Manual Therapy Journal - Volume 14, Issue 3, Pages 241-352 (June 2009)

VOLUME 14 NUMBER 3 PAGES 241–352 June 2009

Editors

International Advisory Board

Ann Moore PhD, GradDipPhys, FCSP, CertEd, FMACP Clinical Research Centre for Health Professions University of Brighton Aldro Building, 49 Darley Road Eastbourne BN20 7UR, UK Gwendolen Jull PhD, MPhty, Grad Dip ManTher, FACP Department of Physiotherapy University of Queensland Brisbane QLD 4072, Australia

K. Bennell (Melbourne, Australia) K. Burton (Huddersfield, UK) B. Carstensen (Frederiksberg, Denmark) J. Cleland (Concord, NH, USA) M. Coppieters (Brisbane, Australia) E. Cruz (Setubal, Portugal) L. Danneels (Maríakerke, Belgium) I. Diener (Stellenbosch, South Africa) S. Durrell (London, UK) S. Edmondston (Perth, Australia) L. Exelby (Biggleswade, UK) J. Greening (London, UK) A. Gross (Hamilton, Canada) T. Hall (Perth, Australia) W. Hing (Auckland, New Zealand) M. Jones (Adelaide, Australia) B.W. Koes (Amsterdam, The Netherlands) J. Langendoen (Kempten, Germany) D. Lawrence (Davenport, IA, USA) D. Lee (Delta, Canada) R. Lee (London, UK) C. Liebenson (Los Angeles, CA, USA) L. Maffey-Ward (Calgary, Canada) E. Maheu (Quebec, Canada) C. McCarthy (Coventry, UK) J. McConnell (Northbridge, Australia) S. Mercer (Brisbane, Australia) P. Michaelson (Luleå, Sweden) D. Newham (London, UK) J. Ng (Hung Hom, Hong Kong) S. O’Leary (Brisbane, Australia) N. Osbourne (Bournemouth, UK) M. Paatelma (Jyvaskyla, Finland) N. Petty (Eastbourne, UK) A. Pool-Goudzwaard (The Netherlands) M. Pope (Aberdeen, UK) G. Rankin (London, UK) E. Rasmussen Barr (Stockholm, Sweden) D. Reid (Auckland, New Zealand) A. Rushton (Birmingham, UK) M. Shacklock (Adelaide, Australia) D. Shirley (Sydney, Australia) C. Snijders (Rotterdam, The Netherlands) P. Spencer (Barnstaple, UK) M. Sterling (Brisbane, Australia) M. Stokes (Southampton, UK) P. Tehan (Melbourne, Australia) M. Testa (Alassio, Italy) P. van der Wurff (Doorn, The Netherlands) P. van Roy (Brussels, Belgium) O. Vasseljen (Trondheim, Norway) B.Vicenzino (Brisbane, Australia) M. Wessely (Paris, France) A. Wright (Perth, Australia) M. Zusman (Perth, Australia)

Associate Editor’s Darren A. Rivett PhD, MAppSc, (ManipPhty) GradDipManTher, BAppSc (Phty) Discipline of Physiotherapy Faculty of Health The University of Newcastle Callaghan, NSW 2308, Australia E-mail: [email protected] Deborah Falla PhD, BPhty(Hons) Department of Health Science and Technology Aalborg University, Fredrik BajersVej 7, D-3, DK-9220 Aalborg Denmark Email:[email protected] Tim McClune D.O. Spinal Research Unit. University of Huddersfield 30 Queen Street Huddersfield HD12SP, UK E-mail: [email protected]

Editorial Committee Timothy W Flynn PhD, PT, OCS, FAAOMPT RHSHP-Department of Physical Therapy Regis University Denver, CO 80221-1099 USA Email: [email protected] Masterclass Editor Karen Beeton PhD, MPhty, BSc(Hons), MCSP MACP ex officio member Associate Head of School (Professional Development) School of Health and Emergency Professions University of Hertfordshire College Lane Hatfield AL10 9AB, UK E-mail: [email protected] Case Reports & Professional Issues Editor Jeffrey D. Boyling MSc, BPhty, GradDipAdvManTher, MCSP, MErgS Jeffrey Boyling Associates Broadway Chambers Hammersmith Broadway London W6 7AF, UK E-mail:[email protected] Book Review Editor Raymond Swinkels PhD, PT, MT Ulenpas 80 5655 JD Eindoven The Netherlands E-mail: [email protected]

Visit the journal website at http://www.elsevier.com/math doi:10.1016/S1356-689X(09)00018-6

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 241e242 www.elsevier.com/math

Editorial

Standardized clinical data collection and agreed outcome measurement Clinicians increasingly need comprehensive data concerning their performance in healthcare delivery in order to enhance their skills. This is consistent with the professional need and desire for mechanisms to support reflective practice and personal professional growth as well as to, demonstrate evidence of continuing professional development, Alongside this issue is the pressure from national health organizations, managers of health service departments and practices, together with clinical interest groups, to provide data on which to underpin financial commitment to health service provision. The concept of standardized baseline clinical data collection is growing in popularity with many individual services adopting varied approaches to data collection. Recently in the United Kingdom the Department of Health has developed a minimum data set for voluntary use in National Health Service hospital trust departments. The data set consists of comprehensive appointment and referral data focused predominantly on waiting times. This initiative was designed to help attain a reduction of waiting times for patients reaching definitive treatment to 18 weeks or less across a range of services and combinations of services. This initiative is of significant relevance to musculoskeletal outpatient physiotherapy services as physiotherapy treatment can often occur during the pathway to definitive surgery or indeed can represent ‘‘the’’ definitive treatment. It has been mooted that the use of the minimum data set may become mandatory by the year 2010. This initiative has been accompanied by the development of a health service-wide outcome metrics system which is currently being trialed in a small number of hospitals within the UK. This is known as the Care Records Service which is designed to standardize the outcome measures used across healthcare provision within all clinical settings and between disciplines involved in healthcare delivery. This has met with some resistance and considerable debate as in many cases there is a feeling that the outcomes currently included in the metrics system do not necessarily serve the interests of a wide range of 1356-689X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.03.002

health professionals, especially allied health professionals and indeed will not capture at all, patient focused outcomes for patients who have experienced healthcare delivery by health professionals other than those directly involved in surgical or medical interventions. The system is still under development. Effectively, this is a database of validated reliable outcome measures which have been tested for reliability and validity in some contexts, but not in many others! The difficulty of determining a small collection of valid and reliable outcome measures deemed appropriate for use for all patients, even for one type of physiotherapy specialty, has proved to be very difficult, particularly when trying to incorporate patients’ needs, desires and expectations. This exercise in the UK has again highlighted the need for rigorously developed outcome measures which are relevant to patients and therapists across specialties and which are simple to use and which are suitable for routine use within day-to-day practice. To this end it would be a great strength across all musculoskeletal physiotherapy services if an appropriate outcome measure could be agreed upon at national or even international levels and which would serve the needs of patients, therapists and health service managers. Standardized data collection has been on a number of agendas for some time and is gaining momentum to serve political needs, to underpin tendering for service provision, to highlight profile of patients accessing local and regional services and in helping to establish which kind of expertise is needed in the workforce to fulfill local healthcare provision needs. It is also essential to establish training needs within a local workforce. Rigorous standardized data collection can highlight the effectiveness of the service in fulfilling the desirable outcomes of musculoskeletal care in outpatient physiotherapy settings, such as, return to work and keeping individuals at work and the attainment of joint, ie patient and therapist joint treatment goals. Coupled with validated outcome measures, standardized data collection can prove a strong and viable mechanism for

242

Editorial / Manual Therapy 14 (2009) 241e242

ensuring maximum patient benefit of services delivered, underpinning service improvements, for maximizing skills available within the clinical workforce and indicating the cost efficiency of the cost in question. Importantly this method of data collection can also indicate how local health service delivery is fulfilling government and national health service agendas. The important point here is local ‘buy in’ and feelings of ownership. Those involved in standardized clinical data collection must all be aware of the possible benefits to them as individuals and to their patients and the service in which they operate. Whatever tool is used to collect data, it must be developed rigorously using consensus methods. It is vital to have strong data to support our service delivery and outcome measurement. The data must be understandable and acceptable to other professions as well as health service managers. It must increase musculoskeletal physiotherapists’ and other physiotherapists’ visibility in health service delivery, justify and

underpin current practices and indeed justify, if relevant, the expansion of musculoskeletal services. The important questions that all musculoskeletal physiotherapists and other practitioners should ask is: What data do we need to collect in order to justify our services and what would be the most appropriate outcome measure to use to indicate the effectiveness of our services? This activity can be a very useful in-service training activity whilst developing consensus and also helps individual staff to feel ownership of whatever system is established. The importance of this work cannot be underestimated. Equally, sharing of any data collected across areas of service delivery, and across hospital trusts and/or individual practices, can create healthy debate and stimulate intellectual and clinical growth. Ann Moore* Gwen Jull *Corresponding author. Tel./fax: þ44 1273 643766. E-mail address: [email protected]

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 243e251 www.elsevier.com/math

Original Article

Perceptions and use of passive intervertebral motion assessment of the spine: A survey among physiotherapists specializing in manual therapy Emiel van Trijffel a,*, Rob A.B. Oostendorp b, Robert Lindeboom a, Patrick M.M. Bossuyt a, Cees Lucas a a

Department of Clinical Epidemiology, Biostatistics and Bioinformatics, Academic Medical Centre, University of Amsterdam, Amsterdam, The Netherlands b Research Centre for Allied Health Sciences, Department of Quality of Care Research, University Medical Centre St Radboud, Nijmegen, The Netherlands

Received 11 September 2007; received in revised form 28 January 2008; accepted 7 February 2008

Abstract Manual therapists commonly use passive intervertebral motion (PIVM) assessment within physical examination. Data describing the use and interpretation of this manual diagnostic procedure, as well as therapists’ perception of related importance and confidence, are lacking. A survey was conducted among Dutch physiotherapists specializing in manual therapy (MT) using a 13-item, selfadministered, structured questionnaire. Three hundred and sixty-seven questionnaires were analysed. Response rate from the postal part of the survey was 56%. Dutch manual therapists most frequently apply passive segmental motion assessment to the cervical region and they prefer three-dimensionally coupled motions. They consider end-feel or, to a lesser extent, provocation of patient’s pain as decisive for diagnostic conclusions. Respondents believe that these spinal motion tests are important for treatment decisions and are confident in their conclusions drawn from it. These perceptions were largely stable across subgroups of therapists with different gender, age, experience, and educational background. Weekly amount of work related to spinal disorders was positively associated with perceived importance and confidence. Reported use and interpretation of PIVM assessment and related perceptions could only partly be substantiated by evidence. Results from this survey will help researchers design studies better reflecting daily practice in MT. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Manual therapy; Range of motion, articular; Physical examination; Spine

1. Introduction The Dutch Association for Manual Therapy describes manual therapy (MT) as a specialization within * Corresponding author. Department of Clinical Epidemiology, Biostatistics and Bioinformatics, Academic Medical Centre, University of Amsterdam, PO Box 22700, 1100 DE Amsterdam, The Netherlands. Tel.: þ31 10 4527409. E-mail address: e.vantrijff[email protected] (E. van Trijffel). 1356-689X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.02.005

physiotherapy characterized by the analysis, interpretation and treatment of complex health problems resulting from arthrogenic, muscular and neurogenic disorders of the spinal column and extremities using specific manual diagnostic and manual therapeutic techniques (Dutch Association for Manual Therapy, 2005). Contrary to many other countries, in The Netherlands, MT is considered a post-graduate (non-university) specialization within physiotherapy providing practitioners additional knowledge and skills for manual diagnosis and

244

E. van Trijffel et al. / Manual Therapy 14 (2009) 243e251

high-velocity thrust interventions (Oostendorp et al., 2004; Oostendorp et al., 2006). Dutch physiotherapists specializing in MT (manual therapists) have explicitly been profiling themselves as specialists in the care of health problems arising from spine-related disorders (Dutch Association for Manual Therapy, 2001). MT is characterized by the skill of therapists to induce articulatory movements manually in joints of spinal motion segments, like, for instance, passive physiological and accessory movements (Farrell and Jensen, 1992; Maher and Latimer, 1992; Dutch Association for Manual Therapy, 2005; Van Ravensberg et al., 2005). From a diagnostic perspective, judging the quantity and quality of passive segmental intervertebral joint motion contributes to classification of patients (Jull et al., 1994). Little is known about how manual therapists and physiotherapists use and interpret passive intervertebral motion (PIVM) assessment within clinical decisionmaking. Dutch manual therapists significantly more often detected impairments of joint mobility than Dutch physiotherapists did (Van Ravensberg et al., 2005; Oostendorp et al., 2006). Manual therapists participating in these studies believed ‘joint range-of-motion’ and ‘manual end-feel’ are relevant indicators of such impairments. A survey among orthopaedic certified specialists from the American Physical Therapy Association revealed that ‘segmental mobility testing or pain provocation’ was often used for the diagnosis of clinical lumbar instability (Cook et al., 2005). Australian physiotherapists rated the presence of an ‘excessively free end-feel’ on passive motion testing as highly important in the recognition of minor cervical instability (Niere and Torney, 2004). However, it remains unclear how manual therapists use, judge, and interpret PIVM assessment within their diagnostic reasoning leading to therapy decisions. In addition, it is unknown to what extent they believe this diagnostic procedure is important for decision-making or how confident they are in their conclusions drawn from it. A cross-sectional study using a self-administered survey questionnaire was conducted to describe and explore the use of PIVM assessment by Dutch physiotherapists specializing in MT and, additionally, to identify factors associated with therapists’ perception of related importance and confidence.

2. Methods 2.1. Survey instrument We developed a 13-item, structured questionnaire aimed at exploring the following three domains: demographic and professional characteristics, the use of PIVM assessment, and perceived importance and confidence related to PIVM assessment (Table 1).

Table 1 Survey instrument consisting of 13 items divided into three domains. Domain

Items

1. Demographic and professional characteristics

Gender Age Weekly amount of work related to spinal disorders MT educational background Experience in MT

2. Use of PIVM assessment in daily practice

Most frequently examined spinal region Most frequently applied type of movement Most decisive clinical finding Scale(s) used for categorizing clinical findings Term(s) used for recording of identified impairments of function of motion segments

3. Perceived importance and confidence related to PIVM assessment

Importance of PIVM assessment for therapy decisions Confidence in reaching correct diagnostic conclusions with PIVM assessment Confidence in reaching the same diagnostic conclusions with PIVM assessment compared to a random colleague

Note: MT: manual therapy, and PIVM: passive intervertebral motion.

In the second domain, two open-ended questions were used inviting respondents to describe types of scales used for classifying clinical findings and terms used for recording of identified impairments of function of motion segments in patient records. Furthermore, see Appendix for definitions of types of movements applied for PIVM assessment (Huijbregts, 2002; Cook, 2003; Brisme´e et al., 2006; Cook et al., 2006). In the third domain, respondents rated their perceived importance and confidence on a seven-point rating scale. 2.2. Procedure The questionnaire was tested for interpretability in two groups of manual therapists constituting consultation platforms. These platforms are part of the quality assurance program of the Royal Dutch Society for Physical Therapy and generally consist of up to 15 therapists working on quality improvement and assurance (Van der Wees et al., 2003). These testing rounds led to minor rephrasing of two items. Completing the questionnaire took 3e5 min. The final version of the questionnaire was sent by e-mail to all practices in The Netherlands listed under ‘manual therapists’ in the Yellow Pages and the Telephone Guide databases (2006) with a link to their e-mail address (September 2006). Potential respondents were requested to complete the questionnaire and return it by e-mail within 3 weeks. Consequently, from a single practice, more than one manual therapist could potentially respond. A reminder, accompanied by a new copy of the questionnaire, was sent after 1 month. Next, questionnaires were sent by post to all 23 MT consultation

245

E. van Trijffel et al. / Manual Therapy 14 (2009) 243e251

platforms in The Netherlands (November 2006). Members were asked to complete the questionnaires during their next meeting and return these using a prepaid and pre-addressed envelope. After 2 months, a reminder was sent in which the opportunity was given to request for new copies of the questionnaire. Finally, a random selection of 200 manual therapists out of 2796 (as at 1 January 2007) registered in the Quality Register of the Royal Dutch Society for Physical Therapy (2007) received a copy of the questionnaire by post (February 2007). Simultaneously, a random sample of 200 practices for MT listed in the Telephone Guide database (2007) also received one questionnaire each by post. Practices involved in the e-mail survey were excluded. Wherever possible, personal addressing was used. Respondents were asked to complete and return the questionnaire within 3 weeks using a prepaid and pre-addressed envelope. No reminder was sent. We incorporated methods that have been proven to increase response rates to postal questionnaires (Edwards et al., 2003). Potential respondents were informed by means of a cover letter explaining the purpose of the study. In case of multiple choice items, they were explicitly requested to select one answer only. It was also pointed out to them that data processing would be carried out anonymously. They were explicitly asked not to return questionnaires twice. 2.3. Statistical analysis Absolute and relative frequencies were used to describe categorical data. Ordinal data relating to the perceived importance and confidence items from the third domain of the questionnaire were additionally described with their medians and interquartile ranges (IQR). Normally distributed numerical data were summarized by their means and standard deviations. In case of non-normal distribution, median and range were presented. Answers to the two open-ended questions in the second domain were recorded and ranked according to reported frequency. Internal consistency reliability of the domain containing the importance and confidence items relating to conclusions drawn from PIVM assessment was calculated using Cronbach’s alpha. An alpha > 0.70 indicates homogeneity of the domain and consistency in scoring among respondents (Streiner and Norman, 2003). Rasch rating scale analysis was used to examine reliability of the rating scale structure of the importance and confidence items using an item response theory measurement model and OPLM, a computer software program for Rasch measurement models (Verhelst et al., 1995). Guided by this rating scale analysis, rating categories were dichotomized to obtain the best discrimination between respondents’ perceptions. Subsequently, univariate logistic regression was performed to identify demographic and professional

characteristics of respondents that were associated with perceived importance and confidence. Strengths of associations were expressed as odds ratios (OR) with their 95% confidence intervals and corresponding p-values. An OR of 1 indicates no association between the importance or confidence item and the demographic or professional characteristic, while an OR much greater or less than 1 indicates stronger associations. All analyses were carried out in SPSS (version 14.0). Missing data were not replaced. Multiple answers to multiple choice questions were handled as missing.

3. Results 3.1. Response rates From 858 citations found in the Yellow Pages and 1079 found in the Telephone Guide, 178 and 128 practices, respectively, had a link to their e-mail address. Twenty-eight practices responded within 3 weeks and returned 33 questionnaires. After the reminder, another 21 questionnaires (21 practices) were received bringing the e-mail response rate to 16% (49/306). Ten consultation platforms responded by returning 68 questionnaires. After the reminder, another 22 questionnaires were sent by two platforms yielding a response rate of 52% (12/23). Finally, 223 (56%) completed questionnaires were returned by post. In total, 367 questionnaires, containing 31 (0.7%) missing data, were analysed. 3.2. Descriptive findings Demographic and professional characteristics of the survey sample are summarized in Table 2.

Table 2 Demographic and professional characteristics of the survey sample (n ¼ 367). Male gender Mean age (SD) Weekly amount of work related to spinal disordersa MT educational background SOMT MT Utrecht (Van der Bijl) Maitland’s Concept Vrije University Brussels Master MT Orthopaedic MT Other More than one Experience in MTa

281 (76.6%), missing 0 46.1 (8.0) yrs, missing 0 24.0 (1.0e55.0) h, missing 6

241 39 19 3

(66.8%) (10.8%) (5.3%) (0.8%)

31 5 23 14.0

(8.5%) (1.4%) (6.4%), missing 6 (1.0e40.0) yrs, missing 7

Note: MT: manual therapy, SD: standard deviation, and SOMT: Stichting Opleiding Manuele Therapie. a Data described as median (minimumemaximum).

246

E. van Trijffel et al. / Manual Therapy 14 (2009) 243e251 Table 3 Ranked, absolute frequencies of 10 most frequently reported terms for recording of identified impairments of motion segments in patient records (n ¼ 367).

30%

Block Restriction Motion restriction Restricted Functional impairment Hypomobility Hypermobility Instability Hypofunction Dysfunction

20%

10%

n=110

n=102

n=62

n=13

C0-C3

C2-T1

C7-T4

T4-T10

0%

n=6 T10-L2

n=64 L1-S1

Fig. 1. Bar chart showing absolute (in bars) and relative ( y-axis) frequencies of most frequently examined spinal region using PIVM assessment (n ¼ 357).

Two hundred and seventy-four respondents (76.8%) reported applying PIVM assessment most frequently to the cervical region, i.e. motion segments C0eT4 (Fig. 1). When using segmental motion assessment, almost 80% (291/366) of manual therapists most frequently apply three-dimensional (coupled) physiological motions, while 23 (6.3%) indicated they use one-dimensional physiological movements primarily and about 11% (39/366) prefer accessory motion assessment. Forty-eight percent of respondents (176/367) considered perceived resistance at the end of the movement (end-feel) as the most decisive clinical finding from PIVM assessment for making diagnostic conclusions about impairments of joint function of motion segments, while 22.6% (83/367) preferred provocation or reduction of pain or other symptoms for this purpose. Forty-eight (13.1%) therapists reported to judge PIVM primarily on range of motion and 10.4% (38/367) relied on perceived resistance during movement. Ninety-two manual therapists (25.1%) stated they made explicit use of scales for categorizing clinical findings from PIVM assessment. For classifying end-feel (19 times), scales were used with terms like ‘hard’, ‘empty’, ‘springy’, and ‘stiff’ reported most often. Visual analogue scales (22 times) were the scale of choice for measuring patient’s pain. Nine therapists reported their use of Maitland’s movement diagram for grading mobility and a three-point scale (hypomobileenormalehypermobile) for this purpose was mentioned seven times. In total, 67 different terms were given for the recording of identified impairments of motion segments in

77 47 38 34 28 19 17 13 8 6

patient records. Table 3 shows 10 most frequently reported terms. Some respondents additionally record segmental level (35 times) or motion direction (36) of impairments, or both (24). In Table 4, frequencies of scores on the importance and confidence items are presented. Eighty-one percent (296/367) of respondents believed that diagnostic conclusions from PIVM assessment were reasonably or very important for deciding on MT as a treatment option (IQR ‘reasonably important’ to ‘very important’). With respect to perceived confidence in diagnostic conclusions drawn from PIVM assessment, 198 therapists (54.0%) were reasonably confident that they would reach a correct diagnosis about impairments of function of motion segments (IQR ‘somewhat confident’ to ‘reasonably confident’), while 251 (68.4%) were somewhat or reasonably confident that they would reach the same conclusions as a random colleague (IQR ‘neutral’ to ‘reasonably confident’). Cronbach’s alpha for the total domain was 0.75, indicating that, on the whole, respondents were consistent in their reporting of perceptions. 3.3. Inferential findings Rating scale analysis indicated that collapsing the ‘reasonably important’ and ‘very important’ and the ‘reasonably confident’ and ‘very confident’ categories versus the collapsed remaining five categories, offered the best differentiation between respondents’ scores on perceived importance and confidence regarding the use of PIVM assessment. ORs representing strengths of associations between the three recoded dichotomous variables on the one hand and demographic and professional characteristics on the other are shown in Table 5. Weekly amount of work related to spinal disorders was positively associated with all perceptions of importance and confidence. This means, for example, that for every additional weekly hour spent on treating patients with health problems arising from disorders of the vertebral column, there was a 3% higher chance (odds) to believe

Table 4 Frequencies of scores on perceived importance and confidence related to PIVM assessment (n ¼ 367). How important to you are diagnostic conclusions from PIVM assessment for deciding on manual therapy as a treatment option? Very unimportant 5 (1.4%)

Reasonably unimportant 3 (0.8%)

Somewhat unimportant 2 (0.5%)

Neutral 5 (1.4%)

Somewhat important 56 (15.3%)

Reasonably important 198a (53.9%)

Very important 98 (26.7%)

How confident are you by using PIVM assessment in reaching the correct diagnostic conclusions with regard to impairments of motion segments? Very unconfident 5 (1.4%)

Reasonably unconfident 10 (2.7%)

Somewhat unconfident 16 (4.4%)

Neutral 29 (7.9%)

Somewhat confident 99 (26.9%)

Reasonably confident 198a (54.0%)

Very confident 10 (2.7%)

Very unconfident 11 (3.0%) a

Reasonably unconfident 20 (5.4%)

Somewhat unconfident 31 (8.4%)

Neutral 45 (12.3%)

Somewhat confident 113a (30.8%)

Reasonably confident 138 (37.6%)

Very confident 9 (2.5%)

Median score.

Table 5 Univariate logistic regression analysis using scores on importance and confidence regarding PIVM assessment as dependent factors and demographic and professional characteristics as independent explanatory variables (n ¼ 367). Characteristic

Male gender Age Weekly amount of work related to spinal disorders MT educational backgrounda Maitland’s Concept Orthopaedic MT Experience in MT

Importance of PIVM assessment for therapy decisions

Confidence in reaching correct diagnostic conclusions with PIVM assessment

Confidence in reaching the same diagnostic conclusions with PIVM assessment compared to a colleague

OR

95% CI

p-Value

OR

95% CI

p-Value

OR

95% CI

p-Value

1.31 0.98 1.03

[0.69, 2.49] [0.95, 1.01] [1.01, 1.06]

0.412 0.205 0.033*

0.90 1.03 1.04

[0.55, 1.46] [0.99, 1.05] [1.02, 1.07]

0.665 0.067 <0.0001*

0.91 1.00 1.03

[0.56, 1.50] [0.97, 1.02] [1.01, 1.06]

0.716 0.863 0.007*

1.03 0.66 0.98

[0.29, 3.70] [0.27, 1.64] [0.95,1.01]

0.964 0.374 0.197

1.24 1.31 1.02

[0.47, 3.25] [0.60, 2.86] [0.99, 1.05]

0.667 0.495 0.180

3.03 2.15 1.00

[1.15, 7.99] [1.01, 4.57] [0.97, 1.03]

0.025* 0.047* 0.877

E. van Trijffel et al. / Manual Therapy 14 (2009) 243e251

How confident are you by using PIVM assessment in reaching the same diagnostic conclusions as a random colleague with the same educational background?

Note: CI: confidence interval, MT: manual therapy, OR: odds ratio, PIVM: passive intervertebral motion, and SOMT: Stichting Opleiding Manuele Therapie. *Significant at the 0.05 level. a Reference category: SOMT (results from MT Utrecht (Van der Bijl) and Vrije University Brussels Master MT not shown). 247

248

E. van Trijffel et al. / Manual Therapy 14 (2009) 243e251

diagnostic conclusions from PIVM assessment are ‘reasonably important’ or ‘very important’ for therapy decisions. Similarly, therapists trained according to the orthopaedic MT principles and Maitland’s Concept were more than twice and three times, respectively, more likely to be ‘reasonably confident’ or ‘very confident’ in reaching the same diagnostic conclusions as their colleagues compared to respondents educated by the Stichting Opleiding Manuele Therapie (SOMT).

4. Discussion This survey found Dutch physiotherapists specializing in MT most frequently apply PIVM assessment to the cervical region and prefer three-dimensionally coupled motions. They consider end-feel or, to a lesser extent, pain or other symptoms and range of motion as the decisive clinical finding for diagnostic conclusions concerning impairments of motion segments. Practitioners believe that this manual diagnostic procedure is important for deciding on MT as a treatment option and they are confident in their conclusions drawn from it. These reported perceptions were largely stable across subgroups of therapists with different gender, age, experience, and educational background. The majority of respondents reported applying PIVM testing most often to the cervical region with about 31% choosing the upper cervical spine in particular. A systematic review showed acceptable inter-examiner reliability of passive assessment of motion in segments C1eC2 and C2eC3 (Van Trijffel et al., 2005). This has been confirmed for assessment of rotation mobility of C1eC2 in later studies (Cleland et al., 2006; Ogince et al., 2007). Another study showed a high level of reliability for lateral gliding examination of C0eC1 (Piva et al., 2006). However, findings from PIVM assessment were not included in clinical prediction rules for guiding manipulative treatment of patients with neck pain (Tseng et al., 2006; Cleland et al., 2007). Segmental hypomobility of lumbar motion segments, on the other hand, has been recognized within a validated prediction rule as a predictor of a successful outcome after spinal manipulation in patients with low-back pain (Flynn et al., 2002; Childs et al., 2004). The lumbar spinal column was reported by only 20% of our sample as the region most frequently examined. Dutch manual therapists prefer to use three-dimensionally coupled movements for passive segmental motion assessment. Cramer et al. (2006) concluded that, although all spinal motions are indeed coupled motions, motion patterns are complex and coupling differs from one segment to the other. Coupling behaviour of the lumbar and thoracic spine has been shown to be inconsistent with respect to directions in which side-bending and axial rotations are associated (Cook, 2003; Legaspi

and Edmond, 2007; Sizer et al., 2007). With respect to the cervical spine, there is full agreement about coupling behaviour of motion segments C2eT1, but variation exists in patterns of C0eC1 and C1eC2 (Cook et al., 2006). Inter-examiner reliability of passive three-dimensional movement tests was poor for L4eL5, while in the mid-thoracic spine (T6eT7) fair to substantial agreement beyond chance was obtained (Brisme´e et al., 2005, 2006). Because of all these variations and the unpredictability of coupling biomechanics in pathological states, authors have cautioned to be reticent in the use of three-dimensional motion assessment in patients (Panjabi et al., 1994; Harrison et al., 1998). We found considerable variation among the sample with respect to which clinical finding from PIVM assessment would be decisive for diagnostic conclusions about impairments of function of motion segments. Jull et al. (1994) proposed to guide detection of dysfunctional spinal segments by assessing tissue stiffness via the presence of muscle reactivity or abnormal thicker through range resistance. In Maitland’s Concept, change in resistance perceived by the therapist during movement combined with pain reported by the patient is used to construct movement diagrams (Maitland et al., 2005). Only a small proportion of respondents chose resistance perceived during passive motion testing as an important diagnostic phenomenon. A recent survey among manual physical therapists in New Zealand and the USA revealed that passive accessory lumbar segmental motion testing is performed to assess pain response and quality of resistance, and physiological motion testing is used to assess quality of motion path (Abbott et al., 2007). No consensus exists on which clinical finding e or combination of clinical findings e is appropriate to identify impairments of function of motion segments nor on which scale to use for categorizing findings. Likewise, 67 different terms were identified for the recording of impairments of motion segments. It seems that manual therapists suffer from the same lack of uniformity in terminology as their colleagues in chiropractic do (Walker and Buchbinder, 1997). Dutch practitioners are confident to be correct in their conclusions drawn from PIVM assessment. Reaching correct diagnoses about impairments of function of motion segments reflects the validity e or diagnostic accuracy e of the test procedure. Evidence of accuracy of segmental motion testing is accumulating gradually but does not permit definitive conclusions (Najm et al., 2003; Humphreys et al., 2004; Abbott et al., 2005; Ferna´ndez-de-las-Pen˜as et al., 2005; Fritz et al., 2005b; Ogince et al., 2007). Respondents are somewhat less confident in reaching the same diagnostic conclusions from PIVM assessment as compared to a random colleague. This means that they are less confident in the inter-examiner reliability of passive segmental motion testing than in its diagnostic accuracy. Nevertheless,

E. van Trijffel et al. / Manual Therapy 14 (2009) 243e251

249

reported levels of confidence in reliability were not in accordance with available evidence. Inter-examiner reliability of segmental intervertebral motion tests has been found to be unacceptably low (Seffinger et al., 2004; Van Trijffel et al., 2005; May et al., 2006; Stochkendahl et al., 2006). Seffinger et al. (2004) concluded that assessing regional range of spinal motion was more reliable than segmental examination. Several authors have questioned the clinical usefulness and necessity of identifying impairments of joint mobility at specified spinal levels in order to make treatment decisions (Troyanovich and Harrison, 1998; Huijbregts, 2002; Haas et al., 2003; Fritz et al., 2005a).

5. Conclusions

4.1. Limitations of this study

Acknowledgements

Low response rates in survey research reduce sample size and precision as well as threatening validity as nonresponders may differ systematically from responders (Kessler et al., 1995). Among health professionals, response rates to mail surveys vary widely, from 16% to 91% (Lusk et al., 2007). The response rate from the postal part of our survey was comparable to rates among other care providers (Asch et al., 1997; Russell et al., 2004). Response to the e-mail component, on the contrary, was at the very low end of the range. We did not collect data on non-responders and data on the distribution of characteristics of manual therapists registered by the Dutch Association for Manual Therapy or the Royal Dutch Society for Physical Therapy were not available. It might be possible that non-responders would have scored systematically different with respect to their use and perceptions of PIVM assessment which could have biased our results. Given the large sample size achieved, we assume distribution of type of educational background in our survey sample to be correctly reflecting the total population of Dutch manual therapists. Furthermore, MT education in The Netherlands is strongly embedded within international concepts. In these traditional concepts, passive joint motion assessment holds a prominent place (Farrell and Jensen, 1992). Therefore, we suppose that results of this study will to a certain extent be generalizable to populations of manual therapists outside The Netherlands. Our opinion is partly supported by Abbott et al. (2007) showing that manual physical therapists from New Zealand and the USA believe passive accessory and physiological motion testing is accurate for estimating the quantity of movement present at a lumbar segment and segmental motion findings are important for treatment selection. Finally, respondents could potentially have returned more than one questionnaire each. Because priority was given to anonymous data processing, we were unable to control this possible threat to validity in our results.

The authors wish to thank all manual therapists for providing us the data by completing and returning questionnaires.

Dutch physiotherapists specializing in MT showed substantial consistency in reporting their use, interpretation, and related perceptions of importance and confidence regarding PIVM assessment. However, these findings could only partly be substantiated by evidence. The role and position of PIVM testing of the spine within the diagnostic pathway as a whole need further clarification to allow more useful evaluation of its diagnostic value (Bossuyt et al., 2006). We aim that the results of this survey will guide future research to better reflect daily practice in MT.

Appendix Definitions of types of movements applied for passive segmental intervertebral motion assessment of the spine One-dimensional physiological movements: Moving one vertebra on another in the sagittal (flexion/extension), frontal (side-bending), or transverse (rotation) anatomical plane. Three-dimensional physiological movements: Moving one vertebra on another in the sagittal, frontal and transverse anatomical planes simultaneously. An emphasis can be placed on any of these single components. Coupling of side-bending and rotation can be in the same direction or in opposite directions. Accessory movements: Moving one vertebra on another using translatory motions associated with physiological motions.

References Abbott JH, McCane B, Herbison P, Moginie G, Chapple C, Hogarty T. Lumbar segmental instability: a criterion-related validity study of manual therapy assessment. BMC Musculoskeletal Disorders 2005;6:56. Abbott JH, Flynn TW, Fritz JM, Hing WA, Reid D, Whitman JM. Manual physical assessment of spinal segmental motion: intent and validity. Manual Therapy 2007; doi:10.1016/j.math.2007. 09.011. Asch DA, Jedrziewski MK, Christakis NA. Response rates to mail surveys published in medical journals. Journal of Clinical Epidemiology 1997;50(10):1129e36. Bossuyt PM, Irwig L, Craig J, Glasziou P. Comparative accuracy: assessing new tests against existing diagnostic pathways. British Medical Journal 2006;332:1089e92. Brisme´e J-M, Atwood K, Fain M, Hodges J, Sperle A, Swaney M, et al. Interrater reliability of palpation of three-dimensional segmental motion of the lumbar spine. The Journal of Manual and Manipulative Therapy 2005;13(4):215e20. Brisme´e J-M, Gipson D, Ivie D, Lopez A, Moore M, Matthijs O, et al. Interrater reliability of a passive physiological intervertebral

250

E. van Trijffel et al. / Manual Therapy 14 (2009) 243e251

motion test in the mid-thoracic spine. Journal of Manipulative and Physiological Therapeutics 2006;29(5):368e73. Childs JD, Fritz JM, Flynn TW, Irrgang JJ, Johnson KK, Majkowski GR, et al. A clinical prediction rule to identify patients with low back pain most likely to benefit from spinal manipulation: a validation study. Annals of Internal Medicine 2004;141(12):920e8. Cleland JA, Childs JD, Fritz JM, Whitman JM. Interrater reliability of the history and physical examination in patients with mechanical neck pain. Archives of Physical Medicine and Rehabilitation 2006;87(10):1388e95. Cleland JA, Childs JD, Fritz JM, Whitman JM, Eberhart SL. Development of a clinical prediction rule for guiding treatment of a subgroup of patients with neck pain: use of thoracic spine manipulation, exercise, and patient education. Physical Therapy 2007;87(1):9e23. Cook C. Coupling behaviour of the lumbar spine: a literature review. The Journal of Manual and Manipulative Therapy 2003;11(3):137e45. Cook C, Brisme´e J-M, Sizer PS. Factors associated with physiotherapists’ confidence during assessment of clinical cervical and lumbar spine instability. Physiotherapy Research International 2005;10(2):59e71. Cook C, Hegedus E, Showalter C, Sizer PS. Coupling behaviour of the cervical spine: a systematic review of the literature. Journal of Manipulative and Physiological Therapeutics 2006;29(7):570e5. Cramer G, Budgell B, Henderson C, Khalsa P, Pickar J. Basic science research related to chiropractic spinal adjusting: the state of the art and recommendations revisited. Journal of Manipulative and Physiological Therapeutics 2006;29(9):726e61. Dutch Association for Manual Therapy. Vision document. Amersfoort, The Netherlands. Available from: http://www.nvmt.nl/ upload/Visiedocument.pdf; June 2001 [in Dutch] [accessed 9.03.07]. Dutch Association for Manual Therapy. Professional competence profile for the manual therapist. Amersfoort, The Netherlands. Available from: http://www.nvmt.nl/upload/BCP.ENG.pdf; May 2005 [accessed 2.02.07]. Edwards P, Roberts I, Clarke M, DiGuiseppi C, Pratap S, Wentz R, et al. Methods to increase response rates to postal questionnaires. The Cochrane Database of Methodology Reviews 2003;(4). doi:10.1002/14651858.MR000008.pub2. Art. No.: MR000008 [accessed 16.08.06]. Farrell JP, Jensen GM. Manual therapy: a critical assessment of role in the profession of physical therapy. Physical Therapy 1992;72(12):843e52. Ferna´ndez-de-las-Pen˜as C, Downey C, Miangolarra-Page JC. Validity of the lateral gliding test as tool for the diagnosis of intervertebral joint dysfunction in the lower cervical spine. Journal of Manipulative and Physiological Therapeutics 2005;28(8):610e6. Flynn T, Fritz J, Whitman J, Wainner R, Magel J, Rendeiro D, et al. A clinical prediction rule for classifying patients with low back pain who demonstrate short-term improvement with spinal manipulation. Spine 2002;27(24):2835e43. Fritz JM, Whitman JM, Childs JD. Lumbar spine segmental mobility assessment: an examination of validity for determining intervention strategies in patients with low back pain. Archives of Physical Medicine and Rehabilitation 2005a;86(9):1745e52. Fritz JM, Piva SR, Childs JD. Accuracy of the clinical examination to predict radiographic instability of the lumbar spine. European Spine Journal 2005b;14:743e50. 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):1091e6. Harrison D, Harrison D, Troyanovich S. Three-dimensional spinal coupling mechanics: part one. Journal of Manipulative and Physiological Therapeutics 1998;21(2):101e13. Huijbregts PA. Spinal motion palpation: a review of reliability studies. The Journal of Manual and Manipulative Therapy 2002;10(1):24e39.

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. Jull G, Treleaven J, Versace G. Manual examination: is pain provocation a major cue for spinal dysfunction? Australian Journal of Physiotherapy 1994;40(3):159e65. Kessler RC, Little RJA, Groves RM. Advances in strategies for minimizing and adjusting for survey nonresponse. Epidemiologic Reviews 1995;17(1):192e204. Legaspi O, Edmond SL. Does the evidence support the existence of lumbar spine coupled motion? A critical review of the literature. The Journal of Orthopaedic and Sports Physical Therapy 2007;37(4):169e78. Lusk C, Delclos GL, Burau K, Drawhorn DD, Aday LA. Mail versus Internet surveys: determinants of method of response preferences among health professionals. Evaluation and the Health Professions 2007;30(2):186e201. Maher C, Latimer J. Pain or resistance e the manual therapists’ dilemma. Australian Journal of Physiotherapy 1992;38(4):257e60. Maitland G, Hengeveld E, Banks K, English K, editors. Maitland’s vertebral manipulation. 7th ed. Oxford, UK: Elsevier/Butterworth and Heinemann; 2005. May S, Littlewood C, Bishop A. Reliability of procedures used in the physical examination of non-specific low back pain: a systematic review. Australian Journal of Physiotherapy 2006;52(2):91e102. Najm WI, Seffinger MA, Mishra SI, Dickerson VM, Adams A, Reinsch S, et al. Content validity of manual spinal palpatory exams. A systematic review. BMC Complementary and Alternative Medicine 2003;3:1. Niere KR, Torney SK. Clinicians’ perceptions of minor cervical instability. Manual Therapy 2004;9(3):144e50. Ogince M, Hall T, Robinson K, Blackmore AM. The diagnostic validity of the cervical flexion-rotation test in C1/2-related cervicogenic headache. Manual Therapy 2007;12(3):256e62. Oostendorp RAB, Scholten-Peeters GGM, Swinkels RAHM, Bekkering GE, Heijmans MWFGJ, Huijbregts PA, et al. Evidence-based practice in physical and manual therapy: development and content of Dutch national practice guidelines for patients with non-specific low back pain. The Journal of Manual and Manipulative Therapy 2004;12(1):21e31. Oostendorp RAB, Van Berkel LM, Van Ravensberg CDD, ScholtenPeeters GGM, Pool JJM, Swinkels RAHM, et al. Physical therapy and manual physical therapy for patients with non-specific lowback pain: differences in patient characteristics with implications for diagnostic classification. The Journal of Manual and Manipulative Therapy 2006;14(3):E46e55. Panjabi M, Oxland T, Yamamoto I, Crisco J. Mechanical behaviour of the lumbar and lumbosacral spine as shown by three-dimensional load-displacement curves. American Journal of Bone and Joint Surgery 1994;76(3):413e24. Piva SR, Erhard RE, Childs JD, Browder DA. Inter-tester reliability of passive intervertebral and active movements of the cervical spine. Manual Therapy 2006;11(4):321e30. Royal Dutch Society for Physical Therapy. Available from: http:// www.fysiotherapie.nl/index.html; 2007 [in Dutch] [accessed 14.02.07]. Russell ML, Verhoef MJ, Injeyan HS, McMorland DG. Response rates for surveys of chiropractors. Journal of Manipulative and Physiological Therapeutics 2004;27(1):43e8. Seffinger MA, Najm WI, Mishra SI, Adams A, Dickerson VM, Murphy LS, et al. Reliability of spinal palpation for diagnosis of back and neck pain. A systematic review of the literature. Spine 2004;29(19):E413e25. Sizer PS, Brisme´e J-M, Cook C. Coupling behaviour of the thoracic spine: a systematic review of the literature. Journal of Manipulative and Physiological Therapeutics 2007;30(5):390e9.

E. van Trijffel et al. / Manual Therapy 14 (2009) 243e251 Stochkendahl MJ, Christensen HW, Hartvigsen J, Vach W, Haas M, Hestbæk L. Manual examination of the spine: a systematic critical literature review of reproducibility. Journal of Manipulative and Physiological Therapeutics 2006;29(6):475e85. Streiner DL, Norman GR. Health measurement scales. A practical guide to their development and use. 3rd ed. Oxford, UK: Oxford University Press; 2003. Telephone guide. Available from: http://dtgi.detelefoongids.nl; 2006 [in Dutch] [accessed 1.09.06]. Telephone guide. Available from: http://dtgi.detelefoongids.nl; 2007 [in Dutch] [accessed 31.01.07]. Tseng Y-L, Wang WTJ, Chen W-Y, Hou T-J, Chen T-C, Lieu F-K. Predictors for the immediate responders to cervical manipulation in patients with neck pain. Manual Therapy 2006;11(4):306e15. Troyanovich SJ, Harrison DD. Motion palpation: it’s time to accept the evidence. Journal of Manipulative and Physiological Therapeutics 1998;21(8):568e71. Van der Wees PJ, Hendriks EJM, Veldhuizen RJ. Quality assurance in the Netherlands: from development to implementation and

251

evaluation. Dutch Journal of Physical Therapy 2003;113(S):3e6 (Special Issue International WCPT Congress Barcelona 2003). Van Ravensberg CDD, Oostendorp RAB, Van Berkel LM, ScholtenPeeters GGM, Pool JJM, Swinkels RAHM, et al. Physical therapy and manual physical therapy: differences in patient characteristics. The Journal of Manual and Manipulative Therapy 2005;13(2):113e24. Van Trijffel E, Anderegg Q, Bossuyt PMM, Lucas C. Inter-examiner reliability of passive assessment of intervertebral motion in the cervical and lumbar spine: a systematic review. Manual Therapy 2005;10(4):256e69. Verhelst ND, Glas CA, Verstralen HH. OPLM: computer program and manual. Arnhem, The Netherlands: CITO; 1995. Walker BF, Buchbinder R. Most commonly used methods of detecting spinal subluxation and the preferred term for its description: a survey of chiropractors in Victoria, Australia. Journal of Manipulative and Physiological Therapeutics 1997;20(9):583e9. Yellow pages. Available from: http://www.goudengids.nl/search/ Manueel_therapeuten.html; 2006 [In Dutch] [accessed 1.09.06.].

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 252e263 www.elsevier.com/math

Original Article

Disability in patients with chronic patellofemoral pain syndrome: A randomised controlled trial of VMO selective training versus general quadriceps strengthening G. Syme a,*, P. Rowe b, D. Martin c, G. Daly d a

Department of Orthopaedic Surgery, St. John’s Hospital in Howden, Livingston, United Kingdom b HealthQWest, Bioengineering Unit, University of Strathclyde, Glasgow, United Kingdom c Centre for Rehabilitation Science, University of Teesside, Middlesbrough, United Kingdom d Department of Physiotherapy Royal Infirmary of Edinburgh, Edinburgh, United Kingdom

Received 31 October 2007; received in revised form 6 February 2008; accepted 18 February 2008

Abstract This study was a prospective single blind randomised controlled trial to compare the effects of rehabilitation with emphasis on retraining the vastus medialis (VMO) component of the quadriceps femoris muscle and rehabilitation with emphasis on general strengthening of the quadriceps femoris muscles on pain, function and Quality of Life in patients with patellofemoral pain syndrome (PFPS). Patients with PFPS (n ¼ 69) were recruited from a hospital orthopaedic clinic and randomised into three groups: (1) physiotherapy with emphasis on selectively retraining the VMO (Selective); (2) physiotherapy with emphasis on general strengthening of the quadriceps femoris muscles (General ); and (3) a no-treatment control group (Control ). The three groups were then compared before and after an eight-week rehabilitation period. The Selective and General groups demonstrated statistically significant and ‘moderate’ to ‘large’ effect size reductions in pain when compared to the Control group. Both the Selective and General groups displayed statistically significant and ‘moderate’ and ‘large’ effect size improvements in subjective function and Quality of Life compared to the Control group. Knee flexion excursion during the stance phase of gait, demonstrated that there were no statistical significant differences and only ‘trivial’ to ‘small’ effect size differences between the Selective or General groups and the Control group. A large number of PFPS patients can experience significant improvements in pain, function and Quality of Life, at least in the short term, with quadriceps femoris rehabilitation, with or without emphasis on selective activation of the VMO component. Both approaches would seem acceptable for rehabilitating patients with PFPS. It may be appropriate to undertake exercises involving selective activation of the vastus medialis early in the rehabilitation process, however, clinicians should not overly focus on selective activation before progressing rehabilitation, especially in more chronic cases with significant participation restrictions. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Patellofemoral; Patellofemoral pain syndrome; Randomised controlled trial

1. Introduction Patellofemoral pain syndrome (PFPS) is reported to be a common, yet difficult to manage musculoskeletal * Corresponding author. Tel.: þ44 01506 522063; fax: þ44 01506 522064. E-mail address: [email protected] (G. Syme). 1356-689X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.02.007

condition (Ruffin and Kiningham, 1993; Biedert, 2004; Vega et al., 2006), although the incidence and prevalence of PFPS remains to be adequately evaluated (Callaghan and Selfe, 2007). There is debate as to whether rehabilitation should be based on exercises strengthening the quadriceps femoris muscle group or specifically targeting the vastus medialis oblique (VMO) or vastus lateralis (VL) muscles in isolation (Callaghan and Oldham,

G. Syme et al. / Manual Therapy 14 (2009) 252e263

1996). The first approach places emphasis on generally strengthening the quadriceps musculature (Wilk and Reinold, 2001; Witvrouw et al., 2004; Bolgla and Malone, 2005), one theory being that, if the force generated by the VMO is essential to proper patella tracking, then general quadriceps femoris strengthening, especially closed chain exercises (Boling et al., 2006), will bring the VMO up to a ‘threshold’ necessary for optimal tracking (Grabiner et al., 1994). The second approach places emphasis on the ‘selective activation’ and reeducation of the VMO component of the quadriceps femoris muscles (McConnell, 1986; Grelsamer and McConnell, 1998). The premise is that an imbalance in muscle activation timing between the VMO and VL muscles can lead to maltracking of the patella, abnormal loading of the patella mechanism and pain (Neptune et al., 2000; Cowan et al., 2001, 2002a, b). Whether the vastus medialis, or more specifically the VMO component, can be selectively activated is a contentious issue (Goh, 2000; Davies et al., 2001). Mean timing differences between the onset of VMO and VL activity have been reported for some patients with PFPS, with the delay in onset of the VMO relative to the VL reported to be between 15 and 19 ms compared to healthy controls (Cowan et al., 2001). Electromyographic biofeedback has been used to retrain VMO activation (Cowan et al., 2002b; Crossley et al., 2002). However, it is questionable whether such small timing differences can be retrained with this approach given the limited temporal resolution of most electromyographic biofeedback machines and the limitation of human capacity to discriminate between two such closely timed events (Karst and Willet, 1995). Furthermore, it is also debatable whether the VMO is a separate anatomical entity within the vastus medialis muscle, there being proponents for (Bose et al., 1980; Ono et al., 2005) and against this view (Hubbard et al., 1997; Nozic et al., 1997; Peeler et al., 2005). Such equivocal evidence raises doubts about the validity of the use of selective VMO activation in the clinical management of PFPS. Several RCTs investigating PFPS have examined the benefit of selective activation of the VMO by comparing conventional open and closed kinetic chain exercise programs with and without electromyographic biofeedback (Dursun et al., 2001; Yip and Ng, 2006; Qi and Ng, 2007; Ng et al., 2008). Dursun et al. (2001) reported no additional clinical improvement between the groups at three-months follow-up. Yip and Ng (2006) reported no statistical differences in clinical outcome between the groups at two months follow-up, but suggested that the addition of biofeedback may hasten recovery. Improvements in VMO/VL ratios in the biofeedback groups have been reported (Qi and Ng, 2007; Ng et al., 2008), but no significant differences were noted regarding pain (Qi and Ng, 2007). It is notable that these studies lacked a control, the randomisation process was not

253

described, and they did not use all the components of the McConnell-based rehabilitation program (Grelsamer and McConnell, 1998). The study presented here is a pragmatic trial, aiming to compare the general strengthening approach with a more selective VMO approach in the management of PFPS. A no-treatment control group and a blinded block randomisation procedure were used. In accordance with the World Health Organization (WHO, 2001) recommendations, information was collected on impairments of ‘body functions and structures’ and ‘activity and participation’. 2. Methods The study design was a RCT with three groups. The treatment groups consisted of Selective group (VMO selective activation treatment approach), General group (quadriceps femoris strengthening group) and Control group C (no-treatment control group). Following ethical approval (NHS Lothian Research Ethics Committee, Edinburgh) and signed informed consent, patients were recruited from an orthopaedic knee clinic (St. John’s Hospital, Livingston, Scotland) and referred to the researcher by the orthopaedic medical staff. Inclusion and exclusion criteria for the study were based on those outlined in previous studies (Table 1) (Kowall et al., 1996; Thomee´, 1997; Timm, 1998; Harrison et al., 1999; Cowan et al., 2001). To date, attempts to evaluate functional outcome and pathokinesiology in PFPS have used varied motion analysis approaches to assess the knee joint complex. Such studies have highlighted several ‘activity limitations’ in PFPS patients: reduced gait velocity (Powers et al., 1997, 1999) and reduced knee flexion angles at gait midstance (Dillon et al., 1983; Nadeau et al., 1997) and during stair ambulation (Crossley et al., 2005). Given these study findings, sample size was based upon the primary outcome measure of knee flexion excursion during gait (measured with flexible electrogoniometers). This selection can also be supported as walking is a common functional activity and it is arguable that a reduced ability and desire to articulate the knee joint would be associated with knee pain. The calculation used figures from a previous PFPS study (Nadeau et al., 1997) as no information existed for flexible electrogoniometry. Assuming a ‘normal’ mean knee flexion angle of 13.4 at 10% of stride length, with power ¼ 80% and alpha ¼ 0.05, the aim was to recruit a minimum of 28 patients for each group. Five physiotherapists, each with a minimum of five years clinical musculoskeletal experience, volunteered to provide physiotherapy management for the intervention groups. All of the physiotherapists received four one-hour training sessions regarding the physiotherapy management regimes to be used and could subsequently manage either Group A or Group B allocated patients.

254

G. Syme et al. / Manual Therapy 14 (2009) 252e263

Table 1 Study inclusion/exclusion criteria for patients with patellofemoral pain syndrome. Inclusion criteria: Males and females to be included. Age range 16e40 years. Unilateral or bilateral patellofemoral pain longer than three months. Willing to complete an eight-week rehabilitation program and attend the hospital clinic for assessments. Anterior or retropatellar pain reported on at least two of the following activities: prolonged sitting, ascending or descending stairs, squatting, running, kneeling, and hopping/jumping. In addition to the above, at least two of the following clinical examination findings: Patellar pain with manual compression of the patella against the femur. Patellar tenderness with palpation of the posteromedial and posterolateral borders of the patella. Patellar pain during resisted dynamic knee extension. Patellar pain with manual compression of the patella against the femur during isometric knee extension contraction. Exclusion criteria: Previous knee surgery or trauma. Ligamentous instability and/or internal derangement. (Subjects were referred for arthroscopy or Magnetic Resonance Imaging based on the criteria outlined by Acton and Craig (2000).) History of patella subluxation or dislocation or patella laxity. Traumatic lesions. Joint effusion when the midpatellar girth was 105% or more than the non involved knee, where applicable. True knee joint locking and/or giving way. Concurrent medical illness. Inflammatory joint pathology. Infection. Confirmed osteoarthritis of tibiofemoral and/or patellofemoral joints. Knee radiograph abnormalities. Circulatory or neurological abnormalities. Pre/infera or pes anserine bursitis, patella tendonitis, iliotibial tract tendonitis, Osgood Schlatter’s disease, Sinding-Larsen Johansson Syndrome, muscle tears or knee plica. Subjects unable or unwilling to give informed to written consent. Subjects awaiting surgery for another lower limb joint problem(s). History of low back, sacroiliac or ankle/feet problems longer than 3 days duration. Subjects already involved in active lower limb training programs. Malignancy. Pregnancy or breast feeding. Ongoing litigation related to lower limb injuries. No patient could be under the care of or have been under the care of another physiotherapist out with the study in the previous one-year prior to the study commencing.

