2000) and spinal palpation tests (Seffinger et al., 2004). In both reviews, it was concluded that inter-examiner reliability of passive intervertebral motion assessment was low. Seffinger et al. (2004) added that assessing regional range of motion was more reliable than evaluating segmental range of motion. However, in both reviews, criteria for assessing methodological quality of studies were not substantiated by evidence of variation and bias in diagnostic research. Furthermore, none of all the above-mentioned reviews explicitly analysed reliability for individual motion segments. So far, no single study has been able to demonstrate acceptable inter-examiner reliability. A systematic review on this topic is needed to allow for an objective appraisal of existing evidence (Egger et al., 2001). We conducted a systematic review of the available literature to determine inter-examiner reliability of passive assessment of segmental intervertebral motion in the cervical and lumbar spine. In addition, we explored sources of heterogeneity.

257

2. Methods 2.1. Study selection Assisted by a clinical librarian, we developed a structured search strategy to identify relevant studies in the MEDLINE database (through PubMed) published between January 1, 1966 and March 31, 2004 (Box 1). The search and study selection were performed by two reviewers (EvT and QA) independently. Based on information in title and abstract, possibly relevant studies were selected and retrieved as a full article. Studies, or subsets of studies, meeting the following criteria were included:










published as a full article; using a repeated-measures, inter-examiner reliability design; evaluating passive motion assessment of one or more motion segments of the cervical (C0–T4) and lumbar (T12–S1) spine performed by manual practitioners; applying judgement criteria that could either concern the quantity (e.g. range of motion, joint play, restriction) or quality (e.g. end-feel, resistance, stiffness) of motion; presenting estimates of inter-examiner reliability for individual motion segments.

No restrictions were imposed on language and date of publication. Abstracts and theses were not included. Studies evaluating active movements or incorporating other clinical symptoms, like pain, into the judgement process were not considered. The first reviewer performed an additional search in the CINAHL database (1982–March 31, 2004). All of the retrieved article references and relevant reviews were further examined by the first reviewer for additional

Box 1. Search strategy for studies on inter-examiner reliability of passive assessment of intervertebral motion in the cervical and lumbar spine in MEDLINE (through PubMed) using medical subject headings (mh) and text words (tw).

ARTICLE IN PRESS E. van Trijffel et al. / Manual Therapy 10 (2005) 256–269

258

Box 2. Criteria list for assessing methodological quality of studies on inter-examiner reliability.

publications. This strategy was complemented by hand searching of nine journals (January 1990–March 31, 2004). A complete list of journals is available from the authors. Eligibility was checked by the second reviewer. Disagreements were resolved by discussion. If disagreement persisted, the judgement of a third reviewer (CL) was decisive. 2.2. Quality assessment A validated list of criteria for assessing methodological quality of inter-examiner reliability studies was not available. We therefore developed a list of 11 criteria for assessing study quality (Box 2). Seven of the criteria were derived from evidence of variation and design-related bias in diagnostic accuracy studies (Lijmer et al., 1999; Whiting et al., 2004), the Standards for Reporting of Diagnostic Accuracy (STARD) statement (Bossuyt et al., 2003a, b), and a validated tool for assessing quality of studies of diagnostic accuracy included in systematic reviews (QUADAS) (Whiting et al., 2003). Based on theoretical evidence (Cohen, 1960; Maclure and Willett, 1987; Thompson and Walter, 1988; Feinstein and Cicchetti, 1990; Altman, 1991; Haas, 1991a; Brennan and Silman, 1992; Byrt et al., 1993; Rothstein and Echternach, 1993; Streiner and Norman, 2003), items 5, 9, 10, and 11 were added to fit the context of reliability. Criteria were designed to tap domains of external validity (items 1–3), internal validity (items 4–8), and statistical methods used (items 9–11). Scores on items 4 and 7 were assumed to be of decisive importance for internal validity. After a training session, two papers (Potter and Rothstein, 1985; Meijne et al., 1999) were used to evaluate interpretability and applicability of items by all reviewers. EvT and QA, who were not blinded to information on authors and journals, independently assessed methodological quality of all included studies. Items were scored

by answering with ‘‘Yes’’, ‘‘No’’, or ‘‘?’’(unclear because of insufficient information). Items were equally weighted. Inter-reviewer reliability was analysed by calculating percentage agreement and a kappa (k) statistic. Disagreements were resolved by discussion. In case disagreement persisted, CL made the final decision. 2.3. Data extraction We extracted data from the original studies on participants (number, age, gender, clinical characteristics, setting), examiners (number, profession, expertise, pre-training, experience), assessment procedure (subject position, motion segments, motion directions), judgement criteria and scales (quantitative and qualitative classifications), and inter-examiner reliability for individual motion segments (point estimates and estimates of precision). EvT and QA extracted data independently. If disagreement persisted after discussion, consensus was met consulting CL. 2.4. Statistical analysis Qualitative analyses were conducted by examining results on reliability from studies with high methodological quality, as well as by examining characteristics of studies that showed the highest and lowest levels of reliability. Additionally, analyses for subgroups of participants, examiners, assessment procedures, judgement criteria and scales, and motion segments were performed. Analyses were carried out for the cervical spine and lumbar spine separately. Value labels for corresponding ranges of kappa (k) statistics were used as assigned by Landis and Koch (1977) (Box 3). Intraclass correlation coefficients above 0.75 are assumed to indicate an acceptable level of reliability (Burdock et al., 1963, cited by Kramer and Feinstein, 1981).

ARTICLE IN PRESS E. van Trijffel et al. / Manual Therapy 10 (2005) 256–269

Box 3. Value labels for ranges of kappa statistics (k).

3. Results Searching MEDLINE yielded 228 citation postings. Of these, 18 possibly relevant studies (Johnston et al., 1982; DeBoer et al., 1985; Love and Brodeur, 1987; Viikari-Juntura, 1987; Mootz et al., 1989; Nansel et al., 1989; Keating et al., 1990; Binkley et al., 1995; Van Dillen et al., 1998; Fjellner et al., 1999; Hawk et al., 1999; French et al., 2000; Scho¨ps et al., 2000; Smedmark et al., 2000; Van Suijlekom et al., 2000; Christensen et al., 2002; Hicks et al., 2003; Pool et al., 2004) were retrieved as full articles. Eight studies fulfilled all eligibility criteria. Searching CINAHL led to the inclusion of one other study (Strender et al., 1997a). Reference tracing and hand searching yielded 16 more possibly relevant studies (Kaltenborn and Lindahl, 1969; Gonella et al., 1982; Mior et al., 1985; Bergstro¨m and Courtis, 1986; Jull and Bullock, 1987; Boline et al., 1988; Leboeuf et al., 1989; Richter and Lawall, 1993; Maher and Adams, 1994; Inscoe et al., 1995; Schoensee et al., 1995; Haas et al., 1995b; Jull et al., 1997; Strender et al., 1997b; Hanten et al., 2002; Downey et al., 2003), of which eight met the inclusion criteria. From the total of 18 excluded studies, 13 were excluded for reasons of design features (Kaltenborn and Lindahl, 1969), evaluating active movements (Viikari-Juntura, 1987; Van Dillen et al., 1998; Van Suijlekom et al., 2000), incorporating other clinical examination symptoms into the judgement process (Binkley et al., 1995; Jull et al., 1997; Hawk et al., 1999; French et al., 2000; Hanten et al., 2002; Downey et al., 2003), and not examining individual motion segments (Johnston et al., 1982; Love and Brodeur, 1987; Nansel et al., 1989). Another five studies (Jull and Bullock, 1987; Leboeuf et al., 1989; Schoensee et al., 1995; Haas et al., 1995b; Christensen et al., 2002) were initially excluded because estimates of inter-examiner reliability for individual motion segments were not

259

presented. Therefore, first authors were contacted and requested to provide segmental data. In total, 19 studies could be included in this review. Nine studies (DeBoer et al., 1985; Mior et al., 1985; Haas et al., 1995b; Strender et al., 1997b; Fjellner et al., 1999; Scho¨ps et al., 2000; Smedmark et al., 2000; Christensen et al., 2002; Pool et al., 2004) examined reliability for the cervical spine and 10 (Gonella et al., 1982; Bergstro¨m and Courtis, 1986; Boline et al., 1988; Mootz et al., 1989; Keating et al., 1990; Richter and Lawall, 1993; Maher and Adams, 1994; Inscoe et al., 1995; Strender et al., 1997a; Hicks et al., 2003) evaluated the lumbar spine. Study characteristics are given in Table 1 (cervical spine) and Table 2 (lumbar spine). There were no disagreements between reviewers on selection of studies and extraction of data. Methodological quality scores of included studies are presented in Table 3. There were 10 disagreements between reviewers on quality scores, resulting in 95% agreement and interreviewer reliability (k) of 0.93. All disagreements were resolved by discussion, consequently there was no need to consult the third reviewer for a final decision. 3.1. Cervical spine Data on estimates of inter-examiner reliability of passive assessment are given in Table 1 (last column). Inter-examiner reliability for the cervical spine ranged from poor to substantial. Overall, reliability was poor to fair. The study by Smedmark et al. (2000) fulfilled all criteria for external validity. It showed fair to moderate reliability among two physical therapists making judgements on stiffness. In the other study (Pool et al., 2004) that used representative patients, substantial reliability was reached for evaluating motion segment C2–C3 and, overall, reliability was slight to fair. Two studies (Strender et al., 1997b; Smedmark et al., 2000) fulfilled both criteria for internal validity. Strender et al. (1997b) achieved slight reliability for assessing the upper cervical spine in volunteers. The study by Mior et al. (1985) scored positive on the criterion of stability of characteristics. It showed slight reliability among two pre-trained students of chiropractic examining fixations of vertebra C1 in healthy students. The study by Fjellner et al. (1999) did not satisfy this criterion because a large number of tests were involved. In their study with healthy volunteers, estimates of reliability ranged from poor to moderate. Fair to moderate reliability was consistently shown in one study (Smedmark et al., 2000) that was externally and internally valid. The lowest levels of reliability were reached by Christensen et al. (2002), with values of kappa statistics up to 0.42, for prone joint play evaluation of the upper thoracic spine in non-represen-

260

Table 1 Characteristics of included studies (ranked in alphabetical order) on inter-examiner reliability of passive assessment of intervertebral motion in the cervical spine First author (year)

Participants

Examiners

Assessment procedure

Judgement criteria and scales

Estimates of inter-examiner reliability

Christensen et al. (2002)

29 patients referred to Dept. of Cardiology with known or suspected stable angina pectoris+27 controls referred to Dept. of Nuclear Medicine. Taken from original sample of 107 with age range 31–74 yr and 68 (64%) males. University Hospital, Denmark

Two pre-trained chiropractors

Subject seated: MSs T1–T4 in lateral flexion R/ L, rotation R/L

Abnormality (based on end-play restriction and joint play, respectively): absent–present

Seated: T1–T2 PA T2–T3 PA T3–T4 PA Prone: T1–T2 PA T2–T3 PA T3–T4 PA

40 healthy students of chiropractic. Mean age 26.2 yr (range 21–44). 40 (100%) males. College of Chiropractic, US

Three (3 pairs) chiropractors. Range 5–14 yr of experience

47 healthy volunteers by advertising or inquiry. Mean age 37.9 yr (SD79.5, range 18–63) Eight (17%) males. Sweden

Two physiotherapists specialized in orthopaedic manual therapy 6 and 12 yr of experience

Haas et al. (1995b)

73 first year students of chiropractic of which 48 (66%) mild symptomatic. Mean age 27.1 yr (SD75.2) 49 (67%) males. College of Chiropractic, US

Two pre-trained chiropractors (faculty members). 15 yr of experience

Fixation:

Vertebrae C1–C7 in flexion, extension, rotation, lateral flexion R/L (reference cited)

Normal-slight-obvious

Subject seated: MS C0–C1 in flexion, extension MS C1–C2 in rotation R/L MSs C2–T4 in rotation R/L First rib R/L

Range of motion MSs C0–C1/ C1–C2/C2–T4: reduced–normal– increased, first rib: reduced– normal

Subject supine: MSs C2–T4 in flexion, extension

Joint play MSs C2–T4: reduced–normal–increased

Subject left side-lying: MSs C2–T4 joint play. (references cited)

End-feel: MSs C0–C1/ C1–C2: hard–normal–empty

Subject seated:

Restriction (based on hard endplay): absent–present

MS T3–T4 in rotation R/L. (reference cited)

C1–C2 (pooled data) PA 56%, kw 0.23; PA 21%, kw 0.03; PA 38%, kw 0.09 C6–C7 (pooled data) PA 44%, kw 0.40; PA 58%, kw 0.41; PA 49%, kw 0.45 Range of motion: C0–C1 flexion PA 62%, kw 0.00 (CI [0.27, 0.27]), extension PA 87%, kw NC; C1–C2 rotation R PA 62%, kw 0.15 (CI [0.14,0.44]), L PA 79%, kw 0.41 (CI [0.096, 0.72]); C2–T4 ranging from flexion C7–T1 PA 72%, kw 0.16 (CI [0.26,0.062]) to rotation L T2–T3 PA 81%, kw 0.49 (CI [0.22,0.76]); first rib R PA 92%, kw NC, L PA 77%, kw 0.06 (CI [0.21,0.33]) Joint play: ranging from C4–C5 PA 79%, kw 0.05 (CI [0.089, 0.011]) to C3–C4 PA 83%, kw 0.36 (CI [0.27, 0.69]) End-feel: C0–C1 flexion PA 64%, kw 0.01 (CI [0.24,0.26]), extension PA 87%, kw NC; C1–C2 rotation R PA 60%, kw 0.06 (CI [0.21,0.33]), L PA 75%, kw 0.18 (CI [0.075,0.43]) T3–T4 rotation R k 0.03 (SE 0.01), L k NC, either R/L k 0.04 (SE 0.03)

ARTICLE IN PRESS

Fjellner et al. (1999)

Subject seated:

75%, k 0.19 73%, k 0.42 68%, k 0.23

E. van Trijffel et al. / Manual Therapy 10 (2005) 256–269

DeBoer et al. (1985)

Subject prone: MSs T1–T4 joint play (reference cited)

75%, k 0.11 77%, k 0.00 75%, k 0.32

Two pre-trained students of chiropractic (first year of clinical training)

Subject supine: Vertebra C1 in lateral flexion R/L, anterior rotation

Fixation (based on joint play and end-feel): absent-present

C1 PA 62%, k 0.15

Pool et al. (2004)

32 patients with neck complaints. Mean NDI score 15.2 (SD78.3), 56.3% had previous episodes, mean present pain on 11-point scale 4.2 (SD72.3), median duration of pain 13.5 wks. Mean age 45.5 yr (SD79.2). 12 (37.5%) males. Practice for Physical and Manual Therapy, The Netherlands

Two pre-trained physical therapists

Subject supine: MS C0–C1in flexion MS C1–C2 in rotation R/L MSs C2–T2 in lateral flexion R/L

Limitation of movement (based on range of motion and resistance): yes–no

C0–C1 flexion PA 77%, k 0.29 C1–C2 rotation R PA 84%, k 0.20, L PA 90%, k 0.37 C2–T2 ranging from C4–C5 lateral flexion L PA 68%, k 0.09 to C2–C3 lateral flexion L PA 84%, k 0.63

Scho¨ps et al. (2000)

20 patients with cervical spine syndrome: mean age 37 yr (range 21–55), 8 (40%) males+20 healthy volunteers: mean age 33 yr (range 20–49), 10 (50%) males. Clinic for Physical Medicine and Rehabilitation, Germany

Five physicians specialized in manual medicine

MS C0–C1 in lateral flexion R/L, rotation R/L MS C1–C2 in nodding-flexionrotation R/L MS C2–C3 in nodding-flexionlateral flexion R/L MSs C3–C6 in unspecified direction R/L MSs C6–T1 in rotation R/L

Hypomobility: absent-present

C0–C1 lateral flexion R k 0.04, L k 0.13; rotation R k 0.08, L k 0.04 C1–C2 R k 0.22, L k 0.28 C2–C3 R k 0.04, L k 0.34 C3–C4 R k 0.43, L k 0.06 C4–C5 R k 0.03, L k 0.03 C5–C6 R k 0.28, L k 0.17 C6–C7 R k 0.44, L k 0.29 C7–T1 R k 0.26, L k 0.30

Smedmark et al. (2000)

61 patients seeking care for nonspecific neck problems. Age range 20–71 yr. 15 (24.5%) males. Private Clinic, Sweden

Two pre-trained physical therapists specialized in orthopaedic manipulative therapy. Over 25 yr of experience

Subject seated: MS C1–C2 in rotation R/L Subject supine: MS C2–C3 in lateral flexion R/L First rib R/L Subject side-lying:MS C7–T1 in flexion, extension

Stiffness (based on range of motion and end-feel): Yes–no difference when R compared to L. For C7–T1 compared to C6–C7 and T1–T2

C1–C2 PA 87%, k 0.28 C2–C3 PA 70%, k 0.43 C7–T1 PA 79%, k 0.36 first rib PA 70%, k 0.35

Strender et al. (1997b)

50 volunteers of which 25 with complaints in neck-shoulder region. Mean age 41.7 yr (SD710.4, range 21–66). 13 (26%) males. Sweden

Two pre-trained physiotherapists specialized in manual medicine. 21 and 23 yr of experience

Subject supine: MS C0–C1 in lateral flexion–rotation R/L MSs C0–C2 in rotation (in flexion-position) R/L MS C2–C3 in lateral flexion R/L. (reference cited)

Mobility: Yes–no difference when R and L compared

C0–C1 PA 26%, k 0.091 (CI [0.22,0.40]) C0–C2 PA 42.9%, k 0.15 (CI [0.06,0.37]) C2–C3 PA 44%, k 0.057 (CI [0.23,0.35])

Note: CI: 95% confidence interval, k: kappa statistic, kw: weighted kappa statistic, L: left, MS: motion segment, NC: not calculated, NDI: Neck Disability Index, PA: percentage agreement, R: right, SD: standard deviation, SE: standard error.

ARTICLE IN PRESS

59 healthy students of chiropractic. Age range 22–30 yr. College of Chiropractic, Canada

E. van Trijffel et al. / Manual Therapy 10 (2005) 256–269

Mior et al. (1985)

261

262

Table 2 Characteristics of included studies (ranked in alphabetical order) on inter-examiner reliability of passive assessment of intervertebral motion in the lumbar spine First author (year)

Participants

Examiners

Assessment procedure

Judgement criteria and scales

Estimates of inter-examiner reliability

Bergstro¨m and Courtis (1986)

100 healthy students of chiropractic. UK

Two students of chiropractic

Subject seated: Vertebrae L1–L5 in lateral flexion R/L

Fixation (based on hard endfeel): absent–present

L1 L2 L3 L4 L5

Boline et al. (1988)

23 symptomatic LBP Two pre-trained chiropractors patients+27 asymptomatic. 27 (members of campus Motion (54%) males. US Palpation Club). One senior intern and one recent graduate

Subject seated: MSs T12–S1 in flexion, extension, lateral flexion R/L, rotation R/L

Fixation (based on hard endfeel): normal-obvious

N ¼ 50 T12–L1 PA 70%, k 0.31 L1–L2 PA 60%, k 0.02 L2–L3 PA 74%, k 0.02 L3–L4 PA 78%, k 0.31 L4–L5 PA 82%, k 0.19 L5–S1 PA 90%, k 0.05

R R R R R

PA PA PA PA PA

83%, 82%, 88%, 87%, 78%,

L L L L L

PA PA PA PA PA

75% 79% 84% 89% 73%

k k k k k k

0.32 0.05 0.09 0.33 0.25 0.06

Five healthy students of physical therapy. Age range 22–27 yr 0 (0%) males. US

Five pre-trained physical MSs T12–S1 in flexion, lateral therapists. 3, 3, 4, 5 and 20 yr of flexion R/L, rotation R/L experience

Mobility: 7-point scale (with half points) ranging from 0 ¼ ankylosed to 6 ¼ unstable with reference point the expected normal for age, body type and activity level

Only descriptive statistics averaged over all motion directions: mean (SD) T12–L1 ranging from 2.93 (0.18) to 3.23 (0.41) L1–L2 ranging from 2.80 (0.30) to 3.00 (0.00 and 0.28) L2–L3 ranging from 2.60 (0.35) to 2.80 (0.30) L3–L4 ranging from 2.05 (0.51) to 2.85 (0.76) L4–L5 ranging from 2.18 (0.41) to 2.73 (0.26) L5–S1 ranging from 2.23 (0.34) to 3.00 (0.36)

Hicks et al. (2003)

63 patients with current complaints of LBP recruited either as consecutive participants in research on LBP or as patients referred to an outpatient physical therapy clinic. 51 (80.9%) had previous episodes, mean Oswestry score 17.8 (SD711.3, range 92–52). Mean age 36.0 yr (SD710.3, range 20–66). 25 (39.6%) males. US

Four pre-trained examiners of which three physical therapists and one physical therapist/ chiropractor. Specialized in orthopaedic physical therapy (2) and experienced in an orthopaedic setting (2). 4, 5, 6 and 8 years of experience

Subject prone: Vertebrae L1–L5 in anteroposterior direction by applying an anteriorly directed pressure on spinous process.(reference cited)

Mobility: hypermobile-normalhypomobile relative to adjacent motion segments and expectation of the examiner

L1 L2 L3 L4 L5

Inscoe et al. (1995)

Six volunteers currently experiencing LBP (but have not sought care) and a reported history of two or more previous episodes. Mean age 29.3 yr (range 24–34). Two (33.3%) males. US

Two physical therapists specialized in orthopaedic manual therapy. 4–5 yr of experience

Subject right side-lying: MSs T12–S1 in flexion with double leg flexion technique.(reference cited)

Mobility: normal-hypomobilehypermobile relative to the expected normal for age, body type, gender and activity level

T12–L1 PA 33.33% L1–L2 PA 58.33% L2–L3 PA 50.0% L3–L4 PA 41.67% L4–L5 PA 58.33% L5–S1 PA 50.0%

Keating et al. (1990)

21 LBP patients+25 asymptomatic students. Age range 23–60 yr. 20 (43.5%) males. US

Three (3 pairs) pre-trained chiropractors. 2.5, 5 and 10 yr of experience

Subject seated: MSs T11–S1 in flexion, extension, lateral flexion R/L, rotation R/L (reference cited)

Fixation (based on hard endfeel): absent-present

T11–T12 k 0.04; k 0.03; k 0.04 (mean k 0.02) T12–L1 k 0.09; k 0.15; k 0.00 (mean k 0.02) L1–L2 k 0.23; k 0.13; k 0.01 (mean k 0.04) L2–L3 k 0.14; k 0.14; k 0.25 (mean k 0.08) L3–L4 k 0.13; k 0.18; k 0.04 (mean k 0.03) L4–L5 k 0.09; k 0.29; k 0.28 (mean k 0.22) L5–S1 k 0.31; k 0.22; k 0.17 (mean k 0.23)

PA PA PA PA PA

68%, 69%, 52%, 58%, 65%,

kw 0.26 (CI [0.01,0.53]) kw 0.17 (CI [0.13, 0.47]) kw 0.02 (CI [0.25,0.28]) kw 0.11, (CI [0.26,0.35]) kw 0.18 (CI [0.03, 0.49])

ARTICLE IN PRESS

Gonella et al. (1982)

E. van Trijffel et al. / Manual Therapy 10 (2005) 256–269

N ¼ 23 PA 65%, PA 61%, PA 74%, PA 87%, PA 87%, PA 87%,

Maher and Adams (1994)

Mootz et al. (1989)

90 patients with non-specific mechanical LBP. 82% previous history, mean time since onset 45.2 days (SD7100.0, range 1–730). Mean age 45.37 yr (SD714.16, range 21–78). 34 (37.7%) males. Physical Therapy Clinics, Australia

6 (3 pairs) physical therapists specialized in manipulative physiotherapy. Range 8–21 yr of experience

Subject prone:

Stiffness:

L1 PA 0.14 L2 PA 11-point scale ranging from Vertebrae L1–L5 in antero0.40 posterior direction by applying 5 ¼ markedly decreased L3 PA stiffness to 5 ¼ markedly an anteriorly directed force 0.25 over spinous process (reference increased stiffness with 0 ¼ normal stiffness relative to L4 PA cited) 0.00 the expected normal L5 PA 0.25

60 students of chiropractic. US Two pre-trained chiropractors. Subject seated: 7 and 10 yr of experience MSs L1–S1 in flexion, extension, lateral flexion R/L, rotation R/L. 2 sessions. (reference cited)

Fixation (based on hard endfeel): absent-present

Subject seated: Vertebrae L1–L5 in flexion, extension, lateral flexion R/L, rotation R/L. Subject prone: Vertebrae L1–L5 in anteroposterior direction (reference cited)

Strender et al. (1997a)

Four pre-trained of which two physiotherapists specialized in manual medicine and two physicians

Subject side-lying: Mobility: MSs L4–S1 in angular and Decreased–normal–increased translational directions with hips and knees flexed (reference cited)

Physiotherapists’ group: 50 patients. Mean age 37.7 yr (SD711.7, range 16–69). 17 (34%) males. Physicians’ group: 21 patients. Mean age 41.2 yr (SD715.7, range 20–71). 11 (52.4%) males. Private Outpatient Clinic specializing in back pain, Sweden

40%, ICC 0.18; PA 13%, ICC 0.28; PA 27%, ICC 27%, ICC 0.41; PA 26%, ICC 0.54; PA 30%, ICC 43%, ICC 0.73; PA 20%, ICC 0.37; PA 23%, ICC

L1–L2 PA 80%, k 0.06; PA 85%, k 0.05 L2–L3 PA 76.7%, k 0.13; PA 85%, k 0.11 L3–L4 PA 70%, k 0.17; PA 75%, k 0.03 L4–L5 PA 63.3%, k 0.02; PA 61.7%, k 0.08 L5–S1 PA 73.3%, k 0.17; PA 73.3%, k 0.08

Mobility: hypermobile-normal- Seated: L1 ranging from flexion k 0.18 to lateral flexion L k 0.72 hypomobile L2 ranging from flexion k 0.20 to lateral flexion L k 0.69 L3 ranging from lateral flexion L k 0.25 to lateral flexion R k 0.34 L4 ranging from lateral flexion R k 0.11 to rotation L k 0.29 L5 ranging from lateral flexion R k 0.08 to flexion k 0.25 Prone: L1 k 0.14, L2 k 0.18, L3 k 0.11, L4 k 0.08, L5 k 0.17

physiotherapists L4–L5 PA 82%, kw 0.66 (CI [0.45,0.86]) L5–S1 PA 80%, kw 0.75 (CI [0.60,0.90])

Note: CI: 95% confidence interval, ICC: intraclass correlation coefficient, k: kappa statistic, kw: weighted kappa statistic, LBP: low-back pain, L: left, MS: motion segment, PA: percentage agreement, R: right, SD: standard deviation.

