Manual Therapy Journal - Volume 14, Issue 5, Pages 461-584 (October 2009)

VOLUME 14 NUMBER 5 PAGES 461–584 October 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. A. Schmitt (Amersfoort, The Netherlands) M. Shacklock (Adelaide, Australia) D. Shirley (Sydney, Australia) C. Snijders (Rotterdam, The Netherlands) P. Spencer (Barnstaple, UK) M. Sterling (Brisbane, Australia) M. Stokes (Southampton, UK) P. Tehan (Melbourne, Australia) M. Testa (Alassio, Italy) P. van der Wurff (Doorn, The Netherlands) P. van Roy (Brussels, Belgium) O. Vasseljen (Trondheim, Norway) B.Vicenzino (Brisbane, Australia) M. Wessely (Paris, France) A. Wright (Perth, Australia) M. Zusman (Perth, Australia)

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)00108-8

Manual Therapy 14 (2009) 461–462

Contents lists available at ScienceDirect

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

Editorial

Clinical expertise: Learning together through observed practice

The popularity of inservice training programmes and short courses, as well as clinically related MSc courses in higher education, suggest that manual therapy practitioners are only too aware of a need to enhance clinical effectiveness and efficiency. While national professional bodies may consider continuous professional development (CPD) activity as obligatory, practitioners themselves have long been driven by a strong moral responsibility to improve what they do for their patients, investing both their time and finances in their learning. But what sort of learning do they do and is it effective to develop clinical expertise? Typically, CPD activities involve in-service training in the workplace and short courses away from the workplace that focus on relevant literature and research and hands-on skill. Practitioners perform techniques on each other with guidance from someone with expertise. Practitioners then go back into clinical practice and apply their new knowledge and skill to patients. Over time, practitioners see numbers of patients and gain experience and ‘patient mileage’ (Richardson, 1996, 1999). Will this CPD diet of patient experience, in-service training and short courses, result in enhanced clinical practice; will it lead to clinical expertise? To address this question, the literature related to professional learning is briefly reviewed. The notion that patient mileage automatically leads to clinical practice expertise is not supported by the literature (Boud et al., 1993; Stathopoulos and Harrison, 2003; Conneeley, 2005). There may be a number of reasons for this:  practitioners tend to experience what they expect to experience, there is a circular nature to their experience and understanding (Heidegger, 1926/1962; Dewey, 1938/1997). As such, practitioners may be trapped within their existing understanding (Dall’Alba and Sandberg, 2006) making it difficult to learn and enhance their practice. The isolated clinical practice within screened cubicles may further exacerbate this situation.  practitioners may develop automatic, habitual use of examination procedures, treatment techniques and management strategies that they routinely apply to patients (Eraut, 1994, 2005). This may be made worse by departments that value efficiency and patient through-put and provide little time for deliberation (Eraut, 1994).  clinical practice involves managing complex and uncertain problems. This may not enable the practitioner to obtain accurate and specific feedback on their clinical decisions that would help them to learn (Eraut, 1994). While in-service training and short courses may enhance motor skill development, it may have limited potential to enhance clinical 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.06.001

practice expertise. The most obvious reason, is that hands-on skill is only one of many competencies needed for clinical practice. The skills, for example, to develop an effective therapeutic relationship with patients or to take an accurate and comprehensive subjective examination are missing from this diet. Furthermore, formal assessment helps to trigger and motivate learners to learn; its absence may therefore limit learning (Criticos, 1993; Ramsden, 2003). Finally, learning occurs away from patients and is not contextualised within clinical practice. The value of learning experiences to be situated where it is to be applied, is well rehearsed in the literature; to enhance clinical practice, learning needs to occur in clinical practice (Fish and Coles, 1998; Billett, 2004; Dall’Alba and Sandberg, 2006). If a diet of patient mileage, in-service training and short courses is insufficient to learn and enhance clinical practice, what is needed? We would suggest that direct observation of each other with patients, in clinical practice, may be a valuable addition to enhancing clinical expertise. Its value hinges on the fact that learning is occurring in clinical practice and there is immediate and specific feedback of performance to the practitioner. We can consider the benefit of this at different stages of practitioner’s clinical development. It is acknowledged that observation and feedback are an important part of the growth and development of newly qualified practitioners (McInstry, 2005; Toal-Sullivan, 2006; Morley, 2007). For example, Morley (2007) found that observed practice by a more senior practitioner, as part of a formal perceptorship programme, increased confidence and clinical competence. In turn, the observer valued the process as a way of accessing the new practitioners’ thinking, particularly in situations where there was little or no co-working. While observed practice is rarely considered as a learning strategy for more experienced practitioners, there is recent evidence of its worth in UK physiotherapists undertaking a musculoskeletal practice-based MSc (Petty, 2009). Observed practice by a more experienced practitioner was identified as the most powerful learning process to foster their development towards clinical expertise. Why is observation of practice considered so helpful? To address this, the benefits of being observed by a colleague as well as observing a colleague with greater expertise needs to be explored? 1. Being observed by a colleague An observer watching a colleague manage a new or follow up patient appointment is able to stand back from the situation and see the interaction as a whole. While it is an encounter with just one patient, it provides a specific example of the practitioner’s practice and may offer a rich learning experience. The nature of that experience will, in part, be determined by the relative

462

Editorial / Manual Therapy 14 (2009) 461–462

experience of the observed and observer. The value of a practitioner with higher levels of expertise enhancing the practice of someone less experienced through observation is well supported by the literature (Fish and Twinn, 1997; Daloz, 1999; Titchen, 2001). Within physiotherapy, clinical ‘experts’ have consistently recalled the powerful impact of learning in practice with patients in the presence of an expert guide and teacher (Jensen et al., 1999) and remains a requirement for all manual therapy courses that have membership of the International Federation of Manipulative Therapists (IFOMT). However, being observed by a peer with similar levels of knowledge may also be valuable as they share their practice and learn from each other. The practice knowledge, (defined here as all types of knowledge and skill) of the practitioner is, in part, revealed to the observer through the actions and decisions they make with the patient, as well as the debrief afterwards. ‘What did they think about the patient’s recent weight loss?’ ‘Why did they choose the straight leg raise and not the slump?’ ‘How will they progress treatment?’ Like an iceberg, a great deal of practice knowledge tends to remain hidden (Argyris and Schon, 1974; Fish and Coles, 1998). The observer, through questioning, can raise the practitioner’s awareness of this hidden, and perhaps taken for granted knowledge, discuss and share, and thereby enhance it. Another advantage of direct observation is that much of practice knowledge is tacit or difficult to articulate (Eraut, 1994; Fish, 1998; Titchen and Ersser, 2001). Skill in analysing posture or palpation for example, are very difficult to describe in words. These aspects of practice can be readily shared as each practitioner observes or palpates the patient and discusses their findings. Bringing to light all forms of practice knowledge provides the potential for affirmation and enhanced confidence, as well as change and improvement. 2. Observing a colleague The opportunity to observe a colleague with a patient may also provide a valuable learning opportunity. Observational learning is highlighted in the literature and considered a potentially powerful process (Bandura, 1997; Titchen, 2001). The observer may gain confidence seeing similar actions to their own, as well as seeing alternative ways to do things that they then may adopt into their own practice. The degree to which this happens will, in part, depend on the relative experience of the observed and observer. Someone with clinical expertise may become an inspirational role model for a novice practitioner. In this situation, observation may raise awareness of a much higher level of practice and professional behaviour, triggering their need to learn, and inspiring their subsequent professional development. Where the observer is more experienced, the process may offer alternative ways to practice as well as affirm and consolidate their practice knowledge. Where the observed and observer are peers with similar experience and knowledge, observation may provide each other with significant help and support as they grapple with similar issues. Introducing observation of clinical practice in the workplace may be strongly resisted by practitioners. They may feel too vulnerable and fear harsh and negative judgment of their clinical practice. They may feel anxious that if this happens they will lose respect and have promotion blocked. However, while independent practice may protect them from criticism, it prevents them receiving encouragement, support and guidance. It may also limit their potential to develop high levels of clinical expertise. For observation of practice to be successfully implemented in the workplace, it is imperative that everyone involved acknowledges and sensitively manages the power relationship between the observed and the observer. Fundamental to this, is a collaborative, respectful, supportive environment that genuinely desires to

facilitate learning and expertise in each other. Everyone from consultant to newly qualified practitioner needs to continually learn and develop their practice. All aspects of knowledge applied in clinical practice needs to be addressed, not just technical skill. We would argue that direct observation of practice offers a powerful, yet readily available tool in the workplace, to enhance clinical practice and maximise patient outcomes for both individual practitioners and clinical teams. References Argyris C, Schon DA. Theory in practice, increasing professional effectiveness. San Francisco: Jossey-Bass; 1974. Bandura A. Self perceived self efficacy, the exercise of control. New York: W H Freeman and Co; 1997. Billett S. Workplace participatory practices, conceptualising workplaces as learning environments. The Journal of Workplace Learning 2004;16(6):312–24. Boud D, Cohen R, Walker D. Using experience for learning. Buckingham: The Society for Research into Higher Education and Open University; 1993. Conneeley AL. Study at master’s level: a qualitative study exploring the experience of students. British Journal of Occupational Therapy 2005;68(3):104–9. Criticos C. Experiential learning and social transformation for a post-apartheid learning future. In: Boud D, Cohen R, Walker D, editors. Using experience for learning. Buckingham: The Society for Research into Higher Education and Open University; 1993. p. 157–68 [chapter 11]. Dall’Alba G, Sandberg J. Unveiling professional development: a critical review of stage models. Review of Educational Research 2006;76(3):383–412. Daloz LA. Mentor, guiding the journey of adult learners. San Francisco: Jossey-Bass; 1999. Dewey J. Experience and education. New York: Touchstone; 1938/1997. Eraut M. Developing professional knowledge and competence. London: Routledge Falmer; 1994. Eraut M. Editorial, continuity of learning. Learning in Health and Social Care 2005; 4(1):1–6. Fish D. Appreciating practice in the caring professions. Oxford: Butterworth-Heinemann; 1998. Fish D, Coles C. Developing professional judgement in health care. Oxford: Butterworth-Heinemann; 1998. Fish D, Twinn S. Quality clinical supervision in the health care professions, principled approaches to practice. Edinburgh: Butterworth-Heinemann; 1997. Heidegger M. Being and time. Oxford: Blackwell; 1926/1962. Jensen GM, Gwyer J, Hack LM, Shepard KF. Expertise in physical therapy practice. Boston: Butterworth Heinemann; 1999. McInstry C. From graduate to practitioner: rethinking organisational support and professional development. In: Whiteford G, Wright-St Clair V, editors. Occupation and practice in context. Oxford: Elsevier; 2005. p. 129–42 [chapter 8]. Morley MT. A realist evaluation of a preceptorship programme for newly qualified occupational therapists. University of Brighton: Unpublished DOccT; 2007. Petty NJ. Towards clinical expertise: learning transitions of neuromusculoskeletal physiotherapists. University of Brighton: Unpublished DPT; 2009. Ramsden P. Learning to teach in higher education. 2nd ed. London: Routledge Falmer; 2003. Richardson B. Paradigms of practice in physiotherapy and the implications for professional development. East Anglia University: Unpublished PhD; 1996. Richardson B. Professional development, professional knowledge and situated learning in the workplace. Physiotherapy 1999;85(9):467–74. Stathopoulos I, Harrison K. Study at master’s level by practising physiotherapists. Physiotherapy 2003;89(3):158–69. Titchen A. Critical companionship: a conceptual framework for developing expertise. In: Higgs J, Titchen A, editors. Practice knowledge and expertise in the health professions. Oxford: Butterworth Heinemann; 2001. p. 80–90 [chapter 10]. Titchen A, Ersser SJ. The nature of professional craft knowledge. In: Higgs J, Titchen A, editors. Practice knowledge and expertise in the health professions. Oxford: Butterworth Heinemann; 2001. p. 35–41 [chapter 5]. Toal-Sullivan D. New graduates’ experiences of learning to practice occupational therapy. British Journal of Occupational Therapy 2006;69(11):513–52.

Nicola J. Petty* School of Health Professions, University of Brighton, Robert Dodd Building, 49 Darley Road, Eastbourne BN20 7UR, UK  Corresponding author. Tel.: þ44(0)1273 643775; fax: þ44(0)1273 643652. E-mail address: [email protected] (N.J. Petty) Mary Morley South West London and St George’s Mental Health NHS Trust, Springfield University Hospital, 1st Floor, Admissions Block, 61 Glenburnie Road, London SW17 7DJ, UK

Manual Therapy 14 (2009) 463–474

Contents lists available at ScienceDirect

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

Systematic Review

The effectiveness of manual therapy in the management of musculoskeletal disorders of the shoulder: A systematic review Chung-Yee Cecilia Ho a, Gisela Sole a, *, Joanne Munn a, b a b

School of Physiotherapy, University of Otago, 325 Great King Street, Dunedin North, P.O. Box 56, Dunedin 9016, New Zealand School of Physiotherapy, The University of Sydney, P.O. Box 170, Lidcombe NSW 1825, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 July 2008 Received in revised form 23 March 2009 Accepted 27 March 2009

A systematic review of randomised controlled trials (RCTs) was conducted to determine the effectiveness of manual therapy (MT) techniques for the management of musculoskeletal disorders of the shoulder. Seven electronic databases were searched up to January 2007, and reference lists of retrieved articles and relevant MT journals were screened. Fourteen RCTs met the inclusion criteria and their methodological qualities were assessed using the PEDro scale. Results were analyzed within diagnostic subgroups (adhesive capsulitis (AC), shoulder impingement syndrome [SIS], non-specific shoulder pain/dysfunction) and a qualitative analysis using levels of evidence to define treatment effectiveness was applied. For SIS, there was no clear evidence to suggest additional benefits of MT to other interventions. MT was not shown to be more effective than other conservative interventions for AC, however, massage and Mobilizations-with-Movement may be useful in comparison to no treatment for short-term outcomes for shoulder dysfunction. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Shoulder Manual therapy Massage Systematic review

1. Introduction Various physiotherapy approaches have been suggested for shoulder musculoskeletal disorders, including manual therapy (MT), electrotherapy, acupuncture and exercise therapy (Brox, 2003). MT, including massage, joint mobilization and manipulation (such as Maitland, 1991), may be used with the aim of decreasing pain and improving range of motion (ROM), thereby improving function. To date, a number of systematic reviews have evaluated the effectiveness of conservative treatment in shoulder disorders (Van der Heijden et al., 1997; Green et al., 1998; Desmeules et al., 2003; Green et al., 2003; Ejnisman et al., 2004; Grant et al., 2004; Gibson et al., 2004; Harniman et al., 2004; Michener et al., 2004; Faber et al., 2006; Trampas and Kitsios, 2006). Although there was some evidence of an additional benefit of MT with exercise in patients with shoulder impingement syndrome (SIS), conclusions from these reviews (Desmeules et al., 2003; Green et al., 2003; Michener et al., 2004; Faber et al., 2006; Trampas and Kitsios, 2006) were limited due to small number of studies including MT. To our knowledge, there is no systematic review specifically for the effectiveness of MT in addition or in comparison to other

* Corresponding author. Tel.: þ64 3 479 7936; fax: þ64 3 479 8414. E-mail address: [email protected] (G. Sole). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.03.008

conservative interventions for patients with musculoskeletal disorders of the shoulder. Therefore, the purpose of this systematic review was to determine the level of evidence of the effectiveness of MT in the management of shoulder musculoskeletal disorders. 2. Methodology 2.1. Types of studies and participants Studies included randomised controlled clinical trials with language restricted to English or German (Fig. 1). Research papers on humans with disorders of the shoulder girdle, including fractures, dislocation, degenerative/osteoarthritis and orthopedic surgery were included. Studies including subjects with systemic diseases such as rheumatoid arthritis, neurological disorders such as stroke, or shoulder symptoms of spinal origin were excluded. 2.2. Interventions and outcomes Studies where at least one application of MT (manipulation, passive joint or soft tissue mobilization techniques or massage) was applied to either the shoulder girdle, cervical or thoracic spine were included (Paris, 2000; Vernon et al., 2007). Multi-modal interventions were included if the effects of MT could be differentiated from the other interventions. Studies reporting pain, ROM, functional outcomes, patient satisfaction or recovery rate were considered.

464

C.-Y.C. Ho et al. / Manual Therapy 14 (2009) 463–474

Step 1: Computer database search (n = 4311): MEDLINE (n = 936) PUBMED (n = 1384) AMED (n = 322) CINHAL (n = 514) EMBASE (n = 1004) PEDro (n = 151)

Manual search of relevant journals (n = 57): Manual Therapy (n = 34) Journal of Manual and Manipulative Therapy (n = 15) Journal of Manipulative and Physiological Therapeutic (n = 8)

Duplicate articles excluded (n = 1406)

Unrelated articles excluded based on title and article type (n = 2122)

Step 2: Screening of title and abstract for inclusion and exclusion (n = 840)

Irrelevant articles excluded (n = 699) Case report (n = 52)

Uncertainfull article retrieved (n = 58)

Irrelevant articles excluded (n = 58)

Step 3: Relevant articles full article retrieved (n = 30) Articles excluded (n = 17): • Laboratory studies (n = 2) • Case studies (n = 2) • Economic evaluation (n = 1) • Mutli-model (n = 12) Articles included in systematic review (n = 13)

Step 4: Hand search of reference list for potentially relevant articles (n = 1)

Step 5: Quality Assessment using PEDro scale (n = 14)

Step 6: Data extraction and analysis Fig. 1. Flow diagram of study selection process.

2.3. Search strategy An electronic search was performed of MEDLINE (1950 to January 2007), CINAHL (1982 to January 2007), AMED (1985 to January 2007), EMBASE (1988 to January 2007), PUBMED (1950 to January 2007) and PEDro (1950 to January 2007), and included a combination of search terms related to shoulder musculoskeletal disorders and to MT (Appendix I). Supplementary searches were done on the PEDro database, and by hand searching all volumes of three relevant MT journals and reference lists of the included studies. 2.4. Study selection One assessor (CH) screened all titles for relevance and duplication. Two independent assessors (CH and GS) blinded to journal,

authors and institutions screened potentially relevant titles and abstracts for inclusion. Full articles were retrieved if there was insufficient information from the title and abstract to determine relevance. If consensus for study eligibility was not reached, a third assessor (JM) was involved.

2.5. Quality assessment Randomised controlled trials (RCTs) were rated independently by two assessors (CH and JM) using the PEDro scale. Disagreements in scores were resolved by consensus or a third opinion (GS) where required. A study was considered to be of high quality if the PEDro score was greater than five and of low quality if the PEDro score was five or less (Maher et al., 2003).

C.-Y.C. Ho et al. / Manual Therapy 14 (2009) 463–474 Table 1 Levels of evidence by van Tulder et al. (2003). Level of evidence

Description

Strong evidence Moderate evidence

Consistent findings among multiple high-quality RCTs Consistent findings among multiple low-quality RCTs and/or CCTs and/or one high-quality RCT One low-quality RCT and/or CCT Inconsistent findings among multiple trials No RCTs or CCTs.

Limited evidence Conflicting evidence No evidence

465

meta-analysis was not performed. A qualitative analysis using levels of evidence to define treatment effectiveness was performed (Table 1, van Tulder et al., 2003). These levels of evidence criteria considers participants, interventions, controls, outcomes, both high and low methodological quality of the included studies and consistency of findings between studies, and are widely used (Faber et al., 2006; Woodley et al., 2007). 3. Results

RCT: randomised clinical trial; CCT: clinical controlled trial.

3.1. Selection of studies and study characteristics 2.6. Data extension and analysis Data were extracted by one author (CH) for characteristics of participants, shoulder conditions, interventions and outcomes of pain, ROM and function using a standardized form (Fig. 1, Step 6). If a study reported more than one measure for an outcome, the measure most commonly used between studies or deemed to be more representative of function was used. Data were extracted for outcomes immediately following the intervention period (initial follow-up) and, where available, at the final follow-up time point (long-term follow-up). Pain outcomes for overall pain, functional pain, 24-h pain and pain on movement and night pain were considered. For ROM, active (and passive for studies with patients with adhesive capsulitis [AC]) measures of abduction in degrees were extracted. For function, patient satisfaction and functional outcome questionnaires were considered. For continuous variables, the mean difference (95% confidence intervals, CI) between groups was calculated from endpoint scores or change scores (Herbert, 2000a, Clare et al., 2004). For dichotomous outcomes, relative risks (RR, with 95% CI) were calculated (Herbert, 2000b). Results for each study were analyzed within commonly reported diagnostic subgroups. Trials were assessed for clinical heterogeneity with respect to the participants, intervention and outcomes. Due to the wide range of disorders and interventions,

Fourteen RCTs (n ¼ 888 subjects) from 840 citation postings and hand searching results were included (Fig. 1). The studies investigated patients with AC (Table 2), SIS (Table 3) and non-specific shoulder pain/dysfunction (Table 4). Sample sizes ranged from 14 to 172 patients, averaging 64 patients per study. The mean age of patients ranged from 44 to 65 years. 3.1.1. Interventions Interventions included joint mobilizations (Maitland concept) of the shoulder girdle (Bulgen et al., 1984; Conroy and Hayes, 1998; Maricar and Chok, 1999; Vermeulen et al., 2006) mobilization of the upper quarter (Winters et al., 1997; Bang and Deyle, 2000; Bergman et al., 2004), manipulation (Winters et al., 1997; Bergman et al., 2004), Cyriax’ manipulation and deep transverse frictions (GulerUysal and Kozanoglu, 2004), ‘‘Mobilization-with-Movement’’ (MWM) (Teys et al., 2008) or soft tissue massage (Van den Dolder and Roberts, 2003). Bang and Deyle (2000) used a pragmatic combination of joint and soft tissue mobilization techniques based on the upper quartile movement impairment assessed for the individual participant in the experimental group; whereas Conroy and Hayes (1998) used glenohumeral joint mobilizations for the experimental group, but included soft tissue mobilization techniques as part of ‘‘conventional physiotherapy’’ for both participant groups.

Table 2 Study characteristics: adhesive capsulitis. Author/year

Condition

Participants characteristics

Interventions

Outcomes

Bulgen et al. (1984)

MOR: Not stated Four groups: intraarticular injection; mobilizations; ice therapy and no treatment

n ¼ 42, 28 female, 14 male Mobilization group: n ¼ 11 Ice therapy group: n ¼ 12 Steroid group: n ¼ 11 Control group: n ¼ 8 Age ¼ 55.8 (44–74) y DOS ¼ 4.8 (1–12) months

Intervention period: varied between groups. All subjects were taught pendular exercises 2–3 min every hour and pain medication if required. Mobilization group: Maitland’s mobilizations Three times weekly for 6 weeks Ice therapy group: Ice pack followed by PNF Three times weekly for 6 weeks Steroid group: Intra-articular /subacromial injection weekly for 3 weeks Non-treatment group: Pendular exercises and pain medication

Follow-up period: weekly for the first 6 weeks then monthly for a further 6 months Outcome measures: Verbal reports of progress Passive ROM (goniometry): Total flexion Total abduction External rotation Glenohumeral flexion Internal rotation

Binder et al. (1984)

MOR: Not stated Four groups: intraarticular injection; mobilizations; ice therapy and no treatment

n ¼ 40, Gender not stated Mobilization group: n ¼ 11 Ice therapy group: n ¼ 11 Steroid group: n ¼ 10 Control group: n ¼ 8 Age ¼ not stated DOS ¼ not stated

Follow-up from original study (Bulgen et al., 1984)

Follow-up period: 40–48 months after initial presentation Outcome measures: Persistent or recurrent pain/or restriction of movement Passive ROM (goniometry): Total flexion Total abduction External rotation Total rotation (continued on next page)

466

C.-Y.C. Ho et al. / Manual Therapy 14 (2009) 463–474

Table 2 (continued ) Author/year

Condition

Participants characteristics

Interventions

Outcomes

Guler-Uysal and Kozanoglu (2004)

MOR: Not stated Two groups: Cyriax approach; physical modalities

n ¼ 40 Experimental group: n ¼ 20, 15 female, 5 male Age ¼ 53.6  6.9 (43–70) y DOS ¼ 7.6  3.9 (2–12) months Control group: n ¼ 20, 13 female, 7 male Age ¼ 58.4  9.7(44–82) y DOS ¼ 5.6  3.9 (2–12) months

3-week intervention Active stretching and pendulum movements were performed by both groups after each session. Experimental group: Deep transverse frictions and manipulation. 1 h session three times weekly. Control group: Hot packs and shorts wave diathermy. 1 h session 5 times weekly.

Follow-up period: End of 1 and 2 week Outcome measures: Pain using VAS: Spontaneous pain Night pain Pain with motion Passive ROM (goniometry): Flexion Abduction Internal rotation External rotation Recovery rate

Maricar and Chok (1999)

MOR: Not stated Two groups: manual therapy þ exercises and exercises alone

n ¼ 32 Experimental group: n ¼ 16, 7 female, 9 male Age ¼ 57.9  9.5 y Control group: n ¼ 16, 6 female, 10 male Age ¼ 54.9  5.4 y DOS of both groups ¼ average: 3 months

8-week intervention Experimental group: Mobilization of upper quadrant using Maitland Grade IIIþ and IV and exercises Once weekly for 8 weeks Control group: Exercises Once weekly for 8 weeks

Follow-up period: 3, 5, 7, and 8 week Outcome measures: AROM (goniometry) Flexion External rotation Internal rotation Hand-behind-back

Nicholson (1985)

MOR: Toss of coin Two groups: joint mobilization þ exercises and exercises alone

n ¼ 20 Experimental group: n ¼ 10, 6 female, 4 male Age ¼ 51(31–70) 12.16 y DOS ¼ 27.6  33.41 (1–104) weeks Control group: n ¼ 10, 4 female, 6 male Age ¼ 55  16.43 (20–77) y DOS ¼ 30.8  31.28 (3–104) weeks

4-week intervention Experimental group: Gliding and distractive mobilization techniques and exercises Two to three times weekly for 4 weeks Control group: Exercises Repeat the exercises three times daily independently

Follow-up period: Weekly for 4 weeks Outcome measures: Pain questionnaire ROM (goniometry): Active internal rotation Active external rotation Active abduction Passive abduction

Vermeulen et al. (2006)

MOR: Randomnumber generator Two groups: highgrade mobilization (HG) and Low grade (LG)

n ¼ 100 HG mobilizations: n ¼ 49, 32 female, 17 male Age ¼ 51.6 (7.6) y DOS ¼ 8 (5–14.5) months LG mobilizations: n ¼ 51, 34 female, 17 male Age ¼ 51.7 (8.6) y DOS ¼ 8(6–14) months

12-week intervention Subjects might have further treatments as suggested by orthopedic surgeon following intervention period LG mobilizations: Maitland grade I and II joint mobilization Number of sessions 18.6  4.9 HG mobilizaionts: Maitland grade III and IV joint mobilization Number of sessions 21.5  2.5 2 times weekly for 30 min for a maximum of 12 weeks A minimal duration of exposure to the therapy of at least 6 weeks

Follow-up period: 3, 6 and 12 month Outcome measures: Active and Passive ROM (goniometry): Abduction Forward flexion External rotation Shoulder Rating Questionnaire (SRQ) Shoulder Disability Questionnaire (SDQ) Pain using VAS: Pain at rest Pain during movement Pain during the night General Health using SF-36

RCT ¼ randomized controlled trial; MOR ¼ method of randomization; DOS ¼ duration of symptoms; ROM ¼ range of motion; PNF ¼ proprioceptive neuromuscular facilitation; MWM ¼ mobilization with movement; s ¼ seconds; min ¼ minutes; VAS ¼ visual Analogue Scale; y ¼ years; data given as means  SD (range), unless otherwise stated.

MT was used in isolation (Winters et al., 1997; Winters et al., 1999; Van den Dolder and Roberts, 2003; Vermeulen et al., 2006; Teys et al., 2008) or in combination with exercises (Nicholson, 1985; Conroy and Hayes, 1998; Maricar and Chok, 1999; Bang and Deyle, 2000; Guler-Uysal and Kozanoglu, 2004; Citaker et al., 2005), hot packs (Conroy and Hayes, 1998; Citaker et al., 2005) or medical care (Bergman et al., 2004). One study compared high-grade (HG) joint mobilizations, defined as grade III or higher on Maitland grading system (Maitland, 1991), to low grade (LG) in patients with AC (Vermeulen et al., 2006). This study was included as there was consensus amongst the current authors to consider LG mobilizations a control condition as clinical lore would usually indicate the use of high rather than low-grade mobilization techniques with the aim of improving ROM in patients with AC. Control interventions included ice therapy (Binder et al., 1984; Bulgen et al., 1984), electrophysical modalities (Guler-Uysal and Kozanoglu, 2004), exercise

(Nicholson, 1985; Conroy and Hayes, 1998; Maricar and Chok, 1999; Bang and Deyle, 2000), education, and proprioceptive neuromuscular facilitation (PNF) (Citaker et al., 2005). The number of intervention sessions ranged from 3 to 20 (average 11 sessions). Twelve studies investigated immediate effects following intervention, with the follow-up period ranging from 3 days to 4 years. Two studies also investigated long-term effects (Bergman et al., 2004; Vermeulen et al., 2006). Two studies investigated long-term results of subjects included in earlier reported studies (Binder et al., 1984; Winters et al., 1999). 3.1.2. Measures The most common measure was pain (such as visual analogue scales, VAS) and goniometric ROM which were reported in 10 out of 14 studies. Various functional outcome measures were used (Table 2).

C.-Y.C. Ho et al. / Manual Therapy 14 (2009) 463–474

467

Table 3 Study characteristics: shoulder impingement syndrome. Author/year

Condition

Participants characteristics

Interventions

Outcomes

Bang and Deyle (2000)

MOR: Table of random numbers Two groups: Manual therapy þ exercise and exercise alone

n ¼ 52 Manual therapy group: n ¼ 28, 10 female, 18 male Age ¼ 42  10.1 (17–64) y DOS ¼ 5.6  3.7 (1–12) months Exercise group: n ¼ 24, 12 female, 12 male Age ¼ 45  8.4 (24–60) y DOS ¼ 4.4  2.8 (1–12) months

3-week intervention Twice weekly for a total of 6 visits Manual therapy group: standardized flexibility and strengthening program, Joint mobilization of upper quarter and soft tissue massage. Exercise group: Standardized flexibility and strengthening program

Follow-up period: After 6 treatment sessions Outcome measures: Perception of shoulder function: Functional assessment questionnaire (9 categories): Pain using (VAS): Overall pain intensity Raising arm overhead Behind the back activities Reaching across body Lifting with problem arm Lying on shoulder Pushing and pulling Carrying an object with arm at side Performance of usual physical activity, sport or hobby Resisted break test: IR; ER and abduction Active abduction Isometric strength using a stabilized electronic dynamometer: Internal rotation External rotation Abduction

Citaker et al. (2005)

MOR: not stated Two groups: Hot pack þ mobilization þ exercises and hot pack þ PNF þ exercises

n ¼ 40, Gender not stated Mobilization group: n ¼ not stated Age ¼ 52.8  9.86 y DOS ¼ not stated PNF group: n ¼ not stated Age ¼ 55.5  8.95 y DOS ¼ not stated

Length of intervention period: Not stated 20-session treatment followed by 3 weeks of theraband exercises Mobilization group: Manual mobilization, hot packs, theraband exercises and Codman pendulum exercises PNF group: PNF , hot packs, theraband exercises and Codman pendulum exercises

Follow-up period: Unclear, stated as after intervention period Outcome measures: Pain using VAS ROM (goniometry): Flexion Abduction External rotation Internal rotation Hyperextension University of California at Los Angeles Shoulder Rating Scale (UCLA) Categorized into pain, function, AROM, strength and patient satisfaction Total score: 2–35 28 or less ¼ unsatisfactory 29–33 ¼ good 34–35 ¼ excellent

Conroy and Hayes (1998)

MOR: not stated Two groups: joint mobilization þ soft tissue massage and soft tissue massage only

n ¼ 14, 6 female, 8 male Experimental group: n ¼ 7 Age ¼ 55  10.2 y DOS ¼ not stated Control group: n ¼ 7 Age ¼ 50.7  16.5 y DOS ¼ not stated

3-week intervention 3 sessions per week Experimental group: Joint mobilization of subacromial and glenohumeral joints, soft tissue mobilization, hot pack, stretching and strengthening exercise, and patient education Manual therapy: oscillatory pressure of 2–3 oscillations per second, each technique was administered 2–4 times (30 s each) Control group: Soft tissue mobilization, hot pack, stretching and strengthening exercise and patient education

Follow-up period: 3 week Outcome measures: Maximum pain over the preceding 24-hr period (VAS) Pain with subacromial compression test (VAS) AROM (goniometry): Shoulder flexion Abduction Scapular plane elevation Internal rotation External rotation Overhead Function (graded on a 3-point scale): Reach behind head Reach across and around the upper body Touch a mark on the wall that required 135 of shoulder flexion.

RCT ¼ randomized Controlled Trial; MOR ¼ method of randomization; DOS ¼ duration of symptoms; ROM ¼ range of motion; PNF ¼ proprioceptive neuromuscular facilitation; MWM ¼ mobilization with movement; s ¼ seconds; min ¼ minutes; VAS ¼ visual Analogue Scale; y ¼ years; data given as means  SD (range), unless otherwise stated.

468

C.-Y.C. Ho et al. / Manual Therapy 14 (2009) 463–474

Table 4 Study characteristics: non-specific shoulder pain or dysfunction. Author/year

Condition

Participants characteristics

Interventions

Outcomes

Bergman et al. (2004)

MOR: Sealed envelope Two groups: Manipulative therapy þ medical care and medical care alone

n ¼ 150 Manipulative group: n ¼ 79, 42 female, 37 males Age ¼ 48.4  12.4 y DOS: 0–12 weeks ¼ 53, 12–26 weeks ¼ 26 Medical care group: n ¼ 71, 32 female, 39 males Age ¼ 47.8  11.8 y DOS: 0–12 weeks ¼ 50, 12–26 weeks ¼ 21

12-week intervention Manipulative group: Usual medical care and mobilization or manipulative to cervical spine, upper thoracic spine and adjacent ribs The mean duration of a manipulative session 23  13 min maximum of 6 treatments over a 12-week period Medical care group: Usual medical care

Follow-up period: Week 6, 12 , 26 and 52 Outcome measures: Patient-perceived recovery (7-point ordinal scale) Patient’s perception of ‘‘cured’’ Severity of the main complaint during preceding week on an 11-point scale (0 ¼ best 10 ¼ worst) Shoulder pain (4-point ordinal scale): At rest In motion Night pain Sleeping problems caused by pain Inability to lie on the painful side Degree of radiation General pain Shoulder disability questionnaire for the functional status of the shoulder in the preceding 24 h 16 items EuroQol health: 5 items 3-point ordinal scale

Teys et al. (2008)

MOR: Drawing of lots Three groups: MWM; Sham and control

n ¼ 24, 13 female, 11 male Age ¼ 46.1  9.86 (20–64) y DOS ¼ 1–12 months

3-day intervention Experimental group: Mobilization with movement: Postero-lateral glide of glenohumeral joint during elevation 3 sets of 10 repetitions with a rest interval of 30 s between each set. Sham group: Anterior glide with minimal pressure applied. Elevation through half of available pain-free range. 3 sets of 10 repetitions with a rest interval of 30 s between each set. Control group: No manual contact

Follow-up period: Each treatment session Outcome measures: Pain-free AROM (goniometry): Scapular plane elevation Pressure pain threshold using pressure pain algometry and by palpating the most sensitive point located over anterior aspect of the shoulder

Van den Dolder and Roberts (2003)

MOR: Sealed envelope Two groups: Massage and control

n ¼ 29 Massage group: n ¼ 15, 4 female, 11 male Age ¼ 63.1  9.9 y DOS ¼ median 26 (13–26) weeks Control group: n ¼ 14, 5 female, 9 male Age ¼ 65.9  9.2 y DOS ¼ median 30 (23–91) weeks

2-week intervention Massage group: 6 treatments of soft tissue massage around the shoulder Each treatment 15–20 min Control group: No treatment for 2 weeks

Follow-up period: 2 week Outcome measures: Pain intensity using Short Form McGill Pain Questionnaire: 3 sections 1st: A list of 15 words to describe pain 2nd: 100 mm VAS pain experienced over last 24 h 3rd: Present pain index Functional disability using a Patient Specific Functional Disability Measure: Active ROM using photographs: Flexion Abduction Hand-behind-back

Winters et al. (1997)

MOR: Not stated 2 categories: Shoulder girdle and synovial Shoulder girdle: Manipulation and physiotherapy Synovial: Corticosteroid injection; manipulation and physiotherapy

n ¼ 172 Shoulder girdle groups: Manipulation: n ¼ 29, 15 female, 14 male Age ¼ 43.9  12.6 y DOS ¼ median 3 weeks Physiotherapy: n ¼ 29, 18 female, 11 male Age ¼ 46.4  11.2 y DOS ¼ median 4 weeks Synovial groups: Manipulation: n ¼ 32, 17 female, 15 male Age ¼ 46.7  12.1 y DOS ¼ median 9 weeks Physiotherapy: n ¼ 35, 14 female, 21 male Age ¼ 53.1  12.6 y DOS ¼ median 4 weeks Corticosteroid injection: n ¼ 47, 32 female, 15 male Age ¼ 53.5  12.5 y DOS ¼ median 8 weeks

Up to 11-week intervention Manipulation group: mobilization and manipulation of the cervical spine, upper thoracic spine, upper ribs, acromioclavicular joints and glenohumeral joint Once a week with a maximum of 6 treatments Physiotherapy group: Exercise therapy, massage and physical applications Twice a week Injection group: 1–3 injections

Follow-up period: 2, 6 and 11 weeks Outcome measures: Shoulder pain score (4-point scale): Pain at rest Pain during motion Pain during the night Sleeping problems because of pain Inability to lie on affected side Presence of radiated pain Together with a 101 point numerical pain scale Patient’s perception of ‘‘cured’’

C.-Y.C. Ho et al. / Manual Therapy 14 (2009) 463–474

469

Table 4 (continued ) Author/year

Condition

Participants characteristics

Interventions

Outcomes

Winters et al. (1999)

MOR: not stated 2 categories: Shoulder girdle and synovial Shoulder girdle: Manipulation and physiotherapy Synovial: Corticosteroid injection; manipulation and physiotherapy

Questionnaire sent to all 172 subjects, 130 (76%) could be evaluated Gender ¼ not stated Age ¼ not stated DOS ¼ not stated Shoulder girdle groups: Manipulation: n ¼ 18 Physiotherapy: n ¼ 22 Synovial groups: Injection: n ¼ 38 Manipulation: n ¼ 26 Physiotherapy: n ¼ 26

Follow-up from original study (Winters et al., 1997)

Follow-up period: 2–3 years after original study Outcome measures: Persisting, recurrent or new shoulder complaints Patient’s perception of ‘‘cured’’

RCT ¼ randomized Controlled Trial; MOR ¼ method of randomization; DOS ¼ duration of symptoms; ROM ¼ range of motion; PNF ¼ proprioceptive neuromuscular facilitation; MWM ¼ mobilization with movement; s ¼ seconds; min ¼ minutes; VAS ¼ visual Analogue Scale; y ¼ years; data given as means  SD (range), unless otherwise stated.

3.2. Methodological quality PEDro quality scores ranged from 3 to 8 out of 10 (Fig. 2). Eight of the 14 studies scored 6 or more. The most common sources of bias were failure to blind therapists (100% of studies), failure to blind subjects (86% of studies), failure to conceal allocation (79% of studies) and lack of analysis by intention-to-treat (71% of studies). Thirteen of 154 (8%) quality criteria assessed across studies required discussion to reach consensus between assessors. Three criteria required an opinion from the third assessor.

3.3. Effects of manual therapy All results are reported as mean differences (95% CI) for the effect of MT compared to control for outcome measures of pain, ROM and function unless otherwise stated. Study

PEDro scale item number 1

a

Total score /10

2 3 4 5 6 7 8 9 10 11

Bang and Deyle 2000

6

Bergman et al. 2004

8

Binder et al. 1984

3

Bulgen et al. 1984

3

Citaker et al. 2005

4

Conroy and Hayes 1998

7

Guler-Uysal and Kozanoglu 2004 Maricar and Chok 1999

6

Nicholoson 1985

6

Teys et al. 2008

8

van den Dolder and Roberts 2003 Vermeulen et al. 2006

7

Winters et al. 1997

5

Winters et al. 1999

3

4

7

Fig. 2. Pedro score table. aCriteria 1 was not used to calculate the PEDro score. ¼ criteria met. Pedro Scale item. 1. Eligibility criteria were , ¼ criteria not met. specified. 2. Subjects were randomly allocated to groups. 3. Allocation was concealed. 4. The groups were similar at baseline regarding the most important prognostic indicators. 5. There was blinding of all subjects. 6. There was blinding of all therapists who administered the therapy. 7. There was blinding of all assessors who measured at least one key outcome. 8. Measures of at least one key outcome were obtained from more than 85% of the subjects initially allocated to groups. 9. All subjects for whom outcome measures were available received the treatment or control condition as allocated or, where this was not the case, data for at least one key outcome was analyzed by ‘‘intention to treat’’. 10. The results of between-group statistical comparisons are reported for at least one key outcome. 11. The study provides both point measures and measures of variability for at least one key outcome.

3.3.1. Adhesive capsulitis 3.3.1.1. Pain. No differences were found between HG MT and LG MT with respect to pain at initial or long-term follow-up in one highquality trial (Fig. 3) (Vermeulen et al., 2006). These findings are consistent with the high-quality trial of Guler-Uysal and Kozanoglu (2004) (for initial follow-up), comparing MT using the Cyriax approach (Cyriax, 1984) to hot packs and short wave diathermy (Fig. 3). 3.3.1.2. Range of motion. For active ROM, two studies (Nicholson, 1985; Maricar and Chok, 1999) showed that mobilization with exercise was no more effective than exercise alone in the shortterm. Vermeulen et al. (2006) found in a high-quality trial that HG joint mobilizations were more effective than LG mobilizations when active ROM was measured both immediately and 12 months following the intervention period (Fig. 3). For passive ROM, Nicholson (1985) showed that mobilization with exercise was more effective than exercise alone. In contrast, Guler-Uysal and Kozanoglu (2004) found that manipulation with deep transverse frictions following the Cyriax approach (Cyriax, 1984) was no more effective than the application of physical modalities. When long-term effects of MT were investigated, Binder et al. (1984) showed in a low-quality trial that MT was no more effective than intra-articular steroid injection, ice therapy or no treatment. Vermeulen et al. (2006) found HG mobilizations were more effective than LG mobilizations at initial and long-term follow-up. 3.3.1.3. Function. Guler-Uysal and Kozanoglu (2004) did not show a better recovery rate (number of patients who reached 80% of normal shoulder ROM) for patients receiving deep massage and manipulation than patients receiving physical modalities [Relative Risk (95% CI) ¼ 1.5 (1.0–2.0)]. HG mobilizations were more effective in improving shoulder function when compared to LG mobilizations for long-term outcomes but not short-term outcomes (Fig. 3) (Vermeulen et al., 2006). The qualitative analysis defining treatment effectiveness (Table 5) showed moderate evidence that MT was no more effective than other interventions in decreasing pain measures and improving ROM and function. However, there was moderate evidence that HG MT compared to LG MT was more effective for increasing ROM and long-term functional outcomes. 3.3.2. Shoulder impingement syndrome 3.3.2.1. Pain. The addition of pragmatic MT was shown to be effective in reducing pain compared to exercise alone (Bang and Deyle, 2000) and when joint mobilizations were compared to ‘‘conventional’’ physiotherapy alone (Conroy and Hayes, 1998) in

470

C.-Y.C. Ho et al. / Manual Therapy 14 (2009) 463–474

Fig. 3. Distribution of estimates from five studies for the mean difference in effect of manual therapy compared to control (or placebo) on pain (-), AROM (:), PROM (6) and function (A) for patients with adhesive capsulitis. The size of each estimate symbol is proportional to the study’s sample size. The horizontal bars report 95% confidence intervals. Pain and function are measured on a 0–100 scale, ROM is measured in degrees. Positive results indicate a beneficial effect of manual therapy over control. Mob ¼ mobilization; AROM ¼ active range of motion; PROM ¼ passive range of motion. aVermeulen et al. (2006) compared high-grade to low-grade mobilization techniques. Low-grade mobilization techniques were considered as a control condition for the purpose of the systematic review as these grades would not be applied for the aim of increasing ROM.

high-quality trials. Converseley, Citaker et al. (2005), a low-quality trial, reported that joint mobilizations in addition to exercise and modalities were no more effective than exercise, modalities and PNF in improving pain (Fig. 4). 3.3.2.2. Range of motion. Joint mobilizations were no more effective in improving active ROM than conventional physiotherapy alone (Conroy and Hayes, 1998) and PNF (Citaker et al., 2005) for short-term outcomes (Fig. 4). 3.3.2.3. Function. Bang and Deyle (2000) found that pragmatic MT was effective in improving function compared to exercise alone. Similarly, Citaker et al. (2005) showed that joint mobilizations were effective in comparison to PNF. Assessment of function on overhead reaching (Conroy and Hayes, 1998) showed that there was no additional benefit of joint mobilizations to physiotherapy which included soft tissue mobilization techniques (Fig. 4). In summary, there was no clear evidence to suggest additional benefits of MT to other interventions in the management of patients with SIS (Table 5). 3.3.3. Non-specific shoulder pain/dysfunction 3.3.3.1. Pain. The additional effect of MT of the upper quarter to medical care was shown to be effective in reducing pain originating from the shoulder girdle at initial follow-up in a high-quality trail (Bergman et al., 2004). In a low-quality trial (Winters et al., 1997) manipulation was beneficial compared to traditional physiotherapy at initial follow-up. However, manipulation was ineffective in treating shoulder complaints where shoulder disorders were classified as originated from synovial structures when compared to traditional physiotherapy or corticosteroid injection (Winters et al., 1997) (Fig. 5). In addition, Van den Dolder and Roberts (2003) found two-weeks of massage more effective for pain relief compared to

no treatment. Long-term, effects of MT was no more greater than usual medical care (Bergman et al., 2004). 3.3.3.2. Range of motion. MWM were effective for improving shortterm active ROM compared to sham or no treatment in a highquality trial (Teys et al., 2008). Similarly, massage of the shoulder was effective compared to no treatment in a high-quality trial (Fig. 5, Van den Dolder and Roberts, 2003). 3.3.3.3. Function. Massage was effective for improving function compared to no treatment (Fig. 5) (Van den Dolder and Roberts, 2003). However, the addition of MT to usual medical care was no more effective for improving function at initial and long-term follow-up (Bergman et al., 2004). Winters et al. (1997, 1999) investigated patients’ perception of recovery following an 11-week intervention and also 2–3 years later. Manipulation was more effective than traditional physiotherapy for treating shoulder complaints originating from the shoulder girdle [RR (95% CI): 6.7 (2.2–20)]. In the group with synovial shoulder complaints, manipulation was no more effective than traditional physiotherapy. Further, it was ineffective when compared to corticosteroid injection for synovial shoulder complaints [RR (95% CI): 2 (0.9–4.4); 0.5 (0.3–0.9), respectively]. At the 2–3 year follow-up, manipulation was shown to be no more effective in improving function than traditional physiotherapy and injection in both groups [RR (95% CI): 1.2 (0.8–1.8); 0.9 (0.7–1.2); 1 (0.7–1.3), respectively] (Winters et al., 1999). For non-specific shoulder pain/dysfunction, there was moderate evidence to suggest MT was effective in the short-term for increasing ROM when compared to sham type treatment and control groups, and massage was effective when compared to no treatment (Table 5). Moderate evidence suggests that MT is no more effective in improving function in the long-term compared to other interventions.

C.-Y.C. Ho et al. / Manual Therapy 14 (2009) 463–474

471

Table 5 Table of level of evidence for the effectiveness of manual therapy for musculoskeletal disorders of the shoulder. Shoulder pathology

Outcome measures

Follow-up

Evidence

Adhesive capsulitisa

Painb

Initial

Moderate evidence exists to suggest that MT is no more effective for improving pain when compared to other interventions. Moderate evidence exists to suggest that high-grade MT is no more effective than low-grade MT for improving pain. Moderate evidence exists to suggest that high-grade MT is no more effective than low-grade MT for improving pain. Conflicting evidence exists regarding the effect of MT on PROM when compared to other interventions. Moderate evidence exists to suggest that high-grade MT is more effective for improving PROM than low-grade manual therapy. Limited evidence exists to suggest that MT is no more effective for improving PROM when compared to other interventions. Moderate evidence exists to suggest that high-grade MT is more effective for improving PROM than low-grade MT. Moderate evidence exists to suggest that MT is no more effective for improving AROM when compared to other interventions. Moderate evidence exists to suggest that high-grade MT is more effective for improving AROM than high-grade MT. Moderate evidence exists to suggest that high-grade MT is more effective for improving AROM when compared to low-grade MT. Moderate evidence exists to suggest that MT is no more effective for improving recovery when compared to other interventions. Moderate evidence exists to suggest that high-grade MT is no more effective for improving shoulder function than low-grade MT. Moderate evidence exists to suggest that high-grade MT is more effective for improving shoulder function than low-grade MT.

Long-term PROM

Initial

Long-term

AROM

Initial

Long-term Function

Initial

Long-term Shoulder impingement syndrome

Shoulder pain/dysfunction

Pain AROM

Initial Initial

Function

Initial

Pain

Initial

Long-term AROM

Initial

Function

Initial

Long-term

Conflicting evidence exists regarding the effect of MT on pain when compared to other interventions. Moderate evidence exists to suggest that MT is no more effective for improving AROM when compared to other interventions. Conflicting evidence exists regarding the effect of MT on function when compared to other interventions. Conflicting evidence exists regarding the effect of MT on pain when compared to other interventions. Moderate evidence exists to suggest that massage is more effective for improving pain compared to no treatment. Moderate evidence exists to suggest that MT is no more effective for improving pain when compared to other interventions. Moderate evidence exists to suggest that MT is more effective for improving AROM compared to sham or no treatment. Moderate evidence exists to suggest that massage is effective for improving AROM compared to no treatment. Conflicting evidence exists regarding the effect of MT on function compared to other interventions. Moderate evidence exists to suggest that massage is effective for improving function compared to no treatment. Moderate evidence exists to suggest that MT is no more effective in improving function or recovery when compared to other interventions.

AROM ¼ active range of motion; PROM ¼ passive range of motion; MT ¼ manual therapy. a Effect statement for adhesive capsulitis does not include study by Bulgen et al. (1984), because insufficient statistical data of study outcomes were given. They reported ‘‘at the end of treatment, the groups were significantly different at the 2% level, but by the end of the study there was no significant difference between the groups’’. b Effect statement for adhesive capsulitis does not include the study by Nicholson (1985), because the pain scale used was not specified, so the score could not be converted to the scale of 0–100 for effect size calculation. The author reported the change pain score in mean degrees (standard deviation): experimental group ¼ 5.10 (4.56) and control group ¼ 2.90 (4.41) and P value ¼ 0.7201.

4. Discussion This review found inconsistent evidence for the effectiveness of MT for various shoulder disorders compared to control interventions and no treatment, contrasting with other published reviews regarding treatment efficacy for SIS. Green et al. (2003), Michener et al. (2004) and Faber et al. (2006) reported limited evidence suggesting that MT combined with exercise was more effective than exercise alone in patients with SIS, whereas here there was conflicting evidence for the benefit of MT on pain and function. The current inclusion of the study by Citaker et al. (2005), finding that the addition of MT yielded no added benefit in SIS, is likely to have contributed to our differing findings. Conflicting evidence for effects on pain and function in SIS may be explained by variable definitions of MT. Bang and Deyle (2000)

found a pragmatic approach, including joint and soft tissue mobilizations to the individual-specific movement impairment of the upper quadrant to be more effective than therapeutic exercise alone. Conroy and Hayes (1998) included soft tissue mobilizations in both the experimental and the control group, adding joint mobilizations to the former. Different forms of MT may have similar neurophysiological effects, despite differences in mechanical applications (Bialosky et al., 2009). It is thus possible, that these common effects contributed to the lack of significant differences for between-group outcomes by Conroy and Hayes (1998). Based on findings of our review, clinicians should consider incorporating soft tissue and joint mobilization techniques in addition to therapeutic exercises for patients with SIS, based on an individual assessment. Future RCTs should investigate pragmatic approaches to determine the effectiveness of MT in the management of patients with SIS.

472

C.-Y.C. Ho et al. / Manual Therapy 14 (2009) 463–474

Fig. 4. Distribution of estimates from three studies for the mean difference in the effects of manual therapy compared to control (or placebo) on pain (-), AROM (:) and function (A) for patients with shoulder impingement syndrome. The size of each estimate symbol is proportional to the study’s sample size. The horizontal bars report 95% confidence intervals. Pain and function are measured on a 0–100 scale, ROM is measured in degrees. Positive results indicate a beneficial effect of manual therapy over the control. AROM ¼ active range of motion.

Our findings indicate that MT may not be more effective for the management of pain and improving ROM and function for patients with AC than other interventions. However, the studies had a Pedro rating of 6 or less (Binder et al., 1984; Nicholson, 1985; Maricar and Chok, 1999; Guler-Uysal and Kozanoglu, 2004). Vermeulen et al. (2006) found that when comparing high-grade to low-grade joint mobilizations, the former was more effective in improving ROM in the short and the long term, and ROM and function in the long term. In the absence of higher quality RCT, the use of MT in patients with AC still relies predominantly on clinical reasoning, with more support for the aim of improving ROM and function, than for pain management.

The lack of clear description and wide range of MT, further compounded by the difficulty of consistent sub-grouping of patients with unspecific shoulder pain/dysfunction make it difficult to provide clear guidelines for the clinician. The evidence was conflicting or moderate that MT may be more effective than other interventions for pain management and improving ROM and function for patients in this large group. One study investigated the effect of massage alone on shoulder pain with beneficial short-term effects (Van den Dolder and Roberts, 2003). The control group of patients received no treatment, thus the positive findings for the experimental group may have, in part, indicated

Fig. 5. Distribution of estimates from four studies for the mean difference in the effects of manual therapy compared to control (or placebo) on pain (-), AROM (:) and function (A) for patients with non-specific shoulder pain/dysfunction. The size of each estimate symbol is proportional to the study’s sample size. The horizontal bars report 95% confidence intervals. Pain and function are measured on a 0–100 scale, ROM is measured in degrees. Positive results indicate a beneficial effect of manual therapy over control. Exp ¼ experimental; Mani ¼ manipulation; Physio ¼ physiotherapy; AROM ¼ active range of motion.

C.-Y.C. Ho et al. / Manual Therapy 14 (2009) 463–474

placebo effects. However, the authors (Van den Dolder and Roberts, 2003) proposed that the decrease in pain with the massage was greater than what was previously considered to be decrease of pain as a result of placebo effects of treatment (Hrobjartsson and Gotzsche, 2001). A qualitative analysis of levels of evidence according to specific criteria van Tulder et al. (2003) was performed to define treatment effectiveness as meta-analysis was inappropriate because of clinical heterogeneity with respect to the interventions and population groups. The average methodological quality of the included studies was defined as high (mean score  6). The most common sources of bias were failure to blind therapists and subjects. It is difficult to administer MT treatment without distinguishing between the treatments. Blinding of patients is also difficult when divergent treatment techniques are compared. Inability to blind patients may change their responses to treatment and may be affected by the expectations of the assessors, thereby potentially producing biases (Trampas and Kitsios, 2006). When the allocation is not concealed, decisions about participant inclusion may be influenced by knowledge of whether or not the patient receives the treatment condition, potentially producing systematic bias (Trampas and Kitsios, 2006). Lack of analysis of intention-to-treat was another common problem of the included studies, thus potentially biasing results. In summary, for patients with AC, MT was not more effective than other rehabilitative interventions in the short term for decreasing pain, improving ROM and function. However, there was moderate evidence that HG MT was more effective than LG MT for improving ROM and function in the long-term. For patients with SIS, evidence was conflicting for use of MT for decreasing pain and improving function in the short term, with moderate evidence that MT was no more effective for improving ROM in comparison to other interventions in the short term. However, a pragmatic combination of soft tissue and joint mobilization techniques, in addition to therapeutic exercise may be more effective than an exercise programme alone in this group of patients. The evidence was conflicting for MT in the management of unspecific shoulder pain for decreasing pain and improving function in the short term compared to other interventions. There was moderate evidence that MT was no more effective in improving function and decreasing pain in this patient group in the long term. However, massage and MWM techniques were shown to be useful in managing patients with musculoskeletal disorders of the shoulder for short-term outcomes compared to no treatment. Further research of high quality of RCTs with standardized definitions of shoulder diagnosis, clear descriptions of treatment and adequate follow-up periods and sample sizes is recommended. Acknowledgements We acknowledge Mrs Kate Thompson, Medical Library, University of Otago, for advice towards the database search. Appendix I Keywords used for Ovid and Pubmed searches. Phase 1

Phase 2

Phase 3

Phase 4

1. Shoulder

12. Pain

33. Ankle

2. Shoulder fracture

13. Injury

3. Shoulder dislocation 4. Rotator cuff

14. Musculoskeletal disorder 15. or/12–14

17. Musculoskeletal manipulation 18. Spinal manipulation 19. Massage

20. Soft tissue technique

5. Bursitis

16. and/1, 15

36. Stroke or cerebrovascular accident 37. Spinal injury

21. Soft tissue therapy

34. Knee 35. Hip

473

Appendix I (continued) Phase 1 6. Adhesive capsulitis 7. Frozen shoulder 8. Joint instability 9. Sternoclavicular joint 10. Acromioclavicular joint 11. Glenohumeral joint

Phase 2

Phase 3

Phase 4

22. Manual therapy 23. Joint mobilization 24. Spinal mobilization 25. Osteopathic manipulation 26. Chiropractic manipulation

38. Rheumatoid arthritis 39. Hemiplegia

27. Acupressure 28. Traction 29. Physical therapy 30. physiotherapy 31. or/29,30 32. and/28,31

40. Cancer or neoplasm 41. Celebral palsy 42. Reflex sympathetic dystrophy 43. Acupuncture 44. or/33–43 45. or/16, 2–11 46. or/17–27,31,32 47. and/45,46 48. 47 not 44 49. limit 48 to English or German

References Bang MD, Deyle GD. Comparison of supervised exercise with and without manual physical therapy for patients with shoulder impingement syndrome. Journal of Orthopaedic and Sports Physical Therapy 2000;30:126–37. Bergman GJ, Winters JC, Groenier KH, Pool JJ, Meyboom-de Jong B, Postema K, et al. Manipulative therapy in addition to usual medical care for patients with shoulder dysfunction and pain: a randomized, controlled trial. Annals of Internal Medicine 2004;141:432–9. Bialosky JE, Bishop MD, Price DD, Robinson ME, George SZ. The mechanisms of manual therapy in the treatment of musculoskeletal pain: a comprehensive model. Manual Therapy 2009;14(5):531–8. Binder AI, Bulgen DY, Hazleman BL, Roberts S. Frozen shoulder: a long-term prospective study. Annals of Rheumatic Diseases 1984;43:361–4. Brox JI. Regional musculoskeletal conditions: shoulder pain. Best Practice and Research in Clinical Rheumatology 2003;17:33–56. Bulgen DY, Binder AI, Hazleman BL, Dutton J, Roberts S. Frozen shoulder: prospective clinical study with an evaluation of three treatment regimens. Annals of Rheumatic Diseases 1984;43:353–60. Citaker S, Taskiran H, Akdur H, Arabaci UO, Ekici G. Comparison of the mobilization and proprioceptive neuromuscular facilitation methods in the treatment of shoulder impingement syndrome. The Pain Clinic 2005;27:197–202. Clare HA, Adams R, Maher CG. A systematic review of efficacy of McKenzie therapy for spinal pain. Australian Journal of Physiotherapy 2004;50:209–16. Conroy DE, Hayes KW. The effect of joint mobilization as a component of comprehensive treatment for primary shoulder impingement syndrome. Journal of Orthopaedic and Sports Physical Therapy 1998;28:3–14. Cyriax J. Treatment by manipulation, massage and injection. In: Textbook of orthopaedic medicine, Vol. 2. Bailiere Tindall; 1984. Desmeules F, Cote CH, Fremont P. Therapeutic exercise and orthopedic manual therapy for impingement syndrome: a systematic review. Clinical Journal of Sport Medicine 2003;13:176–82. Ejnisman B, Andreoli CV, Soares BG, Fallopa F, Peccin MS, Abdalla RJ, et al. Interventions for tears of the rotator cuff in adults. Cochrane Database Systematic Reviews 2004. CD002758. Faber E, Kuiper JI, Burdorf A, Miedema HS, Verhaar JA. Treatment of impingement syndrome: a systematic review of the effects on functional limitations and return to work. Journal of Occupational Rehabilitation 2006;16:7–25. Gibson K, Growse A, Korda L, Wray E, MacDermid JC. The effectiveness of rehabilitation for nonoperative management of shoulder instability: a systematic review. Journal of Hand Therapy 2004;17:229–42. Grant HJ, Arthur A, Pichora DR. Evaluation of interventions for rotator cuff pathology: a systematic review. Journal of Hand Therapy 2004;17:274–99. Green S, Buchbinder R, Glazier R, Forbes A. Systematic review of randomised controlled trials of interventions for painful shoulder: selection criteria, outcome assessment, and efficacy. British Medical Journal 1998;316: 354–60. Green S, Buchbinder R, Hetrick S. Physiotherapy interventions for shoulder pain. Cochrane Database Systematic Reviews 2003. CD004258. Guler-Uysal F, Kozanoglu E. Comparison of the early response to two methods of rehabilitation in adhesive capsulitis. Swiss Medical Weekly 2004;134:353–8. Harniman E, Carette S, Kennedy C, Beaton D. Extracorporeal shock wave therapy for calcific and noncalcific tendonitis of the rotator cuff: a systematic review. Journal of Hand Therapy 2004;17:132–51. Herbert RD. How to estimate treatment effects from reports of clinical trials. I: continuous outcomes. Australian Journal of Physiotherapy 2000a;46:229–35.

474

C.-Y.C. Ho et al. / Manual Therapy 14 (2009) 463–474

Herbert RD. How to estimate treatment effects from reports of clinical trials. II: dichotomous outcomes. Australian Journal of Physiotherapy 2000b;46: 309–13. Hrobjartsson A, Gotzsche PC. Is the placebo powerless? An analysis of clinical trials comparing placebo with no treatment. New England Journal of Medicine 2001;344:1594–602. Maher CG, Sherrington C, Herbert RD, Moseley AM, Elkins M. Reliability of the PEDro scale for rating quality of randomized controlled trials. Physical Therapy 2003;83:713–21. Maitland GD. Peripheral manipulation. Butterworth Heineman; 1991. Maricar NN, Chok B. A comparison of the effect of manual therapy with exercise therapy and exercise therapy alone for stiff shoulders. Physiotherapy Singapore 1999;2:99–104. Michener LA, Walsworth MK, Burnet EN. Effectiveness of rehabilitation for patients with subacromial impingement syndrome: a systematic review. Journal of Hand Therapy 2004;17:152–64. Nicholson GG. The effects of passive joint mobilization on pain and hypomobility associated with adhesive capsulitis of the shoulder. Journal of Orthopaedic and Sports Physical Therapy 1985;6:238–46. Paris SV. A history of manipulative therapy through the ages and up to the current controversy in the United States. Journal of Manual and Manipulative Therapy 2000;8:66–77. Teys P, Bisset L, Vicenzino B. The initial effects of a Mulligan’s mobilization with movement technique on range of movement and pressure pain threshold in pain-limited shoulders. Manual Therapy 2008;13:37–42. Trampas A, Kitsios A. Exercise and manual therapy for the treatment of impingement syndrome of the shoulder: a systematic review. Physical Therapy Reviews 2006;11:125–42.

van Tulder M, Furlan A, Bombardier C, Bouter L. Updated method guidelines for systematic reviews in the Cochrane Collaboration back review group. Spine 2003;28:1290–9. Van den Dolder PA, Roberts DL. A trial into the effectiveness of soft tissue massage in the treatment of shoulder pain. Australian Journal of Physiotherapy 2003;49:183–8. Van der Heijden GJM, Van der Windt DAW, De Winter AF. Physiotherapy for patients with soft tissue shoulder disorders: a systematic review of randomised clinical trials. British Medical Journal 1997;315:25–30. Vermeulen HM, Rozing PM, Obermann WR, Le Cessie S, Vliet Vlieland TP. Comparison of high-grade and low-grade mobilization techniques in the management of adhesive capsulitis of the shoulder: randomized controlled trial. Physical Therapy 2006;86:355–68. Vernon H, Humphreys K, Hagino C. Chronic mechanical neck pain in adults treated by manual therapy: a systematic review of change scores in randomized clinical trials. Journal of Manipulative and Physiological Therapeutics 2007;30:215–27. Winters JC, Sobel J, Groenier K, Arendzen H, Meyboom-de JB. Comparison of physiotherapy, manipulation, and corticosteroid injection for treating shoulder complaints in general practice: randomised, single blind study. British Medical Journal 1997;314:1320–5. Winters JC, Jorristma W, Groenier KH, Sobel JS, Meyboom-de Jont B, Arendzen HJ. Treatment of shoulder complaints in general practice: long term results of a randomised, single blind study comparing physiotherapy, manipulation, and corticosteroid injection. British Medical Journal 1999;318:1395–6. Woodley BL, Newsham-West RJ, Baxter GD. Chronic tendinopathy: effectiveness of eccentric exercise. British Journal of Sports Medicine 2007;41:188–98.

Manual Therapy 14 (2009) 475–479

Contents lists available at ScienceDirect

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

Original Article

Performance in the cranio-cervical flexion test is altered in elderly subjects Sureeporn Uthaikhup*, Gwendolen Jull Division of Physiotherapy, The University of Queensland, St. Lucia, Queensland 4072, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 April 2008 Received in revised form 11 November 2008 Accepted 2 December 2008

The cranio-cervical flexion test (CCFT) tests the coordination of the deep and superficial cervical flexor muscles during a cranio-cervical flexion task. The test has revealed impairments in muscle function in younger/middle aged patients with various neck pain disorders. Neck pain and headache are common in elders but it is unknown if age alone affects performance in the CCFT. This study compared performance in the CCFT between healthy asymptomatic elderly and younger subjects. Electromyographic (EMG) amplitude in the sternocleidomastoid (SCM), angle of cranio-cervical flexion and ability to target the pressure levels of each test stage were examined in 44 elderly and 39 young participants. The results indicated that the elderly group had higher measures of normalized EMG signal amplitude in the SCM during the test (p < 0.001), greater shortfalls from the target pressures of all stages of the test (p < 0.01), except for the 22 mm Hg stage (p ¼ 0.13), and larger variability of the cranio-cervical flexion range of motion for the five successive stages of the test (particularly at 26, 28 and 30 mm Hg stages) compared to young subjects. Clinicians must be aware of this occurrence when assessing performance in the CCFT in elders with neck pain. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Cranio-cervical flexion test Electromyography Sternocleidomastoid Elderly

The cranio-cervical flexion test (CCFT) assesses the coordination of the deep and superficial cervical flexor muscles during the staged performance of cranio-cervical flexion under low load conditions in a recumbent supine position. Subjects are guided to the five stages of the test with feedback from an air-filled pressure sensor which is positioned behind the neck (Jull et al., 2008). Studies which have compared performance between patients with neck pain disorders and asymptomatic subjects have revealed increased EMG amplitude in the superficial flexors, sternocleidomastoid (SCM) (Jull, 2000; Jull et al., 2004) and anterior scalene (AS) (Falla et al., 2004) muscles in neck pain patients. This increased EMG activity is associated with reduced activation of the deep cervical flexors, the longus capitis and its synergist the longus colli (Falla et al., 2004). This change in motor strategy is accompanied by a change in the pattern of movement. In asymptomatic subjects, it has been shown that there is a linear relationship between the increasing pressure targets of CCFT and the range of cranio-cervical flexion used in each test stage (Falla et al., 2003a). However neck pain patients have been shown to use less range of cranio-cervical flexion motion to perform the task (Falla et al., 2004). The recognition of the dysfunction in the cranio-cervical flexors in patients with various neck disorders (Falla et al., 2004; Jull et al., 2004; Sterling et al., 2004; Zito et al., 2005) has assisted in the diagnosis of neck related

* Corresponding author. Tel.: þ61 7 3365 2275; fax: þ61 7 3365 1622. E-mail address: [email protected] (S. Uthaikhup). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.12.003

disorders such as cervicogenic headache (Amiri et al., 2007; Jull et al., 2007a). It has also led to the inclusion of specific training for these muscles in therapeutic exercise regimes (Jull et al., 2008) and the exercise has been shown to be an effective management strategy (Jull et al., 2002, 2007b). Neck pain and headache are common complaints in elders (Hartvigsen et al., 2006; Haan et al., 2007; Kaniecki, 2007; Manchikanti et al., 2008) and the low load nature of the exercise regime would appear to have potential for management for this age group. However baseline performances in the CCFT on which to judge dysfunction have been derived principally from younger to middle aged asymptomatic populations (Falla et al., 2003a, b; Jull et al., 2004; Cagnie et al., 2008). It is unknown if such data are applicable to the more elderly population. There is evidence to suggest that there could be differences in performance. For example changes in neuromuscular function in the back muscles have been demonstrated in elders compared to younger subjects (Brown et al., 1994) and there is evidence of age-related changes in neuromuscular morphology (Akataki et al., 2002; Vandervoort, 2002; Faulkner et al., 2007; Kim et al., 2007). The performance of the CCFT which requires precision may also be influenced by cognitive factors, learning and motor skill acquisition in the elderly (O’Sullivan et al., 2001; Mattay et al., 2002; Cabeza et al., 2004). The purpose of this study was to determine if older age influences performance in the CCFT by comparing test outcomes between healthy young and elderly groups without neck pain. Three variables were measured, EMG amplitude in the SCM, angle

476

S. Uthaikhup, G. Jull / Manual Therapy 14 (2009) 475–479

of cranio-cervical flexion and ability to target the pressure levels at each of the five stages of the test. 1. Methods 1.1. Subjects Participants in this study included 39 healthy young adults (11 men, 28 women; mean age 26.7  4.1 yrs, range 18–35 yrs) and 44 healthy older adults (14 men, 30 women; mean age 66.4  4.1 yrs, range 60–75 yrs). These age ranges were chosen to have a clear distinction in age between the groups and to ensure that there was little chance of age changes influencing performances in the younger group. All subjects were recruited by advertising in the community and through a university Centre of Ageing. Subjects were eligible if they had no history of neck pain disorders or headache for which they had sought management from a health professional. Ethical approval for the study was gained from the institutional Medical Ethics Committee and informed consent was obtained from all subjects prior to participation. 1.2. Instrumentation and measurements Surface electromyography (EMG) was used to measure the activity of the sternocleidomastoid (SCM) muscles bilaterally during the CCFT (Jull, 2000). In respect to the elders in this study, no attempt was made to measure activity in the deep cervical flexors due to the invasive nature of the nasopharyngeal electrode required for this measurement (Falla et al., 2003b). Ag/AgCl surface electrodes (11 mm-disc, 3 mm-diameter) (Grass Telefactor, Astro-Med Inc., USA) were affixed over the lower one third of the SCM muscles bilaterally (Falla et al., 2002). A universal electrosurgical pad (3 M Health Care, USA) was used as a ground electrode and placed over the upper thoracic spine. The EMG signals were amplified (gain, 1 mV) and sampled at 2 kHz (8 channel Bio Amp, ADInstruments Pty Ltd, Australia). The low pass filter was set at 1 kHz. The CCFT consists of five stages of increasing cranio-cervical flexion range of motion (ROM) (Falla et al., 2003a; Jull et al., 2004). An inflatable air-filled pressure sensor (Stabilizer, Chattanooga Group Inc, USA) was placed behind the subject’s neck to guide the subject to each test stage. It was pre-inflated to a stable baseline pressure of 20 mm Hg. During the CCFT, the subject was required to perform cranio-cervical flexion to target five progressive increases in pressure of 2 mm Hg from the baseline pressure of 20– 30 mm Hg. Recording of pressure measure was obtained by connecting the air-filled pressure sensor to a pressure transducer. Electrical signals from the pressure transducer were amplified and relayed to the computer and a visual feedback device. The visual feedback device consisted of an electronic voltmeter, marked in 2 mm Hg increments from 20 to 30 mm Hg. It was calibrated to display the pressure in the air-filled pressure sensor, based on the pressure transducer output. A digital web camera (Logitech Image Studio 7.3, Australia) was positioned on a tripod parallel to the ground. The focus of the camera was at the level of the subject’s tragus at a distance of 60 cm. Anatomical markers were positioned on the right tragus, the lateral corner of the right eye and at a point 7 cm caudad to the right mastoid process. A biomechanics analysis software program (UQ, Australia) was used to analyse the cranio-cervical flexion range achieved during each stage of the test. 1.3. Procedure Subjects were positioned in supine crook lying with the head and neck in a mid position, defined as the forehead and chin being

aligned parallel to the plinth. Folded layers of towel were used as required to achieve the position. Practice was provided for familiarization with the CCFT. The subject’s skin over the SCM and the upper thoracic spine was prepared by cleaning with abrasive skin prepping gel and an alcohol swab. The surface EMG electrodes were then attached. Markers were placed on the anatomical points for the photographic measures and the pressure sensor was placed suboccipitally behind the subjects’ neck and inflated to 20 mm Hg. The subjects then performed the CCFT to reach the five sequential target pressures. Each pressure target was held for 10 s with an interval of 10 s between each stage of the test. The subjects were then asked to perform a head lift by tucking in their chin and lifting the head just off the bed and maintained for 10 s as a standardized reference contraction for normalization of the EMG amplitude. EMG recordings were made for each task. Photographs were taken in the starting position, at each stage of the CCFT and, at completion of the testing, at the full range of cranio-cervical flexion in the supine position. For the measurement of full range of cranio-cervical flexion, the subject was requested to nod their chin to the limit of available range and the researcher assisted the action with gentle manual guidance to the head to ensure that end of range was reached. 1.4. Data management The EMG amplitudes over the 10 s recording of each stage of the CCFT can be variable, especially at the beginning and conclusion of the test stage. Thus, for the EMG data, the maximum root mean square (RMS) was calculated for each stage of the CCFT using 1-s overlapping sliding window (MATLAB version7.0, Mathworks Inc., USA). A time window was shifted over a stable 5 s EMG data, selected from the 10 s EMG recording period. The RMS values were then calculated within each 100 ms window. The RMS values integrated across the different time windows were again computed using an equivalent 1 s (20 samples) overlapping sliding window and a total of 81 RMS data points were obtained. The highest RMS value obtained over the overlapping sliding windows was ultimately considered as the maximum RMS. The maximum RMS value obtained for each stage was then normalized against the maximum RMS obtained using the same procedure during a standardized head lift. For these calculations, the baseline EMG RMS was subtracted from the maximum RMS obtained during each stage of the CCFT and from the maximum RMS obtained during the head lift. The data for the left and right SCMs were averaged for analysis as there were no differences determined between sides for both groups (both p > 0.05). The mean pressure that subjects obtained over the 10 s holding time at each target of the CCFT was computed by measuring the pressure for the same 5 s from which EMG data were derived. The difference between the mean pressure obtained and the nominated target pressure for each stage was calculated for each group. Total cranio-cervical flexion range was computed by subtracting the angle measured at the full head nod position from the angle of the starting position. The difference between the angle obtained at each stage of the CCFT and the angle of the starting position was calculated and then expressed as a percentage of the full cranio-cervical ROM. 2. Statistical analysis A mixed model ANOVA was used to analyse within and between group differences in the EMG RMS activity in the SCM and in the relative amount of cranio-cervical flexion ROM used in each stage of the test. An independent t-test was used to determine between group differences in the EMG RMS activity and ROM at each pressure stage of the CCFT. A test for linearity between EMG activity and

S. Uthaikhup, G. Jull / Manual Therapy 14 (2009) 475–479

477

ROM with each stage of the CCFT for each group was also conducted using a 2-sided test. An independent t-test was used to determine between group differences in any shortfall in pressure from the nominated target pressure. Data are expressed as mean  standard error (SE). Significance was defined as p < 0.01. All statistical analyses were conducted using SPSS statistical package (version 15.0). 3. Results The results of the ANOVA revealed significant differences in the SCM normalized RMS values between groups (p < 0.001) and stages of the CCFT (p < 0.001). There was an interaction between the group and stage of the CCFT (p < 0.001). The independent t-test revealed that the elderly group had significantly higher SCM normalized RMS values than the young group at all test stages (p < 0.001) with the exception of the 30 mm Hg stage (p ¼ 0.25). The linear analysis revealed a positive linear relationship between the RMS normalized values of the SCM muscle and stages of the CCF test for both groups (both p < 0.001), with an exception at the 30 mm Hg stage for the elderly group (deviation from linearity p ¼ 0.002) (Fig. 1). Fig. 2 presents the differences between the target pressure and mean pressure obtained for each group at each stage of the CCFT. The results of the independent t-test showed that the mean shortfalls in pressure at all stages were significantly greater in the elderly compared to younger group (p < 0.01), except for the 22 mm Hg stage (p ¼ 0.13). The mean values for full range of the cranio-cervical flexion were 14.3  5.2 and 13.8  4.1 in the young and elderly groups, respectively. The analysis of variance revealed a significant difference in the cranio-cervical flexion range between each incremental stage of the CCFT (p < 0.001) but there was no difference between the groups (p ¼ 0.27). There were no interactions between the cranio-cervical flexion ROM and group (p ¼ 0.99). From the analysis of linearity, there was a strong positive linear relationship between the cranio-cervical flexion ROM and successive stages of the CCFT for both groups (both p < 0.001) (Fig. 3). 4. Discussion The results of this study demonstrated a linear relationship between the magnitude of the superficial muscle activity and the range of the cranio-cervical flexion ROM used in the five

Fig. 1. Normalized RMS values (mean and SE) for the SCM in each stage of the CCFT for the young and elderly groups. *Significant difference between the groups, p < 0.001.