Patients were assigned a physiotherapist using a rolling list system from a pre-designated list of physiotherapists. 2.1. Selective group: (vastus medialis selective activation approach) Subjects attended the Department of Physiotherapy at St. John’s Hospital, Livingston for assessment, education modification and progression of the treatment plan. Specific details of the Selective group treatment protocol are shown in Table 2.

to attend a supervised training session twice weekly. Specific details of the General group treatment protocol are shown in Table 3. It was deemed unethical to deny patients the use of knee taping/strapping given the weight of evidence supporting the premise that can reduce patellofemoral pain at least in the short term (Aminaka and Gribble, 2005). To avoid using taping that has been purported to specifically alter VMO activation (Cowan et al., 2002c) an alternative taping/strapping method was sought. Thus, the general quadriceps strengthening group employed a ‘knee sling’ strapping (Crozier, 1989).

2.2. General group: (quadriceps femoris strengthening approach)

2.3. Control group: (no treatment)

Training programs consisting of 2e3 training sessions per week have been indicated to improve strength for the novice trainer (Kraemer and Ratamess, 2004; Loudon et al., 2004). Thus, participants were required

Patients assigned to the control group were assessed then reassessed after a period of eight weeks and received no treatment during the eight-week study period. The control group was advised to refrain from

G. Syme et al. / Manual Therapy 14 (2009) 252e263

255

Table 2 Selective group treatment protocol. Physiotherapists were instructed that they could use all components of the ‘McConnell’ approach, such as VMO muscle re-education, flexibility stretching exercises, patella mobilisation, patella taping, electromyographic biofeedback and commercially available prefabricated orthotics (Grelsamer and McConnell, 1998). Exercises Lower limb exercises emphasising selective activation and retraining of the VMO muscle relative to the VL muscle was undertaken by using a dual channel surface electromyographic biofeedback unit (NeuroTracÔ, Verity Medical, Hampshire, England. Guidelines based on previous research suggested that a minimum of six sessions should be given during the eight-week treatment period (Crossley et al., 2002)). Correction of any dynamic lower limb malalignment (Powers, 2003) and gluteus medius muscle retraining was also encouraged (Grelsamer and McConnell, 1998). Participants were prescribed daily home exercises and provided with standardised home exercise information sheets (Physio Tools McConnell Exercises). Taping Patella taping for pain relief in Group A was as advocated by McConnell (Grelsamer and McConnell, 1998). Non-rigid hypoallergic tape (FixomullÒ Stretch Beiersdorf, Hamburg, Germany) was used to provide skin protection and rigid zinc oxide tape (LeukotapeÒ P Beiersdorf Hamburg, Germany) was used for taping corrections. The aim, if possible, was to achieve an immediate reduction in pain intensity of at least 50% (Grelsamer and McConnell, 1998). Participants were taught to independently apply the taping corrections and were instructed to reapply the tape daily and wear the tape during all waking hours until the pain subsided and exercises could be undertaken pain free. Stretching Soft tissue stretches were included for the quadriceps, hamstrings, iliotibial band, gastrocnemius/soleus and anterior hip structures (Grelsamer and McConnell, 1998). The aim was to maintain the stretches for 30 s and repeat each three times over (Shrier and Gossal, 2000). The patella was mobilised by the physiotherapist (Manske and Davies, 2003) and combined with deep frictional massage where necessary. The aim was three repetitions of 60 s each per treatment session (Crossley et al., 2002). Restrictions Physiotherapists were informed not to use isokinetic training, electrotherapy, acupuncture or place the subject on a regular gymnasium based training program for the lower limb(s). Advice All patients were supplied with an advice sheet about patellofemoral pain prior to the start of their treatment.

undertaking any new forms of exercise programs until reassessed eight weeks following initial assessment. Group C controls were offered treatment at the end of the trial. All participants were advised to continue with existing medication but not to alter or begin any new ‘knee related’ medications. Pre-existing foot orthoses were allowed. No patient could be under the care of or have been under the care of another physiotherapist out with the study in the previous one-year prior to the study commencing. One independent physiotherapist carried out the individual randomisation procedure out of view of the researcher. Block randomisation (in blocks of three) was employed through assigning numbers to permutations of the ABC sequence with each letter representing an arm of the study (Altman and Bland, 1999). These sequence blocks were placed in opaque sealed envelopes and randomly shuffled. For each three-sequence block a new envelope was opened and the patient assigned to the appropriate group based on the sequence order from left to right. This was done without the researcher’s knowledge. The researcher (blind assessor) was not involved in the management of the study participants until their involvement in the study had been completed. Blinding of participants was not possible owing to the nature of the study.

Participants were instructed not to reveal information about the nature of their treatment to the researcher. 3. Outcome measures 3.1. Personal information All outcome measures were completed before and after the eight-week trial period. The following information was recorded: age, height, weight, body mass index (BMI), lower limb dominance, site and duration of pain and previous knee treatment. 3.2. Primary outcome measure Two flexible electrogoniometers were used to measure knee flexioneextension angles with respect to time. The method advocated has been shown to be valid and reliable in healthy subjects (Rowe et al., 2001). Heel and toe foot-switches for both feet were used to aid identification of the stance and swing phases of gait. Data were recorded at 200 Hz. The subjects were asked to undertake two functional activities: (1) level walking between two markers 5 m apart and (2) perform an eccentric step down weightbearing on the painful lower limb from a standard

256

G. Syme et al. / Manual Therapy 14 (2009) 252e263

Table 3 General group treatment protocol. Lower limb exercises were based on widely accepted concentric, eccentric and proprioceptive rehabilitation principles (Thomee´, 1997; Lephart and Fu, 2000; Ellenbecker and Davies, 2001). Exercises The strengthening protocol aimed for 3e5 exercises consisting of 1e3 exercise sets of 10 repetitions, at 60e70% of the one repetition maximum intensity (DeLorme, 1945; Grimby and Thomee´, 2003; Kraemer and Ratamess, 2004). This strengthening protocol was suggested as a guideline and physiotherapists were advised to instruct patients to stop any prescribed exercise if pain intensity exceeded 5 on a 0e10 verbal rating scale, (0 ¼ ‘no pain’ and 10 ¼ ‘worst possible pain’) (Thomee´, 1997) and to adjust training loads accordingly. Correction of any dynamic lower limb malalignment (Powers, 2003) was encouraged during the exercises. Taping A ‘knee sling’ U shaped strapping was applied if necessary, which uses 2.5 cm wide zinc oxide ElastoplastÒ (BSN Medical Ltd., England) to support the knee and patellofemoral joints. The strapping was only applied during the initial pain control stages. Stretching Soft tissue stretches were included for the quadriceps, hamstrings, iliotibial band, gastrocnemius/soleus and anterior hip structures (Grelsamer and McConnell, 1998). The aim was to maintain the stretches for 30 s and repeat each three times over (Shrier and Gossal, 2000). The patella was mobilised by the physiotherapist (Manske and Davies, 2003) and combined with deep frictional massage where necessary. The aim was three repetitions of 60 s each per treatment session (Crossley et al., 2002). Restrictions Physiotherapists were informed not to use isokinetic training, electrotherapy, acupuncture, electromyographic biofeedback training or to specifically try and rehabilitate the VMO muscle during the treatment of this group. Advice All patients were supplied with an advice sheet about patellofemoral pain prior to the start of their treatment.

gymnasium wooden bench (height 31 cm) as far as comfortable without putting full weight onto the opposite lower limb. Subjects performed a prior practice to familiarise themselves with the equipment and functional task. Recordings were taken in quiet standing prior to the functional activities and used as baseline values ‘0 ’ indicating the neutral position of the knee. All subsequent recordings were measured relative to this baseline 0 . During gait, where a number of cycles were available, a cycle was selected from the middle of the data stream to avoid cycles during initiation or termination of the activity. As recommended by Maupas et al. (1999) data analysis focused on knee excursions angles: (1) the knee flexion excursion during the stance phase of gait; and (2) the knee excursion during the step down task (excursion being the difference between maximum and minimum angles). Gait time was calculated as the time taken to walk 5 m. Subjects were instructed to execute all tasks at a self-selected speed. 3.3. Pain, function and Quality of Life questionnaires Participants completed three valid and reliable instruments namely the McGill Pain Questionnaire (MPQ) (Melzack, 1975), the Modified Functional Index Questionnaire (MFIQ) (Selfe et al., 2001a, b) and the Short Form-36 Health Evaluation questionnaire (SF-36) (Jenkinson et al., 1996; Ware et al., 2000). Patients also completed the Patient Generated Index (PGI) (Ruta et al., 1994). The PGI is a patient generated Quality of Life

questionnaire, which asks the same questions of all patients, but allows them to specify their own response (Ruta et al., 1994). The PGI has been reported to be reliable and valid (Ruta et al., 1994, 1999). The NRS-101 pain intensity scale has been shown to be a reliable and valid measure of pain intensity (Jensen et al., 1986). The scale consisted of a scale from 0 to 100 with end points labelled ‘no pain’ and ‘worst possible pain’ with the subject asked to rate their ‘average pain intensity in previous one month’ (PIM). The triple hop test of lower limb function was performed as described by Noyes et al. (1989). The method has been shown to be a reliable measure of lower limb functional performance (Risberg et al., 1995; Bolgla and Keskula, 1997). The subject undertook three consecutive forward hops in a straight line on the dominant limb. The best of three trials was used for analysis. The procedure was then repeated for the non-dominant limb. Personal data and dependent variables measured preintervention and post-intervention for the most painful limb were analysed with SPSS (version 12) for WindowsÒ (SPSS Inc., Chicago, IL). MedCalcÒ software Version 7.6.0.0 (MedCalc Software, Belgium) was used to calculate D’Agostino Omnibus tests of normality. A Confidence Intervals Analysis program version 2.0.0 (Altman et al., 2000) was used to calculate median differences and 95% confidence intervals. An Effect Size Calculator spreadsheet (MicrosoftÒ XP Excel, www Version, University of Durham) (Coe, 2000) was used to calculate standardised mean difference effect size indices. Post-RCT scores from each group were compared using a one-way ANOVA for normally distributed data

G. Syme et al. / Manual Therapy 14 (2009) 252e263

and KruskaleWallis for non-normally distributed data, with alpha ¼ 0.05. Post-hoc analyses were executed using either independent samples t-tests or Manne Whitney U as appropriate, with a limited Bonferroni adjusted p < 0.017 for each group within each variable measured. The standardised mean difference effect size (Cohen’s d ) (Coe, 2000) was chosen as the effect size index. Nonparametric ‘Cohen type’ effect sizes were calculated using a method suggested by Hopkins (2000). In this case the ‘non-normal’ distributed variable is rank transformed without regard to grouping, the mean and standard deviations of the rank transformed data for each group can then be calculated as previously outlined

for the parametric standardised mean difference Cohen d (Hopkins, 2000). Intervention trial data were analysed on an intentionto-treat basis (Hollis and Campbell, 1999). 4. Results One hundred and ten patients were considered eligible on initial screening in the orthopaedic outpatient clinic (Fig. 1). Of these, 22 patients did not take up the service and three did not consent to the trial. A further 16 did not meet the inclusion criteria. Hence, 69 patients were randomised into three groups (Selective n ¼ 23, General n ¼ 23 and Control n ¼ 23). Two from

PFPS patients considered eligible for trial at outpatient clinics (n = 110)

Patients ineligible as they did not meet inclusion/exclusion (n = 16) 5 evidence of meniscal pathology 2 patella tendinopathy 1 congenital foot deformity 1 plicae 2 involved in legal claims 2 significant back pain 1 previous dislocation 1 previous arthroscopy 1 pregnant

22 patients failed to contact department to register for study or arrange routine physiotherapy 3 patients unwilling to participate in study (n = 25)

PFPS patients consented and randomised (n = 69)

Group A Selective ‘vastus medialis oblique activation’ group (n = 23)

Withdrawals (n = 2) 2 withdrew owing to work commitments

Completed 8-week intervention and follow up assessment (n = 21)

257

Group B General ‘quadriceps femoris strengthening’ group (n = 23)

Group C Control ‘no treatment’ group (n = 23)

Withdrawals (n = 1) 1 withdrew no further contact with department

Withdrawals (n = 3) 1 withdrew no further contact with department 1 withdrew owing to pregnancy 1 withdrawn owing to undisclosed insurance claim

Completed 8-week intervention and follow up assessment (n = 22)

Completed 8-week ‘intervention’ and follow up assessment (n = 20)

Fig. 1. Flow diagram showing the progression of patients through the clinical study.

258

G. Syme et al. / Manual Therapy 14 (2009) 252e263

the Selective group, one from the General group and three from the Control group failed to complete the study for various reasons (Fig. 1). There were minimal differences between the groups prior to treatment delivery and none were statistically significant (Table 4). Results of the hypothesis tests between the groups post-treatment delivery are shown in Table 5 and the effect sizes, mean/median differences and 95% confidence intervals shown in Table 6. There were no statistically significant differences among the groups in knee flexion excursion during the stance phase of gait. The step down task demonstrated a statistically significant increase in knee excursion in the General group compared to the Control group and, although not statistically significant, there was also an improvement in the Selective group compared to the Control group. There was no statistically significant difference between the Selective and General groups. The results of the MPQ total scores supported the results of the NRS-101 ratings. Both the Selective and General groups revealed statistically significant reductions in comparison to the Control group. There was no statistical difference between the Selective and General groups. There were no statistically significant differences for the triple hop for distance test between the groups. Null hypothesis testing demonstrated that there were no statistically significant differences between the three groups post-intervention for the MFIQ scores. A review of the median differences and 95% confidence intervals, however, revealed that there was a trend for greater reductions in pain between the Selective compared with the

Control group (median difference 10.0 units, 95% CI 20.0 to 0.0) and between the General group and Control group (median difference 15.0 units; 95% CI 25.0 to 0.0), than between the Selective and General groups (median difference 0.0 units; 95% CI 5.0 to 15.0). There was a statistically significant improvement in SF-36 physical function in the Selective group compared to the Control groups and between the General and Control. There was no statistically significant difference between the Selective and General groups. There were no statistically significant differences between the three groups for SF-36 mental function. SF-36 questionnaire physical function revealed a large effect size improvement in the Selective group compared to the Control group (large ES 1.22; 95% CI 0.59 to 1.84) and a moderate effect size improvement in the General group compared to the Control group (moderate ES 1.04; 95% CI 0.22 to 1.65). There was only a ‘trivial’ difference between the Selective and General groups (trivial ES 0.00; 95% CI 0.58 to 0.58). Only ‘trivial’ to ‘small’ effect size differences were observed across the groups for SF-36 mental function. The PGI demonstrated statistically significantly greater scores in the Selective group compared to the Control group and in the General group compared to the Control group. There was no statistically significant difference between the Selective and General groups. 5. Discussion This study demonstrated that both general quadriceps strengthening and VMO specific training reduced pain and improved activity and participation, but that there was no difference between the approaches. Our results

Table 4 Pre-intervention data. Variable

A ¼ Selective (n ¼ 23)

B ¼ General (n ¼ 23)

C ¼ Control (n ¼ 23)

Age (years) Height (metres) Duration of pain (months) Activity (hours) Body weight (kg s) (back log transformed) Body mass index (kg s/m2) (back log transformed) Gender

28.8 (8.0)a 1.68 (0.77)a 49.0 (37.5)a 1.2 (2.3)a 71.8 (1.2)b

27.3 (7.9)a 1.69 (0.73)a 45.5 (35.3)a 1.4 (2.3)a 72.4 (1.3)b

28.5 (6.4)a 1.69 (0.87)a 50.5 (41.3)a 1.3 (2.3)a 74.5 (1.2)b

25.5 (1.3)b

25.5 (1.3)b

26.2 (1.2)b

13 Females 10 Males 20 Right sided 3 Left sided 5 Right side

13 Females 10 Males 21 Right sided 2 Left sided 7 Right side 3 Left side 4 Both sides, but right worse 7 Both sides, but left worse 2 Both sides equally painful n¼9

15 Females 8 Males 18 Right sided 5 Left sided 3 Right side 4 Left side 7 Both sides, but right worse 5 Both sides, but left worse 4 Both sides equally painful n¼6

Side dominance Painful side

Previous recipient of ‘physiotherapy’ for knee problem a b

2 Left side 6 Both sides, but right worse 10 Both sides, but left worse n¼5

Mean and standard deviation. Geometric mean and geometric standard deviation.

259

G. Syme et al. / Manual Therapy 14 (2009) 252e263 Table 5 Results of ‘body functions and structures’ and ‘activities and participation’ measures. Time point

A ¼ Selective

B ¼ General

C ¼ Control

ANOVA

Pre Post

13.9a (5.9) 14.7a (6.2)

17.8a (3.3) 18.3a (5.2)

14.9a (7.3) 15.2a (5.4)

0.070g

‘Body functions and structures’ measures NRS-101 average pain Pre Post intensity previous month

47.7a (29.6) 21.4a (24.7)

51.3a (29.4) 28.1a (28.5)

59.6a (21.8) 49.3a (22.5)

0.001g,**

A vs B 0.395h A vs C p < 0.001h,* B vs C 0.008h,*

McGill pain questionnaire total score

17.5a (6.1) 9.0c (9.0)

20.1a (8.8) 7.0c (12.0)

21.0a (9.4) 17.0c (14.0)

0.004i,**

A vs B 0.270j A vs C 0.014j,* B vs C 0.003j,*

59.8a (16.7) 62.3a (14.2)

63.0a (10.0) 66.0a (12.9)

55.3a (11.8) 56.1a (11.9)

0.040g,**

A vs B 0.362h A vs C 0.124h B vs C 0.012h,*

Pre Post Pre Post Pre Post Pre Post

4.9a (1.1) 4.5a (0.6) 370.7a (142.1) 374.3a (120.7) 33.0a (13.2) 25.0c (30.0) 44.96a (7.69) 53.11c (10.27)

4.2a (1.0) 4.3a (0.9) 373.0a (137.1) 390.4a (133.9) 34.8a (17.9) 10.0c (35.0) 46.67a (7.85) 54.26c (12.33)

4.3a (0.5) 4.5a (0.8) 348.8a (105.5) 349.9a (119.3) 33.0a (15.3) 30.0c (25.0) 46.45a (8.57) 40.38c (18.46)

Pre Post Pre Post

44.96a (7.69) 46.32c (10.38) 5.0a (1.8) 6.3a (2.8)

46.67a (7.85) 50.32c (7.51) 3.9a (2.4) 6.3a (2.5)

46.45a (8.57) 49.33c (11.79) 3.8a (1.7) 4.0a (1.4)

Variable ‘Primary outcome measure’ Gait minimum to peak stance knee excursion

Pre Post

‘Activities and participation’ measures Step down excursion range Pre Post of motion angle

Gait time (seconds) Triple hop for distance (cm) Modified functional index questionnaire SF-36 physical component summary score

SF-36 mental component summary score Patient generated index

Post-hoc tests (if applicable)

0.530g 0.484i 0.070i 0.001i,**

A vs B 0.991j A vs C p < 0.001j,* B vs C 0.001j,*

0.199i 0.001g,**

A vs B 0.933h A vs C 0.001h,* B vs C p < 0.001h,*

*p < 0.017; **p < 0.05. a Mean and standard deviation. c Median and inter quartile range. g One-way analysis of variance. h Independent samples t-test. i Kruskal-Wallis test. j ManneWhitney U test.

question whether the time and effort in specifically reeducating the VMO muscle is warranted. The conclusions supported previous work, which demonstrated minimal added value of selective VMO training in improving pain and function (Dursun et al., 2001; Yip and Ng, 2006; Qi and Ng, 2007). Contrary to previous findings, there were no statistically significant differences in the electrogoniometer gait parameters (Dillon et al., 1983; Nadeau et al., 1997). The present findings, although slightly underpowered, support suggestions that gait analysis may not be sensitive enough (Chesworth et al., 1989) to detect clinically useful changes in pain and function in patients with PFPS, and that motion analysis measures may not correlate with more prolonged or physically demanding walking tasks involved in participation in ‘normal’ life

situations (Mulder et al., 1998). Gait velocity (Powers et al., 1999), pain severity, methodological variations and sample size may explain differences between studies. There was some evidence of improvement in the step down task, in both the Selective and General groups, with a slightly greater improvement in the General group, but no additional benefit of incorporating selective VMO retraining. The results of the NRS-101 and MPQ total pain scores indicated that both the Selective and General groups demonstrated statistically significant reductions in pain in the short term supporting the work of previous investigators (Gaffney et al., 1992; Thomee´, 1997; Clark et al., 2000; Crossley et al., 2002; Loudon et al., 2004). Despite the chronic nature of PFPS exhibited by the patients involved in this study (mean duration

260

G. Syme et al. / Manual Therapy 14 (2009) 252e263

Table 6 Effect size results for ‘body functions and structures’ and ‘activities and participation’ measures. Variable

Effect sizes and 95% confidence intervals A vs C, Selective vs Control

‘Primary outcome measure’ Gait minimum 0.0 (0.65 to 0.50) to peak stance knee excursion ‘Body functions and structures’ measures NRS-101 average 1.17 (1.80 to 0.55) pain intensity previous month McGill pain 0.8 (1.42 to 0.21) questionnaire total score ‘Activities and participation’ measures Step down 0.48 (0.11 to 1.06) excursion range of motion angle Gait time (seconds) 0.00 (0.58 to 0.58)

Mean/median difference and 95% confidence intervals

B vs C, General vs Control

A vs B, Selective vs General

0.51 (0.08 to 1.10)

0.62 (1.21 to 0.03)

A vs B A vs C B vs C

3.6a (7.6 to 0.3) 0.6a (4.6 to 3.4) 3.1a (0.9 to 7.0)

0.89 (1.49 to 0.28)

0.23 (0.81 to 0.35)

1.03 (1.65 to 0.42)

0.31 (0.27 to 0.89)

A vs B A vs C B vs C A vs B A vs C B vs C

6.7a (24.7 to 11.2) 28.0a (45.9 to 10.0) 21.2a (39.1 to 3.3) 2.0b (2.0 to 7.0) 6.0b (11.0 to 1.0) 8.0b (14.0 to 4.0)

0.8 (0.21 to 1.41)

0.2 (0.85 to 0.31)

0.31 (0.89 to 0.27)

0.03 (0.55 to 0.61)

A vs B A vs C B vs C A vs B A vs C B vs C A vs B A vs C B vs C A vs B A vs C B vs C A vs B A vs C B vs C A vs B A vs C B vs C A vs B A vs C B vs C

3.7a (12.9 to 5.5) 6.2a (3.0 to 15.4) 9.9a (0.7 to 19.1) 0.2a (0.3 to 0.8) 0.0a (0.5 to 0.6) 0.2a (0.8 to 0.3) 16.1a (104.4 to 72.1) 24.4a (63.8 to 112.7) 40.6a (47.7 to 128.8) 0.0b (5.0 to 15.0) 10.0b (20.0 to 0.0) 15.0b (25.0 to 0.0) 0.08b (3.42 to 4.02) 10.75b (5.13 to 17.36) 10.57b (4.66 to 16.76) 4.36b (7.87 to 0.24) 2.64b (7.56 to 3.30) 1.17b (3.51 to 6.19) 0.01a (1.6 to 1.6) 2.3a (1.0 to 3.7) 2.3a (1.1 to 3.5)

Triple hop for distance (cm)

0.28 (0.30 to 0.86)

0.42 (0.16 to 1.00)

0.12 (0.70 to 0.45)

Modified functional index questionnaire

0.58 (1.17 to 0.01)

0.62 (1.22 to 0.03)

0.19 (0.39 to 0.77)

SF-36 physical component summary score SF-36 mental component summary score Patient generated index

1.22 (0.59 to 1.84)

1.04 (0.42 to 1.65)

0.00 (0.58 to 0.58)

0.32 (0.92 to 0.26)

0.22 (0.36 to 0.80)

0.56 (1.15 to 0.03)

1.14 (0.52 to 1.76)

1.28 (0.64 to 1.91)

0.00 (0.58 to 0.58)

Effect size qualifiers: trivial, 0 to 0.19; small, 0.20 to 0.59; moderate, 0.60 to 1.19; large, 1.11 to 1.19; very large, 1.20 to 3.9; extremely large, 4 to N. a Mean difference. b Median difference.

48.4 months) the results indicated that pain could be reduced in some cases, at least in the short term. The triple hop for distance test results were contrary to the findings of Yildiz et al. (2003) who reported statistically significant improvements in triple hop for distance measures in a study investigating isokinetic training in an athletic population with PFPS. However, results from studies involving athletes may not be directly generalisable to non-athletic populations, owing to potential, psychological and physical differences between athletic and non-athletic populations (Filho et al., 2005; Torstveit and Sundgot-Borgen, 2005). Selfe et al. (2001b) proposed that for the MFIQ results to be clinically significant the preepost differences had to be at least 10 units. Both the Selective and General groups had preepost median differences of 10 units and 15 units, respectively. Hence, both intervention groups demonstrated improved function clinically.

From the SF-36 scores participants experienced an improvement in perceived physical function, but not mental function. It is interesting to note that Ware and Kosinski (2001) reported that for SF-36 mental function a value of less than 42 units was predictive of a diagnosis of clinical depression, with a sensitivity and specificity of 73.7% and 80.6%, respectively. Using the 42-unit cut-off point, 22 of 69 (31.9%) of our participants fell within this category perhaps supporting the claims of coexisting psychological impairments in some patients with PFPS (Carlsson et al., 1993; Witonski et al., 1998; Jensen et al., 2005). The PGI measure afforded patients the opportunity to set their own goals in relation to ‘activities and participation’ in daily living. Post-intervention the PGI scores demonstrated a statistically significant improvement in both the Selective and General groups compared to the Control group again supporting the finding of improved perceived function.

G. Syme et al. / Manual Therapy 14 (2009) 252e263

As is a common criticism of many physiotherapy intervention studies (Feine and Lund, 1997) this study had a relatively short intervention period and follow-up. That said, the study did reflect current clinical practice, where patients receive an average of 6e8 treatments sessions. The pragmatic nature of the interventions in this study means that the relative importance of different elements of the rehabilitation program cannot be evaluated. However, physiotherapists often use a range of treatment approaches (Bithell, 2000; Grimmer et al., 2000); therefore, single research interventions are uncharacteristic of ‘routine’ clinical practice (Grimmer et al., 2004). Hence, the nature of the interventions used in this study enhances the external validity and generalisability of the results. Arguably, the inclusion of VMO retraining exercises in the initial stages of the rehabilitation process merely reflects the first stage of a program of ‘quadriceps femoris’ strengthening. Furthermore, Powers (2003) suggested that optimising lower limb alignment of the femur relative to the patella, by enhancing pelvic and femoral control, may be as pertinent as focusing on rehabilitating muscles that directly control the patella. If this is the case then the timing and content of the optimal treatment regime in patients with PFPS remains to be established. Difficulty in establishing the optimum rehabilitation regime, however, is compounded by issues related to the subclassification and diagnosis of ‘PFPS’ (Witvrouw et al., 2005; Na¨slund et al., 2006; Nijs et al., 2006). In conclusion, the study demonstrated that physiotherapy involving either selective VMO retraining exercises or a general quadriceps femoris strengthening program reduced pain, improved function and Quality of Life in PFPS patients. This study did not demonstrate that rehabilitation with selective VMO exercise significantly improves outcome above that provided by general open and closed chain strengthening exercises (Malone et al., 2002; Witvrouw et al., 2004; Bolgla and Malone, 2005). In light of these findings, clinicians should not overly focus on selective activation before progressing rehabilitation, especially in more chronic cases with significant participation restrictions.

Acknowledgements The authors would like to thank Mr. Robert Lee (Medical Statistics Department, Edinburgh University) for statistical advice, the staff (Departments of Physiotherapy and Orthopaedics, St. John’s Hospital, Scotland) for their invaluable support and Mr. R. Burnett FRCS(Ed.)Orth and Mr. G. Lawson FRCS(Ed.) Orth for their assistance regarding recruitment. Professor P. Rowe is part of HealthQWest a research consortium funded by the Scottish Funding Council,

261

NHS Scotland and the Chief Scientists Office of the Scottish Government.

References Acton D, Craig D. Should I scan or should I scope? Knee 2000;7:245e8. Altman DG, Bland JM. How to randomise. British Medical Journal 1999;319:703e4. Altman DG, Machin D, Bryant TN, Gardner MJ. Statistics with confidence. BMJ Books; 2000. Aminaka N, Gribble PA. A systematic review of the effects of therapeutic taping on patellofemoral pain syndrome. Journal of Athletic Training 2005;40:341e51. Biedert RM. Pathogenesis of patellofemoral pain. In: Biedert RM, editor. Patellofemoral disorders diagnosis and treatment. Chichester, England: John Wiley & Sons, Inc.; 2004. p. 55e69. Bithell C. Evidence-based physiotherapy; some thoughts on ‘best evidence’. Physiotherapy 2000;86:58e60. Bolgla LA, Keskula DR. Reliability of lower extremity functional performance tests. Journal of Orthopaedic and Sports Physical Therapy 1997;26:138e42. Bolgla L, Malone T. Exercise prescription and patellofemoral pain: evidence for rehabilitation. Journal of Sport Rehabilitation 2005;14: 72e88. Boling MC, Bolgla LA, Mattacola CG, Uhl TL, Hosey RG. Outcome of a weight-bearing rehabilitation program for patients diagnosed with patellofemoral pain syndrome. Archives of Physical Medicine and Rehabilitation 2006;87:1428e35. Bose K, Kangasuntheram R, Osman M. Vastus medialis oblique: anatomical and physiological study. Orthopaedics 1980;3:880e3. Callaghan MJ, Oldham JA. The role of quadriceps exercises in the treatment of patellofemoral pain syndrome. Sports Medicine 1996;21:384e91. Callaghan MJ, Selfe J. Has the incidence or prevalence of patellofemoral pain in the general population in the United Kingdom been properly evaluated? Physical Therapy in Sport 2007;8:37e43. Carlsson AM, Werner S, Mattlar C-E, Edman G, Puukka P, Eriksson E. Personality in patients with long term patello-femoral pain syndrome. Knee Surgery, Sports Traumatology, Arthroscopy 1993;1:178e83. Chesworth BM, Culham EG, Tata GE, Peat M. Validation of outcome measures in patients with patellofemoral syndrome. Journal of Orthopaedic and Sports Physical Therapy 1989;11:302e8. Clark DI, Downing N, Mitchell J, Coulson L, Syzpryt EP, Doherty M. Physiotherapy for anterior knee pain: a randomised controlled trial. Annals of Rheumatic Diseases 2000;59:700e4. Coe R. What is an ‘Effect Size’? A Guide for Users. World Wide Internet Web. Curriculum, Evaluation and Management Centre, University of Durham, ; 2000 [Last accessed: 16/09/2007]. Cowan SM, Bennell KL, Hodges PW, Crossley KM, McConnell J. Delayed onset electromyographic activity of vastus medialis relative to vastus lateralis in subjects with patellofemoral pain syndrome. Archives of Physical Medicine and Rehabilitation 2001;82:183e9. Cowan SM, Bennell KL, Hodges PW, Crossley KM. Altered vastii recruitment when people with patellofemoral pain syndrome complete a postural task. Archives of Physical Medicine and Rehabilitation 2002a;83:989e95. Cowan SM, Bennell KL, Crossley KM, Hodges PW, McConnell J. Physical therapy alters recruitment of the vasti in patellofemoral pain syndrome. Medicine and Science in Sports and Exercise 2002b;34:1879e85. Cowan SM, Bennell KL, Hodges PW. Therapeutic patella taping changes the timing of vasti muscle activation in people with patellofemoral pain syndrome. Clinical Journal of Sports Medicine 2002c;12:339e47.

262

G. Syme et al. / Manual Therapy 14 (2009) 252e263

Crossley K, Bennell K, Green S, Cowan S, McConnell J. Physical therapy for patellofemoral pain. A randomized, double-blinded, placebo-controlled trial. American Journal of Sports Medicine 2002;30:857e65. Crossley KM, Cowan SM, McConnell J, Bennell KL. Physical therapy improves knee flexion during stair ambulation in patellofemoral pain. Medicine and Science in Sports and Exercise 2005;37: 176e83. Crozier L. Theory of strapping. Course notes, B.Sc. Physiotherapy. Glasgow: Glasgow Caledonian University; 1989 [formerly Queen’s College]. Davies GJ, Manske RC, Cooley K, Fletcher-Klos D, Johnson-Stuhr P. Selective activation of the vastus medialis oblique: what does the literature really tell us? Physiotherapy Canada 2001;Spring: 136e51. Dillon PZ, Updyke WF, Allen WC. Gait analysis with reference to chondromalacia patellae. Journal of Orthopaedic and Sports Physical Therapy 1983;5:127e31. DeLorme T. Restoration of muscle power by heavy resistance exercise. Journal of Bone and Joint Surgery 1945;27:645e67. Dursun N, Dursun E, Kilic Z. Electromyographic biofeedback e controlled exercise versus conservative care for patellofemoral pain syndrome. Archives of Physical Medicine and Rehabilitation 2001;82:1692e5. Ellenbecker TS, Davies GJ. Closed kinetic chain exercises: a comprehensive guide to multiple joint exercises. Leeds, UK: Human Kinetics; 2001. Feine JS, Lund JP. An assessment of the efficacy of physical therapy and physical modalities for the control of chronic musculoskeletal pain. Pain 1997;71:5e23. Filho MGB, Ribeiro LCS, Garcı´ a FG. Comparison of personality characteristics between high-level Brazilian athletes and non-athletes (English version). Revista Brasileira de Medicina do Esporte 2005;11:114ee8. Gaffney K, Fricker P, Dwyer T, Barrett E, Skibinski K, Coutts R. Patellofemoral joint pain: a comparison of two treatment programmes. Excel 1992;8:179e89. Goh AC. Vastus medialis oblique: are we blind to the evidence? Physiotherapy Singapore 2000;3:2e3. Grabiner MD, Koh TJ, Draganich LF. Neuromechanics of the patellofemoral joint. Medicine and Science in Sports and Exercise 1994;26:10e21. Grelsamer RP, McConnell J. The patella e a team approach. USA: Aspen Publishers; 1998. Grimby G, Thomee´ R. Strength and endurance. In: Frontera WR, editor. Rehabilitation of sports injuries. Oxford, UK: Blackwell Press; 2003. p. 258e73. Grimmer K, Bowman P, Roper J. Episodes of allied health outpatient care: an investigation of service delivery in acute public hospital settings. Disability and Rehabilitation 2000;22:80e7. Grimmer K, Bialocerkowski A, Kumar S, Milanese S. Implementing evidence in clinical practice: the ‘therapies’ dilemma. Physiotherapy 2004;90:189e94. Harrison EL, Sheppard MS, McQuarrie AM. A randomized controlled trial of physical therapy treatment programs in patellofemoral pain syndrome. Physiotherapy Canada 1999;Spring:93e106. Hollis S, Campbell F. What is meant by intention to treat analysis? British Medical Journal 1999;319:670e4. Hopkins, WG. A new view of statistics: rank transformation: nonparametric models. World Wide Internet Web; 2000. [Last accessed: 16/09/2007]. Hubbard JK, Sampson HW, Elledge JR. Prevalence and morphology of the vastus medialis oblique muscle in human cadavers. Anatomy Record 1997;249:135e42. Jenkinson C, Layte R, Wright L, Coulter A. The U.K. SF-36: an analysis and interpretation manual, A guide to health status

measurement with particular reference to the Short Form 36 Health Survey. Oxford: Joshua Horgan Print Partnership; 1996. Jensen MP, Karoly P, Braver S. The measurement of clinical pain intensity: a comparison of six methods. Pain 1986;27:117e26. Jensen R, Hystad T, Baerheim A. Knee function and pain related to psychological variables in patients with long-term patellofemoral pain syndrome. Journal of Orthopaedic and Sports Physical Therapy 2005;35:594e600. Karst GM, Willet GM. Onset timing of electromyographic activity in the vastus medialis oblique and vastus lateralis muscle in subjects with and without patellofemoral pain syndrome. Physical Therapy 1995;75:811e23. Kowall MG, Kolk G, Nuber GW, Cassici JE, Stern SH. Patellar taping in the treatment of patellofemoral pain e a prospective randomised study. American Journal of Sports Medicine 1996;24: 61e6. Kraemer WJ, Ratamess NA. Fundamentals of resistance training: progression and exercise prescription. Medicine and Science in Sports and Exercise 2004;36:674e88. Lephart SM, Fu FH. Proprioception and neuromuscular control in joint stability. Leeds: Human Kinetics; 2000. Loudon JK, Gajewski B, Goist-Foley HL, Loudon KL. The effectiveness of exercise in treating patellofemoral-pain syndrome. Journal of Sport Rehabilitation 2004;13:323e42. Malone T, Davies G, Walsh WM. Muscular control of the patella. Clinics in Sports Medicine 2002;21:349e62. Manske RC, Davies GJ. A nonsurgical approach to examination and treatment of the patellofemoral joint, part 2: pathology and nonsurgical management of the patellofemoral joint. Critical Reviews in Physical and Rehabilitation Medicine 2003;15:141e66. Maupas E, Paysant J, Martinet N, Andre´ J-M. Asymmetric leg activity in healthy subjects during walking, detected by electrogoniometry. Clinical Biomechanics 1999;14:403e11. McConnell J. The management of chondromalacia patellae: a long term solution. Australian Journal of Physiotherapy 1986;32:215e23. Melzack R. The McGill pain questionnaire: major properties and scoring methods. Pain 1975;1:277e99. Mulder T, Nienhuis B, Pauwels J. Clinical gait analysis in a rehabilitation context: some controversial issues. Clinical Rehabilitation 1998;12:99e106. Nadeau S, Gravel D, He´bert LJ, Arsenault AB, Lepage Y. Gait study of patients with patellofemoral pain syndrome. Gait and Posture 1997;5:21e7. Na¨slund J, Na¨slund U-B, Odenbring S, Lundeberg T. Comparison of symptoms and clinical findings in subgroups of individuals with patellofemoral pain. Physiotherapy Theory and Practice 2006;22: 105e18. Neptune RR, Wright. IC, van den Bogert AJ. The influence of orthotic devices and vastus medialis strength and timing on patellofemoral loads during running. Clinical Biomechanics 2000;15:611e8. Ng GYF, Zhang AQ, Li CK. Biofeedback exercise improved the EMG activity ratio of the medial and lateral vasti muscles in subjects with patellofemoral pain syndrome. Journal of Electromyography and Kinesiology 2008;18:128e33. Nijs J, Van Geel C, Van der Auwera C, Van de Velde B. Diagnostic value of five clinical tests in patellofemoral pain syndrome. Manual Therapy 2006;11:69e77. Noyes FR, Barber SD, Mooar LA. A rationale for assessing sports activity levels and limitations in knee disorders. Clinical Orthopaedics and Related Research 1989;246:238e49. Nozic M, Mitchell J, de Klerk D. A comparison of the proximal and distal parts of the vastus medialis muscle. Australian Journal of Physiotherapy 1997;43:277e81. Ono T, Riegger-Krugh C, Boostein NA, Eisemann Shimizu M, Kanai A, Otsuka A. The boundary of the vastus medialis oblique and the vastus medialis longus. Journal of Physical Therapy Science 2005;17:1e4.

G. Syme et al. / Manual Therapy 14 (2009) 252e263 Peeler J, Cooper J, Porter MM, Thliveris JA, Anderson JE. Structural parameters of the vastus medialis muscle. Clinical Anatomy 2005;18:281e9. Powers C, Perry J, Hsu A, Hislop HJ. Are patellofemoral pain and quadriceps femoris muscle torque associated with locomotor function? Physical Therapy 1997;77:1063e78. Powers CM, Heino JG, Rao S, Perry J. The influence of patellofemoral pain on lower limb loading during gait. Clinical Biomechanics 1999;14:722e8. Powers CM. The influence of altered lower-extremity kinematics on patellofemoral joint dysfunction: a theoretical perspective. Journal of Orthopaedic and Sports Physical Therapy 2003;33:639e46. Qi Z, Ng GYF. EMG analysis of vastus medialis obliquus/vastus lateralis activities in subjects with patellofemoral pain syndrome before and after a home exercise program. Journal of Physical Therapy Science 2007;19:131e7. Risberg MA, Holm I, Ekeland A. Reliability of functional knee tests in normal athletes. Scandinavian Journal of Medicine and Science in Sports 1995;5:24e8. Rowe PJ, Myles CM, Hillman SJ, Hazelwood ME. Validation of flexible electrogoniometry as a measure of joint kinematics. Physiotherapy 2001;87:479e88. Ruffin IV MT, Kiningham RB. Anterior knee pain: the challenge of patellofemoral syndrome. American Family Physician 1993;47:185e94. Ruta DA, Garrat AM, Leng M, Russell IT, Macdonald LM. A new approach to the measurement of Quality of Life: the patient generated index (PGI). Medical Care 1994;32:1109e26. Ruta DA, Garrat AM, Russell IT. Patient centred assessment of Quality of Life for patients with four common conditions. Quality in Health Care 1999;8:22e9. Selfe J, Harper L, Pedersen I, Breen-Turner J, Waring J. Four outcome measures for patellofemoral joint problems: part 1. Development and validity. Physiotherapy 2001a;87:507e15. Selfe J, Harper L, Pedersen I, Breen-Turner J, Waring J. Four outcome measures for patellofemoral joint problems: part 2. Reliability and clinical sensitivity. Physiotherapy 2001b;87:516e22. Shrier I, Gossal K. Myths and truths of stretching. The Physician and Sportsmedicine 2000;28:57e63. Thomee´ R. A comprehensive treatment approach for patellofemoral pain syndrome in young women. Physical Therapy 1997;77:1690e703.

263

Timm K. Randomized controlled trial of protonics on patellar pain, position, and function. Medicine and Science in Sports and Exercise 1998;30:665e70. Torstveit MK, Sundgot-Borgen J. Low bone mineral density is two to three times more prevalent in non-athletic premenopausal women than in elite athletes: a comprehensive controlled study. British Journal of Sports Medicine 2005;39:282e7. Vega J, Golano´ P, Pe´rez-Carro L. Electrosurgical arthroscopic patellar denervation. Journal of Arthroscopic and Related Surgery 2006;22:1028.e1e3. Ware JE, Kosinski M, Gandek B. SF-36Ò health survey: manual and interpretation guide. Lincoln, RI: Quality Metric Incorporated; 2000. Ware JE, Kosinski M. SF36 physical and mental health summary scales: a manual for users of version 1. Lincoln, Rhode Island: Quality Metric Incorporated; 2001. Wilk KE, Reinold MM. Principles of patellofemoral rehabilitation. Sports Medicine and Arthroscopy Review 2001;9:325e36. Witonski D, Karlinska L, Musial A. Personality characteristics in patients with anterior knee pain syndrome. Medical Science Monitor 1998;4:1019e23. Witvrouw E, Daneels L, Van Tiggelen D, Willems TM, Cambier D. Open versus closed kinetic exercises in patellofemoral pain: a 5year prospective randomized study. American Journal of Sports Medicine 2004;32:1122e30. Witvrouw E, Werner S, Mikkelsen C, Van Tiggelen D, Vanden Berghe L, Cerulli G. Clinical classification of patellofemoral pain: guidelines for non-operative treatment. Knee Surgery, Sports Traumatology, Arthroscopy 2005;13:122e30. WHO. ICF international classification of functioning, disability and health. Geneva, Switzerland: World Health Organization; 2001. Yildiz Y, Aydin T, Sekir U, Cetin C, Ors F, Kalyon TA. Relation between isokinetic muscle strength and functional capacity in recreational athletes with chondromalacia patellae. British Journal of Sports Medicine 2003;37:475e9. Yip SLM, Ng GYF. Biofeedback supplementation to physiotherapy exercise programme for rehabilitation of patellofemoral pain syndrome: a randomized controlled study. Clinical Rehabilitation 2006;20:1050e7.

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 264e269 www.elsevier.com/math

Original Article

Noninvasive analysis of fascicle curvature and mechanical hardness in calf muscle during contraction and relaxation Hsing-Kuo Wang a,*, Yu-Kuang Wu a, Kwan-Hwa Lin b, Tzyy-Yuang Shiang c a

Sports Physiotherapy Laboratory, School and Graduate Institute of Physical Therapy, College of Medicine, National Taiwan University, Floor 3, No. 17, Xuzhou Road, Zhongzheng District, Taipei City 100, Taiwan, ROC b School and Graduate Institute of Physical Therapy, College of Medicine, National Taiwan University, Taiwan, ROC c Institute of Sports Equipment Technology, Taipei Physical Education College, Taipei, Taiwan, ROC Received 27 July 2007; received in revised form 18 February 2008; accepted 28 February 2008

Abstract The purpose of this study was to investigate whether changes of fascicle curvature and muscle hardness of the gastrocnemius muscle during relaxation and isometric contraction could be measured using a noninvasive approach. Seventeen male college students (age 21.0  1.5 years) participated in this study. Measurements were made during the resting state and maximal isometric plantarflexion. Fascicle curvature (m1) of the gastrocnemius medialis was measured by ultrasonography. Muscle hardness (kg/mm) was measured with a myotonometer. Angle of ankle joint ( ), amplitude of electromyographic activities (mV), and plantarflexion force (kg) were simultaneously recorded using an electrogoniometer, surface electromyography (EMG), and a load cell, respectively. Results demonstrated that the joint angle, electromyographic activities, and force at muscle contraction for the myotonometer and ultrasound conditions were not significantly different (all p > 0.05). Hardness and fascicle curvature during maximal isometric plantarflexion were significantly greater than those at rest ( p ¼ 0.002 and p < 0.001, respectively). Correlations between changes in fascicle curvature and changes of muscle hardness that took place between muscle relaxation and maximal contraction were significant (r ¼ 0.832, p ¼ 0.011). This study demonstrates that ultrasonographic and myotonometric measurements are useful to quantify changes in muscle geometry and mechanical properties for muscles during isometric contraction. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Ultrasonography; Muscle architecture; Muscle mechanical property; Muscle contraction

1. Introduction The hardness of human skeletal muscles has been measured noninvasively by hardness meter (Ashina et al., 1998), compartment evaluator (Steinberg and Gelberman, 1994; Steinberg, 2005), and myotonometer (Leonard et al., 2001, 2003, 2004). These instruments measure the displacement of a muscle to which a compressive force is applied perpendicularly. The slope of

* Corresponding author. Tel.: þ886 2 33228134; fax: þ886 2 33228160. E-mail address: [email protected] (H.-K. Wang). 1356-689X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.02.011

the linear part of the forceedisplacement curve is representative of the mechanical hardness or transverse stiffness of the compressed muscle (Steinberg and Gelberman, 1994; Ashina et al., 1998; Steinberg, 2005). Measurements for the muscle hardness or transverse stiffness are validated by a two-layered spring model representing muscles and surrounding fascia (Horikawa et al., 1993). A greater muscle hardness (transverse stiffness) was found in the affected muscles of patients with compartment syndrome (Steinberg, 2005), cerebral palsy (Aarrestad et al., 2004), hemiparesis (Leonard et al., 2001), and those of healthy subjects after performing eccentric exercises (Murayama et al., 2000). Leonard et al. (2004) found that the transverse stiffness

H.-K. Wang et al. / Manual Therapy 14 (2009) 264e269

of myotonometric measurements correlates positively with the activity of surface electromyography (EMG) during different levels of voluntary isometric contraction in healthy subjects. Collectively, these studies indicate that muscle hardness is affected by muscle disorders and levels of activation. However, these measurements are performed on the surface of the skin over the muscle, and whether the underlying architecture of the contracting muscle affects the mechanical hardness is unknown. Ultrasonographic measurements of fascicle geometry of the medial gastrocnemius muscle have been confirmed by direct measurement of cadaveric muscles (Narici et al., 1996). Fascicle curvature is defined as the reciprocal of the radius of the fascicle circle (van Leeuwen and Spoor, 1992). Real-time ultrasound has been used to demonstrate that the fascicle of healthily young volunteers is significantly more curved for the medial gastrocnemius muscle during isometric maximal voluntary contraction (MVC) versus the resting state (Muramatsu et al., 2002). Some researchers have suggested the curved muscle fascicle of contracting muscles as a possible cause to explain the elevated intramuscular pressure during isometric contractions (Hill, 1948; Sejersted et al., 1984). To date, the relationship between real-time alterations in the fascicle curvature and changes of muscle hardness in skeletal muscles in vivo is unknown. Thus, the purpose of this study was to analyze the relationship between changes of muscle fascicle curvature and muscle hardness for the medial gastrocnemius muscle in vivo during relaxation and isometric contraction. Potential confounding effects of angle of the ankle joint (Karamanidis et al., 2005), muscular electrical activities, and maximal plantarflexion force were controlled to ensure a standardized protocol.

2. Material and methods 2.1. Subjects Our institutional review board approved this study, and all subjects provided informed written consent. Inclusion criteria were subjects without any history of knee, leg, and ankle pain during their lifetime. Twenty-one male students were recruited through a sports center at a university. All subjects played in recreational clubs including tennis, volleyball, and basketball. The morphology, architecture, and contractile capacity of human pinnate muscles respond to intensive training (Aagaard et al., 2001). Therefore, four students were excluded to reduce confounding effects of training, because they had been training for their specific sport since teenagers. Data of 17 college students (age 21.0  1.5 years; body height 171.4  7.4 cm; weight 63.8  7.6 kg) were analyzed.

265

2.2. Experiment protocol After a 5-min warm-up session, each subject lay prone on an examination bed with both ankles hanging over the edge of the bed. The lower back and knees of each subject were tightly secured by straps to prevent any trunk, back, or knee movements. All measurements were performed on the right leg, the dominant leg in all subjects, with the foot positioned at 90 relative to the tibia and fixed on a footplate. Leg dominance was determined by asking subjects which leg they used to kick a ball, this was the dominant leg. Because the ultrasound and myotonometer could not be applied at the same time, the subjects were randomly assigned to one of two testing sequences: either the ultrasound first and then the myotonometer 20 min later, or vice versa. During each test, subjects were asked to perform at least two repetitions of a 5-s muscle relaxation and a 5-s maximal isometric plantarflexion, with a 2-min rest between each repetition. In each repetition, three myotonometric and ultrasonographic measures (measure interval 1.5 s) were made in 5 s for (1) muscle relaxation and (2) isometric plantarflexion, at approximately 1, 2.5, and 4 s. Synchronized recordings of the electrogoniometer, surface EMG, and force were made simultaneously. Forces of plantarflexion (kg), plantarflexion angle of ankle joint ( ), and root mean square of electromyographic amplitude (mV) corresponding to the time of three myotonometric or ultrasonographic measurements performed were averaged. Variables were then averaged over the repetitions. The effectiveness of the ankle fixation method was assessed using an electrogoniometer (Sharp Sensor S700, Measurand Inc., Fredericton, Canada). In addition, muscular activities from the gastrocnemius medialis were recorded by EMG using one channel bipolar Ag/ AgCl surface electrodes (Model: MP100WSW, Biopac Systems Inc., Goleta, USA). With the right foot fixed, maximal plantarflexion force (kg) was recorded by a load cell (Model: S6001, Celtron Techniques Inc., Taipei, Taiwan) connected to the footplate. The signals of the eletrogoniometer, surface EMG, and load cell were stored on a personal computer (ASUS D672, Taipei, Taiwan, ROC) at 1200, 1200, and 200 Hz, respectively, via an A/D converter (Model: MP100WSW, Biopac Systems Inc., Goleta, USA). Real-time development or changes of the force, amplitudes of surface EMG, and angle of the ankle joint were displayed on a computer monitor with a software window (AcqKnowledge software, Biopac System Inc., Goleta, USA). During relaxation, no significant electromyographic activities of the muscle were observed. Maximal isometric plantarflexion was defined as the greatest value for the real-time force for which no further increase was observed. The electrogoniometer measured the dorsiflexion/ plantarflexion (in degree) of the ankle joint when the

266

H.-K. Wang et al. / Manual Therapy 14 (2009) 264e269

ends of the goniometer were fixed to the medial side of the lower leg parallel to the tibial bone and to the medial side of the foot parallel to the first metatarsal bone. Electromyographic activity was recorded from the midline of the gastrocnemius medialis muscle, guided by axial-plane ultrasonographic scanning. Skin preparation was taken prior to application of surface electrodes and a portable EMG instrument e Seirra II (Cadwell, Kennewick, WA, USA) e was used to ensure that the interelectrode resistance was below 5 kU. Surface electrodes with a diameter of 5 mm were placed with an interelectrode distance of 20 mm at a site corresponding to the proximal third of the gastrocnemius muscle with the reference electrode placed on the lateral malleolus of the left ankle. The raw electromyographic signal was preamplified and filtered using high- and low-pass filters set at 10 and 500 Hz, respectively (pre-amplifier: common mode rejection ratio ¼ 95 dB; impedance ¼ 100 MU; gain ¼ 350). Amplitudes (root mean square; mV) of the surface EMG were quantified for a 1/10 s epoch corresponding to the middle of each myotonometric measurement or ultrasonographic image. Measurements of muscle hardness were performed on the medial gastrocnemius muscle with a hand-held muscle hardness meter (Fig. 1) myotonometer (Neurogenic Technologies Inc., Missoula, USA). The head of the myotonometer probe (diameter: 1.4 cm) was held in a vertical position on the skin surface of the medial gastrocnemius muscle. The measure point of hardness was located at the medial side to midpoint of two recording electrodes on the gastrocnemius medialis with a distance of 1.5 cm. In each trial, three successive

measurements were performed at 1.5-s intervals with an applied force of approximately 2 kg/s. The amount of force in response to the perturbation was simultaneously recorded with the amount of penetration depth. The myotonometer was connected to a computer (TOSHIBA Satellite M40, Tokyo, Japan; Fig. 1) which provided electrical signals (hardness signal) to the Biopac system for each measurement. These signals were used to synchronize the myotonometer with measurements of root mean square of amplitude, plantarflexion force, and joint angle, using software (LabVIEW 7.1, National Instruments, TX, USA). Pilot work indicated that the curve was linear when the force in response to the perturbation was equal to or greater than 0.25 kg. In this study, muscle hardness was defined as a value obtained from linear portion (from 0.5 to 2.0 kg) of the force responseedepth curve. Hardness (the slope) of the linear part of the curve was calculated using software with best polynomial fitting (MATLAB 7.1, The MathWorks Inc., MA, USA). Intrarater reliability of the myotonometer on the lateral gastrocnemius and biceps brachii muscles varied from 0.84 to 0.99 (Leonard et al., 2003). Ultrasonographic measurements of fascicle curvature were made using a SonoSite 180 plus ultrasound (SonoSite Inc., WA, USA) with a 5e10 MHz broadband linear array transducer. The examiner positioned the probe perpendicular to the surface of the leg in such a manner as to avoid compressing the skin and underlying tissues. The ultrasound probe was placed on the sagittal plane of the gastrocnemius medialis muscle medial and parallel to the long axis of the two recording electrodes with

Fig. 1. Arrangements of myotonometric and ultrasound testing conditions. SE, surface electrodes; EG, electrogoniometer; LC, load cell; FP, footplate; US, ultrasonography; Com-1, recording computer; and Com-2, synchronization computer.