ARTICLE IN PRESS

Five physicians specialized in manual medicine

20%, ICC 0.30; PA 20%, ICC 0.15; PA 23%, ICC

E. van Trijffel et al. / Manual Therapy 10 (2005) 256–269

Richter and Lawall (1993) 35 patients with deep back pain. Rehabilitation Clinic, Germany

20%, ICC 0.32; PA 33%, ICC 0.38; PA 33%, ICC

263

ARTICLE IN PRESS 264

E. van Trijffel et al. / Manual Therapy 10 (2005) 256–269

tative patients. Their estimates could have been biased due to low prevalence. Assessing mobility of motion segment C1–C2 reached at least a fair level of reliability in five studies (DeBoer et al., 1985; Fjellner et al., 1999; Scho¨ps et al., 2000; Smedmark et al., 2000; Pool et al., 2004). Examination of motion segment C2–C3 yielded fair to substantial values of kappa statistics in three studies (Scho¨ps et al., 2000; Smedmark et al., 2000; Pool et al., 2004). 3.2. Lumbar spine Data on estimates of inter-examiner reliability of passive assessment are presented in Table 2 (last column). Inter-examiner reliability for the lumbar spine ranged from poor to substantial. Overall, reliability was poor to fair. Two studies (Maher and Adams, 1994; Hicks et al., 2003) fulfilled all criteria for external validity. Hicks et al. (2003) showed poor to fair reliability among four pretrained examiners making judgements on antero-posterior mobility of vertebrae L1–L5 with subjects in prone position. Using this same assessment procedure, Maher and Adams (1994) did not find acceptable ICC values. The study by Hicks et al. (2003) fulfilled the criteria for internal validity. Systematic error could have biased their estimates. In three studies (Richter and Lawall, 1993; Inscoe et al., 1995; Strender et al., 1997a), stability of characteristics during research was not likely. Richter and Lawall (1993) reported reliability among five physicians ranging from slight to substantial. Strender et al. (1997a) calculated substantial values of weighted kappa statistics for two physiotherapists judging mobility of motion segments L4–L5 and L5–S1 with a sidelying (hips and knees flexed) technique described by Kaltenborn. Substantial reliability was shown by Strender et al. (1997a). Their estimates could have been biased due to low prevalence (L4–L5) and systematic error (L5–S1). The lowest levels of reliability, with predominantly negative values of kappa statistics, were reached by Mootz et al. (1989) for evaluation of fixations in students of chiropractic. Prevalence bias due to limited variation could have influenced their results. Chiropractic seated motion palpation for intervertebral fixations consistently yielded poor to fair inter-examiner reliability in three studies (Boline et al., 1988; Mootz et al., 1989; Keating et al., 1990).

4. Discussion In this systematic review, inter-examiner reliability of passive assessment of segmental intervertebral motion in the cervical and lumbar spine ranged from poor to

substantial. However, overall, reliability was poor to fair. Studies addressing reliability are conducted to evaluate consistency of measurements and to quantify measurement error within or between examiners (Bartko and Carpenter, 1976; Brennan and Silman, 1992; Rothstein and Echternach, 1993; Bland and Altman, 1996a, b; Bruton et al., 2000; Streiner and Norman, 2003). A repeated-measures design consists of one assessment of all subjects by two or more examiners to determine inter-examiner reliability (Haas, 1995a). Inter-examiner reliability reflects a profession’s performance (Haas, 1995a). This systematic review was conducted to contribute resolving uncertainty over consistency among manual practitioners in assessing passive intervertebral motion in the spine. This systematic review has several limitations. In our experience, reliability studies were poorly indexed in databases. The main reason for this may be the inconsistent terminology used in reliability research. In addition, we limited our electronic search for relevant studies to MEDLINE and CINAHL. A quick scan in EMBASE showed only duplicate citation postings. In conclusion, although much effort was put in reference tracing and hand searching, it is not impossible that eligible studies were missed. Furthermore, unpublished studies were not included. Publication bias can form a real threat to internal validity of systematic reviews of reliability studies. Quality assessment was performed by using a criteria list mainly derived from the assessment of diagnostic accuracy studies. No evidence is available on whether these items also apply in the context of reliability. Empirical evidence of bias, especially concerning blinding of examiners and stability of characteristics during research, is lacking. Finally, assigning value labels for ranges of kappa (k) statistics was done in accordance with Landis and Koch (1977). As stated by these authors, this classification is an arbitrary one. Others have questioned its appropriateness (Brennan and Silman, 1992; Lantz, 1997). Using another classification may have yielded different results. Only four (Maher and Adams, 1994; Smedmark et al., 2000; Hicks et al., 2003; Pool et al., 2004) out of 19 included studies used representative patients as participants. Estimates of reliability of a test procedure are intimately linked with the population it was used in (Streiner and Norman, 2003). In order to assure external validity, it is necessary to include patients with neck and low-back pain that are likely to undergo passive motion assessment procedures in daily practice (Rothstein and Echternach, 1993). This issue also deals with the essence of the concept of reliability, for reliability can only exist when individuals vary in the characteristic under study like symptomatic subjects most likely do (Streiner and Norman, 2003). We note that characteristics of representative patients may differ substantially for the

ARTICLE IN PRESS E. van Trijffel et al. / Manual Therapy 10 (2005) 256–269

265

Table 3 Methodological quality scores of included studies (grouped in alphabetical order according to spinal region) First author (year)

External validity

Internal validity

1

2

3

4

5

6

7

8

Statistical methods 9 10 11

Cervical spine Christensen et al. (2002) DeBoer et al. (1985) Fjellner et al. (1999) Haas et al. (1995b) Mior et al. (1985) Pool et al. (2004) Scho¨ps et al. (2000) Smedmark et al. (2000) Strender et al. (1997b)

N N N N N Y N Y N

? Y Y N N ? ? Y Y

Y Y Y Y Y N N Y Y

? ? N ? Y ? ? Y Y

N N ? ? N ? ? ? ?

Y ? N ? ? ? Y ? N

Y Y Y Y ? Y ? Y Y

N Y Y Y Y Y Y Y Y

Y Y ? Y Y Y Y Y Y

N ? ? ? Y ? ? ? N

Y ? Y ? Y ? ? ? N

Lumbar spine Bergstro¨m and Courtis (1986) Boline et al. (1988) Gonella et al. (1982) Hicks et al. (2003) Inscoe et al. (1995) Keating et al. (1990) Maher and Adams (1994) Mootz et al. (1989) Richter and Lawall (1993) Strender et al. (1997a)

N ? N Y N N Y N ? ?

N Y Y Y Y Y Y Y ? Y

Y N N Y Y Y Y Y Y Y

? ? ? Y N ? ? ? N N

N ? N ? N ? ? N N ?

? N ? ? N ? N ? ? N

N Y ? Y Y Y Y Y Y Y

Y Y Y Y Y Y Y Y Y Y

? Y N ? ? Y ? Y Y ?

? ? ? ? ? ? ? N ? N

? ? ? N ? ? ? Y ? N

Numbers correspond to questions in the quality criteria list (Box 2). Note: Y: yes, N: no, ?: unclear.

various health care systems depending on the level of direct accessibility of practitioners. The need to use symptomatic participants has also been emphasized by other reviewers (Keating, 1989; Panzer, 1992; Hestbœk and Leboeuf-Yde, 2000; Huijbregts, 2002). From evidence of two studies (Smedmark et al., 2000; Pool et al., 2004), we found that reliability tended to be higher when representative neck patients were examined. With regard to internal validity, only three studies (Strender et al., 1997b; Smedmark et al., 2000; Hicks et al., 2003) satisfied both criteria of blinding of examiners to each others’ results and stability of joint mobility during research. Estimates of reliability can only be valid when the characteristic under study does not change during research, otherwise true reliability will be underestimated (Rothstein and Echternach, 1993). Where passive motion assessment is concerned, stability of biomechanical properties of connective tissue during the research process forms a key issue. These properties are susceptible to change as a result of natural variation over time or mobilizing effects of the test procedure itself (Rothstein and Echternach, 1993). None of the included studies explicitly dealt with this issue in their design. In the majority of cases, study protocol was poorly reported. Items such as number of tests, number of movement repetitions, forces applied in end-position, motion directions, and time intervals should be con-

sidered and described thoroughly. Some researchers post hoc discussed the possibility of changes in mobility as a result of the assessment procedure (Inscoe et al., 1995; Haas et al., 1995b; Strender et al., 1997a; Fjellner et al., 1999). One of the excluded studies (Binkley et al., 1995) used a Latin square design to correct for systematic differences in characteristics induced by the test. Hestbœk and Leboeuf-Yde (2000), Huijbregts (2002) also recognized the importance of stable characteristics, but they did not use this as a quality criterion. In one internally valid study (Smedmark et al., 2000), fair to moderate reliability was consistently shown when representative patients were examined. Currently, kappa (k) is the statistic of choice for analysing inter-examiner reliability with nominal data (Cohen, 1960; Altman, 1991; Haas, 1991a). Most of our included studies appropriately used kappa statistics. However, the interpretation of kappa is not straightforward. Feinstein and Cicchetti (1990) described two paradoxes of kappa by examining cross-tabulations. The first paradox concerns kappa taking lower values in case of substantial symmetrical imbalance in marginal totals and high percentage agreement. This situation, called limited variation in presence or absence of a characteristic, makes kappa susceptible for prevalence bias (Thompson and Walter, 1988; Brennan and Silman, 1992). In the second paradox, kappa overestimates in

ARTICLE IN PRESS 266

E. van Trijffel et al. / Manual Therapy 10 (2005) 256–269

case of asymmetrical imbalance in marginal totals, which is likely when examiners systematically disagree. As a consequence, comparing kappa values from different studies, let alone pooling them, is unjustifiable (Thompson and Walter, 1988; Byrt et al., 1993). At least a critical appraisal of possible prevalence bias and systematic bias is required. For this purpose, raw data, like cross-tabulations, are indispensable (Brennan and Silman, 1992). The majority of studies reviewed did not adequately report statistical data. Appropriate statistical techniques for pooling kappa statistics recognizing the problems with prevalence and systematic error are not available. Due to this fact as well as the strong clinical heterogeneity across studies, we did not perform a metaanalysis to summarize reliability. The concept of prevalence bias is closely related to the choice of study population. This bias is likely when a homogeneous (e.g. asymptomatic) sample is used. In harmony with the need to include representative patients as participants, as stated earlier, careful attention to the choice for a heterogeneous study population will decrease the risk of prevalence bias (Feinstein and Cicchetti, 1990). Meade et al. (2000) proposed phi (F) as a chance-independent statistic to overcome prevalence problems with kappa. In the two studies (Mootz et al., 1989; Christensen et al., 2002) that showed the lowest levels of reliability, non-representative participants were used and estimates were biased due to low prevalence. With respect to reducing systematic error between examiners, several authors have suggested enhanced standardization of procedures to reduce error and improve reliability (Panzer, 1992; Strender et al., 1997a, b; Anson et al., 2003; Streiner and Norman, 2003). Others (Huijbregts, 2002; Hicks et al., 2003) have argued that training of examiners diminishes external validity. We found no relevant differences in reliability for pre-trained examiners. Eight studies determined estimates of intra-examiner reliability but acceptable levels were not reached. We did not conduct a separate appraisal of internal validity of the intra-examiner reliability designs within our quality assessment. In an intra-examiner reliability design, each examiner performs repeated measurements of each subject. Error within examiners constitutes an integral source of the total amount of error between examiners (Streiner and Norman, 2003). Intra-examiner reliability can be computed from an inter-examiner design whilst still avoiding specific problems with blinding, consistency of error, and instability of characteristics under study (Haas, 1995a). In diagnostic accuracy studies, availability of clinical information from participants to examiners before executing the test has been shown to increase sensitivity (Whiting et al., 2004). This distortion is known as clinical review bias. In the context of inter-examiner reliability research reflecting daily practice of manual

practitioners, this type of bias is likely to occur because the same examiner both gathers clinical information and performs physical examination. Using the QUADAS tool, examiners are allowed to have clinical information as long as this information reflects daily practice (Whiting et al., 2003). In case of analysing reliability with kappa statistics, prior knowledge and expectation may influence calculations (Mior et al., 1985; Feinstein and Cicchetti, 1990; French et al., 2000). Furthermore, we argue that not blinding examiners to clinical characteristics will reduce the view on reliability of the test procedure itself. Therefore, in our quality assessment, we judged the presence of fully blinded examiners as a positive feature. Seven of the included studies (Bergstro¨m and Courtis, 1986; Mootz et al., 1989; Keating et al., 1990; Maher and Adams, 1994; Inscoe et al., 1995; Haas et al., 1995b; Christensen et al., 2002) used marking of spinal levels. To date, results of inter-examiner reliability studies on palpating and nominating spinal levels have been inconclusive (McKenzie and Taylor, 1997; Downey et al., 1999; Billis et al., 2003; Downey et al., 2003). It is unclear whether this pre-conditional skill contributes to another source of error in passive intervertebral motion assessment. We consistently found at least fair levels of interexaminer reliability for assessment of motion segments C1–C2 and C2–C3, but low values of reliability estimates were found for chiropractic lumbar motion palpation. We could not discover other explanations for heterogeneity in reliability. Passive assessment of segmental intervertebral motion in the spine is part of the diagnostic clinical expertise of manual practitioners to guide decisions on a therapeutic strategy for patients with neck and low-back pain (Jull et al., 1994; Maher and Adams, 1994). Hypomobility indicates mobilizing interventions, while hypermobility calls for a stabilizing approach (Hicks et al., 2003). Clinical rationales rest on segmental approaches (Breen, 1992). Evidence collected from studies included in this systematic review indicates that inter-examiner reliability of passive intervertebral motion assessment of the cervical and lumbar spine is low. However, this review has also exposed some shortcomings of research in this area. Only two studies (Smedmark et al., 2000; Hicks et al., 2003) proved to be externally and internally valid, of which one (Smedmark et al., 2000) found fair to moderate reliability. There is a need for new and valid studies to be conducted. Some evidence suggests that passive intervertebral motion assessment can be accurate (Humphreys et al., 2004). In a randomised diagnostic trial on the other hand, Haas et al. (2003) did not find better outcomes for neck pain patients treated with chiropractic manipulations after segmental end-play assessment. Hence, no final conclusions can be drawn yet regarding the clinical usefulness of passive

ARTICLE IN PRESS E. van Trijffel et al. / Manual Therapy 10 (2005) 256–269

spinal motion evaluation. Similarly, the contribution of this diagnostic intervention to the effectiveness of manipulative therapies remains unclear. 5. Conclusions In this systematic review, it was found that interexaminer reliability of passive assessment of segmental intervertebral motion in the cervical and lumbar spine by manual practitioners was low. However, most studies did not fulfil the criteria for external and internal validity. In general, reporting of study protocol and statistical data was inadequate. In addition, only a few of all possible assessment techniques have been investigated so far. We propose the following recommendations for future research:










include representative neck and low-back pain patients as participants that are likely to undergo the assessment procedure in daily practice, instead of students, volunteers, healthy individuals, or samples with a mix of symptomatic and non-symptomatic subjects; give careful consideration ensuring stability of joint mobility during research; determine intra-examiner reliability along in the process; present cross-tabulations when using kappa (k) statistics to allow for appraisal of prevalence bias and systematic bias; report the study by following the STARD statement.

Only when new and valid evidence emerges, uncertainty over diagnostic performance can be resolved and definitive conclusions can be drawn regarding the clinical usefulness of passive intervertebral motion assessment. Until then, questions remain about professional credibility and accountability of this diagnostic procedure within evidence-based clinical decision-making.

Acknowledgements We thank I.I. Riphagen, M.Sc., medical information specialist, Vrije University, Amsterdam, The Netherlands, for her help with the electronic literature search. References Altman DG. Practical statistics for medical research. London: Chapman & Hall; 1991. p. 403–9 (Chapter 14). Anson E, Cook C, Camacho C, Gwilliam B, Karakostas T. The use of an educational model in the improvement of student reliability in finding R1. The Journal of Manual and Manipulative Therapy 2003;11(4):204–12. Bartko JJ, Carpenter WT. On the methods and theory of reliability. The Journal of Nervous and Mental Disease 1976;163(5):307–17.

267

Bergstro¨m E, Courtis G. An inter- and intra-examiner reliability study of motion palpation of the lumbar spine in lateral flexion in the seated position. European Journal of Chiropractic 1986;34:121–41. Billis EV, Foster NE, Wright CC. Reproducibility and repeatability: errors of three groups of physiotherapists in locating spinal levels by palpation. Manual Therapy 2003;8(4):223–32. Binkley J, Stratford PW, Gill C. Interrater reliability of lumbar accessory motion mobility testing. Physical Therapy 1995;75(9): 786–95. Bland JM, Altman DG. Measurement error and correlation coefficients. British Medical Journal 1996a;313:41–2. Bland JM, Altman DG. Measurement error. British Medical Journal 1996b;313:744. Boline PD, Keating JC, Brist J, Denver G. Interexaminer reliability of palpatory evaluations of the lumbar spine. American Journal of Chiropractic Medicine 1988;1(1):5–11. Borghouts AJ, Janssen HJ, Koes BW, Muris JWM, Metsemakers JFM, Bouter LM. The management of chronic pain in general practice: a retrospective study. Scandinavian Journal of Primary Health Care 1999;17:215–20. Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, Irwig LM, Lijmer JG, Moher D, Rennie D, de Vet HCW. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD Initiative. Clinical Chemistry 2003a;49(1):1–6. Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, Irwig LM, Moher D, Rennie D, de Vet HCW, Lijmer JG. The STARD statement for reporting studies of diagnostic accuracy: explanation and elaboration. Clinical Chemistry 2003b;49(1):7–18. Breen A. The reliability of palpation and other diagnostic methods. Journal of Manipulative and Physiological Therapeutics 1992; 15(1):54–6. Brennan P, Silman A. Statistical methods for assessing observer variability in clinical measures. British Medical Journal 1992; 304:1491–4. Bruton A, Conway JH, Holgate ST. Reliability: what is it and how is it measured? Physiotherapy 2000;86(2):94–9. Byrt T, Bishop J, Carlin JB. Bias, prevalence and kappa. Journal of Clinical Epidemiology 1993;46(5):423–9. Christensen HW, Vach W, Vach K, Manniche C, Haghfelt T, Hartvigsen L, Høilund-Carlsen PF. Palpation of the upper thoracic spine: an observer reliability study. Journal of Manipulative and Physiological Therapeutics 2002;25(5):285–92. Cohen J. A coefficient of agreement for nominal scales. Educational and Psychological Measurement 1960;20:37–46. DeBoer KF, Harmon R, Tuttle CD, Wallace H. Reliability study of detection of somatic dysfunction in the cervical spine. Journal of Manipulative and Physiological Therapeutics 1985;8(1): 9–16. Downey BJ, Taylor NF, Niere KR. Manipulative physiotherapists can reliably palpate nominated lumbar spinal levels. Manual Therapy 1999;4(3):151–6. Downey B, Taylor N, Niere K. Can manipulative physiotherapists agree on which lumbar level to treat based on palpation? Physiotherapy 2003;89(2):74–81. Egger M, Davey Smith G, O’Rourke K. Rationale, potentials, and promise of systematic reviews. In: Egger M, Davey Smith G, Altman DG, editors. Systematic reviews in health care: metaanalysis in context. 2nd ed. London: BMJ Publishing Group; 2001. p. 3–19 (Chapter 1). Faas A, Chavannes AW, Koes BW, Van den Hoogen JMM, Mens JMA, Smeele LJM, Romeijnders ACM, Van den Laan JR. Dutch College of General Practitioners Standard ‘‘low back pain’’. Huisarts en Wetenschap 1996;39(1):18–31. Feinstein AR, Cicchetti DV. High agreement but low Kappa: I. The problems of two paradoxes. Journal of Clinical Epidemiology 1990;43(6):543–9.