Fig. 2. Shortfall in pressure from the target pressure (mean and SE) for each stage of the CCFT for the young and elderly groups. *Significant difference between the groups, p < 0.01.

incremental stages of the CCFT, which is in accordance with the findings of previous studies (Jull, 2000; Falla et al., 2003a, b; Jull et al., 2004). Nevertheless, healthy elderly subjects displayed significantly higher EMG RMS activity in the SCM muscles compared to healthy younger subjects and were less able to reach the target pressures of the stages of the CCFT, indicating that older age does influence performance in the CCFT. The lesser activity in the SCM, coupled with the pressure shortfall at the 30 mm Hg stage of the CCFT, suggests that many elders could not perform this final stage of the test. The higher levels of EMG RMS activity in the SCM muscles demonstrated by our elders in the CCFT may reflect the effects of the aging process on the neuromuscular system. Age changes have been observed at the level of the muscle spindle (Swash and Fox, 1972; Kim et al., 2007). There is high muscle spindle density in the deep neck muscles (Liu et al., 2003) which is important for movement detection (Proske et al., 2000). Nevertheless, Boyd-Clark et al. (2002) observed no change with age in spindle distribution and density in the longus colli and multifidus muscles at C5-7

Fig. 3. Percentage relative amount of the cranio-cervical flexion range of motion (mean and 95% confidence intervals) obtained for each stage of the CCFT for the young and elderly groups.

478

S. Uthaikhup, G. Jull / Manual Therapy 14 (2009) 475–479

segments. However differences in the neuromotor control have been shown in other studies of elders. For example, Laursen et al. (2001) studied performance and muscle activity during computer mouse tasks and, similar to our observations, found that elderly compared to young adults had higher levels of EMG activity in forearm, shoulder and neck muscles during the task. There is evidence, with increasing age, of a decrease in number and discharge of motor units, less force steadiness and different compositions of muscle fibre types (Vandervoort, 2002; Tracy et al., 2005) which might be reflected in our findings of elder’s performance in the CCFT. Additionally, the strength of anti-gravity muscles has been shown to be lower in elderly compared to young persons (Takeuchi et al., 2007). The deep cervical flexor muscles have an important anti-gravity role in support of the cervical posture and segments (Mayoux-Benhamou et al., 1994). It is possible that the increased activity in the SCM muscles in the CCFT in our elders reflected declining function of the deep cervical flexors in this age group, akin to the changed pattern of muscle activity between the deep and superficial flexors measured in younger populations with neck pain (Falla et al., 2004). Even though a linear relationship was evident between the progressive stages of the CCFT and the range of cranio-cervical flexion used for each test stage, there was a large variability in the range used by the elderly group (as evident in the larger SE at the 26, 28 and 30 mm Hg test stages for this group). Often they could not reach the target pressures (pressure shortfalls). The variability in elders may reflect changes in the biomechanical response of the cervical segments associated with cervical degeneration (Kumaresan et al., 2001) and the loss of segmental mobility with age (ten Have and Eulderink, 1981). The variability may also have reflected age-related changes in central processing. Motor planning and learning are less efficient in elderly subjects (Wishart and Lee, 1997; Labyt et al., 2004; Seidler, 2007) with a decline in coarse and fine motor performance with increasing age (Smith et al., 1999). The CCFT requires both fine motor and cognitive skills which might be challenging for an elderly population. This could have encouraged some elders to use more gross motor strategies during the test, for example head retraction, which might account for the variability in range and pressure targeting measured as well as the increased EMG activity found in this study. Anxiety is another factor to consider in relation to the differences in the CCFT results between the younger and older groups. Relationships have been found between higher levels of anxiety, poorer motor performance and increased EMG activity (Weinberg and Hunt, 1976; Waersted et al., 1994) as well as a greater decline in performance in the elderly compared to young adults when anxiety levels are higher (Backman and Molander, 1991). The CCFT was designed to place a high demand on motor control but not on mental stress. Hence, a measure of an association between anxiety and the CCFT was not considered in this study. However, it was noted that performance of the CCFT was challenging in the elderly subjects and they often provided feedback on the difficulty of the CCFT task. Thus some level of anxiety may have contributed to the increased SCM muscle activity observed in our elderly subjects. 5. Conclusion This study has determined that healthy elders without neck pain use higher levels of SCM activity in the CCFT compared to their younger healthy counterparts. Clinicians treating elders with neck pain and using the CCFT in assessment must be aware of this occurrence. Further research is required to better understand the mechanisms underlying the alteration of cervical flexor muscles performance in the CCFT in the elderly population and it is essential to conduct research to compare elders with and without neck pain

disorders so that correct interpretations of the CCFT can be made in the clinical setting. References Akataki K, Mita K, Watakabe M, Ito K. Age-related change in motor unit activation strategy in force production: a mechanomyographic investigation. Muscle and Nerve 2002;25(4):505–12. Amiri M, Jull G, Bullock-Saxton J, Darnell R, Lander C. Cervical musculoskeletal impairment in frequent intermittent headache. Part 2: subjects with concurrent headache types. Cephalalgia 2007;27(8):891–8. Backman L, Molander B. On the generalizability of the age-related decline in coping with high-arousal conditions in a precision sport: replication and extension. Journal of Gerontology 1991;46(2):79–81. Boyd-Clark LC, Briggs CA, Galea MP. Muscle spindle distribution, morphology, and density in longus colli and multifidus muscles of the cervical spine. Spine 2002;27(7):694–701. Brown JM, Mills JH, Baker A. Neuromuscular control of lifting in the elderly. Gerontology 1994;40(6):298–306. Cabeza R, Daselaar SM, Dolcos F, Prince SE, Budde M, Nyberg L. Task-independent and task-specific age effects on brain activity during working memory, visual attention and episodic retrieval. Cerebral Cortex 2004;14(4):364–75. Cagnie B, Danneels L, Cools A, Dickx N, Cambier D. The influence of breathing type, expiration and cervical posture on the performance of the cranio-cervical flexion test in healthy subjects. Manual Therapy 2008;13(3):232–8. Falla D, Dall’Alba P, Rainoldi A, Merletti R, Jull G. Location of innervation zones of sternocleidomastoid and scalene muscles-a basis for clinical and research electromyography applications. Clinical Neurophysiology 2002;113(1):57–63. Falla D, Campbell CD, Fagan AE, Thompson DC, Jull GA. Relationship between cranio-cervical flexion range of motion and pressure change during the craniocervical flexion test. Manual Therapy 2003a;8(2):92–6. Falla D, Jull G, Dall’Alba P, Rainoldi A, Merletti R. An electromyographic analysis of the deep cervical flexor muscles in performance of craniocervical flexion. Physical Therapy 2003b;83(10):899–906. Falla D, Jull GA, Hodges PW. Patients with neck pain demonstrate reduced electromyographic activity of the deep cervical flexor muscles during performance of the craniocervical flexion test. Spine 2004;29(19):2108–14. Faulkner JA, Larkin LM, Claflin DR, Brooks SV. Age-related changes in the structure and function of skeletal muscles. Clinical and Experimental Pharmacology and Physiology 2007;34(11):1091–6. Haan J, Hollander J, Ferrari M. Migraine in the elderly: a review. Cephalalgia 2007; 27(2):97–106. Hartvigsen J, Frederiksen H, Christensen K. Back and neck pain in seniorsprevalence and impact. European Spine Journal 2006;15(6):802–6. ten Have HA, Eulderink F. Mobility and degenerative changes of the ageing cervical spine: a macroscopic and statistical study. Gerontology 1981;27(1–2):42–50. Jull G. Deep cervical flexor muscle dysfunction in whiplash. Journal of Musculoskeletal Pain 2000;8(1–2):143–54. Jull G, Trott P, Potter H, Zito G, Niere K, Shirley D, et al. A randomized controlled trial of exercise and manipulative therapy for cervicogenic headache. Spine 2002; 27(17):1835–43. Jull G, Kristjansson E, Dall’Alba P. Impairment in the cervical flexors: a comparison of whiplash and insidious onset neck pain patients. Manual Therapy 2004;9(2): 89–94. Jull G, Amiri M, Bullock-Saxton J, Darnell R, Lander C. Cervical musculoskeletal impairment in frequent intermittent headache. Part 1: subjects with single headaches. Cephalalgia 2007a;27(7):793–802. Jull G, Sterling M, Kenardy J, Beller E. Does the presence of sensory hypersensitivity influence outcomes of physical rehabilitation for chronic whiplash? – A preliminary RCT. Pain 2007b;129(1–2):28–34. Jull G, Sterling M, Falla D, Treleaven J, O’Leary S. Whiplash, headache, and neck pain. Edinburgh: Churchill Livingstone; 2008. Kaniecki RG. Tension-type headache in the elderly. Current Treatment Options in Neurology 2007;9(1):31–7. Kim GH, Suzuki S, Kanda K. Age-related physiological and morphological changes of muscle spindles in rats. The Journal of Physiology 2007;582(2):525–38. Kumaresan S, Yoganandan N, Pintar FA, Maiman DJ, Goel VK. Contribution of disc degeneration to osteophyte formation in the cervical spine: a biomechanical investigation. Journal of Orthopaedic Research 2001;19(5):977–84. Labyt E, Szurhaj W, Bourriez JL, Cassim F, Defebvre L, Destee A, et al. Influence of aging on cortical activity associated with a visuo-motor task. Neurobiology of Aging 2004;25(6):817–27. Laursen B, Jensen BR, Ratkevicius A. Performance and muscle activity during computer mouse tasks in young and elderly adults. European Journal of Applied Physiology 2001;84(4):329–36. Liu JX, Thornell LE, Pedrosa-Domellof F. Muscle spindles in the deep muscles of the human neck: a morphological and immunocytochemical study. Journal of Histochemistry and Cytochemistry 2003;51(2):175–86. Manchikanti L, Manchikanti KN, Cash KA, Singh V, Giordano J. Age-related prevalence of facet-joint involvement in chronic neck and low back pain. Pain Physician 2008;11(1):67–75. Mattay VS, Fera F, Tessitore A, Hariri AR, Das S, Callicott JH, et al. Neurophysiological correlates of age-related changes in human motor function. Neurology 2002;58(4): 630–5.

S. Uthaikhup, G. Jull / Manual Therapy 14 (2009) 475–479 Mayoux-Benhamou MA, Revel M, Vallee C, Roudier R, Barbet JP, Bargy F. Longus colli has a postural function on cervical curvature. Surgical and Radiologic Anatomy 1994;16(4):367–71. O’Sullivan M, Jones DK, Summers PE, Morris RG, Williams SC, Markus HS. Evidence for cortical ‘‘disconnection’’ as a mechanism of age-related cognitive decline. Neurology 2001;57(4):632–8. Proske U, Wise AK, Gregory JE. The role of muscle receptors in the detection of movements. Progress in Neurobiology 2000;60(1):85–96. Seidler RD. Aging affects motor learning but not savings at transfer of learning. Learning and Memory 2007;14(1):17–21. Smith CD, Umberger GH, Manning EL, Slevin JT, Wekstein DR, Schmitt FA, et al. Critical decline in fine motor hand movements in human aging. Neurology 1999;53(7):1458–61. Sterling M, Jull G, Vicenzino B, Kenardy J. Characterization of acute whiplashassociated disorders. Spine 2004;29(2):182–8. Swash M, Fox KP. The effect of age on human skeletal muscle: studies of the morphology and innervation of muscle spindles. Journal of the Neurological Sciences 1972;16(4):417–32.

479

Takeuchi Y, Tanaka Y, Shimomura Y, Iwanaga K, Katsuura T. The effect of aging on the backward stepping reaction as estimated from the velocity of center of foot pressure and muscular strength. Journal of Physiological Anthropology 2007;26(2):185–9. Tracy BL, Maluf KS, Stephenson JL, Hunter SK, Enoka RM. Variability of motor unit discharge and force fluctuations across a range of muscle forces in older adults. Muscle and Nerve 2005;32(4):533–40. Vandervoort AA. Aging of the human neuromuscular system. Muscle and Nerve 2002;25(1):17–25. Waersted M, Bjorklund RA, Westgaard RH. The effect of motivation on shouldermuscle tension in attention-demanding tasks. Ergonomics 1994;37(2): 363–76. Weinberg RS, Hunt VV. The interrelationships between anxiety, motor performance and electromyography. Journal of Motor Behavior 1976;8(3):219–24. Wishart LR, Lee TD. Effects of aging and reduced relative frequency of knowledge of results on learning a motor skill. Perceptual and Motor Skills 1997;84: 1107–22. Zito G, Jull G, Story I. Clinical tests of musculoskeletal dysfunction in the diagnosis of cervicogenic headache. Manual Therapy 2005;11(2):118–29.

Manual Therapy 14 (2009) 480–483

Contents lists available at ScienceDirect

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

Original article

The relation between the application angle of spinal manipulative therapy (SMT) and resultant vertebral accelerations in an in situ porcine model Gregory N. Kawchuk a, *, Stephen M. Perle b a b

University of Alberta, Faculty of Rehabilitation Medicine, 3-44 Corbett Hall, Edmonton, Alberta, Canada T6G 2G4 University of Bridgeport College of Chiropractic, Bridgeport, CT 06604, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 13 February 2008 Received in revised form 21 October 2008 Accepted 8 November 2008

It has been hypothesized that the posterior tissues of the spine are frictionless and therefore allow only the normal force component of spinal manipulative therapy (SMT) to pass to underlying vertebrae. Given this assumption, vertebrae could not be moved in practitioner-defined directions by altering the application angle of SMT. To investigate this possibility, porcine lumbar spines were excised and then SMT applied at 90 to the posterior tissues of the target vertebra. A standard curve was constructed of increasing SMT force versus vertebral acceleration. SMT forces were then applied at 60 and 120 and the resulting accelerations substituted into the standard curve to obtain the transmitted force. Results showed that vertebral accelerations were greatest at a 90 SMT application angle and decreased in all axes at application angles s 90 . The average decrease in transmitted force using application angles of 60 and 120 was within 5% of the predicted absolute value. In this model, SMT applied at a nonnormal angle does not increase vertebral acceleration in that same direction, but acts to reduce transmitted force. This work provides justification for future studies in less available human cadavers. It is not yet known if variations in SMT application angle have relevance to clinical outcomes or patient safety. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Spinal manipulation Acceleration Force

1. Background and purpose Spinal manipulative therapy (SMT) is a therapeutic intervention which can be beneficial for low back and neck pain (Bronfort et al., 2004). SMT involves the application of a high velocity, low amplitude force to a target tissue of the musculoskeletal system. Historically, SMT is applied at a specific angle with the intention that the underlying vertebrae will be moved in that same direction. Most often, the desired direction is described to be parallel to the articular space of the zygapophyseal joints so that the vertebral displacement response is maximized (Edmond, 1993; Gibbons and Tehan, 2000; Isaacs and Bookhout, 2002; Peterson, 2002). This rationale, and the clinical importance attached to applying SMT at the ‘‘correct’’ angle, appear to be universal regardless of the provider’s profession (physical therapy, (Edmond, 1993) chiropractic, (Peterson, 2002; Esposito, 2005) medicine, (Isaacs and Bookhout, 2002) or osteopathy (Gibbons and Tehan, 2000)). While described for decades, the idea that SMT can create vertebral accelerations in practitioner-defined directions has been challenged recently with the following hypothesis: frictionless interfaces between the posterior spinal tissues (no matter their

* Corresponding author. Tel.: þ1 780 492 6891; fax: þ1 780 492 4429. E-mail address: [email protected] (G.N. Kawchuk). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.11.001

orientation) allow only the normal component of SMT forces to pass to underlying spinal structures (i.e., perpendicular to the skin surface) (Bereznick et al., 2002). Given this assumption, only the component of SMT force perpendicular to the underlying tissue planes would act on the underlying spinal structures. If this were the case, vertebrae could not be moved in practitioner-defined directions by altering the application angle of SMT. Instead, the only effect of altering the application angle of SMT away from a normal orientation would be a decrease in the overall vertebral acceleration response. While many studies have used in situ and in vivo approaches to record human vertebral motions resulting from manipulation (Nathan and Keller, 1994; Gal et al., 1995; Gal et al., 1997b; Gal et al., 1997a; Colloca and Fuhr, 1998; Keller et al., 2000, 2003, 2006a; Maigne and Guillon, 2000; Colloca et al., 2004), none have provided data that support or refute this recent hypothesis. To address this deficiency, we have chosen to use a porcine model as it is 1) inexpensive, 2) readily available and 3) can be developed for use as an in vivo model for future studies that explore the relation between parameters of manipulation delivery and various outcomes which cannot be quantified in humans. While we acknowledge that there are distinct differences between the posterior spinal tissues in humans and pigs, the intent of this study is not to extrapolate these results to humans, but to establish if the hypothesis underlying this project is tenable.

G.N. Kawchuk, S.M. Perle / Manual Therapy 14 (2009) 480–483

Given the above, the objective of this study was to determine the relation between the angle of applied SMT and the acceleration response of the underlying target vertebra. It was hypothesized that SMT forces applied at non-normal angle would decrease resultant vertebral accelerations by a predictable magnitude. 2. Methods 2.1. Spine preparation All procedures were approved by the Health Sciences Animal Policy and Welfare Committee of the University of Alberta. Using a reciprocating saw, lumbar spines were removed en bloc from three pigs immediately after euthanasia. All spinal structures between and including L1–L5 were removed intact and all posterior soft tissues were preserved. Spines were then frozen at 20  C and thawed for a minimum of 48 h before experimental use. When thawed, soft tissues and posterior bony elements were removed to expose only the vertebral bodies of the two terminal vertebrae (L1 and L5) while preserving the remaining structures. Using a fluoroscope, the skin surface immediately superficial to the spinous process of the target vertebrae (L3 and L4) was identified and marked. The vertebral bodies of the two terminal (end) vertebrae were then affixed by screws into a frame which suspended the spine between two rigid supports (Fig. 1). Care was taken to mount the spine in its neutral orientation. A small section of the anterior vertebral bodies of L3 and L4 were ground to a flat surface and a plastic block was then glued to the ground anterior surface of each vertebra as an accelerometer attachment point. 2.2. Equipment and data collection Impulse loading to the skin surface of the specimen was applied by a commercially available instrument which was driven by compressed air (Activator Methods International, Phoenix, AZ). This instrument, or ones that are similar, are used by up to 70% of chiropractic practitioners to apply SMT (although not exclusive) (Christensen and Kollasch, 2005). A valve on the device allows for continuous adjustment of input pressure between a minimal and maximal setting (42 psi). The force control valve was divided into 12 increments for standard curve creation (see below). The impulse loading device was modified to accept a load cell (Measurement Specialties, Hampton, VA) within the contact stylus.

481

The accelerometer used in this protocol was fabricated in-house. Specifically, a 0.35 gm accelerometer chip (Measurement Specialties, Hampton, VA) was soldered to a custom integrated circuit board having additional on-board filtering and power management. For this experiment, the sensitivity of the device was set to 199.8 mv/g with a range of 2.5 v. Signals from the accelerometer and load cell were collected on two separate channels by a 16 bit A/D collection system (National Instruments, Austin, TX) at a rate of 20 kHz. Peak force and peak accelerations were considered to be the difference between the initiation of impulse loading and the first inflection of signal. 2.3. Testing For each target vertebra, the accelerometer was affixed by double sided tape to the plastic block glued to the vertebra of interest and aligned so that its x, y and z axes corresponded to the horizontal (superioinferior), horizontal (mediolateral) and vertical (posterioanterior) axes, respectively (Fig. 1). SMT forces were then applied with the impulse loading device to the marked surface of the skin associated with the test vertebra. Specifically, the loading device was maintained in a vertical orientation (90 ) using a guide. As is the case in clinical practice, the device was then depressed into the skin manually until compression of a preload indicator spring inside the device was achieved. The device was then fired three times for every one of 12 force settings. For each force setting, the three applied loads were averaged as was the resultant vertebral accelerations in each axis. From these data, standard curves of the increasing SMT load versus the z-axis vertebral accelerations were plotted, described by a linear regression equation, and the coefficient of correlation calculated. Remaining at the skin contact site employed during standard curve creation, the SMT instrument was then placed in a 60 orientation against a guide and fired three times at the tenth force increment of the device while recording the accelerometer and load cell responses. Because the instrument contacts increase less tissue with increasing angulation, we chose to restrict testing to the angle of greatest instrument/skin contact area (60 ). This process was then repeated in each test vertebra for an applied load oriented at 120 . In this way, the instrument/skin contact angle remained the same, but SMT forces were provided in an opposite direction. For each test vertebra, at each angle, the three applied loads were averaged (Fzapplied) as were the resultant vertebral accelerations in each axis (Ax, Ay, Az). The normal component of the applied force was calculated for each test vertebra (Fzpredicted) by multiplying the average applied load for each test vertebra, at each application angle, by the sine of the application angle. For each angle, Az was then entered into the linear equation of the standardized curve of force vs. acceleration for that particular vertebra. Solving this equation resulted in Fztransmitted: the magnitude of the normal force component needed to cause vertebral accelerations of magnitude Az. 3. Results

Fig. 1. Schematic representation of the experimental preparation, SMT application angles and coordinate system for recording of three dimensional accelerations.

In total, six test vertebrae from three animals were studied (2 per animal). Data are summarized in Table 1. Force applications at 60 and 120 were observed to result in reduced vertebral accelerations in all axes (Ax, Ay, and Az) when compared to accelerations obtained from forces applied in a normal orientation (90 ). In addition, at any angle of SMT application, absolute z-axis accelerations (Az) were always greater in magnitude than those found simultaneously in either the x-plane or y-plane (Ax and Ay). The correlation coefficients for standard curves of increasing force at 90 vs. z-plane vertebral acceleration (one curve for each test vertebra) were as follows: Pig 1/L3 ¼ 0.95, Pig 1/L4 ¼ 0.99, Pig

482

G.N. Kawchuk, S.M. Perle / Manual Therapy 14 (2009) 480–483

Table 1 Summary of data. All values for each test vertebra (n ¼ 6) are averages from three trials. The applied force (Fzapplied) was measured directly from the instrument used to apply manipulation. The predicted force (Fzpredicted) represents the decrease in force magnitude expected to occur based on the application angle. The transmitted force (Fztransmitted) was obtained by substituting the z-axis acceleration response of the vertebra into a standard curve of force vs. acceleration.

Pig 1

Pig 2

Pig 3

Test

Angle

Fzapplied

Fzpredicted

Fztransmitted

% difference between Fzpredicted and Fztransmitted

Vertebra

Degrees

N

N

N

%

L3 L3 L4 L4 L3 L3 L4 L4 L3 L3 L4 L4

60.00 120.00 60.00 120.00 60.00 120.00 60.00 120.00 60.00 120.00 60.00 120.00

179.72 188.65 177.04 167.78 165.22 167.83 168.66 162.44 167.35 162.44 207.18 176.70

155.64 163.37 153.32 145.30 143.08 145.34 146.06 140.68 144.93 140.68 179.42 153.03

145.86 174.94 147.45 153.54 136.89 143.81 150.94 119.21 152.06 153.34 169.68 161.00

5.21 6.15 2.68 5.40 3.25 0.23 3.14 12.77 6.56 8.35 4.65 6.28

Average StDev Min Max

174.25 13.00 162.44 207.18

150.90 11.25 140.68 179.42

150.73 14.57 119.21 174.94

0.59 6.43 8.35 12.77

2/L3 ¼ .85, Pig 2/L4 ¼ 0.91, Pig 3/L3 ¼ .92, Pig 3/L4 ¼ 0.92. The mean correlation coefficients was 0.92  0.05 (std). Example standardized curve data from one test vertebra are plotted in Fig. 2 (r2 ¼ 0.956). The difference between the predicted normal component of the force applied at 60 and 120 and the force required to produce resultant vertebral acceleration in the z-axis ranged from 8.35% to 12.77% with an overall average of 0.59%  6.43% (std). As absolute values, the average difference was 5.39%  3.16% (std). 4. Discussion This study investigated the effects of SMT application angle on the accelerations of underlying vertebrae. Our results support the

hypothesis that SMT forces applied in non-normal orientations reduce the vertebral acceleration response. Specifically, when SMT forces are applied at non-normal angles, the accelerations of the target vertebra are reduced in all directions while those accelerations in the normal axis remained largest. These results suggest that when applied at a non-normal angle, SMT does not increase vertebral accelerations along that same angle. Further support of this statement comes not only from our observation that vertebral accelerations decrease with non-normal SMT application angles, but from our ability to predict the magnitude of this decrease using simple trigonometry. These data support the original suggestion of Bereznick et al. (2002) that the posterior tissues between the point of SMT application and the target vertebra slide freely between each other. Assuming that only the normal component of SMT force is passed on to underlying vertebrae, what explanation is there for non-zero vertebral accelerations in the two horizontal planes? These non-normal accelerations are to be expected as the target vertebra is not floating freely but is coupled to, and limited by its various hard and soft tissue connections. Much like a rollercoaster whose motion is initiated by gravity then governed by its connections to the track, the impact of SMT provides a normal force to a vertebra whose resulting motion is governed by a combination of neutral zone limits, articular boundaries and soft tissue deformations. We speculate that the prominence of Az over Ax and Ay is the result of normal forces applied to the vertebrae that accelerate it within the normal axis through the neutral zone. Accelerations in the remaining horizontal axes then arise as spinal tissues with various orientations, interconnections and stiffness come into play. It should be noted that in our experiment, motion of the accelerometer as a result of SMT can barely be perceived visually, a result supported by the magnitudes of vertebral movement reported in other studies using SMT applied to the skin by an Activator device. (Keller et al., 2006b) Given this information, we have assumed the rotations experienced by the accelerometer as a result of SMT are minimal. With this assumption, we are able to compare changes of acceleration directly between conditions of pre-manipulation and peak manipulation. Although other forms of SMT may cause significant vertebral rotations, the approach used in this experiment

Fig. 2. Standard curve plot of force versus z-axis acceleration from one test vertebra (r2 ¼ 0.956).

G.N. Kawchuk, S.M. Perle / Manual Therapy 14 (2009) 480–483

allows us to reduce the number of variables toward investigating the underlying phenomenon. Future studies aimed at investigating manipulations causing large-magnitude vertebral rotations will need to employ methodologies that consider the effect of gravitational acceleration changing between accelerometer axes. Because our results were obtained in a very specific setting, we anticipate that various clinical scenarios may be proposed where the premise underlying our results may not apply. First, could preSMT compression of posterior spinal tissues increase tissue friction to a point where non-normal SMT forces are transmitted to underlying vertebrae? While a pre-load may indeed compress tissue, Bereznick et al. (2002) tested this theory by applying mass progressively to the thoracic spine of prone subjects. Their results demonstrated, as expected, that increased mass does not alter the frictionless behavior of the posterior tissues. Second, could increased skin tension before SMT application alter a vertebra’s directional response? Because the skin over the spinous processes is highly mobile, this strategy is used currently to maintain the desired SMT contact point. There is however, no evidence to suggest that this strategy would create a global tensioning of tissues significant enough to alter a vertebra’s acceleration response to SMT. In fact, as SMT is applied, it has been shown that there is further movement in the direction of tissue tensioning which suggests that maximal skin tensioning is not achieved prior to SMT (Herzog et al., 2001). Third, it should be noted that the findings of this study cannot be generalized to all applications of SMT. When SMT is used to create global rotations of large spinal regions, the resulting forces may not be applied directly to surfaces of the spine, but accumulate in different regions due to complex interactions between spinal mechanics and clinician/patient force interfaces (Bereznick et al., 2006). Finally, we recognize that differences between human and porcine anatomy prevent these results from being extrapolated to humans. While our results support the hypothesis that a vertebra’s direction of movement cannot be influenced by varying the angle of SMT application, this work must now be performed in more expensive, less available human cadaveric specimens. The results of this study provide justification for these future studies. 5. Conclusion This paper examines the commonly held belief that a manual therapist is able to influence the direction of movement of an underlying vertebra by varying the angle of application of the applied force. Results from this study suggest that SMT applied at non-normal angles does not increase vertebral accelerations in that same direction but acts to reduce transmitted SMT force. This work provides justification for future studies in more expensive, less available human cadaveric specimens. It is not yet known if variations in SMT application angle (and any resulting change in vertebral accelerations) has relevance to clinical outcomes or patient safety. Acknowledgements The authors would like to acknowledge Edmond Lou, Ph.D. for his expertise in accelerometry. Salary support for Greg Kawchuk

483

was provided by the Canada Research Chair Program. Funding for Stephen Perle to travel to the University of Alberta for data collection, analysis and manuscript preparation was provided by the University of Bridgeport College of Chiropractic. References Bereznick DE, Ross JK, McGill S. Location of applied forces during side posture lumbar manipulation where should forces be applied to produce cavitation? Journal of Chiropractic Education 2006;20(1):2. Bereznick DE, Ross JK, McGill SM. The frictional properties at the thoracic skinfascia interface: implications in spine manipulation. Clinical Biomechanics 2002;17(4):297–303. Bronfort G, Haas M, Evans RL, Bouter LM. Efficacy of spinal manipulation and mobilization for low back pain and neck pain: a systematic review and best evidence synthesis. Spine Journal 2004;4(3):335–56. Christensen MG, Kollasch MW. Professional functions and treatment procedures, job analysis of chiropractic. Greeley, CO: National Board of Chiropractic Examiners; 2005. Colloca CJ, Fuhr AW. Movements of vertebrae during manipulative thrusts to unembalmed human cadavers. Journal of Manipulative and Physiological Therapeutics 1998;21(2):128–9. Colloca CJ, Keller TS, Gunzburg R. Biomechanical and neurophysiological responses to spinal manipulation in patients with lumbar radiculopathy. Journal of Manipulative and Physiological Therapeutics 2004;27(1):1–15. Edmond SL. Manipulation & mobilization. St. Louis: Mosby; 1993 [chapter 10–13]. Esposito S. Mechanics of adjustment specificity. In: Esposito S, Philipson S, editors. Spinal adjustment technique: the chiropractic art. St. Ives, Australia: S Philipson and S Esposito; 2005. p. 97–9 [chapter 7]. Gal J, Herzog W, Kawchuk G, Conway P, Zhang YT. Measurements of vertebral translations using bone pins, surface markers and accelerometers. Clinical Biomechanics (Bristol, Avon) 1997a;12(5):337–40. Gal J, Herzog W, Kawchuk G, Conway PJ, Zhang YT. Movements of vertebrae during manipulative thrusts to unembalmed human cadavers. Journal of Manipulative and Physiological Therapeutics 1997b;20(1):30–40. Gal JM, Herzog W, Kawchuk GN, Conway PJ, Zhang Y-T. Forces and relative vertebral movements during SMT to unembalmed post-rigor human cadavers: peculiarities associated with joint cavitation. Journal of Manipulative and Physiological Therapeutics 1995;18(1):4–9. Gibbons P, Tehan P. Manipulation of the spine, thorax and pelvis: an osteopathic perspective. New York: Churchill Livingstone; 2000. p. 53 [chapter Part B]. Herzog W, Kats M, Symons B. The effective forces transmitted by high-speed, low-amplitude thoracic manipulation. Spine 2001;26(19):2105–10. Isaacs ER, Bookhout MR. Bourdillon’s spinal manipulation. 6th ed. Boston: Butterworth Heinemann; 2002. Keller TS, Colloca CJ, Fuhr AW. In vivo transient vibration assessment of the normal human thoracolumbar spine. Journal of Manipulative and Physiological Therapeutics 2000;23(8):521–30. Keller TS, Colloca CJ, Gunzburg R. Neuromechanical characterization of in vivo lumbar spinal manipulation. Part I. Vertebral motion. Journal of Manipulative and Physiological Therapeutics 2003;26(9):567–78. Keller TS, Colloca CJ, Moore RJ, Gunzburg R, Harrison DE. Increased multiaxial lumbar motion responses during multiple-impulse mechanical force manually assisted spinal manipulation. Chiropractic and Osteopathy 2006a;14:6. Keller TS, Colloca CJ, Moore RJ, Gunzburg R, Harrison DE, Harrison DD. Threedimensional vertebral motions produced by mechanical force spinal manipulation. Journal of Manipulative and Physiological Therapeutics 2006b;29(6):425–36. Maigne JY, Guillon F. Highlighting of intervertebral movements and variations of intradiskal pressure during lumbar spine manipulation: a feasibility study. Journal of Manipulative and Physiological Therapeutics 2000;23(8):531–5. Nathan M, Keller TS. Measurement and analysis of the in vivo posteroanterior impulse response of the human thoracolumbar spine: a feasibility study. Journal of Manipulative and Physiological Therapeutics 1994;17(7): 431–41. Peterson DH. Principles of adjustive technique. In: Peterson DH, Bergmann TF, editors. Chiropractic technique: principles and procedures. New York: Mosby; 2002. p. 140–56 [chapter 4].

Manual Therapy 14 (2009) 484–489

Contents lists available at ScienceDirect

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

Original Article

The influence of increasing sacroiliac joint force closure on the hip and lumbar spine extensor muscle firing pattern Hiroshi Takasaki a, b, *, Takeshi Iizawa b, c, Toby Hall d, Takuo Nakamura e, Shouta Kaneko b a

Graduate School of Health Sciences, Sapporo Medical University, South-1, West-17, Chuo-ku, Sapporo, Hokkaido, Japan Shinoro Orthopedic, 4-5-3-9, Shinoro, Kita-ku, Sapporo, Hokkaido, Japan. c School of Health Sciences, Sapporo Medical University, South-1, West-17, Chuo-ku, Sapporo, Hokkaido, Japan d Adjunct Senior Teaching Fellow, School of Physiotherapy, Curtin University, Perth, Western Australia e Department of Physical Therapy, Sapporo Medical University School of Health Sciences, South-1, West-17, Chuo-ku, Sapporo, Hokkaido, Japan b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 30 May 2008 Received in revised form 14 October 2008 Accepted 8 November 2008

The prone hip extension (PHE) test is commonly used in the evaluation of lumbo-pelvic dysfunction. It has been suggested that altered motor control identified by the PHE test can be improved with the application of compression force across the pelvis, to increase force closure on the sacroiliac joint (SIJ). This repeated measure study design investigated the effect of three levels of pelvis compression (0 N, 50 N, 100 N) on the muscle firing pattern during the PHE test in 20 asymptomatic male subjects tested on two occasions 4-weeks apart. The right gluteus maximus, right semitendinosus and bilateral lumbar erector spinae were analyzed using surface electromyography (EMG). Subjects were instructed to perform right hip extension in prone position while maintaining knee-extension in each measurement condition. Compared with the onset of the semitendinosus muscle, gluteus maximus became active 263.3  99.5 ms later with no pelvic compression, 183.5  77.9 ms later with 50 N compression, 91.5  49.7 ms later with 100 N compression. While significant differences (a ¼ 0.05) were found in EMG onset for gluteus maximus under different levels of pelvis compression, this was not the case for the erector spinae, which had an inconsistent pattern of temporal onset and was not influenced by the level of pelvis compression force. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Muscle firing pattern Sacroilliac joint Gluteus maximus

1. Introduction Functional stability of the pelvis is generated by a combination of ‘force closure’ and ‘form closure’. The term ‘form closure’ was coined by Snijders and Vleeming (Vleeming et al., 1990a, b; Snijders et al., 1993a, b) and is used to describe how the joint’s shape contributes to stability. On the other hand, ‘force closure’ refers to other forces acting across the joint to create stability. According to theoretical modeling of force closure effects (Pel et al., 2008), the application of 50 N medial compression force at the anterior superior iliac spine increases SIJ compression force by 52%. Furthermore, it has been said that the stronger the force closure the more form closure is obtained (Snijders et al., 1993a, b). A number of tests have been developed which are said to evaluate the functional stability and control of the pelvis (Buyruk et al., 1995a, b, 1999; Mens et al., 1999, 2001; Lee and Lee, 2004). The PHE test is one commonly used in the evaluation of lumbopelvic function (Lee and Lee, 2004). It has been theorized that the * Corresponding author. Shinoro Orthopedic, 4-5-3-9, Shinoro, Kita-ku, Sapporo, Hokkaido, Japan. Tel.: þ81 011 772 7255; fax: þ81 011 772 7256. E-mail address: [email protected] (H. Takasaki). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.11.003

pattern of activation of muscles during PHE represents the muscle recruitment pattern of hip extension during gait (Lehman et al., 2004). According to Janda (Janda, 1978), the ideal sequence of muscle activation during the PHE test in temporal order is ipsilateral hamstring, ipsilateral gluteus maximus, and contralateral lumbar erector spinae. It has been theorized that aberration in the temporal recruitment pattern of these muscles decreases the stability of the pelvis during gait and thus hinders the body’s mechanical efficiency. One of the most commonly described patterns of dysfunction seen clinically during PHE is too much delay in the recruitment of the gluteus maximus (Sahrmann, 2002). In this case hip extension is achieved by hamstring muscle activation, this creates compensatory anterior pelvic tilt and thus lumbar hyperlordosis. In addition poor gluteus maximus strength and activation is postulated to decrease the efficiency of gait (Janda, 1992, 1996). Moreover, Sahrmann (Sahrmann, 2002) has suggested that if the hamstrings are dominant, and gluteus maximus is inhibited, abnormal displacement of the greater trochanter can be palpated during the PHE, which is a finding reported in cases of hip pain (Sahrmann, 2002). A number of studies have investigated the temporal pattern of muscle recruitment during PHE (Bullock-Saxton et al., 1994; Vogt

H. Takasaki et al. / Manual Therapy 14 (2009) 484–489

and Banzer, 1997; Lehman et al., 2004). To date, no consistent pattern of activation has been found. However, one common thread through these various studies is the consistent report of delayed activation of the gluteus muscle compared with the hamstrings (Vogt and Banzer, 1997; Lehman et al., 2004). It has been said that if the efficiency of PHE was improved with the application of compression force across the pelvis, this might have some effect on improving force closure and therefore improve the muscle firing pattern (Lee and Lee, 2004). However, to date no studies have investigated the influence of compressive force across the pelvis on the PHE test. The purpose of this study was to determine whether compressive force applied across the pelvis influences the muscle firing patterns of the semitendinosus, gluteus maximus, and erector spinae. We hypothesized that applying compression force across the pelvis would reduce the onset delay for gluteus maximus and semitendinosus while having no influence on the erector spinae muscles.

2. Methods 2.1. Pilot study Prior to the main experiment, a preliminary study was designed to determine the amount of force routinely applied across the pelvis to increase force closure during the PHE. Five physical therapists, each with three years post-graduate experience, were asked to apply three kinds of pressure to five participants (five males, Average age: 22.2 (SD ¼ 0.8) years old) across the pelvis, simulating the PHE test (Fig. 1): normal, strong, and maximum. The therapists were required to apply the pressure on the pelvis bilaterally through hand-held dynamometers (mTas, ANIMA Co. Ltd., Tokyo). The average force under normal pressure was 51.8  11.0 N, under strong pressure 98.2  11.5 N and under maximum pressure

485

Fig. 2. Prone hip extension.

143.2  11.2 N. As a number of subjects complained of pain when compression force of 150 N was applied this level of force was not used in the main study. 2.2. Subjects For this study, 20 males (Right leg dominant, Average height: 172.6 (SD ¼ 5.40) cm, Average weight: 64.3 (SD ¼ 5.2) kg, Average age: 22.0 (SD ¼ 1.3) years old) with no history of lumbar, sacroiliac or lower limb injury within the past year were recruited from undergraduate students in Sapporo medical university. Subjects who had a previous history of lumbar surgery, spondylophathies, or arthritic disorders were excluded. Individuals with past episodes of ankle sprain (grade 2 or 3) were also excluded because BullockSaxton et al. reported that ankle sprain influenced the muscle firing pattern of the gluteus maximus (Bullock-Saxton et al., 1994). 2.3. Instrumentation The activation patterns of the right gluteus maximus, right semitendinosus muscle group and bilateral lower erector spinae were assessed by surface EMG. Surface electrodes (Ag/AgCl) were placed in pairs and parallel to the muscle fibers (Cram et al., 1998). For gluteus maximus, electrodes were placed at mid belly between sacral vertebrae and the greater trochanter. For semitendinosus, electrodes were attached at the mid point between the inferior gluteal fold and knee joint line. For lower erector spinae electrodes were placed longitudinally 2 cm lateral to the L3 spinous process, bilaterally.

Condition A

Condition B

PHE

without pelvic compression ×5

PHE

without pelvic compression ×5

10 people

Rest (2min) 10 people

with 100 N compression

Rest (2min)

PHE

with 50 N compression ×5 Rest (2min)

PHE

with 100 N compression ×5 Rest (2min)

Fig. 1. Procedure used in the pilot study to measure compressive force across the pelvis using hand-held dynamometers.

20 people Rest (2min)

PHE

Rest (2min)

PHE

without pelvic compression ×5 Rest (2min)

PHE

with 50 N compression ×5

PHE

without pelvic compression ×5

Rest (2min)

PHE

without pelvic compression ×5

Rest (2min)

PHE

without pelvic compression ×5

Fig. 3. Flow chart showing study protocol. Hatched circles indicate no measurements were taken during those trials.

486

H. Takasaki et al. / Manual Therapy 14 (2009) 484–489

2.4. Procedures The protocol for testing is described in the flow chart in Fig. 2. Due to operational and time constraints, the testing procedure was conducted in two parts with Condition A and B carried out one month apart using the same set of 20 subjects. Condition B was essentially to determine the reliability of repeated measurements. Subjects were instructed to lie on the measurement table in a prone position and perform right hip extension until the lower edge of the patella was raised more than 15 cm from the starting position while maintaining knee-extension (Fig. 3). A standardized compressive force of zero, 50 N, and 100 N was applied across the pelvis by the device shown in Fig. 4. Two experimental conditions were employed; Condition A was PHE under the three different levels of compressive force (0 N, 50 N and 100 N). For Condition A all subjects were initially evaluated with no compression force. Half the sample was then tested with increasing compression force (50 N and 100 N) and the remaining half tested with reducing compression force (100 N and 50 N). The final trials in Condition A were with no compression force. Condition B was PHE without compressive force (Fig. 3). For Condition A and B a set of five trials were obtained for each level of force, with a 2-min rest period between each set. Subjects were instructed to perform hip extension at their natural speed, repeating the movement each time from rest. The mean onset of muscle activity for each set was calculated from the five trials within that set. Before the initiation of data collection, subjects provided written informed consent. This study was approved by Sapporo medical university. 2.5. Statistics EMG data processing was performed using Acknowledge software (Chart v.5.2.1, ADInstruments Pty Ltd., Australia). The EMG

Table 1 Average Gluteus Maximus onset time for each subject relative to semitendinosus muscle firing. Subject

No Compression (1st)

50 N Compression

100 N Compression

No Compression (2nd)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

130.4 137.2 430.8 243.0 327.2 213.2 324.0 164.0 249.0 189.4 213.0 316.4 190.0 342.2 109.2 471.8 316.6 367.6 230.8 299.4

83 85.8 385.4 214.6 212.2 151.4 271.4 109.4 200.6 105.0 150.4 210.6 83.0 245.6 102.2 238.2 229.8 239.4 148.8 203.4

60.2 59.0 260.2 124.2 93.2 60.0 123.8 79.2 104.2 60.5 61.2 106.8 22.2 123.4 60.6 109.6 122.0 88.2 46.4 64.4

61.0 65.0 432.8 131.2 133.8 163.6 244.0 127.4 144.4 164.6 154.8 153.0 79.2 244.6 90.6 264.6 201.0 201.6 175.6 197.2

Mean SD

263.3 99.5

183.5 77.9

91.5 49.7

171.5 84.3

signals were full wave rectified, low- and high-pass filtered, with cut-off frequencies of 500 and 10 Hz, respectively, and recorded at a sampling rate of 1000 Hz (Sakamoto et al., 2009). Muscle activation patterns were described after determining the EMG onset for each muscle. The onset of muscular activity was considered to occur when the value exceeded two standard deviations from the mean value observed at baseline for a 50 ms period (Hodges and Bui, 1996; Brindle et al., 1999). Muscle onsets were calculated with respect to onset of muscle activity for the semitendinosus muscle. A one-way ANOVA was used to determine the influence of pressure across the pelvis on the timing of muscle onset relative to the onset of semitendinosus muscle activity. Statistical analysis was performed using SPSS version 11.5 (SPSS Inc., Tokyo, Japan). Statistical significance was attributed to P values less than 0.05. 3. Results 3.1. Gluteus maximus

Fig. 4. Pelvis compression device.

Under Condition A, with the semitendinosus muscle acting as the relative starting point at 0 ms, gluteus maximus become active 263.3  99.5 ms later with no pelvic compression, 183.5  77.9 ms later with 50 N compression, 91.5  49.7 ms later with 100 N compression, and 171.5  84.3 ms later when no pelvic compression was repeated. In all subjects semitendinosus muscle onset occurred prior to gluteus maximus (Table 1). Significant differences were found in EMG onset between no pelvic compression and 50 N compression (P < 0.05), between no pelvic compression and 100 N compression (P < 0.001) and between the two trials of no pelvic compression (P < 0.01). Additionally, there were significant differences between EMG onset for 100 N compression and 50 N compression (P < 0.01), and between 100 N compression and the second trial of no pelvic compression (P < 0.05) (Fig. 5). Under Condition B, gluteus maximus contracted 270.2  90.3 ms later than semitendinosus muscle during first five sets of PHE, and 218.5  71.2 ms later during the last five sets of PHE. Significant differences were found in EMG onset between the second trial of no pelvic compression in Condition A and the first

H. Takasaki et al. / Manual Therapy 14 (2009) 484–489

487

*p<0.05,‡p<0.01,†p<0.001

2nd

Condition B

1st

2nd

Condition B

1st

No Compression (2nd)

‡

100 N Compression

No Compression (2nd) ‡

‡

†

* ‡

-500

-400

Condition A

50 N Compression

*

No Compression (1st) -100

No Compression (1st)

-300

-200

-100

Condition A

50 N Compression

100 N Compression

-50

Fig. 5. The latency (ms) of gluteus maximus compared with the onset of semitendinosus for condition A and B. Negative values indicate delayed activation.

trial in Condition A (P < 0.01), and between the average of the first five sets of trials in Condition B (P < 0.01) (Fig. 5). 3.2. Ipsilateral erector spinae Under Condition A, ipsilateral erector spinae became active 0.1  39.6 ms before the onset of the semitendinosus muscle with the first trials of no pelvic compression, 27.3  54.3 ms later with 50 N compression, 10.2  31.4 ms later with 100 N compression, and 13.4  37.2 ms later with the second set of trials of no pelvic compression. In 13 out of 20 subjects the ipsilateral erector spinae contracted before semitendinosus muscle (Table 2). There was no significant difference (a ¼ 0.05) between any of these conditions (Fig. 6). Under Condition B, the ipsilateral erector spinae contracted 0.5  37.2 ms before the onset of the semitendinosus muscle during first five sets of PHE, and 1.4  31.1 ms prior in the last five sets of PHE. Significant differences (a ¼ 0.05) were not found between any of Condition A and B (Fig. 6).

50

Latency (ms)

0

Latency (ms)

0

Fig. 6. The latency (ms) of ipsilateral erector spinae compared with the onset of semitendinosus for condition A and B. Negative values indicate delayed activation. There was no significant difference between any condition (a ¼ 0.05).

muscle, with the first trials of no pelvic compression, were 17.7  35.7 ms prior, 2.1  39.9 ms prior with 50 N compression, 2.1  32.4 ms later with 100 N compression, 6.7  40.5 ms prior with the second set of trials of no pelvic compression. Activation prior to semitendinosus muscle was seen in 15 out of 20 subjects (Table 3). There was no significant difference (a ¼ 0.05) between any of these conditions (Fig. 7). Under Condition B, contralateral erector spinae become active 11.4  38.0 ms prior to the onset of the semitendinosus muscle during the first five sets of PHE, and 11.3  30.7 ms prior during the last five sets of PHE. There was no significant differences (a ¼ 0.05) between any of Condition A and B (Fig. 7). 4. Discussion

Under Condition A, the averaged onset of the contralateral erector spinae, compared with the onset of the semitendinosus

This study found gluteus maximus muscle onset was consistently delayed with respect to the semitendinosus muscle during the PHE test. No such consistent pattern of temporal activation was found for either the ipsilateral or contralateral erector spinae, which is in line with previous reports (Lehman et al., 2004). Compression force across the pelvis, appeared to reduce the onset delay of gluteus maximus, but this had no such effect on the erector spinae in asymptomatic subjects. Delayed onset gluteus maximus muscle activation (relative to the hamstring muscles) has been suggested as a significant factor in

Table 2 Average ipsilateral erector spinae onset time for each subject relative to semitendinosus muscle firing.

Table 3 Average contralateral erector spinae onset time for each subject relative to semitendinosus muscle firing.

Subject

No Compression (1st)

50 N Compression

100 N Compression

No Compression (2nd)

Subject

No Compression (1st)

50 N Compression

100 N Compression

No Compression (2nd)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

13.6 6.6 28.4 53.0 31.8 30.2 35.0 29.6 57.8 69.8 79.4 10.0 58 24.0 31.4 31.0 29.2 4.0 22.2 10.6

5.6 3.4 139.6 72.8 81.6 55.8 10.4 15.2 44.0 80.2 48.0 22.8 7.2 12.0 149.6 36 0.8 30.8 5.2 4.8

16.4 24.8 29.4 27.0 24.6 6.8 13.4 9.4 59.0 65.6 47.0 30.8 9.2 0.4 50.2 8.0 10.8 58.0 9.2 17.4

26.2 21.6 12.2 76.4 9.2 31.6 99.6 33.4 28.2 27.8 39.4 0.8 27.4 6.0 70.2 16.0 26.4 16.8 13.0 19.4

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

19.0 33.8 35.4 66.2 57.2 18.8 34.6 15.8 29.8 38.8 32.6 38.6 51.4 31.6 31.4 50.6 57.6 15.8 46.6 19.4

20.2 31.0 23.6 3.8 93.6 65.6 6.2 15.4 0.4 56.4 16.6 0.4 10.2 43.8 90.6 5.4 6.2 13.2 17.6 28.8

28.0 11.8 24.0 4.6 48.8 69.0 7.2 7.4 3.4 37.6 36.8 5.8 15.6 50.8 50.4 0.4 1.0 47.8 1.8 28.8

32.0 2.2 82.4 29.2 5.8 40.4 79.0 42.4 37.4 28.0 45.8 25.4 19.6 42.6 32.6 46.2 41.6 3.4 27.2 30.8

0.1 39.6

27.3 54.3

10.2 31.4

13.4 37.2

Mean SD

17.7 35.7

2.1 39.9

2.1 32.4

6.7 40.5

3.3. Contralateral erector spinae

Mean SD

488

H. Takasaki et al. / Manual Therapy 14 (2009) 484–489

2nd 1st

Condition B

No Compression (2nd) 100 N Compression

Condition A

50 N Compression No Compression (1st) -100

-50

0

50

Latency(ms) Fig. 7. The latency (ms) of contralateral erector spinae compared with the onset of semitendinosus for condition A and B. Negative values indicate delayed activation. There was no significant difference between any conditions (a ¼ 0.05).

the development of hip pain (Sahrmann, 2002; Hungerford et al., 2003; Vogt et al., 2003). The present study and previous studies have reported delayed gluteus maximus activation in normal subjects (Bullock-Saxton et al., 1994; Vogt and Banzer, 1997; Lehman et al., 2004), but failed to reach agreement on the magnitude of the delay. Latency of gluteus maximus contraction in our study of male subjects with no compression was approximately 270 ms, in contrast to 500 ms reported by Sakamoto et al (Sakamoto et al., 2009). In that study the standard deviation was three times larger at 300 ms. This disparity might be explained by gender differences, with participants in Sakamoto et al’s study of both genders. Further research investigating a more diverse sample of normal subjects and patients with hip pain is required to identify whether a greater timing delay occurs in symptomatic subjects when compared to healthy controls. Gluteus maximus activation has a major functional role in the early stance phase of gait, where 60% of body weight is transferred in 0.02 s, resulting in abrupt loading of the forward limb (Anderson and Pandy, 2003). At this point the gluteus maximus compresses the SIJ to provide stability of the pelvis (Hossain and Nokes, 2005). Additionally, it has been suggested that lack of control of the pelvis may further increase the movement of an already mobile lumbar spine segment (McConnell, 2002). It has been established that excessive movement, particularly in rotation of the lumbar motion segment, is a contributory factor to disc injury and torsional forces may irrevocably damage the annulus fibrosus (Farfan et al., 1970; Kelsey et al., 1984). Our study analyzed the muscle firing pattern during hip extension in prone, as it has been suggested that the pattern of activation of muscles during PHE represents the muscle recruitment pattern of hip extension during gait (Lehman et al., 2004). To our knowledge no studies have compared muscle onset patterns between the two positions. There may be major changes to the pattern of activation of the lumbo-pelvic stabilizing muscles in prone compared to the upright posture. Additionally, gravitational influences in both positions are different and may influence both force and form closure and so pelvic stability. Therefore relating our findings of the firing pattern during the PHE test to the normal gait cycle is not possible. Much further research is required comparing the muscle firing pattern during normal gait with that found during the PHE test. The synovial SIJ is supported by the anterior, posterior and interosseous sacroiliac ligaments. While the posterior sacroiliac ligament is strong, the interosseous ligament is the strongest ligament between the sacrum and ilium (Clemete, 1997). When the therapist applies force across the pelvis, the strong sacroiliac ligaments increase the lever arm acting on the SIJ, which theoretically results in much greater compression force at the SIJ surface and increased force closure. Although there are no studies that have directly measured this in cadavers, one study has modeled the forces (of the same magnitude as used in our study) and found

medial compression force on the pelvis increases force closure at the SIJ (Pel et al., 2008). Similarly both Mens and Damen (Damen et al., 2002; Mens et al., 2006) found medially applied pelvis force (using a pelvic belt) at the level of the anterior superior iliac spines produced significantly less SIJ laxity in both healthy women and those with pelvic pain. The SIJ has been reported to be richly innervated (Solonen, 1957; Ikeda, 1991; Grob et al., 1995). Ikeda (Ikeda, 1991) reported that the ventral portion of the SIJ was mainly supplied by ventral ramus of the L5, whereas the lower ventral portion was mainly supplied by ventral ramus of the S2. Thick, thin, and unmyelinated nerve fibers have been reported, which are compatible with a broad repertoire of sensory receptors, indicating encapsulated mechanoreceptors (Ikeda, 1991; Grob et al., 1995). According to Indahl, et al. (Indahl et al., 1999) stimulation of the porcine SIJ capsule elicited activity in the multifidus muscle (L5 level), additionally stimulation of the anterior aspect of the joint elicited responses in quadratus lumborum and gluteus maximus. In our study, the stronger the force applied across the pelvis, the earlier the onset of gluteus maximus activity occurred. In contrast this was not the case for the erector spinae muscles. These findings correspond with previous reports (Indahl et al., 1999) With respect to the results of the latency of gluteus maximus compared with the onset of semitendinosus in Condition A, a significant difference was found between the first and second trial of no compression. In contrast, repeated trials of no compression force in Condition B were not different and were no different to the first trial of no compression in Condition A. It appears that the change in onset of gluteus maximus activity continued even after pelvis compression force was removed. This result, suggests a lasting effect of compression force, at least temporarily. We suggest that stimulation of the SIJ by the compressive force across the pelvis causes reflex changes in muscle activation, which facilitates an earlier onset of gluteus maximus activity. It remains unclear how long this effect is maintained. It is possible that other unidentified factors influenced the muscle firing pattern, besides the SIJ compression/force closure theory. For example in our study standardization of the pelvis force application was achieved by the use of a mechanical device. No subject reported pain with this device, but it is possible that subjects may have felt discomfort or experienced difficulty with movement. It is known that experimentally induced pain changes the pattern of muscle activation around the pelvis, even in normal subjects (Hodges et al., 2003). However negating this possibility is that changes to the firing pattern were maintained even after the compression force was removed, indicating a carry over effect. Many further studies are required to investigate this phenomenon as well as the duration of carry over effect, and whether changes to the muscle firing pattern during PHE influences the pattern of recruitment during normal gait before clinical implications can be considered.

5. Conclusions In asymptomatic males, gluteus maximus muscle onset was consistently delayed with respect to the semitendinosus muscle during the PHE test, but no consistent temporal pattern of activation of the erector spinae was found. Compressive force applied medially across the pelvis significantly reduced the muscle onset delay for gluteus maximus but had no effect on the erector spinae.

References Anderson FC, Pandy MG. Individual muscle contributions to support in normal walking. Gait and Posture 2003;17(2):159–69.

H. Takasaki et al. / Manual Therapy 14 (2009) 484–489 Brindle TJ, Nyland J, Shaprio R, Caborn DN, Stine R. Shoulder proprioception: latent muscle reaction times. Medicine and Science in Sports and Exercise 1999; 31(10):1394–8. Bullock-Saxton JE, Janda V, Bullock MI. The influence of ankle sprain injury on muscle activation during hip extension. International Journal of Sports Medicine 1994;15(6):330–4. Buyruk HM, Stam HJ, Snijders CJ, Vleeming A, Lame´ris JS, Holland WP. The use of colour doppler imaging for the assessment of sacroiliac joint stiffness: a study on embalmed human pelvises. European Journal of Radiology 1995a;21(2):112–6. Buyruk HM, Snijders CJ, Vleeming A, Lame´ris JS, Holland WP, Stam HJ. The measurements of sacroiliac joint stiffness with colour doppler imaging: a study on subjects. European Journal of Radiology 1995b;21(2):117–21. Buyruk HM, Stam HJ, Snijders CJ, Lame´ris JS, Holland WP, Stijnen TH. Measurement of sacroiliac joint stiffness in peripatum pelvic pain patients with doppler imaging of vibrations(DIV). European Journal of Obstetrics, Gynecology, and Reproductive Biology 1999;83(2):159–63. Clemete CD. Anatomy. A regional atlas of the human body. 4th ed. Baltimore: William and Wilkins; 1997 [chapter 4], pp. 257. Cram JR, Kasman GS, Holtz J. Introduction to surface electromyography. 1st ed. Maryland: Aspen Publishers; 1998 [chapter 14], pp. 336–339, 349–352, 368–370. Damen L, Spoor CW, Snijders CJ, Stam HJ. Does a pelvic belt influence sacroiliac joint laxity? Clinical Biomechanics 2002;17(7):495–8. Farfan HF, Cossette JW, Robertson GH, Wells RV, Kraus H. The effects of torsion on lumbar intervertebral joints: the role of torsion in the production of disc degeneration. Journal of Bone and Joint Surgery 1970;52(3):468–97. Grob KR, Neuhuber WL, Kissling RO. Innervation of the sacroiliac joint of the human. Z Rheumatol 1995;54(2):117–22. Hodges PW, Bui BH. A comparison of computer-based methods for the determination of onset of muscle contraction using electromyography. Electromyography and Clinical Neurophysiology 1996;101(6):511–9. Hodges PW, Moseley GL, Gabrielsson A, Gandevia SC. Experimental muscle pain changes feedforward postural responses of the trunk muscles. Experimental Brain Research 2003;151(2):262–71. Hossain M, Nokes LD. A model of dynamic sacro-iliac joint instability from malrecruitment of gluteus maximus and biceps femoris muscles resulting in low back pain. Medical Hypotheses 2005;65(2):278–81. Hungerford B, Gilleard W, Hodges P. Evidence of altered lumbopelvic muscle recruitment in the presence of sacroiliac joint pain. Spine 2003;28(14): 1593–600. Ikeda R. Innervation of the sacroiliac joint. Macroscopical and histological studies. Nippon Ika Daigaku Zasshi 1991;58(5):587–96 [in Japanese]. Indahl A, Kaigle A, Reikerås O, Holm S. Sacroiliac joint involvement in activation of the porcine spinal and gluteal musculature. Journal of Spinal Disorders 1999;12(4):325–30. Janda V. Muscles, central nervous motor regulation and back problems. In: Korr I, editor. Neurobiologic mechanisms in manipulative therapy. 1st ed. New York: Plenum Presss; 1978. p. 27–41. Janda V. Treatment of chronic low back pain. Journal of Manual Medicine 1992;6(5): 166–8.

489

Janda V. Evaluation of muscular imbalance. In: Liebenson C, editor. Rehabilitation of the spine: a practitioner’s manual. 1st ed. Baltimore: Lippincott Williams and Wilkins; 1996. p. 97–112 [chapter 6]. Kelsey JL, Githens PB, White 3rd AA, Holford TR, Walter SD, O’Connor T, et al. An epidemiologic study of lifting and twisting on the job and risk for acute prolapsed lumbar intervertebral disc. Journal of Orthopaedic Research 1984;2(1):61–6. Lee D, Lee L. Diagnosing the limbopelvic-hip dysfunction. In: Lee D, editor. The pelvic girdle: an approach to the examination and treatment of the lumbopelvic-hip region. 3rd ed. Edinburgh: Churchill Livingstone; 2004. p. 81–132 [chapter 8]. Lehman GJ, Lennon D, Tresidder B, Rayfield B, Poschar M. Muscle recruitment patterns during the prone leg extension. BMC Musculoskeletal Disorders 2004;5(3):1–5. McConnell J. Recalcitrant chronic low back and leg pain: a new theory and different approach to management. Manual Therapy 2002;7(4):183–92. Mens JM, Damen L, Snijders CJ, Stam HJ. The mechanical effect of a pelvic belt in patients with pregnancy-related pelvic pain. Clinical Biomechanics 2006;21(2): 122–7. Mens JM, Vleeming A, Snijders CJ, Stam HJ, Ginai AZ. The active straight legraising test and mobility of the pelvic joints. European Spine Journal 1999;8(6):468–73. Mens JM, Vleeming A, Snijders CJ, Koes BW, Stam HJ. Reliability and validity of the active straight leg raise test in posterior pelvic pain since pregnancy. Spine 2001;26(10):1167–71. Pel JJ, Spoor CW, Goossens RH, Pool-Goudzwaard AL. Biomechanical model study of pelvic belt influence on muscle and ligament forces. Journal of Biomechanics 2008;41(9):1878–84. Sakamoto AC, Teixeira-Salmela LF, de Paula-Goulart FR, de Morais Faria CD, Guimara˜es CQ. Muscular activation patterns during active prone hip extension exercises. Journal of Electromyography and Kinesiology 2009;19(1):105–12. Sahrmann S. Diagnosis and treatment of movement impairment syndromes. 1st ed. Missouri: Mosby Inc; 2002 [chapter 4], pp. 121–192. Snijders CJ, Vleeming A, Stoeckart R. Transfer of lumbosacral load to iliac bones and legs. 1: biomechanics of self-bracing of the sacroiliac joints and its significance for treatment and exercise. Clinical Biomechanics 1993a;8(6):285–94. Snijders CJ, Vleeming A, Stoeckart R. Transfer of lumbosacral load to iliac bones and legs. 2: loading of the sacroiliac joints when lifting in a stooped posture. Clinical Biomechanics 1993b;8(6):295–301. Solonen KA. The sacroiliac joint in the light of anatomical, roentgenological and clinical studies. Acta Orthopedica Scandinavica 1957;27(Suppl.):1–127. Vleeming A, Stoeckart R, Volkers AC, Snijders CJ. Relation between form and function in the sacroiliac joint. 1: clinical anatomical aspects. Spine 1990a;15(2): 130–2. Vleeming A, Volkers AC, Snijders CJ, Stoeckart R. Relation between form and function in the sacroiliac joint. 2: biomechanical aspects. Spine 1990b;15(2): 133–6. Vogt L, Banzer W. Dynamic testing of the motor stereotype in prone hip extension from neutral position. Clinical Biomechanics 1997;12(2):122–7. Vogt L, Pfeifer K, Banzer W. Neuromuscular control of walking with chronic lowback pain. Manual Therapy 2003;8(1):21–8.

Manual Therapy 14 (2009) 490–495

Contents lists available at ScienceDirect

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

Original Article

Effect of general flexibility on thumb-tip force generation - implication for mobilization and manipulation Meng-Tzu Hu a, Ar-Tyan Hsu b, c, *, Se-Wei Lin d, Fong-Ching Su d a

Department of Physical Therapy, Tzu Hui Institute of Technology, Nanchou Hsian, Ping Tung, Taiwan Department of Physical Therapy, College of Medicine, National Cheng Kung University, Tainan, Taiwan c Institute of Allied Health Sciences, College of Medicine, National Cheng Kung University, Tainan, Taiwan d Institute of Bioengineering, College of Engineering, National Cheng Kung University, Tainan, Taiwan b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 24 November 2007 Received in revised form 12 September 2008 Accepted 9 October 2008

Pain involving basal joints of the thumb is one of the major occupation-related disorders for orthopedic physiotherapists and manual therapists. The thumb-tip force generation while performing manual techniques may be influenced not only by the specific manual techniques employed but also by general flexibility of the therapist. The purpose of this study was to investigate the influence of general flexibility and different techniques on thumb-tip force generation. Twenty-three subjects with no exposure to manual techniques and 15 physical therapy clinicians with at least 3 years of orthopedic experience participated. The general flexibility of each subject was assessed by Beighton score (BS). Each subject was requested to exert a maximal force on a six-component load cell with the thumb unsupported (T1), with the rest of digits supported (T2), and with interphalangeal joint of the thumb supported by the index (T3).The thumb-tip force was normalized by body weight. The thumb-tip force generation is influenced not only by the differences in technique employed by the therapists, but also by the general flexibility of the therapists. Physiotherapists with excessive thumb flexibility are advised to perform PA glide with IP joint supported to protect the thumb joints from injury. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Thumb Thumb-tip force Manipulation Flexibility

1. Introduction As professionals dependent upon manual work, physiotherapists and manual therapists are susceptible to a variety of occupational musculoskeletal injuries (Bork et al., 1996; Cromie et al., 2000). Among these injuries, pain at the basal joints of the thumb is one of the most common complaints, second only to back pain (Holder et al., 1999; West and Gardner, 2001). There appears to be an association between the use of manipulative techniques and the prevalence of thumb pain, as treating patients with manual techniques requires repetitive exertion of high level forces through the joints of the thumb (West and Gardner, 2001). Common sites of thumb pain for manual therapists are the first metacarpophalangeal (MCP), and the carpometacarpal (CMC) joints (West and Gardner, 2001; Caragianis, 2002; Wajon and Ada, 2003; Wajon et al., 2007). Techniques most frequently associated with the

* Corresponding author. Department of Physical Therapy, College of Medicine, National Cheng Kung University, 1 University Road, Tainan, Taiwan. Tel.: þ886 6 2353535 ext 5931; fax: þ886 6 2370411. E-mail address: [email protected] (A.-T. Hsu). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.10.003

thumb pain were the unilateral and the central posteroanterior (PA) glides (Wajon and Ada, 2003; Wajon et al., 2007). The PA glide manipulation/mobilization techniques are employed in the examination and treatment of vertebral disorders (Maher and Adams, 1995; Hertling and Kessler, 1996; Maitland et al., 2001). These techniques are passive movements imparted through tips or pads of the thumbs of the therapist to the targeted structure in specific directions in the form of a thrust, a sustained pressure or a series of oscillatory movements (Hertling and Kessler, 1996; Maitland et al., 2001). The oscillatory or sustained PA pressure is imposed partially by the body weight and transmitted to the thumbs through arms and forearms. Mobilization/manipulation techniques of various grades and characteristics are used to reduce pain and to increase mobility (Maitland et al., 2001). It has been reported that the average thumb-tip force registered from a grade 3 central PA glide is about 40 N (Chiradejnant et al., 2001). During manipulation or mobilization, thumb muscles are coordinated to provide sufficient support for balancing the poly-articular chain against the resistance encountered at the tip of the thumb. Torques and counter torques created about the joints of the thumb during such exertion require active and passive supports provided by the musculotendinous and

M.-T. Hu et al. / Manual Therapy 14 (2009) 490–495

capsuloligamentous structures of the thumb and, thus, may be influenced by the general flexibility of the individual. Ligamentous strains and cartilaginous degradation may be the primary causes for these occupation-related disorders specific to manual therapists. The alignment of the thumb during the performance of a PA pressure technique and increased CMC laxity have been reported to be associated with the prevalence of the thumb pain (Snodgrass et al., 2003; Wajon et al., 2007) and a higher prevalence of osteoarthritis when taking into consideration the mean age of the respondents (Snodgrass et al., 2003). Lifting or transferring patients, treating a large number of patients in a single day, working in awkward positions or in the same position for long periods of time, and performing manual orthopedic techniques were considered major problems for work related disorders among therapists (Bork et al., 1996). We suspect that therapists with manipulation related thumb pain might have exhibited symptoms with progressive capsular laxity and hypermobility at the joints of the thumb similar to patients reported by Eaton and Littler (1973). Laxity and hypermobility predispose a joint to vicious cycles of acute and chronic synovitis and further hypermobility (Eaton and Littler, 1973). Several studies have investigated the association between joint hypermobility and osteoarthritis of the basal joints of the thumb, and between hypermobility and musculoligamentous lesions (Dı´az et al., 1993; Jonsson et al., 1996). The prevalence of hypermobility was high in patients with osteoarthritis of the first CMC joint (Jonsson et al., 1996), and the occurrence of musculoligamentous lesions was more frequent in the hyperlax and lax individuals than in normal individuals (Dı´az et al., 1993). The purpose of this study was to investigate the influence of clinical experience and general flexibility, in terms of Beighton scores, on the magnitude of thumb-tip force generation in 3 forms of PA glide manipulation/mobilization techniques. 2. Methods 2.1. Subjects A Novice group which included twenty-three healthy subjects (11 females and 12 males, aged 21.1  2.2 years) without clinical experience in manual physical therapy were recruited for this study. An Experienced group included fifteen healthy subjects (10 females and 5 males, aged 28.0  2.2 years) with at least 3 years of clinical experience in orthopedic physical therapy. Those subjects who had any history of musculoskeletal, neurological disorders, or pain involving the use of the thumb in the past years were excluded from this study. The demographic information of both groups of subjects is listed in Table 1. There was no difference between these 2 groups in the basic demographic data except age and clinical experience. Each subject signed an inform consent form. The protocol of this research was approved by the Institutional Review Board of the National Cheng Kung University Hospital.

491

2.2. Instrumentations A six-axis load cell (MC3A-6-100, Advanced Mechanical Technology Inc., Watertown, MA, USA) was used to measure forces and moments applied through the thumb at a 50 Hz sampling rate. The capacity of this load cell is 444.8 N for Fz, 222.4 N for both Fy and Fx, 5.6 Nm for Mz, and 11.3 Nm for both Mx and My (Fig. 1). An instruNet data acquisition system (GW Instrument Inc., 35 Medford St, Somerville, MA 02143, USA) was used to register data collected by the load cell.

2.3. Experimental procedures Before testing each subject read and signed an informed consent form and answered a questionnaire concerning his/her history of pain and disabilities relating to the thumb and the hand. A Beighton score was also obtained from each subject (Beighton et al., 1973). The Beighton criteria included: (1) resting palms on the floor in forward flexion with straight knees; (2) hyperextension of elbows  10 ; (3) hyperextension of knees  10 ; (4) passive apposition of the thumb to the forearm; (5) passive dorsiflexion of the fifth finger  90 (Beighton et al., 1973). The numerical scores for items (2)–(5) were scored bilaterally with a highest possible score of 9. The following thumb-tip force generating techniques were tested in a random order on the six-axis load cell (Fig. 2). 1 The unsupported PA glide (T1, Fig. 2A): The PA glide is performed with the thumb and the rest of digits unsupported. 2 The PA glide with digital support (T2, Fig. 2B): The PA glide is performed with the thumb and the rest of digits supported on the table. 3 The PA glide with the interphalangeal joint of thumb supported by the index (T3, Fig. 2C). Subjects were asked first to exert a maximum force with the thumb of the dominant hand on the force plate and perform each of the three PA glide techniques described previously. The maximal force recorded from each of the three techniques was set as 100% for each technique. A LABVIEW (version 6.1, National Instruments, Austin, TX, USA) program was written to calibrate the six-axis load

Table 1 Demographic data of participants. Basic data

Novice Group (N ¼ 23) Mean (SD)

Experienced Group (N ¼ 15) Mean (SD)

Age (years)a Gender (male/female) Height (cm) Weight (kg) BMI (kg/m2) Clinical experience (years)a

21.1 (2.2) 12/11 168.2 (7.3) 61.5 (17.0) 21.5 (4.3) 0.0 (0.0)

28.0 (2.2) 5/10 162.9 (8.0) 56.9 (12.1) 21.3 (3.3) 4.3 (1.4)

a

Significant difference between two groups, P < 0.01.

Fig. 1. A six-component load cell mounted on an L-shaped base-plate. The upright portion of the L-shaped base-plate provided space for digital rest. The force coordinate system is also shown. Four retro-reflective markers were placed on the surface of the force plate to relate the force plate coordinate system to that of the global.

492

M.-T. Hu et al. / Manual Therapy 14 (2009) 490–495

Fig. 2. PA glide techniques were performed on a load cell on the table. (From left to right were the techniques of unsupported PA glide, T1, PA glide with digits support, T2, and PA glide with IP supported by index finger, T3).

cell and display force levels from 0% to 100% with 25% increments. The subjects were then instructed to exert at a force level matching as close as possible to the target force level displayed on the monitor for a period of 5 s. The thumb-tip force response was recorded. 2.4. Data analysis and statistics The raw force data (50 Hz sampling rate) were low-pass filtered (5 Hz) with a third-order Butterworth filter. The mean value of the maximal forces obtained from the middle 3 s was calculated for statistical analyses. The thumb-tip forces were normalized by each individual’s body weight. Data was analyzed by using the Statistical Package for Social Sciences for Windows (version 11.0, SPSS Inc. Chicago, IL, USA). Both sets of thumb-tip force data were tested for normality of distribution. The Mann–Whitney test was chosen to analyze gender difference on thumb-tip force generation. The Spearman’s rank correlation was used for analyzing ordinal scale comparison between thumb-tip force and the Beighton score in both groups. Subjects were divided into three flexibility levels according to their hypermobility scores: (1) normal individuals, with none or only one positive Beighton criterion; (2) lax individuals, with two or three positive Beighton criteria; (3) hyperlax individuals who fulfilled four or more Beighton criteria. The three-way (2 groups  3 flexibility levels  3 techniques) ANOVA with repeated measures was conducted to investigate the effects of experience, technique and flexibility on the magnitude of the maximal thumb-tip force. The significance level was set at p < 0.05. Bonfferroni post-hoc test was applied when significant differences were revealed by ANOVA. The alpha value was adjusted by using the Bonfferroni approach. 3. Results There was a gender difference in thumb-tip force (in Newtons) generation in T2 (Z ¼ 2.187, P ¼ 0.029) and T3 (Z ¼ 2.246, P ¼ 0.025, Table 2). T1 also exhibited a trend of difference between genders (Z ¼ 1.894, P ¼ 0.058). However, no gender difference was observed in the body weight normalized thumb-tip force (%BW) in any of the 3 techniques employed in the present study (Table 2). Therefore, further tests of main effects of experience, Beighton score and technique on the thumb-tip force generation were based on the body weight normalized thumb-tip force. Results of the three-way ANOVA are presented in Tables 3 and 4. There were significant technique (F ¼ 12.007, P ¼ 0.000, power ¼ 0.993) and flexibility (F ¼ 5.774, P ¼ 0.007, power ¼ 0.833)

main effects on the normalized thumb-tip force production. However, no main effect of experience was observed on the generation of thumb-tip force. No interaction was found among the 3 main effects assessed except between flexibility level and experience (F ¼ 3.513, P ¼ 0.042). In the Experienced group, the mean normalized thumb-tip force generated by individuals with normal flexibility (0.114  0.032 BW) was greater than that of lax individuals (0.060  0.014 BW, P ¼ 0.021, Table 4). The average thumb-tip force produced while performing T3 (0.094  0.033 BW) was greater than those of T1 (0.075  0.028 BW, P < 0.001) and T2 (0.082  0.032 BW, P ¼ 0.004) (Fig. 3). However, there was no difference in the magnitude of force between T1 and T2. Twenty-four of the 38 subjects exhibited hypermobility features (Beighton score  2). Among them fourteen were hyperlax individuals. In the Novice group, 7 of the 23 subjects were lax individuals and 10 were hyperlax. In the Experienced group, 7 of 15 subjects had hypermobility features and 4 of them were hyperlax individuals. Fig. 4 shows that greater forces were produced by individuals with normal flexibility (T1: 0.093  0.026 BW; T2: 0.100  0.036 BW; T3: 0.113  0.034 BW) than by the hyperlax individuals (T1: 0.061  0.026 BW; T2: 0.069  0.029 BW; T3: 0.075  0.021 BW; 95% CI ¼ 0.011w0.056, P ¼ 0.002). Subjects with normal flexibility tended to generate greater thumb-tip force than that of the lax individuals (T1: 0.068  0.024 BW; T2: 0.074  0.021 BW; T3: 0.093  0.029 BW; 95% CI ¼ 0.048w0.001, P ¼ 0.062). Table 5 shows the correlation between the thumb-tip force and the Beighton score. Inverse correlations were found in all three techniques in the Novice group. Both T1 and T3 reach significant level (r ¼ 0.450, P < 0.05 and r ¼ 0.482, P < 0.05, respectively). In the Experienced group, the thumb-tip force was negatively

Table 2 The means and standard deviations of the thumb-tip force generated with 3 different mobilization techniques in male (N ¼ 17) and female (N ¼ 21) subjects. The absolute forces are expressed both in Newton and as a percentage of body weight. Mobilization Technique

Male Mean (SD) Newton

%BW

Female Mean (SD) Newton

%BW

T1 T2 T3

51.47 (26.59)a 55.56 (24.72)b 64.28 (26.23)c

0.076 (0.035) 0.084 (0.034) 0.095 (0.031)

36.82 (10.67) 40.14 (15.28) 46.79 (16.36)

0.073 (0.023) 0.080 (0.032) 0.093 (0.035)

T1: The unsupported PA glide. T2: The PA glide with digital support. T3: The PA glide with the interphalangeal joint supported by the index. No gender difference was observed in the body weight normalized thumb-tip force (%BW) in any of the 3 techniques employed in the present study. a Trend of gender difference in thumb-tip force (N), Z ¼ 1.894, P ¼ 0.058. b Significant gender difference in thumb-tip force (N), Z ¼ 2.187, P ¼ 0.029. c Significant gender difference in thumb-tip force (N), Z ¼ 2.246, P ¼ 0.025.