H.-K. Wang et al. / Manual Therapy 14 (2009) 264e269

a distance of 1.5 cm (the same location where the myotonometer was placed). Proximal and distal edges of the ultrasound probe corresponded to the proximal border of the proximal recording electrode and distal border of the distal recording electrode, respectively. The B mode and a view of near field (depth of view: 3.2 cm) with near-field resolution (axial resolution: 0.7 mm) were set when scanning the curvature. The real-time images were recorded using a camera (Handycam DVD803, Sony, Tokyo, Japan) and stored on a disc at 30 Hz. Software containing simulating switching circuits written using LabVIEW 7.1 (National Instruments, TX, USA) was installed in a computer (TOSHIBA Satellite M40, Tokyo, Japan) to add audio-electrical signals to the camera and Biopac system at beginning and end of measurements for synchronization (Fig. 1). Ultrasonographic images were taken at 1, 2.5, and 4 s within the 5-s period of muscle relaxation and repeated for isometric plantarflexion. Thirty ultrasound images are collected during each of the three measurements (at time point 1, 2.5, and 4 s of the 5-s relaxation/contraction). Three images with the best quality in curvature at each time point are used for the determination of fascicle curvature. While scanning, echoes from the interfascicular space were identified and the fascicular path was defined as an arc of a circle between the superior and deep aponeuroses. Determination of the fascicular curvature was made as described by Muramatsu et al. (2002). Fascicle curvature ¼ jcos 2qd  cos 2qs j=½2$Muscle thicknessðcos qd þ cos qs Þ: For the meaning of each symbol, see Fig. 2. All morphological measurements were made by the same examiner with software from AutoCAD (AutoCAD,

267

Autodesk Inc., CA, USA). Previous work from our laboratory demonstrated that the interday reproducibility of calculating the fascicle curvature on the basis of a coefficient of variation was 10.1%. 2.3. Statistics Bonferroni t-method tested whether variables from surface EMG (mean values of root mean square of amplitude (mV)), electrogoniometer (angle of ankle joint ( )), and load cell (maximal force (kg)) were similar between the myotonometer and ultrasound conditions during isometric plantarflexion. Mean values of muscle hardness and fascicle curvature were compared for the muscle at rest and maximal isometric plantarflexion by paired t test. Correlations between curvature changes and changes of muscle hardness that took place between muscle relaxation and maximal contraction were analyzed by Spearman correlation tests.

3. Results Mean values of root mean square of amplitude (mV), angle of ankle joint ( ), and force of maximal isometric plantarflexion (kg) are summarized in Table 1. No significant differences of these above variables were observed between the myotonometer and ultrasound conditions, suggesting that the configuration was similar for the conditions (all p values > 0.05) (Table 1). Hardness and fascicle curvature of maximal isometric plantarflexion were significantly greater than those of the muscle at rest ( p ¼ 0.002 and p < 0.001, respectively) (Table 2). A significant correlation was found between changes of curvature and muscle hardness in the gastrocnemius medialis (r ¼ 0.832, p ¼ 0.011) (Table 2).

Fig. 2. Ultrasonographic measurements of the fascicle curvature. Fascicle curvature ¼ jcos 2qd  cos 2qs j=½2$Muscle thicknessðcos qd þ cos qs Þ: (Muramatsu et al., 2002). qs is the angle formed between the tangent (L3) of the fascicle at the intersection made by the fascicle and the superficial aponeurosis and the tangent (L1) of the deep aponeurosis at the intersection made by the fascicle and the deep aponeurosis. qd is the angle formed   between the tangent (L2) of the fascicle at the intersection made by the fascicle and the deep aponeurosis and L1. Results showed qs ¼ 10 ; qd ¼ 19 ; pixel distance which equals to 1.0 cm shows 10.01 and muscle thickness ¼ 1.21 cm in this image (pixel distance ¼ 12.12).

268

H.-K. Wang et al. / Manual Therapy 14 (2009) 264e269

Table 1 Mean values (standard error) of root mean square of amplitude of the gastrocnemius medialis, angle of ankle joint, and force of maximal isometric plantarflexion in two measure conditions. Ankle joint ( )

Amplitude (mV)

Force (kg)

Myotonometric measurement Ultrasound measurement

88.6  2.7

0.636  0.394

55.7  12.7

89.0  3.1

0.615  0.479

56.2  13.1

p Value

0.138

0. 214

0.775

4. Discussion Physiotherapists often use palpation to evaluate muscular shapes and firmness in conditions of muscle relaxation or contraction. However, quantitative measurements for muscle contraction status are lacking. We used a rigorous and standardized protocol to reduce influences from the joint position and force (Table 1). Results of this study indicate that contracting muscle displays a significantly greater fascicle curvature and muscle hardness compared to the resting muscle (Table 2). As compared with the report by Muramatsu et al. (2002), the curvature observed in contracting muscles was greater in our study (9.4  3.3 versus 5.5  2.2 m1). This may be due to the different methodologies used for curvature measurements. The measurements in this study were performed with a distance of 1.5 cm to long axis of the gastrocnemius medialis muscle, while Muramatsu et al. (2002) conducted their study on the long axis of muscle. Nevertheless, these results confirm previous findings demonstrating that the muscle curvature at maximum contraction was larger, and amounts of muscle displacement to force applied perpendicular to the contracted muscle are less (Muramatsu et al., 2002; Leonard et al., 2004). This indicates that contraction levels affect the muscle architecture and mechanical properties, and measurements of curvature and muscle hardness are capable of objectively assessing these changes. Both curvature and hardness techniques are recommended for studies interested in exploring muscle architecture or mechanical properties during contraction. In addition, our study is the first to demonstrate a significant correlation between the changes of fascicle

Table 2 Mean values (standard error) of muscle hardness and fascicle curvature of the gastrocnemius medialis at muscle relaxation and isometric maximal plantarflexion. Hardness (kg/mm)

Fascicular curvature (m1)

Relaxation Maximal contraction

0.63  0.14 0.86  0.26

0.7  2.0 9.4  3.3

p Value

0.002

<0.001

curvature and muscle hardness for resting and contracting muscles (r ¼ 0.832, p ¼ 0.011) (Table 2). These findings are particularly important for understanding the effects of in vivo changes of muscle architecture on the alteration of mechanical functions of active human skeletal muscles. The changes of fascicular curvature, from relaxation to maximal isometric contraction, accounted for 70.0% of the changes found in muscle hardness. Possibly, the curving fascicles observed in contracting muscles may lead to an increased hardness. During muscle contraction, fascicles become curved because muscle fibers on the concave side are shorter than fibers on the convex side. These shorter fibers produce a higher tension and a greater compression force to the convex side that generates resistance, much like structural beams reinforce the lining of a tunnel. This could explain how, in terms of changes of muscle architecture, contracting muscles are capable of opposing compression forces with less deformation than relaxed muscles. Ankle joints exert a significant effect of curvature (Muramatsu et al., 2002). To confirm the effects of curvature on muscle hardness, we recommend that future hardness studies are conducted using different degrees of ankle dorsiflexion/plantarflexion. Another explanation for how the changes of fascicular curvature account for 70.0% of the changes in muscle hardness is that the increased curvatures occurring with contraction elevate intramuscular pressure and lead to an increase of muscle hardness. The quantitative muscle hardness is related to the increase of the intramuscular pressure (Steinberg and Gelberman, 1994; Steinberg, 2005). The correlation between fascicle curvature and hardness in contracting skeletal muscles may also be caused by an elevation of intramuscular pressure in vivo (Hill, 1948; Sejersted et al., 1984; van Leeuwen and Spoor, 1992). To test this hypothesis, studies with invasive measurements of intramuscular pressure with the measurement of ultrasonography are required. Muscle pathologies such as delayed onset of muscle soreness (DOMS) following eccentric exercises have been found to have increasing muscle hardness (Murayama et al., 2000). Massage may effectively minimize muscle soreness (Hilbert et al., 2003). Noninvasive measures of hardness and curvature may assist in understanding the underlying mechanism of such treatment. Ultrasonographic and myotonometric measurements may be useful to quantify changes in muscle geometry and mechanical properties following muscle treatments. One limitation of this study is that only maximum contraction was investigated, different levels of contraction were not identified. Another limitation is that the pathological changes and relationships to muscle hardness of the normalized fascicle curvature for various conditions, such as chronic compartment syndrome, may show different correlation coefficients from healthy subjects, and further studies are suggested.

H.-K. Wang et al. / Manual Therapy 14 (2009) 264e269

5. Conclusion In this study, measurements of fascicle curvatures and muscle hardness were capable of discriminating between resting and maximal contracting states. The correlation of the curvature change with hardness suggests that morphological changes of muscles may lead to changes in hardness of skeletal muscles. As these noninvasive techniques provide insight into the mechanisms of contraction, future studies should determine whether measures of curvature and hardness are affected by muscle pathologies or subsequent treatments.

References Aagaard P, Andersen JL, Dyhre-Poulsen P, Leffers AM, Wagner A, Magnusson SP, et al. A mechanism for increased contractile strength of human pennate muscle in response to strength training: changes in muscle architecture. Journal of Physiology 2001;534(2):613e23. Aarrestad DD, Williams MD, Fehrer SC, Mikhailenok E, Leonard CT. Intra- and interrater reliabilities of the myotonometer when assessing the spastic condition of children with cerebral palsy. Journal of Child Neurology 2004;19(11):894e901. Ashina M, Bendtsen L, Jensen R, Sakai F, Olesen J. Measurement of muscle hardness: a methodological study. Cephalalgia 1998;18(2):106e11. Hilbert JE, Sforzo GA, Swensen T. The effects of massage on delayed onset muscle soreness. British Journal of Sports Medicine 2003;37(1):72e5. Hill AV. The pressure developed in muscle during contraction. Journal of Physiology 1948;107(4):518e26. Horikawa M, Ebihara S, Sakai F, Akiyama M. Non-invasive measurement method for hardness in muscular tissues. Medical & Biological Engineering & Computing 1993;31(6):623e7. Karamanidis K, Stafilidis S, DeMonte G, Morey-Klapsing G, Bruggemann GP, Arampatzis A. Inevitable joint angular rotation

269

affects muscle architecture during isometric contraction. Journal of Electromyography and Kinesiology 2005;15(6):608e16. van Leeuwen JL, Spoor CW. Modelling mechanically stable muscle architectures. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 1992;336(1277):275e92. Leonard CT, Stephens JU, Stroppel SL. Assessing the spastic condition of individuals with upper motoneuron involvement: validity of the myotonometer. Archives of Physical Medicine and Rehabilitation 2001;82(10):1416e20. Leonard CT, Deshner WP, Romo JW, Suoja ES, Fehrer SC, Mikhailenok EL. Myotonometer intra- and interrater reliabilities. Archives of Physical Medicine and Rehabilitation 2003;84(6):928e32. Leonard CT, Brown JS, Price TR, Queen SA, Mikhailenok EL. Comparison of surface electromyography and myotonometric measurements during voluntary isometric contractions. Journal of Electromyography and Kinesiology 2004;14(6):709e14. Muramatsu T, Muraoka T, Kawakami Y, Shibayama A, Fukunaga T. In vivo determination of fascicle curvature in contracting human skeletal muscles. Journal of Applied Physiology 2002;92(1):129e34. Murayama M, Nosaka K, Yoneda T, Minamitani K. Changes in hardness of the human elbow flexor muscles after eccentric exercise. European Journal of Applied Physiology 2000;82(5e6):361e7. Narici MV, Binzoni T, Hiltbrand E, Fasel J, Terrier F, Cerretelli P. In vivo human gastrocnemius architecture with changing joint angle at rest and during graded isometric contraction. Journal of Physiology 1996;496(1):287e97. Sejersted OM, Hargens AR, Kardel KR, Blom P, Jensen O, Hermansen L. Intramuscular fluid pressure during isometric contraction of human skeletal muscle. Journal of Applied Physiology 1984;56(2):287e95. Steinberg BD, Gelberman RH. Evaluation of limb compartment with suspected increased interstitial pressure. A noninvasive method for determining quantitative hardness. Clinical Orthopaedics and Related Research 1994;300:248e53. Steinberg BD. Evaluation of limb compartments with increased interstitial pressure. An improved noninvasive method for determining quantitative hardness. Journal of Biomechanics 2005;38(8):1629e35.

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 270e277 www.elsevier.com/math

Original Article

Comparison of visual and ultrasound based techniques to measure head repositioning in healthy and neck-pain subjects Alexandra Roren*, Marie-Anne Mayoux-Benhamou, Fouad Fayad, Serge Poiraudeau, Didier Lantz, Michel Revel Department of Rehabilitation, Cochin Hospital, 27 Rue du Faubourg Saint Jacques, 75679 Paris Cedex 14, France Received 26 March 2007; received in revised form 13 February 2008; accepted 1 March 2008

Abstract Three-dimensional (3D) ultrasound based (US) and usual Revel visual techniques were compared to measure head repositioning ability in 41 healthy subjects and 41 subjects with neck pain. Head repositioning absolute value of the global error (AE) was calculated by both techniques after active head rotations. The AE was 3.6 and 3.7 for healthy subjects and 6.3 and 6.1 for neck-pain subjects for the visual and US techniques, respectively. The AE was higher in neck-pain subjects ( p < 0.001), and a value of 4.5 was identified as a threshold of abnormal repositioning for both techniques. The testeretest reliability, calculated in the neckpain subjects, was moderate (intraclass correlation coefficient [ICC] ¼ 0.68) for both techniques. The correlation between the two techniques for AE was poor for both groups with successive measurement of visual and US techniques (r ¼ 0.32 and 0.46, respectively) but excellent with simultaneous measurement (r ¼ 0.95 for both groups). Moreover, we showed substantial agreement between the techniques in discriminating healthy and neck-pain subjects (kappa ¼ 0.65). The Revel visual technique is more appropriate for clinical practice, but with improved software, the 3D US method could provide additional quantitative and qualitative data invaluable for research. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Cervical spine; Proprioception; Head repositioning; 3D ultrasound motion analysis

1. Introduction The neural control of the headeneck system depends on the convergence of the cues provided by the proprioceptive, vestibular and visual paths (Lakie and Loram, 2006). The vestibular and proprioceptive systems interact (Karnath et al., 1994), and oculo-cervical coupling has been identified (Andre-Deshays et al., 1988). The cervical proprioception contributes importantly to head position and to head orientation in space, as suggested by the wealth of muscular and articular receptors * Corresponding author. Tel. þ33 1 58 41 25 41; fax: þ33 1 58 41 25 45. E-mail address: [email protected] (A. Roren). 1356-689X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.03.002

(McLain, 1994; Proske, 2005). The role of kinesthetic sensitivity disturbance in dizziness following cervical trauma has been emphasized (Treleaven et al., 2005). Even with nonspecific cervical pathologic conditions, alterations of proprioception should disrupt heade neck postural and dynamic control (Revel et al., 1991). A reliable measurement of sense of head and neck position is necessary. Revel et al. (1991) devised a simple test to assess cervical proprioception using head repositioning ability. This test consists of a visual measurement of the error of relocating the head to the initial neutral position after an active cervical rotation. The ability to relocate the head was altered in patients with neck pain, and the authors identified a threshold of repositioning error expressed

A. Roren et al. / Manual Therapy 14 (2009) 270e277

in degrees that discriminates abnormal from normal repositioning. The mean error was >4.5 for 89% of the neck-pain subjects and <4.5 for 89% of the healthy subjects. The head repositioning test was also described as an exercise and used to measure the effect of a neckpain rehabilitation program (Revel et al., 1994). This test has been widely used (Heikkila and Astrom, 1996; De Hertogh et al., 2000; Rix and Bagust, 2001; Kristjansson et al., 2001; Humphreys et al., 2002) and the testeretest reliability has been evaluated in a group of healthy subjects (Pinsault et al., 2006). Recently, an ultrasound based (US) technique was used with various experimental procedures to measure sense of head and neck position in asymptomatic populations (Strimpakos et al., 2006; Lee et al., 2006). We aimed to compare the original Revel visual method with a three-dimensional (3D) US technique in healthy subjects and those with neck pain to evaluate the clinical utility of the techniques and intra-rater reliability in head repositioning assessment.

2. Patients and methods 2.1. Population To be included in the study, healthy subjects and those with neck pain had to be older than 18 years and not have any vestibular, neurological or inflammatory rheumatic disorders. All subjects were patients from a rehabilitation department and were therefore screened by the medical team. We collected data on age, sex, height, weight and sports activity, as well as duration of pain (for neckpain subjects) in months and intensity of the cervical pain at rest and during the tests by a visual analog scale (VAS, 0e100-mm length). French bioethics legislation does not require consent from the Hospital Ethics Committee for this type of study. The study was conducted in compliance with the protocol Good Clinical Practices and Declaration of Helsinki principles. In accordance with the French national law patients gave their written agreement to participate after being informed of the experimental protocol. 2.2. Instrumentation The subject, who wore a helmet (different helmet for Revel visual and US technique), was seated in a supported position on a chair with a backrest in front of a target (similar for both techniques) (Fig. 1A and B). 2.2.1. Revel visual technique For the Revel visual technique, the helmet weighed 250 g and had a light beam on top (Fig. 1C). A circular graduated target (40-cm diameter with concentric circles

271

every centimetre) was divided into four quadrants separated by two axes intersecting at 0 (the horizontal axis in abscissa and vertical axis in ordinate). The target was placed on the wall in front of the subject and adjusted to align with the subject’s reference head position. Opaque goggles (weight 30 g) were used to occlude the subject’s vision during the test. Manufacturers’ details for Revel visual technique have previously been described (Savignat and Roren, 2007). 2.2.2. US technique For the US technique, we used a US motion analysis system to measure the head repositioning error in three cardinal planes (Zebris; Zebris medizintechnik GMBH, D88316, Isny, Germany). The US helmet, weighing 194 g, was fitted with a US tripod transmitter to detect real-time cervical spine motion (Fig. 1D), and data were stored on a computer. A laser pen was fixed on the US helmet to visually determine the initial reference position with the light beam projecting on the target. Moreover, the laser pen allowed for, at each trial, a visual reading of the head repositioning error in addition to the results given by the US technique. The visual reading during the US experiment was defined as the US-visual technique. Because of the design of the helmet, the subject could not wear goggles, so the vision was occluded by taping a square piece of fabric in front of each eye with adhesive tape (Fig. 1D). 2.3. Experimental procedure The experimental procedure was similar with both Revel visual and US techniques. The US technique differed only in use of the US motion analysis system. As previously described (Revel et al., 1991), the subject, vision occluded, was instructed to adopt the selfdetermined neutral headeneck position, defined as the reference position. Then, the target was placed so that the light beam on top of the helmet pointed to the target’s center (zero [0] on the target). The reference position was thus reliable for each trial with the visual or US technique. For each trial, the subject was instructed to memorize the initial reference position, then to perform a maximal cervical rotation (horizontal plane) for about 2 s and to accurately return to the reference position without any speed constraint. Subjects performed 10 trials for each technique. Before each trial, the examiner manually repositioned the subject’s head to the reference position, and for the US technique, the motion analysis system was calibrated. The repositioning absolute value of the global error (AE) was recorded after each trial of head

272

A. Roren et al. / Manual Therapy 14 (2009) 270e277

Fig. 1. (A) Revel visual technique: subject on a chair in relaxed posture in front of the target. (B) US technique: subject on a chair in relaxed posture in front of the target, the transducer sensor stand is placed at 80 cm from the source to detect the US waves for all ranges of motion. By use of triangulation, the measurement was derived from the time delay between the US pulses measured at a sampling rate of 25 Hz. (C) Revel visual technique: helmet with a light beam, opaque goggles. (D) US technique: specially designed headgear with ultrasound tripod transmitter. A laser pen was added on top.

repositioning. No information on head repositioning ability was given to the subject. We assessed unilateral repositioning ability so as to avoid subject fatigue, since the results do not differ after right or left head rotation (Revel et al., 1991; Kristjansson et al., 2001). The direction of cervical rotation and the order of the techniques were selected at random. The experiment lasted about 30 min. We calculated testeretest reliability of the visual and US techniques in the neck-pain population by the same examiner repeating the entire procedure after a 1-h rest. With the Revel visual and the US-visual techniques, for each trial, the head repositioning error was visually and directly measured on the target by the distance between the final position of the spot of light and the initial reference position (projection of the light beam on the center of the target). Because the purpose of the experiment was the evaluation of an error in head repositioning after angular head displacement, the centimetric measurements on the target were converted to degrees from the center of

rotation (taken as the source of the light beam at the top of the helmet, 90 cm from the target). For the US technique, 3D data on cervical motion were processed by use of the Winspine V 1.78 software and expressed in degrees. The head repositioning AE corresponded to the final repositioned position registered by the US motion analysis system. We considered the Pythagoras theorem as an appropriate way to convert the 3D data (US technique) into 2D data (Revel visual technique) (Fig. 2A). 2.4. Statistical analysis Data analysis involved use of Systat 9 software. Head repositioning error was expressed by the AE as mean  standard deviation (SD) of 10 trials of head repositioning. The AE means were compared in healthy and neck-pain subjects and by the Student’s t-test. For both visual and US techniques, the relation between the AE means and quantitative anthropometric

273

A. Roren et al. / Manual Therapy 14 (2009) 270e277

The correlation between AE means for successive measurement with Revel visual and US techniques and for simultaneous measurement with US and US-visual techniques were calculated for healthy and neck-pain groups by the Spearman r. The Kappa statistic was used to calculate the percentage agreement between the Revel visual and the US techniques in terms of threshold of abnormal repositioning. Although no standard for interpretating a reliability coefficient exists, Landis and Koch (1977) suggested the following for the Cohen k statistic in terms of agreement: <0.0, poor; 0.0e0.2, slight; 0.21e0.40, fair; 0.41e0.60, moderate; 0.61e0.80, substantial; and 0.81e1.0, excellent.

3. Results All repositioning errors were expressed by AE as mean  standard deviation (SD) of 10 trials of head repositioning. 3.1. Population data Basic data for the healthy and neck-pain subjects are in Table 1. Fig. 2. (A) Conversion of the 2D data into 3D. Pythagoras equation, qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi d ¼ l ðsin a2 þ sin b2 Þ. H, light beam on the subject’s head; d, distance between the initial reference position O and the final position P; x, abscissa of P; y, ordinate of P; l, distance between the cutaneous projection of the subject’s head center of rotation (localization of the light beam) and the target; a, angle in the sagittal plane; and b, angle in the horizontal plane. (B) Conversion of the measurements in centimetres to degrees. H, light beam on the subject’s head; O, center of the target representing the initial reference position of the head; P, final position of the head; OH, distance between the light beam and the target in the initial reference position; HP, distance between the light beam and the target in the final position. The distance between ‘‘O’’ and ‘‘P’’ represents the head repositioning error in centimetres. The triangle HOP is rectangular, thus the trigonometric formula: tangent q ¼ OP/OH, q ¼ tang1 (OP/OH), was used to convert head repositioning error measured in centimetres (OP) to degrees (q).

parameters was assessed by the Spearman correlation coefficient (r) and that between the AE means and qualitative parameters (sex, sports activity) by Student’s t-test. A p < 0.05 was considered statistically significant. To determine the threshold of abnormal repositioning, we used specificity and sensitivity analysis with receiver operating characteristic (ROC) curve analysis. Testeretest reliability for the head repositioning AE in neck-pain group was assessed with the intraclass correlation coefficient (ICC) for Revel visual and US techniques (Shrout and Fleiss, 1979). In addition, the Bland and Altman (1986) method was used for testing agreement between the AE means of the two experiment sets.

3.2. Head repositioning with Revel visual and US techniques The AE was significantly greater in the neck-pain group than in the healthy group with either the Revel Table 1 Basic data of the healthy and neck-pain subjects.

Anthropometric data Male, n (%) Age (mean  SD), years Body mass index (mean  SD), kg/m2 Sport activities, n (%) Neck pain intensity At rest (mean  SD), mm During test (mean  SD), mm Neck-pain duration (mean  SD), months AE Visual technique (mean  SD), degrees US technique (mean  SD), degrees

Healthy (n ¼ 41)

Neck pain (n ¼ 41)

p

18 (44) 30.5  11.4 22.5  3.8

11 (27) 54.7  14.2 25.1  6.5

0.159 <0.0001 0.017

13 (32)

0 (0)

NA

NA NA

15.9  19.7 26.8  21.8

NA NA

NA

88.2  68.3

NA

3.6  0.8

6.3  2.4

<0.001

3.7  0.9

6.1  2.9

<0.001

Sports activities refer to the percentage of subjects playing sports for at least 3 h a week. Body mass index ¼ weight (kg)/height2 (m2). Neck pain intensity was measured on a 100-mm visual analogue scale. NA, non applicable. p < 0.05 was considered statistically significant.

274

A. Roren et al. / Manual Therapy 14 (2009) 270e277

visual or US technique (6.3  2.4 vs 3.6  0.8 [p < 0.001] and 6.1  2.9 vs 3.7  0.9 [p < 0.001], respectively). For the healthy group, the AE was not related to age (r ¼ 0.1, p ¼ 0.5 and r ¼ 0.004, p ¼ 0.98 for Revel visual and US techniques, respectively), sex ( p ¼ 0.26 and p ¼ 0.51 for Revel visual and US techniques, respectively) or sports activity ( p ¼ 0.38 and p ¼ 0.86 for Revel visual and US techniques, respectively) but was weakly related to body mass index for the Revel visual technique only (r ¼ 0.37, p ¼ 0.02). For the neck-pain group, the AE was not related to body mass index (r ¼ 0.1, p ¼ 0.54 and r ¼ 0.05, p ¼ 0.75 for Revel visual and US techniques, respectively), sex ( p ¼ 0.97 and p ¼ 0.77 for Revel visual and US techniques, respectively), or pain (r ¼ 0.07, p ¼ 0.66 and r ¼ 0.04, p ¼ 0.79 for Revel visual and US techniques, respectively) but was weakly related to age for the Revel visual technique only (r ¼ 0.37, p ¼ 0.02).

predictive values were 0.8 with both techniques. The Revel visual and US techniques showed an 80% and 82% chance, respectively, that neck-pain subjects would reposition the head outside the limit of 4.5 and that healthy subjects would reposition within this zone.

3.3. The threshold value of abnormal repositioning

3.5. Test-retest reliability of the Revel visual and US techniques for neck-pain subjects

Sensitivity and specificity analysis gave a threshold value of 4.5 for Revel visual and US techniques (Fig. 3A and B). Sensitivity was 78% with both techniques and specificity 85% and 78% for the Revel visual and US technique, respectively. Positive and negative

3.4. Comparison of techniques in AE values The correlation between the two techniques in head repositioning AE was poor for both healthy and neckpain subjects (r ¼ 0.32, p ¼ 0.04; and r ¼ 0.46, p ¼ 0.004, respectively). When measurements were taken with the Revel visual and US techniques simultaneously the two techniques showed excellent correlation in AE for both healthy and neck-pain subjects (r ¼ 0.946 and r ¼ 0.952, respectively). For the threshold value of 4.5 for head repositioning AE, the kappa statistic showed substantial agreement (kappa ¼ 0.65) between Revel visual and US techniques to discriminate between healthy and neck-pain subjects in repositioning.

ICC values were similar for both Revel visual and US techniques for the neck-pain subjects (ICC ¼ 0.68). For the US-visual technique, the ICC was 0.62. According to the Bland and Altman plot (Fig. 4A and B), most of the differences were within the confidence interval (3.6 and 4.2 for the Revel visual technique and 3.8 and 5.6 for the US technique) and mean differences between test and retest were small. We considered that X (means of the two tests) and Y (differences between the two tests) were independent for both Revel visual and US techniques (r ¼ 0.35 and p ¼ 0.04 and r ¼ 0.3 and p ¼ 0.05 for Revel visual and US techniques respectively). There was no systematic trend in the plot.

4. Discussion

Fig. 3. The receiver operating characteristic (ROC) curve was constructed for four cutpoints: 3.5 , 4 , 4.5 and 5 . The best tradeoff between sensitivity and sensibility (largest ordinate for the smallest abcissa) was observed for a cutpoint of 4.5 , defined as the threshold of abnormal repositioning. (A) ROC curve with Revel visual technique. (B) ROC curve with US technique.

This study was the first to assess AE in the head repositioning test and the clinical relevance of the test with use of two techniques, the usual Revel visual technique and a 3D US technique, in two populations: healthy subjects and those with neck pain. It was also the first to evaluate the testeretest reliability in a neck-pain group. We demonstrated that head repositioning ability could be quantitatively assessed by either technique, but the Revel visual technique might be more adapted to daily clinical practice. The head repositioning AE we observed in the healthy and neck-pain groups was in agreement with those from previous studies of the Revel visual technique, which ranged from 2.74 to 5.25 for healthy

A. Roren et al. / Manual Therapy 14 (2009) 270e277

Fig. 4. Bland and Altman plot for neck-pain population: testeretest study for the AE plotting agreement of the means of the two tests against the difference between the tests for Revel visual (A) and US (B) technique. The mean of the differences between the two scores (0.3 and 0.9 for Revel visual and US techniques, respectively) was used to calculate the 95% limits of agreement (expressed by dashed lines) for the two scores: 3.6 and 4.2 for Revel visual technique and 3.8 and 5.6 for US technique. Means and differences were expressed in degrees.

subjects (Heikkila and Astrom, 1996; De Hertogh et al., 2000; Rix and Bagust, 2001) and from 4.2 to 6.11 for neck-pain populations (Revel et al., 1991; De Hertogh et al., 2000; Rix and Bagust, 2001). Previous studies of the US technique evaluated only healthy subjects, and the method of calculating the head repositioning error differed from ours in that it was assessed only in the horizontal plane and did not take into account associated motions in other planes (Strimpakos et al., 2006; Lee et al., 2006; Demaille-Wlodyka et al., 2007). In the present study, we transformed the 3D independent data into a 2D global result to compare the AEs obtained by both techniques. In agreement with the previous results, our results showed a significant difference in head repositioning AE between healthy subjects and those with neck pain (Revel et al., 1991; Heikkila and Astrom, 1996). One study found no difference in head repositioning AE between similar groups but included only 11 subjects in each group and gave no instruction about the speed of the movement (Rix and Bagust, 2001). Head positioning ability did not differ significantly by sex, age, pain intensity or sports activity in the healthy group. In agreement with our results, recent results showed in a large sample of healthy volunteers that

275

neck proprioception was unaffected by age (DemailleWlodyka et al., 2007). In the present study, AE was weakly related to age in the neck-pain group for the Revel visual technique. Because cervical osteoarthritis is more frequent with age, distinguishing between the effects of age and neck pain is difficult. Further research into the relation between age and head repositioning ability in a neck-pain population is required. In agreement with previous results, our finding of no correlation between repositioning AE and pain intensity did not reflect the possible influence of nociceptive inputs on the repositioning disability; however, in our sample, the mean intensity of pain at rest and during the tests was low (Revel et al., 1991). A difference in participation in sports between the control group and the patient group could account for the difference in repositioning ability. No one in the neck-pain group played sports, and in the healthy group, the AE measured by either technique was not related to sports activity. However, the sample was small and the range of sports was wide. Therefore, the effect of sports on neck proprioception remains unknown. We could have used the same helmet (the US motion analysis system) for the Revel visual and US techniques and thus obtain only simultaneous measurements. We chose to use different headgears for each technique because we wanted to reproduce exactly the initial test (including the specific helmet) described by Revel et al. (1991). Previous studies assessed testeretest reliability of the head repositioning test in healthy people. Excellent reliability was reported in a small sample of young healthy students with the Revel visual technique, and Bland and Altman plotting revealed satisfactory agreement between two successive measurements of head repositioning AE (Pinsault et al., 2006). With a US technique, testeretest reliability ranged from moderate to high (Lee et al., 2006). A different experimental procedure showed the intra- and inter-examiner studies with poor reliability (Strimpakos et al., 2006). In the present study, the testeretest reliability was assessed in the neckpain group only because it had not yet been studied in a symptomatic group. Moreover, study of the evolution of neck pain and of the effects of treatments requires a reliability assessment of the head repositioning test in the symptomatic population. In neck-pain subjects, the ICC values were moderate with visual, US and US-visual techniques. However, the differences in AE for head repositioning between techniques and between the two sets of experiments still showed a narrow margin considering that we were studying human head repositioning performance. That raises the question of the relevance of the ICC test to assess the reliability of these kinds of measurements. Bland and Altman plotting related to Revel visual and US techniques revealed satisfactory agreement between

276

A. Roren et al. / Manual Therapy 14 (2009) 270e277

the two successive measurements. The variability of the head repositioning ability was larger for subjects with higher repositioning AE. However, this trend was close to the significance threshold so it could be neglected. When head repositioning was assessed with Revel visual and US techniques in random order in succession, the correlation of AE was poor for healthy and neckpain groups but excellent with simultaneous measurement from US-visual and US techniques. Visual reading and US assessment could thus be considered as two different ways to measure the same head repositioning error. The results of correlation of Revel visual, USvisual and US techniques and reliability of Revel visual and US techniques in head repositioning error raise the question of the durability of head repositioning ability in a neck-pain population, which had been little studied until now (Revel et al., 1994). The difference in correlation between successive and simultaneous measurements and the moderate reliability also suggest that some of the items of the experimental procedure could increase the intra-individual variability. The weight and the shape of the two helmets were slightly different. The vision-occluding devices were not identical as well. The ultrasound source emitted a buzz, whereas the Revel visual technique was silent. With simultaneous measurements only (US-visual and US techniques), one could be sure that the experimental conditions remained strictly unchanged. Moreover, the duration of the experiment (one set of 10 trials with one technique immediately followed by one set of 10 trials with the other technique) and the numerous handlings (change of helmets and vision-occluding devices) could interfere with subjects’ head repositioning ability and thus alter correlation and reliability. The human factors of variations and the experimental conditions of carrying out the tests tended also to reduce the reliability between trials in the same experiment set and supported the need for several trials. We considered a comparison of Revel visual and US successive measurements according to a pathological threshold identified by both techniques to be clinically more relevant. In our study, the threshold value of inaccurate repositioning was 4.5 with both techniques and the intra-rater agreement between the two techniques measured separately was substantial. This same threshold was identified previously (Revel et al., 1991). The visual measurements of the head repositioning ability confirmed results of previous studies. Among different tests evaluating sense of head and neck position, only the visual head repositioning test was successful in differentiating healthy subjects from those with neck pain (Kristjansson et al., 2003). The Revel visual technique is easy to handle, rapid and economical, so it fits well with clinical practice. However, data are not recorded and cannot be checked later. A detailed analysis of the repositioning procedure is thus impossible;

however, an alternate repositioning strategy, such as overshooting (i.e., overestimation of the distance for repositioning), has been described (Revel et al., 1991). Because our study aimed to compare the Revel visual and US measurement techniques and because the information provided by the Revel visual technique was incomplete, we did not detail how the head was repositioned. However, we noticed differences in repositioning strategy between the two groups that deserve further study. The error visually measured on the target is in two dimensions and results from the projection of 3D coordinates of the head posture on a vertical plane (the target), so a hypothetical projection error cannot be ruled out. On the whole, the Revel visual technique is useful in daily clinical practice for rehabilitation and to evaluate changes. The 3D US technique is more accurate than the Revel visual technique (Dvir and Prushansky, 2000) and excludes any projection error (Bullitt et al., 1997). It also provides complementary data such as 3D measurement of the range of motion and speed of head movements and allows for a delayed analysis of recording. However, the US technique is expensive, time-consuming and not easily adapted to the assessment of head repositioning ability. The head position is expressed in each of the three dimensions separately, and calculations are needed to assess the global head repositioning error. A systematic calibration of the position was necessary before each trial and took several seconds, during which the examiner maintained the head in the reference position with one hand while managing the data-processing tasks with the other one. Thus, the examiner could provoke some perturbations in the subject’s proprioceptive sensitivity. With modification of the software dedicated to the head repositioning test, this disadvantage could be avoided. The 3D US technique is probably more adapted to research in providing information about the individual procedure in repositioning the head.

5. Conclusion This study confirmed the clinical relevance of the head repositioning test, whatever the technique used, Revel visual or US. Head repositioning global error was similar, whatever be the measurement technique. The head repositioning accuracy was significantly higher in the healthy group than in the neck-pain group, and the same pathological threshold was identified with both techniques. The two techniques are complementary: the Revel visual technique provides an efficient quantitative assessment of head repositioning ability but gives only little information on repositioning strategy, whereas the US technique does not improve the quantitative assessment of head repositioning ability but might well be used to study pathological repositioning strategy.

A. Roren et al. / Manual Therapy 14 (2009) 270e277

Acknowledgements The authors thank the patients and the hospital ancillary workers of the department of physical and rehabilitation medicine (Hoˆpital Cochin, AP-HP) for their support. The mathematical consultation from Mr. Michel Dao (Ph.D.) was highly appreciated.

References Andre-Deshays C, Berthoz A, Revel M. Eye-head coupling in humans. I. Simultaneous recording of isolated motor units in dorsal neck muscles and horizontal eye movements. Experimental Brain Research 1988;69:399e406. Bland JM, Altman DG. Statistical method for assessing agreement between two methods of clinical measurement. The Lancet 1986; 1:307e10. Bullitt E, Liu A, Pizer SM. Three-dimensional reconstruction of curves from pairs of projection views in the presence of error. II. Analysis of error. Medical Physics 1997;24:1679e87. Demaille-Wlodyka S, Chiquet Ch, Lavaste JF, Skalli W, Revel M, Poiraudeau S. Cervical range of motion and cephalic kinesthesis: ultrasonographic analysis by age and sex. Spine 2007;32:254e61. Dvir Z, Prushansky T. Reproducibility and instrument validity of a new ultrasonography-based system for measuring cervical spine kinematics. Clinical Biomechanics 2000;15:658e64. De Hertogh W, Vaes P, Duquet W, Oostendporp R. Kinesthetic function in patients with cervicogenic headache. In: Singer KP, editor. Proceedings of the seventh scientific conference of the international federation of orthopaedic manipulative therapists. Perth, Australia: International Federation of Orthopaedic Manipulative Therapists; 2000. p. 441e5. Heikkila HV, Astrom P-G. Cervicocephalic kinesthetic sensibility in patients with whiplash injury. Scandinavian Journal of Rehabilitation Medicine 1996;28:133e8. Supplement. Humphreys K, Irgens P, Rix G. The effect of rehabilitive exercise programme on cervicocephalic kinaesthesia and reported levels of pain in chronic neck pain subjects. Journal of Bone and Surgery 2002;84-B(suppl. 2). Karnath HO, Sievering D, Fetter M. The interactive contribution of neck muscle proprioception and vestibular stimulation to

277

subjective ‘‘straight ahead’’ orientation in man. Experimental Brain Research 1994;101:140e6. Kristjansson E, Dall’Alba P, Jull G. Cervicocephalic kinaesthesia: reliability of a new test approach. Physiotherapy Research International 2001;6:224e35. Kristjansson E, Dall’Alba P, Jull G. A study of five cervicocephalic repositioning tests in three different subject groups. Clinical Rehabilitation 2003;17:768e74. Lakie M, Loram ID. Manually controlled human balancing using visual, vestibular and proprioceptive senses involve a common, low frequency neural process. The Journal of Physiology 2006;577:403e16. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159e74. Lee HY, Teng CC, Chai HM, Wang SF. Testeretest reliability of cervicocephalic kinesthetic sensibility in three cardinal planes. Manual Therapy 2006;11:61e8. McLain RF. Mechanoreceptor endings in human cervical facet joints. Spine 1994;19:495e501. Pinsault N, Vaillant J, Virone G, Caillat-Mousse Lachens L, Vuillerme N. Test de repositionnement ce´phalique: e´tude de la stabilite´ de la performance. Annales de Re´adaptation et de Me´decine Physique 2006;49:647e51. Proske U. What is the role of muscle receptors in proprioception? Muscle and Nerve 2005;31:780e7. Revel M, Andre-Deshays C, Minguet M. Cervicocephalic kinesthetic sensibility in patients with cervical pain. Archives of Physical Medicine and Rehabilitation 1991;72:288e91. Revel M, Minguet M, Gregoy P, Vaillant J, Manuel JL. Changes in cervicocephalic kinaestesia after a proprioceptive rehabilitation program in patients with neck pain: a randomized controlled study. Archives of Physical Medicine and Rehabilitation 1994;75:895e9. Rix GD, Bagust J. Cervicocephalic kinesthetic sensibility in patients with chronic, nontraumatic cervical spine pain. Archives of Physical Medicine and Rehabilitation 2001;82:911e9. Savignat E, Roren A. Evaluation de la proprioception chez le patient cervicalgique: utilisation du test de repositionement cervicoce´phalique (TRC). Kine´sithe´rapie La Revue 2007;63:21e4. Shrout PE, Fleiss JL. Intraclass coefficients: uses in assessing rater reliability. Psychological Bulletin 1979;86:420e8. Strimpakos N, Sakellari V, Gioftsos G, Kapreli E, Oldham J. Cervical joint position sense: an intra- and inter-examiner reliability study. Gait and Posture 2006;23:22e31. Treleaven J, Jull G, Lowchoy N. Standing balance in persistent whiplash: a comparison between subjects with and without dizziness. Journal of Rehabilitation Medicine 2005;37:224e9.

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 278e282 www.elsevier.com/math

Original Article

Influence of elbow flexion angle on mobilization of the proximal radio-ulnar joint: A motion analysis using cadaver specimens Sadanori Ohshiro a, Egi Hidaka a, Shigenori Miyamoto b, Mitsuhiro Aoki b,*, Toshihiko Yamashita c, Haruyuki Tatsumi d a

Graduate School of Health Sciences, Sapporo Medical University, School of Health Sciences, South-3, West-17, Chuo-ku, Sapporo 060-8556, Japan Department of Physical Therapy, Sapporo Medical University, School of Health Sciences, South-3, West-17, Chuo-ku, Sapporo 060-8556, Japan c Department of Orthopaedic Surgery, Sapporo Medical University, School of Medicine, South-3, West-17, Chuo-ku, Sapporo 060-8556, Japan d First Department of Anatomy, Sapporo Medical University, School of Medicine, South-3, West-17, Chuo-ku, Sapporo 060-8556, Japan

b

Received 30 May 2007; received in revised form 20 February 2008; accepted 28 February 2008

Abstract The purpose of this study was to determine the most effective elbow joint flexion angle for mobilization of the proximal radioulnar joint. Five fresh-frozen cadaveric elbows were used to measure displacement of the radial head in the antero-medial and postero-lateral directions by traction force of 2 kgf and 4 kgf, respectively. Simulation of the gliding of the proximal radio-ulnar joint was performed at four elbow flexion angles (0 , 30 , 60 , 90 ). Data obtained from those flexion angles were compared using one-way repeated measures analysis of variance. Radial head displacement at 60 and 90 during antero-medial gliding were significantly greater than those at 0 and 30 ( p < 0.05) There were no significant differences in radial head displacement among four elbow flexion angles during postero-lateral gliding at 2 kgf and 4 kgf. Our findings suggest that proximal radio-ulnar joint mobilization in the antero-medial direction can be performed effectively at 60 and 90 elbow flexion. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Proximal radio-ulnar joint; Joint mobilization; Motion analysis; Cadaveric study

1. Introduction Anatomically, the elbow is a uni-axial hinge joint formed of two articulations, i.e., the humero-ulnar and humero-radial joints that allow for movement of flexion and extension, and is referred to as the cubital articulation. The cubital articulation and proximal radio-ulnar joint are collectively termed ‘the joint complex’. Pronation and supination are motions of the proximal radio-ulnar joint that play an important role on * Corresponding author. Tel.: þ81 11 611 2111x2874; fax: þ81 11 611 2150. E-mail address: [email protected] (M. Aoki). 1356-689X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.02.012

the motion of the wrist and fingers (Kapandji, 1982). Once this motion sequence of the upper extremity is disrupted, activities of daily living, such as having meals and perineal care, are severely restricted. One of the major causes of limitation in forearm rotation is post-traumatic contracture after fracture or fracture dislocation of the elbow. Even after appropriate treatment of these injuries by experienced orthopaedic surgeons, contracture of the elbow is occasionally observed. In such cases, physiotherapy for the elbow becomes extremely important. Restricted motion of the proximal radio-ulnar joint is considered to be caused by a thickening of the capsule and a reduction in the flexibility of the collateral ligaments (Morrey and An, 1985; Herting and Kessler, 1996).

S. Ohshiro et al. / Manual Therapy 14 (2009) 278e282

279

To solve this problem, joint mobilization of the proximal radio-ulnar joint has been used as an effective manual procedure (Miyamoto, 1985; Butler, 1994; Kisner and Colby, 2002). This procedure consists of gliding the radial head in an antero-medial direction for pronation contracture of the forearm, and in a postero-lateral direction for supination contracture (Maitland, 1991; Herting and Kessler, 1996; Edmond, 2006). However, there has been no report concerning the evidence of the effective elbow flexion angle for proximal radio-ulnar joint mobilization. In this experiment, fresh-frozen cadaveric elbows were used to measure the displacement of the radial head during antero-medial and postero-lateral gliding at four elbow flexion angles (0 , 30 , 60 , 90 ). The purpose of this study was to determine the most effective elbow joint flexion angle for mobilization of the proximal radio-ulnar joint.

2. Materials and methods 2.1. Preparation of the specimens On February 27, 2006, the Ethical Committee of our university approved all types of surgical training, and biomechanical studies using frozen-thawed cadavers donated to the Department of Anatomy. This experiment used five frozen-thawed upper extremities that were obtained from five fresh cadavers aged 60e87 years at death (average, 78.2 years) after the acquisition of informed consent prior to death. Two right elbows and three left elbows were taken from two male and three female specimens. The elbows were kept in a freezer at e20  C after disarticulation at the glenohumeral joint. Specimens with a limited range of elbow and forearm motion were excluded from this study. Thawing of the specimens at room temperature began 12 h prior to the experiment. All soft tissue was removed from the elbow and forearm, sparing the joint capsule, ligaments, and interosseous membrane. Thawing was confirmed through preconditioned movement of the glenohumeral and elbow joint in all directions. 2.2. Set-up of specimens for mechanical analysis In this experiment, a wooden jig, consisting of a table and a square column, was used. The humerus was fixed to the wooden jig so that the long axis of the humerus was perpendicular to the ground. The elbow and forearm were allowed to move freely (Fig. 1). The flexion angle of the elbow was set manually at 0 ,  30 , 60 , and 90 with neutral forearm rotation and neutral wrist flexion. These angles were set using goniometric measurement by an examiner. To simulate

Fig. 1. Set-up for the mechanical analysis of the radial head mobilization. The 3Space Magnetic Track system recorded radial head displacement to an accuracy of 0.2 mm RMS.

mobilization of the proximal radio-ulnar joint, a traction force of 2 kgf and 4 kgf was applied to the radial head perpendicular to the long axis of the radius shaft through tapes around the radial neck. Prior to this experiment, thumb pressure simulating the clinical performance of proximal radio-ulnar joint mobilization was measured by physical therapists using a JAMA pinch gauge (Hydraulic Pinch Gauge, JAMAR, Mississauga, Canada). Since the majority of pressure values were between 2 kgf and 4 kgf, traction forces of 2 kgf and 4 kgf were adopted in this biomechanical study. The traction was set in the antero-medial and posterolateral directions as adopted by gliding techniques in joint mobilization manuals (Maitland, 1991; Herting and Kessler, 1996; Edmond, 2006). A six-degree-of-freedom electromagnetic tracking device (3SPACE FASTRAK; Polhemus, Colchester, Vermont) was used for the precise measurement and monitoring of radial head movement. This device enabled the measurement of the three-dimensional position and orientation of the receivers relative to the absolute coordinates generated by the magnetic transmitter. Within a 250-mm range from the magnetic source, the root mean square (RMS) error in positional accuracy was 0.2 mm (Kitaoka et al., 1997). To determine the center of the capitulum of the humerus, a hole of 5 mm in diameter was created from the anterior aspect of the capitulum. To determine the center of the radial head, a hole of 5 mm in diameter was created from the anterior neck of the radius. Thus, the amount of radial head displacement against the capitulum humeri

S. Ohshiro et al. / Manual Therapy 14 (2009) 278e282

during radial head gliding was measured twice in each direction at 0 , 30 , 60 and 90 of elbow flexion (Fig. 2). Throughout the experiment, which was performed at room temperature (22  C), the specimens were kept moist by spraying with saline solution every 5e10 min. 2.3. Statistical analysis One-way repeated measures analysis of variance (ANOVA) was used to determine difference in the radial head movement among the four sets of conditions; i.e., 0 , 30 , 60 and 90 of elbow flexion. Bonferroni’s multiple comparisons test was used as a post hoc test. Statistical significance was set at a ¼ 0.05.

Radial head displacement (mm)

280

Antero-medial traction of the radial head (2 kgf) 5

*

*

4

*

*

3 2 1 0

0°

30°

60°

90°

Elbow joint flexion angle Fig. 3. Displacement of the radial head during traction in an anteromedial direction by 2 kgf of traction force. *p < 0.05.

4. Discussion 3. Results 3.1. Antero-medial gliding During the application of both 2 kgf and 4 kgf of traction force, radial head displacement at 60 (2.1  0.7 mm and 3.2  1.1 mm) and 90 (2.0  0.3 mm and 3.0  1.0 mm) of elbow flexion were significantly ( p < 0.05) greater than those at 0 (0.7  0.2 mm and 1.1  0.3 mm) and 30 (1.2  0.2 mm and 1.5  0.2 mm), (Fig. 3 and Fig. 4, respectively). 3.2. Postero-lateral gliding No significant differences in radial head displacement between the four flexion angles were observed during the application of traction force at either 2 kgf (1.2  0.7 mm for 0 , 1.3  0.4 mm for 30 , 1.9  0.5 mm for 60 , and 1.5  0.8 mm for 90 of elbow flexion) or 4 kgf (1.5  0.9 mm for 0 , 2.0  0.7 mm for 30 , 2.0  0.4 mm for 60 , and 1.5  0.6 mm for 90 of elbow flexion) (Fig. 5 and Fig. 6, respectively).

Joint mobilization consists of low velocity reciprocating passive motion to produce intra-capsular joint movement with varying amplitudes, and is indicated for use with joints in either a loose-packed or resting position (Kaltenborn, 1993; Neuman, 2002; Takei, 2005). Favorable results have been reported for joint mobilization as a treatment for impaired joint function in the extremities and spine (Miyamoto, 1985; Fabio, 1992). In the field of physical therapy, therapists have used manual procedures for the treatment of rotational contracture of the forearm. These procedures consist of gliding the radial head in an antero-medial direction for pronation contracture of the forearm, and in a postero-lateral direction for supination contracture (Maitland, 1991; Herting and Kessler, 1996; Edmond, 2006). However, there have been no reports concerning the effective elbow flexion angle for proximal radioulnar joint mobilization. In the present study, we determined the most effective elbow flexion angle for mobilization of the proximal radio-ulnar joint by direct measurement in cadaver models. During gliding of the radial head in an antero-medial direction to simulate joint mobilization for pronation contracture of the forearm, displacement of the radial

Radial head displacement

Antero-medial traction of the radial head (4 kgf)

Fig. 2. Measurement of the radial head displacement by 3Space Traction in the antero-medial (a) and postero-lateral (b) directions is shown. Radial head displacement was measured with an electromagnetic tracking device.

5

* *

*

4

* 3 2 1 0

0°

30°

60°

90°

Elbow joint flexion angle Fig. 4. Displacement of the radial head during traction in an anteromedial direction by 4 kgf of traction force. *p < 0.05.

281

Postero-lateral traction of the radial head (2 kgf) 5 4 3 2 1 0

0°

30°

60°

90°

Radial head displacement (mm)

Radial head displacement (mm)

S. Ohshiro et al. / Manual Therapy 14 (2009) 278e282

Postero-lateral traction of the radial head (4 kgf) 5 4 3 2 1 0

0°

Elbow joint flexion angle Fig. 5. Displacement of the radial head during traction in a posterolateral direction by 2 kgf of traction force.

head significantly increased at 60 and 90 of elbow flexion. The reason for this result is considered to be due to the direction of the fibers comprising the lateral collateral ligament of the elbow. The lateral collateral ligament is composed of the radial collateral ligament and the annular ligament, each of which has its own biomechanical properties. The radial collateral ligament blends with the annular ligament of the proximal radio-ulnar joint and the outlying band of the annular ligament blends with the supinator crest of the ulna. The posterior fibers of the radial collateral ligament are tight during elbow flexion, and the anterior fibers of the ligament are tight during elbow extension which simultaneously effects the tension of the annular ligament (Morrey and An, 1985; Olsen et al., 1998; Oatis, 2004). Sixty and 90 degrees of elbow flexion, at which the amount of antero-medial displacement of the radial head was the largest in our study, are intermediate positions of the elbow in which the anterior and posterior fibers of the radial collateral ligament are loose. Reduced tension in both sets of fibers of the ligament may also relax the annular ligament. On the other hand, during gliding of the radial head in a postero-lateral direction to simulate joint mobilization for supination contracture of the forearm, there were no significant differences in radial head displacement among four elbow flexion angles. The reason for this result is considered to be the tension produced by the middle fibers of the radial collateral ligament, as these fibers are thick and remain tight across the total range of the elbow flexion (Olsen et al., 1996, 1998; Lockard, 2006). Increased off-setting of the joint facet on the ulna during postero-lateral gliding of the radial head may contribute to this tightness. There are limitations in this study. First, since the elbow specimens were harvested from aged cadavers, the mechanical properties of the ligaments might be different from those of specimens from younger people in whom elbow contracture tends to occur. Although specimens with a limited range of elbow and forearm motion were excluded from this study, the elbows of

30°

60°

90°

Elbow joint flexion angle Fig. 6. Displacement of the radial head during traction in a posterolateral direction by 4 kgf of traction force.

aged specimens have varying degrees of osteo-arthritic changes. This pathology may affect the displacement of the radial head. Since the anatomical structure of the collateral ligaments and the direction of their fibers are similar between the younger and the aged specimens, the results of this study are considered to be applicable to younger patients (Muraki et al., 2007). Second, in the present study, we measured the displacement of the radial head during the mobilization procedure, but did not measure the strain on the collateral ligaments. To analyze the behavior of the lateral collateral ligaments in the elbow during joint mobilization in greater detail, the strain on each collateral ligament should be measured in future studies. 5. Conclusions Five fresh-frozen cadaveric elbows were used to measure the displacement of the radial head during antero-medial and postero-lateral gliding at four elbow flexion angles. Our findings suggested that proximal radio-ulnar joint mobilization can be effectively performed by gliding in the antero-medial direction at 60 and at 90 of elbow flexion. Elbow flexion angle had no effect on the postero-lateral mobilization of the radial head. Acknowledgment The authors would like to thank Gen Murakami, MD, PhD for providing us the fresh-frozen cadaver specimens. The authors also thank Eiichi Uchiyama, MD, PhD; Daisuke Suzuki, PhD; Takayuki Muraki, PhD; Hiroshi Takasaki, RPT; Hitoshi Miyamoto, RPT, and Misaki Fujii, RPT for their technical assistance. References Butler DS. Physical therapy of the cervical and thoracic spine. 2nd ed. New York: Churchill Livingstone; 1994. pp. 217e244.