ARTICLE IN PRESS 268

E. van Trijffel et al. / Manual Therapy 10 (2005) 256–269

Fjellner A, Bexander C, Faleij R, Strender L-E. Interexaminer reliability in physical examination of the cervical spine. Journal of Manipulative and Physiological Therapeutics 1999;22(8):511–6. French SD, Green S, Forbes A. Reliability of chiropractic methods commonly used to detect manipulable lesions in patients with chronic low-back pain. Journal of Manipulative and Physiological Therapeutics 2000;23(4):231–8. Gonella C, Paris SV, Kutner M. Reliability in evaluating passive intervertebral motion. Physical Therapy 1982;62(4):436–44. Haas M. Statistical methodology for reliability studies. Journal of Manipulative and Physiological Therapeutics 1991a;14(2): 119–32. Haas M. The reliability of reliability. Journal of Manipulative and Physiological Therapeutics 1991b;14(3):199–208. Haas M. How to evaluate intraexaminer reliability using an interexaminer reliability study design. Journal of Manipulative and Physiological Therapeutics 1995a;18(1):10–5. Haas M, Raphael R, Panzer D, Peterson D. Reliability of manual endplay palpation of the thoracic spine. Chiropractic Technique 1995b;7(4):120–4. Haas M, Groupp E, Panzer D, Partna L, Lumsden S, Aickin M. Efficacy of cervical endplay assessment as an indicator for spinal manipulation. Spine 2003;28(11):1091–6. Hanten WP, Olson SL, Ludwig GM. Reliability of manual mobility testing of the upper cervical spine in subjects with cervicogenic headache. The Journal of Manual and Manipulative Therapy 2002; 10(2):76–82. Hawk C, Phongphua C, Bleecker J, Swank L, Lopez D, Rubley T. Preliminary study of the reliability of assessment procedures for indications for chiropractic adjustments of the lumbar spine. Journal of Manipulative and Physiological Therapeutics 1999; 22(6):382–9. Hestbœk L, Leboeuf-Yde C. Are chiropractic tests for the lumbopelvic spine reliable and valid? A systematic critical literature review. Journal of Manipulative and Physiological Therapeutics 2000;23(4):258–75. Hicks GE, Fritz JM, Delitto A, Mishock J. Interrater reliability of clinical examination measures for identification of lumbar segmental instability. Archives of Physical Medicine and Rehabilitation 2003;84:1858–64. Huijbregts PA. Spinal motion palpation: a review of reliability studies. The Journal of Manual and Manipulative Therapy 2002;10(1): 24–39. Humphreys BK, Delahaye M, Peterson CK. An investigation into the validity of cervical spine motion palpation using subjects with congenital block vertebrae as a ‘gold standard’. BMC Musculoskeletal Disorders 2004;5:19. Inscoe EL, Witt PL, Gross MT, Mitchell RU. Reliability in evaluating passive intervertebral motion of the lumbar spine. The Journal of Manual and Manipulative Therapy 1995;3(4):135–43. Johnston WL, Blum GA, Hendra JL, Neff DR, Rosen ME. Passive gross motion testing: Part III. Examiner agreement on selected subjects. Journal of the American Osteopathic Association 1982; 81(5):309–13. Jull G, Bullock M. A motion profile of the lumbar spine in an ageing population assessed by manual examination. Physiotherapy Practice 1987;3:70–81. Jull G, Treleaven J, Versace G. Manual examination: is pain provocation a major cue for spinal dysfunction? Australian Journal of Physiotherapy 1994;40(3):159–65. Jull G, Zito G, Trott P, Potter H, Shirley D, Richardson C. Interexaminer reliability to detect painful upper cervical joint dysfunction. Australian Journal of Physiotherapy 1997;43(2):125–9. Kaltenborn F, Lindahl O. Reproducerbarheten vid ro¨relseunderso¨kning av enskilda kotor. Lakartidningen 1969;66(10):962–5.

Keating JC. Interexaminer reliability of motion palpation of the lumbar spine: a review of quantitative literature. American Journal of Chiropractic Medicine 1989;2(3):107–10. Keating JC, Bergmann TF, Jacobs GE, Finer BA, Larson K. Interexaminer reliability of eight evaluative dimensions of lumbar segmental abnormality. Journal of Manipulative and Physiological Therapeutics 1990;13(8):463–70. Kramer MS, Feinstein AR. Clinical biostatistics LIV. The biostatistics of concordance. Clinical Pharmacology and Therapeutics 1981; 29(1):111–23. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159–74. Lantz CA. Application and evaluation of the Kappa statistic in the design of chiropractic clinical research. Journal of Manipulative and Physiological Therapeutics 1997;20(8):521–8. Leboeuf C, Gardner V, Carter AL, Scott TA. Chiropractic examination procedures: a reliability and consistency study. Journal of the Australian Chiropractors’ Association 1989;19(3):101–4. Lijmer JG, Mol BW, Heisterkamp S, Bonsel GJ, Prins MH, van der Meulen JHP, Bossuyt PMM. Empirical evidence of design-related bias in studies of diagnostic tests. Journal of the American Medical Association 1999;282(11):1061–6. Love RM, Brodeur RR. Inter- and intra-examiner reliability of motion palpation for the thoracolumbar spine. Journal of Manipulative and Physiological Therapeutics 1987;10(1):1–4. Maclure M, Willett WC. Misinterpretation and misuse of the Kappa statistic. American Journal of Epidemiology 1987;126(2):161–9. Maher C, Adams R. Reliability of pain and stiffness assessments in clinical manual lumbar spine examination. Physical Therapy 1994;74(9):801–11. Maher C, Latimer J. Pain or resistance—the manual therapists’ dilemma. Australian Journal of Physiotherapy 1992;38(4):257–60. McKenzie AM, Taylor NF. Can physiotherapists locate lumbar spinal levels by palpation? Physiotherapy 1997;83(5):235–9. Meade MO, Cook RJ, Guyatt GH, Groll R, Kachura JR, Bedard M, Cook DJ, Slutsky AS, Stewart TE. Interobserver variation in interpreting chest radiographs for the diagnosis of acute respiratory distress syndrome. American Journal of Respiratory and Critical Care Medicine 2000;161:85–90. Meijne W, Van Neerbos K, Aufdemkampe G, Van der Wurff P. Intraexaminer and interexaminer reliability of the Gillet test. Journal of Manipulative and Physiological Therapeutics 1999;22(1):4–9. Mior SA, King RS, McGregor M, Bernard M. Intra and interexaminer reliability of motion palpation in the cervical spine. The Journal of the Canadian Chiropractic Association 1985;29(4):195–8. Mootz RD, Keating JC, Kontz HP, Milus TB, Jacobs GE. Intra- and interobserver reliability of passive motion palpation of the lumbar spine. Journal of Manipulative and Physiological Therapeutics 1989;12(6):440–5. Nachemson A, Waddell G, Norlund AI. Epidemiology of neck and low back pain. In: Nachemson A, Jonsson E, editors. Neck and back pain: the scientific evidence of causes, diagnosis, and treatment. Philadelphia, PA: Lippincot, Williams and Wilkins; 2000. p. 165–87 (Chapter 8). Nansel DD, Peneff AC, Jansen RD, Cooperstein R. Interexaminer concordance in detecting joint-play asymmetries in the cervical spines of otherwise asymptomatic subjects. Journal of Manipulative and Physiological Therapeutics 1989;12(6):428–33. Panzer DM. The reliability of lumbar motion palpation. Journal of Manipulative and Physiological Therapeutics 1992;15(8):518–24. Pool JJ, Hoving JL, de Vet HC, van Mameren H, Bouter LM. The interexaminer reproducibility of physical examination of the cervical spine. Journal of Manipulative and Physiological Therapeutics 2004;27(2):84–90.

ARTICLE IN PRESS E. van Trijffel et al. / Manual Therapy 10 (2005) 256–269 Potter NA, Rothstein JM. Intertester reliability for selected clinical tests of the sacroiliac joint. Physical Therapy 1985;65(11):1671–5. Richter T, Lawall J. Zur Zuverla¨ssigkeit manualdiagnostischer Befunde. Manuelle Medizin 1993;31:1–11. Rothstein JM, Echternach JL. Primer on measurement: an introductory guide to measurement issues. Alexandria, VA: American Physical Therapy Association; 1993. p. 73–85. Schoensee SK, Jensen G, Nicholson G, Gossman M, Katholi C. The effect of mobilization on cervical headaches. Journal of the Orthopaedic and Sports Physical Therapy 1995;21(4):184–96. Scho¨ps P, Pfingsten M, Siebert U. Reliabilita¨t manualmedizinischer Untersuchungstechniken an der Halswirbelsa¨ule. Studie zur Qualita¨tssicherung in der manuellen Diagnostik. Zeitschrift fu¨r Orthopa¨die und Ihre Grenzgebiete 2000;138:2–7. Seffinger MA, Najm WI, Mishra SI, Adams A, Dickerson VM, Murphy LS, Reinsch S. Reliability of spinal palpation for diagnosis of back and neck pain. A systematic review of the literature. Spine 2004;29(19):E413–25. Smedmark V, Wallin M, Arvidsson I. Inter-examiner reliability in assessing passive intervertebral motion of the cervical spine. Manual Therapy 2000;5(2):97–101. Streiner DL, Norman GR. Health measurement scales: a practical guide to their development and use. 3rd ed. Oxford: Oxford University Press; 2003. p. 126–52 (Chapter 8). Strender L-E, Sjo¨blom A, Sundell K, Ludwig R, Taube A. Interexaminer reliability in physical examination of patients with low back pain. Spine 1997a;22(7):814–20.

269

Strender L-E, Lundin M, Nell K. Interexaminer reliability in physical examination of the neck. Journal of Manipulative and Physiological Therapeutics 1997b;20(8):516–20. Thompson WD, Walter SD. A reappraisal of the kappa coefficient. Journal of Clinical Epidemiology 1988;41(10):949–58. Troyanovich SJ, Harrison DD. Motion palpation: it’s time to accept the evidence. Journal of Manipulative and Physiological Therapeutics 1998;21(8):568–71. Van Dillen LR, Sahrmann SA, Norton BJ, Caldwell CA, Fleming DA, McDonnell MK, Woolsey NB. Reliability of physical examination items used for classification of patients with low back pain. Physical Therapy 1998;78(9):979–88. Van Suijlekom HA, de Vet HCW, van den Berg SGM, Weber WEJ. Interobserver reliability in physical examination of the cervical spine in patients with headache. Headache 2000;40:581–6. Viikari-Juntura E. Interexaminer reliability of observations in physical examinations of the neck. Physical Therapy 1987;67(10):1526–32. Whiting P, Rutjes AWS, Reitsma JB, Bossuyt PMM, Kleijnen J. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Medical Research Methodology 2003;3(1):25. Whiting P, Rutjes AWS, Reitsma JB, Glas AS, Bossuyt PMM, Kleijnen J. Sources of variation and bias in studies of diagnostic accuracy. A systematic review. Annals of Internal Medicine 2004; 140(3):189–202.

ARTICLE IN PRESS

Manual Therapy 10 (2005) 270–280 www.elsevier.com/locate/math

Original article

A comparison of symptomatic and asymptomatic office workers performing monotonous keyboard work—1: Neck and shoulder muscle recruitment patterns Grace P.Y. Szetoa,b,, Leon M. Strakerb, Peter B. O’Sullivanb a

Department of Rehabilitation Sciences, Hong Kong Polytechnic University, Hong Kong SAR, PR China b School of Physiotherapy, Curtin University of Technology, Perth 6845, Australia Received 7 January 2004; received in revised form 30 November 2004; accepted 4 January 2005

Abstract Work-related neck and upper limb disorders (WRNULD) are common problems among office workers who use computers intensively and maintain prolonged static postures. These disorders have often been attributed to result from sustained muscle activity in the neck–shoulder musculature. The present study examined whether symptomatic subjects exhibited the same muscle activity patterns as asymptomatic controls when they performed a prolonged computer task under the same conditions. Surface electromyography (EMG) of four major neck–shoulder muscles were compared between a Case Group (n ¼ 23) and a Control Group (n ¼ 20) of female office workers. The Case Group had higher activity in the right upper trapezius (UT) while the Control Group had more symmetrical muscle activity between left and right UT. The Case subjects could also be differentiated into ‘‘High Discomfort’’ and ‘‘Low Discomfort’’ sub-groups based on their discomfort scores. The High Discomfort Group had significantly higher right UT activity compared to the Low Discomfort and Control Groups. Results suggested that symptomatic individuals had altered muscle recruitment patterns that persisted throughout the sustained occupational task, while discomfort increased with timeat-task. These findings indicate that altered muscle recruitment patterns observed in the symptomatic subjects preceded the onset of task discomfort, and this finding may have important implications for the etiology of WRNULD. r 2005 Elsevier Ltd. All rights reserved. Keywords: Work-related neck and upper limb disorders; Computer use; Electromyography; Motor control

1. Introduction Work-related neck and upper limb disorders (WRNULD) are common problems in office workers, especially among those who are intensive computer users (Kamwendo et al., 1991; Bergqvist, 1993; Bernard et al., 1994; Tittiranonda et al., 1999). A high prevalence of neck and upper limb complaints in computer using office workers has been consistently reported in both eastern and western countries (Kamwendo et al., 1991; Corresponding author. Department of Rehabilitation Sciences, Hong Kong Polytechnic University, Hong Kong SAR, China. Tel.: +852 27666706; fax: +852 23308656. E-mail address: [email protected] (G.P.Y. Szeto).

1356-689X/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2005.01.004

Bergqvist, 1993; Bernard et al., 1994; Siu and Chan, 1998; Tittiranonda et al., 1999). In a survey of over 600 office workers in Hong Kong, the 12-month period prevalence of WRNULD associated with computer use was over 56% (Siu and Chan, 1998). The proximal neck and shoulder region has consistently had higher prevalence rates than the distal elbow and wrist-hand region (Westgaard et al., 1993; Bernard et al., 1994); hence the present study has focused on neck–shoulder problems. The worldwide trend is for people to use computers for longer periods daily, due to increased computerbased tasks at work as well as increased computer-based leisure activities. Static posture associated with computer work has been identified as a major occupational risk

ARTICLE IN PRESS G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 270–280

factor (Schuldt et al., 1986; Kamwendo et al., 1991; Bernard et al., 1994; Aara˚s et al., 1997; Tittiranonda et al., 1999). In particular, static neck and upper limb postures associated with computer use have been linked with sustained low-level muscle activity in neck– shoulder stabilizers. This may impose substantial biomechanical strain on the musculoskeletal system (Bansevicius et al., 1997; Ha¨gg and Astrom, 1997; Cooper and Straker, 1998; Kleine et al., 1999). Ergonomics studies have often examined the muscle load in healthy painfree subjects and assumed that higher levels of muscle activity during work represented higher risks for developing musculoskeletal discomfort (Aara˚s et al., 1997; Bansevicius et al., 1997; Cooper and Straker, 1998; Kleine et al., 1999). Studies that compared EMG activities between symptomatic and asymptomatic workers have either found no significant differences (Jensen et al., 1993; Nordander et al., 2000; Roe et al., 2001) or only weak associations between high muscle activities and increased discomfort (Ha¨gg and Astrom, 1997; Westgaard et al., 2001). There is mounting evidence that altered muscle recruitment patterns are associated with pain disorders of the low back (Hodges and Richardson, 1996; Edgerton et al., 1997), shoulder (Kilber, 1998), and cervical spine (Bansevicius and Sjaastad, 1996; Jull et al., 1999; Jull, 2000; Nederhand et al., 2000; Sterling et al., 2001). However, in most of these studies muscle recruitment has been studied either at rest, during specific limb movements or while carrying out specific spinal maneuvers, rather than during sustained functional work activities. Furthermore, it is not clear whether the altered muscle recruitment patterns observed, occur secondary to pain or whether they are a mechanism for ongoing strain and pain in these disorders. It is generally agreed that the etiology of WRNULD is multifactorial, and the interactions of intrinsic and extrinsic factors may simultaneously operate within each person (Westgaard, 2000; Kumar, 2001; Forde et al., 2002). Interactions of these various factors may cause different reactions in individuals when they are exposed to different physical and/or psychosocial stresses (Marras et al., 2000; Waersted, 2000). Previous ergonomic research has tended to focus on the influence of external factors such as changes in workstation setting or changes in work tasks performed (Aara˚s et al., 1997; Cooper and Straker, 1998; Kleine et al., 1999). These studies have often assumed that all individuals would respond in the same way. The lack of consistent evidence regarding the etiology of WRNULD may be due to a lack of consideration for these intrinsic factors affecting individual responses. The aim of the present study was to investigate the muscle activities in the neck and shoulder regions in symptomatic and asymptomatic office workers while they performed a prolonged and standardized computer

271

task under standardized conditions. As the external physical factors were standardized, any differences in muscle activity patterns may indicate intrinsic differences between individuals. It was hypothesized that investigating muscle activity patterns in this manner might provide an insight into the underlying mechanisms involved in WRNULD.

2. Subjects and method 2.1. Subjects A total of 43 female office workers were recruited as subjects through convenience sampling. Subjects worked a minimum of 4 h on computers daily, mainly text-typing duties. Those with past traumatic injuries or surgical interventions in their neck and upper limb regions were excluded. Twenty-three subjects were allocated to the Case Group (symptomatic), and 20 to the Control Group (asymptomatic). The allocation of subjects into groups was based on the subject’s past and present musculoskeletal discomforts information gathered from a modified version of the Standardized Nordic Questionnaire (Kuorinka et al., 1987). Case subjects had neck and arm discomfort related to computer use which lasted more than 3 months in the past year and was present in the past 7 days as well as on the day of testing. Control subjects had no or minimal discomfort on the day of testing, and had no discomfort in the past 7 days. If they had reported discomfort in the past 12 months, it was of a short duration (o3 months) and had resolved at least 3 months prior to participation. The experimental procedures were explained to each subject and informed consent was obtained before the experiment began. Subjects could have withdrawn if they felt intolerable discomfort anytime during the testing procedures but none actually withdrew from the study. The study was approved by the human research ethics committees of Hong Kong Polytechnic University and Curtin University. Table 1 shows the general characteristics of the two groups. The subjects were reasonably matched in terms of physical build, handedness and work profiles, except for a significant difference in mean age. 2.2. Variables The independent variables were group (Case vs. Control), side (left vs. right) and time (5 repeated measures during 1-h typing task). The dependent variables were muscle electrical activity and discomfort. Kinematics were also collected and are reported in a separate paper (Szeto et al., 2005).

ARTICLE IN PRESS 272

G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 270–280

Table 1 Subject profiles for the Case and Control Groups

Subjects’ Background Information Age (years) [mean (SD; range)] Body height (cm) [mean (SD; range)] Body weight (kg) [mean (SD; range)] Hand dominancy [count (expected count)] Work experience [mode (range)] Working hours per week [mean (SD; range)] Computer usage at work in hrs/day [mode (range)] Keyboard use in hrs/day [mode (range)] Mouse use in hrs/day [mode (range)] Previous typing training [count (expected count)] Typing method adopted [count (expected count)]

Past Discomfort Duration of discomfort in Past 12 months [mode (range)]

Prevalence of discomfort in Past 7 days [count (expected count)]

*

Case Group (n ¼ 23)

Control Group (n ¼ 20)

Group difference statistics

x ¼ 36:0 (4.6; 29–46) x ¼ 158:3 (6.4; 143.0–170.2) x ¼ 53:4 (12.2; 45.5–98.0) Left ¼ 0 (0.5) Right ¼ 23 (22.5) Mode ¼ 43 years (0–6 months–43 years) x ¼ 42:5 (6.9; 20–60) Mode ¼ 4–6 h (2–4 h–48 h) Mode ¼ 2–4 h (2–4 h–48 h) Mode ¼ 0–2 h (0–2 h–4–6 h) Yes ¼ 15 (15.3) No ¼ 8 (7.7) Proper touchtype ¼ 21 (21.9) Certain fingers only ¼ 1 (0.5) Others ¼ 1 (0.5)

x ¼ 31:3 (7.2; 21–48) x ¼ 157:2 (6.7; 139.0–167.6) x ¼ 52:0 (5.3; 43.0–60.0) Left ¼ 1 (0.5) Right ¼ 18 (18.5) Mode ¼ 43 years (0–6 months–43 years) x ¼ 43:0 (4.1; 30–48) Mode ¼ 2–4 h 4–6 h (2–4 h–48 h) Mode ¼ 4–6 h (0–2 h–6–8 h) Mode ¼ 0–2 h (0–2 h–6–8 h) Yes ¼ 13 (12.7) No ¼ 6 (6.3) Proper touchtype ¼ 19 (18.1) Certain fingers only ¼ 0 (0.5) Others ¼ 0 (0.5)

t ¼ 2:53 P ¼ 0:017* t ¼ 0:48 P ¼ 0:638 t ¼ 0:44 P ¼ 0:661 w2 ¼ 1:62a P ¼ 0:204a Z ¼ 1:58b P ¼ 0:115 t ¼ 0:30 P ¼ 0:767 Z ¼ 1:32 P ¼ 0:187 Z ¼ 20:59 P ¼ 0:558 Z ¼ 0:32 P ¼ 0:746 w2 ¼ 0:05 P ¼ 0:826 w2 ¼ 2:49a

Mode ¼ 46 months (8–30 days–46 months) Yes ¼ 20 (13.1) No ¼ 3 (9.9)

Mode ¼ 0 day (0 day–o3 months)

Z ¼ 3:90 Po0:001**

Yes ¼ 4 (10.9) No ¼ 15 (8.1)

w2 ¼ 18:45 Po0:001**

P ¼ 0:288a

P-value significant at a ¼ 0.05 level; **P-value significance at a ¼ 0.001 level. a 1 cell or more have expected count o5. Likelihood ratio is used. b Mann–Whitney U test.

2.3. Muscle electrical activity The eight muscles studied were the bilateral cervical erector spinae (CES), upper trapezii (UT), lower trapezii (LT) and anterior deltoids (AD). These muscles were selected as they are the major stabilizing muscles of the neck and shoulder region during functional activities and they are accessible to surface EMG (Culham and Peat, 1993; Johnson et al., 1994; Kilber, 1998; Chaffin et al., 1999). For EMG measurement, eight pairs of bipolar Ag–AgCl (3 MTM Infant Red DotTM) surface electrodes of 15 mm diameter (3M Hong Kong Limited, Hong Kong) were placed on the eight muscles with an inter-electrode distance fixed at 20 mm. The locations of the electrodes on the eight muscles are presented in Table 2. The skin was carefully prepared by cleaning the located area with water, fine sand paper and 2% alcohol

(and shaved if necessary) before electrode placement. Electrode impedanceo2 kO was considered acceptable. The Noraxon Telemyo System (Noraxon USA Inc., USA) was used to capture EMG signals (intrinsic frequency of 1000 Hz and a bandwidth of 10–500 Hz). The raw EMG signals were channeled into the Vicon 370 system (Oxford Metrics Ltd., UK) as analogue signals sampled at 1920 Hz. EMG data was collected for a 60-s period for five trials (at the 5th, 20th, 35th, 50th and 60th min) during the 1-h typing session. Prior to the typing trials, EMG normalization procedures were carried out. These were based on those of Aara˚s et al. (1996), which have good demonstrated reliability. Each subject performed three trials of resisted isometric maximum voluntary contractions (MVC) and one trial of sub-maximal ‘‘ramp’’ contraction from 0–30% MVC for each muscle. EMG signals recorded during the typing trials were expressed as percentages of

ARTICLE IN PRESS G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 270–280

273

Table 2 Elements of electrode positions and muscle actions tested in normalization for the neck–shoulder muscles Muscle

Electrode position

Starting position

Muscle action and application of load

Cervical erector spinae (CES)

Distal:1 cm lateral to C5 spinous process

Head in upright position

Neck extension—against transducer at posterior occiput

Proximal: 20 mm above distal Upper trapezius (UT)

Midpoint between electrodes at mid-point between acromion and C7 spinous process

Arm in 01 flexion and abduction Scapula in neutral Elevation

Scapular elevation—against adjustable strap on acromioclavicular Joint

Lower trapezius (LT)

Distal: 2.5–3 cm lateral to T6 Proximal: at 451 parallel to muscle fibres and 20 mm above distal

As above

Scapular retraction—against transducer at the posterior aspect of scapula at lateral half of spine of scapula

Anterior deltoid (AD)

Midpoint between electrodes at 2 cm anterior to midpoint between acromion and deltoid tuberosity

Shoulder in 301 forward flexion, elbow in 751 flexion

Forward flexion of shoulder joint— transducer at just above elbow joint

the EMG activity during MVC (%MEMG). The force exerted in MVC was measured by a strain-gauge transducer connected to an adjustable strap (for UT), or an adjustable metal bar (for CES, LT and AD) for resisting the isometric muscle contraction. Each MVC was performed for a 5 s hold. The 30% MVC was determined from the highest value of the 3 MVC trials, and a line was drawn on the oscilloscope screen to trace the ‘‘ramp’’ from 0% to 30% MVC across 5 s. The ramp contraction was used to check the quality of muscle contraction. All the EMG signals were processed in a specially developed Labview (National InstrumentsTM, Austin, USA) program with a high-pass filter at 20 Hz, a lowpass filter at 200 Hz and notch filters at 50 and 60 Hz to reduce the noise levels. Then the signals were downsampled to 10 Hz root-mean-square (RMS) values. Muscle electrical activity during the typing task was analysed in terms of normalized %MEMG expressed as three levels of Amplitude Probability Distribution Function (APDF) (10th percentile (%le), 50th%le, 90th%le) (Jonsson, 1982). The use of APDF has developed from epidemiological research suggesting risk thresholds for occupational exposure of certain amplitudes at these levels (Jonsson, 1982; Jensen et al., 1993; Ha¨gg and Astrom, 1997; Nordander et al., 2000; Roe et al., 2001; Westgaard et al., 2001). We used APDF to enable comparison of our results with field studies and recommendations for occupational exposures. 2.4. Subjective discomfort Immediately after each EMG capture during the typing trial, subjects verbally rated their trial related discomfort in 10 upper body regions (left and right neck, upper back, shoulders, elbows, wrists/hands) on a

numerical scale of 0–10 with 0 ¼ no discomfort, 1 ¼ minimal discomfort and 10 ¼ extreme/intolerable discomfort. The boundaries of the various upper body regions were adopted from the Standardized Nordic Questionnaire (Kuorinka et al., 1987). The discomfort data were analysed in terms of the summed score (total score of all discomfort areas in a trial). 2.5. Controlled variables The workstation included a standard computer desk with an adjustable slide-out tray for keyboard and an adjustable height swivel chair with no arm rests (See Fig. 1). The subject was instructed to adjust the keyboard tray and the chair in order to assume a position of comfort, with hip, knee, and elbow joints approximately at 901. This posture is in line with the generally recommended ‘‘good posture’’ for computer use (Kroemer and Grandjean, 1997; Chaffin et al., 1999). The display screen height, distance and angle were adjusted to the subject (top of screen at approximately the horizontal eye height), so that the head–neck region was in a reasonably erect posture and the subject’s forearms were supported on the rounded edge of the keyboard tray. The adjustments were made to provide support to the subject’s body and to eliminate unnecessary movements contributing to variations in EMG signals. The keyboard was centered in front of the subject and no mouse was available to the subject. Each subject performed a standardized task of copytyping children’s stories displayed on screen using TypingMaster (Aquarian Technologies, Maldon, Australia). The texts represented the usual left/right-hand usage balance found in English. The subject was instructed to work at her normal pace and ignore any typing errors made.

