This is an HTML version of an attachment to the Official Information request 'Acupuncture evaluation document'.



Pragmatic Evidence Based Review 
The efficacy of acupuncture in the 
management of musculoskeletal pain   
Reviewer  
 
Date Report Completed 
August 2011 
Important Note:  
• 
This report is not intended to replace clinical judgement, or be used as a clinical 
protocol. 
• 
A robust evidence-based review of clinical guidelines, systematic reviews and high 
quality primary evidence relevant to the focus of this report was carried out.  This 
does not however claim to be exhaustive. 
• 
The document has been prepared by the staff of the research team, ACC. The 
content does not necessarily represent the official view of ACC or represent ACC 
policy. 
• 
This report is based upon information supplied up to 31st July 2011 
 
Purpose 
The purpose of the report is to;  
•  Briefly describe traditional Chinese medicine (TCM) acupuncture and western 
medical acupuncture 
•  Report the efficacy of acupuncture for the treatment of injury-related spine, shoulder, 
knee & ankle conditions 
•  Report the comparative efficacy of acupuncture when considering alternative 
conservative treatment interventions for the spine, shoulder, knee & ankle 
•  Report any adverse reactions cited in the literature. 
 
Scope 
This report will be restricted to acupuncture involving various modes of needling (including 
electroacupuncture) for musculoskeletal pain from knee, spine, shoulder and ankle injuries. 
Treatment modalities of TCM like cupping, scraping, Chinese massage, and herbalism will 
not be addressed. 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page    

No distinction will be made between traditional Chinese medical acupuncture and western 
medical acupuncture 
 
Summary Message 
The evidence for the effectiveness of acupuncture is most convincing for the 
treatment of chronic neck and shoulder pain. In terms of other injuries, the evidence is 
either inconclusive or insufficient. The state of the evidence on the effectiveness of 
acupuncture is not dissimilar to other physical therapies such as physiotherapy, 
chiropractic and osteopathy.
 
Key findings 
General 
•  There is insufficient evidence to make a recommendation for the use of acupuncture 
in the management of acute neck, back or shoulder pain  
•  There is emerging evidence that acupuncture may enhance/facilitate other 
conventional therapies (including physiotherapy & exercise-based therapies) 
•  There is a paucity of research for the optimal dosage of acupuncture treatment for 
treating shoulder, knee, neck and lower back pain 
•  Studies comparing effective conservative treatments (including simple analgesics, 
physical therapy, exercise, heat & cold therapy) for (sub) acute and chronic non-
specific low back pain (LBP) have been largely inconclusive 
 
Lower back 
•  The evidence for the use of acupuncture in (sub)acute LBP is inconclusive 
•  There is limited evidence to support the use of acupuncture for pain relief in chronic 
LBP in the short term (up to 3 months) 
•  The evidence is inconclusive for the use of acupuncture for long term (beyond 3 
months) pain relief in chronic LBP  
•  There is no evidence to recommend the use of acupuncture for lumbar disc herniation 
related radiculopathy (LDHR) 
 
Neck  
•  There is good evidence that acupuncture is effective for short term pain relief in the 
treatment of chronic neck pain 
•  There is moderate evidence that real acupuncture is more effective than sham 
acupuncture for the treatment of chronic neck pain 
•  There is limited evidence that acupuncture has a long term effect on chronic neck 
pain 
 
 
Shoulder 
•  There is good evidence from one pragmatic trial that acupuncture improves pain and 
mobility in chronic shoulder pain 
•  There is limited evidence for the efficacy of acupuncture for frozen shoulder 
•  There is contradictory evidence for the efficacy of acupuncture for subacromial 
impingement syndrome 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page    

 
Knee 
•  There is no evidence to recommend the use of acupuncture for injury-related knee 
pain 
 
Ankle 
•  There is no evidence to recommend the use of acupuncture for ankle pain 
 