M.-T. Hu et al. / Manual Therapy 14 (2009) 490–495

Main Effects Technique Experience Flexibility Interactions Technique  Flexibility Technique  Experience Flexibility  Experience Technique  Experience  Flexibility

F

Sig.

Power

12.007 0.448 5.744

0.000 0.508 0.007

0.993 0.100 0.833

0.496 0.169 3.513 0.606

0.739 0.845 0.042 0.660

0.161 0.075 0.613 0.189

Results of the three-way ANOVA (2 group  3 techniques  3 flexibility) showed significant main effect of technique and flexibility level. There is no interaction among factors tested except that of between flexibility level and group. In the Experienced group, the mean thumb-tip force generated by individuals with normal flexibility (0.114  0.032 BW) was greater than that of lax individuals (0.060  0.014 BW, P ¼ 0.021).

correlated and the Beighton score in all 3 techniques, but only T1 showed a significant difference (r ¼ 0.547, P ¼ 0.035). 4. Discussion In the present study, three forms of thumb-tip force generation techniques were tested. T2, thumb-tip force generation with the second to fifth digits supported, is similar to the AP or PA manipulative and related assessment procedures employed in the clinics, and were among the most frequently reported incriminating techniques responsible for the thumb pain in manual therapists (Wajon and Ada, 2003; Wajon et al., 2007). While T1, thumb-tip force generation without digital support, represents a technique that is relatively unstable and is not commonly used by therapists in their clinical practice. T3, on the other hand, is a reasonable alternative to T2 by stabilizing the IP joint of the thumb from the palmar side with the radial aspect of the index finger. Although the posture assumed is similar, T3 is different from the lateral pinch or the key pinch; the latter exerts a pinching force between the thumb pad and the radial side of the index while the former uses the radial side of the index to stabilize the IP joint of the thumb allowing the thumb to exert force through its tip (Smaby et al., 2004; Levangie and Norkin, 2005). The results of the present study demonstrated significant technique and flexibility main effects on the thumb-tip force production. The average thumb-tip force produced while performing T3 was greater than those of T1 and T2. Although the thumb is characterized by its great mobility comparing to other digits, a firm stability is required for applying steady pressure through the tip of the thumb. Greater thumb-tip forces were registered in PA glide with IP support (T3) in both the Novice and the Experienced groups. While applying this technique, the IP joint is supported by the counter force provided by the radial aspect of

Table 4 Descriptive statistics of thumb-tip force (%BW) during performing 3 techniques.

Normal

Novice Experience Lax Novice Experience Hyperlax Novice Experience Average a

T1

T2

T3

0.081  0.032 0.101  0.019 0.074  0.025 0.055  0.017 0.058  0.027 0.069  0.022 0.075  0.028

0.082  0.032 0.113  0.034 0.082  0.020 0.055  0.009 0.063  0.025 0.082  0.038 0.082  0.032

0.095  0.030 0.100  0.033b 0.128  0.032 0.103  0.028 0.073  0.026 0.070  0.015 0.071  0.016 0.071  0.026 0.084  0.033 0.094  0.033a

Force produced by T3 was greater than T2 (P ¼ 0.004) and T1 (P < 0.001). Force produced by individuals with normal flexibility was greater than hyperlax individuals (P ¼ 0.002). b

Normalized Force (%BW)

Table 3 Main effects of experience, flexibility level and technique on normalized thumb-tip force generated.

0.140 0.120 0.100

493

Thumb-tip Force among 3 Techniques **

0.075

0.082

**

0.094

0.080 0.060 0.040 0.020 0.000

T1

T2

T3

Technique Fig. 3. Thumb-tip forces generated in the three techniques employed. (N ¼ 38; T1: unsupported PA glide; T2: PA glide with digits supported; T3: PA glide with IP supported by index finger). Thumb-tip force generated in T3 was greater thanT1 and T2. (** P < 0.01).

the index finger, thus, with a more stable IP joint for the force transmission and a greater thumb-tip force generation. Physiotherapists who maintained IP and MCP joints of their thumb in extended position while performing PA mobilization were reported to have less prevalence of thumb pain (Wajon et al., 2007). The PA glide with IP support (T3) maintained the IP and the MCP joints steadily near the neutral positions by the index directing the compression force proximally with less deforming force as it transmits through the poly-articular chain of the thumb. The Beighton score measures the general flexibility of an individual (Beighton et al., 1973). General flexibility in terms of Beighton score affected the ability of the individual to generate thumb-tip force. In the present study, individuals with normal flexibility generated greater thumb-tip force than the hyperlax individuals and a trend toward the same effect in the lax group. Inverse correlations were found between the thumb-tip force and Beighton score in both the Novice group (in T1 and T3) and the Experienced group (in T1 only). In the Novice group, the smaller the Beighton score, the greater the force transmitted through the thumb tip, irrespective of the degree of stability of the task involved. Such results appear to suggest that flexibility of the thumb when excessive is unfavorable to the thumb-tip force generation. Buckingham et al. (2007) also noted that a greater majority (92%) of physiotherapy students tested in their study either could reach the target force but could not concurrently stabilize their thumbs in the recommended position (56%) or could not reach the target force and also could not maintain their thumbs in the position recommended (36%). The authors suggested that these students could potentially be at risk of long term injury due to their inability to maintain the recommended thumb position while performing PA techniques as the results of an inherent structural instability, an acquired capsuloligamentous laxity due to poor technique, or a lack of dynamic stability provided by the muscles acting on the thumb (Buckingham et al., 2007). In the Experienced group an inverse correlation was also found between the Beighton score and the thumb-tip force generated during the PA glide without digital supported (T1), a technique with inherent instability to the thumb and most likely unfamiliar to the therapists. When applying more stable techniques (T2 and T3), however, experienced physiotherapists were able to compensate for the excessive flexibility, possibly through experience and training. Cartilage degradation of the first CMC articulating surface was reported to be initiated in the radial quadrant of the metacarpal and progressed to the volar quadrant, and in the dorsal-radial quadrant of the trapezium progressed to the volar quadrant in the

494

M.-T. Hu et al. / Manual Therapy 14 (2009) 490–495

Thumb-tip Force among 3 Flexibility Levels

** 0.16 0.113 0.14

T1

0.1

T3

T2

Normalized Force (%BW)

0.093 0.12

0.093 0.074

0.1

0.068

0.069 0.061

0.075

0.08 0.06 0.04 0.02 0

Normal

Lax

Hyperlax

Flexibility Level Fig. 4. Thumb-tip forces in different flexibility features. (normal: N ¼ 14; lax: N ¼ 10; hyperlax: N ¼ 14) Force produced by normal individuals was greater than that of hyperlax individuals (post-hoc test with Bonfferroni.adjustment, ** P ¼ 0.002).

late-stage osteoarthritis (Koff et al., 2003). Therapists with hyperextension laxity of the MCP joint may predispose to the development of arthritis of the CMC joint and may benefit from precautionary measures to reduce or shift loading from areas where cartilage degradation frequently occur and, thus, retard the progression of the basal joint osteoarthritis of the thumb. Such precautionary measures may include fixation of the MCP joint in flexion either by changing techniques or splinting for nonsymptomatic or mild cases to alleviate basal joint symptoms by redirecting the CMC joint forces away from the palmar compartment and onto the healthier dorsal aspect of the joint. Comparing with T1 and T2, T3 generates more thumb-tip force and assumed a more flexed MCP posture. More flexion at the MCP was reported to unload the most palmar surfaces of the CMC joint regardless of the presence or severity of the arthritic disease in this joint (Moulton et al., 2001). Surgical stabilization or metacarpal extension osteotomy may be necessary for severe cases. Besides, during T3 the first dorsal interosseous muscle is also more active comparing with T1 and T2. The activation of the first dorsal interosseous muscle with a fixed index finger will pull the first metacarpal from its proximal attachments in a distal direction; therefore, unload the first CMC joint.

Table 5 Correlations of magnitudes of thumb-tip force and the Beighton Score. Mobilization Technique

Novice Group (N ¼ 23)

Experienced Group (N ¼ 15)

Spearman’s Correlation (rs)

Spearman’s Correlation (rs)

Unsupported PA glide (T1) PA glide with digital support (T2) PA glide with IP joint supported by index (T3)

0.450*

0.547*

0.301

0.327

0.482*

0.426

*P < 0.05.

Some limitations are inherent to this study. Subjects performed the PA glide techniques against a force plate and not on the human soft tissue. In most PA glide techniques employed for the spine are usually bimanual in nature and both thumbs are involved while only one thumb (right dominant hand) was tested in the present study. In a pilot study, we tested T2 and T3 techniques in 8 subjects and compared the force generated by one thumb and by both thumbs. The mean forces generated by one thumb was nearly half (T2: 0.54, T3: 0.56) of those registered in maneuvers by both thumbs. The magnitude of force generated at the thumb tip is also influenced by the postures of the thumb, wrist, forearm, arm and trunk that affected the effectiveness of force transmission through the trunk-upper extremity poly-articular chain. In the present study, subjects were instructed to stand on a platform with height adjustable to her/his comfort during force exertion. They were asked to press down vertically on the top plate of the six-component load cell as hard as they could. The postural positions of the trunk, arm and elbow were not strictly controlled and might have influenced the outcome of the present study. Furthermore, the number of subjects (N ¼ 38) recruited in the present study was relatively small. Although normalization of the thumb-tip force with the body weight helped in eliminating the effect of gender on the generated thumb-tip force, more subjects are needed in the future study. 5. Conclusion The general flexibility of a subject affects the magnitude of thumb-tip force generation in all 3 techniques tested in the Novice group; it, however, appears to affect only T1 in the Experienced group. PA glide with IP supported by the index (T3) generates greater thumb-tip force. It is recommended that a physical therapy student or a therapist with greater flexibility may use PA glide with IP supported by the index to generate greater force and to prevent possible injuries to the thumb.

M.-T. Hu et al. / Manual Therapy 14 (2009) 490–495

Acknowledgements This research was supported by grant (NSC93-2320-B-006-063) from the National Science Council, Taiwan. References Beighton P, Solomon L, Soskolne CL. Articular mobility in an African population. Annals of the Rheumatic Diseases 1973;32(5):413–8. Bork BE, Cook TM, Rosecrance JC, Engelhardt KA, Thomason ME, Wauford IJ, et al. Work-related musculoskeletal disorders among physical therapists. Physical Therapy 1996;76(8):827–35. Buckingham G, Das R, Trott P. Position of undergraduate students’ thumbs during mobilization is poor: an observation study. The Australian Journal of Physiotherapy 2007;53:55–9. Caragianis S. The prevalence of occupational injuries among hand therapists in Australia and New Zealand. Journal of Hand Therapy 2002;15(3):234–41. Chiradejnant A, Maher C, Latimer J. Development of an instrumented couch to measure forces during manual physiotherapy treatment. Manual Therapy 2001;13(3):228–35. Cromie JE, Robertson VJ, Best MO. Work-related musculoskeletal disorders in physical therapists: prevalence, severity, risks, and responses. Physical Therapy 2000;80(4):336–51. Dı´az MA, Este´vez EC, Guijo PS. Joint hyperlaxity and musculoligamentous lesions: study of a population of homogeneous age, sex and physical exertion. British Journal of Rheumatology 1993;32(2):120–2. Eaton RG, Littler JW. Ligament reconstruction for the painful thumb carpometacarpal joint. Journal of Bone and Joint Surgery (Am) 1973;55(8):1655–66. Hertling D, Kessler RM. Management of common musculoskeletal disorders: physical therapy principles and methods. 3rd ed. Philadelphia: J.B. Lippincott; 1996. Holder NL, Clark HA, DiBlasio JM, Hughes CL, Scherpf JW, Harding L, et al. Cause, prevalence, and response to occupational musculoskeletal injuries reported by

495

physical therapists and physical therapist assistants. Physical Therapy 1999;79:642–52. Jonsson H, Valtysdottir ST, Kjartansson O, Brekkan A. Hypermobility associated with osteoarthritis of the thumb base: a clinical and radiological subset of hand osteoarthritis. Annals of the Rheumatic Diseases 1996;55(8):540–3. Koff MF, Ugwonali OF, Strauch RJ, Rosenwasser MP, Ateshian GA, Mow VC. Sequential wear patterns of the articular cartilage of the thumb carpometacarpal joint in osteoarthritis. The Journal of Hand Surgery 2003;28A(4): 597–604. Levangie PK, Norkin CC. Joint structure and function – a comprehensive analysis. 4th ed. Philadelpha: F.A. Davis; 2005. Maher C, Adams RA. Psychophysical evaluation of manual stiffness discrimination. The Australian Journal of Physiotherapy 1995;41(3):161–7. Maitland GD, Hengeveld E, Banks K, English K. Maitland’s vertebral manipulation. 6th ed. Oxford: Butterworth-Heinemann; 2001. Moulton MJ, Parentis MA, Kelly MJ, Jacobs C, Naidu SH, Pellegrini Jr VD. Influence of metacarpophalangeal joint position on basal joint-loading in the thumb. The Journal of Bone and Joint Surgery (Am) 2001;83(5):709–16. Smaby N, Johanson ME, Baker B, Kenny DE, Murray WM, Hentz VR. Identification of key pinch forces required to complete functional tasks. Journal of Rehabilitation Research and Development 2004;41:215–24. Snodgrass SJ, Rivett DA, Chiarelli P, Bates AM, Rowe LJ. Factors related to thumb pain in physiotherapists. The Australian Journal of Physiotherapy 2003;49:243–50. Wajon A, Ada L. Prevalence of thumb pain in physical therapists practicing spinal manipulative therapy. Journal of Hand Therapy 2003;16(3):237–44. Wajon A, Ada L, Refshauge K. Work-related thumb pain in physiotherapists is associated with thumb alignment during performance of PA pressures. Manual Therapy 2007;12:12–6. West DJ, Gardner D. Occupational injuries of physiotherapists in North and Central Queensland. The Australian Journal of Physiotherapy 2001;47(3): 179–86.

Manual Therapy 14 (2009) 496–500

Contents lists available at ScienceDirect

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

Original Article

The effect of chronic low back pain on size and contraction of the lumbar multifidus muscleq Tracy L. Wallwork a, b, Warren R. Stanton c, Matt Freke d, Julie A. Hides a, c, * a

Division of Physiotherapy, School of Health and Rehabilitation Sciences, The University of Queensland, Brisbane, 4072, Australia Private practice, Perth, Australia c UQ/Mater Back Stability Clinic, Mater Health Services, South Brisbane, Qld, 4101, Australia d Physiotherapy department, Second Health Support Battalion, Gallipoli Barracks, Enoggera, Brisbane, Queensland, Australia b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 April 2008 Received in revised form 26 August 2008 Accepted 25 September 2008

Decreases in the size of the multifidus muscle have been consistently documented in people with low back pain. Recently, ultrasound imaging techniques have been used to measure contraction size of the multifidus muscle, via comparison of the thickness of the muscle at rest and on contraction. The aim of this study was to compare both the size (cross-sectional area, CSA) and the ability to voluntarily perform an isometric contraction of the multifidus muscle at four vertebral levels in 34 subjects with and without chronic low back pain (CLBP). Ultrasound imaging was used for assessments, conducted by independent examiners. Results showed a significantly smaller CSA of the multifidus muscle for the subjects in the CLBP group compared with subjects from the healthy group at the L5 vertebral level (F ¼ 29.1, p ¼ 0.001) and a significantly smaller percent thickness contraction for subjects of the CLBP group at the same vertebral level (F ¼ 6.6, p ¼ 0.02). This result was not present at other vertebral levels (p > 0.05). The results of this study support previous findings that the pattern of multifidus muscle atrophy in CLBP patients is localized rather than generalized but also provided evidence of a corresponding reduced ability to voluntarily contract the atrophied muscle. Ó 2008 Elsevier Ltd. All rights reserved.

keywords: Chronic low back pain (CLBP) Multifidus muscle Ultrasound imaging

1. Introduction The lumbar multifidus muscle has been the subject of considerable research using imaging techniques including magnetic resonance imaging (MRI), computerised tomography (CT scanning) and ultrasound imaging (for review, see Stokes et al., 2007). Two aspects of muscle function that can be assessed using imaging techniques are muscle size (MRI, CT, ultrasound imaging) and muscle contraction (ultrasound imaging). The clinical relevance of these techniques is that they allow documentation of morphology and dynamic muscle function in both healthy subjects and those with acute and chronic low back pain (CLBP). Detection of changes in multifidus muscle size and motor control in people with low back pain (LBP) (by comparison with healthy subjects) may provide useful information which can be used to guide rehabilitation approaches. In healthy subjects, the lumbar multifidus muscle has been shown to be symmetrical between sides (Hides et al., 1994, 1995) q Work attributed to: Division of Physiotherapy, School of Health and Rehabilitation Sciences, The University of Queensland, Brisbane, 4072, Australia. * Corresponding author. Tel.: þ61 7 3365 2718; fax: þ61 7 3365 1877. E-mail address: [email protected] (J.A. Hides). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.09.006

and increased incrementally in size on progression caudally (Hides et al., 1995). The cross-sectional area (CSA) of the multifidus muscle has been shown to be larger in male subjects, and age was not related to multifidus size (Stokes et al., 2005). The size of the multifidus muscle has also been documented in people with acute and chronic LBP. In patients with acute unilateral LBP, studies employing ultrasound imaging demonstrated a decrease in the CSA of the multifidus muscle ipsilateral to painful symptoms (Hides et al., 1994, 1996), and the atrophy was predominantly isolated to one vertebral level (L5). Multifidus muscle atrophy has been quantified in MRI and CT studies in terms of both decreased muscle size and alterations in muscle consistency (due to fatty deposits or fibrous/connective tissue infiltration). In a CT study, Danneels et al. (2000) showed that the multifidus muscle was selectively decreased at the lowest lumbar level in patients with CLBP when compared with control subjects. In an MRI study of 78 patients with CLBP, degeneration of the multifidus muscle was present in 80% of the participants, and most commonly was seen at L4–5 and L5–S1 (Kader et al., 2000). Similar patterns of atrophy have been demonstrated using ultrasound imaging (Hides et al., 2008a,b). Dynamic studies of the multifidus muscle are of interest as they provide information related to motor control of the muscle.

T.L. Wallwork et al. / Manual Therapy 14 (2009) 496–500

Viewing the muscle contracting in parasagittal section, two studies have provided feedback of multifidus muscle contraction to patients with acute (Hides et al., 1996) and chronic LBP (Hides et al., 2008a). A randomized controlled trial conducted by Van et al. (2006) showed that provision of visual biofeedback using ultrasound imaging improved the ability to contract the multifidus muscle in healthy subjects. While observation of multifidus muscle contraction using ultrasound imaging has been used in clinical practice for quite some time, it is only recently that the technique has been validated by comparison with fine wire electromyography (EMG). Kiesel et al. (2007) used graded resistance of contralateral upper extremity lifts to produce incremental involuntary contraction of the lumbar multifidus muscles and demonstrated a relationship between increases in muscle thickness and fine wire EMG activity for contractions of 19–43% of maximum. In addition, the effect of pain on multifidus muscle function has been demonstrated experimentally by using a model of induced pain (Kiesel et al., 2008). However, the ability of patients with CLBP to voluntarily contract the multifidus has not been formally assessed. The aim of this study was to compare both the size (CSAs) and the ability to voluntarily contract the multifidus muscle at four vertebral levels in subjects with and without CLBP using real-time ultrasound imaging. 2. Methods 2.1. Subjects Seventeen subjects with CLBP (8 males, 9 females, age range 18–60) and seventeen healthy subjects (8 males, 9 females, age range 18–45) participated in the study (Table 1). CLBP was defined in this study as a history of non-specific LBP for more than 3 months (International Association for the Study of Pain, 2008). Exclusion criteria for the CLBP subjects included histories suggestive of nonmechanical LBP, overt neurological signs, previous lumbar surgery, self-reported pain levels of less than 3 on a visual analogue scale (VAS) and LBP associated with a worker’s compensation or motor vehicle accident claim. Exclusion criteria for all subjects included pregnancy, presence of spinal abnormalities, presence of scoliosis with a rib height difference of greater than 2 cm on forward flexion, histories of severe trauma, spinal or abdominal surgery, reported neuromuscular or joint disease, training involving the back muscles within 3 months and difficulty lying in the prone position. This study was approved by the Medical Research Ethics Committee at the host institution. Informed consent was obtained and the rights of human subjects were protected. 2.2. Procedure Three experienced musculoskeletal physiotherapists were involved in data collection. Examiner 1 was responsible for applying inclusion and exclusion criteria and collection of demographic data. For subjects with LBP a body chart was used to record distribution of symptoms and a VAS was used to assess pain levels experienced over the last week. Height and weight were measured and age, gender and weekly activity levels (<1.5, 1.5–3, >3 h) were

Table 1 Demographics of subjects in Group 1 (CLBP) and Group 2 (Unimpaired). Age

Group 1 (Unimpaired) n ¼ 17 Group 2 (CLBP) n ¼ 17

Weight

Height

Mean

(SD)

Mean

(SD)

Mean

(SD)

33.9 41.9

(11.2) (13.7)

81.2 76.1

(12.5) (16.7)

176.6 174.2

(10.3) (10.3)

497

recorded for all subjects. Examiner 2 assessed multifidus muscle size (CSA) and examiner 3 assessed the ability to voluntarily contract the multifidus muscle (muscle thickness measures). Examiners 2 and 3 were blinded to group allocation of the subjects. The assessors were blinded to each other’s results. 2.2.1. Assessment using ultrasound imaging Ultrasound imaging was conducted using Diasonics Synergy ultrasound imaging apparatus equipped with a 5 MHz curvilinear transducer (GE-Diasonics, Japan). Subjects were positioned in prone lying, with a pillow placed under the abdomen to minimize the lumbar lordosis. The spinous processes from L2–L5 were marked with a pen. Detection of spinous processes was determined manually using the iliac crests as a landmark. The location of the spinous processes was then confirmed using ultrasound imaging by viewing the spinous processes relative to the sacrum in sagittal section. 2.2.2. Assessment of multifidus muscle CSAs CSAs of the multifidus muscle were measured from L2 to L5 vertebral segments. Reliability of performing these measures has been previously reported (Hides et al., 1992, 1994; Stokes et al., 2005; Pressler et al., 2006) and previous clinical trials have shown the highly trained assessor (examiner 2) in the present study to be repeatable and reliable with ultrasound measurements of multifidus muscle CSA (Hides et al., 1992, 1994). The validity of measurements obtained using ultrasound imaging has also been demonstrated by comparison with MRI measurements (Hides et al., 1995). Subjects were instructed to relax the paraspinal musculature, electroconductive gel was applied, and the transducer placed transversely over the spinous process of the vertebral level being measured. This produced images in which the spinous process and laminae could be seen, with multifidus muscles visible on both sides of the spine (Fig. 1A). The echogenic vertebral lamina was used consistently as a landmark to identify the muscle’s deep border. The multifidus muscle is bordered superiorly by the thoracolumbar fascia, and the medial border was provided by the acoustic shadow from the tip of the spinous process of the vertebral level being assessed. The lateral border was formed by the fascia surrounding the multifidus and separating it from the longissimus component of the lumbar erector spinae muscle. Bilateral images of the multifidus muscles were obtained where possible (Fig. 1A), except in the case of larger muscles where left and right sides were imaged separately. The CSA (in cm2) of the multifidus was measured by tracing around the muscle border with the on-screen cursor (Fig. 1B). For consistency, the inner edge of the border was used. 2.2.3. Assessment of multifidus muscle thickness and contraction Prior to testing of contraction of the multifidus muscle, all subjects received an initial explanation. The anatomical location of the multifidus muscle was demonstrated using a model of the lumbar spine, and pictures of the muscle were provided and explained. A demonstration of an isometric contraction of the biceps was performed as a simple example of the type of contraction required. Subjects were further instructed to take a relaxed breath in and out, pause breathing and then try to ‘‘swell’’ or contract the muscle. They were also instructed not to move their spine or pelvis when they contracted the muscle, and the type of muscle contraction required was a slow gentle sustained contraction. To familiarize subjects with the contraction prior to measuring, subjects were asked to perform 3 contractions with tactile and verbal feedback while the examiner manually palpated the multifidus muscle. It was explained to the subjects that during testing they would have 5 s to try to contract the multifidus muscle and hold the contraction. At the end of the 5 s period, the image would be saved on the ultrasound screen, and measurements performed.

498

T.L. Wallwork et al. / Manual Therapy 14 (2009) 496–500

Fig. 1. A. Bilateral transverse ultrasound image at the L4 vertebral level, without CSA tracings. B. Bilateral transverse ultrasound image at the L4 vertebral level with CSA tracings. The CSA (in cm2) of the multifidus was measured by tracing around the muscle border with the on-screen cursor.

The multifidus muscle was imaged in parasagittal (longitudinal) section allowing visualisation of the zygapophyseal joints, muscle bulk and thoracolumbar fascia (Hides et al., 1992, 1995, 1996; Van et al., 2006). The multifidus muscle was imaged on both sides from L2 to L5 vertebral levels. Linear measurements (multifidus muscle thickness measures) using on-screen callipers were made in all cases from the tip of the zygapophyseal joint to the superior border of the multifidus muscle for each vertebral level (Fig. 2). In order to assess multifidus muscle contraction, the difference between the multifidus muscle thickness at rest and during contraction was calculated. A split-screen technique was used to make this measurement more reliable, by allowing anatomical orientation to be maintained in both cases (Fig. 2). Subjects were not allowed to watch the ultrasound monitor or receive feedback about the contractions performed during testing. Prior to the present study, a reliability trial was performed on 10 healthy subjects not involved in the main study (Wallwork et al., 2007). Each subject was positioned in the standard testing position. Three separate ultrasound images were obtained at rest and the anteroposterior (thickness) measurement was conducted on parasagittal images at two vertebral levels by two raters (examiner 3 and an expert). Intraclass correlations (ICC) were used to determine intra-rater and inter-rater reliability. Results of the ICC3,1 for intrarater reliability was 0.89 for L2/3 (95%CI ¼ 0.72–0.97) and 0.88 for L4/5 (95%CI ¼ 0.68–0.97). The results of the ICC2,3 for inter-rater reliability was 0.96 for L2/3 (95%CI ¼ 0.84–0.99) and 0.97 for L4/5 (95%CI ¼ 0.87–0.99) (Wallwork et al., 2007). 2.3. Statistical analysis Those without low back pain have greater capacity to produce relatively larger contractions than those with low back pain, threatening the homogeneity of variance of the two samples. To

address this issue, the data for two males and two females were excluded from the study as outliers (more than 3 standard deviations above the sample mean). In the analyses of the thickness contraction, the data for 30 participants were used (16 in the CLBP group and 14 in the unimpaired group). Due to limited availability of examiner 2, CSA of the multifidus muscle was only possible for 22 of the 30 participants (11 in the CLBP group and 11 in the unimpaired group). A mixed design analysis of covariance (ANCOVA) was used to separately analyse the outcome measures of ‘percent change in multifidus muscle thickness due to contraction’ (called multifidus CSA percent thickness contraction) at each vertebral level. Percent thickness contraction was calculated as; [(contracted thickness  resting thickness/contracted thickness)  100]. In this study there were 7 independent variables: age, weight, height, gender, activity level (coded as low, moderate or high), group (CLBP or unimpaired), and the repeated measures of asymmetry (larger or smaller side). The variables of age, weight and height were treated as covariates in the analyses. Post-hoc contrasts were used to test for differences among the 3 activity levels if the main effect for this factor was statistically significant. For both the dependent variables, measures of ‘size’ and ‘asymmetry’ are of interest. As calculation of average size across ‘side’ is confounded by asymmetry across ‘side’, the data for ‘larger’ and ‘smaller’ side were used rather than calculating the percentage difference between the ‘larger’ and ‘smaller’ sides. In addition, as higher-order interactions between the covariates and factors confounded the analysis, a Type I sums of squares model was used in preference to a Type III model. 3. Results Demographic details for subjects of both groups are shown in Table 1. Results of an initial analyses of variance showed that there was no significant difference between the two groups for the variables of age (F ¼ 3.4, p ¼ 0.07), weight (F ¼ 1.0, p ¼ 0.32) and height (F ¼ 0.5, p ¼ 0.49). Results of a chi-square test showed that both genders (p ¼ 0.63) and all 3 activity levels (p ¼ 0.75) were represented in similar proportions across the 2 groups. 3.1. Multifidus muscle size

Fig. 2. Ultrasound image of the multifidus muscle in parasagittal section at rest and on contraction using a split-screen. Relaxed multifidus muscle thickness ¼ 2.68 cm, contracted value ¼ 3.20 cm.

Results of the ANCOVA showed a significantly smaller CSA of the multifidus muscle for the CLBP group compared to the unimpaired group at the L5 vertebral level (F ¼ 29.1, p ¼ 0.001), and slightly larger size at L2 (F ¼ 5.8, p ¼ 0.047). A small but significant effect (mean net difference of 0.17 cm2) was found for multifidus muscle ‘asymmetry’ at each vertebral level (p < 0.05) but this was similar for both groups (p > 0.05). There were no significant effects for

T.L. Wallwork et al. / Manual Therapy 14 (2009) 496–500

‘gender’ at the L4 and L5 vertebral levels, but male subjects had significantly larger mutifidus muscles than females at the L3 vertebral level (difference of 0.9 cm2) and at L2 (difference of 0.15 cm2). ‘Activity level’ was significantly associated with multifidus size at L3 (F ¼ 5.9, p ¼ 0.03), L4 (F ¼ 11.9, p ¼ 0.006) and L5 (F ¼ 5.4, p ¼ 0.04), but this was similar for both groups (no interaction effect, p > 0.05). There were no significant higher-order interactions. Table 2 shows the estimated marginal means of multifidus CSA for the CLBP and unimpaired groups at each level. Table 3 shows the estimated marginal means of multifidus size for the 3 levels of activity. 3.2. Multifidus muscle thickness and contraction Table 4 shows the thickness measurements of the multifidus muscle for rest and contracted conditions, averaged across left and right sides. Analysis of these data was based on calculation of the percent contraction from rest. Results of the ANCOVA showed a significantly smaller percent thickness contraction for the CLBP group compared to the unimpaired group at the L5 vertebral level (F ¼ 6.6, p ¼ 0.02), but not at other vertebral levels (p > 0.05). A small but significant effect (mean net difference of 2.7%) was found for contraction ‘asymmetry’ at each vertebral level (p < 0.05) but this was similar for both groups (i.e. there was no significant interaction between ‘asymmetry’ and ‘group’, p > 0.05). There were no significant effects for the variables of ‘activity level’ or ‘gender’ and no significant higher-order interactions. Table 5 shows the estimated marginal means (and standard deviations) for the CLBP and unimpaired groups at each level. 4. Discussion 4.1. Multifidus muscle size Results from the current investigation showed a specific and localized pattern of atrophy of the multifidus muscles in the presence of chronic LBP. In this study, atrophy was greatest at the L5 vertebral level, and there was a trend towards significance at the L4 vertebral level. Several previous imaging studies have reported evidence of multifidus muscle atrophy in patients with LBP. Researchers have investigated post-operative patients (Sihvonen et al., 1993), patients with acute/subacute LBP (Hides et al., 1994, 1996) and patients with chronic LBP (Kader et al., 2000; Danneels et al., 2000, 2001; Barker et al., 2004; Hides et al., 2008a,b). In agreement with these previous studies, the pattern of atrophy seen in the chronic LBP patients investigated appeared to be specific and localized in nature. 4.2. Multifidus muscle thickness and contraction The results of this study suggest that the neuromotor control of multifidus was altered at the L5 vertebral in patients with CLBP. Subjects with CLBP were less able than healthy subjects to voluntarily contract the multifidus muscle at the same vertebral level where atrophy was present. The clinical relevance of this finding is

Table 2 Group differences (marginal meansa and standard deviations) in multifidus muscle size (cm2) across vertebral levels L2–L5. L2b

Group 1 (unimpaired) Group 2 (CLBP) a b

L3

Mean

(SD)

Mean

(SD)

Mean

(SD)

Mean

(SD)

1.94 2.40

(0.9) (0.9)

3.09 3.02

(1.3) (1.4)

4.61 3.47

(1.0) (1.1)

5.56 3.81

(1.1) (1.2)

Adjusted for the covariates of age, weight and height. Statistically significant difference at p < 0.05.

Table 3 Activity level differences (marginal meansa and standard deviations) in multifidus size (cm2) across vertebral levels L2–L5. Physical activity per week L2

L4b

L3

Mean (SD) Level 1 activity (<1.5 h) Level 2 activity (1.5–3 h) Level 3 activity (>3 h)

1.78 2.27 2.44

Mean (SD)

(1.1) 2.24 (0.9) 3.09 (1.3) 3.82

L5b

Mean (SD)

(1.5) 2.95 (1.3) 4.29 (1.9) 4.87

Mean (SD)

(1.2) 3.96 (1.1) 5.26 (1.5) 4.84

(1.3) (1.2) (1.7)

a

Adjusted for the covariates of age, weight and height. Statistically significant difference between Activity Level 1 versus Level 2 and Level 3, based on post-hoc contrasts with Bonferroni correction. b

that rehabilitation may need to be specific in order to target localized impairments in motor control. Clinical approaches targeting motor control of muscles including the multifidus, transversus abdominis and pelvic floor have been shown to be effective in randomized clinical trials (RCTs) (Hides et al., 1996; O’Sullivan et al., 1997; Stuge et al., 2004; Goldby et al., 2006). A RCT conducted on subjects with first episode acute LBP provided the first evidence of a localized, segmental impairment in the CSA of the multifidus muscle (Hides et al., 1996). Similar to the findings of the current study, it was reported that subjects could not voluntarily contract the multifidus muscle at the vertebral level where the atrophy of the muscle was observed. A tailored exercise approach targeting the impaired muscle restored muscle size and resulted in lower recurrence rates of LBP (Hides et al., 2001). Ultrasound imaging was used to provide feedback of multifidus muscle contraction (Hides et al., 1996; Van et al., 2006). A motor control approach was also recently successfully employed in a study involving elite cricketers with LBP (Hides et al., 2008b). Results showed that the CSAs of the multifidus muscles at the L5 vertebral level increased with training and these changes were commensurate with a 50% decrease in mean reported pain levels. The finding that subjects who have LBP are less able to contract the multifidus has also been reported in a laboratory study. The effect of pain on multifidus muscle function was demonstrated experimentally using a model of induced pain (Kiesel et al., 2008). Increases in multifidus muscle thickness during arm lifting tasks were significantly reduced by pain in response to injection of saline into the erector spinae muscles. While Kiesel et al. (2008) did not examine voluntary contractions of the multifidus muscle, the findings may support the current clinical practice of using physiotherapeutic modalities to decrease pain prior to commencing rehabilitation of the multifidus muscle, and performance of voluntary multifidus contractions in pain-free positions (Hides et al., 1996). 4.2.1. Limitations and future directions This study has some limitations. The study sample size is small, though comparable with other similar investigations (Hides et al., 1996; Danneels et al., 2000; Van et al., 2006). Thickness measures of the multifidus muscle were obtained in 30 subjects, where CSA measures were only obtained in 22 participants. While this is not ideal, the results from this study in relation to CSA of the multifidus are in line with previous reports (e.g. Danneels et al., 2000; Hides Table 4 Thickness (means and standard deviations) of the multifidus muscle in the rest and contracted state across vertebral levels L2–L5 (mm). L2 Mean

L5b

L4

499

Group 1 (unimpaired) Rest 29 Contracted 31 Group 2 (CLBP) Rest 27.6 Contracted 28.9

L3

L4

L5

(SD)

Mean

(SD)

Mean

(SD)

Mean

(SD)

(5.2) (5.3)

33 34.7

(5.0) (4.8)

35.9 37.7

(5.3) (4.9)

35.9 38.1

(4.8) (4.8)

(4.7) (4.7)

30.5 31.9

(4.5) (5.2)

33.6 34.6

(5.3) (5.4)

33.9 35.0

(5.5) (5.6)

500

T.L. Wallwork et al. / Manual Therapy 14 (2009) 496–500

Table 5 Group differences (marginal meansa and standard deviations) in percentage thickness contraction across vertebral levels L2–L5. L2

Group 1 (unimpaired) Group 2 (CLBP) a b

L3

L5b

L4

Mean

(SD)

Mean

(SD)

Mean

(SD)

Mean

(SD)

6.93 4.10

(6.7) (7.4)

5.20 4.22

(7.0) (7.8)

5.15 2.93

(5.9) (6.5)

6.29 3.05

(6.5) (7.2)

Adjusted for the covariates of age, weight and height. Statistically significant difference at p < 0.05.

et al., 2008a). Furthermore, while CSA of the multifidus muscle was measured, consistency changes in the muscle (fatty deposits or fibrous/connective tissue infiltration) were not assessed. Future studies, especially those employing imaging techniques such as CT scanning and magnetic resonance imaging, could assess this. The main new contribution of this paper is the data pertaining to contraction of the multifidus muscle. The measure could be used in future studies to compare the effectiveness of retraining motor control with and without feedback by ultrasound imaging in subjects with CLBP. 4.2.2. Conclusion Patients with CLBP had significantly smaller multifidus muscles than healthy, asymptomatic subjects at the lowest vertebral level of the lumbar spine. Patients with CLBP also had greater difficulty performing a voluntary isometric multifidus contraction at the same vertebral level. The results of this study support previous findings that the pattern of multifidus muscle atrophy in CLBP patients is localized rather than generalized but goes further in also ascertaining a reduced ability to voluntarily contract the atrophied muscle. These findings lend support to the use of specific muscle retraining programmes for patients with CLBP. Acknowledgements The authors wish to thank the subjects studied, the staff at the UQ/Mater Back Stability Clinic, and its Director, Ms Linda Blackwell, for their assistance, and Ms Margot Wilkes, for assistance with this manuscript. References Barker KL, Shamley DR, Jackson D. Changes in the cross-sectional area of multifidus and psoas in patients with unilateral back pain: the relationship to pain and disability. Spine 2004;29:E515–9. Danneels L, Vanderstraeten G, Cambier D, Witvrouw E, De Cuyper H. CT imaging of trunk muscles in chronic low back pain patients and healthy control subjects. European Spine Journal 2000;9(4):266–72.

Danneels L, Van der Straeten G, Cambier D, Witvrouw E, De Cuyper H. The effects of three different training modalities on the cross-sectional area of the lumbar multifidus. British Journal of Sports Medicine 2001;35:186–94. Goldby LJ, Moore AP, Doust J, Trew M. A randomized controlled trial investigating the efficiency of musculoskeletal physiotherapy on chronic low back disorder. Spine 2006;31(10):1083–93. Hides JA, Cooper DH, Stokes MJ. Diagnostic ultrasound imaging for measurement of the lumbar multifidus in normal young adults. Physiotherapy Theory and Practice 1992;8:19–26. Hides J, Gilmore C, Stanton W, Bohlscheid E. Multifidus size and symmetry among chronic LBP and healthy asymptomatic subject. Manual Therapy 2008a;13(1): 43–9. Hides J, Jull G, Richardson C. Long-term effects of specific stabilizing exercises for first-episode low back pain. Spine 2001;26(11):E243–8. Hides J, Richardson C, Jull G. Multifidus recovery is not automatic after resolution of acute, first-episode low back pain. Spine 1996;21:2763–9. Hides J, Richardson C, Jull G. Magnetic resonance imaging and ultrasonongraphy of the lumbar multifidus muscle: comparison of two different modalities. Spine 1995;20(1):54–8. Hides J, Stokes M, Saide M, Jull G, Cooper D. Evidence of lumbar multifidus muscle wasting ipsilateral to symptoms in patients with acute/subacute low back pain. Spine 1994;19(2):165–72. Hides JA, Wilson S, Stanton W, McMahon S, Sims K, Richardson C. The effect of stabilization training on multifidus muscle size among young elite cricketers. Journal of Orthopaedic and Sports Physical Therapy 2008b;38(3):101–8. International Association for the Study of Pain. Available at: http://www.iasp-pain. org//AM/Template.cfm?Section¼Home; 2008 [accessed April 1, 2008]. Kader D, Wardlaw D, Smith F. Correlation between the MRI changes in the lumbar multifidus muscles and leg pain. Clinical Radiology 2000;55:145–9. Kiesel K, Uhl T, Underwood F, Rodd D, Nitz A. Measurement of lumbar multifidus muscle contraction with rehabilitative ultrasound imaging. Manual Therapy 2007;12(2):161–6. Kiesel KB, Uhl TL, Underwood FB, Nitz A. Rehabilitative ultrasound measurement t of select trunk muscle activation during induced pain. Manual Therapy 2008; 13:132–8. O’Sullivan PB, Twomey LT, Allison GT. Evaluation of specific stabilizing exercise in the treatment of chronic low back pain with radiologic diagnosis of spondylolysis or spondylolisthesis. Spine 1997;22:2959–67. Pressler JF, Heiss GD, Buford JA, Chidley JV. Between day repeatability and symmetry of multifidus cross-sectional area measured using ultrasound imaging. Journal of Orthopaedic and Sports Physical Therapy 2006;36(1):10–8. Sihvonen T, Herno A, Paljarvi L, Airaksinen O, Partanen J, Tapaninaho A. Local denervation atrophy of paraspinal muscles in postoperative failed back syndrome. Spine 1993;18:575–81. Stokes M, Rankin G, Newham DJ. Ultrasound imaging of lumbar multifidus: normal reference ranges for measurements and practical guidance on the technique. Manual Therapy 2005;10:116–26. Stokes M, Hides J, Elliott J, Kiesel K, Hodges P. Rehabilitative ultrasound imaging of the posterior paraspinal muscles. Journal of Orthopaedic and Sports Physical Therapy 2007;37(10):581–95. Stuge B, Veierød MB, Laerum E, Vøllestad N. The efficacy of a treatment program focusing on specific stabilizing exercises for pelvic girdle pain after pregnancy: a two year follow up of a randomized clinical trial. Spine 2004;29(10): E197–203. Van K, Hides JA, Richardson CA. The use of real-time ultrasound imaging for biofeedback of lumbar multifidus muscle contraction in healthy subjects. Journal of Orthopaedic and Sports Physical Therapy 2006;36(12):920–5. Wallwork T, Hides J, Stanton W. Intrarater and interrater reliability of assessment of lumbar multifidus muscle thickness using rehabilitative ultrasound imaging. Journal of Orthopaedic and Sports Physical Therapy 2007;37(10):608–12.

Manual Therapy 14 (2009) 501–507

Contents lists available at ScienceDirect

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

Original Article

Quantitative application of transverse friction massage and its neurological effects on flexor carpi radialis Hsin-Min Lee a, *, Shyi-Kuen Wu b, Jia-Yuan You a a b

Department of Physical Therapy, I-Shou University, Kaohsiung, Taiwan, ROC Department of Physical Therapy, HungKuang University, Taichung, Taiwan, ROC

a r t i c l e i n f o

a b s t r a c t

Article history: Received 17 November 2007 Received in revised form 24 September 2008 Accepted 26 September 2008

The purpose of the study was to determine the effects of transverse friction massage (TFM) on flexor carpi radialis (FCR) motoneuron (MN) pool excitability. Twenty-eight healthy subjects were randomly assigned into massage and control groups. Pre- vs postTFM H-reflex data were collected. Controls received a rest period instead of massage. Massage dose was standardized by a novel electronic method which recorded the massage rate, momentary pressure and total cumulative pressure (energy). Two-way ANOVA of H/M ratios derived from maximal amplitudes of Hoffman reflexes (Hmax) and motor responses (Mmax) was used to analyze neurological effects and group differences. Analysis of pressure/time curve data showed: mean massage rate was 0.501  0.005 Hz; mean duration of massage sessions was 184.6  26.4 s; mean peak pressure was 4.990  1.006 psi. Hmax/Mmax ratios declined from 14.3% to 10.3% for massage (P < 0.01) but showed no change for controls (P > 0.05). In conclusion a novel quantitative approach to the study of massage has been demonstrated while testing the effects of TFM on FCR MN pool excitability. TFM appears to reduce MN pool excitability. The novel method of quantifying massage permits more rigorous testing of client-centered massage in future research. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Massage Quantification H-reflex

1. Introduction Muscle massage is a common modality used by a variety of practitioners in a variety of settings. Most clinical practices combine several massage techniques over a relatively large area of the body to resolve various problems such as muscle soreness (Farr et al., 2002) or limb oedema (Braverman and Schulman, 1999). Unlike other techniques, friction massage is usually applied alone, i.e. without co-application of other techniques, directly to the lesion site and transversely to the direction of the muscle fibers. Transverse friction massage (TFM) is commonly used for localized muscle injuries to relieve pain (Hammer, 1993; Brosseau et al., 2002). Recent meta-analysis research has reported that massage can reduce pain, anxiety, blood pressure and heart rate (Moyer et al., 2004). Theories regarding the mechanisms of massage’s benefits include activation of the parasympathetic nervous system for release of endorphins and serotonin, reduction of fibrosis or scar tissue and improved sleep (Field, 1998; Weerapong et al., 2005). Among the most widely accepted mechanisms for the analgesic effects of massage is blocked nociception, i.e. the inhibition of pain * Corresponding author. Tel.: þ886 7 6151100x7562; fax: þ886 7 6155150. E-mail address: [email protected] (H.-M. Lee). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.09.005

signals in the spinal circuits via the gate-control mechanism (Melzack and Wall, 1996; Field, 2000). Alternately, pain (reduction) may be related to (reduced) muscle tension, as indicated by the close correlation between paraspinal muscle spasm and concurrent pain symptoms in patients with low back pain (Zhu et al., 2000). Neurological mechanisms of increased muscle tension involve spinal motoneuron (MN) pool activity, which is known to change due to intervention of physical modalities such as stretching, cryotherapy and transcutaneous electric nerve stimulation (Vujnovich and Dawson, 1994; Hopkins et al., 2002). MN pool excitability is often studied via Hoffman reflex (H-reflex) testing which measures muscle EMG response to mild electrical shock of the nerve. H-reflex tests have shown that slow-rate petrissage massage over the triceps surae produces short-term reduction of spinal MN pool activity (Morelli et al., 1990, 1991), implying that muscle massage reduces leg muscle tension (Weerapong et al., 2005). But as yet no study has investigated the effect of deep-pressure massage such as TFM on the H-reflex and MN pool excitability of muscles in the upper extremities. The effects of traditional muscle massage are achieved through the transfer of mechanical energy to the body of the patient. For the sake of quantification of TFM application, the energy transmitted can be characterized by factors such as total duration, frequency of

502

H.-M. Lee et al. / Manual Therapy 14 (2009) 501–507

repetitive contact, maximum and average pressures during massage, etc. Previous studies which attempted evaluating massage effects (Dishman and Bulbulian, 2001; Brooks et al., 2005) have focused on time factors involved in the mechanical massage contact, typically employing a regular massage rate (such as 0.5 Hz) and a specific duration of massage (such as 3 min). Few studies have paid attention to the issue of momentary pressure during massage or the total transfer of energy over the time course of the massage session. However, Goldberg et al. (1992) use H-reflex to compare the effects of petrissage massage with two qualitative levels of pressure and found that petrissage with higher pressure evoked greater inhibitory effects on MN pool activity. It is to be noted, however, that Goldberg’s results and interpretations of neurological effects were based on a simple qualitative difference of massage intensity. Without accurate quantitative measurement of pressure intensity and other factors during massage, the interpretation of the relationship between massage and corresponding effects carry less conviction. Furthermore, in clinical practice, the selected massage intensity for each individual varies. Choice of massage intensity traditionally is based on the clinician’s subjective experience and his/her immediate qualitative evaluation of factors such as the flexibility of the area to be treated. Hence we recognize the need for a reliable real-time quantitative approach to measure and record the relevant massage parameters over time. Such an approach is presented below for demonstration and evaluation in a series of TFM experimental applications. Massage variables such as the momentary pressure and the total mechanical energy applied to the patient are accurately obtained. H-reflex data is employed to determine whether TFM reduces MN excitability in the muscles of the upper extremities in healthy young volunteers. The following focuses on the flexor carpi radialis (FCR) which is a primary muscle in the forearm. In the literature, the FCR is mentioned in relation to the carpal tunnel syndrome (CTS) which results in pain, paresthesia and muscle weakness in hand (Muggleton et al., 1999; Racasan and Dubert, 2005). Recently, Jaberzadeh and Scutter (2006) reported increased excitability (i.e. larger H-reflex amplitude) of the FCR MN pool in chronic CTS subjects. Our interest in TFM is related to reports of CTS improvement in response to TFM applied to the FCR. The question of whether such improvement is merely psychological or whether it has an actual organic mechanism, and the nature of any such mechanism, is of interest. Therefore this study investigates quantitatively a set of basic TFM parameters which include contact area, direction, rate, maximum momentary pressure and total cumulative transmitted energy. TFM is applied to the FCR and the resulting change of MN pool excitability is evaluated by the Hoffmann reflex vs motor response (H/M) ratio, as commonly used in the H-reflex literature. The H/M ratio data is collected immediately before and after massage. A similar test procedure is performed for a control group who receive no massage and merely rested. 2. Materials and methods

Table 1 Summary of subjects’ basic data. Group

No. of Age (y/o): Gender Height (cm): Weight (kg): Test side subject mean (SD) (male/female) mean (SD) mean (SD) (right/left)

Massage 14 Control 14

21.6 (0.6) 21.2 (1.0)

(8/6) (7/7)

164.8 (8.1) 165.6 (7.3)

60.2 (10.3) 58.8 (10.4)

(13/1) (12/2)

immediately before and immediately after receiving TFM application of approximately 3 min (duration based on cumulative applied pressure as discussed below). Subjects in the control group underwent identical H-reflex testing under identical conditions, but with a 3-min rest period substituting for the massage period. Both test and control subjects were supine on a bed with the dominant arm outstretched sideways at a 45 angle to the body (45 shoulder abduction and full supination of forearm) throughout the experimental session. Subjects were instructed to keep fully relaxed, engage in no conversation and make no positional changes of head–arm–trunk segments during the experimental session, so as to minimize extraneous factors which could alter the H-reflex amplitude (Zehr, 2002). 2.2. H-reflex measurement The setup for H-reflex measurement is illustrated in Fig. 1. The recording techniques employed for H-reflex generally followed the methodologies of two recent studies (Dishman and Burke, 2003; Christie et al., 2005). Briefly, a monopolar Ag/AgCl recording electrode (Grass F-E9, Astro-Med) was placed over the muscle belly of the FCR at a point 1/3 of the distance from the medial epicondyle to the radial styloid process (Fig. 1(a)). An identical Ag/AgCl reference electrode was placed over the tendon of the FCR near the wrist joint. A self-adhesive ground electrode was secured to the forearm beside the recording electrode. The EMG signal was amplified with band-pass filtering (3–1000 Hz) and a gain of 200 (Grass P511, Astro-Med) (Fig. 1(b)). The filtered EMG signal was sent to a PC via a 16-bit analog-to-digital converter (sampling rate: 2000 Hz). Stimulation electrodes were secured over the path of the median nerve through the cubital area, just beside the biceps tendon and above the cubital crease. The cathode was placed distal to the anode so as to leave a distance of 2.5 cm between the two electrodes. Both electrodes were connected in series with an isolation unit (Grass SIU5, Astro-Med) and a stimulator (Grass S88, Astro-Med) that delivered a square-wave pulse (duration: 0.5 ms) every 10 s to prevent reflex attenuation (Bischoff, 2002). Before the first H-reflex measurement session, H/M recruitment curves were obtained to determine the stimulation amplitude to evoke maximal H-reflex (Hmax) and maximal M response (Mmax). The Mmax stimulation amplitude was defined as when there was no further increase in M amplitude for three successive 5-V increments (Dishman and Burke, 2003). In each H-reflex measurement session, 10 Hmax and 10 Mmax were recorded for later analysis of the effects of TFM (massage group) or rest (control group).

2.1. Subjects and experimental design 2.3. Application of massage Twenty-eight healthy subjects aged 20–23 y/o were enrolled from a college student population and randomly assigned to a massage group (n ¼ 14) or a control group (n ¼ 14). All subjects were asymptomatic over cervical, shoulder and upper extremity regions during the test period. The enrolled subjects signed informed consent agreements for the protocols, which were approved by the local Ethics Committee. The basic characteristics of the subject population are summarized in Table 1. The experiment was divided into three phases: two separate H-reflex measurement sessions and one massage (or rest) session. Subjects in the massage group underwent H-reflex measurement

TFM was applied over a 3 cm  5 cm rectangular area covering the FCR muscle belly in the middle part of forearm (Fig. 1(a) and Fig. 2(a)). An ultra-thin (0.1 cm) flexible pressure sensor (ConTacts C500, Pressure Profile Systems) was mounted on the thumb pad of the physical therapist for electronic real-time recording of massage pressure during TFM (Fig. 2(b)). The therapist applied transverse (medial-to-lateral) TFM over the marked area with flexion movements of the interphalangeal joint. The rate of massage was 0.5 Hz and was timed by a metronome-type beep program from a computer. Selection of the 0.5 HZ massage rate was based on

H.-M. Lee et al. / Manual Therapy 14 (2009) 501–507

503

a Flexor carpi radialis ES +

G

-

Isolation unit

G2

G1 EMG recording

A/D Converter EMG amplifier

Electrical stimulation

Triggering

Stimulator

b

PC

c

Fig. 1. Setup of H-reflex measurement. (a) Recording electrodes (G1, G2), ground electrode (G) and stimulator electrodes (ES) are secured on the surface of FCR and over the median nerve. (b) EMG signals are amplified and sent to a data acquisition subsystem in a PC. When the stimulator sends a 0.5 ms pulse to elicit the H-reflex, a triggering signal simultaneously initiates a 100 ms data acquisition episode. (c) The data acquisition subsystem includes a 16-bit A/D converter and a PC for on-line monitoring of the H-reflex and establishing the H/M recruitment curve to determine the stimulation amplitude of the maximal H-reflex and M response.

earlier work (Sullivan et al., 1991; Benjamin and Tappan, 1998) which suggested that 0.5 Hz was optimal for relaxation, whereas faster rates tended to stimulate the client. In this present study we control the total cumulative mechanical energy applied during the TFM session. Further, we provide an over-time record of the momentary pressure applied by the therapist during TFM performance. Total cumulative massage energy is treated as the sum of the momentary pressures generated during duration of massage, as defined by the equation in Fig. 3(a). While not a true measure of transferred energy, the presented method allowed a simple monitoring program embedded in a PC to alert the therapist when a target value was reached, said value representing the summation of pressure over time. Thus the therapist could adjust the pressure applied to the subject according to the therapist’s experience and the subject’s condition, but the total energy transferred from therapist to the subject was approximately and

quantifiably equal in all cases. This, to our knowledge, is the first time such quantifiable pressure data was used as a control variable in a study of therapeutic massage. The target value for cumulative pressure chosen here was 11,500 psi, which was found from our experience to represent approximately 3 min of TFM on the FCR. A LabVIEW program written by our group monitored the momentary and cumulative pressures, calculating continuously the sum of the pressure signals and comparing the latest sum to the preset value (11,500 psi). When the pressure summation reached or exceeded the target value (Fig. 3(b)), the program immediately caused the computer to emit an auditory signal, telling the therapist to stop. Thus the total cumulative energy of each massage was approximately equal, while the duration of each massage session was allowed to vary in response to variations in the momentary pressure as applied by the physical therapist during different massage sessions.

a

b Finger glove

Ultra-thin pressure sensor

Area of massage

Forearm

Fig. 2. Application and pressure measurement of TFM. (a) A 3  5 cm2 rectangular area is marked on the middle segment of forearm to guide massage application. (b) An ultra-thin pressure sensor is mounted on the thumb pad and secured by a finger glove to record pressure data. The pressure sensor pad is 2.5  2.5 cm2 and covers the contact area of the thumb pad during massage.

504

H.-M. Lee et al. / Manual Therapy 14 (2009) 501–507

Diagram of Massage Energy Control

Summation of pressure signal

Application of massage

No

N

Pressure signal

Σ

P[n]

n=1

Is summation of pressure signal ≥11500 PSI?

P[n] denoted the pressure measured at sample number n, n=1,2,3,…N.

Yes Stop massage

a

b

Fig. 3. Schematic diagram to explain the control of massage energy. (a) The data acquisition program calculates the summation of pressure from sample points 1 to N every one second, sample rate 100 Hz. (b) Once the cumulative amplitude reaches or exceeds the preset target value of 11,500 psi, an audio signal is triggered and the therapist stops massage.

2.4. Data analysis Massage parameters including the rate, duration and average amplitude were calculated for each subject. The rate of a massage session was used to check the TFM rhythm and could be derived from the peak spectrum frequency of pressure signals. The duration of massage was defined as the time to reach the target value of massage energy, as seen in Fig. 3. The peak amplitude of each individual TFM contact was collected and could be calculated from the pressure signals. The average peak pressure of a TFM session was derived by averaging all the peak pressures during a massage session. Average TFM amplitude was used to represent the TFM intensity. The Hmax/Mmax ratio reflected the proportion of the MN pool recruited by the Ia afferents and is commonly regarded as an index of MN pool excitability (Fisher, 1992; Bischoff, 2002). Peak-to-peak amplitudes of 10 Hmax waves and 10 Mmax waves were determined in each measurement session. The Hmax/Mmax ratio was calculated by dividing the average Hmax amplitude by the average Mmax amplitude. 2.5. Statistical analysis Paired t-test of Mmax between two sessions was used to test the reliability of the electrode attachments before and after formal experimental sessions in both massage and control groups. Mixed two-way ANOVA of the Hmax/Mmax ratios between two measurement sessions (repeated measures) and two subject groups was performed to determine the main factors of MN excitability change. Furthermore, post hoc tests were applied to compare group differences in the Hmax/Mmax ratio at baseline and to compare session differences in the Hmax/Mmax ratios in both massage and control groups. P values below 0.01 were considered statistically significant. 3. Results Fig. 4(a) shows a typical data capture of momentary pressurevs-time during massage as detected by the pressure sensor mounted on the thumb of the physical therapist. Fig. 4(b) zooms in on the time scale and shows the detailed pressure sensor data over the 90th–120th s. In this time period fifteen TFM waveforms can be

seen. For each cycle of the pressure curve, this closer zoom shows the relatively slow attack and the relatively fast decay of the pressure curve envelope (waveform). Such stroke-by-stroke quantitative detail has never before been seen in a study of massage. It can also be seen from the captured data that the waveforms of the contact signals during TFM are regular in shape, with 5 repetitions within 10 s as is expected of a 0.5 Hz massage rate. The observed time ratio of pressure (on and off) is about 1 (1 s/1 s). The period of zero pressure between each peak is the TFM release time, during which time the therapist resets his thumb for the next pressure stroke, i.e. during this time the pressure sensor is out of contact with the body of the client. In this example the average rate is 0.503 Hz, as determined by locating the peak value in the frequency spectrum (Fig. 4(c)). The captured peak amplitudes for this example range from 3.027 psi to 5.409 psi, with an average amplitude of 4.415 psi. The total duration of this example is 185 s. The combined massage parameters for all subjects are listed in Table 2. The Table 2 results show that the rate of massage was well controlled by the auditory guidance of the digital metronome, with an average rate of 0.501 Hz and a range of 0.491–0.514 Hz. However the duration of massage varied significantly from session to session, because massage duration depended on the time required to reach a cumulative pressure energy and the peak pressure during TFM was not regulated except by the subjective evaluation of the physical therapist. In consequence, the combined mean duration of massage of all the subjects was 184.6  26.4 s, reflecting a combined peak massage pressure which ranged from 3.452 to 7.189 psi and had a combined mean of 4.990 psi. To exclude the possibility of the altered electrode attachments during the massage or rest periods, the results of paired t-test of Mmax amplitudes are presented. In the massage group, the Mmax amplitudes are 5.660  1.232 V and 5.638  1.240 V for H-reflex recording sessions one and two, respectively. Mmax amplitudes in the control group are 5.661 1.028 V and 5.681 1.087 V for H-reflex sessions one and two, respectively. There are no significant differences in the Mmax amplitudes between the two sessions in either the massage (P ¼ 0.463) or control groups (P ¼ 0.524). To evaluate the effect of TFM massage on FCR MN pool excitability, the amplified H-reflex and M response signals from one typical example of our subjects are illustrated in Fig. 5. The average Hmax amplitudes decrease from 1.395 V (1.206–1.515 V) to 1.198 V (1.100–1.377 V) after TFM. The average Mmax amplitudes are

H.-M. Lee et al. / Manual Therapy 14 (2009) 501–507

a

6

b

6 5

Pressure (PSI)

5

Pressure (PSI)

505

4 3 2 1

4 3 2 1

0

0 0

20

40

60

80

90

100 120 140 160 180

95

100

Time (sec)

c

110

115

120

Parameters of massage

3000

d

Spectrum of applied massage

2500

Amplitude (Volt2)

105

Time (sec)

2000

Rate: 0.503 Hz

Duration: 185 sec

1500

Mean amplitude: 4.415 PSI (Range:

1000

3.027 ~ 5.409 PSI)

500 0

0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

1

Frequency (Hz) Fig. 4. Typical captured pressure/time curves and derived data. (a) Pressure signals during a massage session. (b) Zoomed view of the same curve (90th–120th s) shows pressure detail such as attack and release curves. (c) The average massage rate and variance can be verified by the frequency spectrum of the pressure signal. (d) The analyzed parameters of the pressure/time curve are summarized numerically.

5.140 V and 5.166 V before and after TFM, respectively. The derived Hmax/Mmax ratio declines from 27.1% to 23.2% across the massage session. The Hmax/Mmax ratios of all subjects are pooled to determine by two-way ANOVA the inter-group and the intersession significance of the Hmax/Mmax differences (Table 3). The results show that only the session factor was significant for Hmax/ Mmax ratios (P < 0.01). The post hoc test of group difference further shows that the Hmax/Mmax ratios are not significantly different in sessions one (P ¼ 0.925). This indicates that no group differences of MN pool excitability existed between the massage and control groups at baseline. Importantly, the session difference in massage group showed a significant decrease in Hmax/Mmax ratios (14.3  7.5–10.3  6.4%, P < 0.01, Table 3), contrasting to the zero statistical difference in control group (14.5  7.5–14.5  6.7%, P > 0.05). These observations confirm that decreased Hmax/Mmax ratios are only found in the massage group. 4. Discussion A quantitative approach to the study of massage has investigated the effects of TFM on the excitability of the FCR MN pool. TFM parameters including massage area and direction were Table 2 Parameters of massage application: mean, standard deviation (SD), minima (Min) and maxima (Max) of rate, duration and average amplitude of TFM.

Mean SD Min Max

Rate (Hz)

Duration (s)

Pressure (psi)

0.501 0.005 0.491 0.514

184.6 26.4 140.0 229.0

4.990 1.006 3.452 7.189

standardized. Mounting a thin pressure sensor on the thumb of the physical therapist allowed detailed on-online monitoring of the momentary massage pressure-vs-time curves for each complete massage session. From this data it was possible to calculate in realtime the massage rate, momentary pressure and cumulative massage energy. Processing was rapid enough for use of an alarm program in the monitoring PC so that the massage session could be terminated when a preset energy was reached, thereby guaranteeing equal energy administered during each individual massage session. TFM was applied within a localized 3  5 cm2 area concentrated on the FCR muscle belly. One experimental concern was that active TFM application might disturb the recording electrodes (Fig. 1(a)), but this scenario was excluded by the continuing stability of the Mmax amplitudes before and after TFM (P > 0.05). Massage rate was controlled by the physical therapist listening to and matching the auditory signal of a digital metronome. As shown in Fig. 4(c), this technique produced good massage rate control, as verified by the narrow bandwidth of the TFM spectrum (0.491–0.514 Hz, Table 2). Prior work has shown that massage rate can be a determining factor with regard to whether the massage session produces relaxation or stimulation effects (De Domenico and Wood, 1997). For induction of relaxation effects we followed the 0.5 Hz massage rate found in several prior studies (Sullivan et al., 1991; Benjamin and Tappan, 1998; Morelli et al., 1999). Faster rates are generally believed to produce stimulating rather than relaxing effects via either neurophysiological or psychological pathways (De Domenico and Wood, 1997). For future work we may investigate the effects of TFM on the H-reflex and MN pool activity for a range of different massage speeds and with different total cumulative energies. Such research was impossible prior to the presented quantitative methodology.

506

H.-M. Lee et al. / Manual Therapy 14 (2009) 501–507

a

Before massage

5.097

5.121

5.129

5.056

5.120

2 0 -2

5.113

5.170

2 0 -2

5.191

5.207

2 0 -2

5.207

5.220

2 0 -2

5.167

5.174

2 0 -2

5.128

5.178

2 0 -2

5.153

5.185

2 0 -2

5.174

5.177

1.153

2 0 -2

1.294

1.254

2 0 -2

2 0 -2

1.296

1.100

2 0 -2

2 0 -2

1.506

1.111

1.513

1.195

2 0 -2

1.480

1.189

2 0 -2

1.356

1.132

2 0 -2

1.404

1.174

2 0 -2

1.377

1.292

2 0 -2

1.206

1.377

0

50

100 0

50

100

After massage

5.086

1.515

2 0 -2

Before massage

2 0 -2

Amplitude (Volt)

Amplitude (Volt)

2 0 -2

b

After massage

0

50

1000

50

100

Time (ms)

Time (ms)

Fig. 5. The inhibitory effects of massage on MN pool excitability of the FCR from one typical example. Raw data of maximal H-reflexes and M responses before and after massage are shown in (a) and (b), respectively. The peak-to-peak amplitudes are measured over a time span of 13–33 ms (H-reflexes) and 6–26 ms (M responses), as marked by the vertical doted lines. The derived amplitudes for each H-reflex and M response are also shown in the figure.

In contrast to the specific pre-assigned and carefully-timed massage durations of previous studies (Goldberg et al., 1992; Brooks et al., 2005), the present study did not require a specific duration of massage, but rather required pre-assigned and carefully-measured total cumulative mechanical energy applied during massage. An alternative research variable that could be monitored by the presented system is the momentary pressure. For example, an auditory alarm might trigger each time the momentary pressure reached a preset maximum. The accurate online quantification offered by the presented system makes it a powerful research tool for medical investigation of massage-type physical therapy. In the Table 3 Summarized data and statistical results of Hmax/Mmax ratios for two groups and two measurement sessions. Hmax/Mmax ratio (%): mean (SD) Group

Session 1

Session 2

P value

Massage (n ¼ 14) Control (n ¼ 14)

14.3 (7.5) 14.5 (7.5)

10.3 (6.4) 14.5 (6.7)

0.002* 0.907

Two-way ANOVA

Group factor Session factor

0.404 0.003*

Hmax/Mmax ratios of all subjects were summarized as means and standard deviations (SDs). P values marked with asterisk indicate significant decrease was found by the post hoc test. In ANOVA analysis, P values marked with asterisk indicate a significant factor (P < 0.01) in the main factor test.

present study, however, momentary pressure was solely under the control of the physical therapist. Accordingly, some patients experienced lower average peak pressure than others. In consequence, the total duration of massage was longer since the duration of the massage was determined by monitoring the total amount of massage pressure. The pooled statistics show that TFM duration varied from 140 to 229 s as a result of the differing momentary pressures applied by the therapist. Our therapist was not consciously aware that his massage pressures varied so greatly, only that he had adjusted the massage pressure to suit the muscular flexibility of the massage area. This underscores the need for greater quantification during this type of investigation. Further, measuring the pressure between therapist’s thumb and subject’s forearm during massage allows quantitative data which can analyze not only the amplitude range but also the specific waveform properties of applied TFM. It is found that Hmax/Mmax ratios decreased significantly after TFM application (P < 0.01), whereas the Hmax/Mmax ratios remained essentially unchanged in the control group (P > 0.05). These results suggest that TFM application reduces the excitability of the FCR MN pool in healthy subjects. These results are in general agreement with the previous studies which found that muscle massage leads to a reduction of spinal MN excitability (Morelli et al., 1990, 1991), although Morelli’s work was for a different massage technique. Candidate mechanoreceptors responding to the TFM

H.-M. Lee et al. / Manual Therapy 14 (2009) 501–507

stimulations and relaying the signals to the spinal reflex include the cutaneous receptors (Hagbarth, 1952; Levin and Chapman, 1987) and the deep mechanoreceptors, i.e. free nerve endings and muscle spindles in the FCR muscle tissue (Lundy-Ekman, 1998). By applying a topical anaesthetic to abolish physical sensation in the massage area, prior work has shown that the cutaneous receptors play a minor role for the reduction of MN pool excitability during the concurrent petrissage massage (Morelli et al., 1999). In contrast, regular input to the deep mechanoreceptors is reported to have an essential inhibitory effect on the spinal cord circuits (Morelli et al., 1991, 1999; Sullivan et al., 1991). Studies with indirect evidence have demonstrated that pressure-sensitive and stretch-sensitive free nerve endings in muscle tissue connect to inhibitory interneurons and therefore play a role in reducing MN pool excitability (Rymer et al., 1979; Hidler and Schmit, 2004; Ge et al., 2007). Possibly a change of flexibility (i.e. the loosening of muscle fibers from rearrangement of muscle architecture) induced by mechanical energy input to the muscle helps to relax the muscle, thereby reducing active pain sensation (Weerapong et al., 2005) and also reducing the sensitivity of muscle spindles, hence playing one or more roles in reducing MN pool excitability. These are interesting speculations but need further study via advanced quantitative evaluation of the relations between massage and the neurophysiological consequences. 5. Conclusion The findings of the present study have suggested that TFM is effective in actual neurological reduction of the excitability of the FCR MN pool. This study has also introduced a novel method for quantitative measurement of both momentary and cumulative pressure applied during massage. It is expected that this initial quantitative study of massage pressure and the energy transferred from therapist to client will lead to a significant increase in higher studies in this area and an increased understanding of the complex nature of the various therapeutic improvements reported for the many historically popular massage modalities. Acknowledgements The authors would like to thank the National Science Council of Taiwan for financially supporting the work under Contract No. NSC 94-2320-B-214-003. References Benjamin PJ, Tappan FM. Tappan’s handbook of healing massage techniques: classic, holistic and emerging methods. 4th ed. Stamford: Appleton & Lange; 1998. Bischoff C. Neurography: late responses. Muscle Nerve 2002;Suppl. 11:S59–65. Braverman DL, Schulman RA. Massage techniques in rehabilitation medicine. Phys Med Rehabil Clin N Am 1999;10(3):631–49, ix. Brooks CP, Woodruff LD, Wright LL, Donatelli R. The immediate effects of manual massage on power-grip performance after maximal exercise in healthy adults. J Altern Complement Med 2005;11(6):1093–101.

507

Brosseau L, Casimiro L, Milne S, Robinson V, Shea B, Tugwell P, et al. Deep transverse friction massage for treating tendinitis. Cochrane Database Syst Rev 2002;4:CD003528. Christie AD, Inglis JG, Boucher JP, Gabriel DA. Reliability of the FCR H-reflex. J Clin Neurophysiol 2005;22(3):204–9. De Domenico G, Wood EC. Beard’s massage. 4th ed. Philadelphia: W.B. Saunders Company; 1997. Dishman JD, Bulbulian R. Comparison of effects of spinal manipulation and massage on motoneuron excitability. Electromyogr Clin Neurophysiol 2001;41(2):97–106. Dishman JD, Burke J. Spinal reflex excitability changes after cervical and lumbar spinal manipulation: a comparative study. Spine J 2003;3(3):204–12. Farr T, Nottle C, Nosaka K, Sacco P. The effects of therapeutic massage on delayed onset muscle soreness and muscle function following downhill walking. J Sci Med Sport 2002;5(4):297–306. Field TM. Massage therapy effects. Am Psychol 1998;53(12):1270–81. Field T. Touch therapy. New York: Churchill Livingstone; 2000. pp. 48–91. Fisher MA. AAEM minimonograph #13: H reflexes and F waves: physiology and clinical indications. Muscle Nerve 1992;15(11):1223–33. Ge HY, Collet T, Morch CD, Arendt-Nielsen L, Andersen OK. Depression of the human nociceptive withdrawal reflex by segmental and heterosegmental intramuscular electrical stimulation. Clin Neurophysiol 2007;118(7):1626–32. Goldberg J, Sullivan SJ, Seaborne DE. The effect of two intensities of massage on H-reflex amplitude. Phys Ther 1992;72(6):449–57. Hagbarth KE. Excitatory and inhibitory skin areas for flexor and extensor motoneurons. Acta Physiol Scand Suppl 1952;26(94):1–58. Hammer WI. The use of transverse friction massage in the management of chronic bursitis of the hip or shoulder. J Manipulative Physiol Ther 1993;16(2):107–11. Hidler JM, Schmit BD. Evidence for force-feedback, inhibition in chronic stroke. IEEE Trans Neural Syst Rehabil Eng 2004;12(2):166–76. Hopkins J, Ingersoll CD, Edwards J, Klootwyk TE. Cryotherapy and transcutaneous electric neuromuscular stimulation decrease arthrogenic muscle inhibition of the vastus medialis after knee joint effusion. J Athl Train 2002;37(1):25–31. Jaberzadeh S, Scutter S. Flexor carpi radialis motoneuron pool in subjects with chronic carpal tunnel syndrome are more excitable than matched control subjects. Man Ther 2006;11(1):22–7. Levin M, Chapman CE. Inhibitory and facilitatory effects from the peroneal nerve onto the soleus H-reflex in normal and spinal man. Electroencephalogr Clin Neurophysiol 1987;67(5):468–78. Lundy-Ekman L. Neuroscience: fundamentals for rehabilitation. 2nd ed. Philadelphia: WB Saunders; 1998. Melzack R, Wall PD. The challenge of pain. 2nd ed. New York: Basic Books; 1996. Morelli M, Chapman CE, Sullivan SJ. Do cutaneous receptors contribute to the changes in the amplitude of the H-reflex during massage? Electromyogr Clin Neurophysiol 1999;39(7):441–7. Morelli M, Seaborne DE, Sullivan SJ. Changes in H-reflex amplitude during massage of triceps surae in healthy subjects. J Orthop Sports Phys Ther 1990;12(2):55–9. Morelli M, Seaborne DE, Sullivan SJ. H-reflex modulation during manual muscle massage of human triceps surae. Arch Phys Med Rehabil 1991;72(11):915–9. Moyer CA, Rounds J, Hannum JW. A meta-analysis of massage therapy research. Psychol Bull 2004;130(1):3–18. Muggleton JM, Allen R, Chappell PH. Hand and arm injuries associated with repetitive manual work in industry: a review of disorders, risk factors and preventive measures. Ergonomics 1999;42(5):714–39. Racasan O, Dubert T. The safest location for steroid injection in the treatment of carpal tunnel syndrome. J Hand Surg [Br] 2005;30(4):412–4. Rymer WZ, Houk JC, Crago PE. Mechanisms of the clasp-knife reflex studied in an animal-model. Exp Brain Res 1979;37(1):93–113. Sullivan SJ, Williams LR, Seaborne DE, Morelli M. Effects of massage on alpha motoneuron excitability. Phys Ther 1991;71(8):555–60. Vujnovich AL, Dawson NJ. The effect of therapeutic muscle stretch on neural processing. J Orthop Sports Phys Ther 1994;20(3):145–53. Weerapong P, Hume PA, Kolt GS. The mechanisms of massage and effects on performance, muscle recovery and injury prevention. Sports Med 2005;35(3): 235–56. Zehr PE. Considerations for use of the Hoffmann reflex in exercise studies. Eur J Appl Physiol 2002;86(6):455–68. Zhu Y, Haldeman S, Hseih CY, Wu P, Starr A. Do cerebral potentials to magnetic stimulation of paraspinal muscles reflect changes in palpable muscle spasm, low back pain, and activity scores? J Manipulative Physiol Ther 2000;23(7): 458–64.