282

S. Ohshiro et al. / Manual Therapy 14 (2009) 278e282

Edmond SL. Joint mobilization/manipulation. 2nd ed. London: Mosby Elsevier; 2006. pp. 86e87. Fabio RP. Efficacy of manual therapy. Physical Therapy 1992; 72(12):853e63. Herting D, Kessler RM. Management of common musculo-skeletal disorders. 3rd ed. Philadelphia: JB Lippincott; 1996. pp. 238e239. Kaltenborn FM. The spine. 2nd ed. Minneapolis: Olaf Norlis Bokhandel; 1993. pp. 1e87. Kapandji IA. The physiology of the joints: annotated diagrams of the mechanics of the human joints. 5th ed. New York: Churchill Livingstone; 1982. Kisner C, Colby LA. Therapeutic exercise, foundations and techniques. 4th ed. Philadelphia: FA Davis; 2002. pp. 216e256. Kitaoka H, Luo ZP, An KN. Analysis of longitudinal arch supports in stabilizing the arch of the foot. Clinical Orthopaedics and Related Research 1997;341:250e6. Lockard M. Clinical biomechanics of the elbow. Journal of Hand Therapy 2006;19(2):72e80. Maitland GD. Peripheral manipulation. 3rd ed. Oxford: ButterworthHeinemann; 1991. pp. 183e189.

Miyamoto S. Joint mobilization of a shoulder joint. Journal of Physical Therapy 1985;13(2):187e90 (in Japanese). Morrey BF, An KN. Functional anatomy of the ligaments of the elbow. Clinical Orthopaedics and Related Research 1985;201:84e90. Muraki T, Aoki M, Uchiyama E, et al. Strain on the repaired supraspinatus tendon during manual traction and translation glide mobilization on the glenohumeral joint. Manual Therapy 2007; 12(2):231e9. Neuman DA. Kinesiology of the muscloskeletal system, Missouri: Mosby; 2002. pp. 3e24. Oatis CA. Kinesiology, the mechanics and pathomechanics of human movement. Philadelphia: JB Lippincott; 2004. pp. 186e229. Olsen BS, Nielsen KK, Vaesel MT. Posterolateral elbow joint instability. The basic kinematics. Journal of Shoulder Elbow Surgery 1998;7(2):19e29. Olsen BS, Vaesel MT, Jens O. Lateral collateral ligament of the elbow joint, anatomy and kinematics. Journal of Shoulder Elbow Surgery 1996;5(1):103e12. Takei H. Joint mobilization for bone and joint disease. Physical Therapy Science 2005;20(3):219e25 (in Japanese).

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 283e291 www.elsevier.com/math

Original Article

The testeretest reliability and concurrent validity of the Subjective Complaints Questionnaire for low back pain Jon Joseph Ford a,*, Ian Story b, Joan McMeeken a a

School of Physiotherapy, The University of Melbourne, Parkville, Victoria 3010, Australia b Faculty of Health and Behavioural Science, Deakin University, Australia

Received 25 August 2005; received in revised form 6 February 2008; accepted 27 February 2008

Abstract Physiotherapists commonly record detailed patient information regarding subjective complaints for low back pain (LBP), particularly to assist in the process of classifying patients into specific subgroups. A self-administered Subjective Complaints Questionnaire for LBP (SCQ-LBP) measuring such information was developed for the purposes of future clinical research, particularly in the area of LBP classification. The development comprised literature review, feedback from experienced physiotherapists and pilot questionnaire testing in a patient population. Testeretest reliability of the questionnaire in a self administered format as well as concurrent validity against a suitable reference standard was evaluated. The agreement between the self administered questionnaire compared to when administered by a physiotherapist was also tested as the latter method is the most common form of retrieving subjective complaints in clinical practice. Thirty participants with LBP were recruited and at least moderate testeretest reliability was demonstrated in 56 of the 57 self administered questionnaire items. Preliminary evidence was found supporting the concurrent validity of selected items. At least moderate agreement was demonstrated in 51 of the 57 items when comparing between the self administered and physiotherapist administered conditions. The questionnaire is a useful tool for collecting subjective complaints information, particularly for clinical research on the classification of LBP, however, further research regarding validity is required. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Back pain; Classification; Questionnaire; Validity

1. Introduction Subjective complaints, reported by the sufferer in spoken or written form, are commonly documented as part of the standard assessment of low back pain (LBP). In clinical practice physiotherapists use a detailed subjective complaints interview to retrieve large amounts of * Corresponding author. Tel.: þ61 03 9607 3051; fax: þ61 03 9670 3080. E-mail address: [email protected] (J.J. Ford). 1356-689X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.02.010

information regarding the nature of symptoms, aggravating/easing factors, 24 h symptom behaviour, history and special questions (McKenzie et al., 2003; Maitland, 2005). This subjective complaints information can then be used to classify patients into subgroups that receive specific treatment (Delitto et al., 1995; McKenzie et al., 2003; Maitland, 2005). A common clinical example of using subjective complaints in the classification of LBP is the identification of an inflammatory cause of symptoms using criteria such as constancy of pain and a perception of stiffness in the injured area lasting at least 1 h after

284

J.J. Ford et al. / Manual Therapy 14 (2009) 283e291

waking. Such patients may respond more effectively to anti-inflammatory modalities rather than treatment directed at a mechanical cause of symptoms (McKenzie et al., 2003; Maitland, 2005). Improved classification of LBP has been described as a high research priority by an esteemed international working group (Borkan and Cherkin, 1996). However, a number of review papers show that published classification systems have yet to demonstrate adequate reliability or validity and that methodological problems are common (Riddle, 1998; Petersen, 1999; Ford et al., 2007). One of the issues in the literature evaluating classification systems for LBP is the use of unreliable or poorly validated measurement tools for collecting data relevant to the classification process (Riddle, 1998; Petersen, 1999; Ford et al., 2007). There is relatively little research into reliable and valid methods of measuring subjective complaints. Only one study has reported on a comprehensive tool for measuring subjective complaints in individuals with LBP. The questionnaire developed included items such as the nature of symptoms, aggravating factors and history and variable testeretest reliability was found for the different questionnaire components (Walsh and Coggon, 1991). A 1-year retest period was used to minimise memory effects, however, a potential adverse consequence of this approach on the results may have been a spontaneous change in the nature and/or severity of the LBP. In addition the details of questionnaire wording and scoring were not described. The Multi-perspective Multi-dimensional Pain Assessment Protocol has been developed as a self administered questionnaire for chronic pain. It contains items on patient details, nature of symptoms, aggravating and easing factors, 24 h behaviour and history (Rucker et al., 1996). The questionnaire was developed with a view to maximising content validity and on evaluation, acceptable reliability and validity were demonstrated. However, the items in the questionnaire were not sufficiently comprehensive or specific to be used for the assessment of LBP patients when compared to accepted methods currently in widespread clinical use by physiotherapists (McKenzie et al., 2003; Maitland, 2005). A number of recent papers have evaluated the reliability of a variety of classification systems, however, typically subjective complaints are either poorly defined or insufficiently comprehensive to reflect the clinical practice of physiotherapists (Heiss et al., 2004; Petersen et al., 2004; Bertilson et al., 2006; Dankaerts et al., 2006; Fritz et al., 2006). One paper has been published investigating the reliability of the McKenzie classification system using a specific assessment form that included subjective complaints information (Clare et al., 2005). However, the ability of the form itself to reliably retrieve information from patients on retesting was not evaluated.

On the basis of this insufficient literature, there are limited options for clinical researchers wishing to use a reliable and valid questionnaire that comprehensively assesses subjective complaints for LBP. The aims of this study were therefore to develop a questionnaire for the self administered measurement of subjective complaints in LBP, and to evaluate its testeretest reliability and concurrent validity. Self administered questionnaires are an efficient method of data collection in clinical research. The purpose of this study was not to investigate the value of a particular classification system for LBP, but rather to establish a reliable and valid tool for the measurement of subjective complaints for use in future clinical research, particularly in the area of classification.

2. Method 2.1. Questionnaire development The critical components of our Subjective Complaints Questionnaire for low back pain (SCQ-LBP) comprised nature of symptoms, aggravating and easing factors, 24 h behaviour, and history. These components were predetermined based on clinical approaches commonly used by physiotherapists (McKenzie et al., 2003; Maitland, 2005). The specific questionnaire items were developed in four stages: literature review, operational definition of questionnaire items, peer review of the questionnaire, and pilot testing on individuals with LBP. A literature review was performed using the Current Contents database from years 1970 to 2003 inclusive. This database is a multidisciplinary current awareness web resource providing access to complete bibliographic information from over 8000 of the world’s leading scholarly journals. It includes databases with extensive coverage in the social and behavioural sciences as well as clinical medicine. Key words for the search were ‘‘back pain and validity’’ combined in separate searches with ‘‘subjective’’, ‘‘history’’ and ‘‘questionnaire’’. Full text copies of all identified papers were then retrieved. Further papers were unearthed by cross-referencing from the retrieved papers. On review of this literature, pain drawings as a measure of the nature of symptoms were found to be reliable (Margolis et al., 1986; Uden et al., 1988; Weiner et al., 1998) and some evidence of validity as a tool for classification was demonstrated (Uden and Landin, 1987; Mann et al., 1993; Ohnmeiss et al., 1995) although this latter conclusion has been disputed (Rankine et al., 1998). Preliminary evidence was found for the reliability and validity of a self administered questionnaire measuring aggravating factors in LBP (Roach et al., 1994, 1995, 1997). Studies evaluating the reliability or validity of

J.J. Ford et al. / Manual Therapy 14 (2009) 283e291

24 h symptom behaviour had not been published for LBP. The two studies evaluating the reliability and validity of questionnaires measuring the history of LBP showed poor results (Biering-Sorensen and Hilden, 1984; Burdorf and Laan, 1991) potentially due to the influence of recall issues. Given the limited validation of many of the predefined critical subjective complaints components, items for our SCQ-LBP were chosen from the above literature if reliability and/or preliminary validity had been demonstrated (Margolis et al., 1986; Uden et al., 1988; Roach et al., 1994, 1995, 1997; Weiner et al., 1998). Where no such evidence existed, questionnaire items were selected from publications where subjective complaints were well described, either from the above literature or the descriptive literature on LBP (Boissonnault and Di Fabio, 1996; McKenzie et al., 2003; Maitland, 2005) and chronic injury (Rucker, 1996). Following the literature review an initial draft version of the SCQ-LBP was developed and distributed to two senior coordinators of the musculoskeletal physiotherapy postgraduate program from Universities in Victoria, Australia, and three physiotherapists with at least 5 years clinical experience working predominantly with chronic LBP. The feedback received was primarily in relation to scoring issues and consistency of wording between items. The second draft of the SCQ-LBP was self administered to 13 individuals with chronic LBP. Further modification to the SCQ-LBP was made to items where consistent feedback was given resulting in the final version of the SCQ-LBP (see Manual Therapy website).

2.2. Reliability and concurrent validity of the SCQ-LBP Participants were recruited from patients referred to physiotherapy private practices in Melbourne, Australia. Inclusion criteria were a primary problem of LBP, aged over 18 years, no signs or symptoms of serious nonmusculoskeletal pathology, and literacy in spoken and written English. In order to evaluate testeretest reliability in a self administered condition (SA), the SCQ-LBP was completed by participants twice within a period of 24e48 h. Under this condition, participants were given the questionnaire to complete without assistance. Reliability was tested only for the SA as our aim was to develop a reliable and valid measure that for pragmatic reasons in relation to clinical research was self administered. In order to evaluate concurrent and face validity participants performed specific items from the SCQ-LBP whilst being observed by physiotherapists (PO) and in

285

also completed the full questionnaire in a physiotherapist administered condition (PA). Concurrent validity is a research design used to evaluate the ability of a newly developed measure to predict the results of an existing measure that represents an irrefutable truth or criterion standard (Portney and Watkins, 1993; Anastasi and Urbina, 1997). Often the newer measure is more convenient, less invasive or less costly than the criterion standard. However, in practice such absolute measures are rarely available and a more appropriate comparison is often with a reference standard that is the best available measure of truth (Portney and Watkins, 1993). In the current study the concurrent validity of the self administered SCQ-LBP (as a more convenient and cost effective measure for clinical research) was evaluated for its ability to predict the results of the physiotherapist observed condition (PO) (the reference standard). In the PO, the participant was asked to perform specific aggravating factors from the SCQ-LBP that could be replicated in a clinical environment such as sitting tolerance. The physiotherapist observed and questioned the participant regarding pain response according to a predetermined protocol in order to standardise measurement procedures. Time related variables were measured using a stopwatch. For example, time to onset of lumbar symptoms when in erect sitting, and time to the point where erect sitting could no longer be tolerated were measured. These time values were then compared to the response for the same items in the SA. In the PA a physiotherapist read each questionnaire item to the participant and filled in the participant responses. Clarification regarding questionnaire items was provided by the physiotherapist to the participant as required, and in accordance with a standardised protocol. The PA was intended to replicate the current clinical method of retrieving subjective complaints by physiotherapists (Maitland, 2005). Rather than representing an adequate reference standard, the SA was compared to the PA to evaluate whether the role of practitioner clarification was necessary in a clinical research setting. This comparison could therefore be considered an investigation of the face validity of the self administered SCQ-LBP (Portney and Watkins, 1993; Anastasi and Urbina, 1997). Each participant completed four administrations of the SCQ-LBP (initial SA, retest SA, PO and PA) according to one of six predetermined combinations of possible order, sequentially applied to each participant, to eliminate any effect of the order of administration on the results. To minimise any 24 h variation in the participants’ LBP, all data collection occurred at the same time of day. In addition, to evaluate spontaneous change between data collection sessions the Quebec Back Pain Disability Scale (Kopec et al., 1995) was completed at each questionnaire administration including the PO.

286

J.J. Ford et al. / Manual Therapy 14 (2009) 283e291

2.3. Analysis Calculation of agreement was made on comparisons between the following SCQ-LBP conditions: initial SAeretest SA, initial SAePA and initial SAePO. Kappa statistics were used for all data with four or less response categories. For response categories of five or more weighted Kappa was calculated. Moderate agreement was defined as being greater than or equal to 0.4, and substantial agreement greater than or equal to 0.6 (Landis and Koch, 1977). For interval data mean differences and standard deviations (SDs) were calculated for each condition comparison (Bland and Altman, 1986).

3. Results Thirty participants provided informed consent and completed the data collection. There were no withdrawals from the study. The demographic and clinical characteristics of the sample are presented in Table 1. A one-way repeated measures analysis of variance found no significant within participant difference in the Quebec Back Pain Disability Scale scores across the measurement sessions (df ¼ 3.26, F ¼ 0.42, p ¼ 0.66). This indicates that spontaneous change over Table 1 Demographic and clinical characteristics for each group given as the mean (SD) or the number (%). Characteristic

Mean/SD (N ¼ 30)

Age (years)

42 (9.3)

Gender Male Female

Number/ percentage (N ¼ 30)

11 (37%) 19 (63%)

Weight (kg)

79 (15)

Height (cm)

175 (11)

Education Complete primary Complete secondary Complete tertiary

11 (37%) 10 (33%) 9 (30%)

Compensable Yes No

30 (100%) 0 (0%)

Duration of LBP 2e3 months 4e6 months 7e12 months More than 12 months

1 1 5 23

Past history of lumbar surgery Yes No

6 (20%) 24 (80%)

(3%) (3%) (17%) (77%)

time had no significant effect on disability and is therefore unlikely to have influenced the level of agreement observed. Mean differences and SDs were calculated for the pain drawing data and the level of agreement determined by judgement (Bland and Altman, 1986). The area score method (Parker et al., 1995) was used to count the number of grids where symptoms were marked. The maximum number of grids possible to mark was 264. The mean difference between the area scores in the initial SAeretest SA comparison was 0.60 (SD 6.56) and only five of the 30 participants had a difference of over five points. Higher levels of agreement were demonstrated on the initial SAePA comparison. The maximum number of symptoms descriptors was six on the pain drawing and the initial SAeretest SA comparison difference was 0.07 (SD 1.04). Only four of the 30 participants had a difference of over one point. Similar figures on the initial SAe PA comparison was observed. It was judged that sufficient agreement was observed on the initial SAe retest SA and initial SAePA comparisons for both the pain drawing area score and number of symptom descriptors. Substantial agreement was therefore found in 44 out of 57 items on the initial SAeretest SA comparisons. Moderate agreement was found in a further 12 items, with one item having less than moderate agreement. The results of the testeretest reliability of the SCQ-LBP in the SA are summarised in Table 2. Substantial agreement was found in one out of seven items on the initial SAePO comparisons with moderate agreement in a further four items. Two items had less than moderate agreement. The results of this comparison are presented in Table 3. Substantial agreement was found in 37 out of 57 items on comparison between the initial SAePA with moderate agreement in a further 14 items. However, six items had less than moderate agreement. These results are presented in Table 4.

4. Discussion The SCQ-LBP was developed using the methodology of literature review, review from a panel of experienced physiotherapists, and pilot testing on a sample of individuals with LBP. The testeretest reliability of the developed questionnaire was acceptable. Specifically, at least moderate agreement was found on 56 of the total 57 items on testeretest of the SCQ-LBP in the SA. The concurrent validity of the SCQ-LBP was examined by a comparison of the SA to the PO. In the PO each participant was tested in real time for tolerances to performing specific aggravating factors

287

J.J. Ford et al. / Manual Therapy 14 (2009) 283e291 Table 2 Testeretest reliability of the SCQ-LBP (initial SAeretest SA). SCQ-LBP item (number of participants)

Kappa (95% CI)

Pain drawing item Area of symptoms (30) Number of symptom descriptors (30) Presence of leg pain (30) Presence of dermatomal symptoms (30) Presence of distal paraesthesia (30)

0.55 (0.20e0.90) 0.66 (0.31e1.00) 0.57 (0.22e0.92)

Symptom details Presence of pins and needles (29) Area of most severe pain (23) Presence of foot drop (29) Lumbar symptoms deep or superficial (29) Leg symptoms deep or superficial (29) Symptom relationship (29)

0.73 0.78 1.00 0.72 0.73 0.63

Aggravating factors Onset of symptoms while walking (28) Duration of walking limited by symptoms (27) Onset of symptoms while standing (27) Duration of standing limited by symptoms (29) Onset of symptoms while erect sitting (28) Duration of erect sitting limited by symptoms (29) Onset of symptoms while slumped sitting (29) Duration of slumped sitting limited by symptoms (29) Onset of symptoms while lying supine (29) Duration of lying supine limited by symptoms (28) Onset of symptoms while lying prone (29) Duration of lying prone limited by symptoms (27) Symptoms aggravated by getting into car (28) Components of getting into car that aggravate (24) Symptoms aggravated by sitting to standing (30) Components of sit to stand that aggravate symptoms (29) Symptoms aggravated by cough/sneeze (30) Presence of symptoms worse on one side (30) Most comfortable side to lie on (29) Easing factors Whether standing eases symptoms (29) Whether sitting eases symptoms (29) Whether sitting in a couch eases symptoms (29) Whether walking eases symptoms (29) Whether lying supine eases symptoms (29) Whether crook lying eases symptoms (29) Whether lying on the sore side eases symptoms (29) Whether lying on the good side eases symptoms (27) Whether lying prone eases symptoms (29) History of symptoms Whether current symptoms are the first episode (30) Duration of current symptoms (29) Number of episodes of symptoms (28) Time from first episode of symptoms (28) Whether symptoms chronic or recurrent (28) Mechanism of injury for current symptoms (28) Time of onset of current symptoms (29) Behaviour of current symptoms after onset (28) History of manual handling at work (28) History of over the past three days Time of day when symptoms worst (29) Presence of trouble getting to sleep (29) Presence of waking at night (29) Reason for waking at night (28)

Weighted Kappa (95% CI)

Mean difference (SD) 0.60 (5.54) 0.07 (1.05)

(0.38e1.00) (0.47e1.00) (0.63e1.00) (0.43e1.00) (0.48e0.98) (0.38e0.88) 0.83 0.51 0.75 0.51 0.68 0.49 0.86 0.86 0.56 0.51 0.68 0.58 0.75

(0.56e1.00) (0.26e0.76) (0.51e0.99) (0.26e0.76) (0.44e0.92) (0.24e0.74) (0.61e1.00) (0.61e1.00) (0.31e0.81) (0.24e0.78) (0.44e0.92) (0.31e0.85) (0.50e1.00)

0.77 (0.53e1.00) 0.68 (0.44e0.92) 0.63(0.32e0.94) 0.73 (0.49e0.97) 0.67 (0.34e1.00) 0.72 (0.45 0.99) 0.36 0.86 0.63 0.78 0.79 0.73 0.53 0.42 0.73

(0.03e0.69) (0.51e1.00) (0.26e1.00) (0.41e1.00) (0.42e1.00) (0.38e1.00) (0.28e0.78) (0.15e0.69) (0.38e1.00)

0.65 (0.30e1.00) 0.91 (0.64e1.00) 0.74 (0.47e1.00) 0.89 0.66 0.70 0.75 0.66

(0.65e1.00) (0.35e0.97) (0.48e0.92) (0.53e0.97) (0.44e0.88)

0.83 (0.46e1.00) 0.46 (0.30e0.62) 0.78 (0.53e1.00) 0.83 (0.56e1.00) 0.72 (0.47e0.97) (continued on next page)

288

J.J. Ford et al. / Manual Therapy 14 (2009) 283e291

Table 2 (continued) SCQ-LBP item (number of participants)

Kappa (95% CI)

Method of getting to sleep once woken (25) Type of bed (30) Presence of constant symptoms (30) Presence of stiffness in the morning (30) Duration for stiffness to ease (24)

0.73 1.00 0.43 0.61

Weighted Kappa (95% CI)

Mean difference (SD)

(0.44e1.00) (0.65e1.00) (0.43e0.08) (0.26e0.96) 0.78 (0.51e1.00)

Moderate agreement or above shown in bold, CI ¼ confidence interval.

from the SCQ-LBP including standing, erect sitting, sit to stand and cough/sneeze. The validity of subjective complaints whether measured by self administered questionnaire or interview can be affected by poor participant recall (Feine et al., 1998; Dawson et al., 2002). It is self-evident that the measurement of subjective complaints when observed in real time eliminates issues with recall. Real time measurement of subjective complaints was therefore regarded as a suitable reference standard by which concurrent validity of items from the self administered SCQLBP could be tested. At least moderate agreement was demonstrated in five of the seven items evaluated providing evidence for concurrent validity, however, the confidence intervals (CIs) were significantly lower than for the other comparisons made. The results of this component of the study therefore provide preliminary evidence for the concurrent validity of the SCQ-LBP and further research is indicated. Physiotherapists frequently measure subjective complaints by interview (McKenzie et al., 2003; Maitland, 2005). When investigating the value of a self administered questionnaire, it could be argued that the absence of a physiotherapist, and the

Table 3 Concurrent validity of the SCQ-LBP (initial SAePO). SCQ-LBP item

Kappa (95% CI)

Weighted Kappa (95% CI)

Aggravating factors Onset of symptoms 0.54 while standing (29) Duration of standing 0.59 limited by symptoms (28) Onset of symptoms 0.31 while erect sitting (29) Duration of erect 0.40 sitting limited by symptoms (28) Symptoms aggravated 0.35 by sitting to standing (30) Components of sit to stand 0.67 (0.32e1.00) that aggravate symptoms (25) Symptoms aggravated 0.49 by cough/sneeze (28)

(0.32e0.76) (0.35e0.83) (0.11e0.51) (0.15e0.65)

(0.13e0.57)

(0.25e0.73)

PO ¼ physiotherapist observed, moderate agreement or above shown in bold, CI ¼ confidence interval.

resultant unavailability of question clarification, might result in a relative reduction in the accuracy of information obtained. The current study therefore evaluated the agreement between the SA and PA in order to ascertain whether the SCQ-LBP was equally valid in a self administered format. There were minimal differences in the participant responses between the SA and PA. These results suggest that the self administered SCQ-LBP could be useful for future clinical research into the classification of LBP particularly given the moderate levels of reliability demonstrated. The questionnaire could be distributed by post or online, thereby minimising participant inconvenience and loss to follow up. Evidence for the face validity of the SCQ-LBP was provided by the initial SA-PA comparisons in this study. The face validity of the SCQ-LBP is further strengthened given the common usage of the questionnaire items in physiotherapy practice (McKenzie et al., 2003; Maitland, 2005) as well as in much of the large body of research currently being published in the area of LBP classification (Ford et al., 2007). However, most of the classification systems currently being investigated do not have a comprehensive and reliable list of subjective complaints that can be measured in a self administered format. Further research is required on the concurrent validity of the SCQ-LBP items, however, we believe that given substantial reliability and face validity, the SCQ-LBP in its current form has the potential to make a significant contribution to future classification research. Despite these positive findings, a number of items were found to have lower levels of agreement. One of the aims of the physiotherapy clinical interview is to obtain a description of the response to activities/ positions that aggravate symptoms. This usually includes the duration of the activity or position until symptom onset, and the duration upon which the severity of symptoms necessitate cessation of the activity or position; also known as P2 (Maitland, 2005). In the piloting of the SCQ-LBP, eight out of 13 participants experienced difficulty understanding the wording regarding P2. The wording for these items concerning P2 was modified as part of the SCQ-LBP development. In spite of these changes,

289

J.J. Ford et al. / Manual Therapy 14 (2009) 283e291 Table 4 Face validity of the SCQ-LBP (initial SAePA). SCQ-LBP item (number of participants)

Kappa (95% CI)

Pain drawing item Area of symptoms (30) Number of symptom descriptors (30) Presence of leg pain (30) Presence of dermatomal symptoms (30) Presence of distal paraesthesia (30)

0.85 (0.50e1.00) 0.79 (0.44e1.00) 0.78 (0.43e1.00)

Symptom details Presence of pins and needles (30) Area of most severe pain (24) Presence of foot drop (30) Lumbar symptoms deep or superficial (29) Leg symptoms deep or superficial (30) Symptom relationship (30)

0.59 0.78 1.00 0.51 0.65 0.65

(0.24e0.94) (0.49e1.00) (0.65e1.00) (0.16e0.86) (0.41e0.89) (0.40e0.90)

0.39 (0.06e0.72) 0.56 (0.29e0.83)

Easing factors Whether standing eases symptoms (30) Whether sitting eases symptoms (30) Whether sitting in a couch eases symptoms (30) Whether walking eases symptoms (29) Whether lying supine eases symptoms (30) Whether crook lying eases symptoms (30) Whether lying on the sore side eases symptoms (29) Whether lying on the good side eases symptoms (28) Whether lying prone eases symptoms (29)

0.52 0.67 0.63 0.72 0.73 0.71 0.47 0.23 0.75

History of over the past three days Time of day when symptoms worst (29) Presence of trouble getting to sleep (30) Presence of waking at night (30) Reason for waking at night (30)

Mean difference (SD)

0.50 (3.01) 0.10 (0.84)

Aggravating factors Onset of symptoms while walking (29) Duration of walking limited by symptoms (28) Onset of symptoms while standing (28) Duration of standing limited by symptoms (27) Onset of symptoms while erect sitting (29) Duration of erect sitting limited by symptoms (28) Onset of symptoms while slumped sitting (29) Duration of slumped sitting limited by symptoms (29) Onset of symptoms while lying supine (29) Duration of lying supine limited by symptoms (28) Onset of symptoms while lying prone (29) Duration of lying prone limited by symptoms (27) Symptoms aggravated by getting into car (29) Components of getting into car that aggravate (24) Symptoms aggravated by sitting to standing (30) Components of sit to stand that aggravate symptoms (28) Symptoms aggravated by cough/sneeze (30) Presence of symptoms worse on one side (30) Most comfortable side to lie on (30)

History of symptoms Whether current symptoms are the first episode (30) Duration of current symptoms (29) Number of episodes of symptoms (30) Time from first episode of symptoms (29) Whether symptoms chronic or recurrent (30) Mechanism of injury for current symptoms (29) Time of onset of current symptoms (30) Behaviour of current symptoms after onset (29) History of manual handling work (29)

Weighted Kappa (95% CI)

0.68 0.42 0.72 0.63 0.43 0.38 0.71 0.47 0.35 0.19 0.51 0.35 0.41

(0.43e0.93) (0.17e0.67) (0.48e0.96) (0.38e0.88) (0.19e0.67) (0.13e0.63) (0.47e0.95) (0.23e0.71) (0.10e0.60) (0.06e0.44) (0.27e0.75) (0.10e0.60) (0.16e0.66)

0.83 (0.59e1.00) 0.57 (0.33e0.81) 0.60 (0.29e0.91) 0.58 (0.34e0.82)

(0.17e0.87) (0.32e1.00) (0.30e0.96) (0.37e1.00) (0.38e1.00) (0.38e1.00) (0.22e0.72) (0.02e0.48) (0.38e1.00)

0.65 (0.30e1.00) 0.94 0.74 0.76 0.65 0.75 0.75 0.48

(0.69e1.00) (0.49e0.99) (0.52e1.00) (0.36e0.94) (0.53e0.97) (0.53e0.97) (0.26e0.70)

0.47 0.82 0.75 0.62

(0.31e0.63) (0.57e1.00) (0.50e1.00) (0.37e0.87)

0.81 (0.44e1.00)

(continued on next page)

290

J.J. Ford et al. / Manual Therapy 14 (2009) 283e291

Table 4 (continued) SCQ-LBP item (number of participants) Method of getting to sleep once woken (29) Type of bed (30) Presence of constant symptoms (30) Presence of stiffness in the morning (30) Duration for stiffness to ease (24)

Kappa (95% CI)

Weighted Kappa (95% CI)

Mean difference (SD)

0.76 (0.51e1.00) 1.00 (0.65e1.00) 0.79 (0.44e1.00) 0.61 (0.26e0.96) 0.73 (0.46e1.00)

Moderate agreement or above shown in bold, CI ¼ confidence interval.

relatively lower levels of agreement and associated wider CIs were found in the initial SAeretest SA and initial SAePA comparisons regarding P2 for aggravating factors. The questions regarding the effect of the lying position on symptoms also yielded consistently lower levels of agreement and wider CIs in the initial SAeretest SA and initial SAePA comparisons. This was particularly the case regarding items describing the effect of unilateral symptoms on lying in different positions. Conceptually the principles of P2 and the effect of lying may be too complex in a self administered format for use with the population participating in this study. Caution therefore needs to be exercised when including these items in future clinical research. 4.1. Limitations of the study The population sampled for the current study was primarily those with chronic compensable LBP and further research is required to determine the external validity of the SCQ-LBP in relation to acute and noncompensable LBP populations. Some of the data show relatively modest levels of agreement with wide CIs. It is possible that this was influenced by spontaneous change in the participant’s condition (not detected by the analysis of disability data) or low sample size. Future research on the SCQLBP needs to consider these factors in the planning stage. 5. Conclusion The SCQ-LBP was developed for the purpose of providing a measurement tool for subjective complaints in the classification of LBP. The testeretest reliability of the self administered SCQ-LBP was established as well as preliminary evidence for the concurrent validity of selected items compared with an acceptable reference standard. Moderate agreement between the SA with the PA demonstrate that subjective complaints information can be measured with similar levels of agreement using a self administered format as compared to a physiotherapy interview. The SCQ-LBP may be a useful tool for future research into the classification of LBP. Further research

is required to replicate these findings with larger sample sizes as well as in different populations to investigate external validity. Appendix A. Supplementary data Supplementary data associated with this article can be found in the online version, at doi:10.1016/ j.math.2008.02.010. References Anastasi A, Urbina S. Psychological testing. 7th ed. New Jersey: Prentice Hall; 1997. Bertilson BC, Bring J, Sjoblom A, et al. Inter-examiner reliability in the assessment of low back pain (LBP) using the KirkaldyWillis classification (KWC). European Spine Journal 2006;15: 1695e703. Biering-Sorensen F, Hilden J. Reproducibility of the history of lowback trouble. Spine 1984;9:280e6. Bland J, Altman D. Statistical methods for assessing agreement between two methods of clinical measurement. The Lancet 1986;1:307e10. Boissonnault W, Di Fabio R. Pain profile of patients with low back pain referred to physical therapy. Journal of Orthopedic and Sports Physical Therapy 1996;24:180e91. Borkan J, Cherkin D. An agenda for primary care research on low back pain. Spine 1996;21:2880e4. Burdorf A, Laan J. Comparison of methods for the assessment of postural load on the back. Scandinavian Journal of Work and Environmental Health 1991;17:425e9. Clare HA, Adams R, Maher CG. Reliability of McKenzie classification of patients with cervical or lumbar pain. Journal of Manipulative and Physiological Therapeutics 2005;28:122e7. Dankaerts W, O’Sullivan PB, Straker LM, et al. The inter-examiner reliability of a classification method for non-specific chronic low back pain patients with motor control impairment. Manual Therapy 2006;11:28e39. Dawson EG, Kanim LEA, Sra P, et al. Low back pain recollection versus concurrent accounts: outcomes analysis. Spine 2002;27: 984e93. Delitto A, Erhard R, Bowling R. A treatment-based classification approach to low back syndrome: identifying and staging patients for conservative treatment. Physical Therapy 1995;75:470e89. Feine J, Lavigne G, Dao T, Morin C, Lund J. Memories of chronic pain and perceptions of relief. Pain 1998;77:137e41. Ford J, Story I, O’Sullivan P, et al. Classification systems for low back pain: a review of the methodology for development and validation. Physical Therapy Reviews 2007;12:33e42. Fritz JM, Brennan GP, Clifford SN, et al. An examination of the reliability of a classification algorithm for subgrouping patients with low back pain. Spine 2006;31:77e82.

J.J. Ford et al. / Manual Therapy 14 (2009) 283e291 Heiss DG, Fitch DS, Fritz JM, et al. The interrater reliability among physical therapists newly trained in a classification system for acute low back pain. Journal of Orthopedic and Sports Physical Therapy 2004;34:430e9. Kopec J, Esdaile J, Abrahamowicz M, et al. The Quebec back pain disability scale: measurement properties. Spine 1995;20:141e52. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159e74. Mann N, Brown M, Hertz D, Enger I, Tompkins J. Initial-impression diagnosis using low-back pain patients pain drawings. Spine 1993;18:41e53. Maitland G. Maitland’s vertebral manipulation. 7th ed. Philadelphia: Elsevier; 2005. Margolis R, Tait R, Krause S. A rating system for use with patient pain drawings. Pain 1986;24:57e65. McKenzie AR, May S, McKenzie R. The lumbar spine: mechanical diagnosis and therapy. 2nd ed. Orthopaedic Physical Therapy Products; 2003. Ohnmeiss D, Vanharanta H, Guyer R. The association between pain drawings and computed tomographic/discographic pain responses. Spine 1995;20:729e33. Parker H, Wood P, Main C. The use of the pain drawing as a screening measure to predict psychological distress in chronic low back pain. Spine 1995;20:236e43. Petersen T, Olsen S, Laslett M, et al. Inter-tester reliability of a new diagnostic classification system for patients with non-specific low back pain. Australian Journal of Physiotherapy 2004;50:85e94. Petersen T. Classification of non-specific low back pain: a review of the literature on classification systems relevant to physiotherapy. Physical Therapy Reviews 1999;4:265e81.

291

Portney L, Watkins M. Foundations of clinical research: applications to practice. 1st ed. Norwalk: Appleton and Lange; 1993. Rankine J, Fortune D, Hutchison S, Hughes D, Main C. Pain drawings in the assessment of nerve root compression: a comparative study with lumbar spine magnetic resonance imaging. Spine 1998;23:1668e76. Riddle D. Classification and low back pain: a review of the literature and critical analysis of selected systems. Physical Therapy 1998; 78:708e37. Roach K, Brown M, Dunigan K, Kusek C, Walas M. Testeretest reliability of a low back pain questionnaire. Physical Therapy 1994; 74:S56. Roach K, Brown M, Randi D, Delaney K, Lipprandt H, Rangelli D. The sensitivity and specificity of pain response to activity and position in categorizing patients with low back pain. Physical Therapy 1997;77:730e7. Roach K, Brown M, Ricker E, Altenburger P, Tompkins J. The use of patient symptoms to screen for serious back problems. Journal of Orthopedic and Sports Physical Therapy 1995;21:2e6. Rucker K, Mezler H, Kregel J. Standardization of chronic pain assessment: a multi-perspective approach. The Clinical Journal of Pain 1996;12:94e110. Uden A, Astrom M, Bergenudd H. Pain drawings in chronic back pain. Spine 1988;13:389e92. Uden A, Landin L. Pain drawing and myelography in sciatic pain. Clinical Orthopaedics and Related Research 1987;216:125e30. Walsh K, Coggon D. Reproducibility of histories of low-back pain obtained by self-administered questionnaire. Spine 1991;16:1075e7. Weiner D, Peterson B, Keefe F. Evaluating persistent pain in long term care residents: what role for pain maps? Pain 1998;76:249e57.

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 292e298 www.elsevier.com/math

Original Article

Reliability of accessory motion testing at the carpal joints Filip Ferdinand Staes a,b,*, Kevin James Banks c, Luc De Smet a,b, Kim Josefine Daniels a, Pieter Carels a a

Department of Rehabilitation Sciences, Faculty of Kinesiology and Rehabilitation Sciences, Katholieke Universiteit Leuven, Tervuursevest 101, B-3001 Leuven, Belgium b University Hospitals Leuven, Leuven, Belgium c Rotherham Primary Care NHS Trust, UK Received 17 July 2007; received in revised form 12 March 2008; accepted 10 April 2008

Abstract The testing of accessory motion has become a very important part of manual therapy practice. Its value is in assessing whether joint mobility is ideal or impaired. Despite its use, there is little evidence in the literature to support the reliability of such testing. Most of the research carried out on accessory motion testing has focused on the spine. In view of this we decided to evaluate the intra- and interrater reliability of accessory motion testing of carpal joints. Two skilled therapists tested the available motion and the end-feel response of carpal joints in 30 students and 15 patients on two separate occasions. Pain scores were also obtained. In students a moderate to good percentage of agreement [67e97%] was obtained for motion testing. In patients the percentage of agreement ranged from 60% to 100% and weighted kappa values were between 0.33 and 1.0. Intrarater reliability was better than interrater reliability in both groups. Intra- and interrater agreement on end-feel was very good. Overall, the reliability of accessory motion testing of carpal joints was acceptable. The results suggest that this form of testing can be valuable in the training of manual therapists and in clinical practice. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Reliability; Wrist; Manual therapy; Mobility

1. Introduction Accessory motion testing (AMT) is defined as movement of a joint that can only be performed passively by an external force. It is a method used to evaluate the amount and behaviour of resistance relative to motion and also the corresponding end-feel (Maitland, 1991). In addition, the amount and behaviour of pain relative to motion can be evaluated. AMT is well-established in clinical practice and within many manual therapy * Corresponding author. Department of Rehabilitation Sciences, Faculty of Kinesiology and Rehabilitation Sciences, Katholieke Universiteit Leuven, Tervuursevest 101, B-3001 Leuven, Belgium. Tel.: þ32 16 329125; fax: þ32 16 329197. E-mail address: fi[email protected] (F.F. Staes). 1356-689X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.04.001

training programmes. Research, however, has focused mainly on the testing of spinal intervertebral movement and the limited data demonstrate poor inter- and intrarater reliability (Maher and Adams, 1994; Binkley et al., 1995; Strender et al., 1997; Hicks et al., 2003). There is only a limited amount of data on motion testing of peripheral joints. In a study by Hayes and Petersen (2001), two experienced physical therapists judged the end-feel and the pain to resistance ratio (P/R ratio) when performing passive physiological movements on patients with painful shoulders and painful knees. Intrarater reliability (kappa: 0.76e1.00) was better than interrater reliability (kappa: 0.01 to 0.70) for both end-feel and P/R ratio. Doubt exists whether clinicians can reliably test joint motion which is small in range, such as in the carpal joints

F.F. Staes et al. / Manual Therapy 14 (2009) 292e298

(Chesworth et al., 1998). Performing AMT of the carpal joints is also rather complicated. The wrist itself is structurally complex and the contribution of the radio-carpal and midcarpal joints to the range of flexion and extension of the wrist is still a subject for debate (Patterson et al., 1998). Translation of only 0.5e3.3 mm has been observed at the radio-carpal joint in a cadaver study by Ruby et al. (1988) using biplanar orthoradiographs in five different wrist positions. Intercarpal movement at the proximal row is even less (0.09e0.27 mm). The amount of relative motion of the distal to proximal row ranges from 0.10 to 1.20 mm (Ruby et al., 1988). The intercarpal movement between distal carpal bones ranges from 0.19 to 1.07 mm for flexion and extension. The amount of translation of the carpal bones during movement has been shown to be, at most, 2 mm (Jackson et al., 1994). Previous research has focused mainly on kinematic analysis of wrist motion rather than the reliability of the clinical evaluation of motion. Despite the small translatory ROM available at the wrist and hand, the vast majority of manual therapy educators and therapists seem to accept AMT as a valuable assessment tool. Because there are no reliability studies available, this study aims to evaluate intra- and interrater reliability of AMT of carpal joints in a group of students and/or a group of patients with wrist pathology.

2. Method 2.1. Participants As AMT is used in both manual therapy training programmes and clinical settings, two groups of participants were selected to represent each population. One group consisted of 30 students, 22 women and eight men with a mean age of 21 years (SD: 2 years), and all without any history of injury to the hand, wrist or distal end of the forearm (Table 1). A second group included 15 post-operative patients, eight women and seven men, with a mean age of 38 Table 1 Characteristics of students (n ¼ 30) and patients (n ¼ 15). Characteristics

Students

Patients

Age (years) Mean (SD) Range

21.33 (1.6) 18e24

38.33 (11.02) 18e58

Sex Male (n) Female (n)

8 22

7 8

Dominant hand Left (n) Right (n)

6 24

3 12

SD ¼ standard deviation; n ¼ number.

293

years (SD: 11.0 years) (Table 1). Patients in this group were selected by an orthopaedic surgeon (LDS). The wrist problems were diagnosed as follows: scapho-lunate dissociation (n ¼ 4), rupture of the TFCC complex (n ¼ 5), fracture of the hamate (n ¼ 1), fractures of the distal end of the radius (n ¼ 3) and fractures of the distal end of the ulna (n ¼ 1), subluxation of the carpometacarpal joint (n ¼ 1), carpal tunnel syndrome (n ¼ 1), SLAC wrist (n ¼ 1) and exostosis of the scaphoid (n ¼ 1). Eleven patients received open surgery (BLATT n ¼ 1; resection n ¼ 2; pinning n ¼ 5; bone fusion n ¼ 2; shortening of the ulna n ¼ 1; external fixation n ¼ 2; styloidectomy of the radius n ¼ 1). Two patients underwent arthroscopy, another one received a corticosteroid injection and one patient was told to visit a physical therapist. All participants received information about the procedure either from an investigator or from an orthopaedic surgeon. All participants provided us with their written informed consent. The study was approved by the Ethical Medical Committee of the University Hospitals Leuven. 2.2. Examiners As training and experience may influence the results, two physical therapists with different levels of clinical and teaching experience, but both skilled in the use of AMT, were selected to perform the testing. One therapist had 10 years of clinical experience (therapist A), the other 7 years (therapist B). Therapist A had more experience in working with students, therapist B in treating patients with wrist problems. During testing, the therapists neither were not aware of the details of the patient’s diagnosis, nor were they informed about the other therapist’s findings. 2.3. Questionnaire Each participant in the study was asked to complete a brief questionnaire in order to obtain information about their age, sex, dominant hand, profession, hobbies and, in the patient group, injuries to the wrist. This information was gathered independently by a researcher not involved in the testing itself. 2.4. Motion testing and scoring The two therapists were asked to rate the mobility of the following carpal joints: capitateehamate, capitatee lunate, capitateescaphoid and capitateetrapezoid. The method of AMT was as described by Kaltenborn (1983). Both hands were examined and compared. A rating scale based on three descriptors was used to express the difference in mobility between both hands: (1) no or reduced mobility, (2) comparable mobility and

294

F.F. Staes et al. / Manual Therapy 14 (2009) 292e298

(3) increased mobility. In the students group, the dominant hand was used as reference, in the patient group the injured hand was used. Each motion test was followedup by an assessment of the end-feel for the dominant/injured and non-dominant/non-injured hand, according to the Cyriax classification used in other investigations (Hayes et al., 1994). In order to verify whether the participant’s sensitivity to movement altered because of the testing, participants reported the amount of pain experienced pre- and posttesting on a 0e10 visual analogue scale (VAS), with 0 cm representing no pain and 10 cm representing the worst imaginable pain. Pain scores were taken for both hands. 2.5. Procedure Before the start of the study, the physical therapists received one training session of 30 min on the investigation techniques and scoring system. Prior to the testing, participants were asked to fill in the questionnaire and sign the informed consent form. Participants waited in an area outside the testing room and were then escorted randomly into enclosed testing areas. Randomization was carried out by a researcher not involved in the testing and was based on randomization tables generated by a computer. In the student group, therapists were blinded for the participants’ face by seating the students behind curtains. All students were tested in the same room. Testing in the patient group was carried out in a way that did not allow the therapists to be blinded for the patients’ faces. In both groups, the participant’s dominant hand was marked with a cross (prior to the testing). In the patient group, the injured hand was marked with a circle. Each wrist was held in the resting position, if possible (Kaltenborn, 1983). During the testing, subjects were instructed to answer the therapist’s questions only if they related to pain. Each student and each patient was tested twice by each examiner. The mean time between testing of the same participant was 12 min in the student group and 10 min in the patient group. 2.6. Data analysis Pre- and post-testing pain is reported through the mean scores. Frequencies were used for the three mobility categories. The percentage of agreement and weighted kappa (Kw), with it’s 95% confidence limits, were calculated to assess intra- and interrater agreement of AMT in the student and patient groups. Linear weights were used for the calculation of Kw. The interpretation of Kw was based on the criteria of Landis and Koch (1977). A Kw value lower than 0.2 was considered to be slight, between 0.2 and 0.4 as fair, between 0.4 and 0.6 as moderate, between 0.6 and 0.8

as substantial and when more than 0.8 as almost perfect. As prevalence of the attribute influences the kappa coefficient, the prevalence index was calculated with the following formula: Prevalence index ¼

ja  dj n

where a and d represent the cells of agreement and n the number of paired ratings (Sim and Wright, 2005). The value ranges between 0 and 1. The prevalence index is designed for 2  2 tables. For this study, we worked with 3  3 tables and the prevalence index was calculated for each pairwise comparison (Rogel, 1997). Only the highest prevalence value will be reported. A high prevalence index reflects a high percentage of agreement and lowers the kappa coefficient (Sim and Wright, 2005). Bias reflects the (a)symmetry of disagreement between observers and, for a 2  2 table, is calculated as: Bias index ¼

jb  cj n

where b and c reflect the cells of disagreement between raters and n the number of paired ratings (Sim and Wright, 2005). A high bias index reflects a high level of asymmetry in disagreement. As for this study all values were very low (<0.1) and thus did not significantly influence the kappa coefficient, therefore the values are not reported. Data from the first and second trial of each therapist served to calculate the intrarater reliability. Data from the first trial of each examiner were used for the interrater reliability.

3. Results 3.1. Carpal mobility Tables 2 and 3 present the frequencies and level of agreement for the three categories of carpal joint mobility for each articulation in the student and patient groups. In the student group, comparable mobility between both hands was predominantly found in almost all articulations (Table 2). The percentages of agreement were moderate towards excellent for all articulations, with better results for intrarater agreement compared with interrater agreement. Kw values were low to excellent, but were influenced by the very high prevalence indices which were due to the high prevalence of comparable mobility between both hands. In the patient group, a wider variety of differences in mobility between both hands was observed (Table 3). The intrarater agreement was better than the interrater agreement. Prevalence indices were

Table 2 Frequencies (n) of segmental mobility categories in 30 students comparing the mobility between both hands, using the dominant hand as a reference. Frequencies per trial per therapist

Agreement

Trial 1a

Intratherapist

Trial 2a

Therapist A 1

2 7 4 2

Therapist B

b

b

2

3

28 23 23 27

0 0 3 1

b

Therapist A

1

b

b

2

3

0 4 1 1

24 24 27 28

6 2 2 1

b

1

b

2

3

5 7 3 2

24 22 27 28

1 1 0 0

Therapist B b

b

Therapist A b

Trial 1a

Therapist B

Trial 2a

1

b

2

3

% Kw (95% CI)

PI

% Kw (95% CI)

PI

% Kw (95% CI)

PI

% Kw (95% CI)

PI

0 5 0 1

27 24 28 29

3 1 2 0

73 80 83 97

0.73 0.67 0.77 0.90

83 80 90 90

0.76 0.70 0.87 0.90

77 73 77 83

0.77 0.67 0.73 0.83

77 67 87 87

0.73 0.60 0.87 0.87

0.26 0.57 0.53 0.80

(0.03e0.48) (0.42e0.73) (0.34e0.72) (0.61e1.00)

0.46 0.53 0.51 0.31

(0.24e0.68) (0.35e0.70) (0.24e0.78) (0e0.69)

0.30 0.42 0.39 0.29

(0.07e0.53) (0.25e0.60) (0.19e0.59) (0e0.58)

0.37 0.35 0.35 0.30

(0.16e0.58) (0.19e0.52) (0.05e0.65) (0e0.62)

% ¼ Percentage; Kw ¼ weighted kappa; PI ¼ prevalence index; and CI ¼ confidence interval. a First examination done by each therapist (trial 1)/second examination done by each therapist (trial 2). b Segmental mobility categories with 1 (no or reduced mobility), 2 (comparable mobility) and 3 (increased mobility).

Table 3 Frequencies (n) of segmental mobility categories in 15 patients comparing the mobility between both hands, using the injured hand as a reference. Frequencies per trial per therapist

Agreement

Trial 1a

Intratherapist

Trial 2a

Therapist A b

1

Capitateehamate 9 Capitateelunate 10 Capitateescaphoid 9 Capitateetrapezoid 4

Therapist B

b

b

b

b

2

3

1

2

6 4 5 10

0 1 1 1

8 9 7 6

7 5 8 9

Therapist A b

b

Therapist B

3

b

b

1

2

3

0 1 0 0

8 11 8 5

7 3 6 9

0 1 1 1

b

Intertherapist

Therapist A b

1

2

3

% Kw (95% CI)

PI

8 9 7 5

7 5 7 10

0 1 1 0

80 93 80 80

0.33 87 0.76 (0.61e0.92) 0.47 100 1 (1e1) 0.20 93 0.88 (0.76e0.99) 0.47 93 0.87 (0.74e0.99)

(0.48e0.84) (0.73e0.99) (0.52e0.85) (0.49e0.84)

Trial 1a

Therapist B

b

0.66 0.86 0.68 0.66

F.F. Staes et al. / Manual Therapy 14 (2009) 292e298

Capitateehamate Capitateelunate Capitateescaphoid Capitateetrapezoid

b

Intertherapist

%

Kw (95% CI)

Trial 2a

PI

% Kw (95% CI)

PI

% Kw (95% CI)

PI

0.40 0.27 0.47 0.60

53 73 60 67

0.13 0.13 0.20 0.47

87 80 73 93

0.40 0.40 0.27 0.60

0.33 0.66 0.60 0.47

(0.14e0.52) (0.49e0.84) (0.43e0.77) (0.28e0.66)

0.76 0.75 0.69 0.87

(0.61e0.92) (0.59e0.92) (0.53e0.85) (0.74e0.99)

% ¼ Percentage; Kw ¼ weighted kappa; PI ¼ prevalence index; and CI ¼ confidence interval. a First examination done by each therapist (trial 1)/second examination done by each therapist (trial 2). b Segmental mobility categories with 1 (no or reduced mobility), 2 (comparable mobility) and 3 (increased mobility).