 
 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page    

Background 
Acupuncture has roots in ancient Chinese philosophy.  Traditional Chinese Medicine (TCM) 
acupuncture is based on a number of philosophical concepts, one of which is that any 
manifestation of pain/dysfunction is a sign of imbalance of energy flow within the body.  It is 
in this context that the TCM acupuncturist uses a holistic treatment approach.  TCM 
acupuncture involves inserting needles into traditional meridian points with the intention on 
influencing energy flow within that meridian1.  Acupuncture has been adopted into western 
medicine and treatments; many physicians currently practicing acupuncture reject such pre-
scientific notions described above, using unnamed tender or trigger points to stimulate 
nerves or muscles1.   Further to this acupuncture is also now regularly practiced globally by a 
specialist sub-group of physiotherapists and some other health professionals.  New Zealand 
physiotherapists have been practicing acupuncture since 19722.  
As a technique acupuncture includes the invasive or non-invasive stimulation of specific 
anatomical locations by means of needles or other thermal, electrical, light, mechanical or 
manual methods3.  Acupuncture is most commonly used to treat chronic pain4 5 and is 
currently used for a variety of conditions, including; spinal cord injury6, visceral dysfunction 
The other two studies compared 'traditional acupuncture' with suprascapular nerve block and 
acupuncture 'according to Jing Luo' respectively*6, headaches4, addictions6 emesis developing 
after surgery or chemotherapy in adults The other two studies compared 'traditional acupuncture' 
with suprascapular nerve block and acupuncture 'according to Jing Luo' respectively, nausea 
associated with pregnancy6 and dental pain7; all of which fall outside the scope of this report.  
Acupuncture is also used to treat a number of musculoskeletal conditions, including 
shoulder6, wrist, and lower back pain4 6 7 The other two studies compared 'traditional acupuncture' 
with suprascapular nerve block and acupuncture 'according to Jing Luo' respectively, knee pain4 6, 
neck pain, tennis/golfers elbow and ankle pain6.   
Modern acupuncture includes manual stimulation of needles that are inserted into the skin.  
Various adjuncts are often used including: electrical acupuncture (electrical stimulator 
connected to acupuncture needle), injection acupuncture (herbal extracts injected into 
acupuncture points), heat lamps, and moxibustion with acupuncture (the moxa herb, 
Artemesia vulgaris, is burned at the end of a needle). Dry needling is a technique used to 
treat myofascial pain in any part of the body8, by definition trigger point dry needling (TDN) 
and Intramuscular manual therapy (IMT) are acupuncture techniques3.  Dry needling involves 
the insertion of a needle at specific trigger points, the needle being a solid acupuncture 
needle or a dry injection needle. 
1.  Methodology 
Comprehensive literature searching was carried out focused on the efficacy of acupuncture 
for spine, knee, shoulder and ankle pain.  The databases accessed for the search were, 
Medline®, CINAHL, EMBASE, AMED, PsychINFO, PubMed and Medline-in-process and 
Google.  These databases will capture most, if not all, of the more robust clinical studies that 
may have been reported in the TCM-specific databases. In addition, the databases used 
here are used routinely in evidence-based research for complementary and alternative 
medicines. Of note, the TCM-specific databases contain many case series studies and other 
study designs that would be excluded from this report. 
                                                
* see Green 200537. Green S, et al. Acupuncture for shoulder pain. Cochrane Database Syst Rev 
2005(2):CD005319. for more details 
† see Green 200537. Green S, et al. Acupuncture for shoulder pain. Cochrane Database Syst Rev 
2005(2):CD005319. for more details 
‡ see Green 200537. Green S, et al. Acupuncture for shoulder pain. Cochrane Database Syst Rev 
2005(2):CD005319. for more details 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page    

The search was run on the 31st July 2011 for the period 2000 to present.  Manual searching 
of reference lists was also carried out.  A pragmatic approach was taken initially searching 
for randomised controlled trials (RCTs), systematic reviews and meta-analyses, as the 
highest levels of evidence.  RCT’s are also the trial design of choice when investigating 
treatment efficacy.  
The literature was critically appraised using SIGN9 (see below) grading system for systematic 
reviews and RCTs. 
 
SIGN – LEVELS OF EVIDENCE 
1++  High quality meta-analyses, systematic reviews of RCTs, or RCTs with a very low 
risk of bias 
1+  Well-conducted meta-analyses, systematic reviews, or RCTs with a low risk of bias
1- 
Meta-analyses, systematic reviews, or RCTs with a high risk of bias 
2++  High quality systematic reviews of case control or cohort or studies 
High quality case control or cohort studies with a very low risk of confounding or 
bias and a high probability that the relationship is causal 
2+  Well-conducted case control or cohort studies with a low risk of confounding or 
bias and a moderate probability that the relationship is causal 
2- 
Case control or cohort studies with a high risk of confounding or bias and a 
significant risk that the relationship is not causal 