Manual Therapy 14 (2009) 508–513

Contents lists available at ScienceDirect

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

Original Article

The effects of cervical high-velocity low-amplitude thrust manipulation on resting electromyographic activity of the biceps brachii muscleq James Dunning a, b, *, Alison Rushton b a b

Acupuncture, Spine & Headache Centre, Montgomery, AL, United States School of Health Sciences, University of Birmingham, Edgbaston, Birmingham, B15 2TT, United Kingdom

a r t i c l e i n f o

a b s t r a c t

Article history: Received 7 November 2007 Received in revised form 16 September 2008 Accepted 23 September 2008

There is a gap in the literature regarding the effects of spinal manipulation on extremity muscles that are unconnected to the vertebral column by an origin or insertion. This study investigated the effect of a right C5/6 high-velocity low-amplitude thrust (HVLAT) manipulation on resting electromyographic activity of the biceps brachii muscles bilaterally. A placebo-controlled, single-blind, repeated measures design employed an asymptomatic convenience sample (n ¼ 54) investigating three conditions: HVLAT, sham, and control. HVLAT demonstrated an excitatory effect with increased EMG activity of 94.20% (P ¼ 0.0001) and 80.05% (P ¼ 0.0001) for the right and left biceps respectively. A one-way repeated measures ANOVA revealed a significant difference (P ¼ 0.0001) in the mean percentage change of resting EMG activity, as did post hoc analyses (P ¼ 0.0001) between all three conditions. Subjects not experiencing cavitation post HVLAT demonstrated greater EMG increases for both right (P ¼ 0.0001) and left (P ¼ 0.014) biceps than those experiencing cavitation. The magnitude of mean EMG change for the right biceps was significantly greater than the left (P ¼ 0.011) post HVLAT. This study demonstrates a single HVLAT to the right C5/6 zygapophyseal joint elicits an immediate increase in resting EMG activity of the biceps bilaterally, irrespective of whether or not cavitation occurs. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Cervical manipulation Neurophysiological effects High velocity low amplitude thrust Biceps brachii

1. Introduction Spinal mobilisation and manipulation have been used for more than 2000 years in the treatment of neuromusculoskeletal disorders (Curtis, 1988). The effects of mobilisation and high-velocity lowamplitude thrust (HVLAT) manipulation have been an area of focus for recent research. Several studies have demonstrated that mobilisation and HVLAT of the cervical spine produce hypoalgesic effects through increased pressure pain thresholds in symptomatic and asymptomatic subjects (Cassidy et al., 1992; Vicenzino et al., 1995, 1998; Sterling et al., 2001; Fernandez-de-las-Penas et al., 2007). In addition, several studies have demonstrated mobilisation of the cervical spine in asymptomatic and symptomatic populations stimulates the peripheral sympathetic nervous system resulting in decreased blood flow and skin temperature, and increased skin conductance in the upper extremities (Petersen et al., 1993;

q This research was undertaken at the School of Health Sciences, University of Birmingham, United Kingdom. * Corresponding author. Acupuncture, Spine & Headache Centre, Montgomery, AL 36116, USA. Tel.: þ1 334 356 1670; fax: þ1 334 356 1690. E-mail address: [email protected] (J. Dunning). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.09.003

Vicenzino et al., 1998; Sterling et al., 2001). However, there is conflicting evidence regarding the excitatory (Herzog et al., 1999; Suter et al.,1999; Keller and Colloca, 2000; Symons et al., 2000; Suter et al., 2000; Colloca and Keller, 2001; Dishman et al., 2002; Suter and McMorland, 2002) or inhibitory (Dishman and Bulbulian, 2000; Lehman and McGill, 2001; Lehman et al., 2001; DeVocht et al., 2005) nature of the neurophysiological response that occurs after HVLAT manipulation of the spine. The methodological quality of these studies is poor; with only three studies (Keller and Colloca, 2000; Suter et al., 2000; Dishman et al., 2002) utilising control or placebo groups. In addition, only two studies (Dishman and Bulbulian, 2000; Dishman et al., 2002) administered a single unilateral HVLAT manipulation to each subject; with the remaining studies administering multiple bilateral HVLAT manipulations, and in some studies to multiple spinal regions. Conclusions cannot therefore be made regarding the excitatory or inhibitory nature of reflexive muscular response post HVLAT. HVLAT to segmentally associated zygapophyseal joints has demonstrated transient reflexic contractions of local paraspinal muscles using electromyography in asymptomatic (Herzog et al., 1999; Symons et al., 2000) and symptomatic subjects (Colloca and Keller, 2001). After lumbar HVLAT in LBP subjects, immediate increases in muscle strength of the erector spinae have been

J. Dunning, A. Rushton / Manual Therapy 14 (2009) 508–513

demonstrated (Keller and Colloca, 2000). Equally, however, an immediate reduction in paraspinal electromyographic activity post HVLAT in asymptomatic (Dishman and Bulbulian, 2000) and LBP patients (Lehman and McGill, 2001; DeVocht et al., 2005) has been demonstrated and again it remains unclear whether HVLAT produces an excitatory or inhibitory effect on paraspinal muscle activity. There is a gap in the literature regarding the effects of HVLAT on extremity muscles unconnected to the vertebral column by an origin or insertion. Herzog et al. (1999) assessed the effects of HVLAT to the spine on resting EMG activity of deltoid and found an ipsilateral reflex muscle contraction of deltoid post HVLAT. However, this was a limited study (n ¼ 10) with no control or placebo, and each subject received 11 HVLAT manipulations to the cervical, thoracic, lumbar and pelvic regions. In addition, Herzog et al. (1999) did not report the magnitude of the response, only the percentage of positive responses occurring when the signal increased to at least three times the baseline value. Suter and McMorland (2002) demonstrated an immediate 7–10 N.m increase in elbow flexor torque post HVLAT of the cervical spine; however, again the results must be interpreted cautiously because no control or placebo groups were utilised and multiple HVLAT manipulations were administered on all subjects. Several authors (Indahl et al., 1997; Herzog et al., 1999; Symons et al., 2000; Pickar and Kang, 2006) have proposed that the neurophysiologic pathway of the observed electromyographic response following HVLAT involves activation of the mechanoreceptors in the zygapophyseal joint capsule, spinal ligaments, and intervertebral disc, the cutaneous receptors, and the muscle spindles and golgi tendon organs within the muscle belly and tendon of the associated muscles. Alteration in afferent discharge rates from the stimulation of these receptors following HVLAT manipulation is thought to cause changes in alpha motorneuron excitability levels with subsequent changes in muscle activity (Indahl et al., 1997; Dishman and Bulbulian, 2000; Suter et al., 2000; Symons et al., 2000). However, this proposal is not fully supported by their research (Herzog et al., 1999; Symons et al., 2000) as only Pickar and Kang (2006) directly measured mechanoreceptor or proprioceptor activity. Furthermore, Pickar and Kang (2006) only measured the muscle spindle discharge rates in non-human subjects. There has been some debate in the literature surrounding the role of cavitation (an audible ‘‘pop’’ or ‘‘crack’’) during HVLAT and the observed effects. Herzog et al. (1993a) found reflex responses in the paravertebral muscles irrespective of whether cavitation was achieved. Likewise, Dishman and Bulbulian (2000) found similar reflexic responses in the lumbar spine following either mobilisation without cavitation or manipulation with cavitation, and proposed that the velocity dependent facet joint mechanoreceptors were not implicated in the neurophysiologic response. In contrast, Suter et al. (1994) were not able to elicit electromyographic reflex responses from non-cavitation thrust manipulations given at a low-velocity, i.e. at a rate greater than 1 s compared with 100–150 ms for highvelocity thrusts; however, no control or placebo conditions were employed and the findings cannot be attributed to the intervention. The question therefore remains whether the cavitation phenomenon is required to facilitate a neurophysiological response in resting muscle activity post HVLAT. To date, no controlled study has investigated the effects of cervical HVLAT manipulation on resting muscle activity more distal than the deltoid (Herzog et al., 1999; Suter and McMorland, 2002) or on contralateral upper extremity muscle activity. The purpose of this study was to characterise the nature (excitatory or inhibitory) and the magnitude of any change in resting electromyographic activity of the biceps brachii muscle post C5/6 HVLAT ipsilaterally and contralaterally. In addition, the relationship to joint cavitation

509

was explored. The biceps brachii muscle was selected as it is anatomically unconnected to the area of intervention through origin or insertion, but is segmentally linked from a neurophysiological perspective. 2. Methods 2.1. Subjects A convenience sample of 54 asymptomatic undergraduate physiotherapy and nursing students (39 female and 15 male) with a mean age of 22.13  4.68 years were recruited. Mean mass was 65.71 kg (SD 12.49) and mean height was 1.70 m (SD 0.091). Subjects were included if aged 18–40 years. Exclusion criteria included neck pain in the last 6 months; a history of trauma or surgery to the cervical spine or upper extremities; upper extremity referred pain, radiculopathy or peripheral neuropathy; or any contraindication to cervical HVLAT manipulation (Hartman, 2001; Gibbons and Tehan, 2003; Kerry and Taylor, 2006; Kerry et al., 2008). The most recent literature suggests that pre-manipulative cervical artery testing is unable to identify those individuals at risk of vascular complications from cervical HVLAT manipulation (Kerry and Taylor, 2005; Kerry et al., 2008), and any symptoms detected during pre-manipulative testing may be unrelated to changes in blood flow in the vertebral artery, so that a negative test neither predicts the absence of arterial pathology nor the propensity of the artery to be injured during cervical HVLAT, with testing neither sensitive or specific (Licht et al., 2000; Magarey et al., 2004; Kerry and Taylor, 2005; Kerry and Taylor, 2006; Kerry et al., 2008). Screening questions for cervical artery disease were negative, and pre-manipulative screening was not used. The study was approved by the Ethics Committee of the School of Health Sciences of The University of Birmingham, and written informed consent was obtained from all the subjects prior to testing. 2.2. Equipment Resting electromyographic recordings of the biceps brachii muscle were made pre and post C5/6 HVLAT using the DelSysÒ Surface EMG system (DeLuca, 1997, 2002, 2003). Detection electrode surfaces were made of pure silver (>99.5%) in the form of parallel bars 10 mm long and 1 mm wide with an inter-detection surface spacing of 1.0 cm. This small electrode size and interdetection surface spacing minimised cross-talk susceptibility from adjacent muscles (DeLuca, 1997, 2002). In addition, considering an average nerve conduction velocity of 4.0 m/s (Basmajian, 1985) and using the stated electrode size and inter-detection spacing, a bandwidth between 20 and 450 Hz was used to capture the full frequency spectrum of the biceps brachii EMG signal and suppress noise at higher frequencies (DeLuca, 1997, 2002). 2.3. Procedure Each subject was positioned supine on a plinth with their lower limbs straight and head and neck in a neutral position on a single pillow. The subjects’ arms rested on the plinth with the elbows bent at 90 and fingers interlocked over the abdomen to limit movement of the upper limbs during and between interventions. In order to minimise skin impedance between electrodes, the skin was wiped with alcohol swabs (DeLuca, 1997, 2002, 2003). Then 10 mm electrodes were placed on the longitudinal midline of the biceps brachii muscle bilaterally mid-way between the origin and insertion point (DeLuca, 2002). The short head of the biceps brachii originates from the apex of the coracoid process and the long head arises from the upper margin of the glenoid cavity; the two muscle bellies are closely applied to each other in the middle and lower brachium and

J. Dunning, A. Rushton / Manual Therapy 14 (2009) 508–513

2.4. Data and statistical analyses DelSysÒ EMG Analysis software was used to rectify the raw bipolar signal to calculate the mean rectified absolute values, or average rectified value (ARV) for each 30 s data segment for both the right and left biceps brachii muscle in all three conditions for each subject. This process resulted in each subject having 12 ARV means (648 ARV means in total) encompassing a pre and post value for the control, sham, and HVLAT conditions contributing six means for the left and six means for the right biceps. The mean percentage of change (i.e. post-intervention minus pre-intervention, divided by pre-intervention, and multiplied by 100%) in resting EMG activity of the biceps brachii muscle was calculated. Data for both the right and left biceps was included in the analysis using SPSS 14.0. A one-way repeated measures ANOVA tested for differences in the mean percentage change in resting EMG activity of the biceps brachii muscle between the three conditions. Post hoc analyses (Bonferroni pairwise comparisons) were subsequently performed. A paired t-test investigated ipsilateral and contralateral differences. An independent t-test investigated differences between those subjects who demonstrated cavitation and those that did not. The level of significance was set at 0.05 for all statistical procedures. 3. Results 3.1. Magnitude of EMG response The mean percentage change of resting EMG activity of the right biceps brachii in the three conditions was 4.18% (control), 21.12% (sham), and 94.20% (HVLAT); and 2.16%, 17.15% and 80.04%, respectively, for the left. The error chart in Fig. 1 displays the means

110 100 90 80 70 60 50 40 30 20 10 0

ft Le

ig H

VL

AT

R H

VL

AT

am Sh

R am Sh

ht

ft Le

ht ig

ef lL tro on C

lR on tro

Control–Sham–HVLAT Sham–Control–HVLAT Sham–HVLAT–Control Control–HVLAT–Sham HVLAT–Control–Sham HVLAT–Sham–Control

C

1–9 10–18 19–27 28–36 37–45 46–54

ig

Order of conditions Subjects Subjects Subjects Subjects Subjects Subjects

t

-10 ht

95

Table 1 Subject allocation to order of conditions.

and Keller, 2001; Lehman and McGill, 2001; Lehman et al., 2001; Suter and McMorland, 2002; Colloca et al., 2003; Marshall and Murphy, 2006), there are no studies supporting the notion that changes in resting EMG activity of the paravertebral muscles post HVLAT last any longer than 4–5 min in duration (DeVocht et al., 2005). This informed an 8 min ‘wash-out’ period to minimise any carry-over effect between the control, sham and HVLAT conditions.

change in EMG activity of biceps

insert as one tendon into the radial tuberosity (Gray, 1995). The electrode detection bars were aligned perpendicular to the length of the muscle fibres to allow intersection of most of the same muscle fibres by both detection bars and provide an EMG signal that reflected the activity of a fixed set of muscle fibres (DeLuca, 1997, 2002). The reference electrode (2 cm  2 cm) was placed on the dorsum of the right hand (DeLuca, 2002). The DelSysÒ EMG software was set to collect data at a sampling rate of 2000 Hz per channel (Herzog et al., 1999; Symons et al., 2000; Suter and McMorland, 2002; DeVocht et al., 2005). Prior to any data collection, subjects were instructed not to move any part of their body and to ‘‘relax as fully as possible’’. Before each condition was administered (control, sham or HVLAT), baseline resting EMG activity levels of the right and left biceps brachii muscles were recorded for a ‘pre’ 30 s segment, followed by a 1 min rest period (wherein the subject remained relaxed and supine with fingers interlocked over the abdomen), and then a ‘post’ 30 s segment (‘during/after’ ¼ post) was initiated by a research assistant using a manual trigger on the computer to initiate EMG data collection at the moment the manipulative physiotherapist contacted the subjects head and neck region. During this ‘post’ 30 s data segment, one of the three experimental conditions was administered and all three conditions were applied to all 54 subjects. The HVLAT manipulation to the right C5/6 segment (Hartman, 2001; Gibbons and Tehan, 2003) was performed by the manipulative physiotherapist placing the anterolateral aspect of the proximal phalanx of the right index finger over the posterolateral aspect of the articular pillar at the right C5/6 segment while the therapist’s other hand cradled the subjects head on the left. Extension, ipsilateral side-bend, contralateral side-shift and contralateral rotation of C5 on C6 were introduced to engage the barrierdthat is, until a firm crisp end-feel could be felt by the therapistdthen an HVLAT was administered into left rotation in an arc towards the left eye. The head was then repositioned on the pillow into the same starting neutral position and all hand contact was removed for the remainder of the ‘post’ 30 s data collection interval. It was recorded if cavitation occurred. The sham manipulation to the right C5/6 segment was administered using the same ‘set-up’ as the HVLAT manipulation; however, once the barrier was engaged, the head was re-positioned to neutral with no thrust applied. The control condition consisted of no manual contact for 30 s. Six sequencing orders were possible; and subjects, irrespective of gender, were randomly allocated to one of the sequencing orders (see Table 1). In order to minimise any carry-over effect from one intervention to the next, an 8 min ‘‘wash-out’’ period was used between all conditions. DeVocht et al. (2005) found changes in resting electromyographic activity of the paravertebral muscles post spinal HVLAT to stabilise back to pre-treatment levels within several seconds to 4–5 min; and to date, although several studies have demonstrated immediate changes in EMG activity post spinal HVLAT (Herzog et al., 1999; Dishman and Bulbulian, 2000; Keller and Colloca, 2000; Suter et al., 2000; Symons et al., 2000; Colloca

CI for the mean

510

Fig. 1. Mean and 95% CI for the percentage of change in resting EMG activity of the right and left biceps brachii muscles following a control condition, a sham manipulation to the right C5/6 segment, and an HVLAT manipulation to the right C5/6 segment.

3.2. Ipsilateral and contralateral effect The mean percentage change in resting EMG activity following HVLAT to the right C5/6 segment was 94.20% and 80.04% for the right and left biceps brachii muscles, respectively (see Fig. 2), with a mean difference of 14.16%. The right biceps brachii muscle therefore experienced a greater increase in resting muscle activity than the left. A paired t-test demonstrated this difference between the mean EMG change of the right and left biceps brachii muscles to be significant (t ¼ 2.645, P ¼ 0.011). 3.3. Cavitation effect Thirty-two of the 54 subjects demonstrated joint cavitation following the HVLAT condition. The mean percentage change in resting EMG activity of the right biceps brachii muscle post HVLAT was 79.79% and 115.16% for the cavitation and no cavitation groups, respectively (see Fig. 3). Similarly, the mean percentage change for the left biceps brachii muscle post HVLAT was 69.61% and 95.20% for the cavitation and no cavitation groups, respectively. An independent t-test demonstrated a significant difference between the cavitation and no cavitation groups both on the right (t ¼ 3.817, P ¼ 0.0001) and on the left (t ¼ 2.744, P ¼ 0.014). 4. Discussion The findings of this study provide support for previous studies demonstrating an excitatory effect of HVLAT on motor activity (Suter et al., 1999; Keller and Colloca, 2000; Suter et al., 2000; Symons et al., 2000; Colloca and Keller, 2001; Dishman et al., 2002), and more specifically on segmentally associated muscles of the

Table 2 Parameter estimates. Type of Manipulation

Mean

Std. error

Control right Control left Sham right Sham left HVLAT right HVLAT left

4.18 2.16 21.12 17.15 94.20 80.04

1.31 1.09 2.09 2.92 5.10 5.20

95% Confidence interval Lower bound

Upper bound

6.80 4.34 16.92 11.29 83.97 69.61

1.56 0.02 25.31 23.01 104.42 90.47

511

110 100 90 80 70 60 HVLAT Right

HVLAT Left

Fig. 2. Mean and 95% CI for the percentage of change in resting EMG activity of the right and left biceps brachii muscles following HVLAT manipulation to the right C5/6 facet joint.

upper limb (Herzog et al., 1999; Suter and McMorland, 2002). However, in both of these studies multiple HVLAT manipulations were administered on each subject, no control or placebo groups were employed, and small sample sizes of 10 (Herzog et al., 1999) and 16 subjects (Suter and McMorland, 2002) were used. Furthermore, Herzog et al. (1999) did not report the magnitude of the response in the deltoid muscle (only the percentage of positive responses), and Suter and McMorland (2002) measured elbow flexor torque and muscle inhibition changes during maximal voluntary contractions, rather than at rest. Therefore, this is the first controlled study to demonstrate an excitatory effect, and quantify its magnitude on the resting EMG activity of an upper limb muscle following a single HVLAT manipulation to the cervical spine. It has been postulated that HVLAT manipulation activates mechanosensitive afferents (mechanoreceptors) in the intervertebral discs, zygapophyseal joints, spinal ligaments, paravertebral muscles (proprioceptors) and skin (Indahl et al., 1997; Herzog et al., 1999; Symons et al., 2000; Pickar and Kang, 2006). Alteration in afferent input from the stimulation of these receptors is thought to cause changes in alpha motorneuron excitability levels with subsequent increases in muscle activity (Dishman and Bulbulian, 2000; Suter et al., 2000). In this study, the increase in resting EMG

95% CI for the mean % change in EMG activity of right biceps after HVLAT

and 95% confidence intervals for the percentage change in resting EMG activity for each condition, and Table 2 illustrates the parameter estimates post each condition. Resting EMG activity of the biceps brachii muscle increased in 94% (n ¼ 51) of subjects following a single HVLAT to the right C5/6 facet joint, with a slight decrease observed in 6% (n ¼ 3) of subjects. The one-way repeated measures ANOVA demonstrated a significant difference for mean percentage change of resting EMG activity of the biceps brachii muscle (F ¼ 223.28, P ¼ 0.0001). Bonferroni post hoc analyses demonstrated significant differences (P ¼ 0.0001) between all three conditions. For the right biceps, mean percentage change and pairwise comparison between HVLAT and control conditions was 98.38% (P ¼ 0.0001, 95% CI: 84.08–112.68), between sham and control conditions was 25.30% (P ¼ 0.0001, 95% CI: 19.63–30.97), and between HVLAT and sham was 73.08% (P ¼ 0.0001, 95% CI: 59.43–86.73). Similar trends were demonstrated for the left biceps brachii muscle, and pairwise comparison between HVLAT and control conditions was 82.19% (P ¼ 0.0001, 95% CI: 67.06–97.33), between sham and control conditions was 19.31% (P ¼ 0.0001, 95% CI: 10.10–28.52), and between HVLAT and sham was 62.89% (P ¼ 0.0001, 95% CI: 49.18–76.59).

95% CI for the mean % change in EMG activity of the biceps

J. Dunning, A. Rushton / Manual Therapy 14 (2009) 508–513

140 130 120 110 100 90 80 70 60 Cavitation

No Cavitation

Fig. 3. Mean and 95% CI for the mean percentage of change in resting EMG activity of the right biceps brachii muscle following HVLAT manipulation of the right C5/6 facet joint between those subjects experiencing cavitation and those not experiencing cavitation.

512

J. Dunning, A. Rushton / Manual Therapy 14 (2009) 508–513

activity of the biceps brachii muscle occurred whether or not the C5/6 facet joint demonstrated the cavitation phenomenon. Therefore, in agreement with the findings of Herzog et al. (1999), this study supports the hypothesis that the neurophysiological reflexic increase in resting EMG activity of the biceps brachii muscle depends on the magnitude of force applied (Conway et al., 1993; Herzog et al., 1993b) and/or the rate of change in force application (acceleration) during the thrusting impulse (Colloca and Keller, 2001; Pickar and Kang, 2006), rather than the occurrence of the cavitation phenomenon itself. The findings of this study demonstrate that HVLAT manipulation to the right C5/6 facet joint significantly increased the resting electromyographic activity of both the right and left biceps brachii muscles. This is consistent with the findings of Colloca and Keller (2001) who observed a contralateral neuromuscular reflex response in the lumbar erector spinae muscles following HVLAT manipulation to the lumbar spine. These findings are in contrast to Symons et al. (2000) who found the increase in resting EMG activity to always occur ipsilaterally and in muscles that had either their origin or insertion at the vertebral level that was manipulated. The results of this study demonstrate a non-local response and furthermore, an ipsilateral and contralateral response. The nonlocal response found in this study is in agreement with the findings of Herzog et al. (1999) that found increased EMG activity in the deltoid muscle. The notion that muscle inhibition, or decreased motor activity, can occur in muscle groups that are not directly connected to the spine, such as the quadriceps or biceps muscles as a result of lumbopelvic or cervical joint dysfunction is increasingly supported within the literature (Suter et al., 1999, 2000; Suter and McMorland, 2002). Therefore, although this study examined the outcomes in a population of asymptomatic subjects, facilitation of resting motor activity in the elbow flexor muscles post HVLAT to the cervical spine as demonstrated in this study, may still have clinical implications for rehabilitation practitioners. The findings contribute to the suggestion that for optimal management of patients with cervical pain and upper extremity weakness suspected to be of an arthrogenic nature (Suter et al., 2000; Liebler et al., 2001; Sterling et al., 2001; Suter and McMorland, 2002), the application of HVLAT manipulation to the segmentally associated facet joints in the cervical spine may be a beneficial approach before traditional strength training is initiated. Previously, Suter and McMorland (2002) found, when compared with a normal sample, that most patients with chronic neck pain demonstrated more than 5% inhibition of the biceps brachii muscles; and furthermore muscle inhibition bilaterally was reduced to control levels following one treatment session of HVLAT to C5/6 and C6/7 levels. More specifically, Suter and McMorland (2002) reported an immediate increase in elbow flexor torque of 7–10 N.m during maximal isometric contractions and a 4.3–11.1% decrease in elbow flexor muscle inhibition following a single session of HVLAT. However, Suter and McMorland (2002) did not report the side manipulated (right and/or left) or the number of HVLATs delivered to each patient; with no placebo or control groups employed. There were several limitations to this study that need to be acknowledged. No verification existed to ensure that the actual motion segment that was manipulated was indeed the C5/6 level, and this is problematic as the literature reports poor levels of accuracy and specificity of many HVLAT manipulation procedures (Beffa and Mathews, 2004; Ross et al., 2004). In addition, the magnitude of the thrusting force of the HVLAT applied to the C5 vertebrae was not standardised between subjects, and exact replication of electrode placement within the centre of the longitudinal midline of the muscle (DeLuca, 2002, 2003) was not verified.

This study also highlights areas for further research. It would be useful to investigate longer duration recordings of electromyographic activity in order to elucidate the longer term effects of HVLAT. In addition to resting electromyographic activity, measurement of outcomes that represent immediate and longer term changes in the functional capacity of muscles post HVLAT should be investigated. In order to assess the actual clinical relevance of these findings, future studies should employ a symptomatic population with neck pain and/or upper limb dysfunction. 5. Conclusion This study has demonstrated that a single HVLAT manipulation to the cervical spine elicits a measurable short term increase in resting electromyographic activity in a remote area not directly connected by any musculoskeletal structures to the cervical spine but segmentally and neuroanatomically associated. The results suggest that HVLAT to the cervical spine immediately increases the resting electromyographic activity of the biceps brachii muscle, but does not address the duration of this increase. In addition, HVLAT to the right C5/6 zygapophyseal joint immediately increased resting motor activity of both the right and left biceps brachii muscles, and this increase occurred irrespective of whether the cavitation phenomenon was present. Acknowledgements The assistance of Cesar Fernandez-de-las-Penas during manuscript review is gratefully acknowledged. References Basmajian JV. Muscles alive: their functions revealed by electromyography. 5th ed. Baltimore, MD: Williams & Wilkins; 1985. Beffa R, Mathews R. Does the adjustment cavitate the targeted joint? An investigation into the location of cavitation sounds. Journal of Manipulative Physiological Therapeutics 2004;27(2):118–22. Cassidy JD, Lopes AA, Yong-Hing K. The immediate effect of manipulation versus mobilisation on pain and range of motion in the cervical spine: a randomised controlled trial. Journal of Manipulative and Physiological Therapeutics 1992;15(9):570–5. Colloca CJ, Keller TS. Electromyographic reflex responses to mechanical force, manually assisted spinal manipulative therapy. Spine 2001;26(10):1117–24. Colloca CJ, Keller TS, Gunzburg R. Neuromechanical characterization of in vivo lumbar spinal manipulation. Part II. Neurophysiological response. Journal of Manipulative and Physiological Therapeutics 2003;26(9):579–91. Conway PJ, Herzog W, Zhang Y, Hasler EM, Ladly K. Forces required to cause cavitation during spinal manipulation of the thoracic spine. Clinical Biomechanics 1993;8:210–4. Curtis P. Spinal manipulation: does it work? Occupational Medicine 1988;3(1): 31–44. DeLuca CJ. The use of surface electromyography in biomechanics. Journal of Applied Biomechanics 1997;13(2):135–63. DeLuca CJ. DelSysÒ surface electromyography: detection and recording. DelSys Inc.; 2002. DeLuca CJ. DelSysÒ fundamental concepts in EMG signal acquisition. DelSys Inc.; 2003. DeVocht JW, Pickar JG, Wilder DG. Spinal manipulation alters electromyographic activity of paraspinal muscles: a descriptive study. Journal of Manipulative and Physiological Therapeutics 2005;28(9):465–71. Dishman DJ, Bulbulian R. Spinal reflex attenuation associated with spinal manipulation. Spine 2000;25(19):2519–25. Dishman DJ, Ball KA, Burke J. Central motor excitability changes after spinal manipulation: a transcranial magnetic stimulation study. Journal of Manipulative and Physiological Therapeutics 2002;25(1):1–9. Fernandez-de-las-Penas C, Perez-de-Heredia M, Brea-Rivero M, MiangolarraPage JC. Immediate effects on pressure pain threshold following a single cervical spine manipulation in healthy subjects. Journal of Orthopaedic and Sports Physical Therapy 2007;37(6):325–9. Gibbons P, Tehan P. Manipulation of the spine, thorax and pelvis: an osteopathic perspective. Edinburgh: Harcourt Publishers Ltd.; 2003. Gray H. Anatomy: descriptive and surgical. 15th ed. New York, NY: Barnes & Noble Books; 1995. 361–362. Hartman L. Handbook of osteopathic technique. 3rd ed. Cheltenham: NelsonThornes; 2001.

J. Dunning, A. Rushton / Manual Therapy 14 (2009) 508–513 Herzog W, Zhang YT, Conway PJ, Kawchuk GN. Cavitation sounds during spinal manipulative treatments. Journal of Manipulative and Physiological Therapeutics 1993a;16(8):523–6. Herzog W, Conway PJ, Kawchuk GN, Zhang Y, Hasler EM. Forces exerted during spinal manipulative therapy. Spine 1993b;18(9):1206–12. Herzog W, Scheele D, Conway PJ. Electromyographic responses of back and limb muscles associated with spinal manipulative therapy. Spine 1999;24(2): 146–52. Indahl A, Kaigle A, Reikeras O, Holm S. Interaction between the porcine lumbar intervertebral disc, zygapophyseal joints, and paraspinal muscles. Spine 1997;22(24): 2834–40. Keller TS, Colloca CJ. Mechanical force spinal manipulation increases trunk muscle strength assessed by electromyography: a comparative clinical trial. Journal of Manipulative and Physiological Therapeutics 2000;23(9):585–95. Kerry R, Taylor A, Mitchell J, McCarthy C, Brew J. Manual therapy and cervical arterial dysfunction, directions for the future: a clinical perspective. Journal of Manual and Manipulative Therapy 2008;16(1):39–48. Kerry R, Taylor A. The vertebral artery test. Manual Therapy 2005;10(4):297–8. Kerry R, Taylor A. Cervical arterial dysfunction assessment and manual therapy. Manual Therapy 2006;11(3):243–53. Lehman GJ, McGill SM. Spinal manipulation causes variable spine kinematic and trunk muscle electromyographic responses. Clinical Biomechanics 2001;16(4): 293–9. Lehman GJ, Vernon H, McGill S. Effects of a mechanical pain stimulus on erector spinae activity before and after a spinal manipulation in patients with back pain: a preliminary investigation. Journal of Manipulative and Physiological Therapeutics 2001;24(6):402–6. Licht PB, Christensen HW, Hoilund-Carlsen PF. Is there a role for pre-manipulative testing before cervical manipulation? Journal of Manipulative and Physiological Therapeutics 2000;23(3):175–9. Liebler EJ, Tufano-Coors L, Douris P, Makofsky HW, McKenna R, Michels C, Rattray S. The effect of thoracic spine mobilisation on lower trapezius strength testing. The Journal of Manual and Manipulative Therapy 2001;9(4): 207–12. Magarey ME, Rebbeck T, Coughlan B, Grimmer K, Rivett DA, Refshauge K. Premanipulative testing of the cervical spine review, revision and new clinical guidelines. Manual Therapy 2004;9(2):95–108.

513

Marshall P, Murphy B. The effect of sacroiliac joint manipulation on feed-forward activation times of the deep abdominal musculature. Journal of Manipulative and Physiological Therapeutics 2006;29(3):196–202. Petersen N, Vicenzino B, Wright A. The effects of a cervical mobilisation technique on sympathetic outflow to the upper limb in normal subjects. Physiotherapy Theory and Practice 1993;9:149–56. Pickar J, Kang Y. Paraspinal muscle spindle responses to the duration of a spinal manipulation under force control. Journal of Manipulative and Physiological Therapeutics 2006;29(1):22–31. Ross JK, Bereznick DE, McGill SM. Determining cavitation location during lumbar and thoracic spinal manipulation: is spinal manipulation accurate and specific? Spine 2004;29(13):1452–7. Sterling M, Jull G, Wright A. Cervical mobilisation: concurrent effects on pain, sympathetic nervous system activity and motor activity. Manual Therapy 2001;6(2):72–81. Suter E, Herzog W, Conway PJ, Zhang YT. Reflex response associated with manipulative treatment of the thoracic spine. Journal of Neuromusculoskeletal System 1994;2:124–30. Suter E, McMorland G, Herzog W, Bray R. Decrease in quadriceps inhibition after sacroiliac joint manipulation in patients with anterior knee pain. Journal of Manipulative and Physiological Therapeutics 1999;22(3):149–53. Suter E, McMorland G, Herzog W, Bray R. Conservative lower back treatment reduces inhibition in knee-extensor muscles: a randomized controlled trial. Journal of Manipulative and Physiological Therapeutics 2000;23(2):76–80. Suter E, McMorland G. Decrease in elbow flexor inhibition after cervical spine manipulation in patients with chronic neck pain. Clinical Biomechanics 2002;17:541–4. Symons BP, Herzog W, Leonard T, Nguyen H. Reflex responses associated with activator treatment. Journal of Manipulative and Physiological Therapeutics 2000;23(3):155–9. Vicenzino B, Gutschlag F, Collins D, Wright A. An investigation of the effects of spinal manual therapy on forequarter pressure and thermal pain thresholds and sympathetic nervous system activity in asymptomatic subjects. Adelaide: Moving In On Pain; 1995. Vicenzino B, Collins D, Benson H, Wright A. An investigation of the interrelationship between manipulative therapy-induced hypoalgesia and sympathoexcitation. Journal of Manipulative and Physiological Therapeutics 1998;21(7):448–53.

Manual Therapy 14 (2009) 514–519

Contents lists available at ScienceDirect

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

Original Article

Sex differences in the pattern of innominate motion during passive hip abduction and external rotation Melanie D. Bussey a, *, Stephan Milosavljevic b,1, Melanie L. Bell c, 2 a

School of Physical Education, University of Otago, PO Box 56, Dunedin 9013, New Zealand Centre for Physiotherapy Research, School of Physiotherapy, University of Otago, PO Box 56, Dunedin 9013, New Zealand c Senior Lecturer, Biostatistics, Department of Preventive and Social Medicine, University of Otago, Dunedin 9013, New Zealand b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 29 November 2007 Received in revised form 16 September 2008 Accepted 23 September 2008

The objective of the study was to evaluate sex differences in the pattern of innominate motion about the sacroiliac joint (SIJ) during hip movement. Although the magnitude of intrinsic SIJ motion is influenced by joint congruence and ligament elasticity sex differences in pelvic joint kinematics are under-investigated. Forty healthy and active males and females between the ages of 18 and 35 were recruited. 3D motion of the innominate bones and femur were recorded with a magnetic tracking device as the hips were loaded in standardised increments of 10 in 3 positions – external rotation (ER), abduction (AB), and combined external rotation and abduction (AB þ ER). While females had greater overall innominate motion, two distinct sex dominant patterns emerged. Patterns of innominate motion also differed when load was applied to the dominant rather than non-dominant limb. As the main motion within the pelvis is intrinsic, the results of the present study point to a differing viscoelastic response and different movement strategies to passive load between the sexes. In addition, careful attention to limb dominance should be considered when testing SIJ motion. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Sacroiliac joint Range of motion Kinematics Pelvis Hip joint

1. Introduction The amount of passive motion that develops at a joint can be influenced by intrinsic factors such as articular shape, congruence, and capsular and ligamentous laxity. Due to the requirements of pregnancy and childbirth, the female pelvic joints (both SIJ’s and symphysis pubis) are thought to be more lax than their male counterparts (MacLennan et al., 1986; Kristiansson et al., 1996). However, differences in the pelvic joints of males and females extend beyond the hormonal influences. Vleeming et al. (1990) determined that the articular surfaces of the female sacroiliac joint (SIJ) are smoother than the male SIJ and have a lower coefficient of friction, allowing the surfaces to slide more easily on one another. Furthermore, the articular surfaces of the female SIJ are shorter and more angled than the male SIJ (Brunner et al., 1991). Thus, there are inherent sex differences in pelvic joint kinematics. While research has shown that the range of motion between males and females is significantly different (Sturesson, 1997) no research could be found

* Corresponding author. Tel.: þ64 3 479 8981; fax: þ64 3 479 8309. E-mail addresses: [email protected] (M.D. Bussey), stephan.milosavljevic@ otago.ac.nz (S. Milosavljevic), [email protected] (M.L. Bell). 1 Tel.: þ64 3 479 7193; fax: þ64 3 479 8414. 2 Tel.: þ64 3 479 7236; fax: þ64 3 479 7298. 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.09.004

that investigated sex differences in the pattern of motion through successive incremental hip displacements. There is a strong link between the mobility of the hip/pelvis and low back pain (LBP), and studies have shown that participants with LBP also tend to have significant bilateral differences in the magnitude of internal to external rotation of the hip (Friberg, 1983; Offierski and White, 1983; Mellin, 1988; Barbee-Ellison et al., 1990; Chesworth et al., 1994; Gombatto et al., 2006). Further, there are differences in the range of lateral rotation of the hip between male and female LBP sufferers (Gombatto et al., 2006). As asymmetry of hip abduction and external rotation was previously reported in patients with inflammation of the sacroiliac joints (LeBan et al., 1978) Cibulka et al. (1998) investigated the association between LBP and SIJ involvement to determine whether there was a distinguishable difference in hip motion symmetry. They found that patients without SIJ involvement showed greater external and reduced internal hip rotation in both left and right sides whereas those with SIJ involvement had greater hip external rotation than internal rotation, on only one side (Cibulka et al., 1998). The purpose of the present study was to determine whether sex differences exist in motion of the left and right innominate bones when one hip is incrementally moved into increasing amounts of axial rotation, abduction, or a combination of axial rotation and abduction. Based on the background research, we hypothesized that (i) women would have greater overall innominate range of

M.D. Bussey et al. / Manual Therapy 14 (2009) 514–519

motion (ROM); (ii) innominate motion would increase with hip rotation but patterns would differ between males and females; and (iii) that there would be no bilateral asymmetry within each sex.

515

Table 1 The intraclass correlation coefficient (ICC) with 95% confidence intervals and standard error of measurement (SEM) calculated for males and females for each innominate angle. Innominate angle

2. Methods 2.1. Subjects

ICC Loaded Unloaded

Forty healthy and active subjects between the ages of 18 and 35 volunteered for this study approved by the University of Otago Human Ethics Committee. Subjects were 21 females (25.0  3.1 yrs) with a mean BMI of 20.7  2.0 kg/m2 and 19 males (23.3  4.3 yrs) with a mean BMI of 22.5  2.7 kg/m2. All subjects participated in an average of 8.9 h ( 4.1) of vigorous activity per week. At the time of data collection, they were all free from hip or low back disorders and gave informed consent for their participation. This study incorporated a randomized block design of three load conditions with subjects allocated to three random-order blocks ([1 2 3], [2 3 1], [3 1 2]), such that all subjects completed all conditions, but in a different order. 2.2. Procedure Kinematic data were collected with a magnetic tracking device3, consisting of a transmitter, four receivers, a digitizer and a systems electronics unit. Measurement error of the system in the x, y and z coordinates of each of the 4 pelvic points was 0.02 mm (SD 0.84 mm) on the x-axis, 0.07 mm (SD 0.82 mm) on the y-axis and 0.03 (SD 0.99 mm) on the z-axis. The global average value of imprecision in the measurement of a point for intra-observer reliability was 0.80 mm (SD 1.47 mm). A global coordinate system was established by mounting the transmitter to a rigid wooden support. The receivers were mounted to thermoplastic frames and secured firmly to the thighs and over the S1 spinous process with doublesided tape and VelcroÒ support straps. An anatomically relevant reference system for identifying the hip joint centre was defined with a predicative method based on each subject’s pelvic and lower limb anthropometrics (Bush and Gutowski, 2003). A hip rotation frame (as described in Bussey et al., 2009) was used to standardize the rotational increments applied to the femurs in three anatomical hip positions: external rotation - ER, abduction – AB, and a combination of external rotation and abduction called ER þ AB. By standardizing the femur rotation we were capable of standardizing the external load applied to the innominate, although the internal load may vary due to individual anatomical differences in muscle and ligament stiffness. A maximum of six incremental rotations (10 each) for both ER and AB were used, for each participant. A palpation and digitizing technique known to accurately and reliably measure innominate motion (Bussey et al., 2009) was used to calculate motion of the innominate bones in reference to their initial static positions. This technique required the palpation and digitizing (using the tracking stylus) of the anterior superior iliac spines (ASIS) and posterior superior iliac spines (PSIS) at each incremental rotation. Each palpated landmark was digitized several times in the reference position (hip at 0 ); the leg was passively rotated in 10 increments and the procedure repeated. The results of the measurement test-retest reliability analysis conducted for the present study are in Table 1. The motion of the innominate in the sagittal and transverse planes of the pelvis reference system was calculated as angular displacement between the reference

3 Polhemus 3Space FastrackÒ, 40 Hercules Drive, P.O. Box 560, Colchester, VT 05446.

Sagittal

F M F M F M

0.992 0.807 0.939 0.826 0.987 0.893

SEM (0.941,0.992) (0.501,0.978) (0.781,0.993) (0.491,0.980) (0.946,0.980) (0.560,0.992)

0.04 0.08 0.05 0.08 0.02 0.12

position and each subsequent 10 hip rotation (of ER, AB and ER þ AB) (Bussey et al., 2009). 2.3. Kinematics Using anatomically relevant local coordinate axes derived from digitized bony landmarks data were reduced using standard matrix transformations to determine the rotational matrix of the femur with respect to the pelvis and the pelvis with respect to the lumbar spine. Innominate bone motion was defined by the angular displacement of the innominate bones from a neutral position. Sagittal plane motion was calculated as a composite angle between innominates rotating about the sagittal-horizontal axis of the pelvis (described in Bussey et al., 2009). Transverse plane motion was calculated as absolute displacement of left and right bones individually about the vertical axis. For this calculation the innominate on the same side as the hip being moved is referred to as loaded with the innominate on the contralateral (fixed) side referred to as unloaded. Thus SIJ motion is described in three angles of innominate movement, Loaded, Unloaded (in the transverse plane) and Sagittal (in the sagittal plane). 2.4. Data analysis Data reduction was undertaken with a purpose-written MatlabÒ 4 routine and analysis consisted of computations of the mean and standard deviation of the innominate bone and hip range of motion measured across all participants for each trial. To evaluate the patterns of innominate motion between males and females we described each participant’s innominate angles as a function of hip rotation. Each participant’s data were evaluated to determine the best polynomial fit (either a first or second order). All statistical analyses were performed in SAS v9 (SAS Institute, Cary NC). The non-independent data of this repeated measures design were accommodated using linear mixed models (Fitzmaurice et al., 2004). The outcome variables were the three innominate angles: loaded, unloaded, and sagittal. We considered the following factors in our models: hip position (ER, AB and ER þ AB); side (left, right); sex (male, female) and incremental rotations (six 10 steps). Exploratory analyses accepted these rotations as continuous variables with good data fit allowing computation of linear models. Random coefficient models were used with intercept and rotation as effects, allowing each participant’s angles to have their own linear trajectory as a function of rotation. We begun with full models including all interactions, and used backwards selection to choose a final model. Estimates of sex differences were then calculated from these models. To avoid type I errors we used a cutoff value of a ¼ 0.01 for interaction terms.

4

The MathWorks, Inc., Natick, MA.

516

M.D. Bussey et al. / Manual Therapy 14 (2009) 514–519

3. Results The means and standard deviations of the mean maximal innominate and passive hip range of motion in each position of ER, ABD and ABER are presented in Table 2. Mean hip rotation ranged from 10 to 53 of ER and 9.8 to 56 of ABD. Both sexes were capable of greater ROM in the ABER condition with abduction ranging from 10 to 58 and external rotation ranging from 50 to 75 . Further, while we made every effort to limit the amount of coupled rotation of the hip using the standardization frame, there was still some evidence of coupled rotation. For example, ER was coupled with hip flexion (maximal 9 ), and ABD was coupled with some ER of the hip (maximal 15 ). We consider this motion coupling is inevitable to allow for maximal ROM without physiological disruption at the joint. We evaluated each individual’s innominate motion as a function of hip rotation to determine whether the pattern was best estimated as linear (first order polynomial) or non-linear (second order polynomial). Overall the first order polynomial was the most reasonable fit to the raw data, as shown in Fig. 1. From this analysis, we made several generalizations, which may be observed in Fig. 1: first, it is clear from the positive slopes of the linear line estimates for ER and AB that innominate angles increase with hip rotation but do not in the ER þ AB position. Second, for all hip positions mean sagittal innominate angles were greater then loaded and unloaded angles for both sexes. Third, that for all hip positions the female line estimates for both loaded and sagittal are above the males, suggesting that females have greater mean innominate angles than males. Finally, that there is little difference between males and females in unloaded innominate angles. Some of our observations were confirmed as statistically interesting by the linear mixed model. First, it was apparent that the slopes of the lines for the loaded angles are statistically significantly different between the sexes (Table 3). Female loaded angles appear to increase at a greater rate than males. From Fig. 1 it appeared that magnitudes of sagittal innominate angles are greater for women, however, the slope estimates were not statistically distinguishable from the males (Table 3). This means that while females may have greater initial sagittal angles, the rate at which their angles increase is no different from the rate of increase in males. Yet our statistical analysis revealed a sex by side interaction effect, which says that there is a difference in sagittal angles between males and females but it is dependent upon the side (Table 3). Upon further investigation, we found that females had larger sagittal angles on the left and males had larger sagittal angles on the right (Table 2). It appeared that male and female innominate bones responded differently depending on which hip was under load. The women tended to experience larger loaded angles

coupled with smaller sagittal angles when the right leg was stressed but larger sagittal angles coupled with smaller loaded angles when the left leg was stressed (Table 2). In males, load applied to the right leg resulted in greater sagittal angles but made no difference in loaded angles (Table 2). In order to explore the nature of the difference(s) in innominate bone motion between males and females we compared the direction of the load side (same side as the stressed limb) and non-load side (contralateral to stressed limb) innominate bones. Firstly, we isolated the sagittal angle or motion about the sagittal-horizontal axis. Both males and females displayed similar patterns of motion about the sagittal-horizontal axis in that the innominate bones were found to rotate reciprocally, rotating in opposite directions about the axis (Fig. 2B) under load. However, when the motion about the vertical axis was isolated, it appeared the males and females had different strategies for achieving maximal rotation. In almost all hip positions and on both sides, males experienced reciprocal rotation about the vertical axis, where, for example, when the left hip was stressed the left (loaded) innominate experienced counter clockwise rotation and the right (unloaded) innominate experienced clockwise rotation about the vertical axis (Fig. 2A). Conversely, in maximal hip positions, females experienced a unilateral rotation of the innominate bones about the vertical axis, particularly on the right side, where both the loaded and unloaded innominate bones rotated in a clockwise direction about the axis (Fig. 2A). 4. Discussion Our results support the hypothesis that women have a greater overall innominate ROM and are in agreement with Sturesson and colleagues who found that male SIJ ROM was 30–40% less than female SIJ ROM (Sturesson and Uden, 1989; Sturesson, 1997; Sturesson et al., 2000a, 2000b). However, these researchers never postulated as to why males and females differed in SIJ ROM. In the present study, the methods we used cannot explain the disparity in ROM between sexes since both males and females were placed in the same position with the pelvis unloaded in a neutral posture, and both groups were stressed with the same magnitude and direction of external force. The only way that the method would have influenced the outcome was if the structural differences in the pelvis and hip gave one group a mechanical advantage over the other. Due to the passive nature of the study, we assume that muscle activity cannot account for changes in joint stiffness. Therefore, the results of the present study point to differing viscoelastic responses in the pelvic ring between sexes. Based on the understanding that males and females differ in SIJ geometry and possible ligament elasticity (Vleeming et al, 1990;

Table 2 Mean Maximal hip angles (deg) and the corresponding Mean Maximal Innominate Angles for Females (F) and Males (M) measured in each hip position (St.D). Hip Position

External Rotation (ER)

Mean Max Hip Angle

F M

Abduction (AB)

F M

Combination (ER þ AB)

F M

Mean Max Innominate Angle Loaded

Unloaded

Sagittal

52.3 (3.6) 53.0 (5.1) 49.1 (3.1) 50.3 (3.5)

Left Right Left Right

1.5 3.0 1.6 1.6

(0.3) (0.3) (0.3) (0.4)

1.7 2.2 2.3 1.5

(0.3) (0.7) (0.3) (0.7)

4.0 3.2 2.6 3.4

(0.5) (0.6) (0.6) (0.7)

52.7(3.9) 55.5 (3.5) 48.6 (4.3) 51.7 (2.8)

Left Right Left Right

2.2 2.6 1.4 1.3

(0.3) (0.4) (0.3) (0.3)

1.7 1.4 1.7 1.5

(0.6) (0.4) (0.3) (0.4)

3.8 2.7 3.4 3.6

(0.6) (0.7) (0.6) (0.6)

74.5 (7.2) þ 57.5 (4.5) 75.2 (6.1) þ 58.6 (2.8) 72.7 (4.3) þ 48.6 (2.0) 70.8 (12.0) þ 52.1 (3.3)

Left Right Left Right

1.9 1.6 1.6 1.9

(0.3) (0.3) (0.3) (0.3)

1.7 1.8 2.2 1.6

(0.3) (0.5) (0.7) (0.3)

3.3 2.7 2.5 2.3

(0.6) (0.5) (0.6) (0.7)

M.D. Bussey et al. / Manual Therapy 14 (2009) 514–519

517

Fig. 1. Male and female pelvic displacement angles displayed as a linear function of hip rotation for each of the ER, AB and ER þ AB positions. Note only 50 of rotation are shown to standardize the graphs because not all positions had participants who reached 60 of rotation.

Bechtel, 1998), we hypothesized that applying a passive load to the hip in the same manner, would result in different patterns of motion between sexes. Indeed, we found that females experienced a greater rate of increase in the loaded innominate motion, which suggests females and males differ in their viscoelastic response to load directed about the vertical axis. Further, the female SIJ appear to be less stiff about the sagittal-horizontal axis as females had slightly greater initial sagittal angles but similar sagittal slope Table 3 Estimated mean angles and 95% confidence intervals from a backwards selected mixed model which began with all possible interactions of position, side, sex, and rotation, and which used random effects of intercept and rotation. Sex

mean

95% CI for the difference

p-value

Loaded (degrees)

female male

1.63 1.29

(0.093, 0.60)

0.008

Unloaded (degrees)

female male

1.24 1.16

(-0.12, 0.29)

0.4

Sagittal* (degrees)

female left male left female right male right

3.08 2.10 2.31 2.47

(0.26, 1.72)

0.009

(-0.88, 0.57)

0.7

*p-value for sex  side interaction < 0.0001.

estimates as males. Thus, females experience a greater range of motion initially but after the initial rotation, the rate of change in innominate angle is the same as in males, and viscoelastic responses do not differ when load is directed about the sagittalhorizontal axis. Therefore, it appears that a more likely explanation for the sex disparity is the combination of differing articular surface geometry (Bechtel, 1998) and viscoelastic behaviour resulting in differing responses to the standardised load applied in the present research. There also appears to be a great deal of individual variation in the pattern of motion of the innominate bones as load increased. The deformation about the vertical axis was expected to be characterized by both the loaded and unloaded innominate bones moving unilaterally in the direction of applied load since it was hypothesized in a previous study (Bussey, 2004) that this pattern was most likely to reduce the dislocation torque on the pubic symphysis when placed under large deformation load. However, the pattern of innominate motion tended to be sexdependent. Females tended to experience external and posterior rotation of the load side innominate bone accompanied by internal and neutral or anterior rotation of the unloaded side innominate bone (namely a unilateral pattern of motion about vertical axis and a reciprocal pattern about the sagittal-

518

M.D. Bussey et al. / Manual Therapy 14 (2009) 514–519

Fig. 2. A) Depiction of axial rotation (about the vertical axis) of the innominate solid arrows show the unilateral (female) pattern while striped arrows show the reciprocal (male) pattern. B) Depiction of transverse plane rotation of the innominate striped arrows show the reciprocal pattern.

horizontal axis). Males, on the other hand, tended to experience external and posterior rotation of the load side innominate bone accompanied by external and anterior rotation of the unloaded side innominate bone (thus a reciprocal pattern about both the vertical and sagittal-horizontal axes). We feel that the unilateral pattern of the female innominates allowed them greater end range of motion about the vertical axis by reducing the stress on the pubic symphysis. Previous research has suggested that bilateral asymmetry in hip external rotation and abduction is associated with LBP (BarbeeEllison et al., 1990; Chesworth et al., 1994; Gombatto et al., 2006). Therefore, we did not expect bilateral asymmetry to occur within this healthy population. Indeed, the present study found no significant differences in hip range of motion between left and right sides. However, there were some bilateral effects in the innominate range of motion. In female’s passive rotation of the right limb led to a greater innominate ROM about the vertical axis, whereas rotation of the left limb led to a greater innominate ROM about the sagittalhorizontal axis. Males displayed a slightly different pattern under right limb rotation having greater innominate ROM about the sagittal-horizontal axis. These findings point to differing dominant axes for each innominate but further differing dominant axes according to sex. The idea that there might be a different dominant axis for each innominate bone has previously been proposed (Lavignolle et al., 1983; Plochocki, 2002) but has not been demonstrated. The motion of the innominate bones is unique within the body, as they are connected in a three link closed kinetic chain, where motion about one axis occurs in conjunction with/or at the expense of motion about another axis. For this reason, some researchers have used a dominant three-dimensional axis (a helical axis) to describe the motion of the innominate bones (Sturesson and Uden, 1989; Jacob and Kissling, 1995). However, while helical axes are probably a more superior way to describe the 3D rotation of the innominate bones, they also require greater precision in measurement and are more difficult to interpret clinically. 5. Conclusions Males and females appear to have different load transfer patterns across the pelvic ring, although further research is required to

explore their significance in low back and pelvic pain and injury. Asymmetry in innominate bone motion occurs when passively displacing either the dominant or non-dominant leg, at least in these healthy individuals. Further, these findings suggest that clinicians should at least recognise limb dominance when testing SIJ mobility and specifically before labelling a joint dysfunctional due to hypo- or hyper-mobility. Further research on LBP patients is required to determine if and how the pattern of innominate motion is affected by pain and how the patterns of innominate motion relate to the motion of the hip in injured individuals. Acknowledgments This research was supported by funding from the School of Physical Education, University of Otago. The authors thank all subjects for consenting to take part in this project and also thank Associate Professor Peter Milburn for his support. References Barbee-Ellison JB, Rose SJ, Sahrmann SA. Patterns of hip rotation range of motion: comparison between healthy subjects and patients with low back pain. Physical Therapy 1990;70:537–41. Bechtel RH. Biomechanical properties of the axial interosseous ligament and surface topology of the human sacroiliac joint. Unpublished PhD thesis, University of Maryland, College Park; 1998. Brunner C, Kissling R, Jacob HAC. The effects of morphology and histopathologic findings on the mobility of the sacroiliac joint. Spine 1991;16(9):1111–7. Bush TR, Gutowski PE. An approach for hip joint center calculation for use in seated postures. Journal of Biomechanics 2003;36(11):1739–43. Bussey MD. Motion characteristics of the pelvic-hip complex under passive static loads. Unpublished PhD, University of Otago, Dunedin; 2004. Bussey MD, Bell ML, Milosavljevic S. The influence of hip abduction and external rotation on sacro-iliac motion. Manual Therapy 2009;14(5):520–5. Chesworth BM, Padfield BJ, Helewa A. A comparison of hip mobility in patients with low back pain and matched healthy subjects. Physiotherapy Canada 1994;46: 267–74. Cibulka MT, Sinacore DR, Cromer GS, Delitto A. Unilateral hip rotation range of motion asymmetry in patients with sacroiliac joint regional pain. Spine 1998;23(9):1009–15. Fitzmaurice GM, Laird NM, Ware JH. Applied longitudinal analysis. New Jersey: Wiley; 2004. Friberg O. Clinical symptoms and biomechanics of lumbar spine and hip joint in leg length inequality. Spine 1983;8:643–51. Gombatto SP, Collins DR, Sahrmann SA, Engsberg JR, Van Dillen LR. Gender differences in pattern of hip and lumbopelvic rotation in people with low back pain. Clinical Biomechanics 2006;21(3):263–71.

M.D. Bussey et al. / Manual Therapy 14 (2009) 514–519 Jacob HAC, Kissling RO. The mobility of the sacroiliac joints in healthy volunteers between 20 and 50 years of age. Clinical Biomechanics 1995;10(7): 352–61. Kristiansson P, Svardsudd K, Von Schoultz B. Serum relaxin, symphyseal pain, and back pain during pregnancy. American Journal of Obstetrics and Gynecology 1996;175(5):1342–7. Lavignolle B, Vital JM, Sengas J, Destandau J, Toson B, Bouyx P, et al. An approach to the functional anatomy of the sacroiliac joints in vivo. Anatomia Clinica 1983;5:169–76. LeBan MM, Meerschaert JR, Taylor RS, Tabor HD. Symphyseal and sacroiliac joint pain associated with pubic symphysis instability. Archives of Physical Medicine and Rehabilitation 1978;59:470–2. MacLennan AH, Nicolson R, Green RC, Bath M. Serum relaxin and pelvic pain of pregnancy. Lancet 1986;2:243–5. Mellin G. Correlation of hip mobility with degree of back pain and lumbar spinal mobility in chronic low back pain patients. Spine 1988;13:668–70.

519

Offierski CM, White DJ. Hip-spine syndrome. Spine 1983;8:316–21. Plochocki JH. Directional bilateral asymmetry in human sacral morphology. International Journal of Osteoarchaeology 2002;12:349–55. Sturesson B. Movement of the sacroiliac joint: A fresh look. In: Vleeming A, Mooney V, Dorman T, Snijders CJ, Stoeckart R, editors. Movement stability and low back pain: the essential role of the pelvis. New York: Churchill Livingstone; 1997. p. 171–6. Sturesson BGS, Uden A. Movements of the sacroiliac joints: a roentgen stereophotogrammetric analysis. Spine 1989;14(2):162–5. Sturesson B, Uden A, Vleeming A. A radiostereometric analysis of movements of the sacroiliac joints during the standing hip flexion test. Spine 2000a;25(3): 364–8. Sturesson B, Uden A, Vleeming A. A radiostereometric analysis of the movements of the sacroiliac joints in the reciprocal straddle position. Spine 2000b;25(2):214–7. Vleeming A, Volkers A, Snijders J, Stoeckart R. Relation between form and function in the sacroiliac joint. Part II: Biomechanical aspects. Spine 1990;15:133–5.

Manual Therapy 14 (2009) 520–525

Contents lists available at ScienceDirect

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

Original Article

The influence of hip abduction and external rotation on sacroiliac motion Melanie D. Bussey a, *, Melanie L. Bell b,1, Stephan Milosavljevic c, 2 a

School of Physical Education, University of Otago, PO Box 56 Dunedin, New Zealand Department of Preventive and Social Medicine, University of Otago, Dunedin, New Zealand c Centre for Physiotherapy Research, School of Physiotherapy, University of Otago, PO Box 56, Dunedin, New Zealand b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 29 November 2007 Received in revised form 3 August 2008 Accepted 25 August 2008

Although the sacroiliac joint (SIJ) is conventionally accepted as a sagittal joint with little mobility in other planes, recent research has shown evidence for reduced hip abduction and axial rotation in patients with sacroiliac pain. A sample of healthy individuals was investigated to determine whether innominate motion about the sacroiliac joint can be predicted from abduction and external rotation displacement of the femur. The motion of the innominate and femur were tracked as the hip was passively rotated by standardized increments of 10 into (1) abduction; (2) external rotation; and (3) a combination of external rotation and abduction. Although sagittal and transverse plane innominate motion both increased significantly as the hip was rotated further into either abduction or external rotation, external rotation was the strongest predictor of change in innominate angle. A combination of external rotation and abduction led to greater increases in these innominate angles at a smaller degree of hip rotation. The results support the use of abduction and external rotation hip displacements (both singularly and in combination) for assessing SIJ mobility at least in the axes investigated. Further research that investigates the use of these tests in people with SIJ disorders is warranted. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Sacroiliac joint Range of motion Kinematics Pelvic bones Pelvis Hip joint Femur

1. Introduction The pelvis is the bony link between the spine and lower limbs, transferring load from the trunk to the legs and vice versa. The sacroiliac joints (SIJs) transfer load between the innominates and spine and act to attenuate forces (Miller et al., 1987). Abnormal force attenuation is a likely contributor to SIJ dysfunction and may also contribute to unexplained low back pain (LBP) (Egund et al., 1978; McGill, 1987; Porterfield and DeRosa, 1991; Jacob and Kissling, 1995). Previous SIJ studies have found minor joint rotations and translations in the three anatomical planes (Egund et al., 1978; Sturesson et al., 1989; Jacob and Kissling, 1995). Grieve (1982) theorized these small rotations were a result of insufficient load application in end range hip positions and higher loads would be required to create significant ligament elongation and maximum innominate rotation. Sturesson et al. (2000b) investigated SIJ motion during weight bearing in a maximal hip reciprocal straddle position and found no increase in SIJ range of motion (ROM).

* Corresponding author. Tel.: þ64 3 479 8981; fax: þ64 3 479 8309. E-mail addresses: [email protected] (M.D. Bussey), melanie.bell@ otago.ac.nz (M.L. Bell), [email protected] (S. Milosavljevic). 1 Tel.: þ64 3 479 7236; fax: þ64 3 479 7298. 2 Tel.: þ64 3 479 7193; fax: þ64 3 479 8414. 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.08.009

However Sturesson (1997) had previously found consistent increases in SIJ ROM in prone testing postures, with a unilateral load applied via the hip. Thus, the SIJs appear to be capable of greater ROM in prone due to the reduced gravitational effect on form closure. The temporal properties of such SIJ loading have been investigated in cadavers with results suggesting that a time dependent creep response should be considered when investigating the influence of loading on the SIJ (Vleeming et al., 1992). The SIJ is traditionally considered a sagittal joint with little mobility outside this plane (Jacob and Kissling, 1995; Snijders et al., 1997; Sturesson et al., 2000a). However there is evidence for reduced abduction and axial rotation of the hip in patients diagnosed with sacroiliac pain (LeBan et al., 1978; Fowler, 1986; Cibulka et al., 1998). Recently Bussey et al. (2004) used CT scanning to demonstrate transverse plane SIJ motion during maximal abduction–external rotation of the hips in prone lying. Despite this evidence for non-sagittal movement there has been no research investigating functional relationships between hip abduction, hip rotation and innominate movements at the SIJ. Theoretically a functional relationship would be demonstrated by a cause and effect association where innominate motion is determined by hip rotation (Sauerbrei and Royston, 2002). Our aim was to determine whether innominate motion can be predicted from incremental abduction and external rotation displacements of the femur and if so, determine which position is the best predictor of innominate motion.

M.D. Bussey et al. / Manual Therapy 14 (2009) 520–525

2. Methods 2.1. Design A cross-sectional study of healthy individuals where a randomized block design of three hip positions, external rotation (ER), hip abduction (AB) and the combination (ER þ AB) was used. Participants were allocated to three random-order blocks ([ER, AB, ER þ AB]; [AB, ER þ AB, ER]; [ER þ AB, ER, AB]) for testing. Within each of these hip positions, participants’ lower limbs were rotated in six 10 increments or until end range was reached. Thus our independent variables were hip position (ER, AB and ER þ AB), and rotation (six 10 increments), while our dependent variable was pelvic angle defined in three separate outcome measures of innominate motion: firstly, movement of the loaded innominate in the transverse plane, secondly movement of the unloaded innominate in the transverse plane and finally movement of both innominates relative to each other in the sagittal plane. 2.2. Participants This study was approved by the Otago Human Ethics Committee. Forty participants aged between 18 and 35 were recruited. There were 20 females (25.0 years SD 3.1) with a mean BMI of 20.7, SD 2.0 kg/m2 and 20 males (23.3 years SD 4.3) with a mean BMI of 22.5 SD 2.7 kg/m2. Thirty-six of the 40 participants were right hand dominant and all were actively involved in 8.9 hours (SD 4.1) of physical activity weekly. Participants were included if they were healthy and free of back, hip or pelvic pain /injury. 2.3. Procedure Kinematic data were collected with a magnetic tracking device,1 consisting of a transmitter, four receivers, a digitizer and a systems electronics unit. Measurement error of the system in the x, y and z coordinates of each of the four pelvic points was 0.02 mm (SD 0.84 mm) on the x-axis, 0.07 mm (SD 0.82 mm) on the y-axis and 0.03 (SD 0.99 mm) on the z-axis. A global coordinate system was established by mounting the transmitter to a rigid wooden support. The receivers were mounted to thermoplastic frames and secured to the femurs and the S1 vertebra with double-sided tape and VelcroÒ support straps. A hip rotation frame (Fig. 1) was used to standardize the rotational increments applied to the femurs. Standardizing these rotational increments allowed standardization of the external load2 applied to the innominate. The frame was constructed such that only motions in the frontal and transverse planes were allowed, thus loading the innominate about the long and anteroposterior axes. A palpation and digitizing technique known to accurately and reliably measure innominate motion (Bussey et al., 2004) was used to calculate motion of the innominate bones in reference to their initial static positions. This involved the palpation and digitization of the anterior (ASIS) and posterior (PSIS) iliac spines with the tracking stylus. Each palpated landmark was digitized several times in the reference position (hip at 0 ); the leg was passively rotated in 10 increments and the procedure repeated (Fig. 1). The motion of the innominate in the sagittal and transverse planes of the pelvis reference system was calculated as angular displacement between the reference position and each subsequent 10 hip rotation.

1 Polhemus 3Space FastrackÒ, 40 Hercules Drive, PO Box 560, Colchester, VT 05446, USA. 2 Internal load may vary due to individual anatomical differences in muscle and ligament stiffness.

521

Transverse plane motion was calculated as absolute displacement of left and right bones individually about the long axis. For this calculation the innominate on the same side as the hip being incrementally displaced is referred to as loaded with the innominate on the contralateral side referred to as unloaded (Fig. 2). Sagittal plane motion was calculated as a composite angle between innominates rotating about the mediolateral axis of the pelvis (Fig. 2). Thus SIJ motion in this study is operationally defined in three angles of innominate movement with respect to a fixed sacrum: loaded, unloaded (in the transverse plane) and sagittal (in the sagittal plane). The three anatomical hip positions were applied in a random order for each subject. For ER, the thigh was locked into neutral and only the lower leg was rotated to apply an increasing magnitude of hip external rotation (Fig. 1A). For hip AB, only the thigh component of the frame was rotated to apply an increasing magnitude of hip abduction displacement (Fig. 1B). For hip ER þ AB, both components of the frame were used to apply an increasing magnitude of abduction combined with external rotation to the hip joint until end range was reached (Fig. 1C). A maximum of six incremental rotations (10 each) for both ER and AB were held for 2 min to allow optimal soft tissue response at the SIJ (Vleeming et al., 1992). At each incremental rotation the ASIS and PSIS landmarks were digitized using the tracking stylus.

2.4. Data analysis Data reduction involved MatlabÒ3 computation of innominate and hip angles measured across all participants for each trial. All statistical analyses were performed in SAS v94. The non-independent data of this repeated measures design was accommodated using linear mixed models (Fitzmaurice et al., 2004). Three models were fit. The outcome variables were innominate angles: loaded, unloaded, and sagittal. We considered the following factors in our models: hip position (ER, AB and ER þ AB); side (left, right); sex (male, female) and incremental rotations (six 10 steps). Exploratory analyses accepted these rotations as continuous variables with good data fit allowing computation of linear models. Random coefficient models were used with intercept and rotation as effects, allowing each participant’s angles to have their own linear trajectory as a function of rotation. We began with full models that included all interactions, and used backwards selection to choose a final model. To avoid type I errors we used a cut-off value of a ¼ 0.01 for interaction terms. Although sex is not the focus of this research, it was included in the models to improve precision. To test the difference in the initial changes (from 0 to 10 ) in the pelvic angles between the three different hip positions and sides, a two-factor ANOVA (the factors are hip position, ER, AB, ER þ AB and side: L, R) was used for each of the three outcomes. A separate reliability study was conducted to examine the test– retest reliability of the frontal and sagittal plane innominate measures using the intra-class correlation coefficient (ICC) and standard error of measurement (SEM) (Atkinson and Nevill, 1998). The global average value of imprecision in the measurement of a point for intra-observer reliability was 0.80 mm (SD 1.47 mm). The test–retest reliability revealed ICCs of 0.977 (95% CI: 0.941–0.992) for the loaded, 0.971 (95% CI: 0.941–0.994) for the unloaded, and 0.993 (95% CI: 0.982–0.998) for the sagittal angles. The SEM was 0.04 for loaded angles, 0.05 for unloaded angles, and 0.02 for sagittal angles.

3 4

The MathWorks, Inc., 3 Apple Hill Drive, Natick, MA 01760-2098, USA. SAS Institute, Cary, NC, USA.

522

M.D. Bussey et al. / Manual Therapy 14 (2009) 520–525

Fig. 1. Participant set-up in standardizing frame for the right side ER (A), AB (B) and ER þ AB (C) positions. An anatomical reference system for the hip was defined using palpated and digitized bony landmarks including: greater trochanter (GT), medial and lateral femoral epicondyles (ME and LE), ASIS and PSIS, and L1 and L5 for the lumbar spine. The hip joint centre was defined with a predicative method that used a series of regression equations based on each individual’s pelvic anthropometrics (Seidel et al., 1995).

3. Results Mean maximal angles and standard deviations (SD) for the innominates and hip are presented in Table 1. For all participants maximal hip rotation was achieved within either five or six increments in all positions. Mean hip rotation was calculated for each increment and ranged from 10.4 (in rotation 1) to 56.2 (in rotation 5) in ER, and from 9.8 (in rotation 1) to 58.4 (in rotation 6) of AB (Table 1). The combined position of ER þ AB led to greater external rotation displacement of the hip showing obvious signs of motion coupling. Interestingly there were also signs of motion coupling within the single hip positions as well. In the ER position the external rotation was coupled mainly with a secondary hip flexion effect (mean of 9 at maximal ER) and in the AB position abduction was coupled with lateral axial rotation of the hip (mean of 15 at maximal AB). To determine whether rotation of the hip can be used to predict innominate angle we looked at the linear plots of innominate bone angle against hip rotation (Fig. 3). Fig. 3 shows that each of the innominate angles for both the ER and AB positions yield positive slopes, which indicate that innominate angles increase with hip rotation. However, this graph also shows that the estimated lines for both loaded and sagittal innominate angles in the combined

position ER þ AB are close to flat (slope close to 0). Hip rotation was a highly statistically significant predictor for each of the innominate angles (p < 0.0001) as shown in Table 2. Additionally, there are significant rotation by position interaction effects for each innominate angle, which means that the slope for rotation depends on hip position. Our second study aim was to determine which hip position was the best predictor of innominate angle. Table 2 presents the slope estimates for hip rotation vs. innominate angle for all hip positions. The slope estimates for each of the angles differ significantly by hip position. The AB position had a slightly larger slope for sagittal plane innominate angle than ER, although these slopes were not statistically different (p ¼ 0.5). Yet, the ER position was the best predictor of loaded innominate angle as shown by the magnitude of the slopes. Interestingly, the slope estimates for rotation in both the AB and ER positions were larger on the right than on the left side. The loaded innominate angles showed a statistically significant rotation by position by side interaction effect, meaning that the slope for rotation differs by position but the difference is dependent upon the side. It appears that the loaded right side innominate under load experienced greater motion with increasing hip rotation than the loaded left side innominate, as indicated by the larger slope estimates for the right side AB and ER positions.

Fig. 2. Calculation of innominate bone angles in transverse (A) and sagittal (B) planes. (A) Transverse plane motion of left (blue) and right (red) innominates about the long axis of the pelvis. (B) Sagittal plane motion of the innominates where motion is described as the difference between left and right rotations about the mediolateral axis. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

M.D. Bussey et al. / Manual Therapy 14 (2009) 520–525

523

Table 1 Mean maximal innominate and passive hip angles measured in each hip position with standard deviations. Angles

Innominate

Loaded (degrees) Unloaded (degrees)

Hip

Sagittal (degrees) Passive (degrees)

Hip position

L R L R L R L R

External rotation (ER)

Abduction (AB)

Combination (ER þ AB)

3.41 3.38 3.08 2.39 6.40 4.45 52.5 56.2

2.46 3.46 2.88 2.82 7.62 5.58 54.5 58.4

3.30 4.01 2.98 3.37 7.05 5.39 75.4 75.2

The combined position of ER þ AB was not shown to be a good predictor of innominate angle as the slopes were not significantly different from zero (Fig. 3 and Table 2). However, we did find that the combined position yielded significantly more innominate bone rotation for both loaded and sagittal innominate motion in the initial hip rotation (first 10 ) (Table 3), indicating that the majority of the movement occurs within the initial rotation. 4. Discussion These prone lying results demonstrate a linear functional relationship between innominate motion and rotation of the hip in positions of abduction and external rotation. These linear patterns are similar to the patterns of SIJ load displacement previously described by Rothkotter and Berner (1988). We also found observable differences in slope between the left and right innominates, suggesting that joint stiffness varies both within and between individuals. Previous research in weight bearing has

(1.09) (1.01) (0.88) (0.55) (2.65) (1.14) (5.1) (3.7)

(0.77) (0.63) (0.92) (0.81) (2.75) (1.50) (3.5) (5.1)

(0.48) (0.44) (0.61) (0.84) (1.66) (0.82) (11.2) þ 55.9 (3.8) (6.4) þ 58.6 (5.2)

explored the magnitude of SIJ motion change with increased positional load at the hip (Sturesson et al., 1989, 2000b; Smidt et al., 1997). These studies possibly showed no increase in SIJ motion since tissue mechanics dictates that viscoelastic properties of ligaments require application of load for a substantial time in order to maximize the creep response and develop optimal tissue elongation (Rothkotter and Berner, 1988; Vleeming et al., 1992; Wang and Dumas, 1998). We addressed this temporal loading factor by holding incremental rotations for 2 min where Vleeming et al. (1992) has demonstrated this to be sufficient for allowing maximal creep deformation at the SIJ. These studies were also conducted in a standing posture, where paradoxically this joint is most stable, and where increased stability equals decreased mobility. As our investigations were conducted in prone lying with temporal loading, we hypothesized maximal SIJ mobility as demonstrated with the change in innominate angles. Our findings suggest that transverse plane patterns of innominate angle differ with increasing rotation in the AB and ER þ AB hip

Fig. 3. Pelvic displacement angles as functions of rotation and hip position (external rotation, abduction, combination). The loaded graph shows each of the positions by side because this model estimated a statistically significant interaction of rotation  position  side. This can be interpreted as the change in angles with rotation (slopes) differs by position, but this difference depends on the side.