295

296

F.F. Staes et al. / Manual Therapy 14 (2009) 292e298

not as high as in the students group, therefore exerting less of an influence on kappa values. 3.2. End-feel and pain The end-feel of all articulations was described as ‘‘capsular’’ in all student participants leading to a 100% agreement within and between therapists. In patient participants, the end-feel of the injured hand was, generally, rated as ‘‘capsular’’ also. Only on two occasions the end-feel was categorized as ‘‘hard’’. In six patients, the end-feel changed during the testing. Good to excellent (ranging from 73.3% to 100%) agreement was found. The mean pain score for the patients’ injured hand before each examination was 2.29, the mean post-testing pain score was 2.65. This shows that there was no significant difference between pre- and post-testing pain. Four patients had no pain at all. Four of the remaining patients reported a higher post-testing pain score. The mean pre- and post-testing pain score was 2.2 and 3.3, respectively. One patient had pain in both hands. None of the students perceived pre-testing pain. Only two participants experienced mild pain (VAS score of 1) after the examination and one participant reported a different sensation [not pain] in the wrist after the examination.

been due to either a learning effect or, possibly, the presence of a ‘‘recall bias’’ as therapists were not blinded for the patient’s faces. However, ‘‘recall bias’’ did not appear to be a factor which concerned the therapists when they were interviewed after the trial. In the student group as well as in the patient group, intrarater agreement was better than interrater agreement. This is in accordance with previous research (Gonnella et al., 1982; Hayes and Petersen, 2001). The interpretation of mobility testing is susceptible to sources of error. Therapists have their own frame of reference based on education and experience with patients, therefore influencing their evaluation (Riddle, 1992). It is still difficult to find good operational definitions for the criteria used to clinically evaluate the ROM. The therapists only received a short period of training. More intensive training might have improved reliability. Another source of possible error is the position of the wrist. Although therapists were asked to hold the participants’ wrists in the resting position whenever they could, this was not always possible in the patient group because of pain and limitation of movement. Even small differences in positioning could alter the therapist’s perceptions and therefore influence agreement. 4.2. End-feel

4. Discussion 4.1. Carpal joint mobility The purpose of this study was to investigate the reliability of AMT of carpal joints in a group of students and a group of patients with wrist or hand conditions. In the student group, a moderate to good percentage of agreement (67e97%) was found for motion testing. The percentages of agreement might have been influenced by the limited variability in carpal joint mobility. That is, a rating of ‘‘comparable mobility’’ between hands was reported most. The high prevalence of this category also explains the high prevalence index that leads to a reduction of the Kw value. It is possible to correct kappa for prevalence and bias (the prevalenceadjusted bias-adjusted kappa) for 2  2 tables, but such correction is not applicable to 3  3 tables. The small variability of mobility between both hands could have been due to the limited age range in our students group. However, the students selected for this study represent the student population which receives training in AMT. In the patient group, agreement was fair to very good. Percentages ranged from 60% to 100% and Kw from 0.33 to 1.0. Higher agreement between therapists was found during the second trial. This could have

The second purpose of this study was to test the intraand interrater reliability for end-feel categories in the student and patient groups. Intra- and interrater agreement of end-feel was very good. As in the study by Chesworth et al. (1998) all of our students and most of our patients were assigned to only one end-feel category, in our case the ‘‘capsular’’ end-feel category. 4.3. Implications for practice This study is important for the teaching of manual therapy and for clinical practice. The therapists in our study were highly skilled practitioners. Our study does not allow us to draw conclusions about the testing of students of manual therapy as raters. We do believe, however, that it is important for highly skilled teachers to be able to produce reliable findings so that they can provide feedback to their students. The findings in the patient group indicate that therapists can test joint motion reliably within their own frame of reference. Although, based on lower interrater reliability, it should be accepted that, interpretations may differ between therapists, probably because of their different frames of reference (Riddle, 1992). An important difference between this study and clinical practice was that therapists were not permitted to take a history from the patient. Interview data might affect further judgments

F.F. Staes et al. / Manual Therapy 14 (2009) 292e298

about the patient’s condition. Having information about the patient prior to examination, as is usual in daily practice, allows therapists to generate hypotheses that may create expectations about test findings.

4.4. Methodological considerations Thirty students and 15 patients participated. Little is known about the optimal sample size required for reliability studies using percentages of agreement and kappa. Sim and Wright (2005) discussed the sample size and the number of raters that might be necessary for reliability studies for a two-rater study and a dichotomous variable. They presented a table indicating how many subjects should be included, taking into account: the proportion of positive ratings, the kappa value to detect, and the null hypothesis. The number of subjects required ranged between eight 2164 persons. However, their choice of choosing the null hypothesis as a parameter in decision-making can be discussed, as Tooth and Ottenbacher (2004) have done. They mentioned that kappa is not suitable for null hypothesis testing, and that power calculations are not relevant. They stated that size and stability of kappa are of more importance and that power analysis and use of confidence intervals [CI] are less instructive. In our study, wide ranges of CI were observed in the student group. The CI in the student group may have been affected by the (very) high prevalence index for most observations. Due to the nature of the investigation, one may expect that this finding would be confirmed even in larger groups of subjects. Moderate to substantial kappa values, with more but more stable CI, were found in the patient group. This might indicate that the sample size was suitable for this purpose. The use of a three scale scoring system may have enhanced results. Other studies used 7- to 11-point scales (Gonnella et al., 1982; Maher and Adams, 1994; Binkley et al., 1995). In line with Hicks et al. (2003), we preferred a 3-point scale as it reflects, in our opinion, how judgments are made in clinical practice. Therapists are usually interested in limited or increased mobility when making their therapeutic decisions. As mentioned earlier, Riddle (1992) suggested that each therapist has his own frame of reference. Providing a large number of scoring categories might therefore have made it even more difficult to reach agreement. In daily practice, this would reduce effective communication between therapists. The dominant and injured hands were marked. This is not routine practise in AMT but marks were used to overcome the fact that the therapists were not permitted to interview the patients. Having been asked to compare the motion of carpal joints in both hands, therapists needed to know which hand was the reference. Therefore, this was why marks were used.

297

Therapists judgments were based on a maximum of five joint movements at each articulate in order to reduce a mobilizing effect. However, in some cases therapists reported that, in the second trial, patients had more pain or that the end-feel had changed, both of which might indicate that there was a testing effect. Each therapist tested each subject twice. In the case of the students’ testing, therapists were blinded for the subject’s faces, therefore recall bias was avoided. Therapists could not be blinded for the patients’ faces. Therefore, one might argue that a recall bias could not be excluded. However, the time between these two tests, for each therapist, was about 20e24 min. Between these two measurements, other patients were tested (3e4 more patients for each therapist). This reduced the chances of recall bias. As mentioned earlier, therapists also indicated that, before the second test they were not able to remember the findings of the first test. 5. Conclusions In this study, intrarater reliability and to a lesser degree interrater reliability were generally acceptable for carpal joint mobility judgments in a group of students and a group of patients. Limited variability in mobility findings might have affected reliability. Reliability of end-feel judgments was very good, but was also affected by limited variability in the end-feel categories. References Binkley J, Stratford PW, Gill C. Interrater reliability of lumbar accessory motion mobility testing. Physical Therapy 1995;75(9): 786e95. Chesworth BM, MacDermid JC, Roth JH, Patterson SD. Movement diagram and ‘‘end-feel’’ reliability when measuring passive lateral rotation of the shoulder in patients with shoulder pathology. Physical Therapy 1998;78(6):593e601. Gonnella C, Paris SV, Kutner M. Reliability in evaluating passive intervertebral motion. Physical Therapy 1982;62(4): 436e44. Hayes KW, Petersen C, Falconer J. An examination of Cyriax’s passive motion tests with patients having osteoarthritis of the knee. Physical Therapy 1994;74(8):697e709. Hayes KW, Petersen CM. Reliability of assessing end-feel and pain and resistance sequence in subjects with painful shoulders and knees. Journal of Orthopaedic and Sports Physical Therapy 2001;31(8):432e45. 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(12):1858e64. doi:10.1016/S0003-9993(03)00365-4. Jackson WT, Hefzy MS, Guo H. Determination of wrist kinematics using a magnetic tracking device. Medical Engineering and Physics 1994;16(2):123e33. Kaltenborn FM. Mobiliseren van extremiteitsgewrichten, onderzoek en basisbehandelingstechnieken. Oslo: Norlis; 1983. Landis RJ, Koch GG. A one-way components of variance model for categorical data. Biometrics 1977;33:671e9.

298

F.F. Staes et al. / Manual Therapy 14 (2009) 292e298

Maher C, Adams R. Reliability of pain and stiffness assessments in clinical manual lumbar spine examination. Physical Therapy 1994;74(9):801e9. Maitland GD. Peripheral manipulation. 3rd ed. London: Butterworths; 1991. Patterson RM, Nicodemus CL, Viegas SF, Elder KW, Rosenblatt J. High-speed, three-dimensional kinematic analysis of the normal wrist. The Journal of Hand Surgery 1998;23(3): 446e53. Riddle DL. Measurement of accessory motion: critical issues and related concepts. Physical Therapy 1992;72(12):865e74. Rogel A. Analyse de concordance entre plusieurs observateurs. Doctoral thesis, Universite´ de Paris 7; 1997. Ruby LK, Cooney WP, An KN, Linscheid RL, Chao EY. Relative motion of selected carpal bones: a kinematic analysis of the normal wrist. Journal of Hand Surgery 1988;13A(1):1e10. Sim J, Wright CC. The kappa statistic in reliability studies: use, interpretation, and sample size requirements. Physical Therapy 2005;85(3):257e68. Strender LE, Lundin M, Nell K. Interexaminer reliability in physical examination of the neck. Journal of Manipulative and Physiological Therapeutics 1997;20(8):516e20. Tooth LR, Ottenbacher KJ. The K statistic in rehabilitation research: an examination. Archives of Physical Medicine and Rehabilitation 2004;85:1371e6.

Filip F. Staes, PhD, PT is currently working as an Associate Professor in Musculoskeletal Rehabilitation. He is the Head of the Research Centre for Musculoskeletal Rehabilitation of the Faculty of Kinesiology and Rehabilitation Sciences, Katholieke Universiteit Leuven, Belgium. Kevin J. Banks, BA, MCSP, MMACP, Senior Teacher IMTA is currently working as a senior lecturer in Physiotherapy at Sheffield Hallam University and as a physiotherapist at the Rotherham Primary Care NHS Trust, United Kingdom. Luc De Smet, PhD, MD is an Associate Professor in Orthopaedic Surgery at the University Hospitals Leuven, Belgium. He’s also a member of the Research Centre for Musculoskeletal Rehabilitation of the Faculty of Kinesiology and Rehabilitation Sciences, Katholieke Universiteit Leuven, Belgium. Kim Josefine Daniels, MSc, PT is a private practitioner in Manual Therapy and part-time educational assistant at the Department of Rehabilitation Sciences, Faculty of Kinesiology and Rehabilitation Sciences, Katholieke Universiteit Leuven. Pieter Carels, MSc, PT is a private practitioner. This paper was part of his Master’s Thesis.

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 299e305 www.elsevier.com/math

Original Article

Reliability of intra- and inter-rater palpation discrepancy and estimation of its effects on joint angle measurements Cristiane Shinohara Moriguchi a, Letı´ cia Carnaz a, Luciana Cristina Cunha Bueno Silva a, Luis Ernesto Bueno Salasar b, Rodrigo Luiz Carregaro a, Tatiana de Oliveira Sato a, Helenice Jane Cote Gil Coury c,* a

Department of Physiotherapy, Federal University of S~ ao Carlos, Rodovia Washington Luis, km 235, CP 676, CEP 13565-905, S~ ao Carlos, SP, Brazil b Department of Statistics, Federal University of S~ ao Carlos, Rodovia Washington Luis, km 235, CP 676, CEP 13565-905, S~ ao Carlos, SP, Brazil c Laboratory of Preventive Physiotherapy and Ergonomics (LAFIPE), Department of Physiotherapy, Federal University of S~ ao Carlos, Rodovia Washington Luis, km 235, CP 676, CEP 13565-905, S~ ao Carlos, SP, Brazil Received 3 April 2007; received in revised form 18 February 2008; accepted 13 April 2008

Abstract This study presents data on the intra- and inter-rater reliability of palpation on normal and overweight subjects and shows the influence of palpation discrepancy on angular variability for a collected data set, using computer simulation. Thirty healthy males were recruited. Two physiotherapists identified 12 anatomical landmarks that enabled measurement of eight joint angles. Palpation discrepancy was determined by photographic recordings under ultraviolet light. Angular discrepancies were determined from photos of the subject’s orthostatic posture. A computer simulation was developed to predict expected angular variation according to observed palpation discrepancy. The results showed that the inter-rater reliability was lower than the intra-rater reliability for both palpation and angle measurements. Palpation of the greater trochanter (GT), anterior superior iliac spine (ASIS), seventh cervical vertebra (C7) and femoral epicondyle (FE) showed larger discrepancies. The overweight group presented a significant difference in palpation discrepancy for ASIS (P < 0.03). Angular variations were associated with palpation discrepancies for trunk flexion (TF), hip flexion (HF) and pelvic inclination (PI). Therefore, measurements should be performed by a single rater, rather than by different raters, if reliable angular measurements are intended. Specific anatomical landmarks require careful identification. Simulation was useful for providing estimates of variations due to palpation discrepancy. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Photogrammetry; Reliability; Modeling; Physiotherapy/methods

1. Introduction One crucial point for postural and movement analysis is to accurately identify bony landmarks, since an initial error can be propagated to subsequent measurements. * Corresponding author. Tel.: þ55 16 3351 8634; fax: þ55 16 3361 2081. E-mail address: [email protected] (H.J.C.G. Coury). 1356-689X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.04.002

Palpation errors can affect the quality of measurements that are used to determine disabilities, asymmetries and impairments, and could consequently mislead decisionmaking processes. The therapist’s ability and the subject’s characteristics, such as percentage of body fat, are some of the factors that could influence palpation reliability. Although there have been some reports about palpation reliability, the effects of palpation on angular measurements have not been well explored (Billis et al., 2003;

300

C.S. Moriguchi et al. / Manual Therapy 14 (2009) 299e305

Holgrem and Waling, 2008). France and Nester (2001) evaluated the effect of error in locating the anterior superior iliac spine (ASIS), patellar centre (PC) and tibial tuberosity (TT), on the measured value of the quadriceps (Q) angle. Della Croce et al. (1999) also evaluated anatomical landmark palpation and its effect on joint angle estimation. However, neither pelvic and trunk angles nor neck flexion (NF) were evaluated in these previous studies. Considering the importance of these angles for clinical evaluation, the present study was designed to address these matters. Our initial hypothesis was that palpation variations may differ between body regions and that these variations might affect joint angle measurements. Thus, the objective of the present study was to determine the intra- and inter-rater reliability of anatomical landmark palpation for normal weight and overweight subjects and to investigate the influence of palpation discrepancy on angular variability, by means of computer simulation.

position. Foot marks on the platform guided subject positioning. (b) A wooden device that included a ruler was attached to a digital camera to guide the positioning for the ultraviolet light recordings. This device standardised the distance between the camera lenses and the subject’s body, thereby providing a known measurement. (c) Two splints were manufactured, consisting of two rods fixed perpendicularly to each other (i.e. in an L shape). These splints were attached to the subject’s elbow by means of Velcro straps, to standardise the upper limb position and allow the pelvic and femur markers to be viewed. 2.2.2. Other equipments Digital camera (Sony, MVCeFFD91, 1024  768 pixels resolution); surface markers of 25 mm in diameter; ultraviolet light; fluorescent pen to make marks that are seen only under ultraviolet light (‘‘invisible pen’’). 2.3. Procedures

2. Materials and methods 2.1. Subjects A convenience sample of 30 male university students, between 18 and 30 years old, was recruited for this study. This sample size was calculated using the GraphPad StatMate 2 software. The information needed for the sample size calculation was obtained from pilot tests: standard deviation of the sample, relevant differences to be identified, number of groups and test to be applied. The subjects were recruited according to body mass index (BMI) and formed two groups: normal weight group (n ¼ 15) with BMI ranging from 18.5 to 24.99 kg/m2 and overweight group (n ¼ 15) with BMI greater than or equal to 25 kg/ m2. This cut-off point of 25 kg/m2 for overweight was defined in accordance with recommendations from the World Health Organization (WHO, 2008). The exclusion criteria were (1) BMI less than 18.5 or greater than 34 kg/m2; (2) recent injuries or pain causing hypersensitivity or intolerance to manual palpation on any body part; and (3) balance disorders (positive Romberg test) or dizziness. Subjects within the parameters for the normal and overweight groups who agreed to participate in the study were informed about the objectives and procedures of the research and signed an informed consent form. This study was approved by the institution’s Research Ethics Committee. 2.2. Materials and equipment 2.2.1. Devices built for the study (a) A rotating wooden platform was built, equipped with a roller bearing system to allow smooth rotation. The rotation system could be locked at each 90 rotation

The following bony landmarks were evaluated: fifth metatarsal (5MT), lateral malleolus (LM), lateral femoral epicondyle (FE), greater trochanter (GT), ASIS, TT, PC, seventh cervical vertebra (C7), mastoids (MT), ulnar styloid process (US), lateral humeral epicondyle (HE) and acromion (AC). The structures were identified in a random order. Since the C7 is an odd point, the number of possible comparisons was smaller, in comparison with the bilateral points. Consequently, the angles that involved in this structure were evaluated only twice, either for intra-rater (only the second rater) or inter-rater assessment, avoiding excessive number of markers. Also, in order to balance the subjects within the BMI groups, they were randomly subdivided into two groups, for the intra-rater assessment (first and second raters are defined below) and the inter-rater reliability assessment. Two trained physiotherapists performed the palpation procedures. They were trained according to the recommendations in the literature (Gross et al., 2002; Van Sint Jan and Della Croce, 2005), which were used to standardise the anatomical points and palpation protocol. The training involved pilot tests and discussion of the outcomes until achieving concordance in palpation procedures, and it lasted about 20 h. The physiotherapists fixed the surface markers on the subject’s skin with double-sided tape and demarcated the boundaries with the ‘‘invisible’’ fluorescent pen. Use of this pen prevented identification of any demarcation made by the previous rater on the bony landmarks. This methodology had already been used by Billis et al. (2003). The data collection procedures were subdivided into three phases. In the first phase, one rater performed the palpation bilaterally and attached surface markers

C.S. Moriguchi et al. / Manual Therapy 14 (2009) 299e305

and demarcated marker boundaries using the invisible pen. The subject was then instructed to stand erect on the rotating platform and three photos were taken, showing three body planes (frontal and right and left sagittal planes). In the second phase, the other rater performed the same procedures, but only on one side of the subject’s body. The side evaluated by this second rater (right or left) was randomly selected. Two photos were taken (frontal and sagittal) for angular measurements. Data from the two raters were used for inter-rater comparison for one half-body. In the third phase, the first rater returned to the room and repeated the same procedures on the side opposite to the one evaluated by the second rater. Another two photos were then taken for intra-rater comparison. There were 10-min breaks between the phases. Finally, each anatomical landmark demarcated with the invisible pen was photographed under ultraviolet light. For the next subject, these procedures were repeated with the order of raters inverted. For the photos on the rotating platform, subjects were instructed to adopt a standardised orthostatic posture. The elbow posture was standardised by means of the elbow splints and foot position by the foot markers. The subject did not move between the frontal and sagittal photos, since the examiner turned the platform around and set the orthogonal plane necessary for each evaluation (see Fig. 1). The camera was placed on a tripod, at a height of 0.85 m and at a distance of 2.5 m from the subjects. The photos, which were taken under ultraviolet light, were identified with anatomical landmark subtitles (see Fig. 1C).

301

The joint angles were measured by means of photogrammetry, in accordance with Whistance et al. (1995). The measured angles were plantar flexion (PF), knee flexion (KF), hip flexion (HF), pelvic inclination (PI), trunk inclination (TI), trunk flexion (TF) and neck flexion (NF). The quadriceps angle (Q) was also measured (France and Nester, 2001). 2.4. Data analysis The angular and linear measurements were made using AutoCADÒ software (version 2000). The linear measurements were obtained from ultraviolet photos of each anatomical landmark. The marker boundaries were circled using the pen, and their centres were automatically identified by an AutoCAD command. The distance between the two centres provided the palpation discrepancy, in millimetres. The data were analysed descriptively using the SPSS software (version 10.0), and the significance was set at 5% (P < 0.05). Since the data did not present normal distribution, nonparametric statistics were used to test differences between the normal and overweight groups (ManneWhitney test). Multiple linear regression analysis was carried out to investigate the influence of palpation discrepancies and the group effect on angular variation. A model with angular variation as the response was fitted. The palpation discrepancies and groups (normal weight and overweight) were taken to be covariates. Previously, we had tested the model assumptions of linearity, normality, homocedasticity and

Fig. 1. Reflective markers on anatomical landmarks. (A) Frontal plane; (B) left sagittal plane; and (C) ‘‘invisible’’ pen demarcation under ultraviolet light.

302

C.S. Moriguchi et al. / Manual Therapy 14 (2009) 299e305

independence of errors, using the ShapiroeWilk, GoldfeldeQuandt and DurbineWatson tests, respectively. If the P-value from all the tests was greater than the significance level adopted (0.05), the assumptions were accepted. Otherwise, we tried a response transformation using square roots. If the assumptions were accepted, the significance of the covariates was tested using an F test (ANOVA). No automatic covariate selection procedure, like the stepwise method, was used. All relevant covariates were included in the regression model using the enter method. This analysis was carried out using the R statistical software (http:// www.R-project.org). 2.4.1. Computational simulation to evaluate the influence of palpation discrepancy on angular variation A computer simulation was developed using the MatLab software (version 7.0.1, MathWorks Inc., Natick, MA, USA), to predict the maximum expected angular variation according to the observed palpation discrepancy. In this model, three points (P1, P2 and P3) were required. These points represented anatomical landmarks and formed an angle in which the vertex was at P2. To perform the simulation, the following information was needed: distances from P1 to P2 and from P2 to P3; palpation discrepancies for each anatomical landmark; and an estimated measurement for the angle. All these required measurements were obtained by photogrammetry for each subject in the sample. Comparisons between the measured distances and anthropometric data available showed similar values (Winter, 1990). Around each P1, P2 and P3, the simulation generated 10,000 points that were randomly distributed inside a circumference of radius equal to the observed palpation discrepancy at the respective point. For each three points generated around P1, P2 and P3, the algorithm calculated an angle. Thus, 10,000 angular measurements were obtained and, from this sample, the maximum angular variation due to palpation discrepancy was estimated. Following this, the estimated (simulation) and observed (photogrammetry) angular variations were compared. For the inter-rater comparisons, which had larger amounts of data, the observed and estimated angular variations were correlated using Pearson’s coefficient.

3. Results Two groups of 15 subjects each took part in the study. The normal weight group presented a mean age of 24.0  3.3 years, mean height of 1.72  0.07 m and mean weight of 66.9  7.6 kg. The overweight group presented a mean age of 23.8  3.1 years, mean height of 1.76  0.06 m and mean weight of 88.7  9.5 kg.

The inter-rater palpation discrepancy was greater than the intra-rater discrepancy. Some anatomical landmarks presented greater discrepancies, such as the GT, FE, ASIS, C7, MT and AC. Other anatomical landmarks, like the HE, US, PC, TT, LM and 5MT showed small discrepancies (Fig. 2). The intra-rater comparisons for the second rater and the inter-rater comparisons showed significantly larger palpation discrepancies for overweight subjects in relation to normal weight subjects only for the ASIS landmark (P ¼ 0.03 and 0.001, respectively). The angular variability was also higher for the interrater than for the intra-rater comparison. Some joint angles, such as PF, KF, TI and NF, showed variations of less than 10 for both the intra-rater and the inter-rater comparisons. Larger variations (greater than 10 ) were found for the HF, PI, Q and TF angles. There were no significant differences between the normal weight and overweight groups for the measured angles (Fig. 3).

intrarater (No. 1) acromion humeral epicondyle styloid process mastoids 7th vertebra patella centre tibial tuberosity ASIS greater trochanter femoral epicondyle lateral malleolus 5th metatarsal

normal weight overweight 0

5

10

15

20

25

30

35

40

30

35

40

intrarater (No. 2) acromion humeral epicondyle styloid process mastoids 7th vertebra patella centre tibial tuberosity ASIS greater trochanter femoral epicondyle lateral malleolus 5th metatarsal 0

5

10

15

20

25

interrater acromion humeral epicondyle styloid process mastoids 7th vertebra patella centre tibial tuberosity ASIS greater trochanter femoral epicondyle lateral malleolus 5th metatarsal 0

5

10

15

20

25

30

35

40

palpation variation (mm) Fig. 2. Means and standard deviations for palpation discrepancies in normal weight and overweight groups, for the intra-rater comparisons for both raters and for the inter-rater comparisons.

303

C.S. Moriguchi et al. / Manual Therapy 14 (2009) 299e305

intrarater (No. 1) neck flexion

normal weight overweight

trunk flexion trunk inclination quadriceps angle pelvic inclination hip flexion knee flexion plantar flexion 0

5

10

15

20

25

20

25

intrarater (No. 2) neck flexion trunk flexion trunk inclination quadriceps angle pelvic inclination hip flexion

The multiple linear regression results showed that the angular variations in TF, HF and PI were influenced by palpation discrepancy at the ASIS and GT. The adjusted R2 were 0.55, 0.49 and 0.46, respectively. Table 1 compares the observed and estimated angular variations. In general, the estimated variation was higher than the observed variation, except for three angles for the intra-rater assessment of the second rater (PF, KF and HF) and two angles for the inter-rater assessment (PF and KF). The estimated angles derived from the maximum palpation discrepancy showed small variations for PF, KF and TI (within 5 for intra-rater and 10 for inter-rater comparisons). The Pearson results showed the most extreme correlations for the Q (0.02) and PI (0.76) angles. The quadriceps angle results showed no linear relationship between the observed and estimated angular variations, while the PI variations showed a linear relationship between the observed and estimated variations. A linear relationship was also identified for the HF and TF angles (not illustrated).

knee flexion plantar flexion 0

5

10

15

interrater neck flexion trunk flexion trunk inclination quadriceps angle pelvic inclination hip flexion knee flexion plantar flexion 0

5

10

15

20

25

angle variation (°) Fig. 3. Means and standard deviations for angular variations for normal weight and overweight groups, for the intra-rater comparisons for both raters and for the inter-rater comparisons.

4. Discussion The intra-rater reliability was higher than the interrater reliability, for both the palpation and angular variations. This indicates that discrimination of bony landmarks could be prone to misinterpretation by different raters. Similar results were also reported by Della Croce et al. (1999) and Billis et al. (2003), thus suggesting that reliable palpation measurements can be better achieved when procedures are performed by the same examiner. Some anatomical landmarks showed larger discrepancies during palpation, for both intra- and inter-rater comparisons. Several characteristics may explain this variability, such as the location of the structure in relation to the skin surface, and the size and morphology of the structures (Della Croce et al., 1999; Lewis et al., 2002; Holgrem and Waling, 2008). Della Croce

Table 1 Comparison between observed and estimated angular variations ( ). Angle ( )

Intra-rater

Inter-rater

Rater No. 1

PF KF HF PI Q TI TF NF

Rater No. 2

Observed

Estimated

Observed

Estimated

1.77 1.43 3.73 3.93 5.13 e e e

2.40 1.47 4.32 4.82 6.75 e e e

2.80 2.53 6.27 4.80 4.47 0.85 4.69 2.54

2.50 1.70 5.15 5.42 6.84 0.89 5.40 4.10

Observed

Estimated

4.30 2.68 10.03 8.13 5.57 1.33 6.47 4.67

4.03 2.50 10.64 11.62 9.21 2.47 11.89 8.70

304

C.S. Moriguchi et al. / Manual Therapy 14 (2009) 299e305

et al. (1999) pointed out that anatomical landmarks are often not clearly identifiable discrete points but, rather, relatively large and curved areas. Moreover, morphological features (such as surface irregularities) may also affect palpation reliability. Thus, these characteristics could explain the large variability found in the GT, ASIS, FE, PC and AC. Other anatomical landmarks require careful identification. The patella is prone to errors due to the bulky insertion of the quadriceps. The TT varies in shape and prominence, and may be perceived as not having a natural centre (France and Nester, 2001). The seventh cervical spinal process was another structure that is difficult to identify, since there are no clearly described procedures. Some authors have reported that C7 is the most prominent vertebra, while others have indicated that this structure is the last vertebra that moves during NF/extension (Gross et al., 2002). Thus, hypomobility of the cervical column and morphological differences in spinous processes may mislead C7 identification. Significant differences between the normal and overweight groups were found only for ASIS, probably due to abdominal fat. In fact, the lack of influence of BMI on palpation discrepancy has been identified in other study (Harlick et al., 2007), which confirms that the BMI was not a sufficient criterion for differentiating between the groups in relation to palpation difficulty in our study. Similarly, Kushner and Blatner (2005) reported that the BMI did not allow distinction between the composition of lean and fat tissue and therefore could lead to erroneous interpretations. The regression analysis reported here showed a relationship between palpation discrepancy and angular variability for some angles. Della Croce et al. (2005) also demonstrated that the reliability of joint kinematics is dependent on the precision of the palpation procedures. For most of the results reported in the present study, the observed variations were lower than the ‘‘predicted’’ variations. The standardised procedures established seem to have successfully controlled for some photogrammetric sources of error. Only for the lower limb angles (PF, KF and HF) the estimated variation was smaller than the observed variation. In this case, some photometric errors might be added to the palpation variations. Perhaps, the lower limb position could change from one photo to another because of postural adjustments, e.g., oscillation during quiet standing, caused by tibialis anterior and gastrocnemius muscle activity, and on a minor scale, rectus femoris and semitendinosus activity (Madigan et al., 2006). The observed and estimated angular variations showed a linear relationship and the Pearson correlation coefficient ranged from moderate to high for PI (r from 0.64 to 0.76; P < 0.05). For the Q angle, no linear relationship was found (r from 0.02 to 0.21; P > 0.05). In this case, after inspection of the photos, it can be supposed that the PC

position might be an extra source of variation. Since the Q angle is highly dependent on the PC, i.e. the angular vertex, and this bone is embedded in the quadriceps tendon (sesamoid bone), the quadriceps muscle might have presented a different level of contraction when the photos were taken, thus causing patella movement and possible displacement of the marker. Also, the subjects may have had some perception of instability when standing on the rotating platform, which could explain the occurrence of this muscle contraction. Hence, the quadriceps contraction level and surface stability also have to be controlled in order to achieve precise measurements of the Q angle. France and Nester (2001) also investigated the effect of errors in identifying the ASIS, PC and TT on the Q angle, by means of data manipulation. In the present study, the simulation developed the initial idea proposed by France and Nester (2001), in as much as the distance between the anatomical landmarks and the palpation discrepancy associated with these points were taken into consideration. Another advance in the present simulation was the consideration of a very large number of angles (10,000) that could be formed by different placements of landmarks as a result of palpation discrepancy. This leads to more accurate estimation of the influence of palpation discrepancy on angular measurements. The small variations identified for some angles, such as PF, KF and TI, suggest that the calculations of these angles are reliable and repeatable, since they are based on more prominent and easily identified bony landmarks. One limitation to this study was the small number of raters. Della Croce et al. (1999) tested the reliability of six raters, while Billis et al. (2003) investigated 30 physiotherapists with different levels of training and experience. Increasing the number of raters could lead to a better estimate of the palpation discrepancy. Furthermore, these data refer to healthy subjects and cannot be generalised to patients (Harlick et al., 2007). Future studies should be conducted specifically to test possible differences between healthy subjects and subjects with impairments or particular disabilities.

5. Conclusion The intra-rater reliability was higher than the interrater reliability, thus indicating that measurements based on palpation by a single trained rater are advisable. Different palpation discrepancies between the anatomical landmarks suggest that rater training should give special attention to some specific structures. Since palpation discrepancy can alter the measured joint angles, the procedures should be conducted carefully in order to achieve reliable measurements for postural and motion analysis purposes. Finally, the computer simulation was shown to be a useful procedure for estimating the maximum variation due to palpation.

C.S. Moriguchi et al. / Manual Therapy 14 (2009) 299e305

References 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):223e32. Della Croce U, Capozzo A, Kerrigan DC. Pelvis and lower limb anatomical landmark calibration precision and its propagation to bone geometry and joint angles. Medical and Biological Engineering and Computing 1999;37:155e61. Della Croce U, Leardini A, Chiari L, Cappozzo A. Human movement analysis using stereophotogrammetry. Part 4: assessment of anatomical landmark misplacement and its effects on joint kinematics. Gait and Posture 2005;21(2):226e37. France L, Nester C. Effect of errors in the identification of anatomical landmarks on the accuracy of Q angle values. Clinical Biomechanics 2001;16(8):710e3. Gross JM, Fetto J, Rosen E. The cervical spine and thoracic spine. In: Gross JM, Fetto J, Rosen E, editors. Musculoskeletal examination. 2nd ed. Edinburgh: Blackwell Publishing; 2002. p. 43 [chapter 4]. Harlick JC, Milosavljevic S, Milburn PD. Palpation identification of spinous processes in the lumbar spine. Manual Therapy 2007;12(1):56e62.

305

Holgrem U, Waling K. Inter-examiner reliability of four static palpation tests used for assessing pelvic dysfunction. Manual Therapy 2008;13(1):50e6. Kushner RF, Blatner DJ. Risk assessment of the overweight and obese patient. American Dietetic Association 2005;105(Suppl. 1): S53e62. Lewis J, Green A, Reichard Z, Wright C. Scapular position: the validity of skin surface palpation. Manual Therapy 2002;7(1): 26e30. Madigan ML, Davidson BS, Nussbaum MA. Postural sway and joint kinematics during quiet standing are affected by lumbar extensor fatigue. Human Movement Science 2006;25:788e99. Van Sint Jan S, Della Croce U. Identifying the location of human skeletal landmarks: why standardized definitions are necessary e a proposal. Clinical Biomechanics 2005;20(6):659e60. Winter DA. Anthropometry. In: Winter DA, editor. Biomechanics and motor control of human movement. 2nd ed. New York: John Wiley; 1990. p. 52 [chapter 3]. Whistance RS, Adams LP, Van Geems BA, Bridger RS. Postural adaptations to workbench modifications in standing workers. Ergonomics 1995;38(12):2485e503. World Health Organization (WHO). Available from: http:// www.who.int/bmi/index.jsp?introPage¼intro_3. html; 2008.

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 306e313 www.elsevier.com/math

Original Article

Inclusion of thoracic spine thrust manipulation into an electro-therapy/thermal program for the management of patients with acute mechanical neck pain: A randomized clinical trial Javier Gonza´lez-Iglesias a, Cesar Ferna´ndez-de-las-Pen˜as a,b,*, Joshua A. Cleland c,d,e, Francisco Alburquerque-Sendı´ n a,f, Luis Palomeque-del-Cerro a, Roberto Me´ndez-Sa´nchez a,f a

Escuela de Osteopatı´a de Madrid, Madrid, Spain Department of Physical Therapy, Occupational Therapy, Physical Medicine and Rehabilitation of Universidad Rey Juan Carlos, Alcorco´n, Madrid, Spain c Department of Physical Therapy, Franklin Pierce College, Concord, NH, USA d Physical Therapist, Rehabilitation Services, Concord Hospital, Concord, NH, USA e Faculty, Manual Therapy Fellowship Program, Regis University, Denver, CO, USA f Department of Physical Therapy, Universidad de Salamanca, Salamanca, Spain b

Received 18 July 2007; received in revised form 13 March 2008; accepted 11 April 2008

Abstract Our aim was to examine the effects of a seated thoracic spine distraction thrust manipulation included in an electrotherapy/thermal program on pain, disability, and cervical range of motion in patients with acute neck pain. This randomized controlled trial included 45 patients (20 males, 25 females) between 23 and 44 years of age presenting with acute neck pain. Patients were randomly divided into 2 groups: an experimental group which received a thoracic manipulation, and a control group which did not receive the manipulative procedure. Both groups received an electrotherapy program consisting of 6 sessions of TENS (frequency 100 Hz; 20 min), superficial thermotherapy (15 min) and soft tissue massage. The experimental group also received a thoracic manipulation once a week for 3 consecutive weeks. Outcome measures included neck pain (numerical pain rate scale; NPRS), level of disability (Northwick Park Neck Pain Questionnaire; NPQ) and neck mobility. These outcomes were assessed at baseline and 1 week after discharge. A 2-way repeated-measures ANOVA with group as between-subject variable and time as within-subject variable was used. Patients receiving thoracic manipulation experienced greater reductions in both neck pain, with between-group difference of 2.3 (95% CI 2e2.7) points on a 11-NPRS, and perceived disability with between-group differences 8.5 (95% CI 7.2e9.8) points. Further, patients receiving thoracic manipulation experienced greater increases in all cervical motions with between-group differences of 10.6 (95% CI 8.8e12.5 ) for flexion; 9.9 (95% CI 8.1e11.7 ) for extension; 9.5 (95% CI 7.6e11.4 ) for right lateral-flexion; 8 (95% CI 6.2e9.8 ) for left lateral-flexion; 9.6 (95% CI 7.7e11.6 ) for right rotation; and 8.4 (95% CI 6.5e10.3 ) for left rotation. We found that the inclusion of a thoracic manipulation into an electrotherapy/ thermal program was effective in reducing neck pain and disability, and in increasing active cervical mobility in patients with acute neck pain. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Neck pain; Spinal manipulation; Thoracic spine; Electrotherapy

* Corresponding author. Facultad de Ciencias de la Salud, Universidad Rey Juan Carlos, Avenida de Atenas s/n, 28922 Alcorco´n, Madrid, Spain. Tel.: þ34 91 488 88 84; fax: þ34 91 488 89 57. E-mail address: [email protected] (C. Ferna´ndez-de-las-Pen˜as). 1356-689X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.04.006

J. Gonza´lez-Iglesias et al. / Manual Therapy 14 (2009) 306e313

1. Introduction Approximately 25% of all outpatient physical therapy visits consist of patients with symptoms involving the neck region (Jette et al., 1994). It has been found that nearly half of the individuals with neck pain will experience debilitating symptoms (Gummesson et al., 2006). Over a third of patients with neck pain will develop chronic symptoms lasting more than 6 months (Cote et al., 2004), and nearly a third who experience a first time onset of neck pain will continue to report continued healthcare utilization for their symptoms at a 10-year follow-up (Enthoven et al., 2004). Physical therapists utilize a number of interventions in the management of neck pain including joint manipulation (non-thrust and thrust), exercises, massage, thermo-therapy or electrotherapy (American Physical Therapy Association, 2001). However, robust evidence to support the use of many of these therapeutic strategies for neck pain is lacking (Kjellman et al., 1999; Gross et al., 2000; Hoving et al., 2001). The Philadelphia Panel Clinical Practice Guidelines concluded that many commonly used interventions for patients with neck pain lack sufficient evidence to justify their clinical use (Brosseau et al., 2001). Recently, evidence has begun to emerge for the use of manual procedures directed at the thoracic spine for patients with mechanical neck pain (Cleland et al., 2005, 2007a,b; Ferna´ndez-de-lasPen˜as et al., 2004, 2007a). Cleland et al. (2005) found that thoracic thrust manipulation results in immediate improvements in neck pain at rest as measured by the visual analogue scale, compared to patients receiving a placebo manipulation. Further, it has also been found that at short-term follow-up patients receiving thoracic manipulation exhibit superior outcomes to patients receiving non-thrust techniques (Cleland et al., 2007a). The importance of investigating the effectiveness of thoracic spinal manipulation is necessary considering the fact that the thoracic spine is the region of the spine most often manipulated, despite the fact that more patients complain of neck pain (Adams and Sim, 1998). Further, decreased mobility in the thoracic spine has been shown to be related to the presence of neck pain symptoms (Norlander et al., 1996, 1997; Norlander and Nordgren, 1998), so it is possible that manipulation of the thoracic spine may alter the biomechanics of the cervical region and decrease mechanical stress. Finally, it has previously been identified that either cervical mobilization (Vicenzino et al., 2001) or manipulation (Ferna´ndez-de-las-Pen˜as et al., 2007b) induces an activation of descending inhibitory mechanisms; hence, thoracic spine thrust manipulations may also result in a reduction of neck symptoms. It should be noted that the aforementioned studies solely investigated the effects of thoracic thrust manipulation (with the exception of one study which used range of motion exercise). More often physical therapists use

307

a multi-modal treatment approach (exercise, manual therapy, electro-therapy, etc.) in the management of neck pain which may include thrust techniques directed at the thoracic spine. To date only one study has investigated the effects of thoracic spine manipulation incorporated into a physical therapy management program. Fernandez-de-las-Pen˜as et al. (2004) reported that patients with whiplash-associated disorders receiving thoracic thrust manipulation as a component of a physical therapy program experienced a greater reduction in symptoms than subjects whose physical therapy did not include manipulation. To date no studies have explicitly investigated the effects of thoracic manipulation when it is added to a program including electro-therapy and thermal agents in patients with mechanical neck pain. Hence, the purpose of this study was to examine the effects of a seated thoracic distraction manipulation when added to a program including electrotherapy/thermal modalities on neck pain, disability, and cervical mobility.

2. Materials and methods 2.1. Subjects Forty-five patients, 20 males and 25 females, between 23 and 44 years of age (mean 34; SD 4 years) with acute mechanical neck pain referred by their primary care physician to a physical therapy clinic participated in this study. For the purpose of this study mechanical neck pain was defined as generalized neck or shoulder pain with mechanical characteristics (including symptoms provoked by neck postures, neck movement, or palpation of the cervical musculature) of less than 1 month in duration. Exclusion criteria included the following: (1) contra-indication to manipulation; (2) history of whiplash or cervical surgery; (3) diagnosis of cervical radiculopathy or myelopathy; (4) diagnosis of fibromyalgia syndrome (Wolfe et al., 1990); (5) having undergone spinal manipulative therapy in the previous 2 months; or (6) less than 18 or greater than 45 years of age. The patient history for each patient was solicited from their primary care physician to assess the presence of any exclusion criteria or ‘‘red flags’’ (e.g. infection, osteoporosis). This study was supervised by the Escuela de Osteopatı´ a de Madrid (EOM).The research project was approved by the local human research committee (EOM). All subjects signed an informed consent prior to participation in the study. 2.2. Procedure Patients completed self-report measures and received a standardized history and physical examination by an

308

J. Gonza´lez-Iglesias et al. / Manual Therapy 14 (2009) 306e313

experienced manual physical therapist. Demographic data included age, gender, past medical history, location, nature and onset of symptoms. At the first visit, patients reported their level of neck pain and completed the Northwick Park Neck Pain Questionnaire (NPQ). Cervical mobility was assessed by an assessor blinded to the treatment allocation of the patients. 2.3. Outcome measures Neck pain was assessed with an 11-point numerical pain rate scale (Jensen et al., 1999) (NPRS; range 0 ¼ no pain, to 10 ¼ maximum pain). The Spanish version of the NPQ was used to assess subjects’ perceived level of disability as a result of their neck pain (Gonzalez et al., 2001). The NPQ is a self-administered questionnaire that includes 9 sections on typical daily activities that may be affected by the patient’s neck pain: intensity, sleeping, numbness, duration, reading, television, carrying, work, social role, and driving. Each section is scored on a scale from 0 to 4, with 4 representing the greatest disability, and the total score is obtained by summing the scores for the 9 sections (possible score 0e36) (Leak et al., 1994). Cervical range of motion was assessed with a cervical goniometer (Performance Attainment Associates), which has been shown to be a reliable method of measurement (Jordan, 2000) with an intra-tester reliability (ICC) ranging from 0.7 to 0.9, and an inter-tester reliability ranging from 0.8 to 0.87 (Peolsson et al., 2000). Neck mobility was assessed in a relaxed sitting position. All subjects were asked to sit comfortably on a chair with both feet flat on the floor, hips and knees positioned at 90 angles, and buttocks positioned against the back of the chair. The goniometer was placed on the top of the head. Once the goniometer was set in the neutral position, the patient was asked to move the head as far as possible in a standard fashion: forwards (flexion), backwards (extension), right lateral-flexion, left lateral-flexion, right rotation, and left rotation. Three trials were recorded for each type of movement, and the mean was employed in the analysis. This method of assessment has been described elsewhere in detail (Ferna´ndez-de-las-Pen˜as et al., 2006). Outcome measures were captured at baseline and 1 week after discharge form physical therapy by an assessor blind to group assignment. One week after discharge of the last session, patients again completed the 11-point NPRS and the NPQ. 2.4. Allocation Following the baseline examination, patients were randomly assigned to receive the electrotherapy/thermal program with or without thoracic spine manipulation. Concealed allocation was performed by using a computer-generated randomized table of numbers created

prior to the beginning of the study. Individual, sequentially numbered index cards with the random assignment were prepared. The index cards were folded and placed in sealed opaque envelopes. 2.5. Treatment Both groups received 6 sessions of a standard electrotherapy/thermal program during 3 consecutive weeks. The thoracic thrust manipulation was applied once per week for the 3 consecutive weeks and only in the experimental group. Patients were blinded to the treatment allocation group, without revealing that the inclusion of a specific intervention (thoracic spine manipulation) was being evaluated. The adequacy of subject blinding was assessed by a post-questionnaire. 2.6. Electrotherapy/thermal program There is no consensus regarding the most effective program for the management of acute mechanical neck pain (Brosseau et al., 2001). A Cochrane Review found that the evidence for treatment of neck pain by different forms of electrotherapy is either lacking or conflicting (Kroeling et al., 2005). Nevertheless, Chiu et al. (2005) found that the application of transcutaneous electrical nerve stimulation (TENS) combined with other physical approaches was effective for improving neck muscle strength, neck pain and perceived disability. In the present study, the standardized program included the application of superficial thermo-therapy and electro-therapy as follows: an infrared lamp (Philips System, 250 W), located 50 cm distant from the patient’s neck, was applied for 15 min. After superficial thermotherapy, TENS (Uniphy phyaction 782) with a frequency of 100 Hz and 250 ms stimulation was applied for 20 min using two 4  6 cm electrodes placed bilaterally to the spinous process of C7 vertebra. 2.7. Thoracic spine thrust manipulation Patients in the experimental group received a seated thoracic spine ‘‘distraction’’ manipulation once per week for 3 consecutive weeks as follows. The patient was seated with the arms crossed over the chest and hands passed over the shoulders. The therapist placed his or her upper chest at the level of the patient’s middle thoracic spine and grasped the patient’s elbows. Gentle flexion of the thoracic spine was introduced until slight tension was felt in the tissues at the contact point between the therapist’s chest and patient’s back. Then, a distraction thrust manipulation in an upward direction was applied (Fig. 1) (Gibbons and Tehan, 2000). Since the manipulation was applied when a tension was felt in the tissue at the contact point, cavitation often occurred at the end of range (motion barrier). During

309

J. Gonza´lez-Iglesias et al. / Manual Therapy 14 (2009) 306e313

assessed by means of the KolmogoroveSmirnov test (P > 0.05). Baseline features were compared between groups using independent t-tests for continuous data, and c2 tests of independence for categorical data. A 2-way repeated-measures analysis of variance (‘‘ANOV A’’) with group (experimental or control) as betweensubject variable and time (preepost test measurements) as within-subject variable was used to analyse the effects of the interventions. Separate ANOVAs were performed with pain (NPRS), disability (NPQ) and neck mobility as the dependent variables. We used intention-to-treat analysis with subjects analysed in the group to which they were allocated. Between-group effect sizes were calculated using the Cohen d coefficient (Cohen, 1988). An effect size greater than 0.8 was considered large, around 0.5 was moderate and less than 0.2 was small. The statistical analyses were conducted at a 95% confidence level. A P value less than 0.05 was considered as significant.

3. Results

Fig. 1. Thoracic spine thrust manipulation.

the manipulation the therapist listened for a cracking or popping sound. If no popping was heard on the first attempt, the therapist repositioned the patient, and performed a second manipulation. A maximum of 2 attempts were performed on each patient. This procedure was the same as that employed in previous studies addressing the effectiveness of thoracic spine thrust manipulation in patients with mechanical neck pain (Cleland et al., 2007a,b).

The total number of subjects screened, reasons for ineligibility and drop out can be seen in Fig. 2. Twentythree patients, 10 men and 13 women, age 23 to 42 (mean age: 34  5 years) were assigned to the experimental group; and 22 patients, 10 men and 12 women, age

73 patients with neck pain screened for eligibility criteria

Excluded (n=28) Older than 45 years (n=11) Previous whiplash (n=7) Chronic neck pain (n=8) Decline to participate (n=2)

2.8. Sample size determination The sample size and power calculations were performed using Spanish software (Taman˜o de la Muestra, 1.1Ó). The calculations were based on detecting differences of 1.13 units in a 11 numerical pain rate scale at post-data, assuming a standard deviation of 0.69 (data taken from Cleland et al., 2005), a 2-tailed test, an alpha level of 0.05, and a desired power of 80%. These assumptions generated a sample size of 20 subjects per group. 2.9. Statistical analysis Data was analysed with the SPSS package (version 13.0). A normal distribution of quantitative data was

Randomized (n=45)

Allocated to intervention Thoracic Spine Manipulation (n=23)

Allocated to intervention No Manipulation (n=22)

Loss to follow-up (n=0)

Loss to follow-up (n=0)

Analysed (n=23)

Analysed (n=22)

Fig. 2. Flow diagram of subject recruitment throughout the course of the study.

310

J. Gonza´lez-Iglesias et al. / Manual Therapy 14 (2009) 306e313

24e44 (mean 34  6 years) formed the control group. No significant differences were found for gender (P ¼ 0.7), age (P ¼ 0.9), neck pain intensity (P ¼ 0.4), perceived disability (P ¼ 0.4), cervical range of motion (P > 0.1), or duration of symptoms (P ¼ 0.5) between groups, so both groups were comparable in all respects at the start of the study. Baseline data of each group are detailed in Table 1. 3.1. Neck pain and perceived disability assessment The 2-way repeated-measures ANOVA found a significant group  time interaction for both pain intensity (F ¼ 28.3; P < 0.001) and disability (F ¼ 28.3; P < 0.001). Subjects receiving thoracic spine manipulation experienced greater reductions in both neck pain, with a between-group difference of 2.3 points (95% CI 2e2.7) on the NPRS, and disability, with between-group differences of 8.5 points (95% CI 7.2e9.8) in the NPQ. Further, large between-group effect sizes (d > 1) were found for both NPRS vs. time (d ¼ 1.8) and NPQ vs. time (d ¼ 1.75) in favour of the experimental group. 3.2. Active cervical range of motion assessment The analysis of variance found a significant group  time interaction for all cervical motions: flexion (F ¼ 45.4; P < 0.001); extension (F ¼ 66.4; P < 0.001); right (F ¼ 39.5; P < 0.001) and left (F ¼ 27.2; P < 0.001) lateral-flexion; right (F ¼ 28.9; P < 0.001) and left (F ¼ 22.2; P < 0.001) rotation. Patients receiving thoracic thrust manipulation experienced greater increases in all cervical motions with between-group differences of 10.6 for flexion (95% CI 8.8e12.5 ); 9.9 for extension (95% CI 8.1e11.7 ); 9.5 for right lateral-flexion (95% CI 7.6e11.4 ); 8 for left lateral-flexion (95% CI 6.2e9.8 ); 9.6 for right rotation (95% CI 7.7e11.6 ); and 8.4 for left rotation (95% CI 6.5e10.3 ). Again, large between-group effect sizes were also found for all

Table 1 Demographic features of both groups at the beginning of the study.

Gender (male/female) Age (years) Duration of symptoms (days) Neck pain at rest (NPRS) Cervical flexion (degrees) Cervical extension (degrees) Left lateral-flexion (degrees) Right lateral-flexion (degrees) Left rotation (degrees) Right rotation (degrees) Northwick value (degrees)

Control group

Experimental group

10/12 34  6 17  5 53.6  6.3 44.7  5.3 58.8  5.6 40.2  4.5 39.4  4.9 57.8  5.4 56.1  6.6 27.1  2.7

10/13 34  5 18  6 55.6  8.7 45.6  4.3 59.1  8.1 39.1  4.6 36.2  5.1 59.2  6.4 55.8  7.3 27.8  3.1

Values are expressed as mean  standard deviation.

neck movements (1.5 < d < 1.7) in favour of the thoracic spine manipulation group. Table 2 summarizes preepost intervention data for each outcome in both groups, and Table 3 shows the comparison of pree post changes in either group. Finally, in the post-study questionnaire, none of the participants accurately reported which group they believed they were allocated to.

4. Discussion The results of our study demonstrated that patients with acute mechanical neck pain receiving an electrotherapy/thermal program plus thoracic thrust manipulation experienced a significantly greater reduction in pain and disability as well as an increase in cervical mobility compared to a group that received electrotherapy/thermal only. The effect sizes were large for all of the dependent variables assessed in favour of the thoracic spine thrust manipulation group. Additionally, it should be noted that between-group differences for pain achieved by the thoracic spine thrust manipulation group was not only statistically significant but also clinically meaningful as it exceeded the minimum clinically important difference (MCID) on the NPRS, identified as 2 points (Childs et al., 2005). Although the MCID for the NPQ has not been reported (Pietrobon et al., 2002), within-group improvements were significantly greater for subjects in the experimental group. The current results further substantiate the findings of previous studies (Cleland et al., 2005; Ferna´ndez-de-lasPen˜as et al., 2004, 2007a), all of which demonstrated that thoracic thrust manipulation resulted in changes in pain, disability and cervical mobility in different populations of patients with neck pain. While the effect sizes in this study were large, they could have potentially been greater if the inclusion criteria had included a specific subgroup of patients who are likely to exhibit a rapid and dramatic improvement from thoracic manipulation (Cleland et al., 2007a). Cleland et al. (2007a) recently developed a clinical prediction rule with 6 variables from patients with mechanical neck pain. This study identified 6 predictor variables (symptom duration <30 days, no symptoms distal to the shoulder, looking up does not aggravate symptoms, Fear-Avoidance Beliefs Physical Activity subscale score <12, decreased upper thoracic spine kyphosis (T3eT5), and cervical extension <30 ). If 3 of the 6 variables were present, the probability of experiencing a successful outcome improved from 54% to 86% (þLR 5.5). In the present study, patients with acute (less than 30 days) neck pain were included, so our patients presented with at least 1 of the predictors identified by Cleland et al. (2007a). The physiological mechanism associated with the benefits of thrust manipulation is beyond the scope of

311

J. Gonza´lez-Iglesias et al. / Manual Therapy 14 (2009) 306e313 Table 2 Within preepost values of both groups for each outcome measure. Control group

Neck pain at rest Cervical flexion Cervical extension Left lateral-flexion Right lateral-flexion Left rotation Right rotation Northwick value

Experimental group

Pre-intervention

Post-intervention

P value

Pre-intervention

Post-intervention

P value

5.37 44.7 58.8 40.2 39.4 57.8 56.1 27.1

4.3 45.6 60.1 41.6 41.0 58.4 56.3 22.9

<0.001 NS <0.01 <0.01 <0.001 NS NS <0.001

5.6 45.6 59.1 39.1 36.2 59.2 55.8 27.8

2.3 57.2 70.3 48.5 47.2 68.2 65.6 15.2

<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001

(0.6) (5.3) (5.6) (4.5) (4.9) (5.4) (6.6) (2.7)

(0.8) (5.4) (4.8) (4.4) (4.2) (5.1) (5.9) (2.9)

(0.9) (4.3) (8.2) (4.6) (5.1) (6.4) (7.4) (3.1)

(1) (5.8) (8.6) (5.5) (5.6) (6.3) (7.8) (4.1)

Values are expressed as mean  standard deviation; NS, non significant (P > 0.05).

the present study and remains to be fully elucidated. Further, both biomechanical and neuro-physiological (either segmental or central) mechanisms have been suggested. For instance, the biomechanical link between the cervico-thoracic spine and neck pain described by Norlander et al. (1996, 1997) may be one reason why thoracic spine manipulation is beneficial for patients with neck pain. It is also possible that spinal manipulative therapy has inherent qualities that can alter the biomechanics of the treated region (thoracic spine), and it is likely that those segments are bio-mechanically related to the cervical region. One mechanism could be that the manipulative procedure may induce a reflex inhibition of pain or reflex muscle relaxation by modifying the discharge of proprioceptive group I and II afferents (Pickar, 1999). It is also plausible that thrust manipulation decreases pain and spasm while increasing mobility through changes in muscle electrical activity (Shambaugh, 1995); reduced muscle spasm (Johansson and Sojka, 1991) or increased inter-segmental joint play subsequent to a spinal manipulation (Cassidy et al., 1992; Norlander et al., 1997, 1998). Further, mechanical stimulus induced by the manipulative procedure may also alter concentrations of inflammatory mediators (Sambajon et al., 2003), or trigger segmental inhibitory mechanisms (Wall, 2006). Finally, activation of descending inhibitory pathways may explain the decreased cervical symptoms after the application of a manipulation in

another region (Ferna´ndez-de-las-Pen˜as et al., 2007b). Nevertheless, it seems that more than 1 mechanism likely explains the effects of spinal manipulative therapy (Pickar, 2002), and there is insufficient evidence to claim a major role for either peripheral or central mechanisms. Future research is clearly necessary to determine if mechanisms by which manipulation exerts its effects are either mechanical or neuro-physiologic or both. Traditional manual therapy philosophies have focused on using a biomechanical approach to assessing joint dysfunction followed by treatment based on biomechanical theoretical constructs (Jull and Moore, 2002). It is often believed that manual therapists must accurately identify a segmental impairment through careful palpation of vertebral movement or alignment and, once identified, treat the particular impairment by applying a specific amount of force to a single segment in a specific direction. However, all patients in the experimental group received the identical thrust manipulation regardless of the clinical presentation. While this may seem counter-intuitive to some philosophies, based on the lack of evidence to support the use of the biomechanical models, and substandard levels of reliability with palpation techniques (Cleland et al., 2006), we selected to deliver 1 specific technique to all patients. Based on our results, we cannot say whether patients treated with a particular thrust technique selected by the physical therapist would have had better outcomes. However,

Table 3 Inter-group comparison of the changes (preepost scores) between both groups.