ARTICLE IN PRESS G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 270–280

274

Case and Control Groups, though there was significantly more activity in the right side for both groups (Fig. 2). 3.2. Subjective discomfort The Case subjects reported discomfort most frequently in the right shoulder, followed by the left shoulder, both sides of the neck, right wrist/hand and left wrist/hand. The mean of the summed discomfort scores over all trials was 13.8 (77.8) for the Case group and 1.4 (72.3) for the Control group. The discomfort scores were significantly different between groups (F 1;41 ¼ 47:92; po0:0001), and across time (F 3:1;126:9 ¼ 13:94; po0:0001), and there was also a significant group  time interaction (F 3:1;126:9 ¼ 5:20; p ¼ 0:002). 3.3. Post hoc analysis

Fig. 1. Experimental setup with subject in the typing position.

2.6. Data management Four mixed model MANOVAs were used to examine the effects of group (between-subject factor), side and time (within-subject factors) on the 3 levels of APDF, one for each of the 4 muscles. Univariate group  side  time analyses were also conducted on the 50th%le APDF (median). The discomfort results were also analysed using a group and time RANOVA model. The dependent variable was the summed discomfort score for each trial.

3. Results 3.1. Muscle activities of Case and Control Groups Table 3 summarizes the results of the multivariate and univariate analyses of the EMG data. These analyses showed there was a difference in the CES muscle activity between groups, with right CES activity being higher in the Control Group (see Fig. 2), and the difference increasing over time (see Fig. 3). In contrast the right UT had significantly greater activity in the Case Group, while the Control Group had greater symmetry in activity between the left and right UT muscles (Fig. 2). There were no differences in the LT and AD between

Further examination of the Case group results suggested two sub-groups. Subjects with ‘High Discomfort’ (mean summed discomfort score over 5 trials412) showed increasing discomfort over time whilst those with ‘Low Discomfort’ did not (see Fig. 4). When the mean values of the median muscle activities for the 8 muscles were examined in the High and Low Discomfort Groups, differences in CES and UT muscle activities became even more apparent. The Low Discomfort Group had muscle activities similar to the Control Group, and that differed from the High Discomfort Group (see Fig. 5). As both the CES and UT were potential synergists acting on the cervical spine, EMG amplitude ratios were computed to determine if a synergistic relationship between the muscles could be observed (Edgerton et al., 1997). Pairwise contrasts of the right UT/CES ratio showed that the High Discomfort Group was significantly higher than the Control Group (t40 ¼ 2:15; p ¼ 0:048), and the Low Discomfort Group (t40 ¼ 2:37; p ¼ 0:032). The other amplitude ratios comparing the left UT and CES, or the ratios between the two sides of the same muscles, showed no significant difference between groups. These ratio results were consistent with the raw results. Subjects in the High Discomfort Group had a significantly higher UT with lower CES activity in the right side; while subjects in the Low Discomfort and Control Groups had the reversed pattern—with significantly lower UT and higher CES activity on the right side. Post hoc analysis was also carried out to determine the relationship between UT activity and discomfort scores by body area. It showed a positive correlation between increased activity in the right UT with right and left neck and right shoulder discomforts (rho ¼ 0.623, 0.458, 0.427, p ¼ 0:000120:007). In contrast, left UT was not correlated with right and left neck and right shoulder discomforts (rho ¼ 0.229, 0.063, 0.234, p ¼ 0:15720:707).

ARTICLE IN PRESS G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 270–280

275

Table 3 Summary of Multivariate (10th, 50th and 90th APDF percentile) and univariate (50th APDF percentile) EMG amplitudes for 4 muscles Muscle

Effect

Multivariate (10th%le, 50th%le and 90th%le APDF)

Univariate (50th%le APDF)

CES

Group Side Time Group  side Group  time Time  side Group  side  time

F 3;39 ¼ 3:34; P ¼ 0:027 F 3;39 ¼ 2:43; P ¼ 0:079 F 12;30 ¼ 2:76; P ¼ 0:012 F 3;39 ¼ 1:18; P ¼ 0:329 F 12;30 ¼ 1:27; P ¼ 0:287 F 12;30 ¼ 0:91; P ¼ 0:550 F 12;30 ¼ 1:38; P ¼ 0:230

F 1;41 ¼ 0:65; P ¼ 0:425 F 1;41 ¼ 4:30; P ¼ 0:044 F 2:2;90:1 ¼ 3:26; P ¼ 0:039 F 1;41 ¼ 1:61; P ¼ 0:212 F 2:2;90:1 ¼ 1:40; P ¼ 0:252 F 1:6;64:0 ¼ 3:55; P ¼ 0:045 F 1:6;64:0 ¼ 1:03; P ¼ 0:346

UT

Group Side Time Group  side Group  time Time  side Group  side  time

F 3;39 ¼ 1:03; P ¼ 0:391 F 3;39 ¼ 1:91; P ¼ 0:144 F 12;30 ¼ 1:27; P ¼ 0:280 F 3;39 ¼ 2:43; P ¼ 0:080 F 12;30 ¼ 1:34; P ¼ 0:246 F 12;30 ¼ 1:22; P ¼ 0:312 F 12;30 ¼ 1:01; P ¼ 0:462

F 1;41 ¼ 0:76; P ¼ 0:308 F 1;41 ¼ 6:02; P ¼ 0:019 F 2:1;86:9 ¼ 1:27; P ¼ 0:287 F 1;41 ¼ 4:99; P ¼ 0:031 F 2:1;86:9 ¼ 1:13; P ¼ 0:330 F 1:6;64:9 ¼ 0:61; P ¼ 0:511 F 1:6;64:9 ¼ 0:56; P ¼ 0:535

LT

Group Side Time Group  side Group  time Time  side Group  side  time

F 3;39 ¼ 0:25; P ¼ 0:859 F 3;39 ¼ 0:20; P ¼ 0:898 F 12;30 ¼ 1:05; P ¼ 0:430 F 3;39 ¼ 0:25; P ¼ 0:859 F 12;30 ¼ 1:04; P ¼ 0:443 F 12;30 ¼ 1:36; P ¼ 0:237 F 12;30 ¼ 1:40; P ¼ 0:218

F 1;41 ¼ 0:02; P ¼ 0:896 F 1;41 ¼ 0:26; P ¼ 0:616 F 3:4;139:3 ¼ 1:21; P ¼ 0:308 F 1;41 ¼ 0:00; P ¼ 0:992 F 3:4;139:3 ¼ 1:11; P ¼ 0:352 F 3:5;145:7 ¼ 2:04; P ¼ 0:100 F 3:5;145:7 ¼ 1:15; P ¼ 0:333

AD

Group Side Time Group  side Group  time Time  side Group  side  time

F 3;39 ¼ 1:40; P ¼ 0:256 F 3;39 ¼ 2:35; P ¼ 0:087 F 12;30 ¼ 0:58; P ¼ 0:844 F 3;39 ¼ 0:04; P ¼ 0:989 F 12;30 ¼ 1:08; P ¼ 0:409 F 12;30 ¼ 0:55; P ¼ 0:867 F 12;30 ¼ 0:90; P ¼ 0:557

F 1;41 ¼ 0:40; P ¼ 0:528 F 1;41 ¼ 4:14; P ¼ 0:048 F 3:3;133:9 ¼ 1:24; P ¼ 0:296 F 1;41 ¼ 0:05; P ¼ 0:827 F 3:3;133:9 ¼ 0:44; P ¼ 0:738 F 2:6;107:3 ¼ 0:47; P ¼ 0:678 F 2:6;107:3 ¼ 0:61; P ¼ 0:590

 P-value significant at a ¼ 0:05 level.

55 Case

50

Control

50th%leAPDF (% MEMG)

45 40 35 30 25 20 15 10 5 0 right CES

left CES

right UT

left UT right LT Muscle

left LT

right AD

left AD

Fig. 2. Comparison of the mean (+SD) 50th% APDF of the 8 muscles in the Case and Control Groups.

4. Discussion The present study has shown important differences in EMG activity patterns of major neck–shoulder muscles

between the two subject groups. In particular, the Case subjects showed higher activity levels in the right UT muscle while the Control subjects showed higher activity in the right CES. These results suggest that there were

ARTICLE IN PRESS G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 270–280

276

50th%leAPDF (%MEMG)

30 25 20 15 10

Case Group

Control Group

5 0 T1

T2

T3

T4

T5

T1

T2

T3

T4

T5

Trial right CES left CES

right UT left UT

Fig. 3. Comparison of Mean 50th%APDF (+SD) of CES and UT over 5 trials in the Case and Control Groups.

Mean summed score

30 Case High Discomfort Low Discomfort Control

25 20 15 10 5 0 T0

T1

T2

T3

T4

T5

Trial

Fig. 4. Group means of summed discomfort scores comparing Case Group (with High–Low Discomfort sub-groups) and Control Group.

different muscle activation strategies present in different individuals while maintaining the static neck–shoulder posture in performing the same sustained computer task. The group difference in muscle activities became even more apparent when Case subjects were split into subgroups based on their levels of discomfort during the trial. These results suggest that the highly symptomatic individuals employed higher levels of activity in the right UT muscle in maintaining the static neck–shoulder posture; while those with mild or no symptoms were able to perform the same task with more even distribution of UT activity bilaterally and at a lower level. The present finding of altered muscle recruitment patterns highlight the importance of the synergistic roles of the UT and CES muscles in the postural control of the cervical spine. These findings support other reports of altered patterns of muscle recruitment in the cervical spine musculature in the presence of pain disorders of the cervical spine (Hall and Quintner, 1996; Jull, 2000;

Sterling et al., 2001, 2003; Falla et al., 2003; Jull et al., 2004). Bansevicius and Sjaastad (1996) reported the presence of increased UT motor activity in subjects with cervocogenic headache. Hall and Quintner (1996) also showed increased UT activity with sensitized neural tissue, in patients with painful cervical radiculopathy; and Nederhand et al. (2000) demonstrated increased activity in the UT muscle during single arm tasks. In subjects with whiplash injuries as well as those with insidious or chronic neck pain, decreased activity of the deep neck flexors have been reported with associated increased activity of the superficial neck flexors (Jull, 2000; Sterling et al., 2001; Falla et al., 2003; Jull et al., 2004). Together these studies provide evidence that complex patterns of altered recruitment occur in the presence of pain, with inhibition of the deep local musculature (deep neck flexors) in some circumstances, and associated dominant patterns of activation of the superficial musculature (superficial neck flexors and UT). In the present study the focus was on the posterior cervical musculature, where a similar pattern was observed, with increased muscle activity observed in the UT muscle with associated reduced activity in the CES, in subjects with pain. The converse was the case in subjects without pain. The correlation of neck–shoulder discomforts with increased UT muscle activity further supports this relationship. The CES muscles can be classified as part of the ‘‘local muscle system’’ (Bergmark, 1989) of the cervical spine that are anatomically designed for the regional control of extension moments of the head and neck. It could be hypothesized that the higher CES activities in the Control Group may reflect a more efficient strategy for controlling the extension moment of the cervical spine when compared to the activation of the UT observed in the symptomatic group.

ARTICLE IN PRESS G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 270–280

277

70 High Discomfort

50th%le APDF (%MEMG)

60

Low Discomfort Control

50 40 30 20 10 0 rightCES

leftCES

rightUT

leftUT

Muscle Fig. 5. Mean (+SD) 50th%APDF of the CES and UT muscles comparing High Discomfort, Low Discomfort and Control Groups.

Altered muscle recruitment patterns associated with musculoskeletal pain disorders have also been linked with deviations in normal movement patterns and joint control (Griegel-Morris et al., 1992; Dall’Alba et al., 2001; Sterling et al., 2001; Szeto et al., 2002). In the present group of symptomatic office workers, it would be expected that cervical postures and movements may also display abnormal patterns of control and these are being reported in another paper (Szeto et al., 2005). Many past studies focusing on occupational tasks have measured electrical activity in the UT alone, without studying the relationship of this muscle to other synergists (Jensen et al., 1993; Ha¨gg and Astrom, 1997; Nordander et al., 2000; Roe et al., 2001). Kleine et al. (1999) reported higher amplitudes in UT muscles than the CES and AD muscles, but did not compare UT recruitment patterns to the other muscles studied. In the present study, we demonstrated significantly higher amplitude ratios of UT/CES in the High Discomfort Group compared to the Low Discomfort Group. Calculation of ratios may be valuable in detecting abnormal neck and shoulder muscle recruitment pattern relationships, as has been reported in studies of low back pain patients (Edgerton et al., 1997; O’Sullivan et al., 1997). However the mere finding of altered muscle recruitment or movement patterns and their association with musculoskeletal pain disorders, does not determine the cause of these findings nor their effect on the pain disorder. It is well documented that chronic musculoskeletal pain is a multi-dimensional problem that is associated with, and influenced by, a complex array of both physical and psychosocial factors (Linton, 2000; Marras et al., 2000; Forde et al., 2002; Waddell, 2004). It is possible that the increased activity in UT observed in the current study was a result of a complex of different factors that may vary for different individuals. These factors may include reflex muscle activity in response to

the presence of chronic pain disorder or neural tissue sensitization (Mense, 1993; Hall and Quintner, 1996), altered posturing of the head/neck and upper limb, as well as intrinsic motor control differences in the subjects influenced by individual genetic, adaptive and cognitive processes (Linton, 2000; Marras et al., 2000; Forde et al., 2002). Westgaard (2000) and Westgaard et al. (2001) proposed that inter-individual differences in EMG activity patterns were more likely to be a feature of intrinsic, person-based differences rather than purely responses to external work demands, and our present results would support this concept as the external environment was standardized for all subjects. These intrinsic differences may involve motor control mechanisms which may be related to the development of WRNULD. Past established models on musculoskeletal pain such as the vicious cycle model (Travell and Simons, 1983; Johansson and Sojka, 1991) and the pain-adaptation model (Lund et al., 1991) do not seem to fit with our present results. The bilateral and widespread location (well beyond the anatomical boundary of the upper trapezius muscle) and rising levels of discomforts over time contrasted the fairly constant patterns of muscle activity, suggesting that the UT muscle is not the primary source of pain. These findings do not support the ‘‘vicious cycle’’ model which predicts that muscle pain and muscle tension escalate together in a selfperpetuating cycle (Travell and Simons, 1983; Johansson and Sojka, 1991). In the present study, the muscle recruitment patterns were static in spite of rising discomfort levels in the ‘‘High Discomfort’’ Group. While the vicious cycle model could not explain the increase in symptoms during the experimental task, it is possible that the trapezius hyperactivity, observed in the chronically symptomatic subjects in this study, developed secondary to the onset of pain and then became a programmed motor response associated with a chronic

ARTICLE IN PRESS 278

G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 270–280

pain disorder. Sterling et al. (2001) concurred that altered patterns of neuromuscular activation resulting in loss of joint stability and normal movement control may persist into periods of ‘‘chronicity’’ which may possibly be the reason for ongoing symptoms. On the other hand, the pain adaptation model (Lund et al., 1991) predicts that pain inhibits muscle activity, although this model mainly applies to phenomena elicited by experimental pain. While evidence of inhibition of the deep neck flexor muscles in subjects with cervical spine pain exists, this is associated with increased activity of the superficial neck flexors (Jull, 2000; Sterling et al., 2001; Falla et al., 2003; Jull et al., 2004). In a similar manner the deficits in the CES observed were associated with increased activation of the UT in the ‘‘High Discomfort’’ subjects. These findings suggest complex changes within the synergistic control of head–neck posture that are consistent despite increasing symptoms. It is also interesting to note that the increased activity observed in the UT was not associated with a deficit in activity in LT. These findings do not fully support the simplistic model of pain adaptation, but the complex changes that exist in muscle recruitment patterns associated with the presence of pain. It is also known that muscle recruitment patterns are influenced by changes in joint kinematics of the anatomical regions that they control (Bergmark, 1989; Edgerton et al., 1997; Sterling et al., 2001; Forde et al., 2002). Analysis of the associated kinematics of the head/ neck and upper limb that accompanied the typing task are presented in another paper (Szeto et al., 2005). Given that the ergonomic setup was controlled for, any altered kinematics would likely reflect inherent differences within individuals rather than the influence of the workstation. Regardless of the cause or internal attributes that result in the altered muscle recruitment changes, it can be argued that the observed findings in the High Discomfort Group reflect sub-optimal motor recruitment patterns for the typing task. Increased activity in the UT muscle, would be likely to expose the cervico-thoracic region to increased compressive forces which may be a mechanism for maintenance of ongoing tissue strain and nociception in these subjects. On face value, increased activation of the UT and a reduction in CES muscle activity appears to be a maladaptive response to a pain disorder of the cervical spine as high-level UT muscle activity does not appear well designed for this task. It was interesting to note that the Low Discomfort subjects did not display the same altered muscle recruitment patterns as the High Discomfort Group. These findings suggest that different sub-groups with neck and arm pain may exist which have different intrinsic factors associated with what appears to be a similar pain disorder. In contrast, the Control subjects demonstrated relatively lower and more symmetrical muscle activity patterns, suggesting that they had the ability to keep the

muscles in a more relaxed state in their daily work with computers. From the outcome in terms of discomfort scores, it would suggest that the Control Group has adopted or maintained a more efficient motor recruitment strategy. To further test these possible mechanisms and hypotheses, prospective longitudinal studies investigating the development of these pain disorders and clinical outcome studies to test specific motor learning interventions will need to be conducted. Further research is also required to investigate whether differences in the deep segmental and anterior cervical musculature co-exist in this population, as well as investigate the relationship between these findings with other intrinsic physical factors as well as psychological and personality traits.

5. Conclusion The present results have demonstrated important inter-individual differences in muscle activity patterns in the neck–shoulder stabilizers when the task performed and the workstation factors were standardized. In particular, subjects with more severe symptoms (the High Discomfort Group) had significantly increased activity in the right UT muscle as well as significantly increased UT/CES amplitude ratio on the right. This pattern of muscle activation appears to be mal-adaptive as the UT muscle is not anatomically well designed to be a postural stabilizer of the cervical spine for a typing task. In contrast, the Control subjects as well as those in the Low Discomfort sub-group had a more symmetrical distribution of muscle activities, and their higher utilization of the CES muscles may be a more efficient muscle activation strategy for the task. These results suggested that altered muscle recruitment patterns may be an important factor in some computer workers with musculoskeletal symptoms. However there may be multiple factors contributing to the muscle activation patterns observed which could not be fully explained from the present findings.

Acknowledgements The authors would like to acknowledge the Occupational Safety and Health Council of Hong Kong for the research grant that has made this project possible. We would also like to thank Mr. Paul Davey for developing the Labview program for processing the EMG data. References Aara˚s A, Veierod MB, Larsen S, Ortengren R, Ro O. Reproducibility and stability of normalized EMG measurements on musculus trapezius. Ergonomics 1996;39(2):171–85.

ARTICLE IN PRESS G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 270–280 Aara˚s A, Fostervold KI, Ro O, Thoresen M. Postural load during VDU work: a comparison between various work postures. Ergonomics 1997;40(11):1255–68. Bansevicius D, Sjaastad O. Cervicogenic headache: the influence of mental load on pain level and EMG of shoulder–neck and facial muscles. Headache 1996;36:372–8. Bansevicius D, Westgaard RH, Jensen C. Mental stress of long duration: EMG activity, perceived tension, fatigue, and pain development in pain-free subjects. Headache 1997;37:499–510. Bergmark A. Stability of the lumbar spine. A study in mechanical engineering. Acta Orthopaedic Scandinavica 1989;230(60):20–4. Bergqvist U. Health problems during work with visual display terminals. Scandinavian Journal of Work, Environment & Health 1993;20:27–33. Bernard B, Sauter S, Peterson M, Fine L, Hales T. Job task and psychosocial risk factors for work-related musculoskeletal disorders among newspaper employees. Scandinavian Journal of Work, Environment & Health 1994;20:417–26. Chaffin DB, Andersson GB, Martin BJ. Occupational biomechanics. 3rd ed. New York: Wiley; 1999. p. 418–26. Cooper A, Straker L. Mouse versus keyboard use, a comparison of shoulder muscle load. International Journal of Industrial Ergonomics 1998;22:351–7. Culham E, Peat M. Functional anatomy of the shoulder complex. Journal of Orthopaedic and Sports Physical Therapy 1993;18(1): 342–50. Dall’Alba PT, Sterling MM, Treleaven JM, Edwards SL, Jull GA. Cervical range of motion and discriminates between asymptomatic persons and those with whiplash. Spine 2001;26(19):2090–4. Edgerton VR, Wolf SL, Levendowski DJ, Roy RR. EMG activity in neck and back muscles during selected static postures in adult male and females Physiotherapy Theory and Practice 1997;13:179–95. Falla D, Rainoldi A, Merletti R, Jull G. Myoelectric manifestations of sternocleidomastoid and anterior scalene muscle fatigue in chronic neck pain patients. Clinical Neurophysiology 2003;114:488–95. Forde MS, Punnett L, Wegman DH. Pathomechanics of work-related musculoskeletal disorders: conceptual issues. Ergonomics 2002; 45(9):619–30. Griegel-Morris P, Larso K, Mueller-Klaus K, Oatis CA. Incidence of common postural abnormalities in the cervical, shoulder, and thoracic regions and their association with pain in two age groups of healthy subjects. Physical Therapy 1992;72(6):425–31. Ha¨gg GM, Astrom A. Load pattern and pressure pain threshold in the upper trapezius muscle and psychosocial factors in medical secretaries with and without shoulder/neck disorders. International Archives of Occupational & Environmental Health 1997;69:423–32. Hall T, Quintner J. Responses to mechanical stimulation of the upper limb in painful cervical radiculopathy. Australian Journal of Physiotherapy 1996;42:277–85. Hodges PW, Richardson CA. Inefficient muscular stabilization of the lumbar spine associated with low back pain: a motor control evaluation of transverses abdominis. Spine 1996;21:2640–50. Jensen C, Nilsen K, Hansen K, Westgaard RH. Trapezius muscle load as a risk indicator for occupational shoulder–neck complaints. European Journal of Applied Physiology and Occupational Physiology 1993;64:415–23. Johnson G, Bogduk N, Nowitzke A, House D. Anatomy and actions of the trapezius muscle. Clinical Biomechanics 1994;9:44–50. Jonsson B. Measurement and evaluation of local muscular strain in the shoulder during constrained work. Journal of Human Ergology 1982;11:73–88. Johansson H, Sojka P. Pathophysiological mechanisms involved in genesis and spread of muscular tension in occupational muscle pain syndromes: a hypothesis. Medical Hypothesis 1991;35:196–203. Jull G. Deep cervical neck flexor dysfunction in whiplash. Journal of Musculoskeletal Pain 2000;8:143–54.