524

M.D. Bussey et al. / Manual Therapy 14 (2009) 520–525

Table 2 Model summaries and slope estimates for rotation (in degrees of innominate angle per 10 increase in rotation) and 95% confidence intervals. Innominate angle

Loaded (degrees)

Model summarya

Hip position

L R

External rotation (ER)

Abduction (AB)

Combination (ER þ AB)

Effect

p-value

0.21 (0.10, 0.32) 0.37 (0.26, 0.49)

0.070 (0.040, 0.18) 0.32 (0.21, 0.44)

0.070 (0.020,0.16) 0.010 (0.10,0.070)

Position Side Sex Rotation Position  side Rotation  position Rotation  side Rotation  position  side Position Side Sex Rotation Rotation  position

<0.0001 0.8 0.01 <0.0001 0.009 <0.0001 0.009 0.001 0.005 0.9 0.5 <0.0001 0.005

Position Side Sex Rotation Sex  side Rotation  position

<0.0001 0.05 0.1 <0.0001 <0.0001 <0.0001

Unloaded (degrees)

0.18 (0.10, 0.25)

0.23 (0.15, 0.31)

0.090 (0.030, 0.15)

Sagittal (degrees)

0.41 (0.27,0.55)

0.48 (0.34,0.63)

0.060 (0.050,0.17)

Where the slope for rotation depended on side, left (L) and right (R) side estimates are shown separately. a From a backwards selected mixed model which began with all possible interactions of position, side, sex, and rotation, and which used random effects of intercept and rotation. Interactions are indicated with the  symbol.

positions, particularly in the loaded innominate. Generally we also found that the right side SIJ was capable of a greater range of motion than the left as demonstrated in the loaded innominate angles. Vleeming et al. (1992) also observed this phenomenon in their anatomical study of four pelvises where one showed a similar bilateral asymmetry (Vleeming et al., 1992). This result might be a dominance effect associated with the predominantly right handed participant sample. Previous research suggests that hand dominance may have an effect on the surface geometry of the SIJ, where asymmetric variation has been seen in size of sacral articulations and ala (Bechtel, 1998; Plochocki, 2002). The articular surface geometry of our sample is unknown and it is also unknown whether side differences were due to such sacral asymmetries. However, SIJ mobility is ultimately limited by position within a three-link closed kinematic chain, and articular surface geometry will have an effect on the range and direction of motion available to the innominate bones. The second aim of the study was to determine which hip position was better at predicting innominate displacement. While the AB position was slightly better at predicting sagittal plane motion the ER position resulted in the most consistent overall (including left and right) hip rotation slope estimates, and the greatest loaded innominate bone displacement about the long axis of the pelvis. Previous researchers have applied load to the SIJ only in the sagittal plane, testing flexion/extension of the hip or spine, directing maximal load about the mediolateral pelvic axis (Egund et al., 1978; Sturesson et al., 1989; Smidt et al., 1995). In this study load was directed about the long and anteroposterior axes of the

pelvis and femur via external rotation and abduction respectively, giving a clearer understanding of maximal SIJ ROM available outside the sagittal plane. Our findings fit a basic assumption that the innominate bone will generally displace in the same direction as the hip. Thus, the innominate bones act as an extension of the femur and may be called upon to extend the hip range of motion at least in these tested directions (albeit to a small degree). A final finding of the study was the effect of the combined ER þ AB position. While maximum innominate angle for the ER and AB positions was usually reached by the fourth or fifth hip rotation increment, the maximal innominate angle for the ER þ AB position was reached within the first or second hip rotation. This disparity points to differing joint stiffness in response to the direction of the load applied to the joint but this requires verification using Richardson et al.’s (2002) method of testing laxity. Although SIJ stiffness may be affected by activation of the hip and back musculature (van Wingerden et al., 2004) we consider that hip positions were conducted with passive participants and, with the knee held at 90 , we minimized active muscle recruitment. Even though eccentric tension in the piriformis has also been hypothesized as a factor in SIJ stiffness (Snijders et al., 2006), the influence of eccentric muscle activity on joint stiffness is unknown and cannot be accounted for. Axial collinearity between femoral and innominate rotational movement during the ER þ AB rotation should also be considered. As the SIJ is not planar, the main axis of rotation is not in line with the principal axes of the pelvis; rather it is an oblique axis directed either through the SIJ and sacrum or SIJ and innominate, depending

Table 3 Initial changes in pelvic angles, from 0 to 10 in each of the starting hip positions, and 95% confidence intervals. Hip position

Loaded (degrees) Unloaded (degrees) Sagittal (degrees)

L R L R L R

p-value

External rotation (ER)

Abduction (AB)

Combination (ER þ AB)

Side

Position

Side  position

1.21 0.90 0.92 0.91 1.61 1.14

0.93 0.86 0.71 0.81 1.79 1.48

1.27 1.75 1.14 1.14 2.68 2.43

0.8

0.0002

0.03

0.7

0.03

0.9

0.2

0.0005

0.9

(0.92,1.50) (0.61,1.20) (0.63,1.22) (0.61,1.21) (1.00,2.23) (0.52,1.76)

(0.64,1.22) (0.56,1.16) (0.41,1.00) (0.52,1.11) (1.18,2.40) (0.86,2.10)

(0.97,1.56) (1.45,2.04) (0.85,1.44) (0.85,1.44) (2.06,3.29) (1.81,3.05)

p-values from the terms of the three two-factor ANOVAs are shown. These indicate, for example, that there is a borderline significant difference (p ¼ 0.03) in the loaded pelvic angle between hip positions, which varies with side (L versus R).

M.D. Bussey et al. / Manual Therapy 14 (2009) 520–525

on whether the load is directed from the spine or from the lower limbs (Wilder et al., 1980; Mitchell, 1995). The motion of the innominate during the ER þ AB position was thus comprised of axial rotation (in the direction of hip displacement) and reciprocal sagittal innominate displacement on the sacrum. When these movements are combined they can be described threedimensionally (3D) as a single rotation about an oblique axis directed through the loaded SIJ. During ER þ AB positioning the 3D axis of the hip would also be described as an oblique axis, passing through the innominate bone. This also supports an underlying assumption of collinearity of axes leading to greater motion in the direction of applied load. 5. Conclusions In the prone testing position, innominate bone motion increased with increasing hip rotation in the AB and ER positions, demonstrating unique movement properties for the SIJ in that innominate motion increases in proportion to external load applied. The findings of this study thus support the use of positions with reduced form closure, such as the prone lying position for testing the movement integrity of the SIJ – at least in this asymptomatic sample. Clinical texts agree with such an argument (Hoppenfeld and Hutton, 1976; Magee, 1992; Prentice and Arnheim, 2002), describing that tests of joint function should be conducted in a loaded or close-packed position, while the loose-packed or unloaded position should be used for tests of joint integrity. Our results show that the pelvic joints are considerably stiffer when motion is not directed about the main axis of rotation for the innominate bone. Uni-planar positioning (ER and AB) of the femur did not produce significant displacements of the innominates until about 30–40 of hip rotation yet multi-planar (ER þ AB) position led to significant innominate displacement within 10–20 . Sacroiliac joint involvement in LBP is often overlooked or missed in clinical diagnosis with no singular gold standard test for identifying SIJ contribution to LBP. A clear gap in the clinical testing of the SIJ is the lack of physiological movement investigations that examine the integrity of the inert tissue of the SIJ in a functional way. Abduction and external rotation are essential components of much lower limb function and yet are also known to be a component of movement interaction with the innominate. As recent evidence has shown that low back pain can be associated with limitations of hip range of motion it will be of interest to investigate whether the hip motion tests used in this study would be of use in discriminating which structures (hip, SIJ or lumbar) are implicated in the given disorder. Such investigations will help determine future clinical applicability as either assessment tools or treatment procedures for manual therapists. References Atkinson G, Nevill AM. Statistical methods for assessing measurement error (reliability) in variables relevant to sports medicine. Sports Medicine 1998;26(4):217–38. Bechtel RH. Biomechanical properties of the axial interosseous ligament and surface topology of the human sacroiliac joint. Unpublished PhD thesis, University of Maryland, College Park, 1998. Bussey MD, Yanai T, Milburn P. A non-invasive technique for assessing innominate bone motion. Clinical Biomechanics 2004;19(1):85–90.

525

Cibulka MT, Sinacore DR, Cromer GS, Delitto A. Unilateral hip rotation range of motion asymmetry in patients with sacroiliac joint regional pain. Spine 1998;23(9):1009–15. Egund N, Olsson TH, Schmid H, Selvik G. Movements in the sacroiliac joints demonstrated with roentgen stereophotogrammetry. Acta Radiologica Diagnosis 1978;19:833–45. Fitzmaurice GM, Laird NM, Ware JH. Applied longitudinal analysis. New Jersey: Wiley; 2004. Fowler C. Muscle energy techniques for pelvic dysfunction. In: Grieve GP, editor. Modern manual therapy of the vertebral column. Edinburgh: Churchill Livingstone; 1986. p. 805–14. Grieve EFM. Mechanical dysfunction of the sacro-iliac joint. International Rehabilative Medicine 1982;5(1):46–52. Hoppenfeld S, Hutton R. Physical examination of the spine and extremities. New York: Appleton-Century-Crofts; 1976. Jacob HAC, Kissling RO. The mobility of the sacroiliac joints in healthy volunteers between 20 and 50 years of age. Clinical Biomechanics 1995;10(7):352–61. LeBan MM, Meerschaert JR, Taylor RS, Tabor HD. Symphyseal and sacroiliac joint pain associated with pubic symphysis instability. Archives of Physical Medicine and Rehabilitation 1978;59:470–2. Magee J. Orthopedic physical assessment. 3rd ed. Philadelphia: WB Saunders Company; 1992. McGill SM. A biomechanical perspective of sacro-iliac pain. Clinical Biomechanics 1987;2:1–7. Miller JAA, Schultz AB, Andersson GBJ. Load-displacement behavior of sacroiliac joints. Journal of Orthopaedic Research 1987;5:92–101. Mitchell FJ. The muscle energy manual, vol. 1. East Lansing: MET Press; 1995. Plochocki JH. Directional bilateral asymmetry in human sacral morphology. International Journal of Osteoarchaeology 2002;12:349–55. Porterfield JA, DeRosa R. Mechanical low back pain. Philadelphia: WB Saunders Company; 1991. Prentice WE, Arnheim D. Arnheim’s principles of athletic training: a competencybased approach. 11th ed. Boston: McGraw-Hill; 2002. Richardson CA, Snijders CJ, Hides JA, Damen L, Pas MS, Storm J. The relationship between the transversely oriented abdominal muscles, sacroiliac joint mechanics and low back pain. Spine 2002;27(4):399–405. Rothkotter HJ, Berner W. Failure load and displacement of the human sacroiliac joint under in-vitro loading. Archives of Orthopaedics and Trauma Surgery 1988;107:283–7. Sauerbrei W, Royston P. Determination of functional relationships for continuous variables by using a multivariable fractional polynomial approach. In: Medical data analysis, vol. 2526. Springer Berlin: Heideberg; 2002. p. 53–60. Seidel GK, Marchinda DM, Dijkers M, Soutas-Little RW. Hip joint center location from palpable bony landmarks: a cadaver study. Journal of Biomechanics 1995;28(8):995–8. Smidt GL, McQuade K, Wei S, Barakatt E. Sacroiliac joint kinematics for reciprocal straddle positions. Spine 1995;20(9):1047–54. Smidt GL, McQuade K, Wei S, Barakatt E, Sun T, Stanford W. Sacroiliac motion for extreme hip positions: a fresh cadaver study. Spine 1997;22(18):2073–82. Snijders CJ, Hermans PF, Kleinrensink GJ. Functional aspects of cross-legged sitting with special attention to piriformis muscles and sacroiliac joints. Clinical Biomechanics (Bristol, Avon) 2006;21(2):116–21. Snijders CJ, Vleeming A, Stoeckart R, Mens JMA, Kleinrensink GF. Biomechanics of the interface between spine and pelvis in different postures. In: Vleeming A, Mooney V, Dorman T, Snijders CJ, Stoeckart R, editors. Movement stability and low back pain: the essential role of the pelvis. New York: Churchill Livingstone; 1997. p. 103–13. Sturesson B. Movement of the sacroiliac joint: a fresh look. In: Vleeming A, Mooney V, Dorman T, Snijders CJ, Stoeckart R, editors. Movement stability and low back pain: the essential role of the pelvis. New York: Churchill Livingstone; 1997. p. 171–6. Sturesson B, Selvik G, Uden A. Movements of the sacroiliac joints: a roentgen stereophotogrammetric analysis. Spine 1989;14(2):162–5. Sturesson B, Uden A, Vleeming A. A radiostereometric analysis of movements of the sacroiliac joints during the standing hip flexion test. Spine 2000a;25(3):364–8. Sturesson B, Uden A, Vleeming A. A radiostereometric analysis of the movements of the sacroiliac joints in the reciprocal straddle position. Spine 2000b;25(2):214–7. van Wingerden JP, Vleeming A, Buyruk HM, Raissadat K. Stabilization of the sacroiliac joint in vivo: verification of muscular contribution to force closure of the pelvis. European Spine Journal 2004;13:199–205. Vleeming A, Van Wingerden JP, Dijkstra PF, Stoeckart R, Snijders CJ, Stijnen T. Mobility in the sacroiliac joints in the elderly: a kinematic and radiological study. Clinical Biomechanics 1992;7:170–6. Wang M, Dumas GA. Mechanical behavior of the female sacroiliac joint and influence of the anterior and posterior sacroiliac ligaments under sagittal loads. Clinical Biomechanics 1998;13:293–9. Wilder DG, Pope MH, Frymoyer JW. The functional topography of the sacroiliac joint. Spine 1980;5(6):575–9.

Manual Therapy 14 (2009) 526–530

Contents lists available at ScienceDirect

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

Original Article

Gender and site of injection do not influence intensity of hypertonic saline-induced muscle pain in healthy volunteers Lisa Loram*, Elienne Horwitz, Alison Bentley Brain Function Research Group, School of Physiology, University of the Witwatersrand Medical School, 7 York Road, Parktown 2193, South Africa

a r t i c l e i n f o

a b s t r a c t

Article history: Received 29 November 2007 Received in revised form 26 August 2008 Accepted 22 September 2008

The aim of the study was to determine whether the same stimulus to different muscles results in comparable pain and whether gender has any influence on the pain. We compared the quality and intensity of muscle pain induced by a hypertonic saline injection into the tibialis anterior (leg) muscle to that after an injection into the lumbar erector spinae (back) muscle in both male (n ¼ 10) and female (n ¼ 10) volunteers. Hypertonic or isotonic saline was injected into the leg and back muscles and pain intensity (visual analogue scale, VAS) and pain quality (McGill Pain Questionnaire) were measured. Pressure pain tolerance around the site of injection and on the contralateral side was measured. Hypertonic saline injection induced significant muscle pain in the back and leg compared to isotonic saline (P < 0.05, ANOVA). The site of injection did not influence the quality of pain but there was a gender bias in the descriptive words chosen (c2 test, P < 0.05) and female subjects were more sensitive to pressure than male subjects. Experimentally induced muscle pain is equivalent in intensity and quality in the leg and back muscle. Gender does not influence muscle pain intensity but does influence sensitivity to pressure and the description of pain. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Hypertonic saline Muscle pain Gender

1. Introduction Muscle pain is a common morbidity affecting all people at some stage of their lives with a prevalence of at least 20% within the western world (Zedka et al., 1999; Capra and Ro, 2004; Brooks, 2005). Clinical diagnosis of the site of muscular injury is based upon the history of the patient’s pain and assessment of the intensity, quality and duration of the local and referred pain (Maitland, 2000). Muscle pain can be induced experimentally in numerous muscle groups by intramuscular injection of hypertonic saline (Koelbaek Johansen et al., 1999; Zedka et al., 1999; Leffler et al., 2000; Ge et al., 2003, 2004a,b; Slater et al., 2003; Hwang et al., 2005; Masri et al., 2005; Gibson et al., 2006; Schmidt-Hansen et al., 2007,) resulting in local and referred spontaneous pain, and an increased sensitivity to a noxious mechanical stimulus or mechanical hyperalgesia, all of which are characteristics of pain present in chronic and acute muscle pathologies (Kellgren, 1938; Graven-Nielsen et al., 1997a,b; Zedka et al., 1999; Graven-Nielsen and Mense, 2001; Capra and Ro, 2004). Manual therapists base their treatment on their assessment of the patients’ pattern of pain. Therefore, it is important to identify

* Corresponding author. Tel.: þ1 720 224 2860; fax: þ1 303 492 2967. E-mail address: [email protected] (L. Loram). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.09.002

whether the same stimulus or injury to a muscle results in comparable muscle pain patterns at different sites of injury. Research investigating the mechanism and characteristics of muscle pain in human experimental models classically use a hypertonic saline injection into the tibialis anterior muscle. However, injury and pathology limited to this muscle is relatively rare and a more clinically relevant site to induce experimental pain would be the lumbar muscles, such as would occur in patients with lower back pain. Therefore, it is important to identify whether the same stimulus or injury to one muscle results in comparable muscle pain patterns at a different site of injury. Using an experimental model of muscle pain, such as injection of hypertonic saline would assist in understanding whether the intensity, and indeed quality, of perceived pain in all muscle groups are comparable. Therefore, the aim of this study was to determine whether an intramuscular injection of hypertonic saline-induced a different intensity, quality and duration of pain in the lumbar erector spinae muscle compared to that of the tibialis anterior muscle. In addition to comparing different sites of pain, gender may influence both the perception and prevalence of muscle pain (Ge et al., 2004a,b; van Selms et al., 2005; Nie et al., 2007). However, it is possible that these differences may be site specific and thus the secondary aim of this study was to identify whether gender influences the intensity and quality of experimentally induced muscle pain in the two different muscle sites.

L. Loram et al. / Manual Therapy 14 (2009) 526–530

2. Materials and methods 2.1. Subjects Ten male (age 25  4 years, mass 72  9 kg and height 1.80  0.07 m, mean  SD) and 10 female (age 24  2 years, mass 57  6 kg and height 1.65  0.07 m) healthy volunteers participated in the study. Volunteers were excluded if they had any palpable trigger points or tender points in the muscles to be injected, if they had a history of chronic musculoskeletal pain or were taking analgesic or anti-inflammatory medication during the study and for up to one week before the initiation of the study. Subjects were asked to refrain from exercise at least 24 h before each injection to prevent any exercise-induced muscle pain. We obtained written informed consent from all subjects before the commencement of the study, and the study was approved by the University of the Witwatersrand Committee for Research on Human Subjects (M040529). A priori, we decided that a clinically significant difference between gender and site of injection would be 10 mm throughout the VAS curve. Using a pilot sample, we added 10 mm throughout the curve and determined the area under the curves. Based on the variability of preliminary data and the calculated area under the curve, we calculated a sample size of 10 per group, required to obtain a power of 0.8 and an alpha of 0.05, using the freeware, G*Power (Faul et al., 2007). 2.2. Intramuscular injections Each subject underwent two sessions at least a week apart. At each session, the subject received one leg and one back injection, on contralateral sides, in random order. An hour before administration of each injection, we applied a topical anaesthetic cream (EMLA 5%, AstraZeneca, Johannesburg, South Africa) to the areas of skin marked for injection. Participants and the researcher conducting the pain measures were both blinded to the type of injection. A qualified medical doctor (AB) injected 0.5 ml of 0.9% isotonic saline or 5% hypertonic saline (SABAX, Johannesburg, South Africa), with a 25 Gauge needle, 20 mm into the anaesthetised site of the tibialis anterior muscle or lumbar erector spinae muscle. With the subject lying supine, we injected into the belly of the tibialis anterior muscle one third of the distance down the leg from the lateral epicondyle of the knee to the lateral malleolus, as described previously (Graven-Nielsen et al., 1997a,b). With the subject lying prone, we injected into the lumbar muscle 40 mm lateral to the second lumbar vertebra. The second injection was administered at least 20 min after the first injection when all pain had subsided from the first injection. 2.3. Assessment of muscle pain The subjects marked their pain intensity on a 100 mm visual analogue scale (VAS) anchored with ‘‘no pain’’ and ‘‘worst pain ever experienced’’. Subjects were asked to mark the VAS at the time of injection (time 0), and every minute for the first 6 min, and at 8 min and 10 min after injections. Once the pain had subsided, the subjects were asked to complete a McGill Pain Questionnaire to investigate the quality of the perceived pain. 2.4. Assessment of pressure pain tolerance Once the pain had subsided, we assessed the subjects’ sensitivity to pressure in the region injected with hypertonic saline. Each subject’s pressure pain tolerance was measured using a pressure algometer (100 mm2 probe, Somedic, Astra, Sweden) applied 20 mm above and below each injection site and at the same site contralaterally. Pressure pain tolerance was taken as the pressure at

527

which the subject could no longer tolerate the applied increasing pressure. The mean of two measurements taken at both sites was recorded as the pressure pain tolerance (KPa). The measurement taken on the non-injected side was taken as a control measurement. 2.5. Statistical analysis The VAS data were normalised using the arcsine transform and analysed with two-way repeated-measures ANOVA to compare the VAS following hypertonic saline with injection site and time as the main effects, and the VAS following hypertonic saline injection with gender and time as main effects for both sites of injection. For the hypertonic saline injections only, two-way repeated-measures ANOVA was used to compare the area under the VAS curve, peak VAS, duration of pain on the VAS and the time at which peak pain occurred, with site and gender as main effects. Student–Newman– Keuls post-hoc test was used when significant effects were found. After data analysis of the VAS scores, means and standard deviations were back-transformed and represented in millimeters from the left anchor. From the McGill Pain questionnaire, the words chosen to describe the quality of the pain by at least 50% of either the male or female subjects were analysed with Fisher’s exact test to detect any significant association between the words chosen to describe the pain experienced in the leg to the pain experienced in the back and between male and female subjects. The pressure pain tolerance was analysed using two-way ANOVA with concentration of saline and site of injection as main effects, irrespective of gender, and concentration of saline and gender as main effects, irrespective of site of injection. Significance was set at P < 0.05. 3. Results 3.1. Intensity of muscle pain The VAS following intramuscular injection of hypertonic saline into the tibialis anterior muscle (leg) and the lumbar erector spinae muscle (back), produced significantly more pain than that after an isotonic saline injection, 1–3 min after injection (ANOVA, P < 0.0001, SNK, Fig. 1). The remaining results pertain to the hypertonic saline injections only. 3.1.1. Site of injection There was no significant difference in the area under the curve (ANOVA, P ¼ 0.37), peak pain (ANOVA, P ¼ 0.25), the time at which the peak pain occurred (ANOVA, P ¼ 1.0) or the duration of the pain (ANOVA, P ¼ 0.78) between the two sites of injections (Fig. 2). 3.1.2. Gender There was no significant difference in the area under the curve (ANOVA, P ¼ 0.90), peak pain (ANOVA, P ¼ 0.90), the time at which the peak pain occurred (ANOVA, P ¼ 0.20) or the duration of the pain (ANOVA, P ¼ 0.44) between male and female subjects (Fig. 2). 3.2. Quality of muscle pain 3.2.1. Site of injection In order to identify an association between the words selected and the site of injection, we compared the two sensory words, ‘‘cramping’’ and ‘‘sore’’, selected by 50% or more of the subjects, to describe the pain induced by a hypertonic saline injection in the leg and back. There was no association between the words selected and the site of injection (Fisher’s exact, P ¼ 1.0, Table 1). ‘‘Annoying’’ was the only emotive word selected by 50% or more of the subjects. There was no significant difference between the number of subjects selecting ‘‘annoying’’ to describe the pain at the different sites of

528

L. Loram et al. / Manual Therapy 14 (2009) 526–530

80

Females

hypertonic leg isotonic leg hypertonic back

VAS (mm)

60

isotonic back

40

20

injection. Therefore the average of the two set of control values were averaged and reflected in the figure. Irrespective of gender, there was no significant difference in the sensitivity to pressure applied to the leg or back (ANOVA, P ¼ 0.23), or between the pressures applied to the control site, after isotonic saline injection or hypertonic saline injection (ANOVA, P ¼ 0.98). We found a significant difference in the sensitivity to pressure applied to the muscle between males and females (ANOVA, P ¼ 0.03), which was present in the control site and remained after injection irrespective of site of injection or concentration of saline injection. 4. Discussion

0

5

10

15

20

Time (min) 80

Males

VAS (mm)

60

40

20

0

0

5

10

15

20

Time (min) Fig. 1. Visual analogue scale score following a bolus (0.5 ml) hypertonic saline (5%) or isotonic saline (0.9%) intramuscular injection into the leg and back muscle at time 0 in female (top panel, n ¼ 10) and male (bottom panel, n ¼ 10) healthy volunteers (mean, SEM). Hypertonic saline injection, regardless of the site of injection, produced significantly more pain 1–3 min after injection than that after isotonic saline injection (P < 0.0001). There was no significant difference in pain intensity between male and female subjects (P > 0.05).

injection (sign test, P > 0.05). ‘‘Spreading’’ and ‘‘nagging’’ were selected by 50% or more of the subjects and belong to the miscellaneous word categories in the McGill Pain questionnaire (categories 17–20). We found no significant association between the words selected and the site of injection (Fishers exact, P ¼ 1.0). 3.2.2. Gender There was a significant association between females selecting ‘‘cramping’’ and males selecting ‘‘sore’’, regardless of the site of injection (Fishers exact, P ¼ 0.02, Table 1). ‘‘Annoying’’ was the only emotive word selected with no significant difference between the number of males versus the number of females selecting the word (sign test, P > 0.05). There was no significant difference between the number of male and female subjects selecting ‘‘spreading’’ and ‘‘nagging’’, belonging to the miscellaneous word categories in the McGill Pain questionnaire (Fisher’s exact test, P ¼ 0.14).

3.3. Pressure pain tolerance Fig. 3 shows the pressure pain tolerance when pressure was applied to the leg or back muscles after an isotonic or hypertonic saline injection as well as control values, which are those obtained from the contralateral side to the site of the isotonic saline injection, which was not significantly different to the contralateral, control site measurement taken following the hypertonic saline

We compared the intensity and the quality of pain induced in male and female volunteers following isotonic saline (0.9%) and hypertonic saline (5%) intramuscular injections into the tibialis anterior muscle and lumbar erector spinae muscle. No significant difference in intensity of perceived pain, in terms of total pain experienced (area under the curve), duration of pain and peak pain was found between male and female volunteers or between the two sites of injection. However, the words to describe the sensory component of the pain, irrespective of the site of injection, was different between males and females, where males described the pain as ‘‘sore’’ and females described the pain as ‘‘cramping’’. In addition, pressure pain tolerance was lower in the female subjects than in the male subjects, irrespective of whether the pressure was applied to the leg or the back and whether the participants received a painful stimulus before the pressure pain tolerance was measured. Therefore, the intensity of experimentally induced muscle pain is not influenced by gender or site of injection, but the pain is described differently by males and females; and females are more sensitive to noxious pressure applied to muscle. We may have detected significant differences in intensity of perceived pain following intramuscular hypertonic saline injections between males and females and between sites of injection if we measured the visual analogue scale continuously rather than in intervals. However, the trend of our curve is comparable to curves found in control subjects of previous studies where the subjects receive hypertonic saline injections in various muscles including the tibialis anterior muscle (Koelbaek Johansen et al., 1999; Gibson et al., 2006; O’Neill et al., 2007). Our sample size was small compared to previous studies investigating gender differences (Chesterton et al., 2003; Ge et al., 2004a,b, 2006). Therefore, the absence of a significant difference may have been insufficient power statistically. It is possible that future studies with larger sample sizes may detect significant differences between groups. The absence of a difference in pressure pain tolerance may have been from the timing of the measurement. If we had measured pressure pain tolerance during the pain rather than afterwards we may have found some differences between sites as has been found previously (Graven-Nielsen et al., 2003; Gibson et al., 2006). While the tibialis anterior muscle is the conventional site for inducing experimental muscle pain with an intramuscular hypertonic saline injection (Graven-Nielsen et al., 1997a,b; Babenko et al., 1999; Bajaj et al., 2001; Weerakkody et al., 2003; O’Neill et al., 2007), numerous other studies have investigated the pain response following hypertonic saline injection into other muscles such as infraspinatus (Koelbaek Johansen et al., 1999; O’Neill et al., 2007), extensor carpi radialis brevis (Slater et al., 2005) and paraspinal multifidus (Cornwall et al., 2006). Although some studies have divided hypertonic injections into two different muscles, the aim of these studies was to compare the pain intensity between controls and the patients with musculoskeletal pathology (Koelbaek Johansen et al., 1999; O’Neill et al., 2007). However, when comparing the area under the VAS curve for the control subjects, there was no significant difference between hypertonic saline

L. Loram et al. / Manual Therapy 14 (2009) 526–530

female (back)

male (back)

female (leg)

male (leg)

Peak VAS (mm)

100

400

200

80 60 40 20

0

0

8

20

6

15

Time (min)

Time of peak VAS (min)

AUC (AU)

600

529

4

10

2

5

0

0

Fig. 2. Box and whisker plots with 95% confidence intervals, of area under the curve (top left), peak VAS (top right (mm)), time of peak VAS (bottom left (min)), and total duration of pain (time) (bottom right (min)) as calculated from the VAS following hypertonic saline injection (5%) into the leg or back muscle in male (n ¼ 10) and female (n ¼ 10) healthy volunteers. There was no significant difference, in any of the variables, between male and female subjects or between the injection into the leg and back muscles (P > 0.05).

injections into the infraspinatus muscle and the tibialis anterior muscle (Koelbaek Johansen et al., 1999). Ours is the first study, therefore, to directly compare hypertonic saline injections in the leg and the back in normal volunteers and we have shown that the intensity of experimentally induced pain in the two sites is comparable. However, whether our results can be extrapolated to other muscles, for example the small muscles of the hand, remains to be investigated. Not only was the pain intensity not different between the two sites of injection, but also neither was the pressure pain tolerance. In addition, there was no difference between the pressure pain tolerance after hypertonic saline injection or isotonic saline. Previous studies investigating the pressure pain threshold, but without hypertonic saline injection, in 10 healthy volunteers, have found significantly lower pressure pain threshold in the trapezius muscle compared to that of the lumbar muscles and gluteal muscles (Potter et al., 2006). Therefore, pressure applied to different muscles may exert different sensitivities dependent on their location, but we found no difference in pressure pain tolerance between the tibialis anterior muscle and the lumbar erector spinae muscle.

Table 1 The percentage of subjects who selected words used by at least 50% of male or female subjects, in one of the sites of injection, to describe the muscle pain experienced after a hypertonic saline (5%) injection into the leg and back in male and female subjects. Leg

Cramping Sore Annoying Spreading Nagging

Back

Male

Female

Male

Female

20 40 60 50 30

70 30 30 10 40

30 60 40 50 30

70* 10* 30 30 50

*P < 0.05 between male and female subjects irrespective of the site of injection.

Although we found no difference in pressure pain tolerance between sites of injection, we found that the female subjects were significantly more sensitive to pressure than males, irrespective of the site. There is a higher prevalence rate of lower back pain in females compared to that of males (Schneider et al., 2006) and females are more sensitive to both pressure pain in the trapezius muscle (Ge et al., 2006) as well as mechanical sensitization from a glutamate injection into the masseter muscle (Svensson et al., 2003). We did not account for the female’s menstrual cycle for these trials. However, previous studies have shown that menstrual cycle did not influence pressure pain threshold (Sjolund and Persson, 2007). It is possible that experiencing menstrual pain may contribute to the ability to describe skeletal muscle pain. In our study, the female subjects described the muscle pain as ‘‘cramping‘‘ while the males described the pain as ‘‘sore’’. Eighty percent of females described dysmenorrhoea using the word ‘‘cramping’’ (Brodie and Niven, 2000), while males have not had a similar pain experience. Further work is required to identify the underlying mechanism for the differences in gender. When assessing the signs and symptoms of a patient with clinical muscle pain, the intensity of muscle pain at different sites may be comparable between male and female patients. The visual analogue score may be too rudimentary to detect significant differences between the sexes or different sites. However, it is possible that a continuous measurement of the VAS may be able to detect more subtle changes. The comparable muscle pain may be described differently by males and females, and the treatment of muscle pain, for example trigger point therapy, female patients may not tolerate the same degree of pressure as male patients. In conclusion, we have found that experimentally induced muscle pain by hypertonic saline injection into the belly of the muscle is equivalent in intensity and quality in the tibialis anterior muscle and the lumbar erector spinae muscle, in the sample size we tested. In addition, gender does not influence muscle pain intensity but does influence sensitivity to pressure and the description of the pain.

530

L. Loram et al. / Manual Therapy 14 (2009) 526–530

Leg

Pressure tolerance (kPa)

1200 1000 800 600 400 200 0

control

isotonic Male

hypertonic Female

Pressure tolerance (kPa)

Back 1000 800 600 400 200 0

control

isotonic Male

hypertonic Female

Fig. 3. Pressure pain tolerance (KPa) in male (open bars, n ¼ 10) and female subjects (solid bars, n ¼ 10) following pressure applied to the leg and back muscle at least 10 min after isotonic saline (0.9%), hypertonic saline (5%) intramuscular injection or no injection, as measured on the contralateral side following an isotonic saline injection (control). All values are mean and standard error of the mean. The pressure pain tolerance of the females was significantly lower than that of the males irrespective of the stimulus before the pressure pain tolerance was tested.

References Babenko V, Graven-Nielsen T, Svensson P, Drewes AM, Jensen TS, Arendt-Nielsen L. Experimental human muscle pain induced by intramuscular injections of bradykinin, serotonin, and substance P. European Journal of Pain 1999;3(2): 93–102. Bajaj P, Bajaj P, Graven-Nielsen T, Arendt-Nielsen L. Osteoarthritis and its association with muscle hyperalgesia: an experimental controlled study. Pain 2001;93(2):107–14. Brodie EE, Niven CA. Remembering an everyday pain: the role of knowledge and experience in the recall of the quality of dysmenorrhoea. Pain 2000;84(1): 89–94. Brooks P. Issues with chronic musculoskeletal pain. Rheumatology 2005;44:831–3. Capra N, Ro J. Human and animal experimental models of acute and chronic muscle pain: intramuscular algesic injection. Pain 2004;110:3–7. Chesterton L, Barlas P, Foster N, Baxter G, Wright C. Gender differences in pressure pain threshold in healthy humans. Pain 2003;101:259–66. Cornwall J, John Harris A, Mercer SR. The lumbar multifidus muscle and patterns of pain. Manual Therapy 2006;11:40–5. Faul F, Erdfelder E, Lang A-G, Buchner A. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods 2007;39:175–91. Ge H-Y, Madeleine P, Cairns B, Arendt-Nielsen L. Hypoalgesia in the referred pain areas after bilateral injections of hypertonic saline into the trapezius muscles of

men and women: a potential experimental model of gender-specific differences. Clinical Journal of Pain 2006;22:37–44. Ge H-Y, Madeleine P, Arendt-Nielsen L. Sex differences in temporal characteristics of descending inhibitory control: an evaluation using repeated bilateral experimental induction of muscle pain. Pain 2004a;110(1–2):72–8. Ge H-Y, Wang K, Madeleine P, Svensson P, Sessle BJ, Arendt-Nielsen L. Simultaneous modulation of the exteroceptive suppression periods in the trapezius and temporalis muscles by experimental muscle pain. Clinical Neurophysiology 2004b;115(6):1399–408. Ge H-Y, Madeleine P, Wang K, Arendt-Nielsen L. Hypoalgesia to pressure pain in referred pain areas triggered by spatial summation of experimental muscle pain from unilateral or bilateral trapezius muscles. European Journal of Pain 2003;7(6):531–7. Gibson W, Arendt-Nielsen L, Graven-Nielsen T. Referred pain and hyperalgesia in human tendon and muscle belly tissue. Pain 2006;120(1–2):113–23. Graven-Nielsen T, Jansson Y, Segerdahl M, Kristensen JD, Mense S, Arendt-Nielsen L, et al. Experimental pain by ischaemic contractions compared with pain by intramuscular infusions of adenosine and hypertonic saline. European Journal of Pain 2003;7(1):93–102. Graven-Nielsen T, Mense S. The peripheral apparatus of muscle pain: evidence form animal and human studies. Clinical Journal of Pain 2001;17:2–10. Graven-Nielsen T, Arendt-Nielsen L, Svensson P, Jensen T. Stimulus-response functions in areas with experimentally-induced referred muscle pain–a psychophysical study. Brain Research 1997a;744:121–8. Graven-Nielsen T, Arendt-Nielsen L, Svensson P, Staehelin Jensen T. Quantification of local and referred muscle pain in humans after sequential i.m. injections of hypertonic saline. Pain 1997b;69(1–2):111–7. Hwang M, Kang YK, Kim DH. Referred pain pattern of the pronator quadratus muscle. Pain 2005;116(3):238–42. Kellgren J. Observations on referred pain arising from muscle. Clinical Science 1938;3:175–90. Koelbaek Johansen M, Graven-Nielsen T, Schou Olesen A, Arendt-Nielsen L. Generalised muscular hyperalgesia in chronic whiplash syndrome. Pain 1999;83(2):229–34. Leffler A-S, Kosek E, Hansson P. Injection of hypertonic saline into musculus infraspinatus resulted in referred pain and sensory disturbances in the ipsilateral upper arm. European Journal of Pain 2000;4(1):73–82. Maitland G. Vertebral manipulation. 6th ed. London: Butterworth-Heinemann; 2000. Masri R, Ro JY, Capra N. The effect of experimental muscle pain on the amplitude and velocity sensitivity of jaw closing muscle spindle afferents. Brain Research 2005;1050(1–2):138–47. Nie H, Arendt-Nielsen L, Kawczynski A, Madeleine P. Gender effects on trapezius surface EMG during delayed onset muscle soreness due to eccentric shoulder exercise. Journal of Electromyography and Kinesiology 2007;17(4):401–9. O’Neill S, Manniche C, Graven-Nielsen T, Arendt-Nielsen L. Generalized deep-tissue hyperalgesia in patients with chronic low-back pain. European Journal of Pain 2007;11(4):415–20. Potter L, McCarthy C, Oldham J. Algometer reliability in measuring pain pressure threshold over normal spinal muscles to allow quantification of anti-nociceptive treatment effects. International Journal of Osteopathic Medicine 2006;9(4):113–9. Schmidt-Hansen PT, Svensson P, Bendtsen L, Graven-Nielsen T, Bach FW. Increased muscle pain sensitivity in patients with tension-type headache. Pain 2007;129(1–2):113–21. Schneider S, Randoll D, Buchner M. Why do women have back pain more than men? A representative prevalence study in the Federal Republic of Germany. Clinical Journal of Pain 2006;22:738–47. Slater H, Arendt-Nielsen L, Wright A, Graven-Nielsen T. Sensory and motor effects of experimental muscle pain in patients with lateral epicondylalgia and controls with delayed onset muscle soreness. Pain 2005;114(1–2):118–30. Slater H, Arendt-Nielsen L, Wright A, Graven-Nielsen T. Experimental deep tissue pain in wrist extensors – a model of lateral epicondylalgia. European Journal of Pain 2003;7(3):277–88. Sjolund BH, Persson AL. Pressure pain threshold changes after repeated mechanonociceptive stimulation of the trapezius muscle: possible influence of previous pain experience. The Journal of Pain 2007;8(4):355–62. Svensson P, Cairns B, Wang K, Hu J, Graven-Nielsen T, Arendt-Nielsen L, et al. Glutamate-evoked pain and mechanical allodynia in the human-masseter muscle. Pain 2003;101:221–7. van Selms MKA, Wang K, Lobbezoo F, Svensson P, Arendt-Nielsen L, Naeije M. Effects of masticatory muscle fatigue without and with experimental pain on jaw-stretch reflexes in healthy men and women. Clinical Neurophysiology 2005;116(6):1415–23. Weerakkody NS, Percival P, Hickey MW, Morgan DL, Gregory JE, Canny BJ, et al. Effects of local pressure and vibration on muscle pain from eccentric exercise and hypertonic saline. Pain 2003;105(3):425–35. Zedka M, Prochazka A, Knight B, Gillard D, Gauthier M. Voluntary reflex control of human back muscles during induced pain. Journal of Physiology 1999;520: 591–604.

Manual Therapy 14 (2009) 531–538

Contents lists available at ScienceDirect

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

Original Article

The mechanisms of manual therapy in the treatment of musculoskeletal pain: A comprehensive model Joel E. Bialosky a, *, Mark D. Bishop a, Don D. Price b, Michael E. Robinson c, Steven Z. George a a

University of Florida, Department of Physical Therapy, Gainesville, FL 32610-0154, United States University of Florida, Department of Dentistry, Gainesville, FL 32610-0154, United States c University of Florida, Department of Clinical and Health Psychology, Gainesville, FL 32610-0154, United States b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 28 August 2007 Received in revised form 8 August 2008 Accepted 23 September 2008

Prior studies suggest manual therapy (MT) as effective in the treatment of musculoskeletal pain; however, the mechanisms through which MT exerts its effects are not established. In this paper we present a comprehensive model to direct future studies in MT. This model provides visualization of potential individual mechanisms of MT that the current literature suggests as pertinent and provides a framework for the consideration of the potential interaction between these individual mechanisms. Specifically, this model suggests that a mechanical force from MT initiates a cascade of neurophysiological responses from the peripheral and central nervous system which are then responsible for the clinical outcomes. This model provides clear direction so that future studies may provide appropriate methodology to account for multiple potential pertinent mechanisms. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Manual therapy Rehabilitation Pain

1. Introduction Available evidence suggests manual therapy (MT) as effective in the treatment of musculoskeletal disorders including low back pain (Licciardone et al., 2003; Childs et al., 2004), carpal tunnel syndrome (Rozmaryn et al., 1998; Akalin et al., 2002), knee osteoarthritis (Deyle et al., 2000), and hip osteoarthritis (MacDonald et al., 2006). Moreover, recent studies have provided even stronger evidence when participants are classified into sub-groups (Childs et al., 2004; Cleland et al., 2006). Despite the literature supporting its effectiveness, the mechanisms of MT are not established leading to a National Institutes of Health (NIH) call to specifically address this shortcoming (Khalsa et al., 2006). A better understanding of the mechanisms of MT is necessary for several reasons. First, recent evidence suggests successful outcomes in MT are dependent on identifying individuals likely to respond rather than identification of a specific lesion. Subsequently, clinical prediction rules based on clusters of signs and symptoms have been proposed to identify responders to MT (Flynn et al., 2002; Cleland et al., 2007). While helpful in directing clinical practice, an explanation is lacking as to why such patterns of signs and symptoms predicts successful clinical outcomes. Subsequently, the biological plausibility of current clinical prediction rules may not be established

* Correspondence to: Joel E Bialosky, University of Florida, Department of Physical Therapy, PO Box 100154, Gainesville, FL 32610-0154, United States. Tel.: þ1 352 870 9116; fax: þ1 352 273 6109. E-mail address: [email protected]fl.edu (J.E. Bialosky). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.09.001

leading to concern for chance associations rather than causation. Highlighting this concern, only one clinical prediction rule (Flynn et al., 2002) has, to our knowledge, been validated with a follow up study (Childs et al., 2004). An understanding of the mechanisms behind MT could assist in the identification of individuals likely to respond to MT by allowing a priori hypotheses as to pertinent predictive factors for future clinical prediction rules and a better understanding of the factors which are determined as predictive. A second benefit of the identification of MT mechanisms is the potential for increased acceptance of these techniques by healthcare providers. Despite the literature supporting the effectiveness of MT in specific musculoskeletal conditions, healthcare practitioners at times provide or refer for MT at a lower than expected rate (Jette and Delitto, 1997; Li and Bombardier, 2001; Bishop and Wing, 2003). The lack of an identifiable mechanism of action for MT may limit the acceptability of these techniques as they may be viewed as less scientific. Knowledge of mechanisms may promote more appropriate use of MT by healthcare providers. The intention of this manuscript is to present a comprehensive model to guide future studies of MT mechanisms. For our purposes, MT includes a variety of techniques used in clinical practice for the treatment of musculoskeletal pain which target the skeletal system, soft tissue, and nervous system (Table 1). 2. Need for a comprehensive model MT likely works through biomechanical and/or neurophysiological mechanisms. A limitation of the current literature is the

532

J.E. Bialosky et al. / Manual Therapy 14 (2009) 531–538

Table 1 Categorization of MT techniques. MT technique

Definition

Desired outcomes

 Passive movement of a joint beyond the normal range of motion  Passive movement of a joint within its normal range of motion

 Improved range of motion  Decrease muscle spasm  Decreased pain

 Swedish massage

 Stroking and kneading of the skin and underlying soft tissue

 Deep tissue massage

 Deep stroking and pressure across the muscles and soft tissue

 Trigger point massage

 Deep pressure to areas of local tenderness

 Shiatsu massage

 Varying, rhythmic pressure from the fingers

          

Joint biased

 Manipulation  Mobilization

Soft tissue biased

Improve circulation Decrease muscle spasm Relaxation Re-align soft tissue Break adhesions Increase range of motion Release muscle spasm Remove cellular exudates Improve Circulation Decrease muscle spasm Relaxation

Nerve biased

 Neural dynamics

 Passive, combined movement of the spine and extremities, within their normal range of motion, in ways to elongate or tension specific nerves.

 Improve range of motion  Decrease pain

Classification of MT techniques referenced in manuscript along with specific examples of each. Proposed model is general and accounts for all techniques regardless of their theorized anatomical emphasis. Adapted from NCCAM website (http:nccam.nih.gov/, 2007).

failure to acknowledge the potential for a combined effect of these mechanisms. For example, prior studies have noted individual biomechanical (Gal et al., 1997; Coppieters and Butler, 2007) and neurophysiological effects (Vicenzino et al., 1998; Suter et al., 1999; Dishman and Bulbulian, 2000; DeVocht et al., 2005) associated with MT; however the potential interaction of these effects is frequently overlooked. Combined effects may be important to consider as the biomechanical parameters of a given MT may produce unique or dose dependent neurophysiological responses. For example, associated hypoalgesic response (McLean et al., 2002) and EMG response (Colloca et al., 2006) have an observed dependence on the force and force/time profile of a given MT. Additionally, prior studies often focus on a single neurophysiological mechanism without consideration for competing explanations. For example neuromuscular changes such as decreased resting EMG activity (DeVocht et al., 2005) and decreased muscle inhibition (Suter et al., 1999; Suter and McMorland, 2002) have been associated with MT and theorized to occur due to stimulation of the mechanoreceptors or proprioceptors producing a spinal cord mediated effect (Suter et al., 2000; Suter and McMorland, 2002). While helpful in establishing the groundwork for the mechanistic study of MT, conclusions based on studies designed in this fashion may fail to consider other potentially pertinent mechanisms. Psychological factors have an observed association with muscular response in individuals with low back pain (Thomas et al., 2008) and MT has an observed effect on these psychological factors (Williams et al., 2007). Subsequently, outcomes reported in the prior studies (Suter et al., 1999; Suter and McMorland, 2002; DeVocht et al., 2005) could be explained by a descending supraspinal mediating effect due to changes in psychological factors such as fear. A consideration of the interaction between biomechanical and multiple potential neurophysiological effects necessitates a comprehensive model to synthesize the current literature and direct future research. 3. Proposed model We propose the following model which provides a compilation of the existing mechanistic literature of MT as a framework for

interpreting current and conducting future mechanistic research (Fig. 1). Briefly, this model suggests a mechanical stimulus initiates a number of potential neurophysiological effects which produce the clinical outcomes associated with MT in the treatment of musculoskeletal pain. 3.1. Mechanical stimulus Biomechanical effects are associated with MT as motion has been quantified with joint biased MT (Gal et al., 1997; Colloca et al., 2006) and nerve biased MT (Coppieters and Alshami, 2007; Coppieters and Butler, 2007); however, the direct implication on clinical outcomes is questionable. First, only transient biomechanical effects are supported by studies which quantify motion (Gal et al., 1997; Colloca et al., 2006; Coppieters and Alshami, 2007; Coppieters and Butler, 2007) but not a lasting positional change (Tullberg et al., 1998; Hsieh et al., 2002). Second, biomechanical assessment is not reliable. Palpation for position and movement faults has demonstrated poor reliability (Troyanovich et al., 1998; Seffinger et al., 2004) suggesting an inability to accurately determine a specific area requiring MT. Third, MT techniques lack precision as nerve biased techniques are not specific to a single nerve (Kleinrensink et al., 2000) and joint biased technique forces are dissipated over a large area (Herzog et al., 2001; Ross et al., 2004). Additionally, different kinetic parameters are observed between clinicians in the performance of the same technique (Hessell et al., 1990; Ngan et al., 2005) and the choice of technique does not seem to matter as much as identifying an individual likely to respond (Kent et al., 2005; Cleland et al., 2006). Finally, studies have reported improvements in signs and symptoms away from the site of application such as treating cervical pain with MT directed to the thoracic spine (Cleland et al., 2005; Cleland et al., 2007) and lateral epicondylitis with MT directed to the cervical spine (Vicenzino et al., 1996). Collectively, the literature suggests a biomechanical effect of MT; however, lasting structural changes have not been identified, clinicians are unable to reliably identify areas requiring MT, the forces associated with MT are not specific to a given location and vary between clinicians, choice of technique does not seem to affect outcomes, and sign and symptom responses occur in areas separate

J.E. Bialosky et al. / Manual Therapy 14 (2009) 531–538

533

Non Specific Responses • Placebo/Expectation • Psychological measures Fear Catastrophizing Kinesiophobia

Imaging

PAIN

Pain-Related Brain Circuitry

Pain Modulatory Circuitry ACC Amygdala PAG RVM

Endocrine Response • B-endorphins • Opioid response

Rating Autonomic Response • Skin temperature • Skin conduction • Cortisol levels • Heart rate

Imaging

Mechanical Stimulus

Tissue

Decrease Spasm Increase range of motion

Peripheral Nervous System

Spinal Cord

Neuromuscular Responses • Motoneuron Pool • Afferent Discharge • Muscle activity Hypoalgesia • Temporal summation • Selective blocking of neuro-transmitters

Inflammatory mediators

Imaging ACC = anterior cingular cortex; PAG = periaqueductal gray; RVM = rostral ventromedial medulla Fig. 1. Comprehensive model of the mechanisms of MT. Figure key: The model suggests a transient, mechanical stimulus to the tissue produces a chain of neurophysiological effects. ¼ an association between a construct and its Solid arrows denote a direct mediating effect. Broken arrows denote an associative relationship which may include: measure . Bold boxes indicate the measurement of a construct.

from the region of application. The effectiveness of MT despite the inconsistencies associated with a purported biomechanical mechanism suggests that additional mechanisms may be pertinent. Subsequently, we suggest, that as illustrated by the model, a mechanical force is necessary to initiate a chain of neurophysiological responses which produce the outcomes associated with MT. 3.2. Neurophysiological mechanism The proposed model accounts for the complex interactions of both the peripheral and central nervous system which comprise the pain experience. Current mechanistic studies of MT in humans are frequently unable to directly observe the central or peripheral nervous system. Subsequently, in the absence of direct observation, conclusions are drawn from associated neurophysiological responses which indirectly implicate specific mechanisms. Studies have measured associated responses of hypoalgesia and sympathetic activity following MT to suggest a mechanism of action mediated by the periaquaductal gray (Wright,1995) and lessening of temporal summation following MT to suggest a mechanism mediated by the dorsal horn of the spinal cord (George et al., 2006) The model makes use of directly measurable associated responses to imply specific neurophysiological mechanisms when direct observations are not possible. The model categorizes neurophysiological mechanisms as those likely originating from a peripheral mechanism, spinal cord mechanisms, and/or supraspinal mechanisms. 3.3. Peripheral mechanism Musculoskeletal injuries induce an inflammatory response in the periphery which initiates the healing process and influences pain processing. Inflammatory mediators and peripheral

nociceptors interact in response to injury and MT may directly affect this process. For example, Teodorczyk-Injeyan et al. (2006) observed a significant reduction of blood and serum level cytokines in individuals receiving joint biased MT which was not observed in those receiving sham MT or in a control group. Additionally, changes of blood levels of b-endorphin, anandamide, N-palmitoylethanolamide, serotonin (Degenhardt et al., 2007) and endogenous cannabinoids (McPartland et al., 2005) have been observed following MT. Finally, soft tissue biased MT has been shown to alter acute inflammation in response to exercise (Smith et al., 1994) and substance P levels in individuals with fibromyalgia (Field et al., 2002). Collectively, these studies suggest a potential mechanism of action of MT on musculoskeletal pain mediated by the peripheral nervous system for which mechanistic studies may wish to account. 3.4. Spinal mechanisms MT may exert an effect on the spinal cord. For example, MT has been suggested to act as a counter irritant to modulate pain (Boal and Gillette, 2004) and joint biased MT is speculated to ‘‘bombard the central nervous system with sensory input from the muscle proprioceptors (Pickar and Wheeler, 2001).’’ Subsequently, a spinal cord mediated mechanism of MT must be considered and is accounted for in the model. Direct evidence for such an effect comes from a study (Malisza et al., 2003b) in which joint biased MT was applied to the lower extremity of rats following capsaicin injection. A spinal cord response was quantified by functional MRI during light touch to the hind paw. A trend was noted towards decreased activation of the dorsal horn of the spinal cord following the MT. The model uses associated neuromuscular responses following MT to provide indirect evidence for a spinal cord

534

J.E. Bialosky et al. / Manual Therapy 14 (2009) 531–538

Non Specific Responses • Placebo/Expectation • Psychological measures Fear Catastrophizing Kinesiophobia

Imaging

PAIN

Pain-Related Brain Circuitry

Pain Modulatory Circuitry ACC Amygdala PAG RVM

Endocrine Response • B-endorphins • Opioid response

Rating Autonomic Response • Skin temperature • Skin conduction • Cortisol levels • Heart rate

Imaging

Mechanical Stimulus

Tissue

Decrease Spasm Increase range of motion

Peripheral Nervous System

Spinal Cord

Inflammatory mediators

Neuromuscular Responses • Motoneuron Pool • Afferent Discharge • Muscle activity Hypoalgesia • Temporal summation • Selective blocking of neuro-transmitters Imaging

Fig. 2. Pathway for a spinal cord mediated effect of MT from George et al. (2006). Figure key: Proposed model pathway of study by George et al. (2006) suggesting a spinal cord mediating effect of MT. Bold arrows indicate suggested mechanism. Note mediating effect is suggested to be through the spinal cord due to measurement of the associated relationship of temporal summation. Also note, the design of this study neglects to consider potential supraspinal mediated effects.

mediated mechanism. For example, MT is associated with hypoalgesia (Vicenzino et al., 2001; Mohammadian et al., 2004; George et al., 2006), afferent discharge (Colloca et al., 2000; Colloca et al., 2003), motoneuron pool activity (Bulbulian et al., 2002; Dishman and Burke, 2003), and changes in muscle activity (Herzog et al., 1999; Symons et al., 2000) all of which may indirectly implicate a spinal cord mediated effect. 3.5. Supraspinal mechanisms Finally, the pain literature suggests the influence of specific supraspinal structures in response to pain. Structures such as the anterior cingular cortex (ACC), amygdala, periaqueductal gray (PAG), and rostral ventromedial medulla (RVM) are considered instrumental in the pain experience (Hsieh et al., 1995; Vogt et al., 1996; Derbyshire et al., 1997; Iadarola et al., 1998; Peyron et al., 2000; Moulton et al., 2005; Guo et al., 2006; Bee and Dickenson, 2007; Oshiro et al., 2007; Staud et al., 2007). Subsequently, the model considers potential supraspinal mechanisms of MT. Direct support for a supraspinal mechanism of action of MT comes from Malisza et al. (2003a) who applied joint biased MT to the lower extremity of rats following capsaicin injection. Functional MRI of the supraspinal region quantified the response of the hind paw to light touch following the injection. A trend was noted towards decreased activation of the supraspinal regions responsible for central pain processing. The model accounts for direct measures of supraspinal activity along with associated responses such as autonomic responses (Vicenzino et al., 1998; Sterling et al., 2001; Delaney et al., 2002; Moulson and Watson, 2006; Zhang et al., 2006) and opiod responses (Vernon et al., 1986; Kaada and Torsteinbo, 1989) to indirectly imply a supraspinal mechanism. Additionally, variables such as placebo, expectation, and psychosocial factors may be pertinent in the mechanisms of MT (Ernst, 2000; Kaptchuk, 2002). For

example expectation for the effectiveness of MT is associated with functional outcomes (Kalauokalani et al., 2001) and a recent systematic review of the literature has noted that joint biased MT is associated with improved psychological outcomes (Williams et al., 2007). For this paper we categorize such factors as neurophysiological effects related to supraspinal descending inhibition due to associated changes in the opioid system (Sauro and Greenberg, 2005), dopamine production (Fuente-Fernandez et al., 2006), and central nervous system (Petrovic et al., 2002; Wager et al., 2004; Matre et al., 2006) which have been observed in studies unrelated to MT. 4. Implementation of comprehensive model The comprehensive model delineates potential mechanisms associated with pain relief from MT allowing researchers to identify domains of interest their studies are designed to evaluate and potential mechanisms not adequately considered. The model is intended to highlight differing possibilities when conclusions are drawn which may be further explored in subsequent studies. For example, studies have reported hypoalgesia following MT (Mohammadian et al., 2004; George et al., 2006). George et al. (2006) suggested a spinal cord mediated mechanism due to associated hypoalgesia of temporal summation. The model indicates that while monitoring a spinal cord mediating effect (temporal summation), the potential for a peripheral or supraspinal mediating effects was not considered (Fig. 2). A recent study attempted to replicate these prior findings while accounting for potential supraspinal influence (Bialosky et al., 2008). Specifically, a spinal cord mediated effect was measured through an associated response of temporal summation. Additionally, a potential supraspinal mechanism (expectation) was manipulated by randomly assigning participants to receive an instructional set stating MT was expected to either

J.E. Bialosky et al. / Manual Therapy 14 (2009) 531–538

increase, decrease, or have no effect on their pain perception. The model pathway of this study is visualized in Fig. 3. In addition to guiding research, the model also allows clinicians to visualize the potential multiple mechanisms likely involved in the clinical effects of MT. The clinical use of MT is frequently dependent upon a purported biomechanical mechanism in evaluation and treatment. For instance, a clinical examination may focus on locating a mal-aligned joint or a hypomobile joint or soft tissue. An MT technique may then be used as treatment to impart a specific movement to the observed dysfunction. Clinical outcomes are then attributed to alleviation of the biomechanical fault. Such practice is common and has lead to many continuing education dollars and valuable clinic time spent in search of biomechanical dysfunction of questionable validity (Seffinger et al., 2004) and treatments of questionable specificity (Ross et al., 2004). The model provides visualization of what the current literature suggests as mechanisms pertinent to MT and while acknowledging a biomechanical effect allows clinicians to consider other potential mechanisms in the MT evaluation and treatment of individuals with musculoskeletal pain.

5. Limitations of proposed model The model is intended to be applicable to all forms of MT. While the biomechanical application of joint biased, soft tissue biased and nerve biased MT are different, the related neurophysiological responses are similar and adequately encompassed within the model given the current state of knowledge. The proposed model provides a platform to empirically test hypotheses related to different biomechanical and neurophysiological effects specific to types of MT, an area that is currently lacking in the literature. The

535

proposed comprehensive model is intended to explain the mechanisms of MT on musculoskeletal pain. MT has a postulated role in the treatment of disorders of other body systems such as asthma (Balon and Mior, 2004) and high blood pressure (Plaugher and Bachman, 1993); however, those effects are beyond the scope of the current model. Finally, this model is strictly intended to guide research questions regarding the mechanisms of MT. A body of literature already exists suggesting the effectiveness of MT. The proposed model is intended to compliment and provide underlying explanations to the existing body of literature suggesting the effectiveness of MT.

6. Future directions A limitation in the current literature is the failure to account for the non-specific mechanisms associated with MT in the treatment of musculoskeletal pain. A number of neurophysiological responses associated with MT are also associated with non-specific effects such as placebo (Fig. 4). Current study designs have not adequately accounted for non-specific effects, and subsequently, their role in the clinical outcomes associated with MT is unknown. Future mechanistic studies in MT should consider determining the influence of non-specific effects. The model presents a guide to design future mechanistic studies so that all relevant possibilities are included. The model is based primarily on associated responses as the current body of mechanistic literature is lacking in studies which directly observe regions of interest. As technology improves, the means to directly observe specific regions is becoming possible. More recent studies in the acupuncture literature have reported direct observation of the spinal cord (Wang et al., 2006; Chen et al., 2007) and supraspinal centers (Dougherty et al., 2008; Fang et al., 2008) in

Non Specific Responses • Placebo/Expectation • Psychological measures Fear Catastrophizing Kinesiophobia

Imaging

PAIN

Pain-Related Brain Circuitry

Pain Modulatory Circuitry ACC Amygdala PAG RVM

Endocrine Response • B-endorphins • Opioid response

Rating Autonomic Response • Skin temperature • Skin conduction • Cortisol levels • Heart rate

Imaging

Mechanical Stimulus

Tissue

Decrease Spasm Increase range of motion

Peripheral Nervous System

Inflammatory mediators

Spinal Cord

Neuromuscular Responses • Motoneuron Pool • Afferent Discharge • Muscle activity Hypoalgesia • Temporal summation • Selective blocking of neuro- transmitters Imaging

Fig. 3. Pathway considering both a spinal cord and supraspinal mediated effect from Bialosky et al. (2008). Figure key: Proposed model pathway of study by Bialosky et al. (2008) which considers both a spinal cord and supraspinal mediating effect of MT. Bold arrows indicate suggested mechanism. Note mediating effect is suggested to be through both the spinal cord due to measurement of the associated relationship of temporal summation and through a supraspinal mechanism due to measurement of the associated relationship of expectation.

536

J.E. Bialosky et al. / Manual Therapy 14 (2009) 531–538

Non Specific Responses • Placebo/Expectation • Psychological measures Fear Catastrophizing Kinesiophobia

Imaging

Pain-Related Brain Circuitry

PAIN (Price et al., 1999) (Benedetti et al., 2003) (Vase et al., 2002)

Tissue

Autonomic Response • Skin temperature • Skin conduction • Cortisol levels • Heart rate

(Lanotte et al., 2005) (Pollo et al., 2003) (Johansen et al., 2003)

(Price et al., 2007) (Kong et al., 2006) (Bingel et al., 2006)

Peripheral Nervous System

Spinal Cord

(Price et al., 2002)

Decrease Spasm Increase range of motion

Endocrine Response • B-endorphins • Opioid response

Rating Imaging

Mechanical Stimulus

Pain Modulatory Circuitry ACC Amygdala PAG RVN

(Amanzio et al., 2001) (Benedetti et al., 2006) (Zubieta et al., 2005)

Neuromuscular Responses • Motor Neuron Pool • Afferent Discharge • Muscle activity Hypoalgesia (Goffaux et al., 2007)

Inflammatory mediators (Matre et al., 2006)

Imaging

(Goebel et al., 2002)

Fig. 4. Comprehensive model for the mechanisms of MT illustrating similar neurophysiological activity in response to non-specific effects such as placebo and expectation. A limitation of the current mechanistic literature in MT is the failure to adequately account for non-specific effects such as placebo and expectation. Italicized references are examples of studies from the placebo and expectation literature which have reported similar neurophysiological effects as have been associated with MT. These similarities emphasize the potential for nonspecific effects to play a significant role in the mechanisms behind MT and the need to specifically address these factors in future studies.

response to treatment. Similar studies are possible in MT and will allow direct observation of the nervous system response to MT with a subsequent improved understanding of where the techniques exert their effect. Interdisciplinary collaboration has been recommended in the study of the mechanisms of MT (Khalsa et al., 2006). The comprehensive model provides a framework for such efforts to study both specific sections of the model and their interaction. For example, a team of researchers could work together including a manual therapist to provide treatment, a biomechanist to monitor the biomechanical parameters of the studied MT, an endocrinologist to monitor peripheral inflammatory mediators, a neurophysiologist to monitor potential spinal cord and supraspinal mechanisms, and a psychologist to monitor the influence of non-specific effects such as expectation, fear, and catastrophizing. 7. Conclusion The mechanisms behind the clinical effectiveness of MT are not established. Limitations of prior mechanistic studies are the study of individual mechanisms without regard for others and a failure to adequately account for non-specific effects. We have proposed a comprehensive model to consolidate the current research and guide future research into the mechanisms of MT. Acknowledgements The project was supported by Grant Number R-21 AT002796-01 from the National Institutes of Health – National Center for Complimentary and Alternative Medicine (SZG, MDB, MER, DDP). This manuscript was written while JEB received support from the National Institutes of Health T-32 Neural Plasticity Research Training Fellowship (T32HD043730).

References Akalin E, El O, Peker O, Senocak O, Tamci S, Gulbahar S, et al. Treatment of carpal tunnel syndrome with nerve and tendon gliding exercises. Am J Phys Med Rehabil 2002;81:108–13. Amanzio M, Pollo A, Maggi G, Benedetti F. Response variability to analgesics: a role for non-specific activation of endogenous opioids. Pain 2001;90:205–15. Balon JW, Mior SA. Chiropractic care in asthma and allergy. Ann Allergy Asthma Immunol 2004;93:S55–60. Bee LA, Dickenson AH. Rostral ventromedial medulla control of spinal sensory processing in normal and pathophysiological states. Neuroscience 2007;147: 786–93. Benedetti F, Amanzio M, Vighetti S, Asteggiano G. The biochemical and neuroendocrine bases of the hyperalgesic nocebo effect. J Neurosci 2006;26:12014–22. Benedetti F, Pollo A, Lopiano L, Lanotte M, Vighetti S, Rainero I. Conscious expectation and unconscious conditioning in analgesic, motor, and hormonal placebo/nocebo responses. J Neurosci 2003;23:4315–23. Bialosky JE, Bishop MD, Robinson ME, Barabas JA, George SZ. The influence of expectation on spinal manipulation induced hypoalgesia: an experimental study in normal subjects. BMC Musculoskelet Disord 2008;9:19. Bingel U, Lorenz J, Schoell E, Weiller C, Buchel C. Mechanisms of placebo analgesia: rACC recruitment of a subcortical antinociceptive network. Pain 2006;120:8–15. Bishop PB, Wing PC. Compliance with clinical practice guidelines in family physicians managing worker’s compensation board patients with acute lower back pain. Spine J 2003;3:442–50. Boal RW, Gillette RG. Central neuronal plasticity, low back pain and spinal manipulative therapy. J Manipulative Physiol Ther 2004;27:314–26. Bulbulian R, Burke J, Dishman JD. Spinal reflex excitability changes after lumbar spine passive flexion mobilization. J Manipulative Physiol Ther 2002;25: 526–32. Chen YX, Kong KM, Wang WD, Xie CH, Wu RH. Functional MR imaging of the spinal cord in cervical spinal cord injury patients by acupuncture at LI 4 (Hegu) and LI 11(Quchi). Conf Proc IEEE Eng Med Biol Soc 2007;2007:3388–91. 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. Ann Intern Med 2004;141:920–8. 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 2007;87:9–23. 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;10:127–35.

J.E. Bialosky et al. / Manual Therapy 14 (2009) 531–538 Cleland JA, Fritz JM, Whitman JM, Childs JD, Palmer JA. The use of a lumbar spine manipulation technique by physical therapists in patients who satisfy a clinical prediction rule: a case series. J Orthop Sports Phys Ther 2006;36:209–14. Colloca CJ, Keller TS, Gunzburg R. Neuromechanical characterization of in vivo lumbar spinal manipulation. Part II. Neurophysiological response. J Manipulative Physiol Ther 2003;26:579–91. Colloca CJ, Keller TS, Gunzburg R, Vandeputte K, Fuhr AW. Neurophysiologic response to intraoperative lumbosacral spinal manipulation. J Manipulative Physiol Ther 2000;23:447–57. Colloca CJ, Keller TS, Harrison DE, Moore RJ, Gunzburg R, Harrison DD. Spinal manipulation force and duration affect vertebral movement and neuromuscular responses. Clin Biomech (Bristol, Avon) 2006;21:254–62. Coppieters MW, Alshami AM. Longitudinal excursion and strain in the median nerve during novel nerve gliding exercises for carpal tunnel syndrome. J Orthop Res 2007;25:972–80. Coppieters MW, Butler DS. Do ‘sliders’ slide and ‘tensioners’ tension? An analysis of neurodynamic techniques and considerations regarding their application. Man Ther 2007. Degenhardt BF, Darmani NA, Johnson JC, Towns LC, Rhodes DC, Trinh C, et al. Role of osteopathic manipulative treatment in altering pain biomarkers: a pilot study. J Am Osteopath Assoc 2007;107:387–400. Delaney JP, Leong KS, Watkins A, Brodie D. The short-term effects of myofascial trigger point massage therapy on cardiac autonomic tone in healthy subjects. J Adv Nurs 2002;37:364–71. Derbyshire SW, Jones AK, Gyulai F, Clark S, Townsend D, Firestone LL. Pain processing during three levels of noxious stimulation produces differential patterns of central activity. Pain 1997;73:431–45. DeVocht JW, Pickar JG, Wilder DG. Spinal manipulation alters electromyographic activity of paraspinal muscles: a descriptive study. J Manipulative Physiol Ther 2005;28:465–71. Deyle GD, Henderson NE, Matekel RL, Ryder MG, Garber MB, Allison SC. Effectiveness of manual physical therapy and exercise in osteoarthritis of the knee. A randomized, controlled trial. Ann Intern Med 2000;132:173–81. Dishman JD, Bulbulian R. Spinal reflex attenuation associated with spinal manipulation. Spine 2000;25:2519–24. Dishman JD, Burke J. Spinal reflex excitability changes after cervical and lumbar spinal manipulation: a comparative study. Spine J 2003;3:204–12. Dougherty DD, Kong J, Webb M, Bonab AA, Fischman AJ, Gollub RL. A combined [11C]diprenorphine PET study and fMRI study of acupuncture analgesia. Behav Brain Res 2008. Ernst E. Does spinal manipulation have specific treatment effects? Fam Pract 2000;17:554–6. Fang J, Jin Z, Wang Y, Li K, Kong J, Nixon EE, et al. The salient characteristics of the central effects of acupuncture needling: limbic–paralimbic–neocortical network modulation. Hum Brain Mapp 2008. Field T, Diego M, Cullen C, Hernandez-Reif M, Sunshine W, Douglas S. Fibromyalgia pain and substance P decrease and sleep improves after massage therapy. J Clin Rheumatol 2002;8:72–6. 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;27:2835–43. Fuente-Fernandez R, Lidstone S, Stoessl AJ. Placebo effect and dopamine release. J Neural Transm Suppl 2006:415–8. Gal J, Herzog W, Kawchuk G, Conway PJ, Zhang YT. Movements of vertebrae during manipulative thrusts to unembalmed human cadavers. J Manipulative Physiol Ther 1997;20:30–40. George SZ, Bishop MD, Bialosky JE, Zeppieri Jr G, Robinson ME. Immediate effects of spinal manipulation on thermal pain sensitivity: an experimental study. BMC Musculoskelet Disord 2006;7:68. Goebel MU, Trebst AE, Steiner J, Xie YF, Exton MS, Frede S, et al. Behavioral conditioning of immunosuppression is possible in humans. FASEB J 2002;16:1869–73. Goffaux P, Redmond WJ, Rainville P, Marchand S. Descending analgesia – when the spine echoes what the brain expects. Pain 2007. Guo W, Robbins MT, Wei F, Zou S, Dubner R, Ren K. Supraspinal brain-derived neurotrophic factor signaling: a novel mechanism for descending pain facilitation. J Neurosci 2006;26:126–37. Herzog W, Kats M, Symons B. The effective forces transmitted by high-speed, low-amplitude thoracic manipulation. Spine 2001;26:2105–10. Herzog W, Scheele D, Conway PJ. Electromyographic responses of back and limb muscles associated with spinal manipulative therapy. Spine 1999;24:146–52. Hessell BW, Herzog W, Conway PJ, McEwen MC. Experimental measurement of the force exerted during spinal manipulation using the Thompson technique. J Manipulative Physiol Ther 1990;13:448–53. Hsieh CY, Vicenzino B, Yang CH, Hu MH, Yang C. Mulligan’s mobilization with movement for the thumb: a single case report using magnetic resonance imaging to evaluate the positional fault hypothesis. Man Ther 2002;7:44–9. Hsieh JC, Belfrage M, Stone-Elander S, Hansson P, Ingvar M. Central representation of chronic ongoing neuropathic pain studied by positron emission tomography. Pain 1995;63:225–36. Iadarola MJ, Berman KF, Zeffiro TA, Byas-Smith MG, Gracely RH, Max MB, et al. Neural activation during acute capsaicin-evoked pain and allodynia assessed with PET. Brain 1998;121:931–47 [part 5]. Jette AM, Delitto A. Physical therapy treatment choices for musculoskeletal impairments. Phys Ther 1997;77:145–54. Johansen O, Brox J, Flaten MA. Placebo and nocebo responses, cortisol, and circulating beta-endorphin. Psychosom Med 2003;65:786–90.

537

Kaada B, Torsteinbo O. Increase of plasma beta-endorphins in connective tissue massage. Gen Pharmacol 1989;20:487–9. Kalauokalani D, Cherkin DC, Sherman KJ, Koepsell TD, Deyo RA. Lessons from a trial of acupuncture and massage for low back pain: patient expectations and treatment effects. Spine 2001;26:1418–24. Kaptchuk TJ. The placebo effect in alternative medicine: can the performance of a healing ritual have clinical significance? Ann Intern Med 2002;136:817–25. Kent P, Marks D, Pearson W, Keating J. Does clinician treatment choice improve the outcomes of manual therapy for nonspecific low back pain? A metaanalysis. J Manipulative Physiol Ther 2005;28:312–22. Khalsa PS, Eberhart A, Cotler A, Nahin R. The 2005 conference on the biology of manual therapies. J Manipulative Physiol Ther 2006;29:341–6. Kleinrensink GJ, Stoeckart R, Mulder PG, Hoek G, Broek T, Vleeming A, et al. Upper limb tension tests as tools in the diagnosis of nerve and plexus lesions. Anatomical and biomechanical aspects. Clin Biomech (Bristol, Avon) 2000; 15:9–14. Kong J, Gollub RL, Rosman IS, Webb JM, Vangel MG, Kirsch I, et al. Brain activity associated with expectancy-enhanced placebo analgesia as measured by functional magnetic resonance imaging. J Neurosci 2006;26:381–8. Lanotte M, Lopiano L, Torre E, Bergamasco B, Colloca L, Benedetti F. Expectation enhances autonomic responses to stimulation of the human subthalamic limbic region. Brain Behav Immun 2005;19:500–9. Li LC, Bombardier C. Physical therapy management of low back pain: an exploratory survey of therapist approaches. Phys Ther 2001;81:1018–28. Licciardone JC, Stoll ST, Fulda KG, Russo DP, Siu J, Winn W, et al. Osteopathic manipulative treatment for chronic low back pain: a randomized controlled trial. Spine 2003;28:1355–62. MacDonald CW, Whitman JM, Cleland JA, Smith M, Hoeksma HL. Clinical outcomes following manual physical therapy and exercise for hip osteoarthritis: a case series. J Orthop Sports Phys Ther 2006;36:588–99. Malisza KL, Gregorash L, Turner A, Foniok T, Stroman PW, Allman AA, et al. Functional MRI involving painful stimulation of the ankle and the effect of physiotherapy joint mobilization. Magn Reson Imaging 2003a;21:489–96. Malisza KL, Stroman PW, Turner A, Gregorash L, Foniok T, Wright A. Functional MRI of the rat lumbar spinal cord involving painful stimulation and the effect of peripheral joint mobilization. J Magn Reson Imaging 2003b;18:152–9. Matre D, Casey KL, Knardahl S. Placebo-induced changes in spinal cord pain processing. J Neurosci 2006;26:559–63. McLean S, Naish R, Reed L, Urry S, Vicenzino B. A pilot study of the manual force levels required to produce manipulation induced hypoalgesia. Clin Biomech (Bristol, Avon) 2002;17:304–8. McPartland JM, Giuffrida A, King J, Skinner E, Scotter J, Musty RE. Cannabimimetic effects of osteopathic manipulative treatment. J Am Osteopath Assoc 2005;105:283–91. Mohammadian P, Gonsalves A, Tsai C, Hummel T, Carpenter T. Areas of capsaicin-induced secondary hyperalgesia and allodynia are reduced by a single chiropractic adjustment: a preliminary study. J Manipulative Physiol Ther 2004; 27:381–7. Moulson A, Watson T. A preliminary investigation into the relationship between cervical snags and sympathetic nervous system activity in the upper limbs of an asymptomatic population. Man Ther 2006;11:214–24. Moulton EA, Keaser ML, Gullapalli RP, Greenspan JD. Regional intensive and temporal patterns of functional MRI activation distinguishing noxious and innocuous contact heat. J Neurophysiol 2005;93:2183–93. Ngan JM, Chow DH, Holmes AD. The kinematics and intra- and inter-therapist consistencies of lower cervical rotational manipulation. Med Eng Phys 2005; 27:395–401. Oshiro Y, Quevedo AS, McHaffie JG, Kraft RA, Coghill RC. Brain mechanisms supporting spatial discrimination of pain. J Neurosci 2007;27:3388–94. Petrovic P, Kalso E, Petersson KM, Ingvar M. Placebo and opioid analgesia – imaging a shared neuronal network. Science 2002;295:1737–40. Peyron R, Laurent B, Garcia-Larrea L. Functional imaging of brain responses to pain. A review and meta-analysis (2000). Neurophysiol Clin 2000;30:263–88. Pickar JG, Wheeler JD. Response of muscle proprioceptors to spinal manipulativelike loads in the anesthetized cat. J Manipulative Physiol Ther 2001;24:2–11. Plaugher G, Bachman TR. Chiropractic management of a hypertensive patient. J Manipulative Physiol Ther 1993;16:544–9. Pollo A, Vighetti S, Rainero I, Benedetti F. Placebo analgesia and the heart. Pain 2003;102:125–33. Price DD, Craggs J, Verne GN, Perlstein WM, Robinson ME. Placebo analgesia is accompanied by large reductions in pain-related brain activity in irritable bowel syndrome patients. Pain 2007;127:63–72. Price DD, Milling LS, Kirsch I, Duff A, Montgomery GH, Nicholls SS. An analysis of factors that contribute to the magnitude of placebo analgesia in an experimental paradigm. Pain 1999;83:147–56. Price DD, Staud R, Robinson ME, Mauderli AP, Cannon R, Vierck CJ. Enhanced temporal summation of second pain and its central modulation in fibromyalgia patients. Pain 2002;99:49–59. Ross JK, Bereznick DE, McGill SM. Determining cavitation location during lumbar and thoracic spinal manipulation: is spinal manipulation accurate and specific? Spine 2004;29:1452–7. Rozmaryn LM, Dovelle S, Rothman ER, Gorman K, Olvey KM, Bartko JJ. Nerve and tendon gliding exercises and the conservative management of carpal tunnel syndrome. J Hand Ther 1998;11:171–9. Sauro MD, Greenberg RP. Endogenous opiates and the placebo effect: a metaanalytic review. J Psychosom Res 2005;58:115–20.