Neck pain at rest Cervical flexion Cervical extension Left lateral-flexion Right lateral-flexion Left rotation Right rotation Northwick value

Pre-post values of the control group (%; CI)

Pre-post values of the experimental group (%; CI)

Values of F and P

9.4 0.9 1.2 1.4 1.5 0.6 0.2 4.1

32.8 11.6 11.2 9.5 11.1 8.9 9.8 12.6

F ¼ 183.1; P < 0.001 F ¼ 135.4; P < 0.001 F ¼ 126.4; P < 0.001 F ¼ 77.2; P < 0.001 F ¼ 99.4; P < 0.001 F ¼ 82.2; P < 0.001 F ¼ 98.9; P < 0.001 F ¼ 166.1; P < 0.001

(95; (95; (95; (95; (95; (95; (95; (95;

7.2e11.4) 0.2e1.9) 0.4e2.2) 0.5e2.4) 0.5e2.6) 0.5e1,6) 1.3e1.6) 3.4e4.8)

(95; (95; (95; (95; (95; (95; (95; (95;

29.9e35.8) 10.0e13.1) 9.6e12.8) 7.9e11.1) 9.4e12.7) 7.4e10.6) 8.4e11.2) 11.4e13.8)

Values are expressed as mean (95% confidence interval). P values come from the interaction value of the ANOVA test for time (pre, post) and group (control, experimental).

312

J. Gonza´lez-Iglesias et al. / Manual Therapy 14 (2009) 306e313

Kent et al. (2005) observed that in studies investigating the effectiveness of manual therapy in the management of low back pain in which the clinician had no choice of which techniques to use, the outcomes in the shortterm were superior to studies in which the clinicians were allowed to select which techniques to use for each particular patient. A limitation of this study includes the short-term follow-up of 1 week only. Future studies should seek to investigate the long-term benefits of thoracic thrust in patients with acute neck pain. Further, future clinical trials should investigate the effectiveness of different thoracic thrust manipulation techniques to determine which is the most efficacious. It should also be recognized that all patients were recruited at 1 physiotherapy clinic, so the patients may not be representative of the general population with neck pain. Future studies should consist of multi-centre trials with long-term follow-up. It is possible that the cracking or popping sound during the thoracic manipulation could have created a placebo effect on those patients allocated to the experimental group. Nevertheless, this situation is difficult, if not impossible, to control in a manipulation study. Finally, physical therapy programs for the management of mechanical neck pain usually include other modalities (e.g. exercise, mobilizations, muscle energy, etc.) rather than only electrotherapy or thermal agents, so future studies including other physical therapy interventions are recommended.

5. Conclusion We found that the inclusion of thoracic manipulation combined with a standard electrotherapy/thermal program results in significantly greater reductions in neck pain and disability as well as increases in neck mobility in the short-term in patients with acute mechanical neck pain. Our findings suggest that when treating young adults with acute mechanical neck pain clinicians should consider the findings of this trial in their decisionmaking. Future studies are needed to investigate the long-term effects of thoracic spine thrust manipulation in patients with neck pain.

References Adams G, Sim J. A survey of UK manual therapist’ practice of and attitudes towards manipulation and its complications. Physiother Res Int 1998;3:206e27. American Physical Therapy Association. Guide to physical therapist practice. 2nd ed. Alexandria: American Physical Therapy Association; 2001. Brosseau L, Tugwell P, Wells GA. Philadelphia panel evidence-based clinical practice guidelines on selected rehabilitation interventions for neck pain. Phys Ther 2001;81:1701e17.

Cassidy JD, Quon JA, LaFrance L, Yong-Hing K. The effect of manipulation on pain and range of motion in the cervical spine: a pilot study. J Manipulative Physiol Ther 1992;15:495e500. Childs JD, Piva S, Fritz JM. Responsiveness of the numeric pain rating scale in patients with low back pain. Spine 2005;30:1331e4. Chiu TT, Hui-Chan CW, Chein G. A randomized clinical trial of TENS and exercise for patients with chronic neck pain. Clin Rehabil 2005;19:850e60. Cleland JA, Childs JD, McRae M, Palmer JA, Stowell T. Immediate effects of thoracic manipulation in patients with neck pain: a randomized clinical trial. Man Ther 2005;10:127e35. Cleland JA, Childs JD, Fritz JM, Whitman M. Inter-rater reliability of the history and physical examination in patients with mechanical neck pain. Arch Phys Med Rehabil 2006;87:1388e95. Cleland JA, Childs JD, Fritz JM, Whitman JM, Eberhart SL. Development of a clinical prediction rule for guiding treatment of a subgroup of patients with neck pain: use of thoracic spine manipulation, exercise and patient education. Phys Ther 2007a;87: 9e23. Cleland JA, Glynn P, Whitman J, Eberhart SL, MacDonald C, Childs JD. Short-term response of thoracic spine thrust versus non-thrust manipulation in patients with mechanical neck pain: a randomized clinical trial. Phys Ther 2007b;87:431e40. Cote P, Cassidy J, Carroll LJ, Kristman V. The annual incidence and course of neck pain in the general population: a population-based cohort study. Pain 2004;112:267e73. Cohen J. Statistical power analysis for the behavioural sciences. Hillsdale: Lawrence Erlbaum Associates; 1988. Enthoven P, Skargren E, Oberg B. Clinical course in patients seeking primary care for back or neck pain: a prospective 5-year follow-up of outcome and health care consumption with subgroup analysis. Spine 2004;29:2458e65. Fernandez-de-las-Pen˜as C, Fernandez-Carnero J, Plaza Ferna´ndez A, Lomas-Vega R, Miangolarra-Page JC. Dorsal manipulation in whiplash injury treatment: a randomized controlled trial. J Whiplash Related Dis 2004;3(2):55e72. Ferna´ndez-de-las-Pen˜as C, Alonso-Blanco C, Cuadrado ML, Pareja JA. Forward head posture and neck mobility in chronic tension type headache: a blinded, controlled study. Cephalalgia 2006;26:314e9. Ferna´ndez-de-las-Pen˜as C, Palomeque-del-Cerro L, RodriguezBlanco C, Go´mez-Conesa A, Miangolarra-Page JC. Changes in neck pain and active range of motion after a single thoracic spine manipulation in subjects presenting with mechanical neck pain: a case series. J Manipulative Physiol Ther 2007a;30:312e20. Ferna´ndez-de-las-Pen˜as C, Pe´rez de Heredia M, Brea M, Miangolarra JC. Immediate effects on pressure pain threshold following a single cervical spine manipulation in healthy subjects. J Orthop Sports Phys Ther 2007b;37:325e9. Gibbons P, Tehan P. Manipulation of the spine, thorax and pelvis. Edinburgh: Churchill Livingston; 2000. Gonza´lez T, Balsa A, Sa´inz de Murieta J, Zamorano E, Gonza´lez I, Martı´ n-Mola E. Spanish version of the Northwick Park Neck Pain Questionnaire: reliability and validity. Clin Exp Rheumatol 2001;19:41e6. Gross A, Aker PD, Goldsmith CH, Peloso P. Physical medicine modalities for mechanical neck disorders. Cochrane Database Syst Rev 2000 (2):CD000961. Gummesson C, Isacsson S, Isacsson A, Andersson HI, EktorAndersen J, O¨stergren P-O, et al. The transition of reported pain in different body regions: a one-year follow-up study. BMC Musculoskelet Disord 2006;7:17. Hoving JL, Gross AR, Gasner D, Kay T, Kennedy C, Hondras MA, et al. A critical appraisal of review articles on the effectiveness of conservative treatment for neck pain. Spine 2001;26:196e205.

J. Gonza´lez-Iglesias et al. / Manual Therapy 14 (2009) 306e313 Jensen MP, Turner JA, Romano JM, Fisher L. Comparative reliability and validity of chronic pain intensity measures. Pain 1999;83: 157e62. Jette AM, Smith K, Haley SM, Davis KD. Physical therapy episodes of care for patients with low back pain. Phys Ther 1994;74:101e10. Johansson H, Sojka P. Patho-physiological mechanisms involved in genesis and spread of muscular tension in occupational muscle pain and in chronic musculoskeletal pain syndromes: a hypothesis. Med Hypotheses 1991;35:196e203. Jordan K. Assessment of published reliability studies for cervical spine range of motion measurement tools. J Manipulative Physiol Ther 2000;23:180e95. Jull G, Moore A. Are manipulative therapy approaches the same? Man Ther 2002;7:63. Kent P, Marks D, Pearson W, Keating J. Does clinician treatment choice improve the outcomes of manual therapy for non-specific low back pain? A meta-analysis. J Manipulative Physiol Ther 2005;28:312e22. Kjellman GV, Skargren EI, Oberg BE. A critical analysis of randomised clinical trials on neck pain and treatment efficacy: a review of the literature. Scand J Rehabil Med 1999;31:139e52. Kroeling P, Gross A, Houghton PE, Cervical Overview Group. Electrotherapy for neck disorders. Cochrane Database Syst Rev 2005;(2):CD004251 Leak AM, Cooper J, Dyer S, Williams K, Turner-Stokes L, Frank A. The Northwick Park Neck Pain Questionnaire, devised to measure neck pain and disability. Br J Rheumatol 1994;33:469e74. Norlander S, Aste-Norlander U, Nordgren Sahlstedt B. Mobility in the cervico-thoracic motion segment: an indicative factor of musculo-skeletal neckeshoulder pain. Scand J Rehabil Med 1996;28:183e92.

313

Norlander S, Gustavsson B, Lindell J, Nordgren B. Reduced mobility in the cervico-thoracic motion segment e a risk factor for musculoskeletal neckeshoulder pain: a two-year prospective follow-up study. Scand J Rehabil Med 1997;29:167e74. Norlander S, Nordgren B. Clinical symptoms related to musculoskeletal neckeshoulder pain and mobility in the cervico-thoracic spine. Scand J Rehabil Med 1998;30:243e51. Peolsson A, Hedlund R, Ertzaard S, Oberg B. Intra- and inter-tester reliability and range of motion of the neck. Physiother Can 2000;52:233e42. Pickar JG. An in vivo preparation for investigating neural responses to controlled loading of a lumbar vertebra in the anesthetized cat. J Neurosci Methods 1999;89:87e96. Pickar JG. Neuro-physiological effects of spinal manipulation. Spine J 2002;2:357e71. Pietrobon R, Coeytaux R, Carey T, Richardson W, DeVellis R. Standard scales for measurement of functional outcome for cervical pain or dysfunction. Spine 2002;27:515e22. Sambajon VV, Cillo JE, Gassner RJ, Buckley MJ. The effects of mechanical strain on synovial fibroblasts. J Oral Maxillofac Surg 2003;61:707e12. Shambaugh Jr GE. Manipulating medicine. Lancet 1995;345:1574. Vicenzino B, Paungmali A, Buratowski S, Wright A. Specific manipulative therapy treatment for chronic lateral epicondylalgia produces uniquely characteristic hypoalgesia. Man Ther 2001;6:205e12. Wall PD. The dorsal horn. In: Wall PD, Melzack R, editors. Textbook of pain. 2nd ed. Edinburgh: Churchill Livinsgtone; 2006. Wolfe F, Smythe HA, Yunus MB, Bennett RM, Bombardier C, Goldenberg DL, et al. The American College of Rheumatology 1990 criteria for classification of fibromyalgia: report of the multicenter criteria committee. Arthritis Rheum 1990;33:160e70.

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 314e320 www.elsevier.com/math

Original Article

Mechanical or inflammatory low back pain. What are the potential signs and symptoms? Bruce F. Walker a,*, Owen D. Williamson b a

School of Chiropractic and Sports Science, Faculty of Health Sciences, Murdoch University, 6150 Murdoch, Western Australia, Australia b Department of Epidemiology and Preventive Medicine, Monash University, Alfred Hospital, Melbourne, Victoria, Australia Received 7 November 2007; received in revised form 12 March 2008; accepted 10 April 2008

Abstract Non-specific low back pain (NSLBP) is commonly conceptualised and managed as being inflammatory and/or mechanical in nature. This study was designed to identify common symptoms or signs that may allow discrimination between inflammatory low back pain (ILBP) and mechanical low back pain (MLBP). Experienced health professionals from five professions were surveyed using a questionnaire listing 27 signs/symptoms. Of 129 surveyed, 105 responded (81%). Morning pain on waking demonstrated high levels of agreement as an indicator of ILBP. Pain when lifting demonstrated high levels of agreement as an indicator of MLBP. Constant pain, pain that wakes, and stiffness after resting were generally considered as moderate indicators of ILBP, while intermittent pain during the day, pain that develops later in the day, pain on standing for a while, with lifting, bending forward a little, on trunk flexion or extension, doing a sit up, when driving long distances, getting out of a chair, and pain on repetitive bending, running, coughing or sneezing were all generally considered as moderate indicators of MLBP. This study identified two groups of factors that were generally considered as indicators of ILBP or MLBP. However, none of these factors were thought to strongly discriminate between ILBP and MLBP. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Low back pain; Inflammatory; Mechanical; Signs; Symptoms

1. Introduction Low back pain (LBP) is a common problem with point prevalence ranging from 12% to 33%, 1-year prevalence 22e65% and lifetime prevalence 11e84% (Walker, 2000). While LBP is usually self-limiting, it can persist resulting in a substantial personal, social and economic burden (Walker et al., 2003). In the majority of cases, a specific diagnosis for LBP cannot be defined on the basis of anatomical or physiological abnormalities. Although imaging strategies can be employed to exclude serious * Corresponding author. Tel.: þ61 08 93601297; fax: þ61 8 9360 1299. E-mail address: [email protected] (B.F. Walker). 1356-689X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.04.003

causes of LBP (such as tumours and infections), anatomical abnormalities, such as those associated with the aging process, are commonly observed in otherwise asymptomatic, healthy individuals (Deyo, 2002). While specific therapies can be employed to correct identifiable anatomical or physiological abnormalities, non-specific low back pain (NSLBP) can only be treated empirically. Systematic reviews (Van Tulder et al., 2000; Assendelft et al., 2004) have described the benefit of a broad range of physical and pharmacological interventions over natural history or placebo therapies, but have conceded that effect sizes are small, with little difference in outcomes observed when alternative therapies are compared. This apparent lack of effect may, at least in part, be due to the tendency to treat NSLBP as a homogenous condition, rather than

315

B.F. Walker, O.D. Williamson / Manual Therapy 14 (2009) 314e320

a heterogeneous collection of as yet undefined but differing conditions, some of which might respond and others that do not respond to a particular therapy. There is therefore a need to identify subgroups within the broad classification of NSLBP, and given the failure of classification on the basis of anatomical and physiological abnormalities, attempts have been made to identify subgroups on the basis of symptoms and physical signs (Kent et al., 2005). This syndromic approach has been limited in the past because of the poor inter-rater reliability of proposed classifications. More recently, however, several subgroup classification systems have been demonstrated to have moderate or good inter-rater reliability (Fritz and George, 2000; Flynn et al., 2002; Kilpikoski et al., 2002; Fritz et al., 2006). Subsequent randomised controlled trials (Fritz et al., 2003; Childs et al., 2004; Long et al., 2004; Brennan et al., 2006) have indicated that patients with NSLBP who receive treatment matched to subgroup classifications have better outcomes than those who receive alternative therapies. It therefore seems likely NSLBP does represent a heterogeneous collection of conditions and that the identification of subgroups can result in improved outcomes through directed therapies. NSLBP is commonly described as being ‘‘mechanical’’ (Batt and Todd, 2000; Chaudhary et al., 2004; Valat, 2005) or ‘‘inflammatory’’ (Saal, 1995; Ross, 2006). Although these labels have no universally accepted definitions, there is evidence to support the involvement of both mechanical and inflammatory factors in the generation of LBP (Biyani and Andersson, 2004; Hurri and Karppinen, 2004; Igarashi et al., 2004; Abbott et al., 2006; Al-Eisa et al., 2006; Ross, 2006). Further, there are two distinct types of treatment for LBP that seem to follow this nosological separation. That is, ‘‘mechanical’’ treatments such as mobilisation, manipulation, traction and exercise are contrasted with notionally ‘‘anti-inflammatory’’ treatments like non-steroidal anti-inflammatory medications and corticosteroid injections. There are studies that examine signs and symptoms of specific inflammatory arthritides of the spine such as ankylosing spondylitis (AS) (Rudwaleit et al., 2006). But once conditions like AS have been ruled out there are no studies that determine whether or not inflammatory low back pain (ILBP) and mechanical low back pain (MLBP) subgroups can be differentiated within the NSLBP classification. It would therefore seem useful to attempt to divide LBP sufferers into groups that may respond more readily to two types of treatment, mechanical or inflammatory. If this were possible the number of inappropriate therapy decisions could be decreased. The aims of this study were to identify common symptoms or signs that may allow discrimination between ILBP and MLBP and determine whether the different groups involved in the management of LBP interpret these signs and symptoms in a similar manner.

2. Methods Prior to the commencement of the study, the authors designed a questionnaire listing 26 symptoms and signs relating to LBP. The signs and symptoms were drawn from the a priori knowledge of the authors to be possibly related to LBP. The questionnaire was then pre-tested on a group of four practitioners: a spine surgeon, rheumatologist, chiropractor and manipulative physiotherapist, resulting in the addition of a further question. The final 27 signs and symptoms are found in Table 1. The questionnaire also contained an additional row for ‘‘other’’ signs and symptoms beyond the 27 nominated. This row could be filled out at the discretion of the respondent if they thought that there were other associated factors. Those surveyed were asked ‘‘Please circle the number (0e10) which in your opinion best matches the sign or symptom as being from [mechanical]/[inflammatory] low back pain.’’ Responses were assessed on an 11-point semantic differential scale (Streiner and Norman, 2003) requiring the participants to indicate the degree, from strongly disagree (0) to strongly agree (10), with which they associated each symptom or sign with ILBP and/or MLBP. Participants were instructed to use the middle number (5) to indicate neither disagree nor agree and to leave the answer scale blank to indicate ‘‘don’t know’’. Respondents were advised that it was important to assume that all serious causes of LBP were excluded, including cancer, infection and associated systemic disease. In this study the low back was defined as the area between the costal margins and inferior gluteal folds. A convenience sample of health professionals experienced in the diagnosis and treatment of LBP were surveyed. The sample included both orthopaedically and neurosurgically trained spine surgeons, rheumatologists, medical practitioners with a special interest in musculoskeletal medicine, chiropractors and manipulative

Table 1 Potential signs and symptoms of ILBP or MLBP. Morning pain on waking Intermittent pain during day Pain later in the day Straight leg raising hurts Pain wakes the person up Pain on sitting for a while Pain when standing for a while Pain when lifting Pain bending forward a little Burning pain Aching pain Stabbing pain Constant pain Pain on trunk flexion

Pain on trunk extension Pain on lateral bending Palpatory pain of muscles Palpatory pain of spinous process Stiffness after resting (includes sitting) Morning and afternoon pain Doing a sit up is painful Driving long distances is painful Pain on walking more than 50 m Pain on running Pain on repetitive bending Pain getting out of a chair Pain on cough or sneeze

316

B.F. Walker, O.D. Williamson / Manual Therapy 14 (2009) 314e320

physiotherapists. Key informants identified from within each group provided the names of Australian practitioners who were highly regarded within their professions and likely to have an informed opinion about the topic. The Dillman method (Dillman, 1978) was used for the dissemination of the questionnaire, explanatory information, and follow up procedures. The sample population initially received a herald postcard, then 2 weeks later the questionnaire followed by a reminder which was sent to non-responders after 2, 4 and 6 weeks. At 4-weeks a second questionnaire was also included with the reminder. Each questionnaire was coded to identify the profession of the respondent. The questionnaires had no other identifying information recorded on them and were anonymous. The study had ethics approval from James Cook University and Monash University. The same mode of data collection was used for the entire sample. Data were analysed using SPSS/PC Version 14 (SPSS Inc., Chicago). The median score and 10th and 90th centiles were calculated for each statement by ILBP and MLBP. This method is often used when the data have a skewed distribution (Altman and Bland, 1994). The significance of median scores of agreement was subjectively set and the scores are shown in Table 2. For example a median score of 8 or more was regarded as indicating high levels of agreement that the symptom or sign was an indicator of ILBP or MLBP. While a median score of 2 or less was regarded as indicating high levels of disagreement that the symptom or sign was an indicator of ILBP or MLBP. In addition, the difference between ILBP and MLBP scores was calculated for each question, by respondent, and a median difference in scores of 4 or more was regarded as potentially indicating that the question could be used to potentially differentiate between ILBP and MLBP. Non-parametric statistics were used to compare paired responses to statements (Wilcoxon ranked sign test) and score differences by profession (KruskaleWallis test). Table 2 Median scores and their relative significance. Median score

Significance

10 9 8 7 6 5 4 3 2 1 0

Absolute agreement Very high agreement High agreement Moderate agreement Weak agreement Neutral Weak disagreement Moderate disagreement High disagreement Very high disagreement Absolute disagreement

A chi-squared analysis was used to compare response rates by profession. Given the multiple comparisons between profession groups, a Bonferroni correction was applied, hence p < 0.005 was interpreted as indicating differences between profession groups.

3. Results One hundred and thirty-four questionnaires were sent out. Five were returned as undeliverable leaving 129 possible respondents. Of these, 105 respondents (81%) completed the questionnaire, comprising 29 spine surgeons, 28 rheumatologists, 25 medical practitioners with a special interest in musculoskeletal medicine, 26 chiropractors and 26 manipulative physiotherapists. There was no difference in response rates between the professional groups (c24 ¼ 6.072; p ¼ 0.194). Several respondents completed the ‘‘other’’ signs and symptoms row which allowed the addition of a new sign or symptom. When these were analysed there were no new signs and symptoms but instead minor variations or repetition of signs and symptoms from the existing list. Morning pain on waking (median ¼ 8) demonstrated high levels of agreement as an indicator of ILBP. Pain when lifting (median ¼ 8) demonstrated high levels of agreement as an indicator of MLBP. Constant pain, pain that wakes, and stiffness after resting (median  7) were generally considered as moderate indicators of ILBP, while intermittent pain during the day, pain that develops later in the day, pain on standing for a while, pain bending forward a little, pain on trunk flexion or extension, pain doing a sit up, pain when driving long distances, pain getting out of a chair, and pain on repetitive bending, running, coughing or sneezing (median  7) were all generally considered as moderate indicators of MLBP (Table 3). There was, however, no consistency of agreement either between or within professional groups. No statements were associated with a median score of 3 or less indicating significant disagreement that any symptom or sign was not an indicator of ILBP or MLBP to some extent. Those signs and symptoms with a median score between 4 and 6 (weak or no agreement) are also seen in Table 3. No statements were associated with a median score of more than 7 for both ILPB and MLBP suggesting that no statement indicated both types of pain, while no statements were associated with a median score of 3 or less for both ILBP and MLBP suggesting that no statement excludes both types of pain. Although there was a statistically significant difference ( p < 0.05) in paired responses to all statements apart from that relating to aching pain, no statements were found to be associated with a score of 7 or greater for one type of LBP and 3 or less for the other indicating that none of the factors were thought to strongly discriminate between ILBP and

317

B.F. Walker, O.D. Williamson / Manual Therapy 14 (2009) 314e320 Table 3 Twenty-seven signs and symptoms. Sign or symptom

ILBP (median, 10, 90 centiles)

MLBP (median, 10, 90 centiles)

Difference ILBP  MLBP (median, 10, 90 centiles)

Significance of difference by profession ( p value)

Morning pain on wakinga Intermittent pain during dayd Pain later in the dayd Straight leg raising hurts* Pain wakes the person upc Pain on sitting for a whilee Pain when standing for a whiled Pain when liftingb Pain bending forward a littled Burning paine Aching paine Stabbing paine Constant painc Pain on trunk flexiond Pain on trunk extensiond Pain on lateral bendinge Palpatory pain of musclese Palpatory pain of spinous processe Stiffness after resting (includes sitting)c Morning and afternoon paine Doing a sit up is painfuld Driving long distances is painfuld Pain on walking more than 50 me Pain on runningd Pain on repetitive bendingd Pain getting out of a chaird Pain on cough or sneezed

8 4 5 5 7 5.5 5 4 5 5 6 5 7 5 5 5 5 5 7 6 5 5 5 5 5 5 4

4 7 7 6 4 6 7 8 7 5 6 6 5 7 7 6 5 6 5 5 7 7 6 7 7 7 7

4 2.5 2 2 3 0 1 3 2 0 0 1 2 1 1 1 0 0 2 0 2 1 1 2 2 1 2

0.338 0.001 0.124 0.002 0.043 0.063 0.096 0.024 0.012 0.001 0.130 0.283 0.000 0.028 0.008 0.006 0.000 0.192 0.035 0.443 0.098 0.495 0.002 0.000 0.011 0.020 0.001

(3, 10) (1, 7) (1.5, 7.5) (1, 8) (3, 9) (2, 8) (2, 8) (1.2, 8) (2, 8) (2, 8) (3, 8) (2, 8) (3, 9) (2, 8) (1.5, 8) (1.5, 7.5) (1, 7.7) (1, 8) (5, 10) (3.5, 8.5) (2, 7) (2, 8) (1, 8) (1, 7) (1.5, 8) (2, 8) (1, 8)

(1, 8) (5, 9) (4, 9) (3, 9) (1, 7) (4, 8) (4, 8) (5, 9) (4, 9) (2, 7) (3, 8) (3, 8) (3, 7) (5, 9) (4.5, 9) (5, 8.5) (3, 8) (3, 8) (2, 8) (2, 8) (5, 9) (5, 8) (3, 8) (5, 9) (5, 9) (5, 9) (2.5, 9)

(3, 8) (7, 1) (6, 2.5) (7, 2) (1, 7) (5, 2) (5, 1) (7, 0) (6, 2) (2, 5) (4, 3) (6, 3) (2, 2) (6, 1) (6, 2) (6, 2) (5, 3) (4, 2) (2, 7) (2, 5) (6, 0) (5, 1) (6, 2.8) (6.5, 1) (6, 0) (6, 1) (7, 2.6)

Survey results. a High level of agreement as an indicator of ILBP. b High level of agreement as an indicator of MLBP. c Moderate indicators of ILBP. d Moderate indicators of MLBP. e Variables not considered indicative of either inflammatory or mechanical.

MLBP. The only statement that was associated with a difference in response of 4 or greater was that relating to morning pain on waking; suggesting that this was the only statement that was thought to generally distinguish ILBP from MLBP. There were significant differences between professions with respect to many of the statements being able to distinguish ILBP from MLBP (Table 3). For example, rheumatologists were more likely to regard constant pain and pain that wakes a person as inflammatory and pain on straight leg raising, lifting, running, repetitive bending, coughing and sneezing as mechanical than the other groups. Physiotherapists were more likely to regard pain on lifting or repetitive bending as mechanical. Medical practitioners with a special interest in musculoskeletal medicine did not agree as strongly that ‘‘pain wakes me up’’ or that ‘‘constant pain’’ is a sign of ILBP.

4. Discussion Although NSLBP is commonly described as being mechanical or inflammatory in nature and is treated

by mechanical and anti-inflammatory therapies, there have been no previous attempts to distinguish these subgroups on the basis of symptoms or clinical signs. However, Rudwaleit et al. (2006) did study the clinical history of 101 AS patients and 112 patients without AS thereafter labeled as MLBP patients. In their methods they used an external reference standard known as the New York Criteria (Van der Linden et al., 1984) to diagnose AS. They found four factors that potentially separated the two groups, these were morning stiffness greater than 30 min, improvement with exercise but not with rest, awakening because of back pain in the second half of the night and alternating buttock pain. However, despite some similarity in their results, their study differs from ours insofar as they compared a specific inflammatory arthritide (AS) with all other cases of back pain which they tagged MLBP. In contrast we asked expert respondents to compare non-specific ILBP with non-specific MLBP. In our questionnaire there were no pre-determined definitions or external reference standards (other than exclusions) to categorise non-specific ILBP or MLBP. Indeed this was the reason for our study, to measure the opinion

318

B.F. Walker, O.D. Williamson / Manual Therapy 14 (2009) 314e320

of experts about the extent to which MLBP and ILBP can be distinguished by signs and symptoms. Our study demonstrated some evidence that a number of signs and symptoms are possible indicators of ILBP or MLBP. However, there was no clear agreement either within or between professions regarding whether statements based on common signs and symptoms of LBP are either indicative of, or can distinguish between inflammatory or mechanical causes of LBP. An ideal statement for inclusion in an instrument that distinguishes between ILBP and MLBP would have a high score for one form of LBP, a low score for the other form, a significant difference between the scores for both forms and no significant difference between professions with respect to interpretation. None of the studied statements met each of these criteria. Although morning pain on waking (median difference ¼ 4) and pain that wakes the person up (median difference ¼ 3) were thought to be broadly indicative of ILBP and pain on lifting (median difference ¼ 3) was thought to be broadly indicative of mechanical pain, this was not universally recognised either within or between professional groups. Of these, morning pain on waking is commonly used as a marker of pain due to inflammation (Garrett et al., 1994; Yazici et al., 2004). The fact that morning pain is used as a marker of disease severity in inflammatory spondyloarthopathies such as AS (Garrett et al., 1994) could explain why several respondents suggested that this marker should have been expanded in our survey to reflect the length of time the pain lasted in the morning. The relationship between inflammation and pain, however, is not clear. Although a recent study found that the mean intensity of pain over 24 h was independently associated with high levels of high sensitivity C reactive protein in patients with acute sciatica (less than 8 weeks), this association was not found in patients with chronic LBP (Stu¨rmer et al., 2005). Similarly, the relationship between pain that wakes a patient up and inflammation is not clear. Sleep disturbance is commonly reported in people with non-specific chronic pain, as well as those with inflammatory arthritis (Menefee et al., 2000). The mechanisms by which pain and inflammation cause sleep disturbance have not, however, been well described and may differ. Although the levels of inflammatory cytokines, such as interleukin-6 may alter sleep behaviour (Mullington et al., 2001), there did not appear to be an association between improvements in pain and joint stiffness, and improvements in sleep disturbance, in a small group of patients being treated for rheumatoid arthritis with non-steroidal anti-inflammatory drugs (Lavie et al., 1991). Although pain on lifting is commonly thought to represent mechanical pain, the relationship between spinal load and pain is not clear. Whilst there is strong evidence that work activities such as lifting, bending,

twisting and vibration are a risk factor for the onset and reporting of NSLBP, overall it appears that the size of the effect is less than that of other individual factors (Waddell and Burton, 2000). It is postulated that load, posture and creep may alter the mechanical properties of the spine, resulting in stress concentration in innervated tissues such as the intervertebral discs, facet joints and ligaments (Adams et al., 2002), but there is little direct evidence that such factors are important in NSLBP (Waddell, 2004). In overview the results could be interpreted to suggest that movement or activity-related symptoms are more broadly indicative of MLBP and that pain at rest is more indicative of ILBP. Interestingly no variable was considered to represent both ILBP and MLBP and using our analysis, 10 variables were not considered indicative of either ILBP or MLBP. While it is possible that varying educational paradigms could explain variability between professional groups, it does not obviously explain the variability we found within groups. As the key participants (experts) in this study were selected on their academic and professional standing, it is likely that these differences will be transmitted down through the ranks of each profession and sustains the inadequacy of the evidence. The strength of this study is its good response rate and its generalisability to a wide range of practitioners; however, the study does have some limitations. First, the respondents were not randomly selected from within their professional groups, therefore one cannot generalise the results to the entire population of professionals in each group. However, our purposeful intention was to get the opinion of approximately 20 experts from each group. In this way the answers to our primary questions are more likely to have content validity. Secondly, the best method for defining subgroups within the broad diagnosis of NSLBP has not been established. The approach of this study was to suggest two possible subgroups, ILBP and MLBP and investigate whether experts within relevant professional groups could independently agree on certain symptoms and signs. This approach highlighted the variation within and between participating groups. A similar approach would be to use the Delphi technique (using an iterative/consensus method) to define a set of symptoms and signs that could be measured in trials of mechanical and anti-inflammatory therapies. If symptoms and signs could be used to define subgroups of patients with ILBP and MLBP, trials could be conducted to determine if those who receive subgroup specific treatment do better with the subgroup-specific treatment rather than non-specific treatment, thereby confirming the validity of the subgroups. Despite the limitations of this study, it is clear that considerable diversity of opinion exists regarding symptoms or signs that might be used to distinguish between MLBP and ILBP, both within and between the professional groups involved in the management of NSLBP.

B.F. Walker, O.D. Williamson / Manual Therapy 14 (2009) 314e320

NSLBP is commonly labelled, conceptualised and managed as being inflammatory and/or mechanical in nature and this study identified two groups of factors that were generally considered as indicators of ILBP or MLBP. However, we identified few, if any, signs or symptoms that members of professions involved in the management of NSLBP could highly agree distinguished between these aetiologies. While the general absence of agreement regarding signs and symptoms of ILBP and MLBP does not invalidate the pathophysiological paradigms of mechanical and inflammatory pains, it does, however, signal the need for further research. This research should be aimed at testing the 17 indicators identified for their ability to predict the outcome of mechanical and anti-inflammatory treatments of LBP. If further study establishes that they are able to predict the outcome of the two treatment types, the number of inappropriate decisions to use either may be decreased. Acknowledgments The authors acknowledge that this paper was first presented at the Spine Society of Australia Conference 2006 and that the abstract is published in the conference proceedings, Journal of Bone and Joint Surgery, 88B, Supp III: 448. References Abbott JH, Fritz JM, McCane B, Shultz B, Herbison P, Lyons B, Stefanko G, Walsh RM. Lumbar segmental mobility disorders: comparison of two methods of defining abnormal displacement kinetics in a cohort of patients with non-specific mechanical low back pain. BMC Musculoskelet Disord 2006;7:45. doi:10.1186/ 1471-2474-7-45. Adams MA, Bogduk N, Burton K, Dolan P. The biomechanics of back pain. Edinburgh: Churchill Livingstone; 2002. Al-Eisa E, Egan D, Deluzio K, Wassersug R. Effects of pelvic skeletal asymmetry on trunks movement. Three dimensional analysis in healthy individuals versus patients with mechanical low back pain. Spine 2006;31(3):E71e9. Altman DG, Bland JM. Quartiles, quintiles, centiles and other quantiles. BMJ 1994;309:996. Assendelft WJJ, Morton SC, Yu Emily I, Suttorp MJ, Shekelle PG. Spinal manipulative therapy for low-back pain. Cochrane Database Syst Rev 2004;(1). doi:10.1002/14651858.CD000447.pub2. Art. No.: CD000447. Batt ME, Todd C. Five facts and five concepts for rehabilitation of mechanical low back pain. Br J Sports Med 2000;34(4):261. Biyani A, Andersson GB. Low back pain: pathophysiology and management. J Am Acad Orthop Surg 2004;12:106e15. Brennan GP, Fritz JM, Hunter SJ, Thackeray A, Delitto A, Erhard RE. Identifying subgroups of patients with acute/subacute ‘‘nonspecific’’ low back pain results of a randomized clinical trial. Spine 2006;31:623e31. Chaudhary N, Longworth S, Sell PJ. Management of mechanical low back pain e a survey of beliefs and attitudes in GPs from Leicester and Nottingham. Eur J Gen Pract 2004;10(2):71e2. Childs JD, Fritz JM, Flynn TW, Irrgang JJ, Johnson KK, Majkowski GR, Delitto A. A clinical prediction rule to identify

319

patients with low back pain most likely to benefit from spinal manipulation. Ann Intern Med 2004;141:920e8. Deyo RA. Diagnostic evaluation of LBP; reaching a specific diagnosis is often impossible. Arch Intern Med 2002;162:1444e7. doi:10.1002/14651858.CD000447.pub2. Dillman DA. Mail and telephone surveys. The total design method. New York: John Wiley & Sons; 1978. Flynn T, Fritz J, Whitman J, Wainner R, Magel J, Reindeiro D, Butler B, Garber M, Allison S. A clinical prediction rule for classifying patients with low back pain who demonstrate short-term improvement with spinal manipulation. Spine 2002;27:2835e43. Fritz JM, George S. The use of a classification approach to identify subgroups of patients with acute low back pain. Spine 2000;25: 106e14. 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. Spine 2003;28: 1363e72. Fritz JM, Brennan GP, Clifford SN, Hunter SJ, Thackeray A. An examination of the reliability of a classification algorithm for subgrouping patients with low back pain. Spine 2006;31:77e82. Garrett S, Jenkinson T, Kennedy LG, Whitelock H, Gaisford P, Calin A. A new approach to defining disease status in ankylosing spondylitis: the bath ankylosing spondylitis disease activity index. J Rheumatol 1994;21:2286e91. Hurri H, Karppinen J. Discogenic pain. Pain 2004;112:225e38. Igarashi A, Kikuchi S, Konno S, Olmarker K. Inflammatory cytokines released from the facet joint tissue in degenerative lumbar spinal disorders. Spine 2004;19:2091e5. Kent P, Marks D, Pearson W, Keating J. Does clinician treatment choice improve the outcomes of manual therapy for non-specific low back pain? A meta-analysis. J Manipulative Physiol Ther 2005;28:312e22. Kilpikoski S, Airaksinen O, Kankaanpa¨a¨ M, Leminen P, Videman T, Alen M. Interexaminer reliability of low back pain assessment using the McKenzie method. Spine 2002;27:E207e14. Lavie P, Nahir M, Lorber M, Sharf Y. Nonsteroidal antiinflammatory drug therapy in rheumatoid arthritis patients: lack of association between clinical improvement and effects on sleep. Arthritis Rheum 1991;34:655e9. Long A, Donelson R, Fung T. Does it matter which exercise? A randomized control trial of exercise for low back pain. Spine 2004;29:2593e602. Menefee LA, Cohen MJ, Anderson WR, Dogrhamji K, Frank ED, Lee H. Sleep disturbance and non-malignant chronic pain: a comprehensive review of the literature. Pain Med 2000;1:156e72. Mullington JM, Hinze-Selch D, Pollma¨cher T. Mediators of inflammation and their interaction with sleep: relevance for chronic fatigue syndrome and related conditions. Ann N Y Acad Sci 2001;933: 201e10. Ross JS. Non-mechanical inflammatory causes of back pain: current concepts. Skeletal Radiol 2006. Rudwaleit M, Metter A, Listing J, Sieper J, Braun J. Inflammatory back pain in Ankylosing Spondylitis. A reassessment of the clinical history for application as classification and diagnostic criteria. Arthritis Rheum 2006;54(2):569e78. Saal JS. The role of inflammation in lumbar pain. Spine 1995;20: 1821e7. Streiner DL, Norman GR. Health measurement scales. A practical guide to their development and use. 3rd ed. Oxford: Oxford University Press; 2003. Stu¨rmer T, Raum E, Buchner M, Gebhardt K, Schiltenwolf M, Richter W, Brenner M. Pain and high sensitivity C reactive protein in patients with chronic low back pain and acute sciatica. Ann Rheum Dis 2005;64:921e5. Valat JP. Factors involved in progression to chronicity of mechanical low back pain. Joint Bone Spine 2005;72(3):193e5.

320

B.F. Walker, O.D. Williamson / Manual Therapy 14 (2009) 314e320

Van der Linden S, Valkenburg HA, Cats A. Evaluation of diagnostic criteria for ankylosing spondylitis: a proposal for modification of the New York criteria. Arthritis Rheum 1984;27:361e8. Van Tulder MW, Sholten RJPM, Koes BW, Deyo RA. Nonsteroidal anti-inflammatory drugs for low back pain: a systematic review within the framework of the Cochrane collaboration back review group. Spine 2000;25:2501e13. Waddell G, Burton AK. Occupational health guidelines for the management of low back pain at work e evidence review. London: Faculty of Occupational Medicine, Available from: ; 2000 [accessed 06.11.07].

Waddell G. The back pain revolution. 2nd ed. Edinburgh: Churchill Livingstone; 2004. Walker BF. The prevalence of low back pain. A systematic review of the literature from 1966 to 1998. J Spinal Disord 2000;13: 205e17. Walker BF, Muller R, Grant W. Low back in Australian adults: the economic burden. Asia Pac J Public Health 2003;15:79e87. Yazici Y, Pincus T, Kautiainen H, Sokka T. Morning stiffness in patients with early rheumatoid arthritis is associated more strongly with functional disability than with joint swelling and erythrocyte sedimentation rate. J Rheumatol 2004;31:1723e6.

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 321e329 www.elsevier.com/math

Original Article

Relationships between prolonged neck/shoulder pain and sitting spinal posture in male and female adolescents Leon M. Straker*, Peter B. O’Sullivan, Anne J. Smith, Mark C. Perry School of Physiotherapy, Curtin University of Technology, Perth, Australia Received 3 July 2007; received in revised form 31 March 2008; accepted 13 April 2008

Abstract Neck/shoulder pain (NSP) is a common problem for adolescents and posture has been suggested as an important risk factor. The aim of this cross sectional study was to examine the relationship between prolonged NSP and habitual sitting posture in adolescents. The habitual sitting postures of 1593, 14-year-old adolescents with and without prolonged NSP were assessed using sagittal plane digital photographs. Cervicothoracic and lumbopelvic posture angles were calculated from the digital images using motion analysis software. Adolescents reported experience of NSP by questionnaire. Differences between postures of males and females and those with and without prolonged NSP were examined using independent t-tests. The relationships between cervicothoracic and lumbopelvic postures and presence of prolonged NSP were investigated using logistic regression models controlling for gender. Prolonged NSP was reported by 5.3% of the adolescents, with females reporting a higher prevalence rate (6.5%) than males (4.2%). Females also sat more erect with a more lordotic lumbar posture than males. Adolescents with prolonged NSP had more flexed cervicothoracic posture, more erect trunk and more lumbar lordosis. When gender was controlled, only lumbar lordosis was related to the presence of prolonged NSP. Crown Copyright Ó 2008 Published by Elsevier Ltd. All rights reserved. Keywords: Neck pain; Neck/shoulder pain; Sitting posture; Gender; Adolescents

1. Introduction Neck/shoulder pain (NSP), pain felt in the cervical and upper trapezius region (Kuorinka et al., 1987), is a common problem in adults (Croft et al., 2001; Palmer et al., 2001; Hill et al., 2004), incurring high societal and individual costs (Ferrari and Russell, 2003). A less commonly recognised problem is adolescent NSP. Several studies have estimated the prevalence of adolescent NSP (Wedderkopp et al., 2001; Ehrmann Feldman et al., 2002; Hakala et al., 2002; van Gent * Corresponding author. School of Physiotherapy, Curtin University of Technology, GPO Box U1987, Perth, WA 6845, Australia. Tel.: þ61 8 9266 3634; fax: þ61 8 9266 3699. E-mail address: [email protected] (L.M. Straker).

et al., 2003; Murphy et al., 2004; Siivola et al., 2004). However, only two studies have estimated the prevalence of recurrent NSP in adolescents (El-Metwally et al., 2004; Siivola et al., 2004) and none have estimated the prevalence of prolonged NSP, which are both aspects of chronicity. Siivola et al. (2004) noted that 17% of 16-year-olds had experienced recurrent NSP at least once a week during the previous 6 months. Similarly, El-Metwally et al. (2004) reported that 22% of 12-year-olds had experienced NSP both at baseline and one year later. However, this measure did not account for disorders lasting less than a year, and ignores the possibility that the pains could be unrelated. Although the cause of prolonged NSP in adolescents has not been investigated, it is likely to be multi-factorial, including individual physical and psychosocial factors

1356-689X/$ - see front matter Crown Copyright Ó 2008 Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.04.004

322

L.M. Straker et al. / Manual Therapy 14 (2009) 321e329

which are known to be associated with adult prolonged NSP (Makela et al., 1991; Webb et al., 2003; Ylinen et al., 2004). When considering physical factors, there are indications that spinal posture may be associated with adolescent prolonged NSP. In the frontal plane, spinal asymmetry was found to not predict back and neck pain in adults (Dieck et al., 1985) though evidence that scoliosis can cause long-term neurological deficits in adults (Morcuende et al., 2003) does suggest a possible link with chronic pain. In the sagittal plane, adult studies have demonstrated that prolonged NSP may be associated with increased neck flexion postures (Ariens et al., 2001; Harrison et al., 2004) and altered patterns of muscle activity when sitting (Szeto et al., 2005). Moreover, two studies have demonstrated a link between non-prolonged NSP and spinal posture in adolescents (Hertzberg, 1985; Murphy et al., 2004), while altered adolescent posture has been related to computer use (Straker et al., 2007) which in turn has been related to NSP (Harris and Straker, 2000). However, to date, there have been no studies investigating the links between prolonged adolescent NSP and spinal posture. Such a study would need to avoid certain limitations of recent studies in related areas. Previous adult (Ariens et al., 2001; Harrison et al., 2004; Szeto et al., 2005) and adolescent (Hertzberg, 1985) work in this area has not investigated lumbar or pelvic postures in the sagittal plane. However, there is evidence that lumbopelvic posture influences thoracic posture (O’Sullivan et al., 2006b,c) and thoracic muscle activation patterns (O’Sullivan et al., 2002, 2006b), which may potentially influence the cervical spine. Furthermore, Falla et al. (2007) reported that neutral lumbopelvic posture increases cervical spine muscle activation. Some previous work in adults (Harrison et al., 2004) and adolescents (Hertzberg, 1985) did not considered gender as a confounder, despite the influence of gender on both sitting posture and NSP prevalence in adolescents (Siivola et al., 2004; Dunk and Callaghan, 2005) and adults (Bot et al., 2005). Some previous adult studies have also not measured spinal posture in sitting (Griegel-Morris et al., 1992; Harrison et al., 2004; McAviney et al., 2005). Sitting measurements are particularly relevant for a study of adolescents, as adolescents spend over a quarter of their waking hours sitting (Jago et al., 2005) and sitting is a common aggravating factor for prolonged neck pain in adults (Szeto et al., 2005). Therefore the aims of this study were to evaluate the relationships between cervical, thoracic, lumbar and pelvic sagittal sitting postures and adolescent prolonged NSP, with consideration of gender. We restricted our posture measurements to those in the sagittal plane because whilst the optimal sagittal posture is unknown, frontal plane postures are thought to be optimal if straight but abnormal if not.

2. Materials and methods 2.1. Subjects Data from 1593 adolescents (779 females, 814 males) of mean age 14.1 years were collected as part of their participation in the ‘‘Raine’’ child health study (http:// www.rainestudy.org.au/). This is a long-term project on a range of child health and development issues, which started as a pregnancy cohort in which 2979 women were enrolled at or before the 18th week of gestation from the antenatal clinics at King Edward Memorial Hospital for Women (KEMH) between 1989 and 1991. The children have been followed at birth, 1, 2, 3, 5, 8, 10, and now 14 years of age. The criteria for enrolment in the Raine study were gestational age between 16 and 20 weeks, sufficient proficiency in English to understand the implications of participation, an expectation to deliver at KEMH, and an intention to remain in Western Australia so that follow-up through childhood would be possible. A comparison of this cohort with the general population of Western Australia utilising the Western Australian Maternal and Child Health Research Database at Telethon Institute for Child Health Research (Kendall, 2003) found the sample to be reasonably representative of Western Australians of similar age with the exception of higher ‘at risk’ pregnancies. There were no exclusion criteria for this section of the cohort. Of the 2868 subjects originally drawn into the study, 1608 attended for the 14 year data collection. Of these, 15 failed to give NSP data for unknown reasons. Data from the 1593 adolescents with full NSP pain and posture data are presented here. The mean (SD) height for these 1593 subjects was 1.64 m (0.08) and mean weight was 57.6 kg (13.2).

2.2. Procedure Children participating in a regular cohort follow-up completed the questionnaire on a laptop at the assessment centre with a research assistant close by to provide assistance if needed. The questionnaire contained 130 questions concerning a broad range of physical, medical, nutritional, psychosocial and developmental issues. The questions relevant to prolonged NSP are recorded below, with the possible responses in parentheses. Subjects who had experienced NSP lasting more than three months at any point in the past and who had also experienced NSP in the past month were classified as having prolonged NSP. Did your neck/shoulder pain last for more than 3 months? (‘‘yes’’ or ‘‘no’’) Has your neck/shoulder been painful in the last month? (‘‘yes’’ or ‘‘no’’)

L.M. Straker et al. / Manual Therapy 14 (2009) 321e329

The full child questionnaire took about 1 h to complete, and the two questions described occurred in the first half. A physical assessment of the child was carried out after completion of the questionnaire. This measured anthropometric factors, muscle performance, co-ordination and spinal/pelvic posture during sitting, some of which was used for the current study. Spinal sagittal posture was assessed through photographic analysis of visual markers placed on bony landmarks. We have found fair (Fleiss, 1986) inter-rater reliability for most of the posture variables used in this study (ICCs 0.4e0.75), except for cervicothoracic and trunk angle variables, which were good (ICCs 0.75e0.9) (Fleiss, 1986). We have also found excellent (Fleiss, 1986) re-digitisation reliability (ICCs > 0.9) (Perry et al., 2008) which, together with standard errors of measurement (SEMs) of 3.2e6.5 , support its use for large cohort studies. Retro-reflective markers were placed on the right outer canthus, right tragus, C7 and T12 spinous processes, anterior superior iliac spine and greater trochanter. Lateral photographs were taken with each child sitting on a stool (adjusted to their popliteal height) during three different static sitting postures: (a) looking

323

straight ahead, (b) looking down at their lap, and (c) sitting slumped. To minimise parallax error, the camera (Olympus FE-130, Tokyo, Japan) was positioned on a tripod 80 cm from the floor and 250 cm from the subject, with the subject aligned so they were facing perpendicular to camera. Four trained health professionals collected the posture images (image resolution 2594  1944 pixels). Marker points were digitised using a Peak Motus motion analysis system (Version 8; Peak Performance Technologies, Inc., Centennial, CO, USA) and standard head/neck/thoracic angles were calculated, as defined in Fig. 1. Identical or similar angles have been used in other studies (O’Sullivan, 2006; Straker et al., 2008). 2.3. Data analysis Statistical analysis was carried out with SPSS version 13 (SPSS Inc., Chicago, USA). The prevalence of prolonged NSP was calculated as a percentage of the whole cohort. Descriptive statistics were performed on the prolonged pain prevalence data. Chi-squared analysis was used to assess differences between genders for prevalence values. The difference between usual and slumped sitting

Fig. 1. Illustrations of postural angles. Arrows indicate the angle measured. Dashed lines indicate the vertical. Angle definitions were as follows. Head flexion was the angle formed between the vertical and the line from canthus to tragus (measured above intersect); neck flexion was the angle formed between the vertical and the line of tragus to C7 (measured above intersect); craniocervical angle was formed between the lines of canthus to tragus and tragus to C7 (measured anterior to intersect); cervicothoracic angle was formed between the lines of tragus to C7 and C7 to T12 (measured anterior to intersect); thoracic flexion was formed between the vertical and the line of C7 to T12 (measured from vertical above intersect); trunk angle was formed between the lines of C7 to T12 and T12 to greater trochanter (measured posterior to intersect); lumbar angle was formed between the lines of T12 to ASIS and ASIS to greater trochanter (posterior angle); and pelvic angle was formed between the vertical and the line of greater trochanter to ASIS (measured from vertical above intersect).

