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). 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 (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
. 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 (112 N, þLR ¼ 2.3) and unaffected side ( 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 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 ( 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 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