279

Jull G, Barrett C, Magee R, Ho P. Further clinical clarification of the muscle dysfunction in cervical headache. Cephalalgia 1999;19(3): 179–85. Jull G, 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. Kamwendo K, Linton SJ, Moritz U. Neck and shoulder disorders in medical secretaries, Part II: ergonomic work environment and symptom profile. Scandinavian Journal of Surgery 1991;23:135–42. Kilber B. The role of the scapula in athletic shoulder function. American Journal of Sports Medicine 1998;26(2):325–39. Kleine B-U, Schumann N-P, Bradl I, Grieshaber R, Scholle H-C. Surface EMG of shoulder and back muscles and posture analysis in secretaries typing at visual display units. International Archives of Occupational & Environmental Health 1999;72:387–94. Kroemer K, Grandjean E. Fitting the task to the human, 5th ed. London: Taylor & Francis; 1997. Kumar S. Theories of musculoskeletal injury causation. Ergonomics 2001;44(1):17–47. Kuorinka I, Jonsson B, Kilbom A˚, Vinterberg H, Biering-So¨renson F, Andersson A, Jorgensen K. Standardised Nordic Questionnaires for the analysis of musculoskeletal symptoms. Applied Ergonomics 1987;8(3):233–7. Linton SJ. A review of psychological risk factors in back and neck pain. Spine 2000;25(9):1148–56. 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 Physiological Pharmacology 1991;69:683–94. Marras W, Davis KG, Heaney CA, Maronitis AB, Allread WG. The influence of psychosocial stress, gender, and personality on mechanical loading of the lumbar spine. Spine 2000;25(23):3045–54. Mense S. Nociception from skeletal muscle in relation to clinic muscle pain. Pain 1993;54:241–89. Nederhand MJ, Ijzerman MJ, Hermens HJ. Cervical muscle dysfunction in chronic whiplash-associated disorder grade II (WAD II). Spine 2000;25(15):1938–43. Nordander C, Hansson G-A, Rylander L, Asterland P, Bystrom JU, Ohlsson K, et al. Muscular rest and gap frequency as EMG measures of physical exposure: the impact of work tasks and individual related factors. Ergonomics 2000;43(11):1904–19. O’Sullivan P, Twomey L, Allison G, Sinclair J, Miller K, Knox J. Altered patterns of abdominal muscle activation in patients with chronic low back pain. Australian Journal of Physiotherapy 1997;43(2):91–8. Roe C, Bjorklund RA, Knardahl S, Wærsted M, Vollestad NK. Cognitive performance and muscle activation in workers with chronic shoulder myalgia. Ergonomics 2001;44(1):1–16. Schuldt K, Ekholm J, Harms-Ringdahl K, Nemeth G, Arborelius UP. Effects of changes in sitting work posture on static neck and shoulder muscle activity. Ergonomics 1986;29(12):1525–37. Siu ZKF, Chan CCH. Prevalence of musculoskeletal discomforts and the work demand of visual display terminal (VDT). In: Proceedings of the first Pan Pacific conference on rehabilitation, Guangzhou China; 1998. p. 18. Sterling M, Jull G, Wright A. The effect of musculoskeletal pain on motor activity and control. The Journal of Pain 2001;2(3):135–45. Sterling M, Jull G, Vicenzino B, Kenardy J, Darnell R. Development of motor system dysfunction following whiplash injury. Pain 2003;103:65–73. 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 LM, O’Sullivan PB. A comparison of symptomatic and asymptomatic office workers performing monotonous keyboard work, 2: neck and shoulder kinematics. Manual Therapy 2005, in press, doi:10.1016/j.math.2005.01.005.

ARTICLE IN PRESS 280

G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 270–280

Tittiranonda P, Burastero S, Rempel D. Risk factors for musculoskeletal disorders among computer users. Occupational Medicine: State of the Art Reviews 1999;14(1):17–38. Travell JG, Simons DG. Myofascial pain and dysfunction: the trigger point manual. London: Williams & Wilkins; 1983. Waddell G. The back pain revolution. Edinburgh: Churchill Livingston; 2004. Waersted M. Human muscle activity related to non-biomechanical factors in the workplace. European Journal of Applied Physiology 2000;83:151–8.

Westgaard RH. Work-related musculoskeletal complaints: some ergonomics challenges upon the start of a new century. Applied Ergonomics 2000;31:569–80. Westgaard RH, Jensen C, Hansen K. Individual and work-related risk factors associated with symptoms of musculoskeletal complaints. International Archives of Occupational & Environmental Health 1993;64:405–13. Westgaard RH, Vasseljen O, Holte KA. Trapezius muscle activity as a risk indicator for shoulder and neck pain in female service workers with low biomechanical exposure. Ergonomics 2001;44(3):339–53.

ARTICLE IN PRESS

Manual Therapy 10 (2005) 281–291 www.elsevier.com/locate/math

Original article

A comparison of symptomatic and asymptomatic office workers performing monotonous keyboard work—2: Neck and shoulder kinematics Grace P.Y. Szetoa,b,, Leon M. Strakerb, Peter B. O’Sullivanb a

Department of Rehabilitation Sciences, Hong Kong Polytechnic University, Hong Kong SAR, PR China b School of Physiotherapy, Curtin University of Technology, Perth 6845, Australia Received 7 November 2004; received in revised form 30 November 2004; accepted 4 January 2005

Abstract Prolonged static posture has been identified as a major risk factor for work-related neck and upper limb disorders (WRNULD) in computer users. Previous research has mainly examined working postures in healthy pain-free individuals. The present study examined whether symptomatic subjects exhibited the same kinematic patterns as asymptomatic controls during a prolonged computer task. In a Case–Control comparison, female office workers performed the same computer task using the same adjustable computer workstation for 1 h. Three-dimensional (3D) kinematics were measured in the head–neck, thorax and shoulder (upper arm) segments. Case Group subjects (n ¼ 21) displayed trends for increased head–neck flexion angles and greater ranges of movements than the Control Group (n ¼ 17). There were also small but significant differences between groups in side flexion and rotation angles of the head-neck region. The shoulder joints displayed significantly greater flexion and abduction angles on the right in both groups, although no group differences were observed. The increased neck flexion angles were associated with significantly higher activity in the upper trapezius muscle and with neck and shoulder discomfort. The individual differences in postural habits appeared to be independent of the physical environment. These results suggest motor control changes are associated with the presence of WRNULD. r 2005 Elsevier Ltd. All rights reserved. Keywords: Work-related neck and upper limb disorders; Kinematics; Office ergonomics; Motor control

1. Introduction Work-related neck and upper limb disorders (WRNULD) have been associated with long hours of computer work and prolonged periods of holding a static posture (Bernard et al., 1994; Tittiranonda et al., 1999; Ariens et al., 2001). For office workers, static posture is most pronounced in the neck and shoulder region, resulting in increased forward neck flexion posture and increased static muscle tension in the region Corresponding author. Department of Rehabilitation Sciences, Hong Kong Polytechnic University, Hong Kong SAR, PR China. Tel.: +852 27666706; fax: +852 23308656. E-mail address: [email protected] (G.P.Y. Szeto).

1356-689X/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2005.01.005

(Schuldt et al., 1986; Liao and Drury, 2000; Ariens et al., 2001). In a previous field investigation (Szeto et al., 2002), it was found that office workers using computers had increased forward neck flexion compared to their relaxed sitting postures, and this forward flexion was more pronounced in symptomatic persons (about 13% more neck flexion). The consequence of increased forward neck flexion may result in increased tension in the regions postural stabilizing muscles as well as increased compressive forces in the articulations of the cervical spine resulting in higher risk of WRNULD. Many ergonomics studies have focused on the postural effects of changing parts of the computer workstation such as the display screen height and/or keyboard height (Villanueva et al., 1996; Aara˚s et al.,

ARTICLE IN PRESS 282

G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 281–291

1997; Burgess-Limerick et al., 1999; Liao and Drury, 2000). Most of these studies have been conducted on normal healthy persons without work-related musculoskeletal problems. Posture is often studied together with electromyography (EMG) to examine musculoskeletal loading in performing occupational tasks (Vasseljen and Westgaard, 1995, 1997; Aara˚s et al., 1997; Saito et al., 1997; Madeleine et al., 1999). Body posture in performing computer work has been commonly studied in the sagittal plane as two-dimensional (2D) static posture or movements. Photographs of the sagittal profile have been used to measure 2D static postures of the major spinal segments (Raine and Twomey, 1997; Straker et al., 1997). Other studies have used video cameras to capture the sagittal profile and digitized reflective marker locations frame by frame (Braun, 1991; Villanueva et al., 1996; Liao and Drury, 2000; Szeto et al., 2002). This method has been used to provide information about the extent and frequency of movements in addition to static posture (Liao and Drury, 2000; Szeto et al., 2002). More recently, 3D motion analysis systems have been used to analyse motions of body segments in three planes. The advantages and disadvantages of using various motion analysis systems in ergonomic research have been reviewed by Chaffin et al. (1999) and Li and Buckle (1999). In brief, the advantages of these 3D systems are that they can track the movements of multiple body segments simultaneously with EMG capture. In recent years, there has been a rising interest in using these systems to examine the biomechanics of upper limb motions during functional and occupational activities (Anglin and Wyss, 2000; Rau et al., 2000). Past studies have mainly compared within-subject changes in posture in response to workstation changes such as different display screen heights (Aara˚s et al., 1997; Straker et al., 1997; Burgess-Limerick et al., 1999). A recent study by Finley and Lee (2003) examined scapular kinematics in an upright seated position vs. a slouched seated position. These studies, however have only been conducted on healthy pain-free persons. There is a lack of evidence whether symptomatic and asymptomatic persons have different postures when working with the same workstation setting. Szeto et al. (2002) reported consistent differences in the neck and shoulder posture between symptomatic and asymptomatic office workers; but these were based on workers positioned in their own workstations which can involve different settings. Symptomatic persons were found to have increased forward neck flexion and head tilt angles, as well as greater extents of neck movements. These patterns were consistent on repeated measures throughout the day, and were thought to reflect more the subjects’ personal habitual movements and postures rather than the influence of their workstations.

Vasseljen and Westgaard (1995, 1997) compared the postures and muscle activities in 24 matched pairs of office workers in a Case–Control study. Only the upper back and arm postures were reported and there were no significant differences between groups during a 30-min work period. In contrast, Hermans et al. (1998) reported on Case–Control differences in neck postures and muscle activities but the sample sizes were very small. Madeleine et al. (1999) compared working postures and movements in addition to muscle activities between symptomatic and asymptomatic persons in a meatcutting task. Trends for increased amplitudes (extents) of arm and trunk movements were reported in the symptomatic subjects and the authors suggested that there was an interaction between pain and motor control. Other studies that have compared symptomatic and asymptomatic office workers have mainly concentrated on muscle electrical activity and conflicting results have been reported (Jensen et al., 1993; Hagg and Astrom, 1997; Nordander et al., 2000; Roe et al., 2001). If individual differences are important in explaining the development of WRNULD, then there is a need to establish these differences more clearly with symptomatic and asymptomatic office workers performing the same task under the same circumstances. In clinical research, some studies have reported significant differences in relaxed or resting head/neck postures (commonly known as the forward head posture) between symptomatic patients and asymptomatic persons (Braun, 1991; Griegal-Morris et al., 1992). However, large variations in posture exist among individuals, and some studies have reported no significant relationship between posture and symptoms (Raine and Twomey, 1997; Kebaetse et al., 1999). Recent studies examining active range of movement in whiplash patients have shown a significant reduction in end ranges of primary neck movements compared to asymptomatic persons (Dall’Alba et al., 2001). Sterling et al. (2003) also reported reduced ranges of all primary neck movements and altered kinesthetic awareness associated with relocation from right neck rotation, in patients for up to 3 months after whiplash. These findings were associated with increased activities in the superficial neck flexor muscles (Sterling et al., 2003). Researchers have questioned whether posture is an innate characteristic, or a response to the physical environment (Westgaard, 2000; Roe et al., 2001). If it is purely a response to the physical environment, then all persons should have a similar posture when working with the same workstation. Our previous field investigation showing consistent symptomatic–asymptomatic differences would suggest that these differences in posture may reflect either a trait or an individual response to their environment.

ARTICLE IN PRESS G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 281–291

In the present research, we have measured both posture and muscle electrical activity in the neck and shoulder region in symptomatic and asymptomatic office workers in a standardized working environment. This paper focuses on the kinematics in the neck–shoulder region between the symptomatic and asymptomatic groups while performing a 1-h typing task. Muscle activity patterns of neck–shoulder musculature were also measured in the present study and these are reported in a companion paper (Szeto et al., 2005). The relationships between the kinematics of the neck and shoulder segments, the muscle activities in the same region and discomfort are discussed in this paper.

2. Subjects and method

283

subjects were lost from both groups, the results were not biased by this factor. 2.2. Variables The independent variables were group (Case vs. Control), and time (5 repeated measures during a 1-h typing task). Side differences were only assessed for the shoulder kinematics. The dependent variables consisted of 1. Head1 X (flexion/extension)2, Head Y (left/right side flexion)3 and Head Z (left/right rotation)4 2. Thorax X (flexion/extension)2, and Thorax Y (left/ right side flexion)3 3. Left and right Shoulder X (flexion/extension)2, and Shoulder Y (abduction/adduction)5

2.1. Subjects 2.3. Instrumentation and procedures The present research employed a Case–Control quasiexperimental design where subjects were divided into Case and Control Groups based on their past and present discomfort profiles. Each subject performed a 1h typing task using the same adjustable workstation. Female office workers were recruited and assigned into Case (n ¼ 23) and Control (n ¼ 20) Groups. Subjects with current and past complaints of neck and shoulder discomforts within the last 7 days and with discomforts lasting more than 3 months in the past year were assigned into the Case Group. Other subjects without any current discomfort or significant past history were grouped as Controls. Other than a significant difference in their mean age of about 5 years, the subjects were reasonably matched in terms of their physical build, their work background and their experiences with computers. All subjects were experienced touch-typists from a variety of different businesses sampled by convenience (See Szeto et al., 2005 for further subject details). Whilst 23 Case subjects and 20 Control subjects were measured for both EMG and kinematic data during the 1-h typing task, due to technical problems with missing markers, kinematic data were only successfully analysed for 21 Case subjects and 17 Control subjects. In 3D motion analysis, this problem is quite common and if certain important markers could not be tracked properly, the body segment model cannot be constructed. Where body segments were not able to be constructed for just one segment at one time the whole subject had to be discarded due to the repeated measures design. The 21 subjects in the Case Group still had a similar mean age (35.874.6) to that of the full group of 23 (mean age ¼ 36:0  4:6), and the same applied to the Control Group. Hence given a similar number of

The Vicon 370, Version 3.1 (Oxford Metrics, UK), 3D motion measurement and analysis system was used to record the upper body postures and movements during the 1-h typing task. Six infra-red cameras were used to perform the video capture at a sampling frequency of 60 Hz. Before the start of the video capture, the Vicon system was calibrated to determine the exact positions and orientation of the cameras with respect to the laboratory (‘‘global coordinate system’’). Static and dynamic calibration procedures (using Vicon calibration frame and wand, respectively) were carried out to ensure that the ‘‘Calibration Residual’’ values were below 0.1% of the reconstruction volume (Vicon Users’ Manual, 1997). The Vicon 370 system has demonstrated high accuracy and reliability (0.94 mm absolute mean error of marker movements) (Ehara et al., 1997). Reflective markers were placed on the following bony landmarks to define the body segments. Altogether 3 body segments were defined by at least 3 markers in each segment 1. Head–neck segment—defined by 2 markers at two sides of the forehead (lateral to outer canthus of the eye) and 2 markers at bilateral mastoid processes, 2. Thorax segment—defined by marker at the top of the sternum just below the suprasternal notch, and markers at C7 and T8 spinous processes, 3. Shoulder (upper arm) segment—defined by markers at the acromioclavicular joint, at the lateral humeral 1 In the Vicon biomechanical model, the Head and Neck is considered as one segment. 2 Positive(+ve) ¼ flexion, negative(–ve) ¼ extension. 3 –ve ¼ left side flexion. 4 –ve ¼ right rotation. 5 +ve ¼ abduction.

ARTICLE IN PRESS 284

G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 281–291

epicondyle and at midpoint of the posterior shaft of the humerus between lesser tuberosity and olecranon process. Synchronized kinematics and EMG signals were captured at the 5th, 20th, 35th, 50th and 60th min of the typing task. Each data capture was for a 60-s duration. The subject was asked to rate the task related discomfort in 10 upper body regions (left/right neck, shoulder, upper back, elbow and wrist/hand) on a scale of 0–10 immediately after each kinematics and EMG data capture. 2.4. Controlled variables The same computer workstation with adjustable chair height and adjustable keyboard height was used by all the subjects. The subject was instructed to adjust the chair height, screen height and keyboard height, until a reasonably erect and comfortable posture was achieved. This posture is consistent with the ‘‘optimal posture’’ that is commonly recommended by ergonomics textbooks (Kroemer and Grandjean, 1997; Chaffin et al., 1999). The adjustments of height and positions of the furniture and computer equipment were made so that their major body parts were well supported and their overall postures were similar. A standardized typing package, TypingMaster (Aquarian Technologies, Maldon, Australia) was used to display children’s stories on-screen for the subject to perform copy-typing. The subject was instructed to perform the typing task continuously at her normal pace for 1 h. The same keyboard was used by all subjects and it was centered in front of the subject for typing. No mouse use was allowed during the typing task. 2.5. Data management For each data capture trial, the marker trajectories were reconstructed and processed using the Vicon Bodybuilder (Oxford Metrics, UK) to produce Euler’s angles (X ; Y ; Z) for the 3 body segments: head–neck, thorax, and upper arm (shoulder). The data were exported to a Labview (National InstrumentsTM, Austin, USA) program where angle corrections were computed to produce anatomical angles referenced to the vertical. The 10th percentile (%le), 50th%le and 90th%le of the amplitude probability distribution function (APDF) were then computed using the corrected angles. The APDF function has been commonly used in occupational research as a way to quantify risk thresholds for exposure to certain amplitudes at these levels (Jonsson, 1982; Jensen et al., 1993; Hagg and Astrom, 1997; Nordander et al., 2000; Roe et al., 2001). As the APDF levels are commonly used as variables to study EMG, these are also used in this study

to quantify the joint angles. The 50th%le (‘‘median’’) angle was used as an indicator of the average angle (posture) of each movement. The ‘‘range’’ of movement of the segment was calculated as the difference between the 90th%le and the 10th%le in the APDF data. Three mixed model (between-subjects group factor, within-subjects time factor, and for some measures, within-subjects side factor) MANOVAs were performed with the median angles and the ranges of each movement, one for each of the 3 body segments (head, thorax and upper arm). Significant MANOVAs were followed by univariate mixed model ANOVAs to identify which dependent variables contributed to the multivariate results.

3. Results 3.1. Head– neck postures and movements Table 1 shows that the Case subject mean head flexion was 3.91 more than Control subjects with 2.31 greater range. The greater head–neck flexion in Case subjects was maintained over the whole 1 h typing trial (see Fig. 1). The Control subjects had 1.51 more right side flexion and 2.41 more right rotation. A MANOVA found a significant group difference (Table 2) with the univariate analyses showing flexion range, side flexion median and rotation median postures were significantly different between groups. 3.2. Thorax postures and movements Case and Control subject thoracic postures and movements were generally within 11 of each other and a MANOVA found no significant effect of group (nor time). 3.3. Bilateral shoulder postures and movements Case subjects had around 11 more right shoulder flexion and 31 less right shoulder abduction than Control subjects (Table 1). A MANOVA found a significant side difference (Table 2). Subsequent univariate analyses only identified a side difference for flexion (median and range) and abduction (range). 3.4. Post-hoc analysis: high– low Discomfort Group differences When the Case Group subjects were sub-divided into 2 groups according to their discomfort scores (the High Discomfort Group was defined by a mean discomfort score per trial 412), there appeared to be a trend for greater differences in head flexion angles between the two sub-groups. The head flexion angle of the High

ARTICLE IN PRESS G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 281–291

285

Table 1 Descriptive statistics of median angles and ranges of the 3 body segments (SD) comparing Case and Control Groups Head/thorax movementsa

Median angle [mean (SD)] Case

Head X (flex/ext)b Head Y (L/R side flex)c Head Z (L/R rot)d Thx X (flex/ext)b Thx Y (L/R side flex)c Shoulder movements Right median angle

Sh X (flex/ext)b Sh Y (abd/add)e

Range [mean (SD)] Control

67.59 (10.8) –1.14 (1.8) 1.78 (2.4) 15.09 (6.8) –1.43 (3.1) [mean (SD)]

Case

Control

63.74 (12.9) –2.67 (2.2) 4.23 (2.5) 14.41 (6.2) –0.24 (2.9) Left median angle [mean (SD)]

6.77 (3.8) 3.56 (2.1) 4.38 (1.6) 2.43 (0.3) 1.67 (0.1) Right range [mean (SD)]

4.53 (2.1) 2.99 (1.5) 3.41 (1.3) 1.98 (0.2) 0.81 (0.1) Left range [mean (SD)]

Case

Control

Case

Control

Case

Control

Case

Control

7.32 (4.8) 14.03 (5.9)

6.32 (3.3) 17.02 (7.9)

5.20 (4.4) 13.72 (6.3)

5.64 (4.5) 14.85 (4.9)

4.82 (1.8) 3.31 (1.7)

4.77 (1.6) 3.22 (1.2)

3.74 (1.5) 2.76 (0.8)

3.52 (1.4) 2.76 (0.8)

a

All head angles were referenced to the vertical. For comparison, looking straight ahead would be equivalent to about 55751. Positive(+ve) ¼ flexion, negative(ve) ¼ extension. c ve ¼ Left side flexion. d ve ¼ Right rotation. e ve ¼ Adduction. b

Joint Angles (degrees)

70

Case Control

68 66 64 62 60 58 T1

T2

T3

T4

T5

Trial Fig. 1. Head flexion median angles (means+SD) in 5 data capture trials in Case and Control Groups.

Discomfort Group showed a mean difference of about 81 compared to the Low Discomfort Group, and a 61 mean difference compared to the Control Group (Fig. 2). However, this difference was not statistically significant in a one-way ANOVA with pairwise contrasts. Yet the pairwise contrasts for head side flexion and rotation angles showed significant differences comparing the High Discomfort Group to the Controls, although these differences only involved a few degrees. There was no significant difference in the shoulder flexion and abduction angles between the High Discomfort and Low Discomfort Groups, nor between these groups and the Control Group (t35 ¼ 1:01–0.93, P ¼ 0:361–0.610). 3.5. Correlations between kinematics, muscle activity and discomforts Spearman’s rho correlation analyses were conducted to examine the relationships between the head and

shoulder mean angles and the muscle activities of the CES and UT muscles, as well as with the mean discomfort scores of the left/right neck and shoulder areas (see Table 3). The results indicated a significant correlation between head flexion and side flexion angles and right UT activity (rho ¼ 0:366; P ¼ 0:024 and rho ¼ 0:336; P ¼ 0:039). In addition, there was also a significant correlation between right shoulder abduction angle and right UT activity (rho ¼ 0:348; P ¼ 0:032). Head flexion and side flexion were also related to right neck discomfort (rho ¼ 0:323; P ¼ 0:048; rho ¼ 0:392; P ¼ 0:015), head side flexion was related to right shoulder discomfort (rho ¼ 0:561; P ¼ 0:000) and head rotation was related to right neck and right shoulder discomfort (rho ¼ 0:421; P ¼ 0:008; rho ¼ 0:323; P ¼ 0:048).