324

L.M. Straker et al. / Manual Therapy 14 (2009) 321e329

was calculated to determine how close adolescents sat to their end of flexion range. Independent t-tests were used to analyse differences in posture variables between pain and no pain groups, and males and females. The effect of individual posture variables on prolonged NSP after controlling for gender differences was assessed using logistic regression models (entry method). Adequate model fit was assessed using the model chi-squared likelihood-ratio test and HosmereLemeshow goodness of fit test. Analysis of residuals was performed to check for cases with undue influence over model estimates. The strength of the predictive ability of the model was estimated by Nagelkerke R2. Alpha probability level was set at p < 0.05 for all comparisons. 2.4. Ethics The study was approved by the Human Research Ethics Committee of Curtin University of Technology and Princess Margaret Hospital, Perth. 3. Results 3.1. Prevalence A total of 85 subjects (5.3% of the sample of 1593 with complete NSP data) reported both recent NSP and at least one episode of NSP that had lasted three or more months. Prolonged NSP was reported by 51 females (6.5% of 779) and by 34 males (4.2% of 814). The gender difference was significant ( p ¼ 0.035). 3.2. Posture Gender differences in habitual sitting spinal posture were detected, with females in more erect and lordotic postures than males when looking straight ahead (Table 1). The statistically significant differences between genders were in most cases of moderate or large Table 1 Sitting posture angles ( ) when looking straight ahead in males and females [mean(s.d.)]. Posture angle

Males (n ¼ 765)

Females (n ¼ 715)

Gender difference tdf

p-Value

Head flexion Neck flexion Craniocervical angle Cervicothoracic angle Trunk angle Lumbar angle Anterior pelvic tilt

71.6 (10.1) 53.4 (9.9) 161.8 (12.4)

71.2 (9.3) 51.3 (7.1) 160.1 (12.2)

0.771478 4.671377 2.631475

0.442 <0.001* 0.009*

152.4 (8.0)

145.3 (6.9)

18.431465

<0.001*

238.4 (11.5) 134.2 (19.0) 0.3 (16.1)

225.7 (10.7) 21.891475 124.0 (16.3) 11.121467 9.2 (13.7) 11.501475

<0.001* <0.001* <0.001*

*p < 0.05.

effect size. For example, the 10.2 difference in lumbar angle between genders represented 0.58 of the averaged standard deviation of the two groups, which is interpreted as a small/moderate gender effect, while the 12.7 difference in trunk angle between genders represented 1.14 of the averaged standard deviation of the two groups, which is interpreted as a moderate/large gender effect (Hopkins, 2007). 3.3. Prolonged NSP and posture Adolescents with prolonged NSP sat with a more flexed cervicothoracic angle (prolonged NSP mean (SD) of 146.9 (8.6), no prolonged NSP 149.1 (8.3), tdf ¼ 2.191463, p ¼ 0.028), a more extended trunk angle (prolonged NSP 229.4 (12.5), no prolonged NSP 232.4 (12.8), tdf ¼ 1.981463, p ¼ 0.048), a more lordotic lumbar angle (prolonged NSP 123.2 (18.6), no prolonged NSP 129.6 (18.4), tdf ¼ 2.911466, p ¼ 0.004), and more anterior pelvic tilt (prolonged NSP 9.7 (16.4), no prolonged NSP 4.3 (15.6), tdf ¼ 2.831466, p ¼ 0.005) than those with no experience of prolonged NSP, when looking straight ahead (see Table 2). These differences represented small to moderate effects of protracted NSP on posture measures (Cohen’s-d: 0.26, 0.24, 0.35 and 0.33, respectively; Hopkins, 2007). There was no statistically significant difference in head flexion ( p ¼ 0.689), neck flexion ( p ¼ 0.555) or craniocervical angle ( p ¼ 0.915) between adolescents who had and had not experienced prolonged NSP, when sitting looking straight ahead (see Table 2). A similar pattern of postures were observed across the two groups when subjects were sitting looking down, though this time the differences between cervicothoracic angle and trunk angle were not significant (Table 3). There were no significant pain differences for the variable representing how close to the end of range adolescents sat (difference between usual and slump sitting) (Table 4) as there were differences in slump posture similar to usual posture (data not shown). 3.4. Prolonged NSP, posture and gender As there were moderate to large gender differences in posture variables, logistic regression was used to examine the predictive value of each posture variable in which there was a statistically significant difference between those subjects with and without prolonged NSP, controlling for gender. When subjects were looking straight ahead, both increased anterior pelvic tilt ( p ¼ 0.019, Odds ratio (Exp (b)) ¼ 1.02, 95%CI: 1.00e1.03, Nagelkerke’s R2 ¼ 0.021) and decreased lumbar angles (increased lordosis) ( p ¼ 0.014, Odds ratio (Exp(b)) ¼ 0.99, 95%CI: 0.97e 1.00, Nagelkerke’s R2 ¼ 0.017) were weakly predictive of prolonged NSP after controlling for gender, whilst

325

L.M. Straker et al. / Manual Therapy 14 (2009) 321e329

Table 2 Sitting posture angles ( ) when looking ahead in subjects with and without prolonged neck/shoulder pain, overall and with genders separated [mean(s.d.)]. Males

Head flexion Neck flexion Craniocervical angle Cervicothoracic angle Trunk angle Lumbar angle Pelvic tilt

Females

Overall

No prolonged NSP (n ¼ 727) mean (s.d.)

Prolonged NSP (n ¼ 29) mean (s.d.)

No prolonged NSP (n ¼ 667) mean (s.d.)

Prolonged NSP (n ¼ 45) mean (s.d.)

No prolonged NSP (n ¼ 1394) mean (s.d.)

Prolonged NSP (n ¼ 74) mean (s.d.)

71.5 (9.9) 53.4 (9.9) 161.8 (12.3)

72.5 (13.8) 52.8 (10.1) 160.2 (14.6)

71.3 (9.4) 51.3 (7.1) 160.0 (12.3)

70.0 (9.0) 51.1 (7.5) 161.2 (11.6)

71.4 (9.7) 52.4 (8.7) 161.0 (12.4)

71.0 (11.1) 51.8 (8.6) 160.8 (12.7)

152.5 (8.0)

151.1 (8.6)

145.4 (6.9)

144.2 (7.6)

149.1 (8.3)

146.9 (8.6)*

238.3 (11.5) 134.3 (19.1) 0.3 (16.1)

238.0 (10.8) 130.1 (17.8) 3.1 (17.4)

225.9 (10.8) 124.4 (16.1) 8.9 (13.6)

223.8 (10.2) 118.7 (17.9)* 13.9 (14.5)*

232.4 (12.8) 129.6 (18.4) 4.4 (15.6)

229.4 (12.5)* 123.2 (18.6)* 9.7 (16.4)*

*Differences between no prolonged NSP and prolonged NSP groups, p < 0.05.

in contrast to the unadjusted analyses, trunk angle ( p ¼ 0.290, Odds ratio (Exp(b)) ¼ 0.99, 95%CI: 0.97e 1.01) and cervicothoracic angle ( p ¼ 0.156, Odds ratio (Exp(b)) ¼ 0.98, 95%CI: 0.95e1.01) were not. When subjects were looking down, both increased anterior pelvic tilt ( p ¼ 0.020, Odds ratio (Exp(b)) ¼ 1.02, 95%CI: 1.00e1.04) and decreased lumbar angle ( p ¼ 0.025, Odds ratio (Exp(b)) ¼ 0.99, 95%CI: 0.97e1.00) were again weakly predictive of prolonged NSP after controlling for gender, whilst trunk angle and cervicothoracic angle were not ( p ¼ 0.332, Odds ratio (Exp(b)) ¼ 0.99, 95%CI: 0.97e1.01; and p ¼ 0.396, Odds ratio (Exp (b)) ¼ 0.99, 95%CI: 0.96e1.02, respectively). Finally, after adjustment for gender, there were no significant associations between prolonged pain and any of the variables representing how close to the end of range adolescents sat (difference between usual and slump sitting). Due to rater effects being found in posture assessment reliability, models including rater were also tested, but did not alter the relationship between pelvic and lumbar posture and prolonged neck and shoulder pain.

There were no significant interaction effects between gender and any posture variable on prolonged NSP. Fig. 2 illustrates the relationship between prolonged NSP, posture variables and gender.

4. Discussion 4.1. Prevalence of prolonged NSP in adolescents Although previous work has suggested that adolescent NSP may be recurrent (El-Metwally et al., 2004; Siivola et al., 2004) the present study is the first to show that it can also be prolonged, lasting over three months in 5% of the sample. These findings indicate that almost one in 20 adolescents report that they are experiencing periods of prolonged NSP, which is of concern. Indeed, this figure may be an underestimation. Some subjects may have interpreted brief intervals without NSP as the end of a continuous episode, and so may not have reported any episodes of long-term pain

Table 3 Sitting posture angles ( ) when looking down in subjects with and without prolonged neck/shoulder pain, overall and with genders separated [mean(s.d.)]. Males

Head flexion Neck flexion Craniocervical angle Cervicothoracic angle Trunk angle Lumbar angle Pelvic tilt

Females

Overall

No prolonged NSP (n ¼ 728) mean (s.d.)

Prolonged NSP (n ¼ 29) mean (s.d.)

No prolonged NSP (n ¼ 667) mean (s.d.)

Prolonged NSP (n ¼ 45) mean (s.d.)

No prolonged NSP (n ¼ 1396) mean (s.d.)

Prolonged NSP (n ¼ 74) mean (s.d.)

107.2 (13.3) 71.4 (11.3) 144.2 (11.7)

107.2 (18.3) 69.7 (12.7) 142.4 (14.1)

104.1 (13.1) 67.1 (9.2) 143.0 (12.4)

104.9 (13.9) 67.7 (9.8) 142.9 (8.9)

105.7 (13.3) 69.3 (10.6) 143.6 (12.0)

108.8 (15.7) 68.5 (11.0) 142.7 (11.2)

136.7 (9.4)

136.6 (11.1)

131.2 (8.0)

129.8 (9.7)

134.1 (9.2)

132.5 (10.7)

240.6 (11.6) 134.6 (19.0) 0.1 (16.2)

239.8 (11.5) 130.6 (18.2) 3.3 (18.1)

227.7 (10.9) 124.7 (16.1) 8.5 (13.7)

226.1 (10.3) 119.6 (17.9)* 13.2 (14.6)*

234.4 (13.0) 129.9 (18.4) 4.1 (15.6)

231.5 (12.6) 123.9 (18.7)* 9.3 (16.7)*

*Differences between no prolonged NSP and prolonged NSP groups, p < 0.05.

326

L.M. Straker et al. / Manual Therapy 14 (2009) 321e329

Table 4 Change in sitting posture angles ( ) between normal and slump sitting in subjects with and without prolonged neck/shoulder pain, overall and with genders separated [mean(s.d.)]. Males

Head flexion Neck flexion Craniocervical angle Cervicothoracic angle Trunk angle Lumbar angle Pelvic tilt

Females

Overall

No prolonged NSP (n ¼ 448) mean (s.d.)

Prolonged NSP (n ¼ 16) mean (s.d.)

No prolonged NSP (n ¼ 370) mean (s.d.)

Prolonged NSP (n ¼ 28) mean (s.d.)

No prolonged NSP (n ¼ 818) mean (s.d.)

Prolonged NSP (n ¼ 44) mean (s.d.)

82.5 (16.1) 56.7 (12.9) 25.7 (11.2)

83.8 (17.8) 55.3 (14.6) 28.5 (9.3)

79.5 (17.3) 56.0 (13.5) 23.6 (10.6)

76.7 (19.1) 54.3 (15.4) 22.4 (12.4)

81.2 (16.7) 56.4 (13.2) 24.7 (11.0)

79.2 (18.7) 54.7 (15.0) 24.6 (11.6)

42.7 (10.7)

42.6 (11.8)

37.3 (11.5)

35.0 (14.6)

40.2 (11.4)

37.8 (14.0)

15.2 (9.1) 6.9 (7.9) 6.0 (8.7)

15.4 (8.5) 5.1 (6.1) 6.4 (9.1)

12.5 (8.1) 22.4 (9.2) 12.8 (8.2)

9.5 (8.6) 21.2 (9.0) 11.1 (7.2)

18.5 (9.8) 9.6 (8.6) 8.9 (9.0)

19.1 (9.2) 8.9 (7.3) 8.4 (8.8)

*Differences between no prolonged NSP and prolonged NSP groups, p < 0.05.

punctuated by short pain free periods. We also repeated the analysis described in this paper with just NSP of greater than three months duration regardless of whether they experienced NSP in the last month and

found very similar results. In a follow-up of these adolescents at 17 years of age we are asking separate questions concerning three months of continuous pain and three months of intermittent pain.

Fig. 2. Boxplots displaying the differences between prolonged NSP and no prolonged NSP groups (degrees) in males and females for selected posture variables.

L.M. Straker et al. / Manual Therapy 14 (2009) 321e329

Prolonged NSP was more common in females than in males. This gender difference is consistent with the higher prevalence of non-prolonged NSP in female adult (Bot et al., 2005) and adolescent populations (Siivola et al., 2004) and indicates that this pattern is evident at early adolescence. 4.2. Habitual sitting postures and pain Clear gender differences existed in usual adolescent sitting posture, with females sitting in more erect spinal postures than males, which echoes previous findings in adults (O’Sullivan et al., 2006a). When males and females were analysed together, those with prolonged NSP had similar postural patterns to the whole female group. Since females had a greater prevalence of prolonged NSP, this suggested that some of the differences in postures between pain and non-pain groups could be explained simply by gender. This was supported by fewer significant differences between pain and pain free groups when the genders were split (Table 2), and confirmed by the logistic regression analysis controlling for gender. This demonstrated that with gender controlled, only an increase in lumbar extension angle and increased anterior pelvic tilt in sitting were associated with prolonged NSP in these subjects. This was the case both when looking straight ahead and when looking down, indicating that these findings may be applicable to a range of sitting positions. These are the first reports of such associations in adolescents and highlight the importance of controlling for gender when analysing for associations between posture and pain. Interestingly, the adolescents with pain did not sit closer to their end of flexion range as has been found with adults with low back pain (O’Sullivan, 2006). Males and females appeared to have a similar qualitative relationship between posture and prolonged NSP. This was supported by no significant gender/posture interaction effects and can be seen in Fig. 2. Whether the underlying mechanisms contributing to NSP are the same across genders is, however, unknown. The findings of this study do not support the clinical belief that prolonged NSP in adolescents is associated with altered cervicothoracic postures. This is in contrast to previous adult findings that prolonged neck pain disorders are associated with protracted and flexed head/ neck postures (Harrison et al., 2004). Two other studies (Ariens et al., 2001; Szeto et al., 2005) also showed a non-significant trend for such an effect. As differences in cervicothoracic posture were not observed in the current study population once gender was controlled for, it is possible that gender may have been a confounding factor in the study by Harrison et al. (2004), where gender corrections were not performed. However, Ariens et al. (2001) did adjust for gender and Szeto et al. (2005) used only female subjects, so other factors may

327

be important. We do not believe that our differing results are due to inferior reliability or sensitivity of posture measures, as our reliability was fair to good, and our sample size was large enough to detect small differences. Pain definitions in our study were also similar to those in the other two studies. A more likely reason for the differences in findings was the adolescent age group in our study. Adolescents and adults may have different musculoskeletal tissue characteristics (Lambertz et al., 2003), so it seems reasonable that they might respond differently to similar postures. It is also possible that subgroups with differing cervicothoracic postures exist in the prolonged NSP group in our study, which were lost in the current whole-group analysis due to a washout effect (where effects in different directions for subgroups add together to give the impression of no effect for the combined groups). Such an effect has been reported previously in studies investigating spinal posture and prolonged low back pain (Dankaerts et al., 2006). Our results also differ from adolescent NSP studies not specifically investigating prolonged NSP (Hertzberg, 1985; Murphy et al., 2004), which may result from different mechanisms for prolonged pain. Different posture assessment methods may also contribute to the different findings. This study identified greater lumbar lordosis and the associated anterior pelvic tilt as the only postural difference between those with prolonged NSP and those without. These findings likely reflect that the adolescents with prolonged NSP tended to have more lordotic spines, such that they sat with more lordosis and their slump position was more lordotic than pain free subjects. The fact that there was no difference between usual and slump sitting between groups would suggest that this is a structural difference rather than a motor response to pain, further supported by the fact that there was no report of pain during the posture measure collection. No previous studies have investigated lumbopelvic posture in relation to prolonged NSP, although an increased lumbar lordosis has been associated with a prolonged low back pain subgroup (Dankaerts et al., 2006). It is not clear how this increased lordosis influences NSP although a recent study demonstrated that changes in lumbopelvic posture significantly alter motor control patterns in the neck in pain free subjects (Falla et al., 2007). It is possible that the increased lumbar lordosis in the prolonged NSP subjects observed in this study influences motor control of the neck in a similar manner. However, it is not known whether these changes in lumbopelvic posture contribute to NSP or are a result of NSP in these subjects. We are currently conducting longitudinal work aimed at clarifying causality and mechanism issues. We are also investigating the relationship between posture and low back pain, and the comorbidity of NSP and low back pain.

328

L.M. Straker et al. / Manual Therapy 14 (2009) 321e329

The Nagelkerke’s R2 values of the logistic regression models were very small. The lack of association between NSP and cervicothoracic posture and the weak association with lumbopelvic posture, indicate that posture may have only a small influence on prolonged NSP. Indeed, gender appears to have a greater association with posture than pain. It also suggests that there may be other risk factors of greater importance, or that subgroups exist where postural factors are important, which is a focus of our ongoing research. These findings initially suggest limited support for the application of generic postural education for the management of prolonged NSP in adolescents. However, although the magnitude of the posture changes was very small, it is possible that even small systematic differences in habitual posture early in life may lead to a large increase in accumulated load. Thus the small differences identified in the study, which may not be considered clinically important at one point of time, may have long-term health consequences. Hence further work is necessary before the efficacy of postural education can be challenged. In common with other literature, one limitation of this work is that prolonged NSP has been used as a single outcome variable despite the fact that it may have several different presentations and levels of impairment. No indication as to the intensity of pain or impairment was made from this study although this will be a focus of our longitudinal research on this cohort, with questions assessing the impact of pain on function, school/ work attendance, and the need to take medication or seek professional advice. Another possible limitation is that although our measures showed good sensitivity in terms of being able to detect very small group differences, they may have been limited by the biological variability (extent to which test postures approximated habitual posture), the validity (in comparison to other imaging technologies) and the precision limits of the photographic method. Another limitation is the wholegroup analysis. Different subgroups of NSP may exist that have different anatomical, physiological or psychological mechanisms, and therefore different risk factors that underlie the aetiology of the pain disorder. It is also possible that altered spinal posture interacts differently with these variables. For example, cervicothoracic posture may only be an important risk factor for a subgroup with high exposure to computers and poor motor control. Future work will therefore aim to sub-classify prolonged NSP and examine multivariate relationships within each subgroup.

5. Conclusion Prolonged NSP affected 5% of adolescents, and was more common in females than males. Prolonged NSP was weakly associated with more lordotic lumbopelvic

postures, but the clinical belief that NSP is related to cervicothoracic postures was not supported when gender was included in the model. Acknowledgements We would like to acknowledge funding from the Australian National Health and Medical Research Council (project #323200), the Raine Foundation at the University of Western Australia, Healthway, the Arthritis Foundation of Australia, and the Arthritis Foundation of Western Australia. We would also like to thank Rosemary Austin, Lee Clohessy, Alex D’Vauz, Clare Haselgrove, Monique Robinson, Nick Sloan and Diane Wood for collection and/or initial processing of data. References Ariens G, Bongers P, Douwes M, Miedema M, Hoogendoorn W, 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(3):200e7. Bot S, van der Waal J, Terwee C, van der Windt D, Schellevis F, Bouter L, et al. Incidence and prevalence of complaints of the neck and upper extremity in general practice. Annals of the Rheumatic Diseases 2005;64:118e23. Croft PR, Lewis M, Papageorgiou AC, Thomas E, Jayson MI, Macfarlane GJ, et al. Risk factors for neck pain: a longitudinal study in the general population. Pain 2001;93(3):317e25. Dankaerts W, O’Sullivan PB, Burnett AF, Straker LM. Differences in sitting posture are associated with non-specific chronic low back pain disorders when patients are sub-classified. Spine 2006;31(6):698e704. Dieck GGS, Kelsey JJL, Goel VVK, Panjabi MMM, Walter SSD, Laprade MMH. An epidemiologic study of the relationship between postural asymmetry in the teen years and subsequent back and neck pain. Spine 1985;10(10):872e7. Dunk NM, Callaghan JP. Gender-based differences in postural responses to seated exposures. Clinical Biomechanics 2005;20:1101e10. Ehrmann Feldman D, Shrier I, Rossignol M, Abenhaim L. Risk factors for the development of neck and upper limb pain in adolescents. Spine 2002;27(5):23e8. El-Metwally A, Salminen JJ, Auvinen A, Kautiainen H, Mikkelsson M. Prognosis of non-specific musculoskeletal pain in preadolescents: A prospective 4-year follow-up study till adolescence. Pain 2004;110(3):550e9. Falla D, O’Leary S, Fagan A, Jull G. Recruitment of the deep cervical flexor muscles during a postural-correction exercise performed in sitting. Manual Therapy 2007;12:139e43. Ferrari R, Russell AS. Neck pain. Best practice and research. Clinical Rheumatology 2003;17:57e70. Fleiss JL. The design and analysis of clinical experiments, 1. New York: John Wiley; 1986 [Chapter 1, p. 7]. van Gent C, Dols JJ, de Rover CM, Hira Sing RA, de Vet CW. The weight of schoolbags and the occurrence of neck, shoulder, and back pain in young adolescents. Spine 2003;28(9):916e21. Griegel-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):425e31.

L.M. Straker et al. / Manual Therapy 14 (2009) 321e329 Hakala P, Rimpela A, Salminen JJ, Virtanen SM, Rimpela M. Back, neck, and shoulder pain in Finnish adolescents: national cross sectional surveys. British Medical Journal 2002;325:743e7. Harris C, Straker L. Survey of physical ergonomics issues associated with school children’s use of laptop computers. International Journal of Industrial Ergonomics 2000;26(3):337e46. Harrison DD, Harrison DE, Janik TJ, Cailliet R, Ferrantelli JR, Haas JW, et al. Modeling of the sagittal cervical spine as a method to discriminate hypolordosis: results of elliptical and circular modeling in 72 asymptomatic subjects, 52 acute neck pain subjects, and 70 chronic neck pain subjects. Spine 2004;29(22):2485e92. Hertzberg A. Prediction of cervical and low-back pain based on routine school health examinations. A nine- to twelve-year follow-up study. Scandinavian Journal of Primary Health Care 1985;3(4): 247e53. Hill J, Lewis M, Papageorgiou AC, Dziedzic K, Croft P. Predicting persistent neck pain: A 1-year follow-up of a population cohort. Spine 2004;29(15):1648e54. Hopkins WG. A new view of statistics, ; 2007 [last accessed 6/12/07]. Jago R, Anderson CB, Baranowski T, Watson K. Adolescent patterns of physical activity differences by gender, day, and time of day. American Journal of Preventive Medicine 2005;28(5):447e52. Kendall GE, Children in families in communities: a modified conceptual framework and an analytic strategy for identifying patterns of factors associated with developmental health problems in childhood. Perth: University of Western Australia; 2003 [unpublished PhD thesis]. Kuorinka I, Jonsson B, Kilbom A, Vinterverg H, Biering-Sorensen F, Andersson G, et al. Standardised Nordic questionnaires for the analysis of musculoskeletal symptoms. Applied Ergonomics 1987;18(3):233e7. Lambertz D, Mora I, Grosset JF, Perot C. Evaluation of musculotendinous stiffness in prepubertal children and adults, taking into account muscle activity. Journal of Applied Physiology 2003; 95(1):64e72. Makela M, Heliovaara M, Sievers K, Impivaara O, Knekt P, Aromaa A. Prevalence, determinants, and consequences of chronic neck pain in Finland. American Journal of Epidemiology 1991;134(11):1356e67. McAviney J, Schulz D, Bock R, Harrison DE, Holland B. Determining the relationship between cervical lordosis and neck complaints. Journal of Manipulative and Physiological Therapeutics 2005;28(3):187e93. Morcuende JA, Dolan LA, Vazquez JD, Jirasirakul A, Weinstein SL. A prognostic model for the presence of Neurogenic Lesions in atypical idiopathic scoliosis. Spine 2003;29:51e8. Murphy S, Buckle P, Stubbs D. Classroom posture and self-reported back and neck pain in schoolchildren. Applied Ergonomics 2004;35:113e20.

329

O’Sullivan P, Dankaerts W, Burnett A, Straker L, Bargon G, Moloney N, et al. Lumbopelvic kinematics and trunk muscle activity during sitting on stable and unstable surfaces. Journal of Orthopaedic and Sports Physical Therapy 2006a;36(1):19e25. O’Sullivan P, Dankaerts W, Burnett AF, Farrell GT, Jefford E, Naylor C, et al. Effect of different upright sitting postures on spinal pelvic curvature and trunk muscle activation in a pain-free population. Spine 2006b;31(19):707e12. O’Sullivan PB, Grahamslaw KM, Kendell M, Lapenskie SC, Moller NE, Richards KV. The effect of different standing and sitting postures on trunk muscle activity in a pain-free population. Spine 2002;27:1238e44. O’Sullivan PB, Mitchell T, Bulich P. The relationship between posture and back muscle endurance in industrial workers with flexionrelated low back pain. Manual Therapy 2006c;11:264e71. Palmer KT, Walker-Bone K, Griffin MJ, Syddall H, Pannett B, Coggon D, et al. Prevalence and occupational associations of neck pain in the British population. Scandinavian Journal of Work, Environment & Health 2001;27(1):49e56. Perry M, Smith AJ, Straker LM, Coleman JL, O’Sullivan PB. Reliability of sagittal photographic spinal posture assessment in adolescents. Advances in Physiotherapy 2008;10(2):66e75. Siivola S, Levoska S, Latvala K, Hoskio E, Vanharanta H, KeinanenKiukaanniemi S. Predictive factors for neck and shoulder pain: a longitudinal study in young adults. Spine 2004;29(15):1662e9. Straker L, Burgess-Limerick R, Pollock C, Murray K, Netto K, Coleman J, et al. The impact of computer display height and desk design on 3D posture during information technology work by young adults. Journal of Electromyography and Kinesiology 2008;18(2):336e49. Straker L, O’Sullivan PB, Smith A, Perry MC. Computer use and habitual spinal posture in Australian adolescents. Public Health Reports 2007;122(5):634e43. 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;10(4):281e91. Webb R, Brammah T, Lunt M, Urwin M, Allison T, Symmons D. Prevalence and predictors of intense, chronic, and disabling neck and back pain in the UK general population. Spine 2003;28(11): 1195e202. Wedderkopp N, Leboeuf-Yde C, Andersen LB, Froberg K, Hansen HS. Back pain reporting pattern in a Danish populationbased sample of children and adolescents. Spine 2001;26(17): 1879e83. Ylinen J, Salo P, Nykanen M, Kautiainen H, Hakkinen A. Decreased isometric neck strength in women with chronic neck pain and the repeatability of neck strength measurements. Archives of Physical Medicine in Rehabilitation 2004;85:1303e8.

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 330e337 www.elsevier.com/math

Original Article

The effect of a vastus lateralis tape on muscle activity during stair climbing U. McCarthy Persson*, H.F. Fleming, B. Caulfield University College Dublin, School of Physiotherapy and Performance Science, Health Sciences Centre, Belfield Dublin 4, Ireland Received 27 September 2007; received in revised form 24 April 2008; accepted 6 May 2008

Abstract Recently taping techniques with the primary purpose of altering muscle activity have become a part of clinical physiotherapy practice. A firmly applied tape across the fibres of the vastus lateralis (VL) muscle has been proposed to decrease the VL muscle activity. The primary aim of this study was to assess the effects of an inhibitory muscle tape applied over the vastus lateralis (VL) muscle during stair climbing. Twenty five subjects without lower limb pathology were recruited. Normalised integrated EMG (IEMG) was analysed from VL, vastus medialis obliquus (VMO), biceps femoris (BF) and soleus muscles during stair climbing. The subjects were assessed during three conditions: no tape (untaped), (no tension) control tape and (tensioned tape) VL inhibitory taping application. There was a significant decrease ( p < 0.05) in the VL IEMG during the initial stance phase during both stair ascent and descent. The inhibition if the VL muscle occurred with both control and VL inhibitory tape applied. No significant differences ( p > 0.05) were noted in any of the other muscles assessed. The results demonstrated that there was a significant decrease in the IEMG of the VL both during stair ascent and descent with VL inhibitory tape and control tape applied in normal subjects. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Tape; EMG; Muscle inhibition; Stair climbing

1. Introduction The use of taping and strapping in injury management has been advocated for over a century. Gibney advocated what is now the most common taping technique, taping the ankle, as early as 1895 (Wilkerson, 2002). The use of tape as an adjunct to treating musculoskeletal and sports injuries has now become common practice. The main body of research into tape applications has previously focused on techniques aiming to limit available joint movement. New taping applications with

* Corresponding author. E-mail address: [email protected] (U. McCarthy Persson). 1356-689X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.05.002

a primary purpose to increase or decrease a muscles’ activity pattern have recently emerged. It has been proposed that application of a tape parallel to the fibres of a muscle may increase the muscles’ activity and applying tape firmly perpendicular to the muscle fibres could inhibit the muscle (Morrissey, 2000). Despite the popular use of these techniques in clinical practice there is limited evidence. A firmly applied rigid tape across the fibres of the vastus lateralis (VL) muscle has been proposed to decrease the VL muscle activity (Tobin and Robinson, 2000). There is some evidence that an imbalance exists between the vastus medialis obliquus (VMO) and VL in patellofemoral pain (PFP) (Voight and Wieder, 1991; Cowan et al., 2001). The inhibitory taping application was originally designed to restore quadriceps muscle

U. McCarthy Persson et al. / Manual Therapy 14 (2009) 330e337

balance where VL was proposed to be overactive in relation to VMO among PFP patients (Tobin and Robinson, 2000). There are only two studies published on inhibitory taping of the VL and its effect on muscle activity both studies investigating surface EMG during stair descent (Tobin and Robinson, 2000; Janwantankul and Gaogasigam, 2005). The studies produced divergent results likely to be related to methodological issues. The two studies differ in terms of EMG sampling, processing and taping methods to a great extent. It is therefore difficult to draw direct comparisons and identify how this type of inhibitory tape affects the muscle activity. Previous descriptions of the methods of VL inhibitory tape applications are subjective and variable (Tobin and Robinson, 2000; Janwantankul and Gaogasigam, 2005). The previous studies on VL inhibitory tape analysed the mean EMG activity of the whole muscle activation envelope from one full stance phase during stair descent which did not allow for identification of preand post-foot contact muscle activation. (Tobin and Robinson, 2000; Janwantankul and Gaogasigam, 2005). The initial foot contact period has been not only considered an important phase in stair climbing as the greatest knee moments and patellofemoral joint forces occur at this stage (Costigan et al., 2002), but also for the consideration of vasti muscle onset activation in PFP (Gilleard et al., 1998; Cowan et al., 2002). For this study, it was therefore decided to analyse the initial activation period of muscle pre-contact activity and during the weight acceptance phase (McFadyen and Winter, 1988). The aim of this study was to assess the effects of a VL inhibitory tape application using a repeatable taping method, during controlled stair ascent and descent and performing appropriate EMG data collection and analysis. The null hypothesis was that there would be no difference in muscle activity between control tape VL inhibitory tape and a control condition during stair climbing.

2. Methods 2.1. Subjects A total of 25 healthy volunteers (12 males and 13 females) were recruited between the ages of 19 and 39 (mean 25.8 ± 6.52). Subjects were included if they were aged between 18 and 45, had no previous lifetime history of knee pain, or quadriceps or hamstrings muscle pathology leading to consultation of a health professional. The subjects were excluded if they were unable to ascend or descend stairs, or had a history of an allergic reaction to zinc oxide tape. Informed written consent was received from each participant and an explanation of the study was given prior to the commencement of the study, which was approved by the University

331

College Dublin Human Research Ethics Committee. The subject’s height, weight, body mass index and skin fold thickness were measured and calculated. Skin fold thickness was obtained from calliper (Harpenden, Burgess Hill, UK) measurements at four sites: biceps, triceps, subscapular and suprailiac. The subjects were then allocated to a group of VL inhibitory tape first or control tape first by computer generated random allocation using SPSS (v 11.0) by an independent party. All subjects were tested without tape (control condition) initially prior to randomised tape application. The control condition was excluded from the randomisation process to minimise skin irritation. The subjects’ results were analysed across all three conditions. 2.2. Instrumentation Surface EMG was gathered using pre-amplified EMG electrodes MA-317 (gain 300  2%) (Motion lab systems, Baton Rouge, LA, USA) and the data were recorded on a Biopac MP 100A (Biopac Systems Inc., Santa Barbara, CA, USA). Measurement of foot contact on the step was performed using a heel and toe strike transducer (Biopac Systems Inc., Santa Barbara, CA, USA). The data were recorded on the Biopac MP 100. The data from EMG and heel and toe strike transducer was processed using its associated AcqKnowledge software (Version 3.5.7.). 2.3. Data acquisition The surface EMG electrodes were applied over VMO, VL and biceps femoris (BF) muscles of the right limb. EMG was also sampled from the soleus muscle as a close relationship has previously been established between quadriceps and soleus in response to both cutaneus and nociceptive stimulation (Rossi and Decchi, 1995; Marque et al., 2001). The EMG activity of the biceps femoris was collected due to a previously suggested mechanism of quadriceps inhibition that may occur due to a withdrawal reflex. This reflex increases activation of the hamstrings muscle with a reciprocal inhibition of the quadriceps (Leroux et al., 1995). The electrodes were not removed or altered during the study between the different testing conditions. For electrode placement of the VMO the electrode was placed 4 cm proximal to the medial superior corner of the patella and 3 cm medially at an angle of 55 to the line of the femur. The electrode for the VL was placed 10 cm proximal to the central patella in line with the femur and 6e7 cm laterally at an angle of 15 , both electrodes (Gilleard et al., 1998). The biceps femoris placement was located on the midpoint on a line from the ischial tuberosity to the lateral tibial condyle and along the angle of that line with

332

U. McCarthy Persson et al. / Manual Therapy 14 (2009) 330e337

the knee flexed less than 90 (SENIAM, 1999). The electrode for the soleus muscle was placed 3 cm proximal to the musculotendinous junction of the achilles tendon bisecting the leg in line with the tibia (Hugon, 1973). The EMG was sampled at a rate of 2000 Hz. To decrease skin impedance, the skin was shaven, abraded using sandpaper and cleansed using an alcohol wipe. The electrodes were attached with adhesive tape. The data were recorded on the Biopac MP 100A. Heel and toe strike transducers (Biopac Systems Inc., Santa Barbara, CA, USA) were applied using adhesive tape, one under the calcaneus and the other underneath the metatarsal heads, both sensors placed lateral to the midline. 2.4. Procedure The subject walked a set of three steps up and three steps down. Each step was 30 cm deep, 60 cm wide and the height of the step was 20 cm. The subjects stepped up with the left leg first followed by the right from which the data were collected. During stair descent the subject again initiated the gait with the left leg and data was collected from the following right leg. A metronome was used to standardise the pace of walking at 96 beats per minute (Cowan et al., 2000). Each subject was given as much time as was needed to get familiar with the stairs and the pace of walking prior to data collection. Data were collected during five trials of stair walking for each condition (untaped, VL inhibitory tape, control tape). Each trial of the raw EMG data was visually inspected for any evidence of baseline shift, motion artifact or mains interference. If any of the aforementioned were present, the trial was rejected. The three trials displaying the highest EMG collection quality were subsequently analysed. The pressure sensitive heel and toe switches gave a recording of the contact from the forefoot and heel during the stair walking, detected by a spike or sudden deviation from the baseline. The time of impact was determined to within 0.5 ms and rounded to the nearest millisecond. The skin was marked with two reference lines: (1) anterior superior iliac spine (ASIS) to the midpoint of the superior border of the patella; and (2) greater trochanter to the lateral femoral epicondyle. The midpoint was marked on each line. This area was shaved and wiped with alcohol. The subject was positioned on the side with a pillow between the knees, which were flexed to an angle of 30 . Two lengths of flexible hypoallergenic tape, 5 cm in width, (Fixomull e Beiersdorf, Milton Keynes, UK) were applied superior and inferior to the previously marked midpoints of the reference lines extending past both lines by 2 cm. The control tape (flexible hypoallergenic tape) was laid on without any tension applied. Three strips of 3.8 cm zinc oxide tape were applied with tension on top of the hypoallergenic tape from

the anterior line extending over the lateral line (VL inhibitory tape). Tension was applied to the zinc oxide tape laterally and posteriorly with one hand. The lateral thigh tissues were collected with the other hand while applying a downward pressure with the thumb over the VL between the reference lines causing a furrow in the skin. The tension applied on the tape was standardised to cause a ‘‘skin roll’’ anterior and posterior to the thumb with the same height as the width (w20 mm) of the researchers thumb (McCarthy Persson et al., 2007). A total of three zinc oxide tape strips were applied, starting with the most superior tape above the midline mark, followed by the middle and finished with the distal tape, each overlapping the other by one-third of the tape width (Fig. 1). A 10-min period was kept between the taping conditions to minimise any possible carryover effect. Previous studies have found an immediate return to baseline muscle activity upon removal of tape (Alexander et al., 2003, 2008). Any pain perceived from the application of tape during the stair walking was assessed directly after the trial using the visual analogue scale (VAS). 2.5. Data analysis The raw EMG data (Fig. 2A) were band pass filtered (Blackman 61dB) at 20 Hz (low) and 500 Hz (high) (Fig. 2B). The data were thereafter full-wave rectified (Fig. 2C). The data were averaged over a 15-ms moving window (Fig. 2D) with a sampling rate of 2000 Hz, this equals 30 data points that slide one data point at a time (Swanik et al., 1999; Caulfield et al., 2004). The data were exported to Microsoft Excel 2000. A macro was designed to extract EMG data from 500 ms pre-foot contact and 1000 ms post-contact both for stair ascent and descent and performing normalisation of the data. This time period included all EMG activity prior to foot contact and throughout the stance phase thus

Fig. 1. VL inhibitory taping application.

U. McCarthy Persson et al. / Manual Therapy 14 (2009) 330e337

333

represents a percentage of peak activity related to the linear envelope (% ms). The muscle activity was quantified by calculating the resulting normalised integral EMG activity at step-up and step-down during separate envelopes: - 150 ms prior to impact to include the EMG pre-activation and 150 ms post impact to cover the first burst of stance phase EMG activity. This first burst of activity after foot contact was found to correspond to the initial weight acceptance phase. The first period that was analysed therefore included all pre-contact motor activity, while the second period of muscle activation corresponded to the initial weight acceptance phase as described by McFadyen and Winter (1988). As this taping procedure was designed for treatment of PFP, these periods are of particular clinical relevance. The initial period of the vastii muscle contraction has been one of the most researched due to the proposed alteration in onset of muscle activation in subjects with PFP (Voight and Wieder, 1991; Witvrouw et al., 1996; Gilleard et al., 1998; Cowan et al., 2001). The time of impact was determined from the heel/toe strike transducers on the right foot during step up and step down to within 0.5 ms and rounded to the nearest millisecond.

Fig. 2. (A) Shows a sample of raw EMG of VL during the full stairclimbing task. The first burst represents step-up, second burst e step on platform (not analysed) and last burst represents the step-down phase. (B) The EMG sample band pass filtered 20e500 Hz. (C) The EMG sample rectified. (D) The EMG sample averaged over a 15 ms moving window.

allowing for analyses of selected time frames. It has been observed that the sensitivity of EMG testing can be improved if the average pattern of EMG for a specific functional activity can be constructed (ensemble-averaged profile) and the peak or mean from this pattern is used to normalise the amplitude of muscle activity (Yang and Winter, 1984; Swanik et al., 1999). Therefore the peak value of EMG in each control trial was identified and the mean of these three peak values was used to re-express all EMG records, thereby normalising all conditions with respect to the untaped condition. The mean of the three normalised EMG records for all conditions were subsequently calculated to provide a normalised ensemble averaged EMG for each condition (Caulfield et al., 2004). The area was calculated under the curve during the relevant time periods creating a Riemann integral (IEMG) (Hamill and Knutzen, 1995). IEMG has been recommended as the optimal method for quantifying muscle activation during kinesiological applications (SENIAM, 1999; Caulfield et al., 2004). Normalised IEMG can be expressed as a value that

3. Statistical analysis Statistical analysis was carried out using Statistical Package for Social Sciences (SPSS) version 11.0 and Microsoft Excel 2000. The majority of the data were normally distributed as confirmed by the KolmogoroveSmirnov test, thus allowing parametric tests to be employed. Where data were not normally distributed, non-parametric analyses were carried out. Repeated measures analysis of variance (ANOVA) was employed where the data were normally distributed to determine any differences between the untaped, VL inhibitory tape and control tape conditions for each muscle at 150 ms pre- and post-foot contact during stair ascent and descent. Where the data was not normally distributed, Friedman’s test was employed for non-parametric analysis. Post hoc analysis was carried out using paired t-test for parametric analysis. No post hoc analysis was carried out on non-parametric data due to non-significance of the initial analyses of these data sets. The level of significance was set at p < 0.05.

4. Results The mean  standard deviation (SD) height of the subjects was 1.72  0.077 m, mean body mass

334

U. McCarthy Persson et al. / Manual Therapy 14 (2009) 330e337

70.0  9.25 kg, and the calculated body mass index 23.5  2.76 kg/m2. The mean (SD) skin fold thickness was 54.3  28.33 mm.

Normalised group averaged mean EMG - Step up 80% 70% 60%

Analyses of the normalised IEMG (area underneath the rectified smoothed curve expressed in % $ ms) of the VL, VMO, BF and soleus muscles revealed no significant difference in the pre-foot contact period during step-up across the three testing conditions ( p > 0.05) (Table 1). A significant difference was found in the VL IEMG between the three testing conditions in the 150ms period after foot contact during step-up ( p ¼ 0.002). Post hoc analysis of the step-up indicated that there was a significant decrease of VL IEMG with VL inhibitory tape applied compared to untaped ( p ¼ 0.027) and with control tape versus untaped ( p ¼ 0.001) (Table 1, Fig. 3). No significant difference was found in IEMG of the other muscles (VMO, BF or soleus) assessed during step-up post-foot contact period ( p > 0.05) (Table 1). 4.2. Step-down Analysis of the normalised IEMG during the stepdown of the VL, VMO, BF and soleus muscles revealed no significant difference between VL inhibitory tape, control tape and untaped in the 150 ms pre-foot contact period during step-down ( p > 0.05) (Table 2). A significant difference was found in the VL IEMG between the untaped, VL inhibitory tape and control tape in the 150 ms period after foot contact during step-down ( p ¼ 0.003). Post hoc analysis revealed a significant decrease in IEMG of the VL during the 150 ms period post contact when the VL inhibitory tape ( p ¼ 0.002) and control tape ( p ¼ 0.005) were applied compared to the untaped condition (Table 2, Fig. 4)

%peak

4.1. Step-up

50% 40% 30% Mean Untaped Mean Tape Mean Control tape

20% 10% 0% -150

-100

-50

0

50

100

150

Time (ms) Fig. 3. Normalised group averaged mean EMG of the VL during stepup 150 ms pre- and post-foot contact.

No significant difference was found in IEMG of the other muscles assessed (VMO, BF, soleus) during stepdown post-foot contact period ( p > 0.05) There were no significant differences in IEMG between VL inhibitory tape and control tape during stepup or step-down ( p > 0.05). 4.3. Pain assessment All subjects subjectively experienced a strong pulling sensation from the VL inhibitory tape but not from the control tape. No subjects experienced pain as a result of the tape applications during the stair climbing (VAS ¼ 0).

5. Discussion The results of the study of EMG indicated that there was a significant decrease in the IEMG of the VL both

Table 1 Step-up, means standard deviation (SD) and analyses of the IEMG (% ms). Untaped

VL inhibitory tape

Control tape

Statistical analysesa

F

150 ms pre VL VMO BF Soleus

10.59  5.854 9.69  4.155 64.00  50.370 12.29  7.778

10.25  6.709 9.58  6.096 60.46  49.918 9.94  4.140

11.57  12.057 9.68  5.129 63.48  59.092 10.52  4.751

p ¼ 0.432b p ¼ 0.989 p ¼ 0.583 p ¼ 0.726

1.680c 0.011 0.546 0.323

150 ms post VL VMO BF Soleus

143.19  19.015 143.50  19.015 75.08  50.875 69.99  28.627

133.29  29.069* 137.70  31.033 69.49  51.255 54.61  27.522

128.56  21.013* 135.17  27.008 71.14  54.535 56.73  25.927

p ¼ 0.002 p ¼ 0.094 p ¼ 0.429 p ¼ 0.057

7.226 2.484 0.861 3.378

*Significant difference compared to untaped condition ( p < 0.05). a Repeated measures ANOVA unless other indicated. b Friedman’s test. c Chi square.

335

U. McCarthy Persson et al. / Manual Therapy 14 (2009) 330e337 Table 2 Step-down, means  SD and analyses of the IEMG (% ms). Untaped

VL inhibitory tape

Control tape

Statistical analysesa

F

150 ms pre VL VMO BF Soleus

38.45  18.263 31.09  16.022 83.63  58.515 27.09  12.751

35.77  21.579 29.69  17.96 85.07  63.887 27.01  13.349

36.69  19.706 31.42  15.771 86.33  61.972 27.53  10.853

p ¼ 0.368 p ¼ 0.548 p ¼ 0.801 p ¼ 0.895

1.02 0.46 0.133 0.058

150 ms post VL VMO BF Soleus

99.00  36.543 93.35  35.590 105.05  58.714 75.46  23.965

88.47  29.934* 91.52  34.731 107.53  70.622 73.52  21.543

89.01  31.550* 87.77  32.261 105.71  70.743 77.75  20.430

p ¼ 0.003 p ¼ 0.310 p ¼ 0.838 p ¼ 0.572

6.751 1.202 0.066 0.475

*Significant difference compared to untaped condition ( p < 0.05). a Repeated measures ANOVA unless other indicated.

during stair ascent and descent with VL inhibitory and control tapes applied. This indicates that the proposed effects and the clinical use of the VL inhibitory taping are supported when applied in normal subjects, but also suggests that a non-rigid tape without the applied tension (control tape) has a similar effect. The VL inhibitory tape and control tape applications produced a decrease in the VL IEMG both during step-up and step-down, despite the differences in type of quadriceps muscle contraction. During stair ascent the hip and knee were in flexion at foot contact, the knee joint was extending and the quadriceps contracted concentrically (shortening of muscle). At foot contact during stair descent, the hip and knee was close to neutral, the knee joint was undergoing flexion and the quadriceps contracted during muscle lengthening (eccentric contraction). The muscle inhibition occurring in the VL muscle only, demonstrated that the activity of VL could be selectively decreased relative to other parts of the

Normalised group averaged mean EMG - Step down 80% 70% 60%

%Peak

50% 40% 30% 20%

Mean Untaped Mean Tape Mean Control tape

10% 0% -150.0

-100.0

-50.0

0.0

50.0

100.0

150.0

Time (ms) Fig. 4. Normalised group averaged mean EMG of the VL during stepdown 150 ms pre- and post-foot contact.

quadriceps assessed, i.e. the VMO. It has previously been suggested that inhibition of the quadriceps muscle may be due to activation of a flexion withdrawal reflex (Leroux et al., 1995). The withdrawal reflex increases activation of the hamstrings muscle with a reciprocal inhibition of the quadriceps. In this study there was no statistically significant change in the biceps femoris IEMG in response to tape applied ( p > 0.05), therefore this theory of quadriceps muscle inhibition does not seem likely in this case. There was no significant difference ( p > 0.05) in pre-foot contact VL IEMG muscle activity during the selected 150-ms period during stair ascent or descent. Tobin and Robinson (2000) detected a small non-significant increase in VL EMG in response to control tape. In the present study there was a decrease in EMG to both control tape and VL inhibitory tape. The differences in results between the previous study (Tobin and Robinson, 2000) and this study are difficult to explain, as the VL inhibitory tape led to a decrease in EMG in both studies but the control tape led to a different outcome. The method and site of tape application are the same in both studies except for the tension applied on the rigid tape which was not ‘‘measured’’ in the previous study. A possible reason for this disparity may be due to the differences in study methodology, tape tension and EMG data gathering and analysis as the control tape method and application site was the same in both studies. In the study by Janwantankul and Gaogasigam (2005), an elastic tape was applied parallel and perpendicular to the fibres of VL during stair descent. The study subjects descended a single step ‘‘staircase’’ and stepped backwards up again repeatedly without controlling the pace of gait and used non-normalised EMG data. Using non-normalised EMG data, no significant changes in mean EMG activity was noted. It is possible that the taping method and EMG analysis may have affected the outcome as both Tobin and Robinson (2000) and the present study noted a decrease in the VL EMG activity.

336

U. McCarthy Persson et al. / Manual Therapy 14 (2009) 330e337

Tobin and Robinson (2000) postulated that the VL inhibitory tape might decrease the EMG activity, by stimulation of afferent C-fibres (nociceptive) causing descending alpha-motor neurone inhibition. This does not seem a likely mechanism in this study as no pain was experienced from either of the tapes (VAS ¼ 0) during the stair climbing. As a similar change occurred in muscle inhibition with both the VL inhibitory tape and the control tape in the this study, it is more likely that the alterations noted could be due to cutaneus stimulation from either of the tapes applied over the muscle. The findings of this study noted similar effects on IEMG from both VL inhibitory tape and control tape. This evidence may indicate that the relatively large skin displacement occurring with the VL inhibitory tape is not necessary to alter muscle activity and kinematics in normal subjects. No apparent skin displacement occurred with the application control tape at rest, but it is possible that the change in IEMG could result from the contact of the tape on the skin. A tension of the cutaneus tissues may also have occurred from either tape of the applications during the activity of stair climbing. It has been demonstrated that stimulation of mechanoreceptors in the skin can modify efferent alfa-motor neurone activity (Garnett and Stephens, 1981). It has more recently been found that application of patella tape causing tension on the skin can modify the EMG activity of the VMO and that the motor changes were dependent on the direction of the tension applied (MacGregor et al., 2005). There was no significant change in muscle activity from tape application in the pre-contact phase during these analyses. The pre-contact muscle activity of VL during the stair climbing was very low. It may be possible that the taping only affects the muscle activity during a relatively greater muscle activation such as during the weight acceptance phase. The results of this study indicate that in subjects without knee pathology, the VL muscle can be selectively inhibited by tape applied on the muscle by VL inhibitory or control tape. The decrease in VL IEMG found in this study occurred during the initial weight acceptance phase as the foot comes in contact with the step and the weight is transferred through the lower limb. In the clinical setting VL inhibitory taping technique is used in subjects with patellofemoral pain syndrome (PFP) with a proposed relative increased activity of the VL in relation to the VMO muscle (Mariani and Caruso, 1979; Voight and Wieder, 1991; Thomee et al., 1995). This inhibitory effect of the VL, resulting from VL inhibitory tape and control tape, may be useful to address muscle imbalances in subjects with decreased activity of the VMO or increased activity of the VL. Due to the normalising the EMG data to the control condition rather than a maximal voluntary contraction,

it is not possible to quantify the amount of inhibition resulting from the tape applications. It is therefore difficult to draw conclusions in regards to if the reduction in muscle activity found in this study is clinically relevant. Therefore, future studies need to assess the effects of control tape and VL inhibitory in a patient group with altered vastii activation pattern or PFP.

6. Conclusion This work demonstrates that VL inhibitory tape and control tape decrease EMG activity of the VL during stair climbing in normal subjects. The findings of this study noted similar effects on IEMG from both VL inhibitory tape and control tape. This evidence indicates that the large skin displacement occurring with the VL inhibitory tape is not necessary to alter muscle activity in normal subjects. Future studies are needed to assess the VL inhibitory and control taping techniques in subjects with PFP or abnormal quadriceps muscle activity.

References Alexander CM, Stynes S, Thomas A, Lewis J, Harrison PJ. Does tape facilitate or inhibit the lower trapezius. Manual Therapy 2003;8(1):37e41. Alexander MA, McMullan M, Harrison PJ. What is the effect of taping along or across a muscle on a motorneurone exitability? A study using the triceps surae. Manual Therapy 2008;13:57e62. Caulfield B, Crammond T, O’Sullivan A, Reynolds S, Ward T. Altered ankle-muscle activation during jump landing in participants with functional ankle instability of the ankle joint. Journal of Sport Rehabilitation 2004;13:189e200. Costigan PA, Deluzio KJ, Wyss UP. Knee and hip kinetics during normal stair climbing. Gait and Posture 2002;16(1):31e7. Cowan S, Bennell K, Hodges P. The test-retest reliability of the onset of concentric and eccentric vastus medialis obliquus and vastus lateralis electromyographic activity in a stair stepping task. Physical Therapy in Sport 2000;1:129e36. Cowan SM, Bennell KL, Hodges PW, Crossley KM, McConnell J. Delayed onset of electromyographic activity of vastus medialis obliquus relative to vastus lateralis in subjects with patellofemoral pain. Archives of Physical Medicine and Rehabilitation 2001;82:183e9. Cowan SM, Bennell KL, Hodges PW. Therapeutic patellar taping changes the timing of vasti muscle activation in people with patellofemoral pain syndrome. Clinical Journal of Sports Medicine 2002;12(6):339e47. Garnett R, Stephens JA. Changes in the recruitment of motor units produced by cutaneus stimulation in man. Journal of Physiology 1981;311:463e73. Gilleard W, McConnell J, Parsons D. The effect of patellar taping on the onset of vastus medialis oblique and vastus lateralis muscle activity in persons with patellofemoral pain. Physical Therapy 1998;78:25e32. Hamill J, Knutzen KM. Linear kinematics, Biomechanical basis of human movement. Baltimore: Williams and Wilkins; 1995. 346e347. Hugon M. Methodology of the H-reflex in man, vol. 3. Basel: Karger; 1973.

U. McCarthy Persson et al. / Manual Therapy 14 (2009) 330e337 Janwantankul P, Gaogasigam C. Vastus lateralis and vastus medialis obliquus muscle activity during the application of inhibition and facilitation taping techniques. Clinical Rehabilitation 2005;19:12e9. Leroux A, Belanger M, Boucher JP. Pain effect on monosynaptic and polysynaptic reflex inhibition. Archives of Physical Medicine and Rehabilitation 1995;76(6):576e82. MacGregor K, Gerlach S, Mellor S, Hodges PW. Cutaneus stimulation from patella tape causes a differential increase in vasti muscle activity in people with patellofemoral pain. Journal of Orthopedic Research 2005;23:351e8. Mariani P, Caruso I. An electromyographic investigation of subluxation of the patella. Journal of Bone and Joint Surgery 1979;61B:169e71. Marque P, Nicolas G, Marchand-Pauvert V, Gautier J, SimonettaMoreau M, Pierrot-Deseilligny E. Group I projections from intrinsic foot muscles to motor neurones of leg and thigh muscles in humans. Journal of Physiology 2001;536(1):313e27. McCarthy Persson JU, Hooper ACB, Fleming HE. Repeatability of skin displacement and pressure during ‘‘inhibitory’’ vastus lateralis muscle taping. Manual Therapy 2007;12:17e21. McFadyen BJ, Winter DA. An integrated biomechanical analysis of normal stair ascent and descent. Journal of Biomechanics 1988;21(9):733e44. Morrissey D. Proprioceptive shoulder taping. Journal of Bodywork and Movement Therapies 2000;4(3):189e94. Rossi A, Decchi B. Cutaneous nociceptive facilitation of Ib heteronymous pathways in the lower limb motorneurones in humans. Brain Research 1995;700:164e72.