538

J.E. Bialosky et al. / Manual Therapy 14 (2009) 531–538

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:E413–25. Smith LL, Keating MN, Holbert D, Spratt DJ, McCammon MR, Smith SS, et al. The effects of athletic massage on delayed onset muscle soreness, creatine kinase, and neutrophil count: a preliminary report. J Orthop Sports Phys Ther 1994;19:93–9. Staud R, Craggs JG, Robinson ME, Perlstein WM, Price DD. Brain activity related to temporal summation of C-fiber evoked pain. Pain 2007;129:130–42. Sterling M, Jull G, Wright A. Cervical mobilisation: concurrent effects on pain, sympathetic nervous system activity and motor activity. Man Ther 2001;6:72–81. Suter E, McMorland G. Decrease in elbow flexor inhibition after cervical spine manipulation in patients with chronic neck pain. Clin Biomech (Bristol, Avon) 2002;17:541–4. Suter E, McMorland G, Herzog W, Bray R. Decrease in quadriceps inhibition after sacroiliac joint manipulation in patients with anterior knee pain. J Manipulative Physiol Ther 1999;22:149–53. Suter E, McMorland G, Herzog W, Bray R. Conservative lower back treatment reduces inhibition in knee-extensor muscles: a randomized controlled trial. J Manipulative Physiol Ther 2000;23:76–80. Symons BP, Herzog W, Leonard T, Nguyen H. Reflex responses associated with activator treatment. J Manipulative Physiol Ther 2000;23:155–9. Teodorczyk-Injeyan JA, Injeyan HS, Ruegg R. Spinal manipulative therapy reduces inflammatory cytokines but not substance P production in normal subjects. J Manipulative Physiol Ther 2006;29:14–21. Thomas JS, France CR, Sha D, Wiele NV. The influence of pain-related fear on peak muscle activity and force generation during maximal isometric trunk exertions. Spine 2008;33:E342–8. Troyanovich SJ, Harrison DD, Harrison DE. Motion palpation: it’s time to accept the evidence. J Manipulative Physiol Ther 1998;21:568–71. Tullberg T, Blomberg S, Branth B, Johnsson R. Manipulation does not alter the position of the sacroiliac joint. A roentgen stereophotogrammetric analysis. Spine 1998;23:1124–8. Vase L, Riley III JL, Price DD. A comparison of placebo effects in clinical analgesic trials versus studies of placebo analgesia. Pain 2002;99:443–52.

Vernon HT, Dhami MS, Howley TP, Annett R. Spinal manipulation and betaendorphin: a controlled study of the effect of a spinal manipulation on plasma beta-endorphin levels in normal males. J Manipulative Physiol Ther 1986;9:115–23. Vicenzino B, Collins D, Benson H, Wright A. An investigation of the interrelationship between manipulative therapy-induced hypoalgesia and sympathoexcitation. J Manipulative Physiol Ther 1998;21:448–53. Vicenzino B, Collins D, Wright A. The initial effects of a cervical spine manipulative physiotherapy treatment on the pain and dysfunction of lateral epicondylalgia. Pain 1996;68:69–74. Vicenzino B, Paungmali A, Buratowski S, Wright A. Specific manipulative therapy treatment for chronic lateral epicondylalgia produces uniquely characteristic hypoalgesia. Man Ther 2001;6:205–12. Vogt BA, Derbyshire S, Jones AK. Pain processing in four regions of human cingulate cortex localized with co-registered PET and MR imaging. Eur J Neurosci 1996;8:1461–73. Wager TD, Rilling JK, Smith EE, Sokolik A, Casey KL, Davidson RJ, et al. Placeboinduced changes in FMRI in the anticipation and experience of pain. Science 2004;303:1162–7. Wang WD, Kong KM, Xiao YY, Wang XJ, Liang B, Qi WL, et al. Functional MR imaging of the cervical spinal cord by use of electrical stimulation at LI4 (Hegu). Conf Proc IEEE Eng Med Biol Soc 2006;1:1029–31. Williams NH, Hendry M, Lewis R, Russell I, Westmoreland A, Wilkinson C. Psychological response in spinal manipulation (PRISM): a systematic review of psychological outcomes in randomised controlled trials. Complement Ther Med 2007;15:271–83. Wright A. Hypoalgesia post-manipulative therapy: a review of a potential neurophysiological mechanism. Man Ther 1995;1:11–6. Zhang J, Dean D, Nosco D, Strathopulos D, Floros M. Effect of chiropractic care on heart rate variability and pain in a multisite clinical study. J Manipulative Physiol Ther 2006;29:267–74. Zubieta JK, Bueller JA, Jackson LR, Scott DJ, Xu Y, Koeppe RA, et al. Placebo effects mediated by endogenous opioid activity on mu-opioid receptors. J Neurosci 2005;25:7754–62.

Manual Therapy 14 (2009) 539–543

Contents lists available at ScienceDirect

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

Original Article

Athlete compliance to therapist requested contraction intensity during proprioceptive neuromuscular facilitation Peter W. Sheard a, *, Paul M. Smith b, Tim J. Paine a a b

School of Physical Education and Sports Sciences, University of Bedfordshire, Luton, UK Cardiff School of Sport, University of Wales Institute Cardiff, Cardiff, Wales, UK

a r t i c l e i n f o

a b s t r a c t

Article history: Received 9 April 2008 Received in revised form 7 August 2008 Accepted 27 August 2008

Contraction intensities between 10 and 100% maximal voluntary contraction (MVC) have been proposed in varying muscle energy technique (MET) and proprioceptive neuromuscular facilitation (PNF) postisometric relaxation (PIR) protocols. The current study was undertaken to determine if athletes were able to comply with differing therapist requested contraction intensities during (PNF) stretching protocols. Thirty-six university athletes were recruited and MVC was established at hip extension, hip adduction, and horizontal shoulder adduction. Target PIR contractions were set at 20, 50 and 100% MVC and monitored throughout the contractions with a strain gauge dynamometer. Athletes were not able to match the target contraction values at 20 and 100% MVC (P  0.001). When examined for consistency across the three component contractions within each of the three PIR protocols, the athletes demonstrated widely variable scores (coefficient of variation (CV) ¼ 23.2–36.4% at 20% MVC; CV ¼ 19.3–29.4% at 50% MVC; and, CV ¼ 9.4–14.5% at 100% MVC). Our findings indicate that this group of athletes displayed a poor level of compliance to varying therapist requested contraction intensities with respect to both accuracy and consistency. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Muscle stretching exercise Rehabilitation Range of motion Articular Patient compliance

1. Introduction A survey conducted by Surburg and Schrader (1997) suggested that of nine proprioceptive neuromuscular facilitation techniques (PNF) investigated, the two most commonly used are the contractrelax (CR) and hold-relax (HR) flexibility interventions. As suggested by Chalmers (2004), these techniques focus on the initial contraction and subsequent relaxation of the target muscles (CR) or of the opposing muscles (HR). The terms target and opposing are preferred to those of agonist and antagonist that are more commonly used in the literature; although often in contradictory ways. Whilst the efficacy of PNF is generally accepted (Alter, 1996; Shrier and Gossal, 2000; Rowlands et al., 2003; Chalmers, 2004) there exists a broad spectrum of variations in application of the technique. Whether an athlete is moved from contraction to new point of bind before or after the relaxation phase, duration of the contraction phase, duration of the relaxation phase, ratio of contraction to relaxation durations, and intensity of contraction are all potential variables. The variable associated with contraction

intensity is infrequently monitored and controlled, relying instead on the subjective ‘feel’ of the athlete and therapist to estimate the contraction intensity. Due to the importance of this variable, several studies (Magnusson et al.,1996; Burke et al., 2000; Feland and Marin, 2004) employed mechanical devices to ensure that the contraction produced by the subject matches that requested by the therapist. Although a range of contractile intensities, from 10 to 100%, are proposed in the literature (Table 1), no study has explicitly documented the athletes’ ability to comply with the proposed contraction intensity in the absence of quantitative, mechanical monitoring. Therefore, the principal objective of this investigation was to examine athlete compliance with three different therapist requested contraction intensities (20, 50 and 100% of maximum voluntary contraction, MVC). Contraction forces were produced in the absence of feedback during a standardised post-isometric contraction PNF protocol at three different joint complexes. Both therapists and participants were blind to the force of contraction; forces were recorded for later compliance analysis. 2. Materials and methods

* Corresponding author. School of Physical Education and Sports Sciences, University of Bedfordshire, Park Square, Luton, Bedfordshire LU1 3JU, UK. Tel.: þ44 (0)1234 400 400. E-mail address: [email protected] (P.W. Sheard). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.08.006

2.1. Participants During recruitment, five athletes were excluded from participation due to history of traumatic hip, knee (n ¼ 2), shoulder or

540

P.W. Sheard et al. / Manual Therapy 14 (2009) 539–543

Table 1 The range of contraction intensities proposed when applying post-isometric relaxation (PIR) type proprioceptive neuromuscular facilitation (PNF). Author

Contraction intensities indicated (%)

Lewit, 1999 Chaitow and DeLany, 2002 Feland and Marin, 2004 Burke et al., 2000 Padua et al., 2000 Schmitt et al., 1999 Bonnar et al., 2004 Cornelius and Hinson, 1980 Funk et al., 2003 Moore and Hutton, 1980 Osternig et al., 1990 Padua et al., 2000 Pincivero et al., 2003 Spernoga et al., 2001

10, 20 20, 50 20, 60, 100 50, 75 50, 75 75 100 100 100 100 100 100 100 100

elbow injury in the past 6 months or hip flexion measures that exceeded 80 (n ¼ 3). The implementation of these exclusion criteria resulted in a screened sample of 36 university field sport athletes comprised of 24 men (20.4  0.8 years, 182.4  7.7 cm, 83.3  14.1 kg) and 12 women (20.5  0.8 years, 168.1  5.7 cm, 65.4  9.3 kg). Inclusion criteria was limited to apparently healthy individuals, aged 18–45 years, currently active in university field sports, with a minimum of 5 h of sport practice, training and/or competition per week. The local Institutional Review Board approved all experimental procedures and all subjects provided written informed consent prior to their participation. Sample size was determined following Hopkins’ Eq. (3) (2000a):

8S2 =d2 where S is the standard error of the mean and d is the smallest worthwhile effect. The value for S was estimated at 11.8% (Hopkins, 2000b) derived from test/re-test MVC values of 12 male university field sport athletes, all of who were later recruited as participants for the study. The smallest worthwhile effect was estimated at 10% (Campos et al., 2002) deviation from target contraction in the actual PNF trials. From these values the sample size was n ¼ 11.1 where b ¼ 0.80 and a ¼ 0.05. As such, 12 subjects per joint complex group were utilised. This allowed for completion of Latin Square assignations without ‘cropping’ the Latin Square. 2.2. Procedures Athletes followed the procedure described by Stoll et al. (2002) to establish their MVC strength for hip extension, hip adduction and horizontal shoulder adduction, the three trial sub-groups to which they were assigned via sequence of presentation: athletes were required to repeat two, 5 s MVC with a 30 s passive rest interval between attempts; no incentive was provided by the tester in order that measuring conditions be standardised; the higher of the two contraction forces, as recorded by a calibrated strain gauge dynamometer (ErgoMeter, Globus, Codogne, Italy), represented the participants’ MVC. Target PNF contractions were set at 20, 50 and 100% MVC, randomly assigned against a Latin Square, and monitored throughout the duration of each contraction with the strain gauge dynamometer. Participants’ names were drawn from one ‘hat’ and protocol sequences from a second. The following PNF stretch protocol was used: active motion (hip flexion, hip abduction or horizontal shoulder abduction) to first bind point; build to target contraction (isometric hip extension (Fig. 1), hip adduction (Fig. 2) or horizontal shoulder adduction (Fig. 3) as appropriate) over 5 s; hold target contraction for 7 s; active re-set to new bind point; relax for 12 s (Paine, 2007). This sequence was repeated for three

Fig. 1. Participant position and equipment set-up for hip extension (contraction – solid arrow) for increased hip flexion (stretch – broken arrow).

contractions. Isometric contractions were resisted using anchored straps with the strain gauge in-line rather than by the therapists who supervised the sessions and re-set the strap tensions between contractions. 2.3. Statistics A 3  3  2, three-way, mixed measures analysis of variance (ANOVA) with Bonferroni post-hoc analysis provided insight into the athletes’ ability to comply with the target contraction using Statistics Package for Social Sciences (SPSS Inc., V.12.0.1, Chicago, IL, USA). Typical error of measurement, reported as coefficient of variation (Hopkins, 2000a), was used to establish the intra-subject reproducibility of intensity of contraction between the three contractions within each PNF intervention. 3. Results 3.1. Intensity differentiation Within-group comparisons of forces generated by the athletes at each of the three target intensities (Fig. 4) indicated they could broadly distinguish between ‘easy’ (20% MVC), ‘moderate’ (50%

Fig. 2. Participant position and equipment set-up for hip adduction (contraction – solid arrow) for increased hip abduction (stretch – broken arrow).

P.W. Sheard et al. / Manual Therapy 14 (2009) 539–543

Fig. 3. Participant position and equipment set-up for shoulder horizontal adduction (contraction – solid arrow) for increased shoulder horizontal abduction (stretch – broken arrow).

MVC) and ‘maximal’ (100% MVC) contraction intensities. For all joint complexes, athletes exhibited an ability to produce a significantly different level of muscular contraction against the 20 vs. 50, 50 vs. 100, and 20 vs. 100% MVC requests (all interactions P  0.001). 3.2. Joint complex distinctions Between-group comparisons of joint complex contraction intensity at each of the three target contraction intensities (Fig. 5) shows that the athletes demonstrate a consistent pattern of contractile performance (P ¼ 0.173–1.000) at all but one interaction. At the target intensity of 20% MVC, the shoulder adduction group produced a 16.5% (19.6 vs. 36.1%) lower contraction intensity compared to that of the hip adduction group (P ¼ 0.002; 95% confidence interval (CI) ¼ 27.8 to 5.3%). 3.3. Intensity compliance When all joint complex data were pooled (Fig. 6), it became apparent that athletes were unable to comply with the requested contraction intensities associated with 20% MVC (P  0.001; 95% CI ¼ 4.3 to 11.7% MVC) and 100% MVC (P  0.001; 95% CI ¼ 18.9 to 6.4% MVC).

Fig. 4. Contraction intensities as % MVC at three joint actions. All comparisons of 20 vs. 50 vs. 100% are significant (P  0.001) indicating that athletes can distinguish between ‘easy’ (20% MVC), ‘moderate’ (50% MVC) and ‘maximal’ (100% MVC) contractions.

541

Fig. 5. Contraction intensities as % MVC at three target contraction intensities. All comparisons of between joint intensities are non-significant (P ¼ 0.173–1.000) except at target 20% MVC where shoulder adduction presented significantly lower contraction intensity than hip adduction (P ¼ 0.002). This indicates that there is a similar pattern of contraction intensity across the three joint complex contractions.

At the hip joint, a similar inability to comply with requested contraction intensities was observed for both extension and adduction; an over-performance (28.4 and 36.1%) at 20% MVC, general compliance (42.8 and 50.5%) at 50% MVC and underperformance (83.7 and 87.4%) at 100% MVC. Horizontal shoulder adduction resulted in an acceptable level of compliance at all requested intensities (19.6, 47.2 and 91.0% at 20, 50 and 100% MVC, respectively), but again followed the pattern of under-performance at 100% MVC. 3.4. Contraction reproducibility Results of the separate ANOVA tests suggested that there was an acceptable level of compliance at 50% MVC of hip extension and adduction, and across all three requested intensities for shoulder adduction (Fig. 5). However, when the typical errors of measurement, reported as coefficient of variation (Table 2), were considered it became evident that there was a poor level of consistency. Magnitudes of the repeated contractions were erratic; contraction intensities produced deviated from contraction intensities requested by 0.75 to 5.0 (7.5–50.4%) the value set as the smallest worthwhile effect (10%).

Fig. 6. Mean difference and 95% CI of target vs. actual contraction intensity at 20, 50 and 100% of MVC, at hip extension (hip ext), hip adduction (hip add) and shoulder adduction (shoulder add). Pooled data for the three joint complex actions is also included.

542

P.W. Sheard et al. / Manual Therapy 14 (2009) 539–543

Table 2 Typical error of measurement, reported as coefficient of variation in percent, for three component contractions within a specified PNF protocol at three target contraction intensities, across three joint complex actions.

Table 3 Maximum range of contraction intensities presented by individual participants (n ¼ 1–36) and by groups of participants (n ¼ 3  12) at three target contraction intensities, across three joint complex actions.

Contraction intensity (% MVC)

Contraction intensity (% MVC)

Hip extension

Hip adduction

Shoulder adduction

20 50 100

Subject range (%) 42.8 (41.5–84.3) 49.6 (57.5–107.1) 47.1 (75.5–122.6)

Subject range (%) 36.6 (29.5–66.1) 36.2 (30.4–66.6) 20.6 (80.2–100.8)

Subject range (%) 22.6 (12.9–35.5) 43.3 (55.1–98.4) 51.6 (51.6–103.2)

20 50 100

Group range (%) 79.6 (4.7–84.3) 97.6 (9.5–107.1) 120.2 (33.0–153.2)

Group range (%) 52.3 (13.8–66.1) 61.2 (23.4–84.6) 53.9 (67.5–121.4)

Group range (%) 35.3 (5.2 –40.5) 71.3 (27.1–98.4) 82.3 (51.6–133.9)

20 50 100

Hip extension

Hip adduction

Shoulder adduction

CV%; (95% CL)

CV%; (95% CL)

CV%; (95% CL)

36.4 (28.4–50.4) 29.4 (23.1–40.4) 14.5 (11.5–19.4)

23.2 (18.3–31.6) 28.3 (22.3–38.3) 9.4 (7.5–12.6)

30.7 (24.1–42.2) 19.3 (15.3–26.2) 14.4 (11.5–19.4)

Abbreviations: PNF, proprioceptive neuromuscular facilitation; CV%, coefficient of variation as percent; CL, confidence limit; MVC, maximum voluntary contraction.

4. Discussion On average, across the three component contractions within each of the PNF protocols, the athletes demonstrated the ability to broadly distinguish between ‘easy’, ‘moderate’ and ‘maximal’ contractions, i.e. the magnitude of force associated with 20, 50 and 100% MVC were significantly different (P  0.001) from one another. Toffin et al. (2003) suggested this sense of effort distinction might reflect the ability of the Golgi tendon organs (GTO) to recognise changes in the force encountered rather than the absolute values of those forces produced. Pincivero et al. (2003) demonstrated that during an isometric contraction of the knee extensors, subjects produced forces that were too high at low intensities but forces that were insufficient at higher intensities. This pattern of force production is confirmed in the mean over-production of force at hip extension (28.4  15.3% MVC) and hip adduction (36.1  9.0% MVC) at the 20% MVC target. Similarly, under-production of force was elicited at hip extension (83.7  25.2% MVC), hip adduction (87.4  10.9% MVC), and shoulder adduction (91.0  16.5% MVC) at the 100% MVC target. If Feland and Marin (2004) are correct in their assertion that there is no statistical or clinical difference in the efficacy of different contraction intensities (20, 60 and 100% MVC) on increased ROM, then the production of insufficient force at maximum MVC is not of concern. We concur with Feland and Marin’s assertion that maximum contraction intensities may be associated with the development of a delayed onset of muscle soreness and/or an increased risk of injury. As such, we further suggest that the overproduction of force at the 20% target level may increase the risk of injury recurrence as this lower level of contraction is often recommended in the context of rehabilitation (Lewit, 1999; Chaitow and DeLany, 2002) as opposed to competition environments and associated training prescription. Participants demonstrated a statistical compliance to requested contraction intensity at 50% MVC in all joint complexes and at 20 and 100% MVC at shoulder adduction. This observation must be tempered by the fact that mean contraction intensities are derived from widely divergent scores for each of the three component contractions within the PNF protocol. Within-individual variation across a single requested contraction intensity varies by as much as 2.8 (12.9–35.5% MVC at 20% MVC target in shoulder adduction) and within-group variation by as much as 17.9  (4.7–84.3% MVC at 20% MVC target in hip extension; Table 3). In this context, any apparent overall compliance can be viewed as a statistical artefact as opposed to being representative of any individual’s ability to comply with the therapist requested contraction intensity. Concerns over the ability of an athlete’s ability to comply with requested contraction intensities have been raised in previous studies (Osternig et al., 1990; Jackson and Dishman, 2000). Padua et al. (2000) made an explicit cautionary point that the outcomes of their study may be associated with an overestimation of contraction intensity as the force of contraction was not monitored during exercise, nor were MVC measurements made prior to the initiation

Abbreviations: MVC, maximum voluntary contraction.

of exercise. This contrasts our findings in that the lower target contraction intensity in Padua et al.’s study (2000) nominally equated to 50% MVC. At the 50% MVC target in this study, athlete’s achieved the greatest extent of compliance, albeit with a considerable degree of inter-contraction variability. Our athletes produced mean contractions of 42.8% (P ¼ 0.111; 95% CI ¼ 16.12–1.75%) for hip extension, 50.5% (P ¼ 0.910; 95% CI ¼ 8.43–9.43%) for hip adduction and 47.2% (P ¼ 0.532; 95% CI ¼ 11.70–6.15%) for shoulder adduction. These last values are interesting in that they indicate an underestimation of contraction force at the shoulder (Fig. 6), which was the joint complex investigated by Padua et al. (2000) where their results indicated an overestimation of contraction intensity. In a training environment, interplay between the athlete and the therapist is used in an attempt to achieve the desired target contraction intensities associated with PNF stretching techniques (Alter, 1996; Burke at al., 2000). It would, however, be unfair to use the above information in isolation and suggest that the athletes alone are to blame for non-compliance of target intensities. For example, Harvey et al. (2003) quantified the torque applied by physiotherapists when manually stretching the hamstrings of individuals with spinal cord injuries. As the participants were insensate in the target area they were unable to offer any feedback as to the intensity or discomfort elicited by the physiotherapists. The resilience of the tissue of the individual participants was not an issue in the regulation of intensity of torque applied as the 12 physiotherapists applied up to 40 (3–121 Nm) variability in force to any one of the 15 participants. Uncontrolled forces of up 121 Nm were applied. Compare this to the controlled 98 Nm load applied to healthy participants in the study by Magnusson et al. (1996) and the risk of injury becomes apparent. In a clinical environment we have two major considerations for application of PNF protocol: application to a painful and shortened or a previously injured musculotendonus unit; or, application to an apparently healthy musculotendonus unit to optimise ROM for sporting, or other, activities. In the first instance, the potential for injury or re-injury is of concern to the clinician. In the latter instance, the failure to apply PNF with optimum efficacy should be of concern to the clinician. 5. Conclusion The results of our investigation indicate that athletes appear to show limited ability to comply with therapist requested contraction intensities during a specified PNF protocol. However, general compliance appears to be noticeably better for muscular activity around the shoulder than around the hip. It is important to note that this apparent compliance is more likely to represent a ‘smoothing’ of data when the mean values of the three component contractions within the PNF protocol are compared. Analysis

P.W. Sheard et al. / Manual Therapy 14 (2009) 539–543

of the component contractions indicated that the athletes are unable to comply with the requested contraction intensity and unable to repeat their efforts on subsequent contractions. The clinical importance of these findings is that when using lower contraction intensities, as recommended post-injury, subjects tend to underestimate their force (i.e. over-contract) possibly setting themselves up for re-injury. Acknowledgements The authors gratefully acknowledge the assistance of sports therapists Isis Godfrey-Glynn, Mark George and Darren Rodger for their client handling skills. References Alter MJ. The science of flexibility. 2nd ed. Champaign, IL: Human Kinetics; 1996. Bonnar BP, Deivert RG, Gould TE. The relationship between isometric contraction durations during hold-relax stretching and improvement of hamstring flexibility. J Sports Med Phys Fitness 2004;44(3):258–61. Burke DG, Culligan CJ, Holt LE, MacKinnon NC. Equipment designed to simulate proprioceptive neuromuscular facilitation flexibility training. J Strength Cond Res 2000;14(2):135–9. Campos GE, Luecke TJ, Wedeln HK, Toma K, Hagerman FC, Murray TF, et al. Muscular adaptations in response to three different resistance-training regimens: specificity of repetition maximum training zones. Eur J Appl Physiol 2002;88(1-2):50–60. Chaitow L, DeLany JW. Clinical application of neuromuscular techniques: volume 2 – the lower body. New York: Churcill-Livingstone; 2002. Chalmers G. Re-examination of the possible role of Golgi tendon organ and muscle spindle reflexes in proprioceptive neuromuscular facilitation muscle stretching. Sports Biomech 2004;3(1):159–83. Cornelius WL, Hinson MM. The relationship between isometric contractions of the hip extensors and subsequent flexibility in males. J Sports Med Phys Fitness 1980;20(1):75–80. Feland JB, Marin HN. Effect of submaximal contraction intensity in contract-relax proprioceptive neuromuscular facilitation stretching. Br J Sports Med 2004;38:e18. doi:10.1136/bjsm.2003.010967. Funk DC, Swank AM, Mikla BM, Fagan TA, Farr BK. Impact of prior exercise on hamstring flexibility: a comparison of proprioceptive neuromuscular facilitation and static stretching. J Strength Cond Res 2003;17(3):489–92.

543

Harvey LA, McQuade L, Hawthorne S, Byak A. Quantifying the magnitude of torque physiotherapists apply when stretching the hamstring muscles of people with spinal cord injury. Arch Phys Med Rehabil 2003;84(7):1071–5. Hopkins WG. Measures of reliability in sports medicine and science. Sports Med 2000a;30(1):1–15. Hopkins WG. Reliability from consecutive pairs of trials. Available at: http://www. sportsci.org/resource/stats/xrely.xls. (2000b). Last accessed April 09, 2008. Jackson AW, Dishman RK. Perceived submaximal force production in young adult males and females. Med Sci Sports Exerc 2000;31(2):448–51. Lewit K. Manipulative therapy in rehabilitation of the locomotor system. New York: Butterworth-Heinemann; 1999. Magnusson SP, Simonsen EB, Aagaard P, Dhyre-Poulsen P, McHugh MP, Kjaer M. Mechanical and physiological responses to stretching with and without preisometric contraction in humans. Arch Phys Med Rehabil 1996;77(4): 373–8. Moore MA, Hutton RS. Electromyographic investigation of muscle stretching techniques. Med Sci Sports Exerc 1980;12(5):322–9. Osternig LR, Robertson RN, Troxel RK, Hansen P. Differential responses to proprioceptive neuromuscular facilitation (PNF) stretch techniques. Med Sci Sports Exerc 1990;22(1):106–11. Padua DA, Guskiewicz KM, Prentice WE, Schneider RE, Shields EW. The effect of select shoulder exercises on strength, active angle reproduction, single-arm balance and functional performance. J Sport Rehab 2000;13(1):75–95. Paine T. The complete guide to sports massage. 2nd ed. London: A & C Black; 2007. Pincivero DM, Dixon PT, Coelth AJ. Knee extensor torque, work, and EMG during selectively graded dynamic contractions. Muscle Nerve 2003;28(1):54–61. Rowlands AV, Marginson VF, Lee J. Chronic flexibility gains: effect of isometric contraction duration during proprioceptive neuromuscular facilitation stretching techniques. Res Q 2003;74(1):47–51. Schmitt GD, Pelham TW, Holt LE. A comparison of selected protocols during proprioceptive neuromuscular facilitation stretching. Clin Kinesiol 1999;53(1):16–21. Shrier I, Gossal K. Myths and truths of stretching: individualised recommendations for healthy muscles. Phys Sportsmed 2000;28(8):57–63. Spernoga SG, Uhl TL, Arnold BL, Gansneder BM. Duration of maintained hamstring flexibility after a one-time, modified hold-relax stretching protocol. J Athl Train 2001;36(1):44–8. Stoll T, Huber E, Seifert B, Stucki G, Michel BA. Isometric muscle strength measurement. Stuttgart: Thieme; 2002. Surburg PR, Schrader JW. Proprioceptive neuromuscular facilitation techniques in sports medicine: a reassessment. J Athl Train 1997;32(1):34–9. Toffin D, McIntyre J, Droulez J, Kemenya A, Berthoz A. Perception and reproduction of force direction in horizontal plane. J Neurophysiol 2003;90(5): 3040–53.

Manual Therapy 14 (2009) 544–549

Contents lists available at ScienceDirect

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

Original Article

Validity of the Doppler velocimeter in examination of vertebral artery blood flow and its use in pre-manipulative screening of the neck Lucy C. Thomas a, *, Darren A. Rivett a, Philip S. Bolton b, c a

Discipline of Physiotherapy, School of Health Sciences, Faculty of Health, The University of Newcastle, University Drive, Callaghan 2308, NSW, Australia Discipline of Human Physiology, School of Biomedical Sciences, Faculty of Health, The University of Newcastle, NSW, Australia c Hunter Medical Research Institute, New Lambton, NSW, Australia b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 15 April 2008 Received in revised form 21 August 2008 Accepted 27 August 2008

Pre-existing compromise of one or both vertebral arteries is considered a contraindication to neck manipulation. Current pre-manipulative screening tests may not adequately identify individuals with such compromise. It has been proposed that using a continuous wave ultrasound device (Doppler velocimeter) may assist in identifying patients presenting with flow abnormalities. The aim of this study was to determine the validity and reliability of the use of a velocimeter in detecting altered vertebral artery blood flow. Blood flow in the atlanto-axial segment of seated healthy adult volunteers (n ¼ 60) was examined in the neutral and end-range contralateral rotation positions. Duplex ultrasound scans were performed (n ¼ 58) and identified 17 volunteers (29.3%) with abnormal flow according to predetermined criteria. Three trained physiotherapists blinded to the duplex examination results used a velocimeter to examine the vertebral arteries of the volunteers. The specificity of the velocimeter examination to detect abnormal flow identified by the duplex examination was fair to good (range 0.78–0.88). However, its sensitivity was poor (range 0.25–0.38) and the inter-examiner reliability was poor (k ranged from 0.15 to 0.26). This study suggests that the velocimeter may be neither a valid or reliable tool for the detection of abnormal blood flow in the vertebral arteries. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Vertebral artery Cervical manipulation Cervical vertebrae Ultrasonic diagnosis

1. Introduction Stroke is a rare but devastating consequence of cervical spine manipulation (Thiel et al., 2007; Cassidy et al., 2008), and is thought to be associated with emboli or injury of the vertebral artery (VA) (Frisoni and Anzola, 1991; Lee et al., 1995; Cassidy et al., 2008). In a recent review (Thomas et al., 2008) we critically considered the validity of clinical tests intended to identify people at risk of vertebrobasilar insufficiency associated with cervical spine manipulation. In addition, we considered the merit of using a portable continuous wave Doppler device, known as a velocimeter, to examine the integrity of VA blood flow prior to neck manipulation (Thomas et al., 2008). The criterion standard for examining extracranial arterial patency is magnetic resonance angiography (MRA) (Zwiebel, 2000), however, this is a specialized investigation not readily available to manual therapy practitioners. Duplex ultrasound has also been used to examine VA blood flow and while there can be some

* Corresponding author. Tel.: þ61249218680; fax: þ61249217053. E-mail address: [email protected] (L.C. Thomas). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.08.007

difficulties with accuracy associated with its use, it would appear to offer the most direct and least invasive method of measuring VA blood flow. Haynes (2002) and Rivett (2001) have each proposed the use of a handheld Doppler velocimeter as a simple screening tool to provide an objective means for clinicians to assess VA blood flow. Haynes has reported that the velocimeter has an inter-rater reliability of 0.78 and a sensitivity and specificity of 100% for identifying ‘persistent’ or ‘major reduction’ in VA blood flow when compared to findings from duplex ultrasound (Haynes, 2000; Haynes et al., 2000). We have undertaken and report here, a further study of the sensitivity and specificity of the velocimeter to identify altered blood flow in the VA. In contrast to Haynes (2000), all participants in our study underwent both duplex ultrasound examination and velocimeter examination of the VA using a double blind study format. The specific aims of this study were to determine whether physiotherapists, trained in the use of the Doppler velocimeter, are able to reliably detect VA blood flow ‘abnormalities’ in the atlantoaxial segment of the VA and to determine the sensitivity and specificity of this Doppler velocimeter examination when compared to duplex ultrasound examination.

L.C. Thomas et al. / Manual Therapy 14 (2009) 544–549

2. Method The protocol used in this study was approved by the University of Newcastle Human Research Ethics Committee and the Hunter Area Research Ethics Committee. The study was performed in accordance with the 2007 National Statement on ethical conduct in research involving humans (NHMRC, 2007). The study was undertaken in two stages. Firstly, participants were examined with duplex ultrasound by a qualified ultrasonographer, to determine their VA blood flow status. Secondly, participants were examined by three physiotherapists using a Doppler velocimeter, according to a protocol described by Haynes (2000). 2.1. Recruitment Healthy adult volunteers aged between 18 and 80 years were recruited via advertisement. Exclusion criteria were history of cervical spine instability, inflammatory disease or uncontrolled cardiovascular disease which would normally contraindicate cervical manipulation. Eligible volunteers became participants in the study once they gave their informed consent to take part. Three local physiotherapists were recruited to perform the velocimeter examinations. Inclusion criteria were registration as a physiotherapist in New South Wales, at least five years of clinical experience in the field of manual therapy and familiarity with the current Australian Physiotherapy Association Clinical Guidelines for Pre-manipulative Procedures for the Cervical Spine (Magarey et al., 2000). 2.2. Protocols Each participant underwent examination of their left and right VA with a Phillips HDI 3000 Duplex ultrasound scanner (ATL Philips, Bothel, USA) with colour flow imaging. A 38 mm broadband linear transducer with a frequency of between 5 and 12 MHz (mean 6 MHz) was used by a qualified ultrasonographer experienced in vascular imaging, who had additional training to the specific requirements of the study. To facilitate stabilisation of blood pressure (BP) and heart rate (HR), each participant was asked to rest in a relaxed seated position for 10 min prior to duplex or velocimeter examination. These parameters were measured before and after scanning using a digital BP monitor with an accuracy of 3 mm Hg for BP and 5% for HR (Nissei digital BP monitor, model DS-105E, Nihon Seimitsu Co. Ltd, Japan). Any individual who had high BP (systolic  180 mm Hg) as defined by the Australian Heart Foundation (Heart Foundation, 2004) was excluded from the study because this is a contraindication to cervical manipulation (George et al., 1981) and because their BP could vary due to their condition and confound the results of the ultrasound examinations. A cervical range of motion device (CROM) (Performance Attainment Associates, Minnesota, USA) was used to measure cervical range of rotation during the ultrasound procedures. This device has been demonstrated to have good (ICC 2:1 > 0.80) intratherapist reliability in measurement studies of cervical range of movement (Capuano-Pucci et al., 1991; Youdas et al., 1991; Rheault et al., 1992). 2.3. Duplex examination The participant was seated beside the scanner. The ultrasonographer applied conducting gel to the broadband transducer and placed the probe in the sub-occipital region just posterior to the mastoid process. Using real-time imaging and colour flow mapping, blood flow in the VA was identified between the atlas and axis, where it exits from the axial transverse foramen. After 30 s in the

545

neutral position, the Doppler waveform trace was recorded, with values for peak systolic (PS), end diastolic (ED) and time averaged (TAV) velocities. Each velocity was averaged over five cardiac cycles. The participant was then asked to turn their neck slowly to the contralateral side while the ultrasonographer followed with the transducer, maintaining visual contact with blood flow on the realtime image screen. Once maximum rotation was achieved, adjustments to the cursor for midstream sampling and angle correction (60 ) were made by the ultrasonographer for the new position to ensure optimum signal. This position was maintained for a minimum of 30 s before the Doppler waveform trace and velocity values were again recorded. The participant’s maximum range of neck rotation was simultaneously recorded by another examiner using the CROM device. The participant then returned their neck to the neutral position and rested for 30 s. The procedure was then repeated for the opposite VA. Throughout the examination procedure the participant was carefully monitored for any symptoms of vascular compromise (such as dizziness, paraesthesia) or pain. The entire ultrasound examination procedure was repeated for each VA. Following the duplex examination, participants were invited to attend a local physiotherapy clinic for the velocimeter examination. 2.4. Velocimeter examination The velocimeter examination was performed by each of three physiotherapists using a Huntleigh Super Dopplex II velocimeter (Huntleigh Diagnostics, Perth, Western Australia). This bidirectional continuous wave Doppler velocimeter was used with a 4 MHz handheld probe, designed specifically for detecting blood flow in deep lying vessels. The examiners first undertook a 3 h training workshop during which they were briefed on the experimental protocol for the velocimeter examination of the VA and recording of findings. Each was provided with opportunity to practice the examination technique under the supervision of one of the investigators who had received training from an experienced user of the velocimeter. During the workshop, audio recordings of abnormal VA blood flow were used to train the physiotherapists. At the conclusion of the workshop the examiners were each loaned a velocimeter and asked to use it three to four times per week over a two-month period on different subjects. This was done to emulate the typical frequency of use of pre-manipulative tests by physiotherapists (Magarey et al., 2004). At the conclusion of the training period the same investigator assessed the examiners by means of an oral and practical test to ensure minimum standards of competency. Participants underwent velocimeter examination by each of the trained examiners in random order. One examiner did not examine one participant because the participant was known to the examiner. Participants were first seated and rested for 10 min to facilitate stabilisation of BP and HR which were measured at the conclusion of this acclimatisation period. The CROM was then attached to the participant. Each participant had both VAs examined at the atlantoaxial segment according to the protocol described by Haynes (2000). In brief, the examiner directed the velocimeter probe to the sub-occipital level posterior to the mastoid process and angled to a point midway along an imaginary arc between the eyes of the participant. The examiner sought the characteristic low resistance Doppler auditory signal of the VA which is described as a ‘soft whiplash’ or two beat sound. The probe position was maintained for 30 s to confirm the presence and stability of the blood flow signal. The participant was then instructed to turn their head and neck as far as possible to the contralateral side while the probe was manoeuvred to ensure the VA signal remained present. Maximum range of cervical rotation was recorded with the CROM. The examiner then recorded their velocimeter findings categorising blood flow as being either ‘normal’ or ‘abnormal’ on the basis of the

546

L.C. Thomas et al. / Manual Therapy 14 (2009) 544–549

Table 1 A list of findings on velocimeter examination considered to represent aberrant (‘abnormal’) vertebral artery blood flow.

   

absence of signal in neutral cessation of signal in contralateral rotation marked decrease in signal in contralateral rotation marked increase in signal in contralateral rotation

auditory Doppler signal. Blood flow was categorised as abnormal if any of the criteria previously described by Haynes (2000) (see Table 1) were met. This procedure was repeated for the opposite VA. If an examiner failed to detect VA flow at the sub-occipital portal or lost the signal during rotation, the procedure was repeated at the axial level to confirm the lack of flow or to determine if flow could be detected at this axial (lower) level. In this position the probe was directed at the transverse process of the axis and angled medially and slightly superiorly in order to sample the blood flow end-on (i.e. moving towards the probe) where the artery forms a convex loop inferolaterally emerging from the transverse foramen, as described by Haynes (2002). The examiner was permitted to re-search for flow if they were unable to follow the flow round during rotation. At the conclusion of the velocimeter examination BP and HR were measured again. 2.5. Data analysis All data was tabulated and a descriptive statistical analysis was performed on the group data to determine the range, mean and standard deviation values of the study sample. In order to determine the reliability of the duplex examination, repeat PS, ED and TAV velocity scores were assessed using intraclass correlation coefficients (ICC 2, 1) with 95% confidence intervals (Shrout and Fleiss, 1979). The repeat recordings of PS, ED and TAV flow values in each head position were averaged. Using these averaged values, the difference in blood flow velocity (PS, ED, TAV) between the two head positions was determined by calculating the difference in velocity between the two positions for each individual. Participants were categorised into those with ‘normal’ and ‘abnormal’ VA blood flow based on the results of the duplex examination and categorised as ‘abnormal’ if any of the criteria listed in Table 2 were met. Criteria were derived from the work of Freed et al. (1998) and Zwiebel (2000). Participants were categorised into those with ‘normal’ or ‘abnormal’ VA flow based on the velocimeter examination using the criteria listed in Table 1. The results of the duplex ultrasound and the velocimeter examination for each of the three examiners were compared using a 2  2 contingency table. Sensitivity, specificity and likelihood ratios were calculated for each examiner. Cohen’s kappa statistic was used to evaluate the level of agreement between each examiner’s findings (Fleiss, 1971, 1986).

Table 2 A list of findings on duplex examination considered to represent aberrant (‘abnormal’) vertebral artery blood flow.

 absence of peak systolic flow in neutral or contralateral rotation positions  absence of end diastolic flow in neutral or contralateral rotation positions  50% or more change in peak systolic flow velocity between neutral and rotation positions

Table 3 A table showing intra-class correlation coefficients (ICC 2,1) and 95% confidence intervals (CI) of the two duplex ultrasonographic measurements of peak systolic and end diastolic blood flow velocities in the neutral (neutral position) and the endrange rotated neck positions (rotation position), for the left and right vertebral arteries (VA) of each subject (n ¼ 58). Measure

Neutral position ICC (95% CI)

Rotation position ICC (95% CI)

Left VA peak systolic Left VA end diastolic Right VA peak systolic Right VA end diastolic

0.71 0.83 0.79 0.79

0.72 (0.56–0.83) 0.72 (0.56–0.82) 0.85 (0.75–0.97) 0.77 (0.65–0.86)

(0.56–0.82) (0.73–0.89) (0.67–0.87) (0.67–0.87)

Statistical analysis was undertaken using SPSS 11.0 statistical package for Windows (SPSS Inc., Chicago, Illinois, USA).

3. Results Nineteen males and 41 females entered the study, with an age range of 21–71 years and mean of 45.5 years (15.5). No participants complained of neck pain or reported any symptoms of vertebrobasilar insufficiency entering the study, and all remained asymptomatic throughout both examinations. The mean systolic and diastolic BP for participants was 126.53  16.81 and 82.88  13.54 mm Hg, respectively. No participants demonstrated changes in BP during the study period which would have excluded them from the study. The mean HR was 78.46  16.72 beats per minute (bpm). Cervical rotation during scanning ranged from 69.19  10.67 (R rotation, duplex) to 71.84  9.34 (R rotation, velocimeter) and from 70.71 10.74 (L rotation, duplex) to 70.61 10.81 (L rotation, velocimeter). There was no significant difference in range of motion between the duplex or velocimeter examinations.

3.1. Duplex examination Fifty-eight participants were successfully scanned with duplex ultrasound. Two participants were excluded due to ill-defined Doppler traces. Both excluded participants had short necks with well-developed musculature. This may have caused attenuation of the reflected Doppler signal. The ICCs of the duplex ultrasonographic measurement of PS and ED ranged from 0.71 to 0.85 (see Table 3), suggesting good repeatability (Fleiss, 1986) of the procedure. Of the 58 participants scanned with duplex ultrasound, 17 individuals (29.3%) involving 27 VAs (see Table 4) had blood flow considered to be abnormal according to the pre-determined criteria. Fig. 1 shows typical Doppler traces from two participants in the study.

Table 4 Table showing the number and distribution of vertebral arteries categorised as having, each of, aberrant peak systolic (PS) and end diastolic (ED) velocity in the left (LVA) and right (RVA) vertebral arteries in the neutral and contralateral rotation positions (n ¼ 27). L VA

R VA

Total

Absent PS flow neutral Cessation PS flow rotation 50% increase PS velocity 50% decrease PS velocity

0 0 1 1

0 0 4 0

0 0 5 1

Absent ED flow neutral Cessation ED flow rotation Absent ED flow throughout

3 3 6

0 3 6

3 6 12

14

13

27

Total

L.C. Thomas et al. / Manual Therapy 14 (2009) 544–549

547

Fig. 1. Photograph of the Doppler traces of two vertebral arteries demonstrating (i) 50% increase in peak systolic velocity from the neutral position (panel A) to the contralateral rotation position (panel B) and (ii) loss of end diastolic flow from the neutral position (panel C) to the contralateral rotation position (panel D).

The kappa scores ranged from 0.15 to 0.26 (k was 0.15 for examiners 1 and 2; 0.18 for examiners 1 and 3; and 0.26 for examiners 2 and 3). These values are below 0.40, suggesting poor agreement between examiners (Fleiss, 1986; Bogduk, 1999).

Fig. 2 shows the distribution of percent change in PS velocity for all VAs. Five participants (six arteries) had PS changes of >50%. No participants demonstrated absence of PS flow in neutral or complete cessation of flow on contralateral rotation. Twelve participants had ED findings considered to be aberrant based on the criteria in Table 2. Table 4 shows the type of abnormal blood flow demonstrated. Abnormalities were equally distributed between sides.

4. Discussion The results of the present study suggest that trained clinicians cannot reliably detect VA blood flow abnormalities in the atlantoaxial region using a Doppler velocimeter. Furthermore, Doppler velocimeter examination is not a valid method of assessing VA blood flow for the types of aberrant blood flow identified in this study. We found the use of the velocimeter by the clinicians to have poor sensitivity (0.25–0.38) in detecting aberrant VA blood flow demonstrated by duplex. This suggests that use of the velocimeter, at least in this study, would result in a high proportion of false negative findings. The specificity of the velocimeter examination was fair to good (0.78–0.88), suggesting the examination would yield few false positive results. Nonetheless, the likelihood ratios varied between the three examiners. In particular, one examiner had a notably different positive likelihood ratio (3.08) compared to the other two examiners (1.22 and 1.25). It is not clear to us why

3.2. Velocimeter examination Fifty-seven participants completed the velocimeter examination. One participant was lost to follow up due to re-location. Table 5 presents a comparison of the individual results from the duplex examination and velocimeter examination findings obtained by each of the physiotherapist examiners. Table 6 summarises the sensitivity, specificity and likelihood ratios of the velocimeter examination compared with the duplex. The specificity of the velocimeter examination was between 0.78 and 0.88. This can be considered to be fair to good (Fleiss, 1986). In contrast, the sensitivity of the velocimeter examination (range 0.25–0.38) was poor (Fleiss, 1986). Examiner 1 had an appreciably higher positive likelihood ratio.

Distribution of Peak Systolic Changes on Rotation

Number of arteries n=116

14 LVA

12

RVA

10 8 6 4 2

>100

90–100

80–90

70–80

60–70

50–60

40–50

30–40

20–30

10–20

0–10

-10–0

-20–10

-30–20

-40–30

-50–40

-60–50

-70–60

-80–70

-90–80

-100–90

<-100

0

% change Fig. 2. Graph showing the distribution of peak systolic velocity (PS) changes from the neutral position to the contralateral neck rotation position for left (LVA) and right (RVA) vertebral arteries.

548

L.C. Thomas et al. / Manual Therapy 14 (2009) 544–549

Table 5 A table showing combined 2  2 contingency tables of the individual results of the comparison between duplex and velocimeter examination findings for each of the physiotherapist examiners. Velocimeter

Examiner 1

Examiner 2

Duplex

þve ve Total

Total þve

ve

6 10

5 36

16

41

Examiner 3

Duplex

Total

þve

ve

11 46

4 11

9 32

57

15

41

Duplex

Total

þve

ve

13 43

4 12

8 33

12 45

56

16

41

57

þve refers to ‘abnormal’ vertebral artery flow.

this difference occurred, other than perhaps examiner 1 may have been more diligent during their clinical practice sessions. These likelihood ratios suggest, at least for two examiners, the velocimeter examination added little to their ability to predict the presence of altered blood flow (Bogduk, 1999). Examiner 1 may have obtained some limited clinical advantage in detecting altered blood flow in the VA, as positive likelihood ratios of greater than 3 are considered clinically useful (Bogduk, 1999; Deeks and Altman, 2004; Herbert, 2004; Perera and Heneghan, 2006). Our study findings are in contrast to the previous study by Haynes (2000) which reported that sensitivity and specificity of using the velocimeter to detect ‘marked’ VA blood flow changes, confirmed by duplex ultrasound examination, were both 1 (i.e. 100% agreement). In contrast, our study found that the velocimeter has a poor sensitivity despite (50%) changes in PS velocity in the VA. Interestingly, Haynes (2000) reported persons with changes in PS velocity of as little as 35% had been identified by ‘marked’ or ‘complete’ loss in velocimeter (audio) signal. In our study, participants had to have a change in PS velocity of 50%. In contrast to Haynes (2000), our study did not identify any participants with complete cessation of Doppler signal. However, our duplex examination did find 29% of the participants (23.28% of arteries) had altered VA blood flow, based on our criteria (see Table 2), during contralateral rotation. Haynes (2000) found 20% (12.5% of arteries) of participants in his study had altered VA blood flow. Haynes’ preselection of patients by velocimeter may have contributed to this discrepancy. Perhaps the difference in velocimeter findings between the two studies may be at least partially attributable to an examiner who had substantial experience (three years) in VA velocimeter examination. Haynes et al. (2000) undertook a further study assessing interexaminer reliability of the velocimeter involving the experienced examiner and a trained novice. The study reported ‘excellent’ (k ¼ 0.78) agreement between examiners. This is in marked contrast to the ‘poor’ (k ¼ 0.15–0.26) agreement among examiners in our study. We are at a loss to explain this large discrepancy between the two studies. It is possible that the training and experience of the examiners using the velocimeter in our study was inadequate in spite of including a formal training period and assessment. Our data suggest, in agreement with Haynes et al. (2000), that the velocimeter is highly operator dependent. This is unacceptable for any diagnostic or screening test. Current pre-manipulative testing procedures lack objective information on blood flow status (Refshauge, 1994). The inclusion

of a Doppler (velocimeter) examination of the VA as part of a premanipulative testing regimen is not an unreasonable proposal given the findings of Haynes (2000). Despite our results, it could be argued that the use of the velocimeter may still allow detection of some individuals with abnormal VA blood flow (Thomas et al., 2007). Notably, our study suggests that the velocimeter would have identified approximately 30% more patients with altered VA blood flow than would have been detected by positional tests described in pre-manipulative guidelines (International Federation of Manual Medicine, 1979; George et al., 1981; Rivett et al., 2006). Moreover, use of the velocimeter may assist in earlier referral of patients with potential VA blood flow alterations to a medical specialist for more comprehensive examination. It has been considered by some authors (Haynes, 2002; Rivett et al., 2003) that marked reduction or loss of VA blood flow on neck rotation can be an indication of biomechanical stress of the VA induced by this position, and should be considered an independent risk factor for stroke (Weintraub and Khoury, 2004). It is proposed that this should be considered a contraindication to neck manipulation as a manipulative thrust might cause further stress of the artery resulting in mechanical damage to the vessel (Haynes, 2000). Importantly, it remains to be demonstrated that alterations in blood flow correlate with higher risk of cerebrovascular compromise following neck manipulation. Given that the individuals did not demonstrate any symptoms or signs of VBI it would seem that alterations in blood flow may in fact represent normal physiological variation and that the collateral flow was adequate to maintain flow to the brain. However, a prudent clinician should attempt to gain as much clinical data about the integrity of the cerebrovascular system, especially blood flow to the brain, before considering the use of neck manipulation (Kerry and Taylor, 2006). Based on our data however, we question the validity of findings obtained from VA velocimeter examination. In conclusion, the use of the velocimeter to detect abnormalities in VA blood flow in the upper cervical spine in neutral and endrange contralateral rotation positions is not supported by the present study. Therefore, its use as a pre-manipulative screening tool cannot be recommended at this time. Further research using a study sample with more pronounced VA blood flow abnormalities, such as total occlusion or cessation of flow in rotation, might yield more positive results and better define the circumstances in which examination with a Doppler velocimeter may reliably assist the clinician to determine the integrity of flow in the VA. Acknowledgements

Table 6 A table of the sensitivity, specificity and likelihood ratios for velocimeter examination for each of the individual physiotherapist examiners compared to the duplex examination findings.

Sensitivity Specificity þve Likelihood ratio ve Likelihood ratio

Examiner 1

Examiner 2

Examiner 3

0.38 0.88 3.08 0.72

0.27 0.78 1.22 0.94

0.25 0.80 1.25 0.94

We are grateful to Greg O’Connor from the Department of Diagnostic Imaging, John Hunter Hospital, NSW Australia for performing the Duplex ultrasound scans. References Bogduk N. Truth in musculo-skeletal medicine. truth in diagnosis – validity. Australasian Musculo-Skeletal Medicine 1999;May:32–9.

L.C. Thomas et al. / Manual Therapy 14 (2009) 544–549 Capuano-Pucci D, Rheault W, Aukai J, Bracke M, Day R, Pastrick M. Intratester and intertester reliability of the cervical range of motion device. Archives of Physical Medicine and Rehabilitation 1991;72:338–40. Cassidy D, Boyle E, Cote P, He Y, Hogg-Johnson S, Silver FL, Bondy SJ. Risk of vertebrobasilar stroke and chiropractic care. Results of a population-based casecontrol and case-crossover study. Spine 2008;33(4S):S176–83. Deeks JJ, Altman DG. Diagnostic tests 4: likelihood ratios. British Medical Journal 2004;329(July):168–9. Fleiss JL. Measuring nominal scale agreement among many raters. Psychological Bulletin 1971;76(5):378–82. Fleiss JL. The design and analysis of clinical experiments. New York: John Wiley & Sons; 1986. Freed KS, Brown LB, Carroll MD. The extracranial cerebral vessels. In: Rumack CM, Wilson SR, Charboneau JW, editors. Diagnostic ultrasound (2nd ed., vol. 1, p. 885–919). St Louis, MO: Mosby Year Book, Inc; 1998. Frisoni G, Anzola G. Vertebrobasilar ischaemia after neck motion. Stroke 1991; 22:1452–60. George PE, Silverstein HT, Wallace H, Marshall M. Identification of the high risk pre-stroke patient. Journal of Chiropractic 1981;Suppl:26–8. Haynes M. Vertebral arteries and neck rotation: Doppler velocimeter and duplex results compared. Ultrasound in Medicine and Biology 2000;26(1):57–62. Haynes M. Vertebral arteries and cervical movement: Doppler ultrasound velocimetry for screening before manipulation. Journal of Manipulative and Physiological Therapeutics 2002;25(9):556–67. Haynes M, Hart R, McGeachie J. Vertebral arteries and neck rotation: Doppler velocimeter interexaminer reliability. Ultrasound in Medicine and Biology 2000;26(8):1363–7. Heart Foundation, Hypertension guide for doctors; 2004. Available from:<www. heart.foundation.org.au>. Herbert R. The accuracy of diagnostic tests. In: Refshauge KM, Gass E, editors. Musculoskeletal physiotherapy. Clinical science and evidence-based practice. Oxford: Butterworth-Heinemann; 2004. p. 109–16. International Federation of Manual Medicine, On the prevention of accidents arising from manipulative therapy of the cervical spine. Declaration of the presidium of the German Association of Manual Medicine; 1979. Kerry R, Taylor A. Cervical arterial dysfunction assessment and manual therapy. Manual Therapy 2006;11:243–53. Lee K, Carlini WG, McCormick GF, Albers GW. Neurologic complications following chiropractic manipulation: a survey of California neurologists. Neurology 1995;45(June):1213–5.

549

Magarey M, Coughlan B, Rebbeck T. Clinical guidelines for pre-manipulative procedures for the cervical spine. Australian Physiotherapy Association & Musculoskeletal Physiotherapy Australia 2000:1–7. Magarey M, Rebbeck T, Coughlan B, Grimmer K, Rivett D, Refshauge KM. Pre-manipulative testing of the cervical spine review, revision and new clinical guidelines. Manual Therapy 2004;9:95–108. NHMRC, N.H. a. M.R.C., National statement on ethical conduct in research involving humans. A. Government; 2007. Available from: <www.nhmrc.gov.au/ publications/synopses/e35syn.htm>. Perera R, Heneghan C. Making sense of diagnostic tests likelihood ratios. Evidence Based Medicine 2006;11(Oct):130–1. Refshauge KM. Rotation: A valid premanipulative dizziness test? Does it predict safe manipulation? Journal of Manipulative and Physiological Therapeutics 1994; 17(1):15–9. Rheault W, Albright B, Byers C, Franta M, Johnson A, Skowronek M, Dougherty J. Intertester reliability of the cervical range of motion device. The Journal of Orthopaedic and Sports Physical Therapy 1992;15:147–50. Rivett D. A valid pre-manipulative screening tool is needed. Australian Journal of Physiotherapy 2001;47:165. Rivett D, Sharples K, Milburn P. Reliability of ultrasonographic measurement of vertebral artery blood flow. New Zealand Journal of Physiotherapy 2003;31(3):119–28. Rivett D, Shirley D, Magarey M, Refshauge KM. APA clinical guidelines for assessing vertebrobasilar insufficiency in the management of cervical spine disorders – 2006. Melbourne, Australia: Australian Physiotherapy Association, Muculoskeletal Physiotherapy; 2006. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychological Bulletin 1979;86(2):420–8. Thiel H, Bolton JE, Docherty S, Portlock JC. Safety of chiropractic manipulation of the cervical spine. a prospective national survey. Spine 2007;32(21):2375–8. Thomas L, Rivett D, Bolton PS. Comments in response to letters to editor regarding article: Thomas LC, et al. pre-manipulative testing and the velocimeter. Manual Therapy (2007). Manual Therapy 2007;13(1):e5–6. Thomas L, Rivett D, Bolton PS. Pre-manipulative testing and the use of the velocimeter. Manual Therapy 2008;13(1):29–36. Weintraub ML, Khoury A. Mechanical compression of the extracranial vertebral artery during neck rotation. Neurology 2004;62(June):2143–4. Youdas JW, Carey JR, Garrett TR. Reliability of measurements of cervical spine range of motion – comparison of three methods. Physical Therapy 1991;71:98–106. Zwiebel WJ. Introduction to vascular ultrasonography. Philadelphia, Pensylvania: Harcourt Health Sciences; 2000.

Manual Therapy 14 (2009) 550–554

Contents lists available at ScienceDirect

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

Original Article

Development of a clinical prediction rule to identify initial responders to mobilisation with movement and exercise for lateral epicondylalgia Bill Vicenzino a, *, Dugal Smith a, Joshua Cleland b, c, Leanne Bisset a, d a

Division of Physiotherapy, School of Health and Rehabilitation Sciences, The University of Queensland St Lucia, Queensland 4072, Australia Department of Physical Therapy, Franklin Pierce University, Concord, NH 03301; USA c Rehabilitation Services, Concord Hospital, Concord, NH; Manual Physical Therapy Fellowship Program, Regis University, Denver, CO, USA d School of Physiotherapy and Exercise Science, Griffith University, Queensland 4222, Australia b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 1 January 2008 Received in revised form 25 July 2008 Accepted 3 August 2008

The aim of this post hoc analysis was to develop a preliminary clinical prediction rule (CPR) for identifying patients with lateral epicondylalgia (LE) likely to respond to mobilisation with movement and exercise (PT). Currently practitioners do not have an evidence-based means to identify such patients a priori. Potential predictive factors were recorded at baseline and reference measures at 3 weeks after treatment was initiated. Participants (n ¼ 64) received standardised PT. After 3 weeks, participants were categorised as having experienced ‘improvement’ or ’no improvement’ with treatment. Factors with univariate relationship (p < 0.15) to ’improvement’ were entered into a step-wise logistic regression model. Receiver operator characteristic curves were used to calculate cut-off points for continuous variables. Analyses resulted in a CPR that included: age (<49 years, þLR ¼ 2.6) as well as pain free grip strength on the affected (>112 N, þLR ¼ 2.3) and unaffected side (<336 N, þLR ¼ 2.1). Probability of improvement rose from 79 to 100% if all three were positive. The CPR did not predict outcome for wait and see (n ¼ 57), indicating it was more accurate for PT. This post hoc analysis has created a Level IV CPR that with further validation will help practitioners identify responders. Future studies are required to validate the rule. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Tennis elbow Clinical prediction rule Mobilisation with movement Grip strength

1. Introduction Lateral epicondylalgia (LE), commonly know as ‘tennis elbow’, is clinically defined as pain over the lateral epicondyle of the humerus that is aggravated by gripping activities and wrist extension (Stratford et al., 1987; Haker, 1993; Pienimaki et al., 2002). Prevalence ranges from 1.3% among the general population to 15% among individuals in employment requiring repetitive gripping (Chiang et al., 1993; Ranney et al., 1995; Shiri et al., 2006). Many conservative treatments are used to manage LE (Smidt et al., 2003; Bisset et al., 2005), but there is limited evidence supporting their efficacy. A recent randomised clinical trial (RCT) highlighted the benefits of a physiotherapy program including joint mobilisation and exercise compared to either corticosteroid injection or wait and see approach (Bisset et al., 2006a; Vicenzino, 2003; Vicenzino and Bisset, 2007). The mobilisation technique used was the mobilisation with movement (MWM) of the elbow as described previously

* Corresponding author. Tel.: þ61 7 3365 2275. E-mail address: [email protected] (B. Vicenzino). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.08.004

(Mulligan, 1999). Despite the fact that physiotherapy exhibited positive outcomes, not all patients with LE responded to this intervention. At present, practitioners do not have the means to differentiate between patients who will respond to this intervention from those who will not. Clinical prediction rules (CPR) are tools that can be used to assist health care practitioners in overcoming the dilemma of identifying responders to a treatment prior to initiation (Laupacis et al., 1997; McGinn et al., 2000). The purpose of a CPR is to improve the practitioner’s ability to accurately predict an outcome to intervention. Clinical prediction rules have been successfully created to identify patients with back pain who are likely to benefit from spinal manipulation (Childs et al., 2004) and lumbar stabilisation programs (Hicks et al., 2005). A CPR with the ability to identify patients with LE likely to respond favourably to physiotherapy would aid clinical decision-making. The first step in creating a CPR is to develop the rule through the identification of possible predictor factors (Childs and Cleland, 2006). The purpose of this post hoc analysis of the data from Bisset et al. (2006a) was to develop a preliminary CPR for identifying patients with LE likely to respond to physiotherapy MWM and exercise intervention early in the rehabilitation program.

B. Vicenzino et al. / Manual Therapy 14 (2009) 550–554

2. Methods Data for this analysis was collected as part of a previous RCT that investigated the effectiveness of physiotherapy, wait and see and corticosteroid approaches to treating LE of greater than 6 weeks duration (Bisset et al., 2006a). The group receiving physiotherapy was used to develop the CPR and the group who followed a wait and see policy were used in a preliminary validation analysis. The groups were similar at baseline (Bisset et al., 2006a) 2.1. Participants Volunteers from the greater Brisbane region of Australia were recruited through advertisements and media releases between March 2002 and April 2004. Volunteers were eligible for inclusion if they were aged between 18 and 65 years, had pain over the lateral elbow of at least 6 weeks duration that was provoked by palpation of the lateral epicondyle, gripping and resisted extension of the wrist, second or third finger (Haker, 1993). Volunteers were excluded from participation if they had been treated by a health care practitioner for their lateral elbow pain in the preceding 6 months; they had bilateral elbow symptoms, cervical radiculopathy, concomitant shoulder, elbow or hand pathology, peripheral nerve involvement, previous surgery of the elbow, a history of elbow dislocation, fracture or tendon rupture, systemic neurological disorders or contraindications to steroids. Self-prescribed analgesia, braces and stretches were not excluding factors. The study received approval from the institutional ethics committee and all 64 participants provided informed consent. 2.2. Treatment protocol Participants in the physiotherapy group received five 30-min treatments over 3 weeks by one of six physiotherapists with postgraduate qualifications in musculoskeletal physiotherapy. All participating therapists received specific training to assure standardisation of treatments. Treatment consisted of MWM, specifically either lateral glide of the elbow or posteroanterior glide of the radiohumeral joint (Mulligan, 1999), and prescription of an exercise program. Participants were taught home exercises and self-MWM, as previously described (Vicenzino, 2003; Vicenzino and Bisset, 2007), and given exercise equipment and an instruction booklet on how to correctly perform exercises at home. The participants following the wait and see policy were given no treatment and as for all participants in the RCT they were discouraged from seeking further treatment and were given information about the disease process, self-management and ergonomic advice. 2.3. Outcome measures An assessor blinded to treatment assignment, recorded outcome measures at baseline and 3 weeks after treatment. Global perceived effect (GPE) was measured on a six-point scale from 0 (‘completely recovered’) to 5 (‘much worse’) (Smidt et al., 2002). Pain was measured via continuous visual analogue scale (VAS) from 0 mm (‘no pain’) to 100 mm (‘worst pain imaginable’) (Carlsson, 1983). 2.4. Potential predictor variables A search of the literature revealed factors that should be considered in this study as potential predictor variables. Severity of pain, duration of symptoms and manual employment have been identified as prognostic of poor outcome in LE (Haahr and Andersen, 2003b; Smidt et al., 2006) and were included a priori in this analysis. Gender and involvement of the dominant arm were also included a priori because some studies show a bias of these factors

551

among people with LE (Pienimaki et al., 2002; Haahr and Andersen, 2003a; Shiri et al., 2007). The assessor who was blinded to treatment assignment performed a clinical examination of each participant at baseline. Duration was recorded in weeks, with those of duration greater than 3 years truncated to 156 weeks. Employment was categorised into three sub-groupings: ‘manual work,’ ‘non-manual work’, and ‘unemployed’. Pain was measured with a VAS and pain free grip strength (PFGS), the latter being a key defining feature of LE and its response to treatment (Thurtle et al., 1984; Stratford et al., 1987; Haker, 1993; Pienimaki et al., 1997; Pienimaki et al., 2002; Bisset et al., 2006b). PFGS was measured using a digital grip dynamometer (MIE, Medical Research Limited, UK; Newtons). The initial effect of MWM on PFGS was also included as a potential predictor variable. Although it has not been previously examined as a possible predictor of treatment outcome, the initial effect of MWM on PFGS is used as a guide clinically (Mulligan, 1999; Vicenzino, 2003). Percentage change in PFGS on the initial application of MWM was calculated from the treating physiotherapist’s records. PFGS was measured before and during the MWM application as described in the literature (Vicenzino, 2003). These records were then used to calculate the percentage change in PFGS during MWM expressed as a percent of pre-treatment PFGS. 2.5. Data analysis Data analysis was performed using the SPSS Version 14.0 statistical software package (SPSS Inc, Chicago, IL, USA). Patients with GPE less than 3 (‘improved’, ‘much improved’ or ‘completely recovered’) were labelled as ‘improved’ and the remainder as ‘no improvement’; ensuring an adequate number of improved cases. Treatment induced changes in pain VAS were calculated for the ‘improved’ and ‘no improvement’ groups and differences between groups were analysed using an independent t-test. Percentage change in PFGS during application of the initial MWM was expressed by a series of dichotomous variables that grouped results at cut-offs of greater than 25, 50, 75 or 100%, respectively (i.e., we tested four possible cut-offs). Potential predictor variables were tested for univariate relationship to ‘improvement’ using independent samples t-tests for continuous variables and c2 tests for categorical variables. Variables with a significance level of p < 0.15 were retained as potential prediction variables (Freedman, 1983). Often studies that set out to develop a CPR set a liberal significance level at this stage to avoid excluding any potential predictor variables (Flynn et al., 2002). For continuous variables with a significant univariate relationship, sensitivity and specificity values were calculated for all possible cut-off points, and then plotted as a receiver operator characteristic (ROC) curve (Deyo and Centor, 1986). The point on the curve nearest the upper left-hand corner represented the value with the best diagnostic accuracy, and this point was selected as the cut-off defining a positive test (Deyo and Centor, 1986). Sensitivity, specificity and positive and negative likelihood ratios (þLR, LR) were calculated for potential predictor variables. Retained potential predictor variables were entered into a step-wise logistic regression model to determine the most accurate set of variables for prediction of treatment improvement. A significance level of 0.05 was necessary to enter the variable into the model and 0.10 was required for removal from the equation to minimise the likelihood of excluding potentially helpful variables (Freedman, 1983). Variables retained in the regression model were included in the CPR for classifying patients with LE likely to benefit from 3 weeks of MWM and exercise. To further ascertain the validity of the developed CPR, and investigate if indeed it predicted response to MWM and exercise rather than the natural history of the disorder, we also calculated diagnostic accuracy statistics and post-test probability of the

552

B. Vicenzino et al. / Manual Therapy 14 (2009) 550–554

Table 1 Potential predictor variables and their univariate relationship to ‘improvement’. Variable Numbera Sex: womena Dominant side affecteda Manual laboura Age (years)b Duration of symptoms (weeks)b Baseline pain on 100-mm VASb Affected pain free grip at baseline (N)b Unaffected pain free grip at baseline (N)b First treatment session: changes with MWM MWMPFG > 25%a,e MWMPFG > 50%a,e MWMPFG > 75%a,e MWMPFG > 100%a,e

All participants

Improved

62 (100) 21 (33.9) 38 (61)

49 (79) 18 (36.7) 30 (61)

20 (43) 48.2 (7.4) 26.3 (28.0)

Not improved

p value

13 (21) 3 (23) 8 (62)

– 0.52d 0.62d

17 (34.6) 47.2 (7.6) 24.8 (24.6)

3 (23) 50.8 (5.8) 29.4 (40.1)

0.51d 0.11c 0.61c

58.3 (24.8)

56.4 (25.1)

65.2 (25.1)

0.27c

127 (67.1)

133.8 (72.2)

98.8 (39.0)

0.10c

320.3 (107.6)

308.7 (106.3)

358 (106)

0.14c

model and used to form the CPR. In order of predictive value, they were: age < 49 years, affected arm PFGS > 112 N and PFGS for the unaffected arm < 336 N (p < 0.01, Nagelkerke’s R2 ¼ 0.45). The final CPR criterion and their accuracy statistics can be found in Table 3. Of the 57 patients who were at least positive for one of the criteria, 46 were improved. Of the 34 patients that were positive for at least two of the three criteria in the CPR 31 were improved. All four of the patients that exhibited three/three criteria experienced improvements. The diagnostic accuracy analyses of the CPR in the group of patients following a wait and see policy (n ¼ 57) revealed that the lower bound estimate for the 95% confidence interval for all positive LRs was below 1 (0.23, 0.42, 0.29 for one/three, two/three and three/three variables present, respectively, Table 3). 4. Discussion

32 24 17 7

(52) (39) (27) (11)

27 19 14 4

(55) (39) (29) (9)

5 5 3 3

(38) (38) (23) (23)

0.13d 0.58d 0.45d 0.16d

p values <0.15 are shown in bold type. a Categorical data expressed as number count (% of respective group). b Continuous data expressed as mean (SD). c Independent samples t-tests. d Chi-square tests. e Cells in these rows refer to counts (% of group) of participants who demonstrated >25, 50, 75 or 100% change from baseline in pain free grip during the Mulligan mobilisation with movement (e.g., in the first cell, 32 participants exhibited a >25% change in pain free grip strength from baseline during the application of the MWM at the first treatment session).

developed CPR in the group of patients who followed a wait and see approach (n ¼ 57) in the primary clinical trial (Bisset et al., 2006a). 3. Results Sixty-four (42 male) participants who underwent physiotherapy completed the 3-week reassessment, however two participants did not have sufficient clinician-recorded treatment data and were excluded from analysis. Participant demographics and initial baseline variables from the clinical examination for the entire sample (n ¼ 62), ‘improvement’ group (n ¼ 49) and ‘no improvement’ group (n ¼ 13) can be found in Table 1. Analysis of pain scores revealed the ‘improved’ group experienced a significantly greater reduction in pain compared to the ‘no improvement’ group (15.3 points, 95% CI ¼ 0.50, 30.1). Four potential predictor variables exhibited a significance level of less than 0.15 (seen in Table 1) and were entered into logistic regression: age < 49 years, PFGS of the affected arm > 112 N, PFGS of the unaffected arm < 336 N and change in PFGS with the first MWM in situ > 25%. The univariate accuracy statistics can be seen in Table 2. Three variables were retained in the final regression

The main aim of this post hoc analysis was to develop a preliminary CPR by determining predictors of a positive response to a treatment program consisting of MWM and exercise for patients presenting with LE. Patients who were younger than 49 years and who had a high PFGS on the affected side (>112 N) and low PFGS on the unaffected side (<336 N) were most likely to respond to this intervention. This preliminary evidence indicates it may well be possible to predict which patients will respond positively to MWM and exercise early in rehabilitation. Importantly when each CPR criterion was met, the probability of improvement increased from 79% pre-test to 100% post-test. Future studies are necessary to determine if this increase is clinically meaningful. The ability to identify patients with 100% accuracy seems unlikely to remain in future studies. It is interesting to speculate about the plausibility of the CPR criteria. Age was the most significant CPR criterion, which may reflect an age-related decline in the response to short-term resistance training (Hameed et al., 2003, 2004). Hameed and colleagues found an age-related insensitivity of skeletal muscle to mechanical loading, implied by an attenuated mechano-growth factor response to high-load resistance training. PFGS also contributed to the CPR, which is consistent with prior studies showing PFGS to be a valid indicator of functional impairment (Pienimaki et al., 2002) and to be sensitive to patient improvement (Stratford et al., 1987). Interestingly, though severity of pain and duration of symptoms are prognostic indicators in LE (Smidt et al., 2006), they did not contribute to the CPR. Future studies may further investigate them for positive predictive power. In contrast to baseline PFGS, the change in PFGS affected by the initial application of MWM was not identified as a predictor in the current CPR. This study was the first, to the authors’ knowledge, to have evaluated the effect of the first MWM application as a predictor of outcome following ongoing physiotherapy. Though undocumented for LE, it is a widely held clinical paradigm that a pain-relieving technique should improve pain free exercise

Table 2 Accuracy statistics (95% confidence intervals) for potential predictor variables for 3-week response and percentage of patients who satisfied each criterion. Variable

Sensitivity (95% CI)

Specificity (95% CI)

Positive likelihood ratio (95% CI)

Post-test probability of improvement (%)c

Percentage of patients that met criteria (%)

Age < 49 years Affected pain free grip > 112 Na Unaffected pain free grip < 336 Na MWMPFG > 25%b

0.61 (0.46, 0.74) 0.53 (0.38, 0.67)

0.77 (0.46, 0.94) 0.77 (0.46, 0.93)

2.6 (0.96, 7.3) 2.3 (0.82, 6.4)

91 90

53 47

0.49 (0.35, 0.63)

0.77 (0.46, 0.94)

2.1 (0.76, 6.0)

89

44

0.75 (0.58, 0.87)

0.5 (0.20, 0.80)

1.5 (0.78, 2.9)

85

52

a b c

N, Newtons. MWMPFG ¼ percent change in pain free grip following the Mulligan mobilisation with movement. The probability of improvement is calculated using the positive likelihood ratios and assumes a pre-test probability of 79%.