4. Discussion 4.1. Group differences in head– neck kinematics The results showed that the Case subjects had different head postures with greater flexion range, less right side flexion and rotation. These differences were maintained throughout the 1-h typing task, in spite of an increase of the symptoms over the same period. The present trends for increased forward head flexion angles in the Case subjects were compatible with previous findings in the field investigation (Szeto et al., 2002) which also found Case subjects in more neck flexion than the Control subjects (about 71, or a difference of about 13%). In the present study, the overall means of head flexion angles had a smaller group difference of only about 41, although this difference was

ARTICLE IN PRESS G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 281–291

286

Table 2 Summary of multivariate and univariate analyses results for the head–neck, thorax and shoulder segments Movement

Effect

Multivariate (median angle and range)

Univariate (median angle)

Univariate (range)

¼ 4:04; P ¼ 0:004

Head X ; Y ; Z

Group Time Group  time

F 6;31 F 13;24 ¼ 1:09; P ¼ 0:447 F 13;24 ¼ 1:04; P ¼ 0:486

Head X (flex/ext)

Group Time Group  time

F 1;36 ¼ 1:00; P ¼ 0:325 F 4;144 ¼ 2:16; P ¼ 0:077 F 4;144 ¼ 0:29; P ¼ 0:885

F 1;36 ¼ 4:74; P ¼ 0:036 F 4;144 ¼ 0:95; P ¼ 0:437 F 4;144 ¼ 0:90; P ¼ 0:468

Head Y (L/R side flex)

Group Time Group  time

F 1;36 ¼ 5:56; P ¼ 0:024 F 4;144 ¼ 1:29; P ¼ 0:279 F 4;144 ¼ 0:48; P ¼ 0:750

F 1;36 ¼ 0:89; P ¼ 0:035 F 4;144 ¼ 0:73; P ¼ 0:517 F 4;144 ¼ 0:91; P ¼ 0:461

Head Z (L/R Rot)

Group Time Group  time

F 1;36 ¼ 9:47; P ¼ 0:004 F 2:4;88:1 ¼ 0:76; P ¼ 0:493 F 2:4;88:1 ¼ 0:42; P ¼ 0:793

F 1;36 ¼ 4:70; P ¼ 0:053 F 4;144 ¼ 1:26; P ¼ 0:292 F 4;144 ¼ 1:57; P ¼ 0:168

Thorax X ; Y

Group Time Group  time

F 1;36 ¼ 0:11; P ¼ 0:742 F 13;24 ¼ 0:99; P ¼ 0:525 F 13;24 ¼ 0:86; P ¼ 0:641

Shoulder X ; Y

Group Time Side Time  group Side  group Time  side Time  side  group

F 4;33 ¼ 0:45; P ¼ 0:773 F 16;21 ¼ 1:23; P ¼ 0:356 F 4;33 ¼ 8:21; P ¼ 0:000 F 16;21 ¼ 0:62; P ¼ 0:833 F 4;33 ¼ 0:44; P ¼ 0:775 F 16;21 ¼ 1:76; P ¼ 0:112 F 16;21 ¼ 1:91; P ¼ 0:083

Shoulder X (Flex/ext)

Group Time Side Time  group Side  group Time  side Time  side  group

F 1;36 ¼ 0:05; P ¼ 0:828 F 4;144 ¼ 1:37; P ¼ 0:246 F 1;36 ¼ 4:30; P ¼ 0:045 F 4;144 ¼ 0:97; P ¼ 0:426 F 1;36 ¼ 1:12; P ¼ 0:298 F 4;144 ¼ 0:73; P ¼ 0:571 F 4;144 ¼ 1:64; P ¼ 0:169

F 1;36 ¼ 0:11; P ¼ 0:739 F 4;144 ¼ 1:87; P ¼ 0:155 F 1;36 ¼ 25:99; P ¼ 0:000 F 4;144 ¼ 0:39; P ¼ 0:813 F 1;36 ¼ 0:15; P ¼ 0:702 F 4;144 ¼ 0:39; P ¼ 0:721 F 4;144 ¼ 0:28; P ¼ 0:887

Shoulder Y (Abd/add)

Group Time Side Time  group Side  group Time  side Time  side  group

F 1;36 ¼ 1:36; P ¼ 0:251 F 4;144 ¼ 1:78; P ¼ 0:136 F 1;36 ¼ 1:31; P ¼ 0:259 F 4;144 ¼ 0:78; P ¼ 0:543 F 1;36 ¼ 0:74; P ¼ 0:395 F 4;144 ¼ 0:18; P ¼ 0:946 F 4;144 ¼ 0:83; P ¼ 0:509

F 1;36 ¼ 0:02; P ¼ 0:893 F 4;144 ¼ 1:87; P ¼ 0:155 F 1;36 ¼ 7:55; P ¼ 0:009 F 4;144 ¼ 1:87; P ¼ 0:155 F 1;36 ¼ 0:07; P ¼ 0:792 F 4;144 ¼ 4:18; P ¼ 0:003 F 4;144 ¼ 1:92; P ¼ 0:110

 Significance level o0.05.

more apparent when the Case Group was sub-divided into the High and Low Discomfort Groups. The difference in head flexion angle between the High Discomfort Group and the Control Group was about 81, very similar to that found in the field study (see Table 4). This group difference may have important clinical implications, even though it failed to reach statistical significance. (Post hoc power calculations indicated power was only 0.303 and 0.342 for High vs. Low and High vs. Control comparisons. A sample size of around 100 subjects would therefore be needed to achieve a power of 0.80). The present results lend some support to the hypothesis that symptomatic subjects (with more severe

discomforts) had consistent differences in terms of their forward head flexion angles when they performed computer tasks. On the other hand, the subjects with mild discomfort (the Low Discomfort Group) presented with similar kinematic patterns to the Control Group suggesting that kinematic differences were not apparent in this group, and a similar pattern was also found in their muscle activities (Szeto et al., 2005). It has been reported that a difference of about 5o in neck flexion angle can have a considerable impact on the neck extensor moment and the muscle forces required from the neck extensors to support the weight of the head (Straker et al., 1997; Burgess-Limerick et al., 1999; Chaffin et al., 1999). Hence, the present finding of about

ARTICLE IN PRESS G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 281–291

Control

80

Right shoulder pain

70 65 60

r ¼ 0:421; P ¼ 0:008 r ¼ 0:122; P ¼ 0:464 r ¼ 0:176; P ¼ 0:292 r ¼ 0:073; P ¼ 0:662 r ¼ 0:177; P ¼ 0:288 r ¼ 0:294; P ¼ 0:073 r ¼ 0:009; P ¼ 0:959 R ¼ 0:348; P ¼ 0:032 r ¼ 0:108; P ¼ 0:518 r ¼ 0:170; P ¼ 0:308

r ¼ 0:073; P ¼ 0:665 r ¼ 0:102; P ¼ 0:541 r ¼ 0:016; P ¼ 0:925 r ¼ 0:162; P ¼ 0:330 r ¼ 0:233; P ¼ 0:159 r ¼ 0:173; P ¼ 0:299 r ¼ 0:175; P ¼ 0:294

CES ¼ Cervical erector spinae. UT ¼ Upper trapezius.  Significance level o0.05.

Left UT Right UT Left CES Right CES

Table 3 Spearmen’s rho correlation among kinematics, muscle activities and discomfort variables

81 difference in head flexion posture between groups may have clinical significance, indicating that symptomatic subjects had to sustain a greater neck extensor moment throughout the typing task. In the present study, we have also found that the High Discomfort Group subjects had significantly higher activity in the upper trapezius (UT) muscle coupled with lower activity in the cervical erector spinae (CES) muscles (Szeto et al., 2005), compared to the Control Group. Furthermore, head flexion and side flexion correlated with the increase in UT muscle activity and right side neck discomfort further supporting this relationship. From an anatomic perspective, the CES provide an extensor moment to balance the flexion moment created from the forward head posture (Keshner et al., 1989; Chaffin et al., 1999). The UT muscle is a long lever muscle designed more for scapular elevation rather than control of the extensor moment controlling a flexed head–neck position (Keshner et al., 1989; Johnson et al., 1994). The greater activity in UT and lower activity in CES in the High Discomfort Group would appear to represent a mal-adaptive motor response to a greater extensor demand being placed on the cervical spine in these subjects. Hence, high amplitudes of UT coupled with reduced CES controlling the neck extensor moment, may result in a greater compressive penalty on the cervical spine as well as resulting in the larger magnitudes of head–neck movements and altered side bending and rotation postures as observed in the Case Group. This is consistent with the present finding of significantly greater head flexion range in the High Discomfort Group than the Low Discomfort and Control Groups.

Right neck pain

Fig. 2. Comparison of the head flexion median angles (+SD) in the High Discomfort, Low Discomfort and Control Groups.

r ¼ 0:213; P ¼ 0:199 r ¼ 0:094; P ¼ 0:575 r ¼ 0:216; P ¼ 0:193 r ¼ 0:011; P ¼ 0:950 r ¼ 0:144; P ¼ 0:389 r ¼ 0:012; P ¼ 0:944 r ¼ 0:185; P ¼ 0:265

HdFlex_mean

r ¼ 0:122; P ¼ 0:466 r ¼ 0:270; P ¼ 0:102 r ¼ 0:165; P ¼ 0:322 r ¼ 0:128; P ¼ 0:445 r ¼ 0:248; P ¼ 0:134 r ¼ 0:059; P ¼ 0:727 r ¼ 0:160; P ¼ 0:338

40

Head X (flexion) Head Y (side flex) Head Z (rot) Right Sh X (flex) Right Sh Y (abd) Left Sh X (flex) Left Sh Y (abd)

Left neck pain

45

r ¼ 0:323; P ¼ 0:048 r ¼ 0:392; P ¼ 0:015

50

r ¼ 0:262; P ¼ 0:113 r ¼ 0:155; P ¼ 0:353 r ¼ 0:269; P ¼ 0:103 r ¼ 0:107; P ¼ 0:522 r ¼ 0:092; P ¼ 0:585 r ¼ 0:075; P ¼ 0:656 r ¼ 0:096; P ¼ 0:568

55

r ¼ 0:366; P ¼ 0:024 r ¼ 0:336; P ¼ 0:039

Joint angles (degrees)

75

r ¼ 0:090; P ¼ 0:593 r ¼ 0:220; P ¼ 0:184 r ¼ 0:094; P ¼ 0:574 r ¼ 0:077; P ¼ 0:645 r ¼ 0:142; P ¼ 0:395 r ¼ 0:283; P ¼ 0:085 r ¼ 0:076; P ¼ 0:651

Low Discomfort

85

r ¼ 0:228; P ¼ 0:169 r ¼ 0:561; P ¼ 0:000 r ¼ 0:323; P ¼ 0:048 r ¼ 0:027; P ¼ 0:872 r ¼ 0:092; P ¼ 0:581 r ¼ 0:138; P ¼ 0:408 r ¼ 0:021; P ¼ 0:899

Left shoulder pain

High Discomfort

287

ARTICLE IN PRESS G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 281–291

288

Table 4 Comparison of head–neck angles [mean (SD)] in previous field study and the present study Head–neck angles and ranges

Case [mean (SD)]

Control [mean (SD)]

Difference between Groups

Field study

Head Tilt angle Neck Flexion angle Head Tilt range Neck Flexion range

60.61 (6.1) 59.31 (10.0) 15.31 (5.7) 7.71 (2.5)

57.11 (6.2) 52.51 (7.1) 12.31 (5.9) 6.11 (2.0)

3.11 6.81 31 1.61

Present study

Head Flexion angle (composite head–neck flexion)

67.61 (10.8)

63.71 (12.9)

3.91 (Case vs. Control)

4.51 (2.1)

8.21 6.21 2.31 0.21 2.21

Head Flexion range

High Discomfort ¼ 69.91 (10.1) Low Discomfort ¼ 61.71 (11.6) 6.81 (3.8) High Discomfort Group ¼ 6.71 (4.2) Low Discomfort Group ¼ 6.91 (3.0)

The relationship between neck angles/movements, muscle activities and discomforts have not been clearly established in previous studies. Conflicting results have been reported in ergonomic studies, with either increased muscle activities with lower neck angles (HarmsRingdahl et al., 1986; Hermans et al., 1998; Turville et al., 1998) or no change in muscle activities with different neck angles during occupational tasks (Aara˚s et al., 1997; Vasseljen and Westgaard, 1997; Hermans et al., 1998; Burgess-Limerick et al., 1999). The significant correlations between right UT activity and head–neck angles in the present study provided important evidence to illustrate the link between muscle activation and postural control. Clinical research has produced some evidence to suggest that postural differences exist between symptomatic and asymptomatic persons in relaxed standing or sitting but there was no consensus among the different studies (Braun 1991; Griegal-Morris et al.,1992; Raine and Twomey, 1997). Sterling et al. (2001, 2003) and Dall’Alba et al. (2001) reported decreased active range of movement in whiplash patients and movement deviations either through-range or when the head was held in sustained postures. In our prior field study (Szeto et al., 2002) it was observed that the ‘‘poking chin’’ posture was more common among the symptomatic subjects when they concentrated on viewing the computer display. This increased forward head movement may have gradually developed into a fixed postural habit whenever these individuals worked with computers; and different muscle control strategies may have also developed concurrently. In the present study, by standardizing the physical environment we were able to identify intrinsic differences in the subjects’ postural patterns (not related to the ergonomic set-up) and these were most obvious in the head–neck postures. The results showed that the highly symptomatic individuals displayed what appears to be a mal-adaptive postural control pattern for the

(High vs. Low Discomfort) (High Discomfort vs. Control) (Case vs. Control) (High vs. Low Discomfort) (High Discomfort vs. Control)

head–neck region. This pattern which has the potential to increase biomechanical loading in the musculoskeletal structures relating to the head–neck region, did not alter in spite of the development of increasing neck and shoulder pain over the 1-h typing period. 4.2. Differences in shoulder kinematics The shoulder posture analysis showed a strong side effect in the shoulder flexion angles, as well as the ranges of shoulder flexion and abduction, but not for groups. These findings may reflect the rising trend for intensive mouse use with the right hand associated with computer use, resulting in altered movement control patterns (increased movement of the right arm) observed during the typing task. It is also interesting to note that there was a correlation between right UT activity and shoulder abduction as well as neck flexion and side flexion. These findings indicate that the UT muscle has a dual role both as a stabilizer of the shoulder/upper limb segments as well as the cervical spine/head segment. 4.3. Factors contributing to altered kinematics The link between altered kinematics and muscle recruitment of the head and neck appear to have a relationship to the discomforts experienced by the subjects. However, the study results do not provide an understanding to the basis of these altered motor control patterns, apart from the fact that they do not have an ergonomic basis. There are many possibilities for the altered kinematic patterns observed in the Case Group. The findings may reflect inherent differences in motor control characteristics in different individuals. On the other hand, they may reflect individual differences in terms of social and psychological factors that could be the underlying or predisposing factors to the development of abnormal postural habits. Non-physical factors such

ARTICLE IN PRESS G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 281–291

as personality types, perceived tension, psychosocial stress and motivational factors can significantly affect the perception of musculoskeletal discomforts (Vasseljen and Westgaard, 1995; Linton, 2000; Marras et al., 2000; Roe et al., 2001). These non-physical factors may have caused some individuals to adopt more ‘‘flexed’’ postures during work, which gradually became ‘‘natural’’ to them. Adoption of a flexed posture may also be part of the ‘‘flexor withdrawal reflex’’ associated with a pain response to sensitization of central and/or peripheral nervous tissues (Sterling et al., 2001). However, it was interesting to note that the altered kinematics and muscle recruitment patterns remained constant over the typing period in spite of increasing symptoms suggesting that a reflex or protective response to pain did not occur in these subjects over the observation period. Further research is needed to explore the differences and associations in the sub-groups among symptomatic individuals that contribute to their differences in postural habits and motor control. The present results suggested that the inter-individual differences in posture and muscle recruitment were likely due to intrinsic mechanisms, rather than simply responses to the ergonomic or physical environmental conditions in performing occupational tasks. The results support that altered kinematics associated with muscle recruitment changes may be an important mechanism related to WRNULD in a sub-group of subjects. Muscle activation is closely related to the control of joint movements and postures and it is difficult to separate the influence of the two components. A schematic

289

diagram depicting the relationship of altered muscle recruitment and altered kinematics is presented in Fig. 3. Whatever the cause for the observed findings, both the altered muscle recruitment patterns and altered kinematics appear to be a poor adaptation for pain of the head–neck region, as they are likely to result in increased compressive loading in the cervical spine, affecting muscles, articular structures such as zygapophyseal joints, connective tissues and neural tissues which are all peripheral generators of referred pain (Bogduk, 1995). Increased loading of already sensitized cervical structures may have contributed to the widespread and bilateral pain patterns observed in the High Discomfort Group subjects in this study (Szeto et al., 2005). It is also interesting to note that not all the symptomatic individuals presented in the same manner suggesting different clinical sub-groups exist with WRNULD. Further research is clearly required to identify characteristics with these different sub-groups, and investigate other physical and psycho-social factors known to be associated with chronic musculoskeletal pain disorders. Prospective studies are required to determine whether these findings predispose the onset of pain or whether they occur secondary to pain and/or other work related stressors.

5. Conclusion The present study has shown differences in head–neck posture in the Case Group compared to the Control High Discomfort Group

Altered (Mal-adaptive) Motor Control Altered Muscle Recruitment Patterns

Physical factors: individual or inherent

Case Group

Physical stressors: e.g. Static posture

Non-physical factors: Psychological and social Control Group

Discomforts

Altered Kinematics

Low Discomfort Group ?? Altered Motor Control (may be less extensive and only activated if sufficiently stressed)

Normal Adaptive Motor Control: Minimal changes in muscle activities & kinematics

Discomforts

Minimal Discomforts

Fig. 3. The Altered Motor Control Model based on the results of altered kinematics and altered muscle recruitment in the present study.

ARTICLE IN PRESS 290

G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 281–291

Group during the performance of a 1-h continuous typing task. The head angles were significantly correlated with muscle activity of the right UT muscle and with neck and shoulder discomfort. These findings suggest that altered motor control strategies for the head and neck are present in symptomatic individuals, and are independent of the ergonomic set-up. Further research is required to determine the exact relationship of the findings to WRNULD and whether specific interventions can positively change them.

Acknowledgements The authors would like to thank the Occupational Safety and Health Council of Hong Kong for funding this research project. We would also like to acknowledge Mr. Paul Davey for developing the Labview program to process the kinematics data.

References Aara˚s A, Fostervold KI, Ro O, Thoresen M. Postural load during VDU work: a comparison between various work postures. Ergonomics 1997;40(11):1255–68. Anglin C, Wyss UP. Review of arm motion analyses. Proceedings of Institution of Mechanical Engineers 2000;214(Part H):541–55. Ariens GAM, Bongers PM, Douwes M, Miedema MC, Hoogendoorn WE, van der Wal G, et al. Are neck flexion, neck rotation, and sitting at work risk factors for neck pain? Results of a prospective cohort study. Occupational and Environmental Medicine 2001;58:200–7. Bernard B, Sauter S, Peterson M, Fine L, Hales T. Job task and psychosocial risk factors for work-related musculoskeletal disorders among newspaper employees. Scandinavian Journal of Work, Environment and Health 1994;20:417–26. Bogduk N. The anatomical basis for spinal pain syndromes. Journal of Manipulative and Physiological Therapeutics 1995;18(9):603–5. Braun BL. Postural differences between asymptomatic men and women and craniofacial pain patients. Archives of Physical Medicine and Rehabilitation 1991;72:653–6. Burgess-Limerick R, Plooy A, Fraser K, Ankrum D. The influence of computer display height on head and neck posture. International Journal of Industrial Ergonomics 1999;23:171–9. Chaffin DB, Andersson GB, Martin BJ. Occupational biomechanics. 3rd ed. New York: Wiley; 1999. Dall’Alba PT, Sterling MM, Treleaven JM, Edwards SL, Jull GA. Cervical range of motion and discriminates between asymptomatic persons and those with whiplash. Spine 2001;26(19):2090–4. Ehara Y, Fujimoto H, Miyazaki S, Mochimaru M, Tanaka S, Yamamoto S. Comparison of the performance of 3D camera systems II. Gait and Posture 1997;5:251–5. Finley M, Lee R. Effect of sitting posture on 3-dimensional scapular kinematics measured by skin-mounted electromagnetic tracking sensors. Archives of Physical Medicine and Rehabilitation 2003;84:563–8. Griegal-Morris P, Larson K, Mueller-Klaus K, Oatis CA. Incidence of common postural abnormalities in the cervical, shoulder, and thoracic regions and their association with pain in two age groups of healthy subjects. Physical Therapy 1992;72(6):425–30.

Harms-Ringdahl K, Ekholm J, Schuldt K, Nemeth G, Arborelius UP. Load moments and myoelectric activity when the cervical spine is held in full flexion and extension. Ergonomics 1986;29(12):1539–52. Hermans V, Hautekiet M, Spaepen AJ. Relation between posture and neck–shoulder muscular effort. In: Kumar S, editor. Advances in occupational ergonomics and safety. New York: IOS Press; 1998. p. 202–5. Hagg GM, Astrom A. Load pattern and pressure pain threshold in the upper trapezius muscle and psychosocial factors in medical secretaries with and without shoulder/neck disorders. International Archives of Occupational and Environmental Health 1997;69: 423–32. Jensen C, Nilsen K, Hansen K, Westgaard RH. Trapezius muscle load as a risk indicator for occupational shoulder–neck complaints. European Journal of Applied Physiology and Occupational Physiology 1993;64:415–23. Johnson G, Bogduk N, Nowitzke A, House D. Anatomy and actions of the trapezius muscle. Clinical Biomechanics 1994;9:44–50. Jonsson B. Measurement and evaluation of local muscular strain in the shoulder during constrained work. Journal of Human Ergology 1982;11:73–88. Kebaetse M, McClure P, Pratt N. Thoracic position effect on shoulder range of motion, strength, and three-dimensional scapular kinematics. Archives of Physical Medicine and Rehabilitation 1999;80:945–50. Keshner EA, Campbell D, Katz RT, Peterson BW. Neck muscle activation patterns in humans during isometric head stabilization. Experimental Brain Research 1989;75:335–44. Kroemer K, Grandjean E. Fitting the task to the human, 5th ed. London: Taylor & Francis; 1997. Li G, Buckle P. Current techniques for assessing physical exposure to work-related musculoskeletal risks, with emphasis on posturebased methods. Ergonomics 1999;42(5):674–95. Liao MH, Drury CG. Posture, discomfort and performance in a VDT task. Ergonomics 2000;43(3):345–59. Linton SJ. A review of psychological risk factors in back and neck pain. Spine 2000;25(9):1148–56. Madeleine P, Lundager B, Voigt M, Arendt-Nielsen L. Shoulder muscle co-ordination during chronic and acute experimental neck–shoulder pain, an occupational pain study. European Journal of Applied Physiology and Occupational Physiology 1999;83:127–40. Marras W, Davis KG, Heaney CA, Maronitis AB, Allread WG. The influence of psychosocial stress, gender, and personality on mechanical loading of the lumbar spine. Spine 2000;25(23): 3045–54. Nordander C, Hansson G-A, Rylander L, Asterland P, Bystrom JU, Ohlsson K, Balogh I, Skerving S. Muscular rest and gap frequency as EMG measures of physical exposure: the impact of work tasks and individual related factors. Ergonomics 2000;43(11):1904–19. 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. Rau G, Disselhorst-Klug C, Schmidt R. Movement biomechanics goes upwards: from the leg to the arm. Journal of Biomechanics 2000;33:1207–16. Roe C, Bjorklund RA, Knardahl S, Waersted M, Vollestad NK. Cognitive performance and muscle activation in workers with chronic shoulder myalgia. Ergonomics 2001;44(1):1–16. Saito S, Miyao M, Kondo T, Sakakibara H, Toyoshima H. Ergonomic evaluation of working posture of VDT operation using personal computer with flat panel display. Industrial Health 1997;35:264–70. Schuldt K, Ekholm J, Harms-Ringdahl K, Nemeth G, Arborelius UP. Effects of changes in sitting work posture on static neck and shoulder muscle activity. Ergonomics 1986;29(12):1525–37. Sterling M, Jull G, Wright A. The effect of musculoskeletal pain on motor activity and control. The Journal of Pain 2001;2(3):135–45.

ARTICLE IN PRESS G.P.Y. Szeto et al. / Manual Therapy 10 (2005) 281–291 Sterling M, Jull G, Vicenzino B, Kenardy J, Darnell R. Development of motor system dysfunction following whiplash injury. Pain 2003;103:65–73. Straker L, Jones KJ, Miller J. A comparison of the postures assumed when using laptop computers and DT computers. Applied Ergonomics 1997;28:263–9. 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 LM, O’Sullivan PB. A comparison of symptomatic and asymptomatic office workers performing monotonous keyboard work—1: Neck and shoulder muscle recruitment patterns. Manual Therapy 2005, in press, doi:10.1016/j.math. 2005.01.004. Tittiranonda P, Burastero S, Rempel D. Risk factors for musculoskeletal disorders among computer users. Occupational Medicine: State of the Art Reviews 1999;4(1):17–38.