337

SENIAM (Surface EMG for Non-Invasive Assessment of Muscles) Research Group. European recommendations for surface electromyography e results of the SENIAM project. Enschede, The Netherlands: Roessingh Research and Development; 1999. Swanik CB, Lephart SM, Giraldo JL, Demont RG, Fu FH. Reactive muscle firing of anterior cruciate ligament-injured females during functional activities. Journal of Athletic Training 1999;34(2): 121e9. Tobin S, Robinson G. The effect of vastus lateralis inhibition taping technique on vastus lateralis and vastus medialis obliquus activity. Physiotherapy 2000;86(4):173e83. Thomee R, Renstro¨m P, Karlsson J, Grimby G. Patellofemoral pain syndrome in young women. Scandinavian Journal of Medicine Science and Sports 1995;5:245e51. Voight ML, Wieder DL. Comparative reflex response times of vastus medialis obliquus and vastus lateralis in normal subjects and subjects with extensor mechanism dysfunction. American Journal of Sports Medicine 1991;19:131e7. Wilkerson GB. Biomechanical and neuromuscular effects of ankle taping and bracing. Journal of Athletic Training 2002;37(4):436e45. Witvrouw E, Sneyers C, Lysens R, Bellemans J. Reflex response times of vastus medialis oblique and vastus lateralis in normal subjects and in subjects with patellofemoral pain syndrome. Journal of Orthopedic and Sports Physical Therapy 1996;24:160e5. Yang JF, Winter DA. Electromyographic amplitude normalization methods: improving their sensitivity as diagnostic tools in gait analysis. Archives of Physical Medicine and Rehabilitation 1984;65 (9):517e21.

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 338e345 www.elsevier.com/math

Original Article

Neckeshoulder muscle activity in general and task-specific resting postures of symptomatic computer users with chronic neck pain Grace Pui Yuk Szeto a,b,*, Leon Melville Straker b, Peter Bruce O’Sullivan b a

Department of Rehabilitation Sciences, Hong Kong Polytechnic University, Hong Kong SAR, China b School of Physiotherapy, Curtin University of Technology, Perth 6845, Australia Received 15 October 2007; received in revised form 17 April 2008; accepted 6 May 2008

Abstract Past research on work-related musculoskeletal disorders (WMSD) has frequently examined the activity of neckeshoulder muscles such as upper trapezius (UT) and cervical erector spinae (CES) during typing tasks. Increased electromyographic activity in these postural stabilising muscles has been consistently found in chronic neck pain patients under different physically stressful conditions. The present study compared muscle activity when female office workers with chronic neck pain (n ¼ 39) and asymptomatic controls (n ¼ 34) adopted two resting postures: (1) with hands on laps versus; and (2) hands on a keyboard. Resting hands on keyboard elicited significantly increased muscle activity in the right UT of subjects with high discomforts (n ¼ 22), similar to that observed during actual typing. In contrast, the asymptomatic controls showed no difference in muscle activity between the resting postures. This result suggested that altered muscle activation patterns were triggered by some anticipatory task demand associated with a task-specific position in some individuals. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Work-related musculoskeletal disorders; Surface electromyography; Computer use; Keyboard; Neck pain; Trapezius

1. Introduction Work-related musculoskeletal disorders (WMSD) are a significant health problem worldwide, with intensive computer users particularly affected (Tittiranonda et al., 1999; Buckle and Devereux, 2002; Gerr et al., 2004). Many studies have examined activity in the trapezius muscle in order to understand the biomechanical exposure in this postural stabiliser and its contribution to WMSD in computer users. Higher amplitudes of muscle activity, less varied muscle activity and fewer periods of no muscle activity have been suggested to be important biomechanical risk factors (Veiersted * Corresponding author. Department of Rehabilitation Sciences, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong SAR, China. Tel.: þ852 27666706; fax: þ852 23308656. E-mail address: [email protected] (G.P. Yuk Szeto). 1356-689X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.05.001

et al., 1990; Ha¨gg and Astrom, 1997; Nordander et al., 2000). Studies on chronic neck pain patients have also shown changes in muscle activation patterns during functional activities compared to healthy control subjects (Madeleine et al., 1999; Falla et al., 2004). Recently, Szeto et al. (2005aee) presented quasiexperimental evidence that individuals with WMSD had different muscle activation patterns and different kinematics during computer use compared to individuals without a disorder. Consistently, patterns of increased muscle activity in the upper trapezius were found in symptomatic participants who performed typing tasks under a variety of conditions. The higher muscle activities in the symptomatic participants remained the same even when pain increased significantly in this group suggestive of inherent altered motor patterning associated with the occupational task. It was also found that symptomatic subjects had about five degrees more neck

G.P. Yuk Szeto et al. / Manual Therapy 14 (2009) 338e345

flexion compared to the asymptomatic controls and this remained unchanged when pain increased during prolonged typing (Szeto et al., 2005b). The etiology of WMSD is likely to be multi-factorial, comprising of a combination of physiological, biomechanical, psychological and social factors (Feuerstein et al., 2004; Gerr et al., 2004), and these factors may offset or modify the influence of other factors. There is growing evidence that similar factors may relate to different types of neck and shoulder pain (Nederhand et al., 2000; Sterling et al., 2001; Falla, 2004; Jull et al., 2004). In studies on whiplash-associated and idiopathic neck pain patients, the more superficial muscles in the neckeshoulder region (namely upper trapezius, sternocleidomastoid and anterior scalene muscles) showed increased activity compared to deeper postural stabilisers like the deep cervical flexors (Falla, 2004). Whilst altered motor patterns may expose the cervical spine to excessive loading forces that may be pain provocative, it is not clear whether the altered patterns are induced in response to specific task demands or are a more generalised manifestation of the pain disorder. Some evidence of increased muscle activity in symptomatic persons during long periods of resting and sleeping was reported. Holte and Westgaard (2002) and Mork and Westgaard (2004) suggested that altered muscle activity patterns may be habitually adopted in individuals with chronic pain. However, these results did not clearly identify specific increases of muscle activity with different postures or tasks, and it is not clear whether such mechanisms were already present before the onset of pain. To clarify whether the reported motor control differences are present for symptomatic individuals just in work specific situations or are a more generalised state, we compared the muscle activation in the cervical erector spinae and upper trapezii muscles of symptomatic and asymptomatic office workers in different resting postures. The posture of resting hands on lap was compared to the task-specific situation of resting hands on a keyboard. We also compared the muscle activity responses during resting to that following an increase in pain related to typing. The results may provide a better understanding of the inherent muscle activation and motor control mechanisms in different individuals when the subject is not fully engaged in functional activities, leading to enhanced management of neck pain disorders.

2. Methods 2.1. Participants Seventy-three female office workers were recruited by convenience sampling. Participants had regular sustained exposure to computer work, with a minimum of

339

2e4 h daily, of mainly clerical tasks. Volunteers were assigned into either Case or Control Groups according to the results of a questionnaire-interview about their work habits and their past and present history of WMSD. Case participants (n ¼ 39) had more than 3 months of discomfort in the past year and had discomfort in the past 7 days and discomfort on the day of testing of more than 2/10 in any one or more areas in the neck, shoulder, upper back, elbow, wrist/hand region. Control participants (n ¼ 34) had less than 3 months of discomfort in the past year, no discomfort in the past 7 days or on the day of testing. Case and Control Groups were reasonably well matched in terms of their physical build and work histories; however, there was a significant difference in their mean age (Case Group ¼ 37.4  6.1; Controls ¼ 30.9  6.5). Subsequent analysis did not reveal any significant covariant effect of age. All participants provided informed consent and the study were approved by the Hong Kong Polytechnic University Human Ethics committee. 2.2. Surface EMG measurements and procedures Electrical muscle activity was collected from the cervical erector spinae (CES) and the upper trapezius (UT) muscles using the Noraxon Telemyo system (Noraxon, USA Inc., USA), with a sampling frequency of 1000 Hz and a bandwidth of 10e500 Hz. Bipolar AgeAgCl surface electrodes (3MÔ Infant Red DotÔ, 15 mm in diameter, 3M Hong Kong Ltd., Hong Kong) were used, with an inter-electrode distance of 20 mm. The skin was carefully prepared to reduce the skin impedance to <2 KU. The locations of electrodes and muscle actions used for amplitude normalisation were previously tabled in Szeto et al. (2005a,c). For the CES muscles, the electrodes were positioned at about 1 cm distance lateral to the C4 and C5 spinous processes bilaterally. For UT muscles, the mid-point between the two electrodes was located at half the distance between the acromion arch and the T1 spinous process. The EMG normalisation procedures consisted of three trials of maximum voluntary exertions (MVE) and one trial of 0e30% ramp contraction. Each contraction lasted for 5 s. A special chair was constructed for this process. For testing the MVE for the CES muscles, the subject had to perform resisted neck extension against a loadcell positioned at the occiput with her maximal effort. The loadcell was fixed on a steel bar which was adjusted to subject’s height. For testing the MVE for the UT muscles, the subject had to perform resisted shoulder elevation against a shoulder strap (connected to a loadcell fixed to the floor) with maximal effort. For the CES muscles, both muscles were tested simultaneously; for the UT muscle the two sides were tested separately. Using a custom Labview (National InstrumentsÔ, Austin, USA) program, the EMG signals

G.P. Yuk Szeto et al. / Manual Therapy 14 (2009) 338e345

were processed with a high-pass filter at 20 Hz, low-pass filter at 200 Hz, and notch filters at 50 and 60 Hz. The raw signals were down-sampled to 10 Hz RMS values and normalised to MVE and expressed as the percentage of maximum EMG (%MEMG). 2.2.1. EMG measurements in two resting postures Muscle activity was collected during two resting conditions. A general resting posture in relaxed sitting was recorded before the participants performed the MVE procedures. They were instructed to rest their hands (palm down) on their thighs in a seated position, and EMG was captured for 15 s. Subjects kept their upper arms by the side of the body with the weight of the upper extremities supported through the hands on the thighs. With their back resting against the chair, participants were instructed to ‘‘relax, keep the head in upright position looking straight ahead, and do not move’’. The task-specific resting posture was assumed after MVE procedures. Participants were seated at a computer workstation and instructed to adjust the seat height, keyboard and monitor position to achieve a comfortable and reasonably erect posture. Participants were instructed to rest their fingers on the home row of the keyboard in a relaxed manner, and EMG was again captured for 15 s. The subject’s forearms were supported on the curved padding of the keyboard tray (about 6 cm front to back) in front of the keyboard, and this forearm support was maintained in the resting EMG data capture as well as during the subsequent typing trials. The same instruction for staying relaxed and motionless was given for the hands on keyboard resting trial. Following the task-specific resting trial, 43 participants performed continuous typing (previously published work) for 1 h (Szeto et al., 2005a) and 41 participants performed three typing trials of 20 min each, one with normal speed and force, one with increased speed and one with increased force (Szeto et al., 2005c). Resting trial data from both studies are combined in this paper due to identical participant recruitment procedures and similar muscle activity measures. Eleven women participated in both studies, which accounted for the total participant number being 73 in the present report. For these 11 participants, the resting EMG data from the first study were used. The same computer workstation was used throughout. Subjects were instructed to adjust the chair height in order to achieve a reasonably erect but relaxed posture. 2.2.2. Subjective discomfort ratings At the start of each experiment, participants were asked to rate their discomfort in 10 upper body areas (left and right neck, upper back, shoulders, elbows, wrist/hands) on a numerical scale of 0e10 (Kuorinka et al., 1987). In the numerical scale, 0 was considered

no discomfort, 1 was minimal discomfort and 10 was extreme/intolerable discomfort. The participants were asked to rate their discomfort before EMG recording of the resting conditions, as well as periodically during the various typing tasks. The discomfort scores from the various body areas were summed at each rating and this was compared between the resting and typing conditions. 2.3. Data analysis The median muscle activity, 50th percentile (%) Amplitude Probability Distribution Function (APDF), was computed for left and right CES and UT for Case and Control subjects during general and task specific resting. Statistical analysis using MANOVA followed by univariate mixed model ANOVA were conducted to test the effects of resting condition (‘‘rest’’), side and group. Subsequent analysis separated the Case Group into those with higher and lower levels of discomfort as performed in related analysis (Szeto et al., 2005aee).

3. Results 3.1. Muscle activity in general and task-specific resting postures The median muscle activity for CES during the general and task-specific resting postures were generally around 15e20%MEMG, with UT levels around 5e 15%MEMG (Fig. 1 and 2). Large group standard deviations existed for all muscles. Right UT appeared to be higher than the left UT for the Case Group in the taskspecific resting posture. The multivariate analysis showed significant effects for rest and side, with a trend for a group  side interaction (see Table 1). The univariate analysis showed no 55 50

50th% APDF (% MEMG)

340

RCES LCES RUT LUT

45 40 35 30 25 20 15 10 5 0 Case

Control

Fig. 1. Muscle activities in the general resting posture (50th% APDF is expressed in terms of %MEMG). RCES, right cervical erector spinae; LCES, left cervical erector spinae; RUT, right upper trapezius; and LUT, left upper trapezius.

50th% APDF (% MEMG)

G.P. Yuk Szeto et al. / Manual Therapy 14 (2009) 338e345 55 50 45 40 35 30 25 20 15 10 5 0

RCES LCES RUT LUT

Case

Control

Fig. 2. Muscle activities in the task-specific resting posture. RCES, right cervical erector spinae; LCES, left cervical erector spinae; RUT, right upper trapezius; and LUT, left upper trapezius.

effects for CES but significant rest, side and group  side interaction effects for UT. This affirmed the higher right UT activity for Cases but not Controls, as observed in Figs. 1 and 2. However, the lack of a group  side  rest interaction suggests that higher right UT activity occurred in both resting conditions, despite RUT being three times higher than LUT in task-specific resting compared with only two times higher in the general resting condition.

3.2. Discomfort levels in Case and Control Groups At the start of each study in the resting conditions, Case participants had a greater mean summed discomfort score of 9.5 (8.2), than Controls (0.5  1.6) (P < 0.001, unequal variances). Case participants also Table 1 Summary of statistical analysis results comparing CES (cervical erector spinae) and UT (upper trapezius) muscle activities in Case and Control Groups. Multivariate

Univariate (CES)

Univariate (UT)

Rest condition ( general F2,70 ¼ 6.01, P ¼ 0.004* vs task-specific)

F1,71 ¼ 0.68, P ¼ 0.413

F1,71 ¼ 7.59, P ¼ 0.007*

Side

F2,70 ¼ 4.36, P ¼ 0.016*

F1,71 ¼ 1.29, P ¼ 0.260

F1,71 ¼ 7.14, P ¼ 0.009*

Group

F2,70 ¼ 0.51, P ¼ 0.601

F1,71 ¼ 0.37, P ¼ 0.544

F1,71 ¼ 0.39, P ¼ 0.536

Rest  side

F2,70 ¼ 2.26, P ¼ 0.112

F1,71 ¼ 2.91, P ¼ 0.093

F1,71 ¼ 2.60, P ¼ 0.111

Rest  group

F2,70 ¼ 1.73, P ¼ 0.185

F1,71 ¼ 0.63, P ¼ 0.429

F1,71 ¼ 1.48, P ¼ 0.228

Side  group

F2,70 ¼ 3.00, P ¼ 0.056

F1,71 ¼ 0.02, P ¼ 0.879

F1,71 ¼ 6.08, P ¼ 0.016*

Rest  side  group

F2,70 ¼ 1.25, P ¼ 0.292

F1,71 ¼ 0.24, P ¼ 0.625

F1,71 ¼ 1.93, P ¼ 0.169

*Significance level P < 0.05.

341

had greater discomfort during the typing trials (P < 0.001). Examining the discomfort score patterns throughout the typing trials within the Case Group showed two distinct sub-groups e High Discomfort Group and Low Discomfort Group. The cut-off point for the High Discomfort Group was a mean score of 12 or above. There were significant differences in the discomfort scores among the three groups (P < 0.010), and between these two pain groups and the Control Group (Dunnett T3, P < 0.010). Examining the locations of the discomforts in the neck and shoulder areas revealed that majority of the Case subjects had bilateral discomforts (n ¼ 17 for both neck areas, n ¼ 14 for both shoulders). For those with unilateral discomforts, there was actually a trend for somewhat higher scores in the left neck and shoulder compared to the right. There were no change of discomfort scores between the two resting trials hence the resting discomforts were examined as one set of data (Tables 2 and 3). 3.3. General and task-specific resting muscle activity in High and Low Discomfort sub-groups The High Discomfort Group showed the greatest increase of right UT activity in the task-specific resting posture, to 21.8%MEMG (29.7) compared to the Low Discomfort Group (6.2%  9.4) and the Control Group (7.5%  10.0). The High Discomfort Group also showed a trend for the most increase in right UT activity from general resting (8.9%MEMG  14.5) to task-specific resting (21.8  29.7). Interestingly, the Low Discomfort Group (n ¼ 21) seemed to have even lower muscle activity than the Control Group (n ¼ 37) in nearly all situations (see Fig. 3 and 4). In the statistical analysis, when the group factor was changed to three levels comparing the High Discomfort, Low Discomfort and Control Groups, the results (Table 4) were similar to some extent to the first analysis in Table 1, except that, the multivariate analysis showed significant effects for the three-way interaction of rest  side  group, and the two-way interaction of side  group, in addition to the significant effects for rest and side factors. The univariate analysis again showed no significant effects for CES for all factors, but significant effects for rest condition, side, side  group and the rest  side  group 3-way interaction for UT (see Table 4). The High Discomfort Group showed a significant increase in RUT activity in the task-specific condition compared to general resting condition (t21 ¼ 2.331, P ¼ 0.030). Using independent sample t-tests, the right UT in the High Discomfort Group (21.8  29.7) was statistically greater than the right UT activity in Low Discomfort Group (6.2  9.4) in the task-specific condition only (t26.2 ¼ 2.32, P ¼ 0.028). This result further supported that the increase in RUT muscle activity in the

342

G.P. Yuk Szeto et al. / Manual Therapy 14 (2009) 338e345

Table 2 Summary of mean discomfort scores of subjects comparing resting discomforts and discomfort scores in different experimental tasks.

Case Group Two sub-groups

High Discomfort Low Discomfort

Control Group

Resting discomfort scores (summed)

Summary discomfort scores in different experimental tasks 1-h typing

Normal typing speed and force

Increased speed

Increased force

9.52  8.24 13.00  9.53 5.71  4.22 0.53  1.55

13.84  7.70 18.60  8.32 4.92  1.99 1.38  4.09

14.16  4.50 25.33  13.75 7.28  4.73 1.08  1.93

26.37  3.44 32.15  13.66 23.00  10.43 1.65  2.50

25.44  2.07 34.35  16.17 19.95  12.15 1.68  2.34

Note: values are mean  sd.

High Discomfort Group was significantly greater in the task-specific condition.

observed during typing (Szeto et al., 2005a), even though subjects were instructed to ‘‘relax and stay motionless’’ and had their forearms and hands fully supported. This muscle activity pattern was observed despite no change in discomfort between the two resting positions. The two ‘‘resting’’ postures in the present study may have activated different motor control mechanisms in different individuals, as resting hands on laps may not be perceived as functional or work-related. In contrast, resting hands on the keyboard involved contact of the keys with the fingertips and, together with the mental image of the imminent work task, may have already triggered some motor response associated with typing work. The observed difference in the High Discomfort Group may also represent an anticipatory motor response triggered by the memory or anticipation of pain related to typing. This may represent a programmed motor response related to a pain provocative task. To further test this, other non-provocative functional tasks would have to be assessed in order to determine if the motor response in this group is specifically related to pain provocation, or simply related to upper limb function in general. The fact that this motor response was not reflected in the Low Discomfort Group may indicate that the

4. Discussion 4.1. Muscle activation in general and task-specific resting postures The present results demonstrate differences in muscle activation patterns when individuals rested their hands on their laps compared to resting their hands on the keyboard. When participants rested their hands on their laps, Case and Control Groups had quite similar muscle activities, except that the High Discomfort Group already had a trend for increased right UT muscle activity. However, when symptomatic participants rested their hands on the keyboard, there was a marked increased in the right UT muscle activity in those individuals who reported high discomfort during the subsequent typing tasks. This suggests that even when the participants were instructed to stay ‘‘relaxed’’ when resting their hands on the keyboard, some motor control mechanisms associated with the imminent typing task were already activated. Further, the response of symptomatic subjects during task-specific resting was similar to that

Table 3 Summary of discomforts in the neck and shoulder areas of Case and Control Groups. Resting discomforts by area

Neck

Shoulders

Left only

Right only

Both sides

Left only

Right only

Both sides

Case Group (n ¼ 39)

Number of subjects Mean score sd

5 5.10 1.98

6 3.57 2.28

17 3.13 1.84

7 5.00 2.33

10 3.56 1.76

14 2.63 1.35

High Discomfort Group (n ¼ 22)

Number of subjects Mean score sd

3 5.33 2.75

4 4.38 2.87

11 3.66 1.95

4 5.63 1.97

6 3.75 1.94

9 3.33 1.16

Low Discomfort Group (n ¼ 17)

Number of subjects Mean score sd

2 4.75 0.35

3 2.50 0.50

5 1.95 0.76

3 4.17 2.93

2 3.00 1.41

6 1.58 0.85

Control Group (n ¼ 34)

Number of subjects Mean score sd

0 0 0

1 1.00 0

3 1.50 0.77

0 0 0

3 1.83 0.29

1 2 0

343

G.P. Yuk Szeto et al. / Manual Therapy 14 (2009) 338e345 55 RCES LCES RUT LUT

50th% APDF (% MEMG)

50 45 40

Multivariate

Univariate (CES)

Univariate (UT)

Rest condition ( general F2,69 ¼ 7.15, P ¼ 0.002* vs task-specific)

F2,69 ¼ 1.06, P ¼ .307

F2,69 ¼ 8.33, P ¼ 0.005*

Side

F2,69 ¼ 5.82, P ¼ 0.005*

F2,69 ¼ 0.82, P ¼ 0.368

F2,69 ¼ 10.28, P ¼ 0.002*

Group (High, Low, Control)

F4,138 ¼ 1.37, F2,70 ¼ 0.92, P ¼ 0.601 P ¼ 0.403

F2,70 ¼ 2.22, P ¼ 0.116

Rest  side

F2,69 ¼ 2.37, P ¼ 0.100

F2,69 ¼ 3.26, P ¼ 0.075

Rest  group

F4,138 ¼ 1.90, F4,138 ¼ 0.32, F4,138 ¼ 2.42, P ¼ 0.114 P ¼ 0.729 P ¼ 0.096

Side  group

F4,138 ¼ 3.12, F4,138 ¼ 0.80, F4,138 ¼ 5.42, P ¼ 0.017* P ¼ 0.453 P ¼ 0.006*

Rest  side  group

F4,138 ¼ 2.81, F4,138 ¼ 1.29, F4,138 ¼ 3.24, P ¼ 0.028* P ¼ 0.283 P ¼ 0.045*

35 30 25 20 15 10 5 0 High Discomfort Group

Low Discomfort Group

Control Group

Fig. 3. Comparing muscle activities in the High and Low Discomfort Group and the Control Group in the general resting posture. RCES, right cervical erector spinae; LCES, left cervical erector spinae; RUT, right upper trapezius; and LUT, left upper trapezius.

presence of higher levels of pain mediate this motor response in some way. The Low Discomfort Group may represent a different sub-group of patients with a different motor response to the neck and shoulder pain disorder. Clinical studies on low back pain have found different ‘‘sub-groups’’ based on EMG patterning (Dankaerts et al., 2007). There are also other factors that should be considered beside the proposed anticipatory feed-forward motor control mechanism. One of these factors would be the postural habits of the individuals, which are likely to be interdependent to muscle activities for controlling movements. Even when subjects were given similar instructions to maintain the two resting postures, there may still be some differences in the individual postures assumed, which may have contributed to some of the variations in their muscle activities. A limitation of the study was that postural data were not collected

55 RCES LCES RUT LUT

50

50th% APDF (% MEMG)

Table 4 Summary of statistical analysis results when Case Group subjects were sub-divided into High Discomfort and Low Discomfort Groups.

45 40 35 30 25 20 15 10 5 0 High Discomfort Group

Low Discomfort Group

Control Group

Fig. 4. Comparing muscle activities in the High and Low Discomfort Groups and the Control Group in the task-specific resting posture. RCES, Right cervical erector spinae; LCES, left cervical erector spinae; RUT, right upper trapezius; and LUT, left upper trapezius.

F2,69 ¼ 2.78, P ¼ 0.099

*Significance level P < 0.05.

concurrently with the muscle activity data during the resting trials, although they were collected during the typing trials. In the 1 h typing trial, the symptomatic group had about five degrees greater neck flexion compared to the asymptomatic group. It was proposed in our previous research that the neck posture changes and the increased muscle activities were all closely related phenomena in the altered motor control mechanisms contributing to the musculoskeletal symptoms of the Case subjects (Szeto et al., 2005b). Although it is known that the high discomfort subjects had higher levels of pain intensity, it is unlikely that the motor response was directly related to pain intensity levels, as the motor response observed during the typing task did not change, in spite of increasing levels of pain reported during the task (Szeto et al., 2005a,c). The evidence for altered muscle activation has been consistent across research on different types of neck pain patients performing upper limb functional tasks (Nederhand et al., 2000, 2003; Falla et al., 2004; Jull et al., 2004). These studies have reported increased motor responses induced by pain provocative tasks and proposed that the muscle hyperactivity phenomenon was common to different types of neck pain regardless of traumatic origin or not. Cognitive data such as fear avoidance beliefs were not assessed in this study, but could be a focus for future research to determine if this plays a role in this process (Fig. 5). Some evidence of muscle activity in a general resting situation came from studies that examined EMG over long durations comparing work, leisure and sleep (Holte and Westgaard, 2002; Mork and Westgaard, 2004). These studies also reported consistently higher muscle activity patterns in symptomatic participants compared

344

G.P. Yuk Szeto et al. / Manual Therapy 14 (2009) 338e345

to asymptomatic individuals. However, muscle activities were averaged over long periods and specific tasks or postures were not identified. Hence these studies could not be directly compared to the results of the present study. Studies on occupational tasks comparing symptomatic individuals with neck and shoulder pain and asymptomatic workers have also reported similar results of consistently higher muscle activities in upper trapezius and other cervical muscles (Madeleine et al., 1999; Ha¨gg and Astrom, 1997). Besides higher activity levels, it was also reported that persons with neck pain had an inability to relax the neck muscles even upon task completion (finger tapping on targeted dots) (Falla et al., 2004). This concept of muscles being unable to relax may result in increased tissue loading in cervical structures and thereby be inherently provocative for chronic neck pain associated with pain sensitive structures. 4.2. Clinical implications The altered motor patterning observed in the current study may be provocative in nature as a higher level of UT activity could expose possibly sensitised cervical spine structures to increased mechanical loading (Graven-Nielsen and Mense, 2001). Increased activation of the UT has also been documented in neck and arm pain of a neurogenic nature and has been reported to represent an adaptive (protective) response to sensitised neural structures (Graven-Nielsen and Mense, 2001; van der Heide et al., 2001). The present study has found much higher muscle activities in the right UT muscle which was also consistently observed during typing trials (Szeto et al., 2005a,c). The unilateral effect may relate to mouse use as results may reflect the rising trend for intensive mouse use with the right hand (Cook et al., 2000). Another

50th% APDF (% MEMG)

55 50

RCES LCES RUT LUT

45 40

factor contributing to higher muscle activation may be due to inherent postural or movement differences in some individuals. It is also possible that those in the High Discomfort Group may have had a poor ergonomic setup in their daily work with computers, contributing to the development of chronic muscle activity changes and pain disorders. This factor would require a field investigation and more long-term observations to confirm the hypothesis. The high discomfort group displayed predominantly a bilateral neck and shoulder pain distribution. While such pain may arise from the local muscles, other possible causes may include loading of sensitised articular structures in the cervical spine such as the zygapophyseal joints, which could have explained the increased generalised pain in the distal forearm regions indicative of a referred pain pattern (Bogduk, 1995). On the other hand widespread pain has also been associated with regional pain disorders involving changes to the central nervous system pain processing (Graven-Nielsen and Mense, 2001). In reality, the contribution of motor control mechanisms to the development of musculoskeletal pain disorders is likely to involve multiple contributing factors such as behavioural responses to work, stress and pain as well as morphological processes affecting the muscles and other physiological functions (Waersted, 2000; Feuerstein et al., 2004). These factors provide the basis for future research. The present results are in line with recent research findings showing altered feed-forward mechanisms in the motor control of neck muscles in patients with chronic neck pain (Falla et al., 2004). It has been suggested that rehabilitative exercise training should incorporate ‘‘task-specific practice’’ such as typing, with variations in speed or range of movement, so that these variations could facilitate the cortical connections associated with such variations in task performance (van Vliet and Heneghan, 2006). If the motor patterns observed in high discomfort participants are provocative for their disorder, then training relaxation of the UT muscle in conjunction with postural reeducation both at rest and during typing task would be a logical management approach.

35 30

5. Conclusions

25 20 15 10 5 0 High Discomfort Group

Low Discomfort Group

Control Group

Fig. 5. Muscle activities of the High and Low Discomfort Groups and Control Group in the 1-h typing study. RCES, right cervical erector spinae; LCES, left cervical erector spinae; RUT, right upper trapezius; and LUT, left upper trapezius.

The present study has compared the muscle activities of UT and CES when individuals rested their hands on their laps versus resting hands on keyboard. Similar muscle activity levels were found in Case and Control Groups with hands resting on laps. However, when resting hands on the keyboard, there was greater muscle activity in the right UT of the Case Group, especially in the High Discomfort Group. This increased muscle activity pattern was similar to that observed during actual

G.P. Yuk Szeto et al. / Manual Therapy 14 (2009) 338e345

typing. This may indicate that symptomatic individuals had pre-programmed motor control mechanisms that were triggered by imminent, as well as actual, work task demands. This finding may have important implications in understanding the physiological mechanisms contributing to the development of WMSD and management.

Acknowledgement This project has been supported by the Departmental Research Grant of the Department of Rehabilitation Sciences, the Hong Kong Polytechnic University. The authors would also like to acknowledge the work of Mr. Paul Davey on developing the Labview Program for analysing the EMG data.

References Bogduk N. The anatomical basis for spinal pain syndromes. Journal of Manipulative and Physiological Therapy 1995;18(9):603e5. Buckle PW, Devereux JJ. The nature of work-related neck and upper limb musculoskeletal disorders. Applied Ergonomics 2002; 33:207e17. Cook C, Burgess-Limerick R, Chang S. The prevalence of neck and upper extremity musculoskeletal symptoms in computer mouse users. International Journal of Industrial Ergonomics 2000;26:347e56. Dankaerts W, O’Sullivan PB, Burnett AF, Straker LM. The use of a mechanism-based classification system to evaluate and direct management of a patient with non-specific chronic low back pain and motor control impairment e a case report. Manual Therapy 2007;12(2):181e91. Falla D. Unravelling the complexity of muscle impairment in chronic neck pain. Manual Therapy 2004;9:125e33. Falla D, Bilenkij G, Jull G. Patients with chronic neck pain demonstrate altered patterns of muscle activation during performance of a functional upper limb task. Spine 2004;29(13):1436e40. Feuerstein M, Shaw WS, Nicholas RA, Huang GD. From confounders to suspected risk factors: psychosocial factors and work-related upper extremity disorders. Journal of Electromyography and Kinesiology 2004;14:171e8. Gerr F, Marcus M, Monteilh C. Epidemiology of musculoskeletal disorders among computer users: lesson learned from the role of posture and keyboard use. Journal of Electromyography and Kinesiology 2004;14:25e31. Graven-Nielsen T, Mense S. The peripheral apparatus of muscle pain: evidence from animal and human studies. Clinical Journal of Pain 2001;17:2e10. 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 and Environmental Health 1997;69:423e32. van der Heide B, Allison GT, Zusman M. Pain and muscular responses to a neural tissue provocation test in the upper limb. Manual Therapy 2001;6(3):154e62.

345

Holte KA, Westgaard RH. Daytime trapezius muscle activity and shouldereneck pain of service workers with work stress and low biomechanical exposure. American Journal of Industrial Medicine 2002;41:393e405. 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;9:89e94. Kuorinka I, Jonsson B, Kilbom A, Vinterberg H, Biering-Sorenson F, Andersson A, et al. Standardised Nordic Questionnaires for the analysis of musculoskeletal symptoms. Applied Ergonomics 1987;18(3):233e7. Madeleine P, Lundager B, Voigt M, Arendt-Nielsen L. Shoulder muscle co-ordination during chronic and acute experimental necke shoulder pain. European Journal of Applied Physiology 1999;79:127e40. Mork PJ, Westgaard RH. The association between nocturnal trapezius muscle activity and shoulder and neck pain. European Journal of Applied Physiology 2004;92:18e25. Nederhand MJ, Hermens HJ, Ijzerman MI, Turk DC, Zilvold G. Chronic neck pain disability due to an acute whiplash injury. Pain 2003;102:63e71. Nederhand MJ, Ijzerman MJ, Hermens HJ, Baten CTM, Zilvold G. Cervical muscle dysfunction in the chronic whiplash associated disorder grade II (WAD-II). Spine 2000;25(15):1938e43. 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):1904e19. Sterling M, Jull G, Wright A. The effect of musculoskeletal pain on motor activity and control. Journal of Pain 2001;2(3):135e45. Szeto GPY, Straker LM, O’Sullivan PB. A comparison of symptomatic and asymptomatic office workers performing monotonous keyboard work e 1. Neck and shoulder muscle recruitment patterns. Manual Therapy 2005a;10:270e80. Szeto GPY, Straker LM, O’Sullivan PB. A comparison of symptomatic and asymptomatic office workers performing monotonous keyboard work e 2. Neck and shoulder kinematics. Manual Therapy 2005b;10:281e91. Szeto GPY, Straker LM, O’Sullivan PB. The effects of speed and force of keyboard operation on neckeshoulder muscle activities in symptomatic and asymptomatic office workers. International Journal of Industrial Ergonomics 2005c;35:429e44. Szeto GPY, Straker LM, O’Sullivan PB. The effects of typing speed and force on motor control in symptomatic and asymptomatic office workers. International Journal of Industrial Ergonomics 2005d;35:779e95. Szeto GPY, Straker LM, O’Sullivan PB. EMG median frequency changes in the neckeshoulder stabilisers of symptomatic office workers when challenged by different physical stressors. Journal of Electromyography and Kinesiology 2005e;15:544e55. Tittiranonda P, Burastero S, Rempel D. Risk factors for musculoskeletal disorders among computer users. Occupational Medicine: State of the Art Reviews 1999;14:17e38. Veiersted KB, Westgaard RH, Andersen P. Pattern of muscle activity during stereotyped work and its relation to muscle pain. International Archives of Occupational and Environmental Health 1990;62:31e41. van Vliet PM, Heneghan NR. Motor control and the management of musculoskeletal dysfunction. Manual Therapy 2006;11:208e13. Waersted M. Human muscle activity related to non-biomechanical factors in the workplace. European Journal of Applied Physiology 2000;83:151e8.

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 346e350 www.elsevier.com/math

Technical and Measurement Report

Simple anatomical information improves the accuracy of locating specific spinous processes during manual examination of the low back Dean R. Phillips a,*, Sue Barnard b, Mark A. Mullee c, Michael V. Hurley d a

School of Health Professions and Rehabilitation Sciences, University of Southampton, Highfield, Southampton SO17 1BJ, England, United Kingdom b School of Health Sciences, The Robert Gordon University, Aberdeen, United Kingdom c Research and Development Support Unit, CCS Division, University of Southampton, United Kingdom d Rehabilitation Research Unit, Physiotherapy Division, King’s College London, United Kingdom Received 19 December 2006; received in revised form 2 January 2008; accepted 27 February 2008

Abstract The objective of the study was to test whether a teaching protocol including simple anatomical information on the surface anatomy of spinous processes, improves physiotherapy students’ ability to accurately locate selected thoracic and lumbar spinal segments e T12 and L3. First year physiotherapy students were allocated to Group 1 (n ¼ 35) and Group 2 (n ¼ 34). Both groups were taught to identify spinous processes by counting up from the sacrum, but Group 2 received supplementary anatomical information on the shapes and vertical length of the tips of L5 to T12 spinous processes. The spinous processes of L3 and T12 were located by two experienced physiotherapists and marked on a model using an invisible skin marker. Volunteer students were asked to locate these spinous processes and accuracy was confirmed using an ultraviolet lamp. Students with supplementary anatomical information (Group 2) were significantly better at locating T12 (difference in proportions 36% (95% confidence interval 14 to 51%)) and both T12 and L3 (difference in proportions 33% (11 to 48%)). Group 2 students were also better than Group 1 students at locating L3 (difference in proportions 28% (4 to 48%)), but the difference was not significant. Including simple anatomical information when teaching manual examination skills improved the accuracy of locating specific low back spinal levels. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Manual examination; Palpation; Accuracy; Low back

1. Introduction Accurate manual examination and location of segmental level of impairment are important for diagnosis and treatment of spinal pathology (Downey et al., 1999). Accurate location of spinal segmental levels is * Corresponding author. Tel.: þ44 023 8059 5305; fax: þ44 023 8059 5301. E-mail address: [email protected] (D.R. Phillips). 1356-689X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.02.009

also important in studies investigating reliability of manual examination skills. Error in identification of bony landmarks may explain the poor reliability of many manual examination techniques (Simmonds and Kumar, 1993a). Physiotherapy students’ ability to accurately locate spinal segments is a skill that will help their manual therapy. However, there is little research on the best methods of teaching students these skills. Studies have shown experienced physiotherapists possess manual examination skills enabling them to

D.R. Phillips et al. / Manual Therapy 14 (2009) 346e350

diagnose the segmental origin of a patient’s spinal pain (Jull et al., 1988; Phillips and Twomey, 1996), aided by their skill at accurately locating specific spinal segments. Whether similar findings would occur in therapists with less clinical experience is unclear. Research investigating manual examination of the spine has commonly shown intra-therapist reliability is better than inter-therapist reliability (Gonnella et al., 1982; Matyas and Bach, 1985; Panzer, 1992; Seffinger et al., 2004). Studies investigating accuracy in locating lumbar spinal segments show physiotherapists are moderately successful at this task (Harlick et al., 2007). Results of landmark location tests (Byfield et al., 1992; McKenzie and Taylor, 1997; Downey et al., 1999), although variable, show acceptable reliability (Seffinger et al., 2004), but others suggest reliability in the lumbar/sacral spine is questionable (Burton et al., 1990; Simmonds and Kumar, 1993b). Studies have not attempted to identify teaching protocols aimed at improving the ability of therapists in locating specific spinal segments. Teaching that facilitates accurate location of spinal segments will aid recording and communication of the segmental origin of a patient’s complaint. More accurate localisation of spinal segments should also improve results of studies investigating intra- and inter-examiner reliability of manual examination. The level of iliac crests is commonly used by manual therapists to locate spinal segments (Harlick et al., 2007). However, a radiological study found they lie in the same plane as the L4/5 disc inter-space in 31.9%; L4 body in 49.5% and L5 body in 18.6% (Walsh et al., 2006). Variability in level of iliac crests can also occur due to trunk side flexion in a patient during examination. Such inherent variability suggests this should be avoided as a technique for locating specific lumbar segments and its use may in part account for differences in accuracy reported between therapists in some studies (Downey et al., 2003; Harlick et al., 2007). One strategy that might improve accuracy of locating spinal segmental levels is to provide simple anatomical information on the surface anatomy of spinous processes. Although scientific confirmation of surface marking size and shape of spinous processes is lacking there is evidence suggesting anatomical consistency in the population. The L5 spinous process vertical height is smallest of the lumbar segments (Downey et al., 2003; Harlick et al., 2007) and has a rounded tip (Standring, 2005). Commonly the vertical height of L3 spinous process is larger than other lumbar segments (clinical observation by lead author), although Downey et al. (2003) reports both L3 and L2 to be similarly long. Photographic evidence in anatomical text confirms the spinous process vertical height at both L3 and L2 are larger than at other lumbar segments and L3 appears to be marginally larger than L2 (Rohen et al., 2006). In addition vertical height

347

of T12 spinous process is noticeably smaller than L1 and with a rounded tip similar to L5 (Rohen et al., 2006). Anatomical text also suggests the lumbar spinous processes L1eL4 are quadrangular in shape (Gunn, 1992). The difference in vertical height of the tip of the spinous processes and their shape offers potential to identify specific spinal segments when palpating, particularly the contrast in vertical height between L4 and L3, L5 and L4, and between T12 and L1; and the contrast in shape between L5 and L4, and T12 and L1. In this study we evaluated whether a teaching protocol that added simple anatomical information on the surface anatomy of the spinous processes, improves an undergraduate physiotherapy students’ ability to accurately locate selected thoracic and lumbar spinal segments (T12 and L3). 2. Methodology 2.1. Participants Participants were recruited from 1st year undergraduate physiotherapy students at the School of Health Professions and Rehabilitation Sciences, University of Southampton (n ¼ 69). At the beginning of their training participants were allocated to two teaching groups within the neuromusculoskeletal module (students with surname beginning AeL were in Group 1 and MeZ in Group 2). Teaching sessions took place in separate practical rooms and students were asked not to discuss their specific training with those in the other groups. Ethical approval was granted by the School’s Ethics Committee. 2.2. Manual examination protocols In session 5 of the neuromusculoskeletal module both groups were taught to identify the surface marking of the spinous processes of L5 to T12 on a model lying prone on a couch, as below: Group 1 manual examination protocol (basic): (1) L5 e slide cranially off the sacrum until a spinous process is located deep at the centre of the lumbosacral depression. (2) L5 e reconfirm by finding the posterior superior iliac spine (PSIS) and draw an imaginary line at a 30 angle (from the horizontal) and the same spinous process should be located. (3) L4 to T12 spinous processes e count up the spine in a cranial direction. Group 2 manual examination protocol (basic plus supplementary anatomical information): (1), (2) and (3) as above, plus; (4) Final confirmation is related to the anatomical knowledge that the tip of the spinous process of

348

D.R. Phillips et al. / Manual Therapy 14 (2009) 346e350

L5 is small and rounded; L4eL1 are quadrangular, with L3 being the longest; T12 is rounded (Gunn, 1992; Standring, 2005; Rohen et al., 2006). Both groups practised their specific protocol for 10 min once per week, for 3 successive weeks. Working in pairs, each had 5 min (15 min total) practice. After the third practice session (and on the same day), volunteer students were randomly tested as to the accuracy of their palpation. On testing, one student (with no spinal pain and previous X-ray confirming five lumbar spinal segments) acted as a model. An experienced manual therapist placed hypoallergenic tape on the skin over the spine (lower sacrum to mid thoracic) of the model. The outline of the spinous processes of T12 and L3 was marked on the tape using marker visible only in ultraviolet (UV) light. A second experienced manual therapist acted as an independent observer to verify accuracy. In addition to confirm their accuracy a vernier with 0.1 mm increments was used by the first experienced manual therapist to measure the surface marking vertical length of the tip of the spinous processes on the model (vernier is a small movable device able to obtain fractional parts of a fixed main scale i.e. mm in this study). These measures and the shape of the tip of the spinous processes were verified by the independent observer and compared to known anatomical knowledge. All testing occurred over one morning with each student tested one after the other. Each volunteer student was asked to locate the spinous processes of T12 and then L3 on the model. During the procedure they were required to wear latex gloves to avoid skin contact with the UV sensitive ink. When each spinous process was found the student used a narrow gauge pointer to indicate its location. An assistant blind to the randomisation process and specific manual examination protocol used by each student confirmed their accuracy using an UV lamp (Berol Detective Lamp, Berol Corporation) i.e. if the pointer was within the outline of the spinous process it was considered to be accurate. During this confirmation the student turned away so they were unaware of their accuracy locating T12 as this could have aided them in finding L3. Skin redness was observed for on the model, but none was visible through the hypoallergenic tape. 2.3. Statistical analysis Yate’s corrected chi-squared test (c2) with a twosided significance level was used to assess for differences in accuracy between Groups 1 and 2 students in locating T12 and L3 spinous processes. To account for multiple testing Bonferroni correction factor was used and P value of less than 0.017 (0.05/3) was thus considered statistically significant.

2.4. Sample size A sample size of 35 in Groups 1 and 34 in Group 2 had 80% power to detect a 35.7% difference in proportions of physiotherapy student’s ability to accurately locate the T12 spinal segment, with a 0.017% two-sided significance level. 3. Results 3.1. Student characteristics All 1st year physiotherapy students agreed to participate and were in their first 3 months of undergraduate physiotherapy training. In Group 1 there were 11 males and 24 females (mean age 20.4 years; SD ¼ 3.8). In Group 2 there were eight males and 26 females (mean age 23 years; SD ¼ 6.2). 3.2. Characteristics of the model/spinous process measurements The model was female aged 30 years; weight 73.5 kg and height 180 cm. Surface marking vertical length of the model’s tip of the spinous processes (measurements taken from the mid point of the inter-spinous space) was T12 ¼ 26 mm (also rounded in shape); L1 ¼ 30.5 mm; L2 ¼ 31.3 mm; L3 ¼ 40 mm; L4 ¼ 28 mm; and L5 ¼ 18.4 mm (also rounded in shape). 3.3. Number of students correctly palpating each level The percentage of students who correctly located T12, L3, and T12 and L3 was 14.3%, 40% and 14.3% for Group 1 and 50%, 67.6% and 47.1% for Group 2, respectively (see Table 1). Group 2 students (basic plus supplementary anatomical information) were significantly better than Group 1 students (basic training only) at locating T12, and T12 and L3. Group 2 students were also better than Group 1 students at locating L3, but the difference was not statistically significant under the significance level set by Bonferroni correction (See Table 1). 4. Discussion The study found that enhancing student physiotherapists’ manual examination training with supplementary anatomical information on the shapes and length of the tips of the spinous processes increased their accuracy of locating T12 and L3 spinal segments. All students wore latex gloves during testing potentially decreasing their accuracy, as product safety information stated skin contact with the UV sensitive ink was to be avoided. Similarly the tape on the model’s skin

349

D.R. Phillips et al. / Manual Therapy 14 (2009) 346e350 Table 1 Number of students correctly palpating each spinous process.

T12 L3 T12 & L3

Group 1 (basic) n ¼ 35

Group 2 (basic plus) n ¼ 34

Difference in proportions (95% Cl)

Yates c2

P value

5 (14.3%) 14 (40%) 5 (14.3%)

17 (50%) 23 (67.6%) 16 (47.1%)

35.7% (14.1 to 50.8%) 27.6% (4.1 to 48.1%) 32.8% (11.3 to 47.9%)

8.55 4.25 7.27

0.003 0.039 0.007

T12 ¼ spinous process of the 12th thoracic vertebra. L3 ¼ spinous process of the 3rd lumbar vertebra. CI ¼ confidence interval.

may have decreased accuracy. If replicating the study, the tape could be applied lateral to the spinous processes and skin contact with the UV sensitive ink could be avoided. However, the position of the tape successfully concealed any redness that may have occurred on the model, preventing any possible visual cues to students. There are three potential limitations to this study. Firstly, students in Group 2 could have told student’s in Group 1 the supplementary anatomical information. If this had occurred results for Group 1 students would have been expected to be better than were observed. It is therefore unlikely significant ‘‘contamination’’ occurred. Secondly, the teaching protocol with supplementary anatomical information was based on anatomical knowledge of the shape and vertical height of the tip of the spinous processes, which needs to be scientifically validated. However, there appears some consistencies across the population in the difference in shape and vertical height length of the tip of L5 to T12 spinous processes in an individual (Gunn, 1992; Downey et al., 2003; Standring, 2005; Rohen et al., 2006; Harlick et al., 2007), with the contrast in their shape and size offering the potential to more accurately locate specific segmental levels with palpation. These findings were used to substantiate the supplementary anatomical information given to students in Group 2 and were found to be consistent on the model used in this study. Thirdly, no gold standard was used to ensure the accuracy of the two physiotherapists who initially identified T12 and L3. Therefore uncertainty exists as to the validity of their assessment. However, they had a minimum of 8 years postgraduate qualifications in manipulative therapy and independently assessed the location. In addition, the same model was used throughout and had radiological confirmation of five spinous processes. Further verification of their accuracy was achieved through a comparison of the surface marking vertical length measurements of the model’s spinous processes and their shape, with known anatomy. It is acknowledged that in a patient population accuracy of locating spinal segments would be affected by sacralisation or lumbarisation (four lumbar vertebrae and six lumbar vertebrae, respectively). Although the age of students and proportion of males/females was different between groups it is not believed these differences have confounded the results. All

students involved in this study had only just started physiotherapy training, so the accuracy of locating spinal segments compared very favourably to other studies (Byfield et al., 1992; McKenzie and Taylor, 1997; Downey et al., 1999; Harlick et al., 2007). With further training and clinical experience this accuracy may improve further. However, the validity of this assumption has been questioned by Seffinger et al. (2004), but the studies they reviewed had not attempted to teach simple anatomical information in order to improve spinal palpation skills. Therefore future studies that include teaching of simple anatomical information need to investigate whether clinical experience improves accuracy of locating spinal segments in the low back.

5. Conclusion Giving simple anatomical information on the shapes and length of the tips of the spinous processes improved students’ ability to accurately identify specific spinal levels. Simple anatomical information should be included when teaching manual examination skills to improve the accuracy of identifying spinal levels in the low back.

Acknowledgment This study was published as a conference abstract: Phillips DR, Barnard S, Hurley M. Simple anatomical information improves the accuracy of locating specific spinous processes during examination of the vertebral column. BSR XIXth AGM and BHPR Spring Meeting, April 2002, Rheumatology, 2002, 41 (Abstract Suppl. 1), p. 82e3; Phillips DR, Barnard S. Test of Validity and Reliability of a Protocol Used to Identify Specific Spinous Processes. Physiotherapy, 2000, 86 (11), p. 591; Phillips DR, Barnard S. Investigation to test the validity and reliability of a protocol used to identify specific spinous processes. In: Singer KP, editor. Proceedings of the 7th Scientific Conference of the IFOMT in conjunction with the 11th Biennial Conference of the MPAA, Perth, Western Australia, 2000. p. 369e73.

350

D.R. Phillips et al. / Manual Therapy 14 (2009) 346e350

References Burton AK, Edwards VA, Sykes DA. Invisible skin marking for testing palpatory reliability. Journal of Manual Medicine 1990;5:27e9. Byfield DC, Mathiasen J, Sangren C. Intra- and inter-examiner reliability of static palpation of specific landmarks in the lumbar spine and pelvis using an invisible skin marking pen. 1991 World Chiropractic Congress Abstracts. Journal of Manipulative and Physiological Therapeutics 1992;15(9):612e3. Downey BJ, Taylor NF, Niere KR. Manipulative physiotherapists can reliably palpate nominated lumbar spinal levels. Manual Therapy 1999;4(3):151e6. Downey B, Taylor N, Niere K. Can manipulative physiotherapists agree on which lumbar level to treat based on palpation? Physiotherapy 2003;89(2):74e81. Gonnella C, Paris SV, Kutner M. Reliability in evaluating passive intervertebral motion. Physical Therapy 1982;62(4): 436e44. Gunn C. Bones and joints. A guide for students. 2nd ed. Edinburgh: Churchill Livingstone; 1992. p. 125 [chapter 9]. Harlick JC, Milosavljevic S, Milburn PD. Palpation identification of spinous processes in the lumbar spine. Manual Therapy 2007;12:56e62. Jull G, Bogduk N, Marsland A. The accuracy of manual diagnosis for cervical zygapophysial joint pain syndromes. Medical Journal of Australia 1988;148:233e6. McKenzie AM, Taylor NF. Can physiotherapists locate lumbar spinal levels by palpation? Physiotherapy 1997;83(5):235e9.

Matyas TA, Bach TM. The reliability of selected techniques in clinical arthrometrics. Australian Journal of Physiotherapy 1985;31(5): 175e99. Panzer DM. The reliability of lumbar motion palpation. Journal of Manipulative and Physiological Therapeutics 1992;15(8):518e24. Phillips DR, Twomey LT. A comparison of manual diagnosis with a diagnosis established by a uni-level lumbar spinal block procedure. Manual Therapy 1996;2:82e7. Rohen JW, Yokochi C, Lu¨tjen-Drecoll E. Color atlas of anatomy. A photographic study of the human body. 6th ed. Stuttgart/London: Schattauer/Lippincott Williams and Wilkins; 2006. p. 193 [chapter 3]. Seffinger MA, Najm WI, Mishra SI, Adams A, Dickerson VM, Murphy LS, et al. Reliability of spinal palpation for diagnosis of back and neck pain. A systematic review of the literature. Spine 2004;29(19):E413e25. Simmonds MJ, Kumar S. Part 1: the fundamental skill of palpation e a review and critique. International Journal of Industrial Ergonomics 1993a;11:135e43. Simmonds MJ, Kumar S. Part II: location of body structures by palpation e a reliability study. International Journal of Industrial Ergonomics 1993b;11:145e51. Standring S, editor. Gray’s anatomy: the anatomical basis of clinical practice. 39th ed. Edinburgh: Churchill Livingstone; 2005. p. 748e9 [chapter 45]. Walsh JC, Quinlan JF, Butt K, Towers M, Devitt AT. Variation in position of the L4/5 disc inter-space from the anatomical landmark: review of 450 radiographs and clinical applications. European Journal of Orthopaedic Surgery and Traumatology 2006;16(3):203e6.

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 351 www.elsevier.com/math

Diary of events NZMPA biennial scientific conference, Heritage Hotel, Rotorua, New Zealand 28, 29 & 30 August 2009. The theme is ‘Striving for Excellence in OMT’ & also celebrating 40 years of Manual Therapy in New Zealand. The conference co-coordinator is Vicki Reid, Phone 0800 646 000 or 09 476 5353 Fax 09 476 5354 e-mail: [email protected] Website: www.nzmpa.org.nz APA Conference Week, Sydney Convention Centre, Sydney, Australia 1-5 October 2009 For more information please visit www.apaconferenceweek09. asn.au NOI International conference UK and Ireland Nottingham UK e April 15e17, 2010 Dublin IRELAND April 21e23, 2010

1356-689X/$ - see front matter doi:10.1016/S1356-689X(09)00021-6

For further details www.noi2010.com Fax þ 3906 51882443 Janet G. Travell, MD Seminar Series, Bethesda, USA For information, contact: Myopain Seminars, 7830 Old Georgetown Road, Suite C-15, Bethesda, MD 20814-2432, USA. Tel.: þ1 301 656 0220; Fax: þ1 301 654 0333; website: www.painpoints.com/seminars.htm E-mail: [email protected] The Diary of events will soon be moving to http://www.elsevier. com/math (click on the journal news tab). If you wish to advertise a course/conference, please contact: Karen Beeton, Associate Head of School (Professional Development), School of Health and Emergency Professions, University of Hertfordshire, College Lane, Hatfield, Herts AL10 9AB, UK. E-mail: [email protected] There is no charge for this service.