B. Vicenzino et al. / Manual Therapy 14 (2009) 550–554

553

Table 3 (a) Criterion of the clinical prediction rule identified by the logistic regression analysis and their accuracy statistics for (b) the mobilisation with movement and exercise group and (c) the group who followed a wait and see policy. (a) Criterion of the clinical prediction rule identified in logistic regression analysis Age < 49 years Affected pain free grip > 112 Na Unaffected pain free grip < 336 Na Positive LRb

Probability of success (%)c

Number in improved group

(b) Mobilisation with movement and exercise 3 0.08 (0.03, 0.20) 1.0 (0.7, 1.0) 2 0.57 (0.42, 0.71) 0.85 (0.54, 0.97) 1 0.98 (0.88, 0.99) 0.46 (0.20, 0.74)

N 3.7 (1.0, 13.6) 1.8, (1.1, 3.0)

100 93 87

4 27 15

0 3 8

(c) Wait and see policy 3 0.18 (0.03, 0.52) 2 0.19 (0.09, 0.35) 0.17 (0.09, 0.30) 1d

1.2 (0.29, 5.0) 3.1 (0.42, 23.0) 1.0 (0.08, 13.6)

19 37 16

2 6 1

9 24 13

Positive criterion

Sensitivityb

Specificityb

0.85 (0.71, 0.93) 0.94 (0.68, 0.99) 0.83 (0.20, 1.0)

Number in non-improved group

a

N, Newtons. 95% CI. Probability of improvement is calculated using the positive likelihood ratios (LR) and assumes a pre-test probability of 79%; R2 ¼ 0.45 for mobilisation with movement and exercise and a pre-test probability of 16% for the wait and see group. d 0.5 was added to each cell in the table to allow for the calculation of LRs. b c

performance by 50% to be of therapeutic value. For example in patellofemoral syndrome where taping rather than MWM is used to alleviate pain with exercise, there is a CPR indicating that the tape must improve the condition by 50% for it to be of benefit (Lesher et al., 2006). In our study, a greater than 25% increase in PFGS on application of the MWM was significantly associated with improvement in the univariate analyses, but this association was not present in the multivariate analysis from which the preliminary CPR was developed. Thus in our preliminary study, response to initial MWM intervention appears not to predict the outcome of continued treatment. This study has developed a Level IV CPR, which by definition should not be immediately taken as the ultimate CPR in guiding clinical practice (McGinn et al., 2000). That is, the study was not definitive, largely due to its explorative nature as a post hoc analysis of a study powered for other hypotheses. It is important to recognise that the rating of ‘improvement’ defined a priori to this study as per the dichotomisation of GPE, was different from the more stringent definition of ‘success’ adopted previously (Smidt et al., 2002; Bisset et al., 2006a). Specifically, ‘completely recovered,’ ‘much improved’ and ‘improved’ were classified as ‘improvement’ in this study, rather than only ‘completely recovered’ and ‘much improved.’ Clinicians should draw inferences from this study with this in mind. It should also be recognised that only four patients (all of which were in the improved group) exhibited three/three criteria on the CPR. Hence, the 100% post-test probability in this study should not be expected to directly apply to other populations, as it is not likely that any variables will predict outcomes with 100% accuracy. Additionally, in such a study design it is possible that the predictors may simply predict improvement to any intervention or no intervention. However, when we investigated the diagnostic accuracy of the CPR in the group which followed a wait and see policy it appears that the predictors are specific to response to MWM and exercise. Prior to being confidently employed in clinical practice, the CPR would need to be validated by its reproduction in different groups of patients or in an alternate clinical setting and by an impact analysis utilising a prospective study design (Childs and Cleland, 2006). References Bisset L, Beller E, Jull G, Brooks P, Darnell R, Vicenzino B. Mobilisation with movement and exercise, corticosteroid injection, or wait and see for tennis elbow: randomised trial. BMJ 2006a;333:939–41.

Bisset L, Paungmali A, Vicenzino B, Beller E. A systematic review and meta-analysis of clinical trials on physical interventions for lateral epicondylalgia. British Journal of Sports Medicine 2005;39(7):411–22. Bisset LM, Russell T, Bradley S, Ha B, Vicenzino BT. Bilateral sensorimotor abnormalities in unilateral lateral epicondylalgia. Archives of Physical Medicine and Rehabilitation 2006b;87(4):490–5. Carlsson A. Assessment of chronic pain. I. Aspects of the reliability and validity of the visual analog scale. Pain 1983;16:87–101. Chiang HC, Ko YC, Chen SS, Yu HS, Wu TN, Chang PY. Prevalence of shoulder and upper-limb disorders among workers in the fish-processing industry. Scandinavian Journal of Work. Environment & Health 1993;19(2):126–31. Childs JD, Cleland JA. Development and application of clinical prediction rules to improve decision making in physical therapist practice. Physical Therapy 2006;86(1):122–31. 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):920–8. Deyo RA, Centor RM. Assessing the responsiveness of functional scales to clinical change: an analogy to diagnostic test performance. Journal of Chronic Diseases 1986;39(11):897–906. 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):2835–43. Freedman DA. A note on Screening regression equations. American Statistician 1983;37(2):152–5. Haahr JP, Andersen JH. Physical and psychosocial risk factors for lateral epicondylitis: a population based case-referent study. Occupational and Environmental Medicine 2003a;60(5):322–9. Haahr JP, Andersen JH. Prognostic factors in lateral epicondylitis: a randomized trial with one-year follow-up in 266 new cases treated with minimal occupational intervention or the usual approach in general practice. Rheumatology 2003b;42(10):1216–25. Haker E. Lateral epicondylalgia: diagnosis, treatment and evaluation. Critical Reviews in Physical and Rehabilitative Medicine 1993;5(2):129–54. Hameed M, Lange KH, Andersen JL, Schjerling P, Kjaer M, Harridge SD, et al. The effect of recombinant human growth hormone and resistance training on IGF-I mRNA expression in the muscles of elderly men. Journal of Physiology 2004;555:231–40. Hameed M, Orrell RW, Cobbold M, Goldspink G, Harridge SDR. Expression of IGF-I splice variants in young and old human skeletal muscle after high resistance exercise. Journal of Physiology 2003;547(1):247–54. Hicks GE, Fritz JM, Delitto A, McGill SM. Preliminary development of a clinical prediction rule for determining which patients with low back pain will respond to a stabilization exercise program. Archives of Physical Medicine and Rehabilitation 2005;86(9):1753–62. Laupacis A, Sekar N, Stiell IG. Clinical prediction rules: a review and suggested modifications of methodological standards. Journal of the American Medical Association 1997;277(6):488–94. Lesher JD, Sutlive TG, Miller GA, Chine NJ, Garber MB, Wainner RS. Development of a clinical prediction rule for classifying patients with patellofemoral pain syndrome who respond to patellar taping. Journal of Orthopaedic and Sports Physical Therapy 2006;36(11):854–66. McGinn TG, Guyatt GH, Wyer PC, Naylor CD, Stiell IG, Richardson WS. Users’ guides to the medical literature XXII: how to use articles about clinical decision rules. JAMA 2000;284(1):79–84. Mulligan BR. Manual therapy: NAGS, SNAGS, MWMS, etc. 4th ed. Wellington, NZ: Plane View Services; 1999. p.142.

554

B. Vicenzino et al. / Manual Therapy 14 (2009) 550–554

Pienimaki T, Siira P, Vanharanta H. Muscle function of the hand, wrist and forearm in chronic lateral epicondylitis. European Journal of Physical Medicine and Rehabilitation 1997;7(6):171–8. Pienimaki T, Tarvainen T, Siira P, Malmivaara A, Vanharanta H. Associations between pain, grip strength, and manual tests in the treatment evaluation of chronic tennis elbow. Clinical Journal of Pain 2002;18(3):164–70. Ranney D, Wells R, Moore A. Upper limb musculoskeletal disorders in highly repetitive industries: precise anatomical physical findings. Ergonomics 1995;38(7):1408–23. Shiri R, Varonen H, Heliovaara M, Viikari-Juntura E. Hand dominance in upper extremity musculoskeletal disorders. Journal of Rheumatology 2007;34(5):1076–82. Shiri R, Viikari-Juntura E, Varonen H, Heliovaara M. Prevalence and determinants of lateral and medial epicondylitis: a population study. American Journal of Epidemiology 2006;164(11):1065–74. Smidt N, Assendelft WJJ, Arola H, Malmivaara A, Green S, Buchbinder R, et al. Effectiveness of physiotherapy for lateral epicondylitis: a systematic review. Annals of Medicine 2003;35(1):51–62.

Smidt N, Lewis M, Van Der Windt DAWM, Hay EM, Bouter LM, Croft P. Lateral epicondylitis in general practice: course and prognostic indicators of outcome. Journal of Rheumatology 2006;33(10):2053–9. Smidt N, Van Der Windt DAWM, Assendelft WJJ, Deville WLJM, Korthals-de Bos IBC, Bouter LM. Corticosteroid injections, physiotherapy, or a wait-and-see policy for lateral epicondylitis: a randomised controlled trial. Lancet 2002;359(9307):657–62. Stratford P, Levy DR, Gauldie S, Levy K, Miseferi D. Extensor carpi radialis tendonitis: a validation of selected outcome measures. Physiotherapy Canada 1987;39(4):250–5. Thurtle O, Tyler A, Cawley M. Grip strength as a measure of response to treatment for lateral epicondylitis [letter]. British Journal of Rheumatology 1984;23:154–5. Vicenzino B. Lateral epicondylalgia: a musculoskeletal physiotherapy perspective. Manual Therapy 2003;8(2):66–79. Vicenzino B, Bisset L. Physiotherapy for tennis elbow. Evidence Based Medicine 2007;12(2):37–8.

Manual Therapy 14 (2009) 555–561

Contents lists available at ScienceDirect

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

Original Article

Inter-examiner reliability of a classification system for patients with non-specific low back pain K. Vibe Fersum a, *, P.B. O’Sullivan b, A. Kvåle a, J.S. Skouen a, c a

Section for Physiotherapy Science, Department of Public Health and Primary Health Care, University of Bergen, Kalfarveien 31, 5018 Bergen, Norway School of Physiotherapy, Curtin University, Bentley 6102, WA, Australia c The Outpatient Spine Clinic, Department of Physical Medicine and Rehabilitation, Haukeland University Hospital, Bergen, Norway b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 20 February 2008 Received in revised form 10 July 2008 Accepted 1 August 2008

There is a lack of studies examining whether mechanism-based classification systems (CS) acknowledging biological, psychological and social dimensions of long-lasting low back pain (LBP) disorders can be performed in a reliable manner. The purpose of this paper was to examine the inter-tester reliability of clinicians’ ability to independently classify patients with non-specific LBP (NSLBP), utilising a mechanism-based classification method. Twenty-six patients with NSLBP underwent an interview and full physical examination by four different physiotherapists. Percentage agreement and Kappa coefficients were calculated for six different levels of decision making. For levels 1–4, percentage agreement had a mean of 96% (range 75–100%). For the primary direction of provocation Kappa and percentage agreement had a mean between the four testers of 0.82 (range 0.66–0.90) and 86% (range 73–92%) respectively. At the final decision making level, the scores for detecting psychosocial influence gave a mean Kappa coefficient of 0.65 (range 0.57–0.74) and 87% (range 85–92%). The findings suggest that the inter-tester reliability of the system is moderate to substantial for a range of patients within the NSLBP population in line with previous research. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Agreement Classification Low back pain Reliability

1. Introduction LBP represents a common and very costly health problem and a definite diagnosis is difficult to achieve in most cases (85%) (Waddell, 2004). As a result, uncertainty regarding treatment of this group of patients is common (Cherkin et al., 1998). A number of studies have shown little or no difference between various physiotherapy treatments for chronic LBP (Delitto et al., 1995; Petersen et al., 1999; Ferreira et al., 2007). Several authors have suggested that these results may reflect the heterogeneity of the NSLBP group, with several distinct subgroups, including psychosocial problems, each with its own potential set of beneficial treatments (O’Sullivan, 2000; Petersen et al., 2002; O’Sullivan, 2005; Dankaerts et al., 2006b). There is growing evidence suggesting that sub-classifying patients and offering them tailored interventions matching their disorder improves patient outcome (Frymoyer et al., 1985; Main and Watson, 1996; O’Sullivan, 1997; Nachemson, 1999; Linton, 2000; Skouen et al., 2002; Fritz et al., 2003; Stuge et al., 2004). It has been proposed that a classification system (CS) for NSLBP should identify the underlying mechanisms driving the disorder within a bio-psycho-social framework, * Corresponding author: Tel.: þ47 55586711. E-mail address: [email protected] (K. Vibe Fersum). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.08.003

enabling specific therapies to be applied so as to favourably influence the outcome of the disorder (O’Sullivan, 2005). A number of CS have been proposed (McKenzie, 1981; Spitzer, 1987; Maluf et al., 2000; Sahrmann, 2001). However, only a few are found sufficiently reliable and valid (Petersen et al., 1999), and even fewer consider the disorder from a bio-psycho-social perspective (Petersen et al., 1999; Ford et al., 2003; McCarthy et al., 2004; O’Sullivan, 2005; Dankaerts et al., 2006b). The Quebec Task Force system was designed to classify all LBP patients to help with clinical decision making, establishing prognosis and evaluating treatment effectiveness (Spitzer, 1987). However, it has not been tested for reliability and does not consider the underlying mechanism (Dankaerts et al., 2006b), except for differentiating somatic from radicular pain. Within this system there is no subgrouping of NSLBP except on the basis of pain area, and no specific treatment is advocated for this large group of patients other than general exercise, therefore limiting its use for physiotherapy assessment and treatment (Padfield et al., 2002). The McKenzie (1981) system is based on information from history taking, and symptom response to generated loading of the lumbar spine. The system has been tested for reliability, and has substantial inter-tester agreement when applied by trained examiners (Kappa coefficients ranging from 0.6 to 0.7) (Kilpikoski et al., 2002).

556

K. Vibe Fersum et al. / Manual Therapy 14 (2009) 555–561

Petersen and co-workers (2004) have proposed a McKenziebased CS with good inter-tester reliability, but it has a pathoanatomical orientation and lacks clear guidelines for management. Sahrmann and co-workers have developed another CS, comprising five categories based on testing of muscular stability, alignment, asymmetry, flexibility of the lumbar spine, pelvis, and hip (Maluf et al., 2000). Reliability of the individual tests used for classification has been shown to vary from fair to almost perfect (Van Dillen et al., 1998, 2003). However, there are no reports on reliability in classification of the patients into the five categories, nor does this system consider patho-anatomical or psychosocial dimensions. Since 1997 Peter O’Sullivan has developed a novel system, based on the Quebeck Task Force, incorporating multiple dimensions in the classification of patients into subgroups based on proposed underlying pain mechanisms. Initially, this mainly targeted a subgroup of patients with localised NSLBP where provocative movement behaviours and positions of the spine, associated with a loss of spinal control, represent a mechanism for ongoing pain. These patients are classified as LBP patients with motor control impairment (MCI). The evidence validating this subgroup is growing (O’Sullivan et al., 1997, 2005; O’Sullivan, 1997, 2000, 2003; Dankaerts et al., 2006a) and the reliability of clinicians to identify these different subgroups has been established (Dankaerts et al., 2006b). Lately, this approach has also incorporated classification of patients with lumbo-pelvic pain and a wider range of pain mechanisms linked to their disorder (O’Sullivan, 2005; O’Sullivan and Beales, 2007a). This system differentiates between specific LBP versus NSLBP. NSLBP is further split into subgroups based on the proposed driving mechanism behind the disorder (Fig. 1). The classification is based on a systematic examination process (subjective history, objective examination and available medical information). Within this system psychosocial factors are accounted for, acknowledging their potential to amplify pain and drive disability. To date the ability of clinicians to agree on this broad classification process has not been formally tested. Validating the system has been a multi-step process, in which establishing inter-tester reliability is crucial. The aim of this study was therefore to examine the inter-tester reliability of clinicians’ ability to independently classify a wide range of patients with NSLBP, utilising an extended mechanism-based classification method lately developed by O’Sullivan. 2. Methods The study was conducted from March 2006 to June 2006, and was approved by the regional ethics committee of medical research in western Norway. 2.1. Patients Patients were recruited consecutively from physiotherapy clinics around Bergen and from The Outpatient Multidisciplinary Spine Clinic, Haukeland University Hospital. After recruitment, a telephone screening was performed, and the first 30 patients that fit the inclusion criteria, were tested (Table 1). Since the patients were tested twice on each occasion, a 0–10 pain numerical rating scale was conducted prior to each testing. If a patient’s pain score changed 2 levels between two examinations on the same day, this was considered to be a threat to the classification validity and the patient would then be excluded. Four patients were excluded after further examination: three did not fulfil the inclusion criteria and one reported a two-level change in pain between examinations on the given day. This left 26 patients participating in the study. See Table 2 for the patients’ characteristics. Prior to the study, design and possible

risks were fully explained to each subject, and all signed a consent form. 2.2. Examiners There were four physiotherapists, each with several years of experience in examination and treatment of LBP patients (mean 12 years, range 7–20 years). Three of the four examiners were physiotherapists with a masters degree in manual therapy. One was the developer of the system. 2.3. Training All the examiners had been educated in the CS during several workshops with the developer, and were using it in their clinical practice. Prior to the study, O’Sullivan explained procedures and classifications were discussed using a series of case studies. The examiners also underwent a pilot training period where O’Sullivan examined and classified six patients, while the three others observed. The aim was to refine the specific criteria for assessment, as well as making testers more familiar with the system. The estimated training time for each therapist ranged from 69 to 140 h, the average being 106.3 h (workshops and pilot study included). 2.4. Clinical procedure A test–retest design was utilised. A classification manual was developed by O’Sullivan prior to the study. The patients underwent a comprehensive interview and full physical examination by each of the four physiotherapists. Rather than assess the reliability of individual tests, this system involved making a disorder classification based on compilation of subjective and physical examination findings in relation to other medical tests and radiological imaging. The subjective assessment included pain area (pain drawing), intensity and nature, pain behaviour (aggravating/easing movements), identification of primary impairments, disability levels, avoidance behaviours, pain coping and pain beliefs. The examination involved assessment of spinal range of movement, analysis of the patient’s primary physical impairments (pain provocative and easing postures, movements and functional tasks). Specific muscle and movement tests were performed to identify the relationship between the control of the lumbo-pelvic region and the pain disorders (O’Sullivan, 2000), as well as specific articular tests for the lumbar spine and pelvic region as indicated to identify the structural source of pain and the presence of movement impairments (MI). These are important elements in the classification of the pain disorder and in determining whether the habitual movements or postures are provocative or protective (O’Sullivan, 2000, 2005; O’Sullivan and Beales, 2007a,b). The process consists of several stages before reaching a classification (Fig. 1): 1. The first part involves screening; determining if the condition is specific LBP or NSLBP (O’Sullivan, 2005). 2. The second stage considers whether specific LBP disorders have an adaptive or maladaptive response to the disorder (O’Sullivan, 2005). If the disorder is classified as non-specific, then consideration of whether the disorder is predominantly centrally or peripherally mediated is made. The presence of localised and anatomically defined pain, associated with specific and consistent mechanical aggravating and easing factors, suggests that physical/mechanical factors are likely to dominate the disorder resulting in a peripheral nociceptive drive. Constant, non-remitting widespread pain, not influenced by mechanical factors, could on the other hand indicate inflammatory or centrally driven pain (O’Sullivan, 2005).

K. Vibe Fersum et al. / Manual Therapy 14 (2009) 555–561 Classification process adapted from Peter O’ Sullivan

Red flag disorders Cancer Infection Inflammatory disorder Fracture

Chronic back pain disorders

Specific back pain disorders

557

Non-specific back pain disorders Level 1

- Spondylolisthesis - disc herniation + radicular pain - degenerative disc + modic changes - foraminal and central stenosis

Adaptive response Patients response to disorder is adaptive / protective

Mal-adaptive Patients response to disorder is mal-adaptive

Centrally mediated back pain

Peripherally mediated back pain

Level 2 Dominant psychosocial factors

Nondominant psychosocial factors

Pelvic girdle pain

Low Back Pain

Level 3

Reduced force closure

Excessive force closure

Control impairment (directional subgroups)

Movement impairment (directional subgroups)

Level 4

Directional subgroups (+ level of dysfunction)

Directional subgroups (+ level of dysfunction)

Level 5

+/- central pain modulation based on contribution of psycho-social factors

Level 6

Management Advise, medical, surgical – as appropriate

Management - Cognitive / Motor learning - Medical

- Multidisciplinary management Psychological (CBT), medical, functional rehabilitation

- Medical management - Functional rehabilitation

- Motor learning within cognitive framework (enhance force closure) - Functional restoration

- Motor learning within cognitive framework (reduce force closure/ relaxation) - Functional restoration

- Motor learning within cognitive framework (enhance control) - Functional restoration

Fig. 1. Classification process adapted from Peter O’Sullivan (O’Sullivan, 2005; O’Sullivan and Beales, 2007a,b).

3. Centrally mediated pain can then be further sub-classified into the presence of non-dominant or dominant psychosocial factors. Peripherally mediated disorders are sub-classified into either LBP or a pelvic girdle pain disorders. 4. Peripherally mediated lumbar spine pain disorders are divided into MI or MCI disorders and peripherally mediated pelvic

girdle pain into excessive or deficit of force closure. Both these classifications have been described in detail elsewhere (O’Sullivan, 2005; O’Sullivan and Beales, 2007a,b). 5. If the lumbar spine is the source of pain, the primary directional provocation bias as well as the symptomatic spinal level is noted.

558

K. Vibe Fersum et al. / Manual Therapy 14 (2009) 555–561

Table 1 Inclusion/exclusion criteria. Inclusion criteria

Exclusion criteria

Patients with non-specific LBP (NSLBP) (6 weeks) Male or female Age between 18 and 65 years

Sick-listed for more than 4 months continuous duration during last year Acute exacerbation of LBP Radicular pain. Positive neural tissue provocation tests (primary peripheral symptoms) Any low limb surgery on the last 3 months

Localised LBP: primarily in the area from T12 to gluteal folds Moderate ongoing LBP, VAS > 2/10 Surgery involving the lumbar spine (fusion) and Oswestry > 14% Mechanical provocation of pain: Pregnancy postures, movement and activities Psychiatric disorders Widespread non-specific pain disorder (no primary LBP focus) Specific diagnoses: active rheumatologic disease, progressive neurological disease, serious cardiac or other internal medical disease

6. The final decision is to indicate if significant psychosocial factors are associated with the disorder, based on all information from the examination process. The evaluation of psychosocial factors considers the presence of underlying fear avoidance behaviour, as well as psychological and social drivers considered to contribute to the pain disorder. Within this reasoning process, consideration is given to whether the patient has adapted in a positive (confrontation, active coping and minimal avoidance behaviours) or negative manner (passive coping and fear avoidance). Each testing took about 1 h. The patient was examined independently twice on two days, within a 1-week period. Each therapist filled out a classification form (see Supplementary Appendix A.1) and put it in a sealed opaque envelope after their patient assessment. After examination the patient completed several questionnaires to formally assess their disorder. This included a pain drawing, a functional assessment chart from the Dartmouth Primary Care Cooperative Information Project (COOP/WONCA), Oswestry Disability Index (ODI), Hopkins Symptoms Check List (HSCL), Fear Avoidance Beliefs Questionnaire (FABQ) and Ørebro Musculoskeletal Pain Screening Questionnaire (Ørebro MSPSQ). 2.5. Analysis After completed examinations, the results were compared and logged. The developer’s classification of each patient was used as the gold standard to which the other results were compared. Kappa coefficients and percentage of agreement were calculated using SPSS 13.0 for Windows. Cohen’s Kappa statistic was used to calculate inter-tester reliability and Landis and Koch’s (1977) values for interpretation of the reliability scores were used. Kappa values <0.20 indicate poor agreement, 0.21–0.40 fair, 0.41–0.60 moderate,

Table 2 Patients’ characteristics. Number of patients Female Male Mean age (years) Mean pain intensity Mean duration (years) Mean Oswestry Mean HSCL Mean Ørebro score

26 11 15 32.4 6/10 4.9 21.2/100 1.53/4 87.5/210

0.61–0.80 substantial, and 0.81–1.00 indicate almost perfect agreement. The data was analysed based on agreement of overall classification (specific LBP vs NSLBP), centrally or peripherally mediated, adaptive or maladaptive movement disorders, and whether it was considered to be a pelvic girdle pain or LBP disorder. Kappa agreement of the primary directional pain provocation, the spinal level of pain provocation and the presence of psychosocial influence on their LBP disorder was calculated. 3. Results In the first part of the classification process, all patients were classified with NSLBP with 98% agreement for this level. All patients in the study had pain arising from a peripheral pain source, with 99% agreement for this. One patient was classified by all four testers as having pelvic girdle pain (100% agreement); the rest were classified as LBP disorders (99% agreement). The fourth level considered increased or decreased force closure for pelvic pain (one patient, 100% agreement), MCI (24 patients, 99% agreement) or MI (one patient, 75% agreement) for low back. In the fifth level, Kappa agreement could be calculated, deciding the directional pattern of provocation (Fig. 2). For the primary direction of provocation, Kappa (K) and percentage agreement had a mean between the four testers of 0.82 (range 0.66–0.90) and 86% (range 73–92%) respectively. Increased familiarity with the system also increased the reliability results (<100 h K ¼ 0.66, >100 h K ¼ 0.90). In the final level of decision making, the mean Kappa coefficient for detecting psychosocial influence was 0.65 (range 0.57–0.74) and the mean agreement 87% (range 85–92%). 4. Discussion The principal finding of our study suggests that therapists with substantial training in this CS (O’Sullivan, 2005) demonstrated fair to excellent agreement (Landis and Koch, 1977) in primary classification of the disorder, identification of directional patterns of provocation and the presence of psychosocial factors associated with the disorder, when applied to a range of NSLBP patients. Our findings are in accordance with a recent study (Dankaerts et al., 2006b), who also found moderate to excellent agreement between testers examining patients within the MCI subgroup. Their study consisted of two separate parts. The first part demonstrated almost perfect agreement between two expert clinicians when classifying 35 patients with MCI identified from a clinical case load, into the various directional patterns (K ¼ 0.96, agreement 97%). In the second part, 25 out of 35 patients with MCI in the first study were randomly selected. These were videotaped and classified into directional groups by 13 other therapists based on the video and subjective complaints of the patients. The agreement between the 13 different raters was moderate to excellent (mean Kappa 0.61, agreement 70%). As in Dankaerts et al.’s study (2006b), familiarity with the CS also influenced the reliability results, demonstrating higher agreement among raters with more CS training. These findings are in line with Strender’s study (1997), concluding that reliability of clinical tests requires sufficient time for examination and conformity of performance, definitions and evaluations. The protocol of our study followed a similar examination procedure as the first part of Dankaerts et al.’s (2006b) study. By including any patient with localised low back pain in our study, we anticipated a more heterogenic NSLBP population with the inclusion of patients with back pain associated with MI as well as pelvic girdle pain disorders. However, 24 out of the 26 patients were classified as having MCI, which is in line with the findings of Dankaerts et al. (2006b). Furthermore, the current study involved four therapists examining the patients versus two in the first part of Dankaerts study (2006b).

en t

0

irm

Correct Incorrect m

tI

en

t

c

lia

0 en

irm

pa

m

tI

en

em

l

ct io

si o

n na l Sa cr oi lia c M Im ov pa em irm e en nt t

ire

Lateral shift n= 8

t

Correct Incorrect

en

Sacroiliac n=4

rm

c

lia

oi

cr

Sa

tid

ul

M

n

0

pa i

ov

M

0

oi

l

0

cr

n na

io

ct

ire

tid

ul

Ex te n

si o

ft

sh i

0

Sa

0

na

n

io

100

Pa ss iv e

Ex te n

al

n

0

io

ct

ire

M

n

io ns

te

Ex

Ac tiv e

te r

12,5

em

ov

M

tid

ul

io

e

iv

ft

hi ns

te

Ex

lS

0

M

ns

n

io

ss

Pa

e

tiv

Ac

La

ex io

Fl

75

te

Ex

ns

te

Ex

ft

hi

0

iv e

e

0

Pa ss

tiv

Ac

ra

te

La

n

io

ex

Fl

100 90 80 70 60 50 40 30 20 10 0

pa

100 90 80 70 60 50 40 30 20 10 0

lS

ra

La te

n

io

ex

Fl

Classification per pattern (%) 100 90 80 70 60 50 40 30 20 10 0

m

tI

oi lia c

Movement impairment n=4

en

0

na l

0

Sa cr

4,5

io

Classification per pattern (%)

25

em

ov

M

Correct Incorrect

Classification per pattern (%)

50

ct

on

ns i

Correct Incorrect

tid ire

0

te

0

Ex

t

4,2

ul

e

n

io

en

irm

c

n

Correct Incorrect

M

iv

0

ns

2,3

ss

2,3

t

75

en

Multidirectional n=44

te

pa

ct io

si o

na l Sa cr oi lia c M Im ov pa em irm e en nt t

ire

n

0

Pa

ift

100 90 80 70 60 50 40 30 20 10 0 m

tI

en

lia

oi

cr

Sa

Active extension n=8

Ex

c

em

ov

M

0

m pa irm

lia

oi

cr

Sa

tid

ul

Ex te n

si o

ft

sh i

8,3

e

tI

en

m

ov e

M

0

Sh

io n

l

na

M

Pa ss iv e

Ex te n

al

n

4,2

ra l

ex

l

4,5

na

io

n

25

io

ct

ire

n

io

io

ns

te

Ex

tid

ul

M

e

iv

ss

Ac tiv e

te r

4,2

ire ct

n

0

io

11,4

ns

n

ft

hi

ns

te

Ex

La

0

te

Ex

Pa

e

tiv

lS

ra

te

ex io

Fl

Flexion pattern n =24

tiv

Ac

La te

Fl

Classification per pattern (%)

iv e

io

ft

hi

ns

te

Ex

Ac

n

io

ex

Fl

79,1

M ul tid

Pa ss

e

lS

La

100 90 80 70 60 50 40 30 20 10 0

tiv

Ac

n

io

ex

Fl

Classification per pattern (%) 100 90 80 70 60 50 40 30 20 10 0

te ra

La

Classification per pattern (%) 100 90 80 70 60 50 40 30 20 10 0

Classification per pattern (%)

K. Vibe Fersum et al. / Manual Therapy 14 (2009) 555–561 559

Correct Incorrect

12,5 0 0

Passive extension n=12 Correct Incorrect

0 0

75

25 0 0

75

25

0

Fig. 2. Classification per different pattern (in %) by all examiners; n ¼ total number of that specific pattern included  4 (total number of examiners).

560

K. Vibe Fersum et al. / Manual Therapy 14 (2009) 555–561

This may explain the greater reliability in this aspect of Dankaerts study, in comparison to ours. With regards to the second part of Dankaerts et al.’s study (2006b), it was acknowledged that the use of previously collected information (both subjective and video) represented a bias for the 13 clinicians. In our study, the testers did not have any prior information regarding the patient’s disorder as this could influence the classification reliability, as different raters may gather information from patients in different ways. Eight subjects in our study out of 26 with disorders classified as peripherally mediated NSLBP were also identified as having moderate, but significant psychosocial factors contributing to their disorders. Analysis of the questionnaire data collected after all assessments, confirmed that these eight patients scored significantly higher on HSCL and Ørebro MSPSQ (p < 0.05). Linton and Hallden (1998) identified potential psychosocial risk factors associated with future sick absenteeism in a study, using the Ørebro as the screening instrument. High total scores were related to outcome and to cut-off points that correctly identified the prognosis of nearly 80% of the patients. Psychosocial factors can modulate pain behaviour, which then can increase disability via fear avoidance, as well as promoting pain levels via central mechanisms (Vlaeyen and Linton, 2000). However there is little evidence to date that physiotherapists can identify these subjects at risk, based on subjective examination. It has been emphasised (Dankaerts et al., 2006b) that the development of a multi-dimensional mechanism-based CS based on a bio-psycho-social framework should be seen as a critical development of a CS. The Quebeck Task Force has been considered by many as the first CS that included biomedical, psychological and social considerations in the classification process (McCarthy et al., 2004). The system used in our study, developed by O’Sullivan, utilises the Quebeck Task Force as an underlying framework, by classifying specific LBP versus NSLBP, the stage of the disorder, and the presence of red and dominant yellow flags. However, patients are sub-classified further, identifying the primary direction of provocation and the proposed underlying mechanism of the disorder. Furthermore, very specific interventions are indicated for the different classifications (O’Sullivan, 2005; O’Sullivan and Beales, 2007a,b). In contrast, the McKenzie CS is a bio-system that lacks validity within a chronic LBP population, as only about 40% of patients have a directional pain preference (Donelson et al., 1990). Consistent with our findings, 45% of the subjects were classified as having MCI with multi-directional pain provocation, suggesting that a unidirectional preference was not present. This lack of uni-directional preference limits the use of directional treatment methods as advocated by McKenzie. Interestingly, 25 of the patients in our study had MCI, and only one had MI. This finding is consistent with reports that impairments of range of motion are often not present in chronic low back pain disorders (Nattrass et al., 1999). However the lack of subjects with MI disorders in this study limits the ability to confirm the reliability of physiotherapists when identifying this subgroup. The Sahrmann CS for NSLBP proposes a single mechanism for LBP (movement dysfunction), but does not consider specific diagnosis of LBP, CNS mediated pain, MIs or psychosocial factors, limiting its application within a chronic LBP setting. Petersen et al. (2004) in contrast proposed a system that demonstrated substantial reliability, but it lacked clear guidelines for management. Reliability can be influenced by many different factors. The participants seemed representative of the population normally seen in primary health care, but the small sample may not be representative of the chronic LBP population. The first part of the classification process in this study was to determine whether the patient’s condition was specific or non-specific. Secondly, an assessment was made to classify the source of the underlying

mechanism as being centrally or peripherally driven. Our study’s inclusion criteria were aimed at subjects with localised NSLBP that was mechanically provoked, making it more likely that they had a peripheral pain disorder. None of our subjects were classified with neurogenic pain. This fits with Bogduk’s study (1995), which concluded that most NSLBP disorders are peripherally mediated, having a pain source that most likely is discogenic or from the facet joint and less commonly from the sacroiliac joint. It can be argued that the Kappa scores could have been higher if all the testing procedures had been standardised. However, the study’s intention was to evaluate the reliability as a result of the whole examination as performed in clinical practice, and standardising the examination for this heterogenic group of patients could have influenced the validity. 5. Conclusion The findings provide evidence that the inter-tester reliability of O’Sullivan’s CS is substantial for a range of patients within the NSLBP population in line with previous research. Using a mechanism-based CS has implications in terms of treatment being directed towards identified subgroups. The use of the CS is currently being evaluated in a randomised controlled trial in order to compare the efficacy of different interventions for any given category. Appendix A. Supplemental material Supplementary information for this manuscript can be downloaded at doi: 10.1016/j.math.2008.08.003. References Bogduk N. The anatomical basis for spinal pain syndromes. Journal of Manipulative and Physiological Therapeutics 1995;18(9):603–5. Cherkin D, Deyo R, Battie M, Street J, Barlow W. A comparison of physical therapy, chiropractic manipulation, and provision of an educational booklet for the treatment of patients with low back pain. New England Journal of Medicine 1998;339:1021–9. 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 2006a;31(6):698–704. Dankaerts W, O’Sullivan PB, Straker LM, Burnett AF, Skouen JS. The inter-examiner reliability of a classification method for non-specific chronic low back pain patients with motor control impairment. Manual Therapy 2006b;11(1):28–39. Delitto A, Erhard RE, Bowling RW. A treatment-based classification approach to low back syndrome: identifying and staging patients for conservative treatment. Physical Therapy 1995;75(6):470–85. Donelson R, Silva G, Murphy K. Centralization phenomenon. Its usefulness in evaluating and treating referred pain. Spine 1990;15(3):211–3. Ferreira ML, Ferreira PH, Latimer J, Herbert RD, Hodges PW, Jennings MD, et al. Comparison of general exercise, motor control exercise and spinal manipulative therapy for chronic low back pain: a randomized trial. Pain 2007;131(1-2):31–7. Ford J, Story I, McKeenen J. A systematic review on methodology of classification system research for low back pain. Musculoskeletal Physiotherapy Australia 13th Biennial Conference, Sydney, Australia, 2003. Fritz JM, Delitto A, Erhard RE. Comparison of classification-based physical therapy with therapy based on clinical practice guidelines for patients with acute low back pain – a randomized clinical trial. Spine 2003;28(13):1363–71. Frymoyer J, Rosen J, Clements J, Pope M. Psychological factors in low back pain disability. Clinical Orthopaedics and Related Research 1985;May;(195):178–84. Kilpikoski S, Airaksinen O, Kankaanpaa M, Leminien P, Viderman T, Alen M. Interexaminer reliability of low back pain assessment using the Mckenzie method. Spine 2002;27(8):207–14. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33(1):159–74. Linton SJ. A review of psychological risk factors in back and neck pain. Spine 2000;25(9):1148–56. Linton SJ, Hallden K. Can we screen for problematic back pain? A screening questionnaire for predicting outcome in acute and subacute back pain. Clinical Journal of Pain 1998;14(3):209–15. Main C, Watson P. Guarded movements: development of chronicity. Journal of Musculoskeletal Pain 1996;4(4):163–70. Maluf KS, Sahrmann SA, Van Dillen LR. Use of a classification system to guide nonsurgical management of a patient with chronic low back pain. Physical Therapy 2000;80(11):1097–111.

K. Vibe Fersum et al. / Manual Therapy 14 (2009) 555–561 McCarthy C, Arnall F, Strimpakos N, Freemont A, Oldham J. The biopsychosocial classification of non-specific low back pain: a systematic review. Physical Therapy Reviews 2004;9:17–30. McKenzie R. The lumbar spine, mechanical diagnosis and treatment. Waikanae, New Zealand: Spinal Publications Ltd; 1981. Nachemson A. Back pain; delimiting the problem in the next millennium. International Journal of Law Psychiatry 1999;22(5-6):473–80. Nattrass CL, Nitsche JE, Disler PB, Chou MJ, Ooi KT. Lumbar spine range of motion as a measure of physical and functional impairment: an investigation of validity. Clinical Rehabilitation 1999;13:211–8. O’Sullivan PB. Evaluation of specific stabilizing exercise in the treatment of chronic low back pain with radiologic diagnosis of spondylolysis or spondylolisthesis. Spine 1997;22(24):2959–67. O’Sullivan PB. Lumbar segmental ’instability’: clinical presentation and specific stabilizing exercise management. Manual Therapy 2000;5(1):2–12. O’Sullivan PB. Lumbar repositioning deficit in a specific low back pain population. Spine 2003;28(10):1074–9. O’Sullivan P. Diagnosis and classification of chronic low back pain disorders: maladaptive movement and motor control impairments as underlying mechanism. Manual Therapy 2005;10(4):242–55. O’Sullivan PB, Beales DJ. Diagnosis and classification of pelvic girdle pain disorders – Part 1: a mechanism based approach within a biopsychosocial framework. Manual Therapy 2007a;12(2):86–97. O’Sullivan PB, Beales DJ. Diagnosis and classification of pelvic girdle pain disorders – Part 2: illustration of the utility of a classification system via case studies. Manual Therapy 2007b;12(2):1–12. O’Sullivan P, Twomey L, Allison G, Sinclair J, Miller K, Knox J. Altered patterns of abdominal muscle activation in patients with chronic back pain. Australian Journal of Physiotherapy 1997;43(2):91–8. Padfield B, Chesworth B, Butler R. Use of an outcome measurement system to answer a clinical question: is the Quebec task force classification system useful in an outpatient setting? Physiotherapy Canada 2002:254–60. Petersen T, Kryger P, Ekdahl C, Olsen S, Jacobsen S. The effect of McKenzie therapy as compared with that of intensive strengthening training for the treatment of

561

patients with subacute or chronic low back pain: a randomized controlled trial. Spine 2002;27(16):1702–9. Petersen T, Olsen S, Laslett M, Thorsen H, Manniche C, Ekdahl C, et al. Intertester reliability of a new diagnostic classification system for patients with non-specific low back pain. Australian Journal of Physiotherapy 2004;50(2): 85–94. Petersen T, Thorsen H, Manniche C, Ekhdahl C. Classification of non-specific low back pain: a review of the literature on classification systems relevant to physiotherapy. Physical Therapy Reviews 1999;4:265–81. Sahrmann SA. Diagnosis and treatment of movement impairment syndromes. Mosby: St Louis; 2001. Skouen JS, Grasdal AL, Haldorsen EM, Ursin H. Relative cost-effectiveness of extensive and light multidisciplinary treatment programs versus treatment as usual for patients with chronic low back pain on long-term sick leave: randomized controlled study. Spine 2002;27(9):901–9. Spitzer WO. Scientific approach to the assessment and management of activityrelated spinal disorders. Spine 1987;7S:S1–55. Strender LE, Sjoblom A, Sundell K, Ludwig R, Taube A. Interexaminer reliability in physical examination of patients with low back pain. Spine 1997;22(7): 814–20. Stuge B, Laerum E, Kirkesola G, Vollestad N. The efficacy of a treatment program focusing on specific stabilizing exercises for pelvic girdle pain after pregnancy: a randomized controlled trial. Spine 2004;29(4):351–9. Van Dillen LR, Sahrmann SA, Norton BJ, Caldwell CA, Fleming DA, McDonnell MK, et al. Reliability of physical examination items used for classification of patients with low back pain. Physical Therapy 1998;78(9):979–88. Van Dillen LR, Sahrmann SA, Norton BJ, Caldwell CA, McDonnell MK, Bloom NJ. Movement system impairment-based categories for low back pain: stage 1 validation. Journal of Orthopaedic and Sports Physical Therapy 2003;33(3): 126–42. Vlaeyen JW, Linton SJ. Fear-avoidance and its consequences in chronic musculoskeletal pain: a state of the art. Pain 2000;85(3):317–32. Waddell G. The back pain revolution. 2nd ed. Edinburgh: Churchill Livingstone; 2004.

Manual Therapy 14 (2009) 562–566

Contents lists available at ScienceDirect

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

Original Article

Trunk muscle response to various protocols of lumbar traction Jacek Cholewicki a, b, c, *, Angela S. Lee a, N. Peter Reeves a, b, Elizabeth A. Calle b a

Department of Orthopaedics and Rehabilitation, Yale University School of Medicine, 333 Cedar St., New Haven, CT 06520-8071, USA Department of Biomedical Engineering, Yale University, New Haven, CT 06520, USA c Michigan State University Center for Orthopedic Research, Ingham Regional Orthopedic Hospital, 2727 S. Pennsylvania Avenue, Lansing, MI 48910, USA b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 3 October 2007 Received in revised form 21 July 2008 Accepted 3 August 2008

The purpose of this study was to compare trunk muscle activity, spinal decompression force, and trunk flexibility resulting from various protocols of spinal traction. Four experiments explored the effects of (1) sinusoidal, triangular, square, and continuous distraction-force waveforms, (2) 0, 10, 20, and 30 of pull angle, (3) superimposed low, medium and high frequency force oscillations, and (4) sham traction. Nineteen healthy subjects volunteered for this study. Surface EMG was recorded during traction and later used in a biomechanical model to estimate spine decompression force. Trunk flexibility was measured before and after each treatment. There were no differences in muscle activity between any of the experimental conditions except the thoracic erector spinae muscle, which had lower EMG during continuous compared to sinusoidal distraction-force waveform (p ¼ 0.02). Thoracic and lumbar erector spinae muscles were significantly less active during sham than real traction (p ¼ 0.01 and p ¼ 0.04, respectively). The estimated L4–L5 spine compression force was 25 N. Trunk flexibility decreased after each experimental session (p ¼ 0.01), and there were no differences between sessions. Our results suggest that the trunk muscle activity is minimal and point toward fluid exchange in the disc as one of the key biomechanical effects of spinal traction. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Spinal loads EMG Flexibility

1. Introduction With low back pain (LBP) remaining one of the most prevalent and costly health problems in Western Society (Andersson, 1999), the search continues for an effective treatment. Because spinal surgery is expensive and not always effective, the management of LBP begins usually with a conservative approach. One such conservative approach is mechanical spinal traction. This type of treatment relies on the application of a continuous or intermittent distraction-force between the pelvis and ribcage. Over 30% of physical therapists surveyed in Ontario, Canada, used spinal traction as the preferred treatment for subacute LBP and acute LBP with sciatica (Li and Bombardier, 2001), which represents the trends in North America. Similarly, lumbar traction is frequently used in the UK despite numerous recommendations suggesting it is ineffective (Harte et al., 2003). These recommendations, based on comprehensive reviews of randomized clinical trials, state that lumbar traction cannot be recommended as a single therapy for LBP with or

* Corresponding author. Michigan State University Center for Orthopedic Research, Ingham Regional Orthopedic Hospital, 2727 S. Pennsylvania Avenue, Lansing, MI 48910, USA. Tel.: þ1 517 975 3302; fax: þ1 517 975 3305. E-mail address: [email protected] (J. Cholewicki). 1356-689X/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.08.005

without sciatica (Harte et al., 2003; Airaksinen et al., 2006; van Tulder et al., 2006a,b; Clarke et al., 2007). However, these reviews also state that the literature does not allow for a firm negative conclusion to be made due to the small number of high quality studies published. Most of the studies had too few subjects, mixed patient population, and other methodological flaws. The exact mechanism through which traction might be effective is not known. It has been suggested that spinal elongation, by increasing intervertebral space, inhibits nociceptive nerve activity, improves mobility, reduces muscle spasm, relieves nerve root compression, and lessens adhesions around the facet joints. None of these mechanisms have been supported sufficiently by empirical data (van der Heijden et al., 1995; Clarke et al., 2007). However, all of these possible mechanisms depend on adequate distractionforce being transmitted directly to lumbar segments. During traction, muscle tension and friction between the body and the support surface should be taken into account in the form of counterforces (van der Heijden et al., 1995). While the counteractive friction force can be eliminated with various technological solutions, such as a split and sliding table, the effects of trunk muscle response to lumbar traction are unknown (van der Heijden et al., 1995; Krause et al., 2000; Clarke et al., 2007). Two previous studies looked only at EMG of sacrospinalis muscles (Hood et al., 1981; Letchuman and Deusinger, 1993). Thus, relaxation of spinal muscles appears to be

J. Cholewicki et al. / Manual Therapy 14 (2009) 562–566

the most important prerequisite for spinal traction to be mechanically effective. The most recent developments in spinal traction involve new technologies that allow for varying angles of pull, varying load duty cycles; waveforms; their frequency; and concurrent application of superimposed oscillations (Shealy et al., 2005). Such a treatment, named Intervertebral Differential Dynamic (IDDÒ) therapy, claims to be more effective in treating patients with LBP than a standard traction technique (Shealy et al., 2005). However, further refinement of IDD therapy requires quantification of trunk muscle activity and the resultant spinal loads under various waveforms, angles of pull, and oscillations. Currently, no studies comparing trunk muscle response to these protocols exist. Therefore, the purpose of this study was to compare trunk muscle activity, spinal decompression force, and trunk flexibility resulting from various protocols available with the Accu-Spina device (North American Medical Corporation, Marietta, GA) used for IDD therapy. The premarket approval for this device was granted by the FDA in 2005 (510(k) #K033231). 2. Methods 2.1. Study design The entire study consisted of four separate experiments, each exploring changes in trunk muscle activity, spinal decompression force, and trunk flexibility during various treatment options available with the Accu-SPINA device (Fig. 1). Each experiment lasted between 24 and 28 min, per manufacturer’s recommendations, and contained all experimental conditions presented in random order: (1) The effects of various distraction-force waveforms (sinusoidal, triangular, square, and continuous). The angle of pull was kept at 10 . (2) The effects of various angle of pull (at 0, 10, 20, and 30 ) using sinusoidal distraction-force waveform. (3) The effect of force oscillations (low, medium and high frequency) superimposed on the square distraction-force waveform. The angle of pull was kept at 10 .

563

(4) The effects of sham traction consisting of lying supine without any distraction-force. It should be noted that it was not possible to investigate all of the independent variables in one experiment because we did not want to expose the subjects to traction longer than the recommended 30-min limit. The difference in sit-and-reach tests performed before and after each experiment served as an indicator of possible changes in the fluid content of intervertebral discs. Because in addition to hip and hamstring, this test also measures low back flexibility; and because the range of motion of the back (i.e. modified Schro¨ber test) reflects diurnal changes in disc hydration, (Wing et al., 1992), we included the sit-and-reach test as one of the outcome measures. Trunk muscle activity was monitored with surface EMG, which was later used in an EMG-assisted spine model to estimate net forces acting on the osteoligamentous spine during traction. Fifteen subjects were tested in each experiment. However, most of the subjects volunteered for more than one experiment and were thus tested multiple times on separate days. In total, 13 males and 6 females, each without a history of LBP, were recruited for all experiments. On average (standard deviation) they were 26.4(6.2) years old, 1.76(0.10) m tall, and weighed 74.3(13.3) kg. All subjects read and signed an informed consent form prior to testing. The protocol for this study was approved by Yale University’s Human Investigation Committee. 2.2. Procedures Prior to traction treatment, all subject performed three trials in a sit-and-reach flexibility test according to a standard protocol (Allen, 1988). This protocol involved sitting on the floor with straight legs braced against a box. With palms facing down, the subject reaches forward along the measuring line on the box as far as possible. The maximum reach was held for 3 s and all three trials were averaged to obtain a flexibility score. The flexibility test was repeated at the end of each traction experiment. After appropriate skin preparation, Ag–AgCl, bipolar, disposable surface EMG electrodes were placed over the following muscles on

Fig. 1. The Accu-SPINA device used in this study (North American Medical Corporation, Marietta, GA). The table was split, such that the lower body of a subject moved with the bottom part of the table on linear bearings during traction.

564

J. Cholewicki et al. / Manual Therapy 14 (2009) 562–566

the right side of the body: rectus abdominis (RA, 3 cm lateral to the umbilicus), external oblique (EO, medial to the mid auxiliary line at the level of the umbilicus), internal oblique (IO, approximately midway between the anterior superior iliac spine and symphysis pubis, above the inguinal ligament), latissimus dorsi (LD, lateral to T9 over the muscle belly), thoracic erector spinae (TE, 5 cm lateral to T9 spinous process), and lumbar erector spinae (LE, 3 cm lateral to L4 spinous process) (Cholewicki and McGill, 1996). Each pair of electrodes was spaced 3 cm center-to-center along the muscle belly. A reference electrode was placed over the 10th rib on the right side. After verifying the quality of EMG signals on an oscilloscope, subjects performed maximum isometric exertions in trunk flexion, extension, and lateral bending on an examination table against the resistance provided manually by one of the investigators. These tasks were designed to elicit maximum voluntary activation (MVA) levels from trunk muscles, for the purpose of EMG normalization (McGill, 1991). For the abdominal muscles, an exertion in a sit-up position was modified from McGill (1991) in that the subjects produced a sequence of maximal efforts in trunk flexion as well as trunk flexion with superimposed left and right torso twists. Next, subjects donned chest and pelvic harnesses and lay supine on the Accu-SPINA table (Fig. 1). The chest harness was affixed to the immovable part of the table, while the pelvic harness was attached to the motorized traction assembly. This assembly moved up or down for adjusting the angle of pull, which was verified with an inclinometer. At this point, 3 s of EMG data were recorded while subjects lay fully relaxed to obtain a baseline EMG value. The exact shapes of all force waveforms applied are presented in Fig. 2. According to the manufacturer’s recommendations, the peak force was set at half body weight plus 44.5 N (10 lb), while the low force was set at half of the peak value. Each traction experiment began with a 60 s ramp-up to the peak force followed by two cycles of a given force waveform application. The bottom part of the split table was then released to slide freely on linear bearings. This release reduced the friction between the person and the table and

allowed the distraction-force to be transmitted to the trunk. The release also marked the beginning of the treatment, which consisted of three cycles of each experimental condition applied consecutively. The EMG data and the distraction-force were recorded with the same data acquisition board on the third cycle of each condition using 1 kHz analog-to-digital conversion (A/D). A 60 s ramp-down concluded each condition. In the sham experiment, EMG data were collected every 5 min. Prior to the A/D conversion, the EMG signals were band-pass limited between 20 and 450 Hz and differentially amplified (input impedance ¼ 100 GU, CMRR > 140 dB). 2.3. Data analysis Mean absolute values of EMG signals were computed between heart beats (QRS waves) in epochs corresponding to the peaks and troughs of the force waveforms. The data were examined for normality using the Anderson–Darling test and corrected with the Box–Cox transformation prior to the statistical analyses, if they were not normally distributed. Repeated measures ANOVAs and Tukey’s post hoc tests (p  0.05) were used to evaluate differences in muscle activities. First, the comparison was made between EMG corresponding to peaks and troughs of the distraction-force. Next, EMG data corresponding to peak force were compared between all experimental conditions in the first three experiments (various waveforms, angle of pull, and oscillations). Finally, we compared the sham and real traction using the EMG collected during the last time point for the sham and the last experimental condition from experiment 1 (various waveforms). Because the data for this comparison came from different testing sessions, we normalized the EMG using the baseline EMG value obtained from the relaxed lying condition. Because these data were not normally distributed, even after the transformation, a non-parametric Kruskal–Wallis test was used. A nested repeated measures (subjects nested within each experiment) ANOVA was used to compare sit-and-reach flexibility before and after each experiment. Before and after condition served as a within-subjects factor and four experiments constituted a between-subjects factor. All analyses were performed using the Minitab statistical software (Minitab Inc., State College, PA). All data were presented as % MVA. The net decompression force transmitted to the osteoligamentous spine was computed as the difference between the sum of all trunk muscle forces and the distraction-force applied to the trunk by the Accu-SPINA device. Muscle forces were estimated based on the level of their EMG activation using the biomechanical model of a lumbar spine system. A detailed description of this model has been previously published (Cholewicki and McGill, 1996). It consists of a rigid pelvis and sacrum, five lumbar vertebrae separated by a lumped parameter disc and ligament equivalent, rigid ribcage and 90 muscle fascicles. Each muscle consists of an active contractile part, a passive parallel elastic element and a passive nonlinear tendon. Forces in all 90 muscle fascicles were calculated with the help of EMG and the cross-bridge bond distribution moment approach (Cholewicki and McGill, 1995). As in the original work, assumptions were made regarding the neural activation of deep muscles not accessible via surface EMG. Psoas and quadratus lumborum were driven with the EMG signals of their synergists (IO and LE, respectively). Left/right muscle activation symmetry was also assumed. 3. Results

Fig. 2. Various waveforms of distraction-forces applied via Accu-SPINA device (left panel). Low, medium and high frequency oscillations are presented in the right panel.

There were no differences between EMG activity corresponding to the peaks and troughs of the distraction-force in any of the six muscles tested (p > 0. 50, DF ¼ 1, F < 0.5). Therefore, only peak force EMG was used for subsequent analyses.

J. Cholewicki et al. / Manual Therapy 14 (2009) 562–566

Within the three traction experiments, no effects of angle of pull (p > 0.06, DF ¼ 3, F < 2.6) or superimposed oscillations (p > 0.36, DF ¼ 2, F < 1.1) were found in any of the six trunk muscles. With respect to waveform, however, a significantly lower EMG activity was present in the TE muscle during constant compared to a sinusoidal distraction-force waveform (ANOVA: p ¼ 0.02, DF ¼ 3, F ¼ 3.6; Tukey’s post hoc: p ¼ 0.02, T ¼ 3.0) (Fig. 3). A comparison between sham and real traction was made using EMG collected at the end of the sham traction and the EMG obtained from the last waveform tested in experiment 1, which gave similar duration of treatment in both cases. Both TE and LE were significantly less active during sham than during real traction (p ¼ 0.01, DF ¼ 3, H ¼ 6.2 and p ¼ 0.04, DF ¼ 3, H ¼ 4.1, respectively) (Table 1). To compute spine decompression force, the counter force (spine compression force) stemming from the activity of all trunk muscles was estimated with a biomechanical model. Because overall muscle activity was very low with little differences between various experimental conditions, two representative cases were considered: sham and sinusoidal traction. The input to the model consisted of the across-subjects average EMG data expressed as % MVA (Table 1). The L4–L5 spine compression force was 218 N for sham and 434 N for the sinusoidal waveform traction. Considering that on average 409 N of peak distraction-force was applied, the spine was decompressed to 25 N during the sinusoidal waveform traction. Trunk flexibility decreased after all of the four experimental sessions (main effect: p ¼ 0.01, DF ¼ 1, F ¼ 7.2). There was no significant interaction between the sessions and flexibility (p ¼ 0.90, DF ¼ 3, F ¼ 0.2), suggesting that flexibility decreased similarly after each session. On average, subjects lost 6 (SD ¼ 2) mm in their reach during a traction or sham session.

4. Discussion The main finding of this study was that the overall trunk muscle activity is very low during traction and varies very little between different protocols of applying distraction-force in healthy subjects. For example, the average overall activity during sinusoidal waveform traction was 0.65% MVA. As expected, this value is lower than 1.7% MVA reported during upright standing (Cholewicki et al., 1997), because the demands on spine stability are lower when lying as compared to standing postures. These results agree with the only two previous studies that looked at EMG activity of sacrospinalis muscle. Hood et al. (1981) found no difference in EMG in healthy subjects between lying supine on a table and applying traction.

4.0 3.5

%MVA

3.0 2.5

*

Sinusoidal Triangular Square Constant

2.0 1.5 1.0 0.5 0.0 RA

EO

IO

LD

TE

LE

Muscle Fig. 3. Comparison of trunk muscle activities (mean (SD)) during traction using various force waveforms. An asterisk indicates significant difference (p < 0.05).

565

Table 1 Average trunk muscle activity (% MVA, mean (SD)) during sham and traction. RA

EO

IO

LD

TE*

LE*

Sham 0.14 (0.32) 0.07 (0.12) 0.73 (2.41) 0.01 (0.03) 0.08 (0.31) 0.13 (0.39) Traction 0.18 (0.20) 0.29 (0.90) 0.17 (0.30) 0.27 (0.22) 1.06 (1.65) 0.84 (1.06) *, Significant difference between two conditions (p < 0.05).

Letchuman and Deusinger (1993) recorded approximately 4% MVA of EMG activity in patients with LBP during traction, but there is always a doubt whether these patients were able to produce true maximum voluntary contractions. Both of these studies recorded higher EMG during the initial traction cycle. After approximately 4–6 min, this activity returned to baseline (Hood et al., 1981; Letchuman and Deusinger, 1993). Because we pre-conditioned the subjects before data collection with a 60 s ramp-up and two cycles (4 min total), we did not find any differences in EMG activity between cycles during the subsequent treatment part. Both studies found less sacrospinalis activity during continuous traction than during intermittent traction, although these differences were not statistically significant (Hood et al., 1981; Letchuman and Deusinger, 1993). These results are again consistent with our finding of significantly lower TE activity during continuous traction compared to the traction with a sinusoidal waveform. No other differences between waveforms, angle of pull, or superimposed oscillations existed in our study. Any possible cumulative effects of EMG responses were circumvented by randomizing the order of conditions tested within each experiment. It is quite likely that patients with LBP would demonstrate different muscle response to traction and this should be the focus of a future study. Patients with LBP demonstrate trunk muscle recruitment patterns that enhance spine stiffness, including greater antagonist co-activation (van Diee¨n et al., 2003). Therefore, it is also possible that in the face of reduced demands for spine stability during traction, patients would relax their muscle co-activation to some extent. Because prolonged muscle co-activation levels exceeding 5% MVA could lead to muscle fatigue and pain, such relaxation would have a positive result and could be one of the mechanisms by which traction might relieve back pain symptoms. This mechanism was proposed earlier for lumbosacral orthoses (Cholewicki, 2004; Cholewicki et al., 2007). The estimated spine compression force was only 434 N during sinusoidal waveform traction. This compressive force was comprised of a passive elastic muscle force component and a very low active component, which some may call muscle tonus (Walsh, 1992). Combined with the peak distraction-force of 409 N, the spine was almost completely decompressed during traction. Ramos and Martin (1994) measured negative 100 mmHg pressure in a few patients’ discs during the application of approximately a 100 lb (445 N) distraction-force. Taking 1500 mm2 as a disc’s crosssectional area, this distraction-force would produce 55 mmHg in our experiment ((434 N–445 N)/0.0015 m2/133 Pa mmHg1). Therefore, both the documented muscle activity and the estimated spine decompression forces appear reasonable in our study. Despite the relatively short duration (approximately 0.5 h) of each experimental session, a significant loss in trunk flexibility occurred. This was likely due to an increase in disc hydration (Adams et al., 1990; Wing et al., 1992). Such changes increase disc height and decrease flexibility of the lumbar spine (Adams et al., 1990; Wing et al., 1992). These phenomena are well documented as diurnal changes during sleep and are considered an important mechanism for nutrient transport to the intervertebral discs (Grunhagen et al., 2006). Although there was no difference in flexibility between real and sham traction, the intermittent force application might be more advantageous for maximizing fluid exchange and nutritional

566

J. Cholewicki et al. / Manual Therapy 14 (2009) 562–566

transport. This could be another biomechanical effect of spinal traction. If differences in fluid flow exist between various distraction-force waveforms used in the Accu-SPINA device, it is possible that they could be detected with MRI modalities. The short treatment duration and rapid effects of fluid flow in our study should not be surprising, because the greatest increase in hydration of the unloaded disc takes place within the first hour of load removal (Costi et al., 2002). In summary, our results suggest that overall trunk muscle response to traction does not pose a great problem for mechanically decompressing the intervertebral disc. The significant changes in trunk flexibility point toward fluid exchange as one of the key biomechanical effects of spinal traction, but this study did not address the overall effectiveness of traction as a treatment for LBP. Acknowledgments This study was supported by a research grant from the North American Medical Corporation (Marietta, GA), the manufacturer of Accu-Spina device. References Adams MA, Dolan P, Hutton WC, Porter RW. Diurnal changes in spinal mechanics and their clinical significance. J Bone Joint Surg Br 1990;72(2):266–70. Airaksinen O, Brox JI, Cedraschi C, Hildebrandt J, Klaber-Moffett J, Kovacs F, et al. European guidelines for the management of chronic nonspecific low back pain. Chapter 4. Eur Spine J 2006;15(Suppl. 2):S192–300. Allen ME. Clinical kinesiology: measurement techniques for spinal disorders. Orthop Rev 1988;17(11):1097–104. Andersson GB. Epidemiological features of chronic low-back pain. Lancet 1999;354(9178):581–5. Cholewicki J. The effects of lumbosacral orthoses on spine stability: what changes in EMG can be expected? J Orthop Res 2004;22(5):1150–5. Cholewicki J, McGill SM. Relationship between muscle force and stiffness in the whole mammalian muscle: a simulation study. J Biomech Eng 1995;117(3): 339–42. Cholewicki J, McGill SM. Mechanical stability of the in vivo lumbar spine: implications for injury and chronic low back pain. Clin Biomech 1996;11(1):1–15. Cholewicki J, Panjabi MM, Khachatryan A. Stabilizing function of trunk flexorextensor muscles around a neutral spine posture. Spine 1997;22(19):2207–12.

Cholewicki J, Reeves NP, Everding VQ, Morrisette DC. Lumbosacral orthoses reduce trunk muscle activity in a postural control task. J Biomech 2007;40(8):1731–6. Clarke JA, van Tulder MW, Blomberg SE, de Vet HC, van der Heijden GJ, Bronfort G, et al. Traction for low-back pain with or without sciatica. Cochrane Database Syst Rev 2007;2:CD003010. Costi JJ, Hearn TC, Fazzalari NL. The effect of hydration on the stiffness of intervertebral discs in an ovine model. Clin Biomech 2002;17(6):446–55. van Diee¨n JH, Cholewicki J, Radebold A. Trunk muscle recruitment patterns in patients with low back pain enhance the stability of the lumbar spine. Spine 2003;28(8):834–41. Grunhagen T, Wilde G, Soukane DM, Shirazi-Adl SA, Urban JP. Nutrient supply and intervertebral disc metabolism. J Bone Joint Surg Am 2006;88(Suppl. 2):30–5. Harte AA, Baxter GD, Gracey JH. The efficacy of traction for back pain: a systematic review of randomized controlled trials. Arch Phys Med Rehabil 2003;84(10): 1542–53. van der Heijden GJ, Beurskens AJ, Koes BW, Assendelft WJ, de Vet HC, Bouter LM. The efficacy of traction for back and neck pain: a systematic, blinded review of randomized clinical trial methods. Phys Ther 1995;75(2):93–104. Hood JC, Hart DL, Smith HG, Davis H. Comparison of electromyographic activity in normal lumbar sacrospinalis musculature during continuous and intermittent pelvic traction. J Orthop Sports Phys Ther 1981;2(3):137–41. Krause M, Refshauge KM, Dessen M, Boland R. Lumbar spine traction: evaluation of effects and recommended application for treatment. Man Ther 2000;5(2): 72–81. Letchuman R, Deusinger RH. Comparison of sacrospinalis myoelectric activity and pain levels in patients undergoing static and intermittent lumbar traction. Spine 1993;18(10):1361–5. Li LC, Bombardier C. Physical therapy management of low back pain: an exploratory survey of therapist approaches. Phys Ther 2001;81(4):1018–28. McGill SM. Electromyographic activity of the abdominal and low back musculature during the generation of isometric and dynamic axial trunk torque: implications for lumbar mechanics. J Orthop Res 1991;9(1):91–103. Ramos G, Martin W. Effects of vertebral axial decompression on intradiscal pressure. J Neurosurg 1994;81(3):350–3. Shealy CM, Koladia N, Wesemann MM. Long-term effect analysis of IDD therapy in low back pain: a retrospective clinical pilot study. Am J Pain Manage 2005;15(3):93–7. van Tulder M, Becker A, Bekkering T, Breen A, del Real MT, Hutchinson A, et al. European guidelines for the management of acute nonspecific low back pain in primary care. Chapter 3. Eur Spine J 2006a;15(Suppl. 2):S169–91. van Tulder MW, Koes B, Malmivaara A. Outcome of non-invasive treatment modalities on back pain: an evidence-based review. Eur Spine J 2006b;15(Suppl. 1):S64–81. Walsh EG. Muscles, masses, and motion, The physiology of normality, hypotonicity, spasticity and rigidity. In: Clinics in developmental medicine no. 125. London: Mac Keith Press; 1992. Wing P, Tsang I, Gagnon F, Susak L, Gagnon R. Diurnal changes in the profile shape and range of motion of the back. Spine 1992;17(7):761–6.

Manual Therapy 14 (2009) 567–571

Contents lists available at ScienceDirect

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

Case Report

Brachial neuritis (Parsonnage–Turner syndrome) – A case study Neasa De Burca a, b, * a b

University Hospital Galway, Ireland Physiotherapy Department, University of Limerick, Limerick, Ireland

a r t i c l e i n f o Article history: Received 10 April 2008 Received in revised form 23 December 2008 Accepted 6 January 2009

1. Introduction Brachial neuritis (BN) is a condition of unknown aetiology characterised by acute onset of severe unilateral shoulder pain followed by flaccid paralysis of shoulder and parascapular muscles (Ashworth, 2007). This condition can be confused clinically with many common musculoskeletal conditions of the neck and shoulder encountered daily in physiotherapy clinics (Spillane, 1943; Parsonnage and Turner, 1948; Tsairis et al., 1972; Helms et al., 1998). Idiopathic BN is a relatively rare condition, but is the most common cause of non traumatic brachial plexopathy (Mullins et al., 2007). Incidence of this condition is thought to be approximately 1.64 cases per 100,000 people and is thought to be male predominant but ratios vary from 2:1 to11.5:1 (Miller and McDonald, 2000). Most commonly it affects people between the ages of 20–60 years (Spillane, 1943; Ashworth, 2007). The aetiology of the condition is unclear. Some studies propose a viral aetiology, while others suggest that various infections precede the onset of the condition (Misamore, 1996). Fifteen percent of cases have been reported as occurring post vaccinations. Other possible hypotheses include immunopathological inflammatory reaction precipitated by infection, surgery, or systemic illness with concurrent injury to the involved nerves (Dillin et al., 1985; Helms et al., 1998). Diagnosis is made from a careful history and physical examination and may be confirmed by clinical neurophysiology testing (Rix et al., 2006). This syndrome may be confused with other more common conditions such as C5 nerve root lesions, suprascapular nerve entrapments, rotator cuff tears, shoulder impingement

* Physiotherapy Department, University of Limerick, Limerick, Ireland. Tel.: þ353 61233773; fax: þ353 61234251. E-mail address: [email protected] 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.01.001

syndrome, calcific tendinopathy and adhesive capsulitis (Helms et al., 1998; Miller and McDonald, 2000). Investigations may include shoulder and cervical spine radiographs. MRI of the affected muscles may also be used. Electrophysiological examination is unnecessary in most cases but may be warranted where there is some ambiguity in the clinical picture, thus identifying any co-existing neurogenic conditions and avoiding any unnecessary surgical intervention (Rix et al., 2006). Prognosis is generally excellent, with Tsairis et al. (1972) reporting that, of 99 patients with brachial plexus neuropathy 89% had recovered by 3 years.

2. Clinical presentation A 52 year old female presented to the Physiotherapy Department complaining of a dull ache over the right scapula of 10 months duration. She reported waking one morning with sudden onset excruciating pain over the scapula which lasted 4 days. She described the pain as constant and burning in nature and scored it 10/10 on the visual analogue scale (VAS) (Scott and Huskisson, 1976). She denied any history of trauma or any change in activities. She denied any dizziness, double vision, dysarthria, dyphagia or drop attacks. Over the next several days the pain woke her from sleep frequently. Treatment consisted of ibuprofen and rest. The constant burning pain subsided over the following week to an intermittent dull ache which she scored as 8/10 on the VAS (see Fig. 1). At that time she noticed winging of the right scapula and weakness of the right arm into abduction. She attended her General Practitioner who carried out routine blood tests which were normal. She tried various medications including acetomenophen and ibuprofen which provided some temporary pain relief. On initial presentation, her main complaint was of weakness of her shoulder leading to difficulty with many activities of daily living.

568

N. De Burca / Manual Therapy 14 (2009) 567–571

P2 • Soreness • Occasional • 3/10 • Aggravates: right rotation, flexion of cervical spine • Eases: heat

P1 • Dull ache • 8/10 • Intermittent • Aggravates - overhead activities - driving x 20 mins - mouse x 20 mins - writing x 30 mins • Eases - holding arm by side - lying supine in bed

• º Anasthesia • º Parasthesia

Fig. 1. Body chart on initial assessment.

The patient reported neck and shoulder stiffness for 15 min in the morning. Her daily pain pattern was activity related with increased activity leading to increased pain. No night pain was reported. Past medical history included an episode of cervical spine pain 2 years previously with associated parasthesia and weakness in her right thumb and index finger. At that time she attended a physiotherapist and reported that treatment consisted of mobilisation and traction of her cervical spine along with a home exercise programme. She was referred to a neurologist who ordered an MRI of her cervical spine followed by nerve conduction studies of the median and ulnar nerves. This MRI was carried out 18 months prior to her presenting to the department and showed minor prolapse of C6/7 disc. The nerve conduction studies were carried out 11 months later (i.e. 2 months after the onset of her shoulder symptoms) and were normal. The patient reported occasional cervical spine soreness since that episode but no recurrence of the parasthesia or weakness. There was no other significant medical history. This patient worked as a Nurse Tutor. Approximately 60% of her time was spent sitting using a computer. The remaining 40% of her day was spent standing teaching. Outside of work, she led a very active lifestyle. She swam breast stroke for 45 min twice per week, walked for an hour 3 times weekly and spent 2–3 h gardening once per week. In recent times her shoulder pain had prevented her from swimming and gardening. On objective examination the patient presented with a forward head posture and anteriorly located glenohumeral joints bilaterally. She had a flattened thoracic kyphosis and reduced lumbar lordosis. There was pronounced medial border winging of her right scapula with a medially rotated inferior angle due to muscle wasting. Examination of active range of movement revealed full range in all directions on the left side. Flexion on the right side was limited to 130 by pain at end of range. Abduction was limited to 120 by weakness with pain throughout the movement. There was an altered scapulohumeral rhythm throughout the range of available flexion and abduction and on return to the starting position. The pain and limitation of range in flexion and abduction was resolved by the therapist manually assisting scapular movement during these movements. External and internal rotations were full and pain free on over pressure. Full pain free movement of both

shoulders was obtained passively. Both shoulders were cleared clinically for rotator cuff injury, instability, labral injury and impingement. Wall push up revealed increased right medial border winging compared to the left side (see Fig. 2). Upper trapezius activation was assessed using shoulder elevation with a two kilogram weight and was found to be similar bilaterally. Middle and lower trapezius activation was assessed in the prone position. Both muscles appeared to have reduced strength and endurance on the right when compared to the left side. Muscle length of the cervical spine and shoulder musculature was assessed and showed reduced length of pectoralis major bilaterally (see Table 1 for details of muscle length and strength testing). Cervical spine movement was examined and right rotation was found to be slightly limited by stiffness, but did not reproduce any symptoms. All other movements were full and pain free on over pressure. Neurological examination revealed a normal gait pattern, dermatomes, myotomes and reflexes bilaterally. Babinski and clonus were normal. Neurodynamic testing and nerve palpation were normal bilaterally. 3. Treatment Treatment was carried out in the chronic phase of the condition over a period of 5 months with a frequency of once per week over the initial 6 weeks and then once every 2–3 weeks thereafter. The primary aim of treatment was to reduce the patient’s pain and disability levels over the time required for the condition to resolve. A second aim was to prevent the development of secondary musculoskeletal dysfunction due to adaptive changes and muscle atrophy. Treatment included postural correction, ergonomic advice focusing on workspace assessment, taping to reposition the scapula (see Fig. 3) and reassurance and education about the condition. A home exercise programme was designed and progressed as appropriate. This programme included stretching of tight structures, scapular awareness work and strengthening of the scapular stabilisers (see Table 2). It was carried out on a daily basis with the patient instructed to continue exercising either to the point of fatigue, or to the point where the quality of the exercise was lost (See Figs. 4–6 below).

Fig. 2. Medial border winging with wall push up.

N. De Burca / Manual Therapy 14 (2009) 567–571

569

Table 1 Chart of muscle innervation, length and strength adapted from Kendall et al (Kendall et al., 2005). Muscle Muscle strength length grading Cervical nerves

Brachial plexus

Axillary Musculocutaneous

Radial

5

N

5

N

5

–

5

–

5

N

5

N

5

N

4

–

0

–

5

–

5 5 5 5

N N N N

5 5

– S

5

S

5

S

5 5 5

N – –

5 5 5 –

N – N –

Cervical Cervical Long Dorsal SupraUpper Thoraco- Lower Lateral C1–8 C1–4 thoracic scapular scapular subdorsal subpectoral C5–8 C4,5 C4,5,6 scapular C5–8 scapular C5–7 C4–7 C5–7 Head and C neck extensors Rectus capitus anterior and lateral Longus capitis Longus C colli Levator scapulae Scaleni C anterior, medius and posterior Sternocleidomastoid Trapezius upper, middle and lower Serratus anterior Rhomboid major and minor Supraspinatus Infraspinatus Subscapularis Latissimus dorsi Teres major Pectoralis major (clavicular) Pectoralis major (sternal) Petoralis minor Teres minor Deltoid Coracobrachialis Biceps Brachialis Triceps Anconeus

Medial Axillary Musculo- Radial pectoral C5,6 cutaneous C5–8, C6–8, C4–7 T1 5 T1

C

C

C

C

C C

C C

4. Outcome Five months after commencing treatment the patient was discharged from the physiotherapy service. On discharge she was pain free, had no medial border winging, had full pain free shoulder range of movement, good scapular control bilaterally and had returned to all her activities of daily living including gardening and swimming without any difficulty. However, she continued to experience P2 occasionally. 5. Discussion This case reflects the typical presentation of BN in the sub acute to chronic stages. The patient showed the classic features of the syndrome and had an obvious non radicular pattern to her pain and weakness. Confusion as to the cause of the symptoms could have

C C C

C C C C

C

C

C C C C C C C C

arisen in this case due to the patient’s past history of right sided cervical spine pain. After the subjective examination a list of potential hypotheses was developed including those in Table 3 above. The objective examination was prioritised in order to confirm or negate the above hypotheses. A local glenohumeral origin to the patient’s pain was negated by thoroughly examining the passive range of the joint, rotator cuff tests, instability tests, labral tests, accessory tests and impingement tests. The possibility of cervical spine pathology contributing to the patient’s symptoms was investigated by assessing available cervical range of motion. As mentioned earlier, cervical imaging studies had previously been carried out on this patient showing changes indicative of a minor C6/7 disc prolapse. It was important however that these changes were interpreted in the context of the clinical picture, as the appearance of abnormality on imaging does not automatically imply causality. As the therapist was aware of the

570

N. De Burca / Manual Therapy 14 (2009) 567–571

Fig. 5. Lower trapezius strengthening. Fig. 3. Taping to reposition scapula.

Fig. 6. Stretching sternal pectoralis major.

Fig. 4. Middle trapezius strengthening.

patient’s past history it was expected that there may have been some restriction in the available cervical range and so particular interest lay in reproducing the patient’s symptoms. Although right cervical rotation was slightly restricted it did not reproduce any symptoms.

The CNS and brachial plexus was negated as a source of the symptoms due to the normal response on neurological testing. Further confirmation was available in the form of the normal nerve conduction studies of the median and ulnar nerves. Thoracic outlet syndrome as a cause of the patient’s symptoms was negated by a negative response to the Adson, Allen and provocative elevation tests. A visceral origin to the patient’s pain was thought improbable due to the mechanical symptom response and normal blood works.