291

Turville KL, Psihogios JP, Ulmer TR, Mirka GA. The effects of video display terminal height on the operator: a comparison of the 151 and the 401 recommendations. Applied Ergonomics 1998;29(4):239–46. Vasseljen O, Westgaard RH. Can stress-related shoulder and neck pain develop independently of muscle activity. Pain 1995;64:221–30. Vasseljen O, Westgaard RH. Arm and trunk posture during work in relation to shoulder and neck pain and trapezius activity. Clinical Biomechanics 1997;12(1):22–31. Vicon 370 Users’ Manual, Version 3.1. Oxford, UK: Oxford Metrics Ltd., 1997. Villanueva MB, Sotoyama M, Jonai H, Takeuchi Y, Saito S. Adjustments of posture and viewing parameters of the eye to changes in the screen height of the visual display terminal. Ergonomics 1996;39(7):933–45. Westgaard RH. Work-related musculoskeletal complaints: some ergonomics challenges upon the start of a new century. Applied Ergonomics 2000;31:569–80.

ARTICLE IN PRESS

Manual Therapy 10 (2005) 292–296 www.elsevier.com/locate/math

Case Report

The T4 syndrome Jenny Louise Conroy, Anthony G. Schneiders School of Physiotherapy, University of Otago, P.O. Box 56, Dunedin, New Zealand Received 7 April 2004; received in revised form 11 January 2005; accepted 20 January 2005

1. Introduction The term ‘T4 syndrome’ is a clinical pattern that involves upper extremity paraesthesia and pain with or without symptoms into the neck and/or head (Maitland, 1986). Mobilization of an upper thoracic vertebrae (commonly T4), reproduces or eliminates these symptoms (Grieve, 1988), although the mechanism for this remains unclear. The presence of a hypomobile thoracic segment may indicate involvement of a synovial joint structure (Bogduk, 1986). However, this is unlikely to somatically refer symptoms to the upper extremity (Grieve, 1988). The sympathetic nervous system may provide a pathway for referral from the thoracic spine to the head and arms, but the link between the sympathetic and the somatic nervous system is not clearly understood (Evans, 1997). The purpose of this case study is to describe a patient who presented with symptoms that closely resembled the clinical condition known as ‘T4 syndrome’. As this syndrome is poorly defined in the literature, this study will detail the history and clinical interpretation of symptoms. This patient responded to thoracic mobilizations and the proposed mechanisms by which thoracic dysfunction are thought to cause upper extremity pain and paraesthesia are discussed.

2. History A 28-year-old female university student presented to physiotherapy with pain across both shoulders (PA) and bilaterally down the arms (PB) (see Fig. 1). These Corresponding author. Tel.: +64 3 4795426; fax: +64 3 4798414.

E-mail address: [email protected] (A.G. Schneiders). 1356-689X/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2005.01.007

symptoms were preceded by paresthesia in the form of tingling in the palm of both hands, and if the pain worsened, a headache (PC) and tingling across the face would develop (see Fig. 1). Limb symptoms were bilateral with the left being more intense. There was no associated dizziness, blurred vision or nausea. Symptoms would begin within a couple of hours of arising in the morning. Sitting in lectures or studying for more than one hour would exacerbate the pain across the patient’s shoulders (PA). Changing position or walking around would not change the level of pain once exacerbated, although a hot shower and massage to the patient’s shoulders and upper back would occasionally relieve it. This patient’s shoulder and arm symptoms started gradually 2 months previously at the end of a period of intense reading and studying slumped into flexion on a couch. Three weeks prior to presentation she had swum for 20 minutes, which was the first swim in 2 years. The following morning the patient woke feeling stiff in her thoracic spine, and over the next 3 days her shoulder and arm pains gradually worsened. The tingling symptoms into the patient’s hands and fingers also started around this time. The pain started interrupting sleep and she presented herself to the Emergency Department at the local hospital. Routine investigations were considered normal and she was sent home with Tilcotil, Tramol and Diazepam. The patient’s symptoms eased with medication, but 4 days later they worsened for no apparent reason. At the visit to the Emergency Department for a second time, she was admitted and underwent an assessment by neurology and orthopaedic specialists. A lumbar spine MRI, cervical spine X-ray and kidney ultrasound excluded further serious pathology, and she was referred for physiotherapy.

ARTICLE IN PRESS J.L. Conroy, A.G. Schneiders / Manual Therapy 10 (2005) 292–296

293

Fig. 1. Body chart and symptom behaviour. Int ¼ intermittent, sup ¼ superficial.

On presentation to the clinic, symptoms had settled, but she was continuing to take pain-relieving medication as needed. Amitriptyline had also been prescribed for anxiety. Past history revealed long-standing anxiety and lower back and left leg pain. Six months previously the patient had a suspected prolapsed lumbar disc and had been prescribed Codeine to be taken as required. An orthopaedic and musculoskeletal consultation had been arranged. The patient considered her general health to be good. However, she believed that anxiety was playing a role in exacerbating symptoms. Due to the nature of her pain the patient could only study part-time, with her goal to return to fulltime study. The patient was not currently active in any form of exercise, but had swum and rockclimbed 2 years previously.

3. Physical examination Postural examination in sitting revealed an increased lumbar lordosis, decreased upper thoracic kyphosis, increased angulation at the cervico-thoracic junction and an increased cervical lordosis. Palpation found the upper fibres of trapezius to be taut and elicited pain bilaterally. Cervical extension and right rotation were full and pain-free. Cervical flexion reproduced pain across both the shoulders (PA) and was limited to within 5 cm from the chest. Left rotation was 601 and side flexion was 401 bilaterally. All thoracic movements were pain-free but restricted early in range, especially left rotation and flexion. Flexion of the thoracic spine occurred from above the T4 vertebra with minimal intervertebral movement below this level. No abnormal neurological signs were detected, although the upper

ARTICLE IN PRESS 294

J.L. Conroy, A.G. Schneiders / Manual Therapy 10 (2005) 292–296

Fig. 2. kPA to T4 in prone lying. R1 ¼ onset of resistance; R2 ¼ resistance limits movement; P1 ¼ onset of pain; P0 ¼ pain at limit of movement; L ¼ limit of the range.

limb neurodynamic test of both arms was limited by pain following scapular depression, shoulder abduction (1101), extension (101), external rotation and elbow extension to 401. Central postero-anterior (kPA) vertebral pressure to T4 reproduced the pain across both the shoulders (PA) and was restricted early in range by resistance (see Fig. 2).

The differential diagnoses for these symptoms include thoracic outlet syndrome, cervical spine dysfunction and cardiac or neurological disease (DeFranca and Levine, 1995). Cardiac pain refers to the left arm and chest. However, it is activity dependant and eases with rest (Evans, 1997). Thoracic outlet syndrome typically exhibits nocturnal symptoms resulting from neurovascular compression. Pulse deficits or venous engorgement and signs of lower brachial compression are common (DeFranca and Levine, 1995). Cervical disc disease may present with neck and arm symptoms and limited cervical range of motion. However, symptoms usually follow a dermatomal or sclerotomal distribution (DeFranca and Levine, 1995). In this report, the patient underwent a kidney ultrasound and lumbar spine MRI to exclude kidney disease or intervertebral disc pathology. The kidney receives autonomic innervation from T10–L1, therefore any symptoms around the thoracolumbar area may be indicative of internal pathology. Furthermore, these investigations would exclude a spinal tumour or metastases, which can mimic thoracic joint pain (DeFranca and Levine, 1995).

5. Treatment Prognostic indicators such as the chronic nature of the condition and the unknown component her anxiety may have had on symptom manifestation, suggested that symptoms would improve but over several months.

4. Clinical interpretation 5.1. Treatment 1 ‘T4 syndrome’ typically presents with unilateral or bilateral glove distribution of paraesthesia into the hands. However, the syndrome is not clearly defined in the literature and Grieve (1994) stated that paraesthesia involving the hand to any extent, but not in a dermatomal distribution, can still be indicative of a T4 dysfunction. Night or early morning pain or paraesthesia is also considered common (Evans, 1997). Typical occupations of those that have a diagnosis of T4 syndrome involve forward stooping and bending or sedentary seated positions, and it is suggested that these people develop postural faults (DeFranca and Levine, 1995). This patient reported pain and paraesthesia in the upper limbs in combination with neck, upper thoracic and cranial pain and no abnormal neurological signs, which are consistent with clinical features reported in the literature (Bogduk, 1986; Grieve, 1994; Evans, 1997). The testing procedure that reproduced her symptoms was a central postero-anterior accessory movement of the T4 vertebra, indicating a T4 dysfunction (Evans, 1997). The clinical entity ‘T4 syndrome’ was used to describe this patient’s symptoms, which were generally consistent with the literature.

Treatment was based on interpretation of the movement diagram, Fig. 2. A grade IIIcentral posteroanterior (PA) mobilization of the fourth thoracic vertebra was performed for 20 s with the patient lying prone to restore normal mobility of the T4/5 joint (Maitland, 1986). On reassessment, thoracic movements were unchanged. However, full left cervical rotation was achieved along with an increase of 101 of side flexion bilaterally. 5.2. Treatment 2 Two days later, the patient reported that her symptoms had not worsened over the course of the day, which was a normal occurrence prior to her seeking treatment. Her resting shoulder pain (PA) and the tingling in her palms were still present. Central posteroanterior mobilizations were repeated to T4 in slight thoracic flexion, as this position would reproduce her symptoms when sitting. Full rotation of both the cervical and thoracic spines was achieved bilaterally, but thoracic flexion remained unchanged. The patient

ARTICLE IN PRESS J.L. Conroy, A.G. Schneiders / Manual Therapy 10 (2005) 292–296

295

6. Discussion

Fig. 3. kPA to T4 in prone lying, 3 weeks from initial assessment. R1 ¼ onset of resistance, R2 ¼ resistance limits movement, P1 ¼ onset of pain, P0 ¼ pain at limit of movement.

was educated on postures to avoid and specific postural corrections in sitting were shown.

5.3. Subsequent treatments Over the next four treatments, the mid-thoracic pain (PD) localized to the T4 location dorsally. Upper limb, neck and head symptoms disappeared and she was able to discontinue her pain medication, with the exception of Amitriptyline that she continued to take for anxiety. Mobilizations were progressed to a grade III in further thoracic flexion. Full range was achieved by central postero-anterior vertebral pressure to T4 by the sixth treatment, three weeks from the initial assessment (see Fig. 3). The patient gained full thoracic and cervical movement, although she continued to hinge from the T4 segment on forward flexion. The shoulder pain on the left (PA) would occasionally develop when studying or writing on the computer. Correcting her posture to sit straight or by using a hot pack over her shoulders eased the shoulder pain (PA). It was decided at this point to refer her onto a Pilates exercise programme to focus on upper back mobility and general trunk stabilization. This was anticipated to help maintain a more normal spinal position since poor and prolonged posturing, as seen in this patient, is associated with changes in muscle length and strength over time (Sahrmann, 2002). Muscle imbalances around the upper thoracic spine may contribute to continued thoracic joint and muscle dysfunction (DeFranca and Levine, 1995). Unfortunately, this patient did not attend her appointment for a Pilates programme and could not be contacted by phone. A follow-up phone call was made 6 months later, but the phone was disconnected.

In this case report, upper limb symptoms and head pain were alleviated by mobilization of the T4 vertebra. This is consistent with the findings of DeFranca and Levine (1995). Postural exercises were also emphasized over the course of treatment. However, static postural measures remained unchanged during treatment. Improvement in symptoms was noted with the mobilizations performed on day 1 and this was maintained until the next reassessment. Specific postural exercises were not introduced until the second treatment, therefore it is suggested that mobilization primarily resolved symptoms. A joint manipulation of the T4 vertebra is an option in treating this syndrome (DeFranca and Levine, 1995). It was not chosen for this patient due to the unknown component her anxiety had on symptoms and because her pain was deemed to be irritable. It is widely considered that the autonomic nervous system provides a pathway for dysfunctions of the thoracic spine to be expressed over the lower cervical spine and down the upper limb (Grieve, 1994; Evans, 1997; Bogduk, 2002), although the interactions between the somatic and autonomic systems are not widely understood. Sympathetic fibres leave the spinal nerve from levels T1–L2 to join the sympathetic chain via the white rami communicantes. They then travel within the sympathetic chain for up to six segments before synapsing on between four and twenty postganglionic neurons. The postganglionic neurons then exit via the grey rami communicantes to rejoin a peripheral nerve and are distributed to the target tissues (Evans, 1997). One preganglionic neuron synapsing with numerous postganglionic neurons in the sympathetic chain therefore interacts with somatic nerve fibres supplying a variety of target tissues. The head and neck are supplied by levels T1–T4, and the upper trunk and upper limb by T1–T9 (Bogduk, 2002). It is therefore postulated that dysfunction of the sympathetic nervous system from T4 could result in referred pain in the head, neck, upper thoracic and upper limbs. Since a possible pathway that connects the upper thoracic to the head and neck exists, the structures at fault need to be identified. Evans (1997) suggested that the joint itself is not the causative factor, but that sustained or extreme postures may lead to relative ischaemia in tissues. The sympathetic nerves also form a vasoconstrictor network on all arterioles and capillaries, and therefore would be stimulated in the presence of ischaemia. The insidious nature of onset, and the prevalence of postural dysfunctions associated with the ‘T4 syndrome’ (DeFranca and Levine, 1995) in the absence of radiographic evidence (Grieve, 1994), supports this theory. However, Grieve (1994) goes on to hypothesize that hypomobile or subclinical degenerative

ARTICLE IN PRESS 296

J.L. Conroy, A.G. Schneiders / Manual Therapy 10 (2005) 292–296

mid-thoracic joints not detected on X-ray signal pain through mechanical and biomechanical irritation of nociceptors in the synovial joint capsule. The proximity of the sympathetic chain to a dysfunctional thoracic joint may predispose the ganglion to mechanical pressure (Menck et al., 2000). Bogduk (1986) proposed that the intervertebral disc could be another source of pain, although the pain pattern observed with a disc lesion is dissimilar to that of the ‘T4 syndrome’ (Maitland, 1986). One way to test the physiological mechanism for the referred pain is to inject the costovertebral, costotransverse or costochondral joint, or the surrounding T4 structures with a noxious substance. Reproduction of symptoms classically seen in a T4 dysfunction, would implicate the underlying source of pain (Bogduk, 1986). This may also help define the type and location of symptoms from a particular structure, hence minimising the ambiguity in the definition of the ‘T4 syndrome’. Mobilization of joints is thought to activate descending inhibitory pain pathways (Zusman, 2002) resulting in a hypoalgesic effect. There is also a close relationship between pain reduction and sympathetic excitation (Vicenzino et al., 1998; Sterling et al., 2001), supporting the role of spinal mobilization as a treatment option for the T4 syndrome. The psychological state of a person can be linked to a somatic dysfunction due to the connections of both psychological and physical aspects of a dysfunction to the central nervous system (Shacklock, 1999). Coexisting factors such as anxiety and lower back pain were not addressed in the treatment of this patient, which could have influenced her symptom presentation. Despite this, treating one dysfunction resulted in an improvement of her upper limb, neck, shoulder and thoracic pains. Manual therapy therefore, may have acted on both these components.

7. Conclusion Unilateral or bilateral upper limb paraesthesia and pain, in association with a dysfunction of an upper thoracic vertebra, may be part of the entity known as ‘T4 syndrome’. It is likely that a hypomobile thoracic

joint or surrounding soft-tissue dysfunction is the causative factor, although scientific evidence for this is currently lacking. This case report found that a symptomatic T4 vertebra responded to central passive mobilizations, improving spinal mobility and reducing pain. The T4 syndrome needs to be considered whenever bilateral upper limb symptoms or paraesthesia are present, as an upper thoracic dysfunction may be the cause.

References Bogduk N. Innervation and pain patterns of the thoracic spine. In: Grant R, editor. Physical therapy of the cervical and thoracic spines. 3rd ed. Edinburgh, Scotland: Churchill Livingstone; 2002. p. 73–84 ch 5. Bogduk N. Research design or what is T4 syndrome. Define exactly what the phenomenon of interest is. New Zealand Journal of Physiotherapy 1986;14(3):9–11. DeFranca CG, Levine LJ. The T4 syndrome. Journal of Manipulative and Physiological Therapeutics 1995;18(1):34–7. Evans P. The T4 syndrome: some basic science aspects. Physiotherapy 1997;83(4):186–9. Grieve GP. The autonomic system in vertebral pain syndromes. In: Grieve GP, editor. Common vertebral joint problems. 2nd ed. Edinburgh, Scotland: Churchill Livingstone; 1988. p. 319–29 ch 7. Grieve GP. Thoracic musculoskeletal problems. In: Boyling J, Palastanga N, editors. Grieve’s Modern Manual Therapy. 2nd ed. Edinburgh, Scotland: Churchill Livingstone; 1994. p. 401–28 ch 29. Maitland GD. Vertebral manipulation, 5th ed. Oxford, England: Butterworth Heinemann; 1986. Menck JY, Requejo SM, Kulig K. Thoracic spine dysfunction in upper extremity complex regional pain syndrome type I. Journal of Orthopaedic and Sports Physical Therapy 2000;30(7):401–9. Sahrmann SA. Diagnosis and treatment of movement impairment syndromes. Mosby, St Louis, Missouri; 2002. Shacklock MO. Central pain mechanisms: a new horizon in manual therapy. Australian Journal of Physiotherapy 1999;45:83–92. Sterling M, Jull G, Wright A. Cervical mobilization: concurrent effects on pain, sympathetic nervous system activity and motor activity. Manual Therapy 2001;6(2):72–81. Vicenzino B, Collins D, Benson H, Wright A. An investigation of the interrelationship between manipulative therapy-induced hypoalgesia and sympathoexcitation. Journal of Manipulative and Physiological Therapeutics 1998;21(7):448–53. Zusman M. Forebrain-mediated sensitization of central pain pathways: ‘non-specific’ pain and a new image for MT. Manual Therapy 2002;7(2):80–8.

ARTICLE IN PRESS

Manual Therapy 10 (2005) 297–298 www.elsevier.com/locate/math

Letters to the Editor

The ‘vertebral artery test’ The ‘‘vertebral artery test’’ (cervical rotation) has been widely touted and embraced worldwide and across professions as one of the methods of screening the vertebro-basilar system prior to manipulative therapy (e.g. recently published APA guidelines) despite the fact that it has no published sensitivity or specificity data. The interesting study by Mitchell et al. (2004) is a welcome addition to the vascular knowledge base for all Manual Therapists who treat the cervical spine either with prescribed exercise, non-manipulative manual techniques or manipulation. The findings of this study are well presented and contribute cumulative knowledge to this vital area. However, we feel that the discussion and conclusion (abstract), which state ‘‘this study supports the use of the vertebrobasilar insufficiency (VBI) test’’ represents a leap of faith, and as such may be misinterpreted by the reader. Mitchell et al. (2004) report on young ‘‘normal’’ females and demonstrate that, in this specific group, a clinical test (sustained end range cervical rotation) produces a haemodynamic response. That is, a reduction of flow in the contra-lateral vertebral arteries (VA) and because those subjects have adequate compensatory mechanisms, no symptoms are produced. The findings are important to us because they show that a clinical test has a demonstrable effect in normal subjects and indeed backs up the theoretical concept of vertebro-basilar compensation. However, this information alone is not enough to support or add credence to the use of the test (as suggested by the authors). Over recent years, many other studies on asymptomatic and symptomatic subjects have reported little agreement between positive objective findings and symptom reproduction (see Terrett, 2001 for an overview of flow study results). The usefulness of a test is judged on how well it can separate those with a disease (or dysfunction) from those without. This concept can be expressed as the test’s likelihood ratio, or sensitivity and specificity (Gilbert et al., 2001). In order for a test to have good sensitivity and specificity, there must be good agreement (positive correlation) between the physical sign (in this case vertebral artery blood flow) and the symptom

(e.g. dizziness, nausea, etc.). The current knowledge base does not come near to supporting this level of correlation and the present study’s findings are not what would be needed to support the utility of the test. In line with our current state of knowledge, clinically there is little evidence that a ‘negative’ test predicts either the absence of arterial pathology or the propensity of the artery to be injured during treatment (i.e. to filter out those ‘‘at risk’’ patients—the rationale for the test). We commend Mitchell et al. (2004) and the authors of previous blood flow studies, for their valuable work which forms a significant addition to knowledge in this complex clinical area. However, we feel that sound judgement is required when drawing conclusions, to allow the reader to clearly discern between results which appear to back up a theoretical basis for testing, and those which actually add support to the utility of the test. The VBI test remains without known sensitivity or specificity and therefore clinical decisions made on the basis of the results of such a test should be carefully thought out. It is our belief (based on an extensive literature search due for publication in the near future) that rather than relying on guidelines or clinical tests of dubious or unknown utility, there is a need for a wider consideration of haemodynamic issues within manual therapy training and education. This should include the theories of vascular injury and thrombogenesis together with the factors that influence them. The paper published in issue 10/2 of Manual Therapy (Thiel and Rix, 2005) debates the issue from an interesting perspective and indeed mirrors many of our thoughts and findings, far better than we could do in a short correspondence. We urge all manual therapists to read it and contribute to this important debate in any way they can.

Reference Gilbert R, Logan S, Moyer VA, Elliott EJ. Assessing diagnostic and screening tests: Part 1. Concepts. Western Journal of Medicine 2001;174(6):405–9. Mitchell J, Keene D, Dyson C, Harvey L, Pruvey C, Phillips R. Is cervical spine rotation, as used in the standard vertebrobasilar insufficiency test, associated with a measureable change in intracranial vertebral artery blood flow? Manual Therapy 2004; 9(4):220–7.

ARTICLE IN PRESS 298

Letters to the Editor / Manual Therapy 10 (2005) 297–298

Terrett AJ. Current concepts—vertebrobasilar complications following spinal manipulation. USA: NCMIC Iowa; 2001. Thiel H, Rix G. Is it time to stop functional pre-manipulative testing of the cervical spine? Manual Therapy 2005;10(2):154–8.

Alan J. Taylor, Roger Kerry Nottingham Nuffield Hospital, 748 Mansfield Road, Woodthorpe, Nottingham NG5 3FZ, UK

doi:10.1016/j.math.2005.02.005

Reply to comments on our paper: ‘‘Is cervical spine rotation, as used in the standard vertebrobasilar insufficiency test, associated with a measurable change in intracranial vertebral artery blood flow?’’ In the interests of pursuing the debate of VBI and related issues, we appreciate the comments made by our colleagues in their letter (This issue), and would like to take this opportunity to address some of the points raised. Our suggestion in the abstract that our results support the use of the VBI test, must be taken in context. The rest of the sentence indicates that the VBI test may be useful ‘‘yin the absence of a more specific, sensitive and valid test,y’’. This, we feel, is clarified in the body of the text. In the Introduction (p. 221, paragraph 4), we note from a review of the literature that ‘‘ythe justification of the use of cervical rotation in isolation as a pre-manipulative screening procedure is tenuous.’’, and in the Discussion (p. 225, paragraph 4), ‘‘the VBI testyis of use for potentially identifying the patients at risky’’. However, we agree with our colleagues in that our findings in this study do not support the use of the VBI test alone, in identifying at-risk patients. We acknowledge that, although the VBI test may, in some patients who show some signs of VBI (i.e. dizziness), suggest that the individual has an already compromised blood flow before end-range cervical spine rotation, it does not identify vascular pathology, where present, in patients for whom the test is negative. More importantly, as our colleagues imply, the test does not identify those patients in whom the vertebral artery wall may be injured, by stretching for example, putting that individual at risk of developing atherosclerosis of that part of the vessel, or thrombi and/or emboli from the atherosclerotic plaques. We suggest in this context (p. 225, paragraph 4) that Doppler sonography also be used, where possible, as a measure of blood flow and, by implication, of vascular patency. This point may have been more strongly doi:10.1016/j.math.2005.03.004

debated in the paper and is worth consideration in practice and for future research. Therefore, we endorse our colleagues’ assertion that the VBI test is still without demonstrated or proven validity, sensitivity and specificity. We state in the Introduction (p. 221, paragraph 4) that ‘‘ydebate has been prompted on the efficacy of end-of-range cervical spine rotation as a test for VBI, and on the validity, specificity and sensitivity of this procedure as a premanipulative screening tool.’’, highlighting the uncertainty about the validity of the VBI test and the effect of cervical spine rotation on blood flow. We would like to stress our assertion that ‘‘yfurther research ywould be of valuey’’ (p. 221, paragraph 4), and add that the need is for more rigorously controlled research. Our colleagues also raise the issue of the need for a greater understanding of haemodynamic theory related to vascular pathology and injury. Many therapists in this field would agree with this, and we have attempted to address this issue in a review paper: ‘‘The vertebral artery: a review of anatomical, histopathological and functional factors influencing blood flow to the hindbrain’’ (Physiotherapy Theory and Practice 2005, 21(1), in press). Although we have attempted, in this study of vertebral artery blood flow, to provide further evidence for the dangers of sustained end-of-range cervical rotation for some patients, we agree that there is still no convincing evidence of a method for identifying all at-risk patients. We would like to join our colleagues in appealing to clinicians and researchers alike, to address this problem and perhaps form collaborative links to help plan further feasible studies to provide some evidence on which to base future practice. Jeanette Mitchell (January, 2005) J. Mitchell Department 3166, University of Wyoming, 1000 E. University Avenue, Laramie, WY 82071, USA E-mail address: [email protected]

ARTICLE IN PRESS 298

Letters to the Editor / Manual Therapy 10 (2005) 297–298

Terrett AJ. Current concepts—vertebrobasilar complications following spinal manipulation. USA: NCMIC Iowa; 2001. Thiel H, Rix G. Is it time to stop functional pre-manipulative testing of the cervical spine? Manual Therapy 2005;10(2):154–8.

Alan J. Taylor, Roger Kerry Nottingham Nuffield Hospital, 748 Mansfield Road, Woodthorpe, Nottingham NG5 3FZ, UK

doi:10.1016/j.math.2005.02.005

Reply to comments on our paper: ‘‘Is cervical spine rotation, as used in the standard vertebrobasilar insufficiency test, associated with a measurable change in intracranial vertebral artery blood flow?’’ In the interests of pursuing the debate of VBI and related issues, we appreciate the comments made by our colleagues in their letter (This issue), and would like to take this opportunity to address some of the points raised. Our suggestion in the abstract that our results support the use of the VBI test, must be taken in context. The rest of the sentence indicates that the VBI test may be useful ‘‘yin the absence of a more specific, sensitive and valid test,y’’. This, we feel, is clarified in the body of the text. In the Introduction (p. 221, paragraph 4), we note from a review of the literature that ‘‘ythe justification of the use of cervical rotation in isolation as a pre-manipulative screening procedure is tenuous.’’, and in the Discussion (p. 225, paragraph 4), ‘‘the VBI testyis of use for potentially identifying the patients at risky’’. However, we agree with our colleagues in that our findings in this study do not support the use of the VBI test alone, in identifying at-risk patients. We acknowledge that, although the VBI test may, in some patients who show some signs of VBI (i.e. dizziness), suggest that the individual has an already compromised blood flow before end-range cervical spine rotation, it does not identify vascular pathology, where present, in patients for whom the test is negative. More importantly, as our colleagues imply, the test does not identify those patients in whom the vertebral artery wall may be injured, by stretching for example, putting that individual at risk of developing atherosclerosis of that part of the vessel, or thrombi and/or emboli from the atherosclerotic plaques. We suggest in this context (p. 225, paragraph 4) that Doppler sonography also be used, where possible, as a measure of blood flow and, by implication, of vascular patency. This point may have been more strongly doi:10.1016/j.math.2005.03.004

debated in the paper and is worth consideration in practice and for future research. Therefore, we endorse our colleagues’ assertion that the VBI test is still without demonstrated or proven validity, sensitivity and specificity. We state in the Introduction (p. 221, paragraph 4) that ‘‘ydebate has been prompted on the efficacy of end-of-range cervical spine rotation as a test for VBI, and on the validity, specificity and sensitivity of this procedure as a premanipulative screening tool.’’, highlighting the uncertainty about the validity of the VBI test and the effect of cervical spine rotation on blood flow. We would like to stress our assertion that ‘‘yfurther research ywould be of valuey’’ (p. 221, paragraph 4), and add that the need is for more rigorously controlled research. Our colleagues also raise the issue of the need for a greater understanding of haemodynamic theory related to vascular pathology and injury. Many therapists in this field would agree with this, and we have attempted to address this issue in a review paper: ‘‘The vertebral artery: a review of anatomical, histopathological and functional factors influencing blood flow to the hindbrain’’ (Physiotherapy Theory and Practice 2005, 21(1), in press). Although we have attempted, in this study of vertebral artery blood flow, to provide further evidence for the dangers of sustained end-of-range cervical rotation for some patients, we agree that there is still no convincing evidence of a method for identifying all at-risk patients. We would like to join our colleagues in appealing to clinicians and researchers alike, to address this problem and perhaps form collaborative links to help plan further feasible studies to provide some evidence on which to base future practice. Jeanette Mitchell (January, 2005) J. Mitchell Department 3166, University of Wyoming, 1000 E. University Avenue, Laramie, WY 82071, USA E-mail address: [email protected]

Manual Therapy (2005) 10(4), 299

Diary of events

conference courses on the major theme of Positive Precise Performance or on the sub-themes of Pain; Lower limb function; Motor control; Musculoskeletal physiotherapy and its relationship to the fitness industry. Submission is online via mpa2005.com.au/submissions. shtml. Closing date for the receipt of submissions is 31 March 2005. Full details www.mpa2005.com.au Email: mpa2005@ meetingplanners.com.au

10–11 November 2005, Warwick The Society For Back Pain Research Meeting . Study Design Workshop . Back Pain and the Intervertebral Disc Abstracts may be submitted on any aspect of research into back pain. The Back Care Medal (formerly National Back Pain Association Medal) and the President’s Medal ('300) will be awarded to the best two papers of the meeting. Accepted abstracts will be published in the Proceedings of the Journal of Bone and Joint Surgery. Student Prize for best paper submitted by PhD student. Instructions for abstracts available at http://www.sbpr.info Closing date for abstract submission is: 31st JULY 2005 Submit abstracts to: SBPRWARWICK2005@BOA. AC.UK

31 March–2 April 2006 6th International Conference on Advances in Osteopathic Research 31st March to 2nd April 2006 FIRST CALL FOR PAPERS contact: www.bcom.ac.uk/research/icaor6

Instructions for abstracts and details of registration may be obtained from: Mrs Hazel Choules Society for Back Pain Research British Orthopaedic Association 35-43 Lincoln’s Inn Fields, London WC2A 3PN Tel: +44(0)20 7405 6507. Fax: +44(0)20 7831 2676 Email: [email protected]

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

Or visit our website at http://www.sbpr.info

Evidence-based manual therapy congress Further information: www.medicongress.com

Thursday, 24th November–Saturday, 26th November 2005 MPA2005–MUSCULOSKELETAL PHYSIOTHERAPY AUSTRALIA 14TH BIENNIAL CONFERENCE Theme: Positive Precise Performance Location: Brisbane Convention and Exhibition Centre, Brisabne, Queensland, Australia Call for Submissions: Musculoskeletal Physiotherapy Australia invites submissions form people interested in presenting papers, posters, workshops or pre or post

Intensive courses in Manual Therapy Further information: http://allserv.rug.ac.be/bvthillo If you wish to advertise a course/conference, please contact: Karen Beeton, Department of Physiotherapy, University of Hertfordshire, College Lane, Hatfield, Herts AL10 9AB, UK. There is no charge for this service.

299

DOI: S1356-689X(05)00139-6

Volume Contents for Vol. 10, 2005 Vol. 10, No. 1 Editorial The use of qualitative research methodologies within musculoskeletal physiotherapy practice A. Grant Review articles Manual therapy treatment of cervicogenic dizziness: a systematic review Susan A. Reid, Darren A. Rivett Can acute low back pain result from segmental spinal buckling during sub-maximal activities? A review of the current literature Richard Preuss, Joyce Fung Original articles Physiotherapy and osteoporosis: practice behaviors and clinicians’ perceptions—a survey Meena M. Sran, Karim M. Khan Shoulder impingement: the effect of sitting posture on shoulder pain and range of motion Michael P. Bullock, Nadine E. Foster, Chris C. Wright Do Norwegian manual therapists provide management for patients with acute low back pain in accordance with clinical guidelines? L.I. Strand, A. Kvale, M. Raheim, A.E. Ljunggren Influence of cranio-cervical posture on three-dimensional motion of the cervical spine Stephen J. Edmondston, Svein-Erik Henne, Winston Loh, Eirik Østvold The impact of neurodynamic testing on the perception of experimentally induced muscle pain M.W. Coppieters, K. Kurz, T.E. Mortensen, N.L. Richards, I.A˚. Skaret, L.M. McLaughlin, P.W. Hodges A descriptive study of the usage of spinal manipulative therapy techniques within a randomized clinical trial in acute low back pain D.A. Hurley, S.M. McDonough, G.D. Baxter, M. Dempster, A.P. Moore Technical and measurement report Measurement of cervical range of motion pattern during cyclic neck movement by an ultrasound-based motion system Shwu-Fen Wang, Chin-Chih Teng, Kwan-Hwa Lin Case reports Neck pain and headache as a result of internal carotid artery dissection: implications for manual therapists Alan J. Taylor, Roger Kerry A patient with severe right frontal headache Henry Tsao Manipulation following regional interscalene anesthetic block for shoulder adhesive capsulitis: a case series Robert E. Boyles, Timothy W. Flynn, Julie M. Whitman List of Reviewers 2004 Dariy of events

1

4

14

21 28

38 44 52

61

68

73 78 80 88 91

Vol. 10, No. 2 Editorial Manual Therapy Journal 10 year anniversary A. Moore, G. Jull Masterclass The management of hamstring injury—Part 1: issues in diagnosis W. Hoskins, H. Pollard Original Articles Size and shape of the posterior neck muscles measured by ultrasound imaging: normal values in males and females of different ages G. Rankin, M. Stokes, D.J. Newham Ultrasound imaging of lumbar multifidus muscle: normal reference ranges for measurements and practical guidance on the technique M. Stokes, G. Rankin, D.J. Newham Immediate effects of thoracic manipulation in patients with neck pain: a randomized clinical trial J.A. Cleland, M.J.D. Childs, M. McRae, J.A. Palmer, T. Stowell Effect of straight leg raise examination and treatment on vibration thresholds in the lower limb: a pilot study in asymptomatic subjects C. Ridehalgh, J. Greening, N.J. Petty Abdominal muscle recruitment during a range of voluntary exercises D.M. Urquhart, P.W. Hodges, T.J. Allen, I.H. Story

300

93

96

108

116 127

136 144

301 Profesional Issue Is it time to stop functional pre-manipulation testing of the cervical spine? H. Thiel, G. Rix Case Reports A shoulder derangement A. Aina, S. May Manipulation following regional interscalene anesthetic block for shoulder adhesive capsulitis: a case series R.E. Boyles, T.W. Flynn, J.M. Whitman Letter to the Editor Diary of events

154

159 164 172 174

Vol. 10, No. 3 Editorial Improving application of neurodynamic (neural tension) testing and treatments: A message to researchers and clinicians M. Shacklock Masterclass Hamstring injury management—Part 2: Treatment W. Hoskins, H. Pollard Original Articles Reliability of palpation of humeral head position in asymptomatic shoulders D. Bryde, B.J. Freure, L. Jones, M. Werstine, N.K. Briffa A normative database of lumbar spine ranges of motion M. Troke, A.P. Moore, F.J. Maillardet, E. Cheek Diagnosis of Sacroiliac Joint Pain: Validity of individual provocation tests and composites of tests M. Laslett, C.N. Aprill, B. McDonald, S.B. Young Technical and Measurement Report Intra- and inter-rater reliability of the anterior atlantodental interval measurement from conventional lateral view flexion/extension radiographs M.D. Westaway, W.Y. Hu, P.W. Stratford, M.E. Maitland Case Report The use of manipulation in a patient with an ankle sprain injury not responding to conventional management: a case report J.M. Whitman, J.D. Childs, V. Walker Letters to the Editor Book Review MACP — Manipulation Association of Chartered Physiotherapist UK announcement Diary of events

175

180

191 198 207

219

224 232 235 236 237

Vol. 10, No. 4 Editorial Why is the recent research regarding non-specific pain so non-specific? C.J. McCarthy, M.C. Cairns Masterclass Diagnosis and classification of chronic low back pain disorders: Maladaptive movement and motor control impairments as underlying mechanism P. O’Sullivan Review Article Inter-examiner reliability of passive assessment of intervertebral motion in the cervical and lumbar spine: A systematic review E. van Trijffel, Q. Anderegg, P.M.M. Bossuyt, C. Lucas Original Articles A comparison of symptomatic and asymptomatic office workers performing monotonous keyboard work—1: Neck and shoulder muscle recruitment patterns G.P.Y. Szeto, L.M. Straker, P.B. O’Sullivan A comparison of symptomatic and asymptomatic office workers performing monotonous keyboard work—2: Neck and shoulder kinematics G.P.Y. Szeto, L.M. Straker, P.B. O’Sullivan Case Report The T4 syndrome J.L. Conroy, A.G. Schneiders Letters to the Editor Diary of events Volume Contents and Author Index Keyword Index Published only online Book Reviews Manipulation at Home; Exercises Based on Osteopathic Structural Examination C. Fawkes Clinical Reasoning for Manual Therapists I. Swinkels-Meewisse Energy Medicine in Therapeutics and Human Performance E. Ernst The list of referees for Volume 10 (2005) will appear in Vol. 11, No. 1

239

242

256

270

281

292 297 299 300 303

e1 e2 e3

Author index A Aina, A., 159 Allen, T.J., 144 Anderegg, Q., 256 Aprill, C.N., 207 B Bossuyt, P.M.M., 256 Boyles, R.E., 80, 164 Briffa, N.K., 191 Bryde, D., 191 Bullock, M.P., 28 C Cairns, M.C., 239 Cheek, E., 198 Childs, J.D., 224 Childs, M.J.D., 127 Cleland, J.A., 127 Conroy, J.L., 292 Coppieters, M.W., 52 D David Baxter, G., 61 Dempster, M., 61 E Edmondston, S.J., 44 F Flynn, T.W., 80, 164 Foster, N.E., 28 Freure, B.J., 191 Fung, J., 14 G Grant, A., 1 Greening, J., 136

J

R

Jones, L., 191 Jull, G., 93

R— heim, M., 38 Rankin, G., 108, 116 Reid, S.A., 4 Richards, N.L., 52 Ridehalgh, C., 136 Rivett, D.A., 4 Rix, G., 154

K Kerry, R., 73 Khan, K.M., 21 Kurz, K., 52 Kvale, A., 38

S

L Laslett, M., 207 Lin, K.-H., 68 Ljunggren, A.E., 38 Loh, W., 44 Lucas, C., 256

M Maillardet, F.J., 198 Maitland, M.E., 219 May, S., 159 McCarthy, C.J., 239 McDonald, B., 207 McDonough, S.M., 61 McLaughlin, L.M., 52 McRae, M., 127 Moore, A., 93 Moore, A.P., 61, 198 Mortensen, T.E., 52

T Taylor, A.J., 73 Teng, C.-C., 68 Thiel, H., 154 Troke, M., 198 Tsao, H., 78 U Urquhart, D.M., 144 V

N Newham, D.J., 108, 116

van Trijffel, E., 256 W

O 0stvold, E., 44 O’Sullivan, P.B., 242, 270, 281

H Henne, S.-E., 44 Hodges, P.W., 52, 144 Hoskins, W., 96, 180 Hu, W.Y., 219 Hurley, D.A., 61

Schneiders, A.G., 292 Shacklock, M., 175 Skaret, I.A, 52 Sran, M.M., 21 Stokes, M., 108, 116 Story, I.H., 144 Stowell, T., 127 Straker, L.M., 270, 281 Strand, L.I., 38 Stratford, P.W., 219 Szeto, G.P.Y., 270, 281

P Palmer, J.A., 127 Petty, N.J., 136 Pollard, H., 96, 180 Preuss, R., 14

302

Walker, V., 224 Wang, S.-F., 68 Werstine, M., 191 Westaway, M.D., 219 Whitman, J.M., 80, 164, 224 Wright, C.C., 28 Y Young, S.B., 207

Keyword index A Abdominal muscles 144 Acute low back pain 38 Aging 68 Atlantoaxial 219 Atlantodental interval 219 C Cervical pain 127 Cervical range of motion 68 Clinical guidelines 38 Computer use 270 D Database 198 Degenerative joint disease 68 Diagnosis 96, 207 E Electromyography 270 Exercises 144 H Hamstring 96, 180 I Instability 219 Intra-and inter-tester 191 J Joint instability 14 K Kinematics 280 L Low back pain 14, 144, 207 Lumbar 198 M Manual therapists 38 Manual therapy 127 Mechanical neck pain 127 Motion assessment 256 Motor control 270, 281 Muscle size 108 Muscle strain 96, 180

N Neck 219 Neck muscles 108 Normative 198 O Office ergonomics 281 P Pain 28 Palpation 191 Physical examination 207 Posture 28 R Radiographs 219 Range of motion 28 Reliability 191, 219, 256 Reproducibility of results 256 Rheumatoid arthritis 219 S Sacroiliac joint 207 Sensitivity 207 Shoulder 191 Shoulder impingement 28 Soft tissue injuries 14 Specificity 207 Spine 198, 256 Sports injury 96, 180 Systematic review 256 T Thoracic spine manipulation 127 Transversus abdominis 144 Treatment 180 U Ultrasonography 108 Ultrasound-based coordinate measuring system 68 V Validity 207 W Work-related neck and upper limb disorders 270, 281 303

ARTICLE IN PRESS

Manual Therapy 10 (2005) e1 www.elsevier.com/locate/math

Book review R. David, Essig-Beatty, Manipulation at Home; Exercises Based on Osteopathic Structural Examination, WVSOM Bookstore, ISBN 0-9766441-1-8, 2004, http://www.wvsom. edu/clinicalsciences/opp/Faculty/ManipulationAtHome.htm (299 pp., $36). Manipulation at Home is a comprehensive text addressing the use of exercise at home to augment the impact of manipulative treatment on a patient. The author has aimed to bring together his diagnostic skills and exercises to accomplish this effect. The text is written for both practitioners and patients. It is divided into seven sections; the initial chapter describes how this book is different to other exercise books in that it aims to combine positions of ease for relieving pain, myofascial stretches for muscle tension and mobilizations for restricted movement. The author goes on to describe his understanding and rationale for the use of these techniques citing a comprehensive selection of osteopathic and chiropractic research studies originating from the USA. A further five chapters cover problems in the lower back, head and neck, thoracic region, shoulder joint and upper extremity, hip joint and lower extremity, respectively. A variety of clear photographs showing positions of ease, stretching and joint mobilizations are

doi:10.1016/j.math.2005.06.012

interspersed throughout the text. In each chapter clear advice is also included about indications and contraindications to the use of these techniques for the guidance of both practitioner and patient. The final chapter focuses on the limitations of this and any other exercise book in terms of compliance with an exercise programme. The greatest part of the book is composed of appendices; these show clear pictures of all the recommended exercises. The final appendix shows the sets of exercises required to reinforce the effect of manipulative treatment. The author explains he has included all the exercises and programmes in the appendices to allow easy reproduction for suitable patients. This is a useful, very readable text for both patients with little knowledge of exercise and practitioners looking for alternative exercises to those they are already familiar with. Carol Fawkes National Council for Osteopathic Research, Clinical Research Centre for Health Professions, University of Brighton, Aldro Bulilding, 49, Darley Road, Eastbourne, East Sussex BN20 7UR, UK E-mail address: [email protected]

ARTICLE IN PRESS

Manual Therapy 10 (2005) e2 www.elsevier.com/locate/math

Book Review M.A. Jones, D.A. Rivett, Clinical Reasoning for Manual Therapists, first ed., Butterworth, Heinemann, ISBN 0750639067, 2004 (£34,99, 460pp.). This book is a comprehensive and practical guide for manual therapists and students who wish to reflect on and improve their clinical reasoning skills. The editors believe that clinical reasoning can be developed, partly, by skilled reflective reasoning. They have brought together authors with knowledge regarding educational theory and clinical experts in the field of manual therapy, which makes it a worthwhile read for students, teachers and clinicians who want to improve and test their clinical reasoning skills. In their definition and reflection on clinical reasoning the latest ideas on patient investigation and management are integrated. The book contains three sections with 1, 23, and 2 chapters, respectively. In the first section, the authors describe how they define clinical reasoning, with all its related issues. In their definition and reflection on clinical reasoning the latest ideas on patient investigation and management are integrated. It outlines and explains a model of clinical reasoning and describes the skills, knowledge and attributes manual therapists or other clinicians need to fully understand their patients’ problems and generating a patient centered health profile from a bio-psychosocial perspective. In the second section, Chapters 2–24, 23 case reports are presented written by a selection of expert manual therapists from all over the world. The editors have not critically appraised the case contributors but tried to elucidate the case contributors way of clinical thinking. This had been done in a structured manner by way of an interview approach after each stage of the investigation of a patient. The editors also discuss the clinical reasoning of the case contributors in this way making

doi:10.1016/j.math.2005.06.010

the experts answers more transparent, placing it in a broader context. The third section contains two chapters. The first concerns educational theories and related issues that are useful in learning and developing clinical reasoning. The last chapter describes pattern recognition related errors and common errors of clinical reasoning and how to improve clinical reasoning capacity. The editors stress the importance of continuously improving the reasoning process. In their view, an expert clinician needs advanced technical skills, but without good clinical reasoning these skills will have no meaning in clinical context. Reading the case reports, the reader is stimulated to answer the questions put forward by the editors in this way testing his/her reasoning thoughts. This book may be helpful in improving knowledge of ones own clinical reasoning skills. The case reports cover a broad range of disorders and show a perfect clinical reasoning and patient management. However, showing possible flaws in clinical reasoning may be of equal value in a learning process, outlining the importance of misinterpretations or missed cues in patient management. This book contributes to the evolution of manual therapy as an advanced discipline with professional education. It may be an individual guide, but might as well be used with colleagues, enhancing discussion and clinical reasoning.

Ilse Swinkels-Meewisse Faculty of Medicine and Pharmacology, Free University Brussels, Belgium Department of Medical, Clinical and Experimental Psychology, University of Maastricht, The Netherlands E-mail address: [email protected]

ARTICLE IN PRESS

Manual Therapy 10 (2005) e3 www.elsevier.com/locate/math

Book Review J. Oschman, Energy Medicine in Therapeutics and Human Performance, Butterworth-Heinemann, Stoneham, MA, ISBN 0750654007, 2003 359pp., £ 19.99. Most books on energy medicine are somewhat vague and, in my view, have nothing to do with energy and little with medicine. James Oschman’s new book is different. Oschman is well-versed in scientific language. The main thesis proposed by the author is as follows. The collagen structures of our connective tissue provide a ‘living matrix’. They serve as a network that conducts energy and information linking virtually every cell with any other. The living matrix is the missing link within the body’s information systems like blood and lymph vessels or nerve fibres. Because it connects anything with everything, it effortlessly explains how alternative therapies work. Oschman frequently uses acupuncture as an example but concedes that many if not all therapies that so far have been inexplicable become plausible through his theory. It was a relief that he calls his views ‘speculations’ because that is precisely what they are. Much of his reference to highly specialized basic research will go over

doi:10.1016/j.math.2005.06.005

the head of the moderately educated clinician. Reading this book, I first harboured a suspicion and later became convinced that Oschman does not really want us to understand the exact details of his text. To use the old analogy, he seems to use science as a drunken man uses a lamp post—not for enlightenment but for support. Readers of Manual Therapy will either love or hate it. Essentially, Oschman’s theory is a web woven out of scientific facts, wild speculation and wishful thinking. The resulting fabric fails to resist scrutiny. Yet many will love this book because it seemingly renders everything plausible that previously was not. However, lack of plausibility or theoretical underpinning is not the true problem of acupuncture or all the other alternative therapies. The real problem is to show with rigorous methodology that they work beyond a placebo response—and that, of course, is quite independent of biological plausibility. Edzard Ernst Complementary medicine, Peninsula Medicine School, University of Exeter & Plymouth, Exeter EX2 4NT, UK E-mail address: [email protected]