Table 2 Home exercise programme. Daily Programme Week 1–6

Progression of exercises Week 7–22

Maintenance Programme At discharge (3–4 times per week)

- Continued as per week 1–6

- Continued as per week 1–6

- Side lying elevation/depression and protraction/retraction and proprioceptive neuromuscular facilitation patterns (PNF). - 4 point kneeling elevation/depression and protraction/retraction and PNF patterns

- 4 point kneeling PNF patterns - 4 point kneeling with gym ball under pelvis - Scapular proprioceptive work in standing

- Continued as per week 1–6 - Increase time spent in neutral position

- Continued as per week 1–6 - Increase time spent in neutral position

- Prone lying middle trapezius and lower trapezius strengthening - Serratus anterior strengthening in standing (PNF)

- Maintenance gym programme consisting of middle and lower trapezius, serratus anterior, latissmus dorsi, biceps and triceps strengthening

- Increased time spent walking/on bike to 30 min - Introduced swimming (breast stroke) at week 16

- Maintenance swimming programme (breast stroke)  20 min

Stretches:

- Sternal and clavicular pectoralis major - Pectoralis minor Proprioception:

- Elevation/depression and protraction/retraction (in sitting using mirror) - Scapular clock Postural correction:

- Lumbopelvic neutral - Neutral scapula (In sitting) Scapular strengthening: - Middle and lower trapezius strengthening in sitting with theraband

Cardiovascular - Bike/walking  20 min daily

N. De Burca / Manual Therapy 14 (2009) 567–571

571

Table 3 Hypotheses post subjective examination. Local

Neurogenic

Visceral

Vascular

Acute calcific Cervical tendonpopathy radiculopathy Rotator cuff tendonopathy

Referred

Non traumatic peripheral nerve lesion e.g. spinal accessory nerve Non traumatic brachial plexus injury

Pancoast tumour

Thoracic outlet syndrome Axillary vein thrombosis

Adhesive capsulitis

Spinal cord tumour

Subacromial bursitis

Thoracic outlet syndrome

By eliminating the other possible causes of the patient’s symptoms, only the possibility of a non traumatic peripheral nerve lesion such as a neuroma or neuritis remained. True scapular winging was present, along with an altered scapulohumeral rhythm which are the hallmarks of a long thoracic nerve lesion (Rix et al., 2006). It was therefore concluded that the patient was suffering from BN affecting the long thoracic nerve. No specific treatment has been shown to be effective in the treatment of this condition (Vanpee et al., 2000). In the early stages pain medication especially opiate based medication may be helpful in reducing pain (Miller and McDonald, 2000). Some literature advocates the use of slings to immobilise the shoulder, reduce pain and prevent stretching of weakened muscles (Miller and McDonald, 2000). Physiotherapy is advocated in some of the literature to maintain passive range, strengthen and prevent complications (Miller and McDonald, 2000; Vanpee et al., 2000; Rix et al., 2006). There is little evidence available to support the use of modalities such as electrotherapy and acupuncture in the treatment of BN (McCarthy et al., 1999). Corticosteroids have not shown any effectiveness in altering the course of the condition (Tsairis et al., 1972). It is possible that secondary musculoskeletal dysfunction and adaptive responses may occur due to the slowly recovering motor deficit and so there may be a role for a management plan aimed at treating the pain, the muscle atrophy and weakness and at the prevention of the development of secondary dysfunction. These principles were adapted in the treatment of this case of BN. This patient made steady progress over the course of her treatment. By using tape to reposition the scapula in a neutral position, her pain was reduced significantly (8/10 to 3/10) within the first month. However, this pain relief only lasted as long as the tape was applied. By the end of the second month the patient reported some carryover following the removal of the tape with pain relief being maintained for 4 to 5 days. The therapist reasoned that the carryover might have in part been due to the patients improving scapular control. BN is a self limiting condition, with Tsairis et al. (1972) in their study of 99 patients reporting functional recovery in 80% of patients at 2 years and 89% of patients at 3 years. This patient presented to the department 10 months after the onset of symptoms, was treated over a period of 5 months, and appeared fully recovered (i.e. had returned to full pain free function) 15 months after the onset of

Bone tumour in younger population

Referral from - gall bladder - duodenal ulcer - spleen

symptoms. Although it is likely that recovery in this case was in part due to natural recovery with time, this patient improved more rapidly than the average. The treatment provided the patient with methods to reduce her pain, prevent postural abnormalities and prevent secondary musculoskeletal adaptations.

6. Conclusion Cervical spine and shoulder pain are complaints encountered on an everyday basis in physiotherapy clinics. It is important that as clinicians developing hypotheses, less common pathologies such as BN are not ignored. This will ensure that patients are not subjected to inappropriate and ineffective treatments by us as therapists, or worse still to unnecessary surgical or medical interventions.

References Ashworth NL. Brachial neuritis. E Medicine. http://www.emedicine.com/pmr/ topic58.htm; 2007 (Last accessed 29/06/2007). Dillin L, Hoaglund FT, Scheck M. Brachial neuritis. Journal of Bone and Joint Surgery American Edition 1985;67:878–80. Helms CA, Martinez S, Speer KP. Acute brachial neuritis: MR imaging appearance report of three cases. Radiology 1998;207:255–9. Kendall FP, McCreary EK, Provance PG, Rogers MM, Romani WA. Muscles: testing and function, with posture and pain. 5th ed. Lippincott Williams & Wilkins; 2005. McCarthy EC, Tsairis P, Warren RF. Brachial neuritis. Clinical Orthopaedics 1999: 39–43. Miller JD, Pruitt S, McDonald TJ. Acute brachial plexus neuritis: an uncommon cause of shoulder pain. American Family Physician 2000;62(9):2067–72. Misamore GW, Lehman DE. Parsonnage Turner Syndrome. Journal of Bone and Joint Surgery 1996;78-A:1405–8. Mullins GM, O’Sullivan SS, Neligan A, Daly S, Galvin RJ, Sweeney BJ, et al. Nontraumatic brachial plexopathies, clinical, radiological and neurophysiological findings from a tertiary centre. Clinical Neurology and Neurosurgery 2007;109:661–6. Parsonnage MJ, Turner JW. Neuralgic amyotrophy. The shoulder girdle syndrome. Lancet 1948;1:973–8. Rix GD, Rothman DC, Robinson A. Idiopathic neuralgic amyotrophy: an illustrative case report. Journal of Manipulative and Physiological Theraputics 2006;29(1):52–9. Scott J, Huskisson EC. Graphic representation of pain. Pain 1976;2:175–84. Spillane JD. Localised neuritis of the shoulder girdle: a report of 46 patients in the MEF. Lancet 1943;2:594–5. Tsairis P, Dyck PJ, Mulder DW. Natural history of brachial plexus neuropathy. Report on 99 patients. Archives of Neurology 1972;27:109–17. Vanpee D, Laloux P, Gillet JP, Esselinckx W. Viral brachial neuritis in emergency medicine. The Journal of Emergency Medicine 2000;18(2):177–9.

Manual Therapy 14 (2009) 572–578

Contents lists available at ScienceDirect

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

Technical and Measurement Report

The validity of Rehabilitative Ultrasound Imaging for measurement of trapezius muscle thicknessq Cliona O’Sullivan a, b, *, Jim Meaney c, Gerard Boyle c, John Gormley a, Maria Stokes d a

Department of Physiotherapy, School of Medicine, Trinity College Dublin, Ireland School of Physiotherapy and Performance Science, University College Dublin, Belfield, Dublin 4, Ireland c St. James Hospital, Dublin 8, Ireland d School of Health Professions and Rehabilitation Sciences, University of Southampton, United Kingdom b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 6 May 2008 Received in revised form 28 October 2008 Accepted 3 December 2008

The purpose of the study was to establish the validity of Rehabilitative Ultrasound Imaging (RUSI) against Magnetic Resonance Imaging (MRI) for measuring trapezius muscle thickness. Participants were asymptomatic subjects recruited from Trinity College Dublin and associated teaching hospitals. Four MRI axial slices were made through each of the C6, T1, T5 and T8 spinous processes, with the subject supine. RUSI was performed immediately after MRI at the same vertebral levels, with the subject prone. Linear measurements of trapezius muscle thickness were made off-line on both the MRI and Ultrasound scans, in three regions: lower, middle and upper trapezius. Bland and Altman limits of agreement and Pearson’s correlation coefficient were used to analyse the relationship between thickness measures taken from MRI and RUSI. Eighteen subjects (9 women) participated, (age-range 21–42 years). Results demonstrated good agreement between MRI and RUSI measurements of the lower trapezius muscle at T8 (r ¼ 0.77) and moderate agreement at T5, (r ¼ 0.62). Results were poor for the middle (T1) and upper (C6) trapezius muscles, (r ¼ 0.22 to 0.52) but may be explained by differences in both positioning and imaging planes between the 2 modalities. It was concluded that RUSI is a valid method of measuring lower trapezius muscle thickness. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Rehabilitative Ultrasound Imaging Magnetic Resonance Imaging Validity Trapezius muscle

1. Introduction Movement of the scapula on the thoracic cage is termed scapulo-thoracic motion and is necessary for optimal upper limb function (Ludewig et al., 1996). The muscles that contribute most to scapulo-thoracic stability and motion are: the upper, middle and lower trapezius muscle and serratus anterior, (Ludewig et al., 1996; Mottram, 1997; Ebaugh et al., 2005). There remains a lack of consensus in the literature about the anatomical orientation of the different portions of the trapezius muscle and its stated functions (Johnson et al., 1994). The anatomical divisions of trapezius as defined by Johnson et al. (1994) are used in this study and are as follows: The upper trapezius arises from the superior nuchal line and the ligamentum nuchae and inserts into the lateral third of the clavicle. The middle trapezius arises from spinous processes of C7 and T1 and inserts into the acromion and spine of the scapula

q This study was supported by a grant from Trinity College Dublin. * Corresponding author. Tel.: þ353 1 7166516; fax: þ353 1 7166501. E-mail address: [email protected] (C. O’Sullivan). 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2008.12.005

respectively. The lower portion of trapezius (from T2 to T10) inserts into the deltoid tubercle of the scapula. Dysfunction of the scapular muscles is common and thought to be a precursor as well as a consequence of shoulder disorders (Smith et al., 2002; Cools et al., 2003). Clinical assessment of scapular muscle function involves observation of the scapula both in static postures and during movement (Mottram, 1997; Myers et al., 2005). However studies examining the validity and reliability of such assessment are lacking (Nijs et al., 2007). Over the past decade, Rehabilitative Ultrasound Imaging (RUSI) has become increasingly popular in the field of neuromusculoskeletal medicine (Whittaker et al., 2007). The validity of using RUSI to measure muscle size has been investigated using Magnetic Resonance Imaging (MRI) and has been established for abdominal muscle thickness (Hides et al., 2006), lumbar multifidus muscle cross-sectional area (CSA) (Hides et al., 1995), cervical multifidus muscle thickness (Lee et al., 2006), infraspinatus muscle thickness (Juul-Kristensen et al., 2000) and finally, supraspinatus muscle thickness and CSA (Juul-Kristensen et al., 2000). The role of RUSI in the assessment of scapular muscle function warrants further study. The purposes of the present paper are: 1) to describe protocols for measuring lower trapezius at the level of T5

C. O’Sullivan et al. / Manual Therapy 14 (2009) 572–578

and T8, middle trapezius at T1 and upper trapezius at C6; and 2) to compare measurements of trapezius muscle thickness taken from MRI scans as the gold standard against ultrasound scans, at the four different levels, in order to investigate the validity of the RUSI technique. 2. Methods 2.1. Subjects Asymptomatic subjects were recruited from the staff and students of Trinity College Dublin and associated teaching hospitals. The inclusion criteria were: healthy subjects, aged between 18 and 50 years, with full active painfree range of motion at the neck and shoulder. The exclusion criteria were: current symptoms or a history of trauma, surgery or pain anywhere in the neck, shoulder, upper back or arms, requiring time off work and/or consultation with a health care practitioner; neurological disorders; involvement in a training program involving the scapular muscles; claustrophobia, metal implants anywhere in the body, or pregnancy. The study was approved by the Trinity College Dublin Faculty of Health Sciences Research Ethics Committee. A letter was sent to subjects GPs informing them of their involvement in the study, at least 1 week prior to participation. A consultant radiologist reviewed all MRI scans once the study was completed, and in the event of serious pathology, would have contacted the participants GP directly. 2.2. Subject preparation Prior to the scanning procedures; subjects changed into a hospital gown and the following spinous processes were identified by the lead investigator using manual palpation and marked with a kohl pencil; C6, T1, T5 and T8. A line was drawn horizontally, perpendicular to the interspinous line, across the back at each of the four points. A Vitamin E capsule, which is easily visualized on MRI, was then attached to each of the 4 spinous processes to act as landmarks and secured using a transparent Tegaderm dressing (Fig. 1). The horizontal lines allowed for accurate placement of the ultrasound transducer once the MRI scan was complete. The posterolateral border of the acromion was also palpated and marked with a kohl pencil. A line was then drawn between this mark and T1; the distance was then measured and divided into

573

thirds, the middle third later being used as the ultrasound imaging site for the belly of the middle trapezius muscle. MRI was performed on a research 3T scanner in the Trinity College Institute of Neuroscience. Subjects were positioned supine and a triangular cushion was placed underneath the knees to reduce the lumbar lordosis. The subjects’ arms were positioned by their sides with the palms facing downwards. Subjects were prewarned about the high levels of acoustic noise from the MRI scanner and were given ear-plugs. Subjects were informed that the scanning time was between 30 and 40 min and were instructed not to move. An alarm was clipped to the hospital gown and subjects were instructed on how to use it during scanning, if they required the assistance of the radiographer. For RUSI, subjects were positioned prone with the head in the midline and the shoulders supported by towels to prevent the shoulder girdle from falling into protraction. A medium sized pillow was placed underneath the subject’s abdomen to reduce the lumbar lordosis and the arms were positioned by their sides with the palms facing the ceiling. All RUSI took place in a room adjacent to the MRI suite in the Trinity Institute for Neuroscience and was performed by the lead investigator within 60 min of the MRI scan being performed. 2.3. Procedures 2.3.1. MRI MRI was performed with a 3T MRI scanner, (Philips Achieva 3T System), using a multichannel spinal coil. To ensure that images would be taken at the same location as the RUSI images, the 4 Vitamin E capsules were visualized from parallel images in the midsagittal plane. Four axial MRI slices were then made through each capsule at an interval of 0.3 mm to ensure coverage of the region of interest. T1-weighted images were obtained using a turbo spinecho sequence. The parameters were as follows: the matrix size was 320  320, TR ¼ 10.2 ms, TE ¼ 6.9 ms, the field of view was 300 mm, the slice thickness was 4 mm and the number of slices at each level was 4. 2.3.2. RUSI An Aquila Pie Data real-time ultrasound scanner (Pie Data Medical, Maastrict, The Netherlands), with an 8-MHz linear transducer (40 mm footprint) was used for all ultrasound scanning. The scanner’s accuracy was confirmed by calibration using a phantom 3 days prior to commencing the study. The right trapezius muscle was imaged first in all subjects. Two scans were performed at each site. Once a good quality image was obtained, it was frozen onscreen and saved to a compact flash card for later analysis off-line. 2.3.3. RUSI of the lower trapezius muscle Lower trapezius was imaged at 2 sites: T8 and T5. The transducer was first placed centrally and horizontally over the spinous process (O’Sullivan et al., 2007), producing a bilateral image of the medial portions of the lower trapezius muscle which resembled a butterfly; (Fig. 2). To image the muscle belly at the level of T8, the transducer was moved laterally, initially to the right, maintaining the lateral edge of the spinous process in view. To image at the level of T5, the transducer was moved laterally to the thickest part of the muscle where the echogenic muscle borders were clear and parallel.

Fig. 1. Skin markings and placement of the Vitamin E capsules.

2.3.4. RUSI of the middle trapezius muscle The middle trapezius muscle was also imaged at 2 sites at the same vertebral level (T1). The first site was the medial portion of the middle trapezius muscle. The transducer was placed centrally and horizontally over the spinous process of T1 and then moved laterally until the triangular shaped junction between the middle

574

C. O’Sullivan et al. / Manual Therapy 14 (2009) 572–578

Fig. 2. RUSI central image of the lower trapezius muscle at the level of T8.

trapezius muscle and the fascia was visualized (Fig. 3). The transducer was tilted in line with curvature of the soft tissue until the muscle borders were clearly visualized. At the second imaging site, the transducer was placed facing in an antero-inferior direction over the middle section of the line drawn earlier between the T1 spinous process and the posterolateral aspect of the acromion. Previous exploratory work on the middle trapezius muscle had shown this site to provide the best visualization of the muscle belly (O’Sullivan et al., 2008). 2.3.5. RUSI of the upper trapezius muscle One site was imaged in this region of the muscle. The transducer was placed centrally over the spinous process at the level of C6 and then tilted in line with the skin curvature so that the triangular shaped medial portion of the muscle could be identified (Fig. 4). The transducer was then moved laterally, keeping the triangular medial portion in view. 2.4. Measurements

Fig. 4. RUSI imaging site for the right upper trapezius muscle at the level of C6 demonstrating the triangular shaped myofascial junction (MFJ).

Vitamin E capsule was measured on each of these two slices and the mean of the measurements was used in the analysis. Linear measurements of trapezius muscle thickness were made from the ultrasound scans off-line using Image J software (available from http://rsb.info.nih.gov/ij/docs/index.html). Each ultrasound scan was measured twice and the mean of 4 measurements made (i.e. 2 measurements from 2 scans at each site) was used in the analysis. For both imaging modalities, the cursor was placed on the inside edge of the muscle border and the site of measurement was the same for MRI and RUSI, (Table 1). The principal investigator performed the measurements on all scans. 2.5. Data analysis The dependent measure for analysis was trapezius muscle thickness. Pearson’s Correlation Coefficient (r) and the r2 value were used to investigate the linear relationship between the two

Linear measurements from the MRI scans were made using Philips DICOM Viewer R1.2 Version 1 Level 1 SP01, (Philips Medical System Nederland BV). Of the four slices taken at each level, two adjacent slices having the clearest visualization of the hyperintense landmark capsules were selected for measurement, (Fig. 5). The thickness of the trapezius muscle adjacent to both sides of the

Fig. 3. RUSI medial imaging site for the left middle trapezius muscle at the level of T1, showing the triangular shaped myofascial junction (MFJ).

Fig. 5. MRI image for lower trapezius at the level of T5, showing clearly the Vitamin E capsule.

C. O’Sullivan et al. / Manual Therapy 14 (2009) 572–578 Table 1 Measurement sites.

Table 3 Comparison of mean muscle thickness between left and right sides obtained by ultrasound imaging.

Muscle

Spinal RUSI level

MRI

Lower trapezius

T8

As for RUSI

Lower trapezius

T5

Middle trapezius T1 (medial imaging site) Middle trapezius T1 (lateral imaging site)

Upper trapezius

C6

3 cm lateral to the lateral edge of the spinous process Thickest point along the belly of the muscle where the muscle borders were parallel 1 cm lateral to the myofascial junction Thickest point along the belly of the muscle where the muscle borders were parallel

2 cm lateral to the triangular myofascial junction at a direction that was perpendicular to the plane of the muscle belly

575

As for RUSI

Muscle

Right mean (SD)

Left mean (SD)

p-value

Lower Trapezius T8 Lower Trapezius T5 Middle Trapezius T1 (Medial) Middle Trapezius T1 (Lateral) Upper Trapezius C6

4(1.2) 5.8(1.5) 4.3(0.8) 8(1.8) 5(1.6)

3.8(0.9) 5.2(1) 4.2(1) 8(2) 4.9(1.7)

¼0.6 ¼0.2 ¼0.7 ¼0.9 ¼0.9

As for RUSI

The midpoint of the muscle belly at a direction that was perpendicular to the plane of the muscle belly As for RUSI

imaging methods (MRI and RUSI) and the extent to which muscle thickness on MRI can be explained by thickness measurements made from RUSI, respectively. Bland and Altman’s limits of agreement were also used to provide a visual representation of the degree of agreement and to allow easy identification of bias and outliers. Independent sample t-tests were used to examine muscle thickness differences between right and left sides. Statistical analyses were performed using SPSS Version12 for Windows (SPSS Inc. Chicago, Illinois).

T8 (r ¼ 0.77; p < 0.001, Table 4), suggesting a high level of agreement between RUSI and MRI for scans taken at this level. Bland and Altman plots for scans taken at the level of T8 demonstrate that the mean differences (C) are close to 0; (right side: 0.1 mm; left side: 0.2 mm). The 95% limits of agreement for the right and left sides are narrow; 1.3–1.5 mm and 1.8 to 1.4 mm respectively, demonstrating very good validity for RUSI of the lower trapezius muscle at the level of T8 (Figs. 6 and 7). The results show a moderate correlation between RUSI and MRI measurements of the lower trapezius muscle taken at T5, (r ¼ 0.62, p < 0.001, Table 4) and a fair correlation of measurements of the upper trapezius muscle taken at C6; (r ¼ 0.52, p ¼ 0.001, Table 4). In the Bland and Altman plots for scans taken at the level of T5, the mean differences (C) are further from 0 (right side: 0.8; left side: 1.8) and the 95% limits of agreement are wider (right side: 1.9 to 3.3; left side: 1.8 to 5.1), suggesting some discrepancies between assessments of muscle size made at this level with MRI and RUSI (Figs. 8 and 9). There was no correlation between measurements of muscle thickness taken from MRI and RUSI images at the level of T1 (middle trapezius). 4. Discussion

3. Results Nine females (mean age 26.7, SD  4.1, range 21–41 years) and 9 males participated (mean age 30, SD  8, range 22–42 years). All participants were right hand dominant. Table 2 outlines the muscle characteristics of each of the 5 sites explored. Analyses of left and right side muscle thicknesses demonstrate no significant difference in trapezius muscle thickness between right and left sides (Table 3). The results show a good correlation between RUSI and MRI measurements of lower trapezius muscle thickness at the level of

Table 2 Measurements of thickness of the lower trapezius, middle trapezius and upper trapezius muscles obtained by ultrasound imaging and MRI; (mean  SD). Muscle All participants thickness Right Left (mm)

Lower Trapezius (T8): 3.8(1.4) 3.4(1.1) MRI 4.0(1.2) 3.8(0.9) RUSI Lower Trapezius (T5): 6.3(1.7) 6.1(2) MRI 5.8(1.5) 5.2(1.0) RUSI Middle Trapezius (Medial) (T1): 5.7(1.4) 6.3(1.8) MRI 4.3(0.8) 4.2(1.0) RUSI

Female Right

Male Left

Right

Left

3.4(1) 3(0.8) 4.7(1.4) 3.2(0.9) 3.3(0.7) 4.7(1)

4.2(0.8) 4.3(0.7)

5.1(1) 5.1(1)

8.2(2.1) 5.8(0.9)

5.5(1) 7.9(1.3) 4.6(0.8) 6.5(1.5)

5.9(0.7) 5.9(1.2) 6.2(1.4) 7(1.6) 4.1(0.8) 3.9(0.5) 4.4(0.8) 4.5(1.3)

Middle Trapezius (Lateral) (T1): 15.4(3.7) 14.9(3.9) 14.5(4.4) 13(4) 16.2(3) 16.4(3.4) MRI 8(1.8) 8(2) 7.4(1.8) 6.9 (1.9) 8.7(1.7) 9(1.7) RUSI Upper Trapezius (C6): 6.2(2.8) 6.7(3) 5.2(2.5) 5.6(2.5) 7.2(2.9) 7.9(3.1) MRI 5.0(1.6) 4.9(1.7) 4.8(1.6) 4.5(1.8) 5.2(1.4) 5.3(1.5) RUSI

This study describes the RUSI technique for lower trapezius at T5 and T8, middle trapezius at T1 and upper trapezius at C6. As reported in anatomical studies using cadavers (Johnson et al., 1994), the axial thickness of the trapezius muscle in the present study (measured from MRI scans) increases steadily from the lower to the upper thoracic spine; (T8: 3.6 mm, T5: 6.2 mm and T1: 15.2 mm). Mean thickness for the middle trapezius muscle (T1 level) at the medial measurement site were much smaller (6 mm) than those of the lateral measurement site but it must be remembered that the medial measurements of middle trapezius were taken 1 cm lateral to the myofascial junction, where the muscle is still quite narrow. The present results demonstrate that the trapezius muscle is thickest at the level of T1 (of the 4 spinal levels investigated), when thickness measurements are made centrally at the belly of the muscle. The results for mean lower trapezius thickness on RUSI at the lower thoracic spine (3.6 mm) are consistent with those of a previous RUSI study of lower trapezius thickness (3.1 mm) in a different group of healthy subjects (O’Sullivan et al., 2007). The primary aim of this study was to investigate the validity of RUSI to measure thickness of the trapezius muscle as compared to

Table 4 Correlation between muscle thickness measurements taken from MRI and RUSI scans (Pearson’s Correlation Coefficient). Muscle

Pearson’s correlation coefficient (r)

P value

r2

Lower Trapezius T8 Lower Trapezius T5 Middle Trapezius T1 (MED) Middle Trapezius T1 (LAT) Upper Trapezius C6 (2 cm)

0.77 0.62 0.25 0.22 0.52

<0.001* <0.001* 0.16 0.25 0.001*

0.6 0.38 0.06 0.05 0.27

576

C. O’Sullivan et al. / Manual Therapy 14 (2009) 572–578

Bland and Altman Plot for scans of the Right Lower Trapezius Muscle at the level of T8

2.0

Difference (mm)

1.5 1.0 0.5 0.0 0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

-0.5 -1.0 -1.5 Mean Thickness (mm)

Fig. 6. Bland and Altman Plot showing the validity of RUSI and MRI for the right lower trapezius imaged at the level of T8. The difference in muscle thickness measured by the two modalities is plotted against the mean muscle thickness for each subject. The middle line shows the mean difference. The 95% upper and lower limits of agreement represent 2 standard deviations above and below the mean difference.

MRI. A good correlation was found between measurements of lower trapezius muscle thickness at the level of T8 (r ¼ 0.77) and a moderate correlation at the level of T5, (r ¼ 0.62) demonstrating that RUSI is a valid technique to measure the thickness of the muscle at these levels. It is probable that a more standardized site of measuring muscle thickness at the level of T5 would have resulted in higher correlation values. The site of measurement was at the thickest point, along the muscle belly laterally, where the muscle borders are parallel. The limitation here is that there is no anatomical reference with which to use when deciding on the point for measurement (e.g. at the level of T8, the lateral border of the spinous process is used). At the level of T5, it is likely that measurements of thickness taken from the 2 different modalities were made at different points along the muscle belly, resulting in differing values for muscle thickness. Using an anatomical

landmark as a reference point would have ensured a standard point of measurement. It was however not possible to use the T5 spinous process as the transducer was only 40 mm in width and in most subjects; lower trapezius was thickest at a point more laterally than this. There was no correlation between measurements of muscle thickness taken from MRI and RUSI images at the level of T1 (middle trapezius) and a fair correlation between measurements taken at the level of C6 (upper trapezius). The lack of correlation at these two sites made may be explained by differences in orientation of the images between MRI and RUSI. In order to obtain a good RUSI image it is necessary to maintain good skin contact (Whittaker et al., 2007) and therefore angle the transducer relative to the curvature of the skin, to eliminate artifact and to visualize clearly the echogenic fascial borders of the muscle. This would inevitably

Bland and Altman Plot for scans of the Left Lower Trapezius at the Level of T8 2.0 1.5

Difference (mm)

1.0 0.5 0.0 0.0

1.0

2.0

3.0

4.0

5.0

6.0

-0.5 -1.0 -1.5 -2.0

Mean Thickness (mm) Fig. 7. Bland and Altman Plot showing the validity of RUSI and MRI for the left lower trapezius imaged at the level of T8. The difference in muscle thickness between the RUSI and the MRI scans is plotted against the mean muscle thickness for each subject. The middle line shows the mean difference. The 95% upper and lower limits of agreement represent 2 standard deviations above and below the mean difference.

C. O’Sullivan et al. / Manual Therapy 14 (2009) 572–578

577

Bland and Altman Plot for scans of Right Lower Trapezius at the level of T5 3.0

Difference

2.0

1.0

0.0 0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

-1.0

-2.0

Mean Thickness (mm) Fig. 8. Bland and Altman Plot showing the validity of RUSI and MRI for the right lower trapezius imaged at the level of T5. The difference in muscle thickness between the RUSI and the MRI scans is plotted against the mean muscle thickness for each subject. The middle line shows the mean difference. The 95% upper and lower limits of agreement represent 2 standard deviations above and below the mean difference.

result in RUSI images that were not conducted in the horizontal plane and therefore different to the plane of the MRI scans, with the ultimate result that the muscle portions imaged are inherently different. Furthermore, on reviewing the raw data, it is evident that there is a greater difference in measurements taken from 2 consecutive MRI slices at the level of T1 and C6 (mean difference: 2.7 mm and 1.8 mm respectively) than at T5 and T8, (mean difference: 0.5 mm and 0.9 mm respectively), suggesting that muscle thickness changes rapidly along the length of the middle and upper portions of the trapezius muscle. This may also account for the lower correlations at these levels. A study investigating the thickness of the trapezius muscle throughout its length from MRI scans is underway. It is also possible that the difference in body postures between the two scanning techniques was influential (MRI in supine and RUSI in prone), as the position of the neck was difficult to

7.0

standardize between prone and supine and could result in muscle length changes, with consequent alterations in muscle thickness. The converse is true for the lower trapezius muscle, where it was possible to place the ultrasound transducer horizontally at the levels of T5 and T8, mimicking the horizontal plane of the MRI scans. It was also easier to standardize the position of the lower thoracic spine in both prone and supine lying. Muscle CSA evaluated using RUSI has demonstrated a high level of correlation with MRI in a number of muscles including lumbar multifidus, (Hides et al., 1995) and vastus lateralis, (Reeves et al., 2004). The large width of the trapezius muscle throughout its length relative to the width of the ultrasound transducer, (40 mm) means that it is not possible to capture the CSA of the trapezius muscle at any vertebral level, using a single image. However; it may be possible to capture CSA of the trapezius muscle using RUSI by imaging consecutive portions of the muscle using skin markers;

Bland and Altman Plot for scans for Left Lower Trapezius at the Level of T5 Series1

6.0

Difference (mm)

5.0 4.0 3.0 2.0 1.0 0.0 0.0

2.0

4.0

6.0

8.0

10.0

12.0

-1.0 -2.0

Mean Thickness (mm) Fig. 9. Bland and Altman Plot showing the validity of RUSI and MRI for the left lower trapezius imaged at the level of T5. The difference in muscle thickness between the RUSI and the MRI scans is plotted against the mean muscle thickness for each subject. The middle line shows the mean difference. The 95% upper and lower limits of agreement represent 2 standard deviations above and below the mean difference.

578

C. O’Sullivan et al. / Manual Therapy 14 (2009) 572–578

and fitting images together as outlined by Reeves et al. (2004) for the vastus lateralis muscle. This method of RUSI would have reduced clinical utility due to the increased time required to complete the procedure and problems with accuracy. Ultrasound scanner technology is advancing and some scanners have the facility to build up a cross-sectional image. However, these machines are costly and not yet widely available. Further research is necessary to determine whether evaluation of trapezius muscle CSA is superior to evaluation of muscle thickness, particularly for the middle and upper portions of the trapezius muscle. It is possible that such a method may yield higher correlation values. However, limitations such as different body postures during MRI and RUSI and tilting of the transducer relative to the curvature of the skin, as outlined above, would remain. Although MRI is widely reported to be the gold standard for muscle tissue imaging, occasionally the muscle borders were difficult to identify on MRI, especially if a subject had very little fat around the muscle tissue. Fat is hyperintense on T1 weighted MRI, allowing easier identification of the muscle borders. In subjects with little or no fat around the muscle tissues; it was difficult to determine where one muscle ended and another began (e.g. middle trapezius and the underlying supraspinatus muscle). On these occasions, the echogenic muscle borders of the trapezius muscle were easier to visualize on RUSI images, which calls to question the assumption of MRI as the gold standard for muscle tissue imaging for rehabilitation purposes. In studies of the lumbar multifidus muscle (Hides et al., 1995), and the deep abdominal muscles (Hides et al., 2006), the authors noted that there was no significant difference in muscle size between sides in cohorts of asymptomatic subjects. This is in line with the present results that show no significant difference in the thickness of the trapezius muscle between right and left sides in healthy subjects. The main limitation of the study was the variation in resting positions between the two imaging techniques, a limitation which has also been experienced by other researchers when exploring other muscle groups (Hides et al., 1995). While every effort was made to standardize both scanning positions, it is likely that variations in joint position, muscle length and soft tissue compression may have influenced muscle thickness, resulting in differing values for both scanning techniques. The results of this study add to the growing body of literature regarding the validity of using RUSI to evaluate muscle size. Clinically, patients presenting with dysfunction of the scapular muscles are difficult to rehabilitate as motor re-training of the phasic actions of these muscles can be challenging, especially in patients with poor proprioception and limited scapular control. Some studies have incorporated RUSI as a biofeedback tool in the re-training of the deep abdominal muscles (Henry and Westervelt, 2005; Kiesel et al., 2007) and there is some evidence that using ultrasound imaging to provide feedback of transversus abdominus muscle activation is superior to clinical instruction alone, in normal subjects, (Henry and Westervelt, 2005). Future studies will examine the contribution of RUSI to rehabilitation of scapular muscle dysfunction. 5. Conclusions This study demonstrates a good level of agreement between lower trapezius muscle thickness measured from images obtained by RUSI and MRI at the level of T8 and a moderate level of agreement at T5, suggesting that RUSI is a valid method of measuring

trapezius muscle thickness at these levels. The poorer levels of agreement for middle and upper trapezius require further examination. Images of the middle and upper trapezius muscle in the exact plane as used during RUSI are difficult to replicate on MRI for the reasons outlined above and therefore it may be impossible to establish validity in these muscles using MRI. In some subjects, RUSI images provided clearer muscle borders than those provided by MRI. This combined with the fact that RUSI of the trapezius muscle can be carried out within a short time-frame, is cheap and safe, suggest that it is a valid method of measuring trapezius muscle thickness in a rehabilitation setting.

Acknowledgements We thank the staff and students of Trinity College Dublin and Associated Teaching Hospitals for their participation in the study, Ms. Aoife Clarke and Ms. Mairead Marnane for assistance with subject preparation and data entry, and Mr Sojo Joseph, radiographer, Trinity College Institute of Neuroscience, Trinity College Dublin.

References Cools AM, Witvrouw EE, Declercq GA, Danneels LA, Cambier DC. Scapular recruitment patterns: trapezius muscle latency with and without impingement symptoms. American Journal of Sports Medicine 2003;31(4):542–9. Ebaugh DD, McClure PW, Karduna AR. Three-dimensional scapulothoracic motion during active and passive arm elevation. Clinical Biomechanics 2005;20:700–9. Henry SM, Westervelt KC. The use of real-time ultrasound feedback in teaching abdominal hollowing exercises to healthy subjects. The Journal of Orthopaedic and Sports Physical Therapy 2005;35(6):338–45. Hides JA, Richardson CA, Jull GA. Magnetic resonance imaging and ultrasonography of the lumbar multifidus muscle. Spine 1995;20:54–8. Hides J, Wilson S, Stanton W, McMahon S, Keto H, McMahon K, et al. An MRI investigation into the function of the transversus abdominus muscle during ‘‘drawing-in’’ of the abdominal wall. Spine 2006;31:E175–8. Johnson G, Bogduk N, Nowitzke A, House D. Anatomy and actions of the trapezius muscle. Clinical Biomechanics 1994;9:44–50. Juul-Kristensen B, Bojsen-Moller F, Holst E, Ekdahl. Comparison of muscle sizes and moment arms of two rotator cuff muscles measured by ultrasonography and magnetic resonance imaging. European Journal of Ultrasound 2000;11:161–73. Kiesel KB, Uhl T, Underwood FB, Nitz A. Rehabilitative ultrasound measurement of select trunk muscle activation during induced pain. Manual Therapy 2007;13(2): 132–8. Lee JP, Tseng WY, Shau YW, Wang CL, Wang HK, Wang SF. Measurement of segmental cervical multifidus contraction by ultrasonography in asymptomatic adults. Manual Therapy 2006;12:286–94. Ludewig PM, Cool TM, Nawoczenski DA. Three-dimensional scapular orientation and muscle activity at selected positions of humeral elevation. The Journal of Orthopaedic and Sports Physical Therapy 1996;24(2):57–65. Mottram SL. Dynamic stability of the scapula. Manual Therapy 1997;2(3):123–31. Myers JB, Laudner KG, Pasquale MR, Bradley JP, Lephart SM. Scapular position and orientation in throwing athletes. The American Journal of Sports Medicine 2005;33(2):263–71. Nijs J, Roussel N, Struyf F, Mottram S, Meeusen R. Clinical assessment of scapular positioning in patients with shoulder pain. Journal of Manipulative and Physiological Therapeutics 2007;30(1):69–75. O’Sullivan C, Bentman S, Bennett K, Stokes M. Rehabilitative ultrasound imaging of the lower trapezius muscle: technical description and reliability. The Journal of Orthopaedic and Sports Physical Therapy 2007;37(10):620–6. O’Sullivan C, Bentman S, Stokes M. Rehabilitative ultrasound imaging of the middle trapezius: an investigation of the reliability of measurement and technique. In: Proceedings of the 9th congress of the international federation of orthopaedic manipulative therapists, Rotterdam 2008. Reeves ND, Maganaris CN, Narici MV. Ultrasonographic assessment of human skeletal muscle size. European Journal of Applied Physiology 2004;91:116–8. Smith J, Kotajarvi BR, Padgett DJ, Eischen JJ. Effect of scapular protraction and retraction on isometric shoulder elevation strength. Archives of Physical Medicine and Rehabilitation 2002;83:367–70. Whittaker JL, Teyhen DS, Elliot JM, Cook K, Langevin HM, Dahl HH, et al. Rehabilitative Ultrasound Imaging: understanding the technology and its applications. Journal of Orthopaedic and Sports Physical Therapy 2007;37(8):434–49.

Manual Therapy 14 (2009) 579–582

Contents lists available at ScienceDirect

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

Professional Issue

The convex–concave rule and the lever law Jochen Schomacher ¨ snacht ZH, Switzerland Dorfstr. 24, CH-8700 Ku

a r t i c l e i n f o Article history: Received 11 January 2009 Received in revised form 23 January 2009 Accepted 23 January 2009

1. Introduction The convex–concave rule is considered an important theory during treatment decision-making (Kirby et al., 2007). According to this rule (Fig. 1) the therapist moves a bone with a convex joint surface opposite to the direction of restricted movement of the distal aspect of the bone (e.g. the head of humerus inferiorly for restricted shoulder abduction). However, a concave joint surface is mobilized in the same direction as the direction of the restricted bone movement (e.g. the tibia condyles anteriorly for restricted knee extension) (Kaltenborn, 2002: p 34). Recent studies are questioning this principle. On 3D reconstructions of helical CT data of 3 asymptomatic shoulders Baeyens et al. (2000) for example observed a posterior translation of humeral head during external rotation in 90 abduction. However, the convex–concave rule predicts an anterior glide for external rotation. This observation could also be made in 3 symptomatic shoulders with minor instability (Baeyens et al., 2001). The same method of 3D reconstructions of helical CT data was used in the analyses of pro- and supination of the forearm (Baeyens et al., 2006). It was found for example a posterior translation of the radial head during supination in the proximal radio-ulnar joint, while the convex–concave rule predicts anterior gliding of the radial head’s joint surface on the radial notch of ulna. Cattrysse et al. (2005) used an electromagnetic tracking device to study coupled motions of the acromioclavicular, glenohumeral and humero-ulnar joints on cadavers. With mathematical calculations they deduced intra-articular kinematics which were not according to the convex–concave principle. Brandt et al. (2007) found in their literature review inconsistent evidence, poor methodological quality, and heterogeneity of the studies, so that no clear

E-mail address: [email protected] 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.01.005

conclusion could be drawn on the direction of translation of the humeral head. Johnson et al. (2007) evaluated the effect of the gliding mobilization in 20 patients with adhesive capsulitis (frozen shoulder). Half of the patients were mobilized anteriorly and the other half posteriorly. They found pain alleviation in both groups, but the posterior mobilization group had better results in range of motion – although the convex–concave rule predicts anterior gliding of the humeral head during external rotation. The aim of this paper is to explain the mechanics of the convex– concave rule and then to discuss possible misinterpretations in the above mentioned studies. 2. Mechanics of the convex–concave rule During the movement of a bone around an axis (¼osteokinematics), its joint surface is doing complex movements described by arthrokinematics (Williams et al., 1989: p 478). The form of the joint surface has been considered to induce its gliding/

fixated bone

moving bone

Gliding of the joint surface (= arthrokinematics)

Movement of the distal bone (= osteokinematics) fixated bone

moving bone Direction of the gliding mobilization

Fig. 1. Convex–concave rule (Kaltenborn, 2002: p 35).

580

/ Manual Therapy 14 (2009) 579–582

FForce

F Force F Weight

F Weight

a

F Weight F Force

b

c

Fig. 2. The three types of levers. (a) lever with two arms: Weight and force are acting on both sides of the axis. Example: hip joint standing on one leg and seen in the frontal plane (scale of Pauwels). (b) lever with one arm: Weight is acting between axis and force on one side of the axis. Example: metatarsofalangeal joint standing on the forefoot seen in the sagittal plane. (c) Lever with one arm: Force is acting between axis and weight on one side of the axis. Example: flexion in the ellbow joint seen in the sagittal plane.

sliding movement: a female (¼concave) joint surface glides in the same direction as the bone movement, while a male (¼convex) surface is gliding in the opposite direction of the bone movement (MacConaill and Basmajian, 1977: p 36 and 37; Williams et al., 1989: 483). Kaltenborn (2002: p 34) has described these mechanics in terms of the convex–concave rule. The mechanical basis of the convex–concave rule is the lever law. A lever is a body (mostly a bar) which can be moved around an axis (Brockhaus, 1993). The bones of the locomotor system represent such levers which are moved by muscles or weight forces around the axis of a joint. There are levers with two arms, where weight and force are acting both sides of the axis (e.g. a seesaw; Fig. 2 a), and levers with one arm, where weight and force are acting on the same side of the axis (e.g. a wheelbarrow and a shovel; Fig. 2 b and c). Movement of a bone with a convex joint surface like the humerus represents movement of a lever with two arms (Fig. 3a). The axis is roughly in the middle of humeral head. One lever arm is the shaft of humerus moving cranially during abduction. The other lever arm is between the axis and the joint surface of humeral head and is moving caudally during abduction. When moving a bone with a concave joint surface like the scapula, the lever system is a lever with only one arm (Fig. 3b). The axis of motion remains in roughly the middle of the humeral head. The bone of scapula and its joint surface are moving in the same direction, because both are on the same side of the axis and therefore on the same lever. The convex–concave rule is a simplification of these mechanics describing the movements of the joint surfaces using their form. However, there are variations of joint surfaces where this simplification is no longer useful. For example during a cadaver study, Lazennec et al. (1994) found in 150 proximal tibiofibular joints the fibular joint surface to be plane in 40%, convex and concave in 57% and round and convex in 3%. The tibial joint surface was not always reciprocally formed, as one would expect, but in 55% plane, in 40% convex and in 5% concave. This illustrates the difficulty in applying

a

humerus moves Abd.

scapula is fixated

the convex–concave rule according to the form of the joint surfaces in this joint. The joint mechanics are determined by the position of the axis of motion and the type of lever (one or two arms). Therefore, Lazennec et al. (1994) suppose an axis in the frontal plane between the lower third and the upper two thirds of the fibula, which represents a lever with two arms. So when the fibular malleolus moves posteriorly during dorsal flexion in the ankle, the head of fibula is moving anteriorly and the other way round (Lazennec et al., 1994). 3. Discussion: movements of the centre of the humeral and radial head or of the joint surface Having the lever law as a mechanical basis, the convex–concave rule can hardly be contradicted. So why is it questioned in different studies? The explanation is a misunderstanding! Baeyens et al. (2000, 2001) described the translation of the centre of the humeral head during external rotation in 90 abduction. Fig. 4a shows a similar movement – horizontal abduction – which is easier to represent graphically. The humerus moves physiologically as a lever with two arms and therefore its joint surface is gliding anteriorly while the bone shaft is moving posteriorly – according to the convex rule. When gliding is restricted, rolling predominates (Kaltenborn, 2002: p 27). Rolling shifts the axis of motion towards the contact point of the joint surfaces. For simplification this is exaggerated in Fig. 4b. This transforms the humerus nearly into a lever with one arm. In this case the joint surface of the humerus is mostly rolling posteriorly and its anterior gliding is restricted. The centre of the head of humerus is now moving posteriorly. The observation of Baeyens et al. (2000, 2001) is, therefore, correct, but they do not describe the movement of the joint surface as does the convex– concave rule. The described shift of the axis of motion happens especially in stiff joints and it has been known for a long time (Fig. 5; Jordan,1963: p 22).

b

humerus is fixated

scapula moves Abd.

Fig. 3. The lever system applied to abduction in the shoulder joint. (a) Abduction of the shoulder with humerus moving: the convex rule as a lever with two arms. (b) Abduction of the shoulder with scapula moving: concave rule as a lever with one arm.

/ Manual Therapy 14 (2009) 579–582

a

581

b

Fig. 4. The lever system applied to horizontal abduction in the shoulder with physiologic and restricted gliding of the humeral joint surface. (a) Horizontal abduction in a horizontal section: in a physiological joint a lever with two arms exists and the humeral joint surface is gliding anteriorly according to the convex rule. (b) Shift of the axis of motion towards the contact point of the joint surfaces because of restricted gliding. Humerus is becoming a lever with one arm and the centre of head of humerus is moving posteriorly. Note the absence of gliding in this extreme example.

The same explanation applies to the study of pro- and supination of Baeyens et al. (2006). In supination the radial head rotates on its axis. This is a two arm lever system. During supination the lateral aspect of the radial head moves posteriorly, while its medial aspect – the articular surface – moves anteriorly on the radial notch of ulna. This is according to the convex–concave rule! However, rolling of the joint surface causes posterior displacement of the centre of the radial head, which was the correct observation of Baeyens et al. (2006). The simplification of the convex–concave rule describes only the gliding of the joint surface of the moving bone. It should be noted that human joints surfaces not only glide but simultaneously roll upon the opposite joint surface (Williams et al., 1989: p 483), which is never fully congruent to the other one (MacConaill and Basmajian, 1977: p 33). In the reasoning model of the convex–concave rule the axis of motion is considered stationary for simplification. However, the rolling component in human joints shifts the axis. This is responsible for the displacement of the centre of the humeral and radial head observed by Baeyens et al. (2000, 2001, 2006). Stiff joints are thought to have restricted gliding and predominant rolling between the joint surfaces (Kaltenborn, 2002: p 27). The cause of this gliding restriction is unclear and may be an increased articular pressure as a consequence of shortening of the joint capsule or increased tension in periarticular muscles. Other

causes are possible such as an altered synovial liquid or loosening of periarticular ligaments or insufficiency of periarticular muscles as in hypermobile joints. Shortening of the joint capsule and muscles is also mentioned by Brandt et al. (2007) as a factor having an influence on arthrokinematic movements. These authors confuse, like Baeyens et al. (2000, 2001, 2006), the displacement of the centre of the humeral head with the movement of the joint surface described by the convex–concave rule (Schomacher, 2008). The increased external rotation after posterior gliding mobilization in the shoulder joint (Johnson et al., 2007) might be explained by the frequently observed anterior positional fault of the humeral head in relation to the acromion (Schomacher, 2007; Bryde et al., 2005). There was no significant difference between anterior and posterior gliding mobilization regarding pain (Johnson et al., 2007). This indicates, that respecting mechanical principles is seen mainly in mechanical parameters like range of motion, while for pain relief many techniques might be done – even ignoring joint mechanics! Finally, it should be mentioned, that the convex–concave rule describes gliding in physiological joints. It is valid also in pathological ones in which the physiological gliding becomes restricted. However, the physiotherapist should not mobilize a pathological joint according to a rule, but treat pathological clinical findings, which are in correlation with the patient’s complaints.

axis of motion

starting position

adhesions

starting position

final position

final position

Fig. 5. Extension of the knee with physiological gliding and with restricted gliding due to adhesions. Note the shift of axis of motion (after Jordan 1963: 22).

582

/ Manual Therapy 14 (2009) 579–582

4. Conclusion The convex–concave rule, introduced by Kaltenborn into manual therapy, is a didactic simplification of the lever law, during rotatory movements of the joints. Movement of a convex bone corresponds to movement of a lever with two arms (¼convex rule) and movement of a concave bone to a movement of a lever with one arm (¼concave rule). The convex–concave rule describes the movement of a pair of mating joint surfaces (arthrokinematics) and not the movement of the bone e.g. the centre of humeral head (osteokinematics), which is often analyzed in biomechanical research. In practice it is important not to transfer the gliding direction from physiological joints to pathological ones for mobilization without a prior examination. Restricted gliding and the associated dysfunctions may have different causes. They must be assessed individually in an examination and the findings must be interpreted in a thorough clinical reasoning process. Acknowledgements The author likes to thank Freddy Kaltenborn for revision of the text and Ola Grimsby for the help with the English language.

References Baeyens J-P, van Roy P, Clarys JP. Intra-articular kinematics of the normal glenohumeral joint in the late preparatory phase of throwing: Kaltenborn’s rule revisited. Ergonomics 2000;43(10):1726–37. Baeyens J-P, van Roy P, de Schepper A, Declercq G, Clarijs J-P. Glenohumeral joint kinematics related to minor anterior instability of the shoulder at the end of the late preparatory phase of throwing. Clinical Biomechanics 2001;16:752–7.

Baeyens J-P, van Glabbeek F, Goossens M, Gielen J, van Roy P, Clarys J-P. In vivo 3D arthrokinematics of the proximal and distal radioulnar joints during active pronation and supination. Clinical Biomechanics 2006;21:S9–12. Brandt C, Sole G, Krause MW, Nel M. An evidence-based review on the validity of the Kaltenborn rule as applied to the glenohumeral joint. Manual Therapy 2007;12(1):3–11. Brockhaus. Der Brockhaus in fu¨nf Ba¨nden. Mannheim – Leipzig: F.A. Brockhaus; 1993. und 1994. Bryde D, Freure BJ, Jones L, Werstine M, Briffa NK. Reliability of palpation of humeral head position in asymptomatic shoulders. Manual Therapy 2005;10(3):191–7. Cattrysse E, Baeyens J-P, van Roy P, van de Wiele O, Roosens T, Clarys J-P. Intraarticular kinematics of the upper limb joints: a six degrees of freedom study of coupled motions. Ergonomics 2005;48(11–14):1657–71. Johnson AJ, Godges JJ, Zimmermann GJ, Ounanian LL. The effect of anterior versus posterior glide joint mobilization on external rotation range of motion in patients with shoulder adhesive capsulitis. Journal of Orthopaedic and Sports Physical Therapy 2007;37(3):88–99. Jordan HM. Orthopedic appliances. Springfield: Charles C. Thomas Publisher; 1963. Kaltenborn FM. Manual mobilization of the joints. In: The extremities, vol. I. Oslo (Norway): Norlis; 2002. Kirby K, Showalter C, Cook C. Assessment of the importance of glenohumeral peripheral mechanics by practicing physiotherapists. Physiotherapy Research International 2007;12(3):136–46. Lazennec JY, Besnehard J, Cabanal J. L’articulation pe´rone´o-tibiale supe´rieure, une anatomie et une physiologie mal connues: Quelques re´flexion physiologiques et the´rapeutiques. Annales de Kine´sithe´rapie 1994;21(1):1–5. MacConaill MA, Basmajian JV. Muscles and movements, a basis for human kinesiology. Huntington, New York: Robert E. Krieger Publishing Company; 1977. Schomacher J. Response to: Brandt C, Sole G, Krause MW, Nel M. An evidencebased review on the validity of the Kaltenborn rule as applied to the glenohumeral joint. Manual Therapy 2007;12(1):3–11. Manual Therapy 2008;13(1):e1–2. Schomacher J. Letter to the editor in response to: Johnson AJ, Godges JJ, Zimmermann GJ, Ounanian LL. The effect of anterior versus posterior glide joint mobilization on external rotation range of motion in patients with shoulder adhesive capsulitis. Journal of Orthopaedic and Sports Physical Therapy 2007;37(7):413 (Authors reply on pp. 414–415). Williams PL, Warwick R, Dyson M, Bannister LH. Gray’s anatomy. Edinburgh: Churchill Livingstone; 1989. pp. 476–485.

Available online at www.sciencedirect.com

Manual Therapy 14 (2009) 583 www.elsevier.com/math

Diary of events APA Conference Week, Sydney Convention Centre, Sydney, Australia 1e5 October 2009 For more information please visit www.apaconferenceweek09. asn.au

MUSCULOSKELETAL DISORDERS IN PRIMARY CARE RESEARCH CONGRESS 2010 Rotterdam 11e13 October 2010 For further information visit www.bsl.nl/primarycare

Aaron Mattes 4 Day Active Isolated Stretching & Strengthening Seminar The Renaissance Hotel, Heathrow 15e18 October 2009 First ever UK Seminar. Further details regarding the contents of the seminar can be found at www.stretchingusa.com & registration details can be found at www.stretchinggb.com. Phone: 020 8897 0377 / 07984 005366. Email: [email protected]

Janet G. Travell, MD Seminar Series, Bethesda, USA For information, contact: Myopain Seminars, 7830 Old Georgetown Road, Suite C-15, Bethesda, MD 20814-2432, USA. Tel.: þ1 301 656 0220; Fax: þ1 301 654 0333; website: www.painpoints.com/seminars/index.html E-mail: [email protected]

Congresso OMT e Does the efficacy of Manual Therapy depend on its specificity? Parma, Italy, April 10the11th April 2010. For further information visit info@fisiodynacom.it NOI International conference UK and Ireland Nottingham UK e April 15e17, 2010 Dublin IRELAND April 21e23, 2010 For further details www.noi2010.com Fax þ 3906 51882443

1356-689X/$ - see front matter doi:10.1016/S1356-689X(09)00111-8

The Diary of events will soon be moving to http://www.elsevier. com/math (click on the journal news tab). If you wish to advertise a course/conference, please contact: Karen Beeton, Associate Head of School (Professional Development), School of Health and Emergency Professions, University of Hertfordshire, College Lane, Hatfield, Herts AL10 9AB, UK. E-mail: [email protected] There is no charge for this service.

Manual Therapy 14 (2009) e5

Contents lists available at ScienceDirect

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

Letter to the Editor

Annual conference letter

Dear Editors, As a final year chiropractic student I write to you to express my optimism as I enter a manual therapy profession having recently attended the 5th Chiropractic, Osteopathy and Physiotherapy Annual Conference 2008 at the Anglo-European College of Chiropractic, Bournemouth. The conference is a UK showcase for the high quality research that undergraduates of the aforementioned manual therapy professions produce. The gathering provides the perfect opportunity for closer professional collaboration and understanding, in research and practice, towards the aim of moving the manual therapy field forward. Collaboration might be considered increasingly important given the growing evidence-based culture within healthcare, and particularly since manual therapy does not attract the same levels of research funding that other areas of healthcare enjoy. Chiropractic, Osteopathy and Physiotherapy each have a great deal to offer patients with musculoskeletal disorders and it is surely to the betterment of these patients that these professions are able to complement each other through the sharing of their knowledge bases. At the conference the day’s programme was richly varied with subject matter and included professional issues, such as general practitioner referral characteristics for low back pain patients, and clinical issues, such as comparing the effectiveness of different treatment modalities for neck pain. The conference was professionally orchestrated by the chairpersons assigned to direct each section of the day’s proceedings and a show of gratitude must go to the organisers for the smooth flow of the day. A highlight of the event has to be the presentation by the keynote speaker, Professor Gordon Waddell, a highly respected and international authority on musculoskeletal conditions. His work has had a great influence on the manual therapy professions and it was an honour to have him as our distinguished guest and deliver an inspiring presentation. How appropriate that he should present his latest

1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.04.004

collaborative work, ‘Vocational Rehabilitation: What Works, for Whom and When?’ (Waddell et al., 2008). Waddell et al. (2008) define vocational rehabilitation as ‘whatever helps someone with a health problem to stay at, return to and remain at work’. As is well known the most detrimental effect of musculoskeletal conditions is commonly on ability to work. Implicit within this definition is the scope for supporting ‘‘whatever helps’’ a patient, and varied management approaches do indeed seem to help different people. Considering this definition leads me to my final point. In his introduction to the conference, Professor Alan Breen had said he often had a recurring nightmare – the landscape was that of a battlefield with three trenches labelled ‘Chiropractic’, ‘Osteopathy’, and ‘Physiotherapy’, each filled with patients. Between the trenches was a no-man’s land, the patients were trapped. It needn’t be like this. By providing an arena for these professions to share and work together, this conference is playing a part towards the release of these patients from any potential professional trenches and towards the safe passage across any potential no-man’s land of manual therapy. That’s why I’m optimistic.

Reference Waddell G, Burton AK, Kendall NAS. Vocational rehabilitation: what works, for whom, and when? Vocation Rehabilitation Task Group – Industrial Injuries Advisory Council. London: The Stationary Office; 2008.

Jonathan Branney* Anglo-European College of Chiropractic, 13-15 Parkwood Road, Bournemouth, BH5 2DF, United Kingdom  Tel.: þ44 1202 436200. E-mail address: [email protected] 8 April 2009

Manual Therapy 14 (2009) e6

Contents lists available at ScienceDirect

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

Letter to the Editor

Reply to Thomas LC, Rivett DA, Bolton PS. Validity of the Doppler velocimeter in examination of vertebral artery blood flow and its use in pre-manipulative screening of the neck

I consider that the paper by Thomas et al. (2008) has some major flaws. The use of duplex scanning to compare with velocimetry is appropriate, as the literature supports the validity of VA duplex, and shows some uncertainty about the superiority of MRA over duplex for assessing VA stenosis (Khan et al., 2007), as well as the lesser accessibility of MRA. However, the authors have relied solely on suboccipital duplex scanning with no evidence in the literature to support its validity, unlike mid-cervical scanning (Khan et al., 2007). Hence, this is a major limitation for their comparison with velocimetry results, regardless of the repeatability of their duplex measurements. Thomas et al. found 23.28% of VAs in their sample positive, using duplex scanning according to their criteria. In my validity study (Haynes, 2000) 4 preselected participants artificially inflated the proportion of positive cases in the validity trial to ensure stability of the kappa statistic. However, if just the results for the non-preselected VAs are analysed, duplex found only 1 positive VA out of 33 arteries (3%). They observed 15.5% positive VAs involving end diastolic velocities, contrasting sharply with Theil et al.’s study (1994) that found none out of 84 VAs. Hence, Thomas et al.’s findings are disproportionate to other studies that used mid-cervical scanning, raising the question about the validity of their duplex technique. The supposed decrease in many of the blood velocities they observed could be due to artefact, induced by greater attenuation of the beam subsequent to bulging of the suboccipital portion of the trapezius and sternocleidomastoid muscles under the probe, with rotation, leading to a high proportion of false positive results. I demonstrated the velocimetry technique to the first author of their study, with supervised instruction, examining only 3 individuals. The author’s minimal training from myself warranted comment in their paper, because it may have impacted on the quality of the subsequent physiotherapists’ instruction. The average proportion of volunteers who were positive for the 3 physiotherapists was 21.2%, much higher than the 3.4% of chiropractic patients with at least one VA Doppler signal detected using a validated velocimetry technique (from data used in Haynes, 2002) or 2.6%

1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.04.006

of VAs (Haynes, 2002). I still observe this low proportion after 13 years experience. The discrepancy suggests that the physiotherapists obtained a high proportion of false positives, reflecting their possibly poor training and/or inexperience. The authors’ finding of low validity for VA velocimetry, contrasts with my validity trial (Haynes, 2000), and five earlier studies (Haynes, 2002), which used arteriography as the ‘‘gold standard’’. These studies had limitations that were avoided in my trial that was ‘‘blinded’’ and involved chiropractic patients. The agreement in results between my trial and the earlier ones revealed no evidence of bias in the early studies, and so needs to be considered. The authors’ conclusion ‘‘. use of the velocimeter as a premanipulative screening tool cannot be recommended at this time’’ lacks legitimacy due to the serious limitations of their study with aberrant results. References Haynes MJ. Vertebral arteries and cervical rotation. Doppler velocimeter and duplex results compared. Ultrasound Med Biol 2000;26:57–62. Haynes MJ. Ultrasound and biomechanical studies of human vertebral arteries. PhD thesis. The University of Western Australia; 2002. p. 120. Khan S, Cloud GC, Kerry S, Marcus HS. Imaging of vertebral artery stenosis: a systematic review. J Neurol Neurosurg Psyc 2007;78(11):1218–25. Theil H, Wallace K, Donat J, Yong-Hing K. Effect of various head and neck positions on vertebral artery flow. Clin Biomech 1994;9:105–10. Thomas LC, Rivett DA, Bolton PS. Premanipulative testing and the velocimeter. Manual Therapy 2008;13(1):29–36. Thomas LC, Rivett DA, Bolton PS. Validity of the Doppler velocimeter in examination of vertebral artery blood flow and its use in pre-manipulative screening of the neck. Manual Therapy 2009;14(5):544–9.

Michael John Haynes* High Wycombe Chiropractic Clinic, 506 Kalamunda Road, High Wycombe, WA 5007, Australia  Tel.: þ69 8 9454 4711. E-mail address: [email protected] 17 January 2009

Manual Therapy 14 (2009) e7–e8

Contents lists available at ScienceDirect

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

Letter to the Editor

Comments in response to letter to the editor

We thank Dr. Haynes for his letter. His comments raise some interesting points for discussion. He is concerned with our choice of sub-occipital duplex scanning of the vertebral artery (VA) and that it is, in his view, a major limitation of our study due to the lack of evidence of validity for scanning of this segment. This raises interesting issues. We chose to perform the evaluation of vertebral artery blood flow at the atlanto-axial segment for specific reasons. Firstly, this segment of the VA is the most tortuous region and most likely to undergo flow abnormalities on cervical spine rotation (Thiel et al., 1994; Rivett et al., 2003). Secondly, it is where most VA incidents have been reported to occur (Haldeman et al., 2002). In contrast to Dr. Haynes’ protocol which compared the velocimeter findings to the mid-cervical duplex evaluation of VA blood flow, we chose to insonate the same region of the VA on which the velocimeter was to be used. This allowed us to make a direct comparison of blood flow in the same region of the VA. It remains to be determined if duplex scanning of the atlanto-axial and midcervical regions of the VA yields comparable blood flow results or if one is more sensitive than the other. We note the reference by Dr. Haynes to the systematic review by Khan et al. (2007) concerning stenosis of the vertebral artery and the absence of studies in this review concerning validity of blood flow parameters detected by duplex scanning. Nevertheless, our results showed good repeatability of blood flow measures and this is consistent with other authors who have used duplex to examine this level (Johnson et al., 2000; Rivett et al., 2003; Mitchell and Kramschuster, 2008). We were able to detect the characteristic flow signal of the VA throughout the duplex examination, and concurrently view the VA in those subjects included in the study (see below). Dr. Haynes is concerned about the discrepancy between our data and his study (Haynes, 2000) and that of Thiel et al. (1994). We identified 29.3% of subjects in our sample with altered VA flow on cervical rotation, which is consistent with a number of studies which have looked at VA blood flow on cervical rotation (Arnetoli et al., 1989; Refshauge, 1994; Licht et al., 1998; Rivett, 2000; Mitchell and Kramschuster, 2008). It is not surprising that there will be variance between studies when there are clear differences in the design and methods including subject selection, subject posture, and scanning protocols, as will be noted between our paper and the two cited by Dr. Haynes (Thiel et al., 1994; Haynes, 2000). We believe it unlikely that in our observation that there was a decrease in blood flow on duplex scanning was due to an artefact. We excluded 2 participants where attenuation of the beam did not 1356-689X/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.math.2009.04.005

allow us to obtain an adequate signal. In the participants in whom blood flow velocity was observed to decrease the VA could still be visualised, making attenuation of the beam due to bulging of musculature, as suggested by Dr. Haynes, unlikely. Furthermore, the decrease in flow velocity detected with duplex scanning on contralateral rotation we observed is not an uncommon finding reported in other studies (Zaina et al., 2003; Mitchell and Kramschuster, 2008). Regarding the training protocol, the instruction of the physiotherapists was performed by the first author who was trained by Dr. Haynes, for which we are most grateful. We acknowledge that one instructor or training program may be more or less successful than another. The protocol adopted here was intended to have face validity by providing instruction and training such as would typically occur in the clinical skill instruction of registered practising physiotherapists in a continuing education setting. However it is important to acknowledge, as reported by Sackett et al. (1991), that the relative rarity of positive subjects, in this case on velocimeter examination, may limit practitioners’ ability to recognise abnormal findings when they do encounter them. As identified above there were design and methodological differences which in and of themselves may account for the differences identified by Dr. Haynes. In our view the definitive study remains to be done, however, based on our study we believe the conclusion as presented in our paper is appropriate. Thomas L, Rivett D, Bolton P. Premanipulative testing and the use of the velocimeter. Manual Therapy 2008;13(1):29–36. References Arnetoli G, Amadori A, Stephani P, Nuzzaci G. Sonography of vertebral arteries in De Kleyn’s position in subjects and in patients with vertebrobasilar transient ischemic attacks. Angiology 1989;40:716–20. Haldeman S, Kohlbeck FJ, McGregor M. Unpredictability of cerebrovascular ischemia associated with cervical spine manipulation therapy. Spine 2002;27(1):49–55. Haynes M. Vertebral arteries and neck rotation: Doppler velocimeter and duplex results compared. Ultrasound in Medicine and Biology 2000;26(1):57–62. Johnson C, Grant R, Dansie B, Spyropolous P. Measurement of blood flow in the vertebral artery using colour duplex Doppler ultrasound: establishment of the reliability of selected parameters. Manual Therapy 2000;5(1):21–9. Khan S, Cloud GC, Kerry S, Markus HS. Imaging of vertebral artery stenosis: a systematic review. Journal of Neurology, Neurosurgery and Psychiatry 2007; 78:1218–25. Licht PB, Christensen HW, Hojgaard P, Hiolund-Carlson PF. Triplex ultrasound of vertebral artery flow during cervical rotation. Journal of Manipulative and Physiological Therapeutics 1998;21:27–31. Mitchell J, Kramschuster K. Real-time ultrasound measurements of changes in suboccipital vertebral artery diameter and blood flow associated with cervical spine rotation. Physiotherapy Research International 2008;13(4):241–54.

e8

Letter to the Editor / Manual Therapy 14 (2009) e7–e8

Refshauge KM. Rotation: a valid premanipulative dizziness test? Does it predict safe manipulation? Journal of Manipulative and Physiological Therapeutics 1994; 17(1):15–9. Rivett D. Vertebral artery blood flow during pre-manipulative testing of the cervical spine. Dunedin: School of Physiotherapy, University of Otago; 2000. p. 321. Rivett D, Sharples K, Milburn P. Reliability of ultrasonographic measurement of vertebral artery blood flow. New Zealand Journal of Physiotherapy 2003; 31(3):119–28. Sackett DL, Haynes RB, Guyatt GH, Tugwell P. Clinical epidemiology – a basic science for clinical medicine. Boston: Little, Brown and Company; 1991. Thiel H, Wallace K, Donat J, Yong-Hing K. Effect of various head and neck positions on vertebral artery blood flow. Clinical Biomechanics 1994;9:105–10. Zaina C, Grant R, Johnson C, Dansie B, Taylor J, Spyropolous P. The effect of cervical rotation on blood flow in the contralateral artery. Manual Therapy 2003;8(2):103–9.

Lucy C. Thomas* Darren A. Rivett Philip S. Bolton The University of Newcastle, University Drive, Callaghan, Newcastle, NSW 2308, Australia  Corresponding author. Tel.: þ61 2 4921 8680; fax: þ61 2 4921 7902. E-mail address: [email protected] 30 March 2009

Manual Therapy 14 (2009) e9

Contents lists available at ScienceDirect

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

Book Review A massage therapist’s guide to lower back & pelvic pain, L. Chaitow, S. Fritz (2008). 256 pp., Price £ 22,99, ISBN: 044310218X Back pain is a problem that affects almost all of us. It therefore contributes considerably to the burden of suffering and it causes high expenditure in healthcare. What is more, we are not very successful in treating it. Numerous methods have been studied but none is totally convincing, at least not in my view. Any treatment is therefore potentially a good one as long as it generates promising results, is not associated with high costs or adverse effects and is liked by patients. Massage fulfills these latter criteria perfectly. The book by Chaitow and Fritz provides brief introductions into the nature of back and pelvic pain and into some of the background

doi:10.1016/j.math.2009.04.009

issues related to massage therapy. It then shows in its main sections the aspiring student how to do it by going through some of the main techniques. Numerous drawings and photos illustrate the book and make it easier to understand. Of course, this book cannot replace proper hands-on training but is a most useful supplement to it. I recommend it to anyone who wants to study massage therapy seriously. Edzard Ernst, MD, PhD, FRCP, FRCP Complementary Medicine, Peninsula Medical School, 25 Victoria Park Road, Exeter, United Kingdom E-mail address: [email protected] 29 April 2009

Manual Therapy 14 (2009) e10

Contents lists available at ScienceDirect

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

Book Review Functional neurology for practitioners of manual therapy, R.W. Beck. Elsevier/Churchill Livingstone (2008). 552 pp., Price £ 59.99, ISBN-13: 978-0-443-10220-2 A book such as this is long awaited. For years manual therapists have been using aspects of neurology in their practices (such as antagonist contraction to relax and assist stretching in the agonist muscle) but more recently there has been an interesting shift to incorporate other less obvious aspects of neurology. This was largely developed and pioneered by Professor Carrick, a chiropractor, but it has largely remained unavailable unless you attend his courses. As with many techniques it has its followers and its antagonists (!) it has not been subject to much in the way of peer reviewed research (as the author himself indicates) and to those of us that know very little about it, this book will be the first time that the technique has entered the wider domain, and the first time that it has come under any degree of scrutiny. The book introduces the discipline of ‘‘functional neurology’’; covers main areas of neurology and introduces specific manipulation. Each chapter begins with a case pertinent to the subject, and a series of study questions. The reader is taken through basic concepts in neurology and then into aspects of functional neurology. However, unless you know your neurology well, it is often difficult to see where one stops and the other begins. The author’s response to this is very likely to be that this is deliberate, as there is no division. However, in my opinion this is its weakness; it is very difficult to know what is generally accepted in the wider medical community as being ‘fact’ and that which remains ‘hypothetical’ and in the realms of ‘functional neurology’, as both circumstances are presented together.

doi:10.1016/j.math.2009.04.008

Because of that, for me at least, It is a near miss, as, if it were presented as ‘this is what is known’ and ‘this is what the functional paradigm suggests/has been observed/researched’ the value of the book would be far greater. I do no mean to detract from the initiative, which I welcome, but the lack of critique in this way or lack of referencing context, in these days of evidence-based practice renders it less useful than it could have so easily been. I hasten to add that I do not believe that we should only practice that which is proven, as the scope of practice would be so severely limited that it would cease to be useful and advances in practice would cease. Notwithstanding, we must strive for evidence-based practice wherever possible, and be very clear about that which is evidence-based and that which is experiential. On the whole it is well written and if you have an interest in this area then it would be a useful adjunct to the book case as it provides a resource for those interested in seeing how neurological assessment may influence manual practice – indeed it is the only current resource to attempt this. However, you would need to temper some of the content with another text; to distil well accepted fact from hypothesis. Having worked for many years on this I am quite sure that the last thing that the author would like me to say is ‘‘please make this simple change in your second edition!’’. Neil Osborne, Senior Clinical Tutor Anglo European College of Chiropractic, Bournemouth, UK E-mail address: [email protected] 21 April 2009

Manual Therapy 14 (2009) e11

Contents lists available at ScienceDirect

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

Book Review Atlas of living and surface anatomy for sports medicine, Philip Harris, Craig Ranson. Churchill Livingstone, Elsevier (2008). 186 pp., Price: £39.99, ISBN: 978-0-443-10316-2 This atlas is very concise. It is less complete than many other atlases of surface anatomy, which sometimes exaggerate many irrelevant details. In this book the authors have carefully restricted the items to those relevant for sports medicine, and sports traumatology in particular. As far as the living anatomy is concerned, many diagnostic tests and clinical problems are included, the answers of which are given at the end of the book. The objective of addressing sports traumatology is well illustrated by the fact that 7 of the 10 chapters (125 out of 174 pages) are devoted to the extremities. The choice of landmarks and tests in the chapter on the head and neck is also illustrative of this approach. The explanation and interpretation of tests is somewhat superficial and straightforward and

doi:10.1016/j.math.2009.04.007

there is no reference to the sensitivity and specificity of the tests. Fortunately the CD-ROM included with the book adequately compensates for the superficial descriptions in the text. For this CD-ROM the subject on whom the tests are illustrated is a normal uninjured subject and not a real patient. This is understandable given the ethical problems associated with using patients. Unfortunately it results in images in which signs to be observed, such as a sulcus or a restriction of amplitude are not visible. Erik Barbaix, MD, MSC Sports Medicine Department of Anatomy and Embryology, Ghent University, Ghent, Belgium E-mail address: [email protected] 21 April 2009

Manual Therapy 14 (2009) e12

Contents lists available at ScienceDirect

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

Book Review Introduction to Research in the Health Sciences, Stephen Polgar, Shane A. Thomas, fifth ed. Elsevier (2008). 334 pp., price £18.99, ISBN: 0443074291 The aims of this textbook are to introduce and explain fundamental principles of research and to demonstrate how evidence produced through research is applied to solving problems in health care. The focus on presenting basic concepts of research design, data collection, quantitative and qualitative methods of analysis, and research synthesis to heath care students allows the authors to achieve these aims. It is not the intent of this book to serve as an advanced research or statistics reference. The book contains 24 well written chapters within 7 sections. Each chapter is concise and ends with a section for self-assessment that includes terms to define, true or false, and multiple choice questions for the reader to demonstrate understanding of the material. Each section is also followed by a discussion section with additional questions to apply the concepts from the relevant chapters. Although the examples in the discussions are from all areas of health care, a student of manual therapy should relate to the discussion following the section on research design in which an example of a person with low back pain is used to illustrate the benefits of different research designs. This edition includes a new chapter which discusses systematic reviews and meta-analyses to provide a more thorough exposure to synthesis of

doi:10.1016/j.math.2009.04.003

contemporary research methods, such as evidence-based practice. There is also an enhanced discussion of qualitative research methods and evaluation in this edition. In the sections on descriptive and interferential statistics, common statistical tests are introduced. Some of the tests are explained in more detail than others, for example chi-square and measures of effect size, with the purpose to illustrate their interpretation. An eighth section provides a glossary of terms, references for further reading, answers to the chapter self-tests and appendices of tables for three statistical tests. The references to additional research texts, although somewhat dated, are helpful. However, a reader looking for additional information on specific topics might be better served by a reference list for each chapter. This book is recommended as a text in introduction to research courses. Students and clinicians who have been previously exposed to research courses or are participating in research projects may find the book a quick review but otherwise will need to use other sources. David A. Scalzitti American Physical Therapy Association, 1111 N Fairfax Street, Alexandria, VA 22314, USA E-mail address: [email protected] 6 April 2009

Manual Therapy 14 (2009) e13

Contents lists available at ScienceDirect

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

Book Review Postural Disorders and Musculoskeletal Dysfunction: Diagnosis, Prevention and Treatment, G. Solberg, second ed. Churchill Livingstone, Elsevier (2008). 304 pp., price: £36.99, ISBN: 978-0-443-10382-7 This 300 page soft cover book focuses predominantly on clinical practice and combines theoretical and practical aspects of human movement and posture. It contains 3 main parts: the first part describes the theoretical background and encompasses basic kinesiology and its application to different joint regions. An overview of common postural disorders involving upper and lower extremities and identification of gait disorders is discussed in Part 2. The final part comprises more than 125 pages on the diagnosis and treatment of postural disorders. It explains how to assess posture and offers different therapeutic approaches utilizing adapted movement for special needs with an emphasis on children with postural disorders. The 4 appendices provide an impression of how a daily protocol can be used. It is a richly illustrated book with black and white photos and figures and the work is backed up by research

doi:10.1016/j.math.2009.04.001

data. The author has refined this second edition by integrating his clinical experience and research work of the past few years. It expands and enriches the use of movement for both, therapy and regular activity to improve postural patterns of daily function. This book is therefore an excellent reference for the analysis and treatment of posture and movement for all professions confronted with either correct or impaired movement. The neuromusculoskeletal therapist, independent of previously learned movement models or educational background, may generate many new ideas for clinical practice from this book. Harry J.M. von Piekartz* Fachhochshule/University of Applied Science, Faculty Physiotherapy, Caprivistrasse 30, ¨ ck(D), Germany 49076 Osnabru  Tel.: þ31 541 294 118; fax: þ31 541 294 002. E-mail address: [email protected] 2 April 2009