Non-analytic studies, e.g. case reports, case series 
4 Expert 
opinion 
 
3. Review of the Literature 
Neck pain and lower back pain (LBP) are two conditions that can be problematic to treat.  
Studies examining effective conservative treatments for (sub)acute and chronic non-specific 
low back pain have been largely inconclusive.  This is also true of neck and thoracic spine 
pain.   
A lot of the literature focuses on chronic spinal pain; there are no high quality trials for the 
treatment of acute spinal pain.   
There is limited evidence to suggest that acupuncture is not an appropriate treatment for any 
spinal condition with suspected neurological involvement10 11. 
3.1 Lower Back Pain 
Chronic spinal pain presents a diagnostic and treatment challenge ,reaching a specific 
diagnosis is often difficult.  Effective conservative treatments for (sub)acute and chronic non-
specific LBP have been largely inconclusive12.  Differing patient populations and 
methodologies make direct comparison of studies problematic often resulting in inconclusive 
findings.   
Studies comparing spinal manipulation, medication, and acupuncture for chronic spinal pain 
revealed that spinal manipulation produced the greatest benefit both in the short13 and long 
term 12; within these studies acupuncture produced ‘consistent’ improvement in outcomes 
although this did not reach statistical significance.  Outcome measures addressed both pain 
and function (Oswestrey scale, Visual Analogue Scale (VAS), lumbar flexion in sitting and 
standing); overall recovery was 27% of the patients receiving spinal manipulation, 9.4% of 
those receiving acupuncture and only 5% of those receiving medication.  It is noteworthy 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page    

here that spinal manipulation is not appropriate for all LBP patients and a range of 
conservative treatment options must always be considered.  In this study it was not possible 
to blind the patient and the therapist to the treatment allocation due to the ‘hands on’ nature 
of manipulation and acupuncture, therefore the placebo effect cannot be discounted.   
In a study14 comparing 3 different acupuncture approaches (individualised, standardised & 
sham) to standard care (inclusive of medications, primary care and physical therapy, non-
study related), all groups treated with acupuncture demonstrated greater improvement in 
dysfunction than standard care14.  The acupuncture groups included in this study all used 
different needle locations and depths, which suggests that this is unimportant in eliciting a 
therapeutic effect and may in-fact represent a placebo or non-specific effect.  This was the 
only study reporting on function; the literature more frequently reports pain relieving effects. 
Itoh et al15 reported that a study group receiving trigger point acupuncture recorded 
significantly less pain (VAS) than a sham control group.  This finding remained true when the 
groups were crossed over following a 3 week washout period.  As acknowledged by the 
authors, the 3 week washout may have been insufficient and therefore a carry over treatment 
effect could not be discounted.  This study does however support the notion that both sham 
and real acupuncture exert positive therapeutic effects on chronic LBP and that real 
acupuncture is more effective than sham. 
A systematic review of acupuncture for chronic LBP16 returned only 5 RCT’s.   A meta-
analysis was not performed due to the wide disparities in design, groups, needling points, 
control groups and how & when pain relief outcomes were measured in these studies.  The 
trials were examined individually, and did not provide definitive evidence to support or refute 
acupuncture as an effective treatment for chronic LBP.  Closer examination of the articles 
included in the review reveals that the results of the RCT’s show a trend towards study 
groups receiving some form of acupuncture intervention show improvement/positive 
treatment effects.  However in agreement with the review author there are some 
methodological issues within the studies that prevent the drawing of definitive conclusions.  A 
systematic review8 concluded that in chronic LBP acupuncture is more effective than no 
treatment or sham treatment at up to 3 month follow up.  It was also reported that 
acupuncture as an adjunct to conventional therapies is more effective than conventional 
therapies alone.  Dry needling is also considered in this review and reported as a useful 
adjunct to other therapies for chronic LBP.  
A larger systematic review 1 inclusive of both acute and chronic LBP focused on the primary 
outcome of short term pain relief reports that acupuncture is described as statistically 
significantly and clinically important and is more effective than sham acupuncture and 
concludes that acupuncture effectively relieves chronic LBP.  It is noteworthy that of the 33 
RCTs included in the review only 22 could be included in the meta-analysis due to the 
heterogeneity across the study samples and methodologies in the remaining 11 RCTs, 4 of 
which were related to chronic LBP.  The quality of the studies included in the meta-analysis 
is variable, as such the findings from this review should only be considered as somewhat 
preliminary.  Future publication of larger trials would have an impact on the evidence overall.   
A more recent systematic review17 inclusive of 6 RCTs not published when previous reviews1 
8 were carried out reported that there is moderate evidence that acupuncture is more 
effective than no treatment and strong evidence of no significant difference between 
acupuncture and sham acupuncture for short term pain relief for chronic LBP.     
 
Considering 3 systematic reviews1 8 17 of reasonable quality the evidence shows a trend 
towards acupuncture being more effective than no treatment, however the evidence remains 
limited.  There are inconsistent findings for acupuncture versus sham acupuncture.  There is 
consistent evidence that acupuncture is a useful adjunct to other conservative treatments 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page    



















4. Hopton A, MacPherson, H.,. Acupuncture for chronic pain: Is acupuncture more than 
an effective placebo? A systematic review of pooled data from meta-analyses. Pain 
Practice
 2010;10(2):94-102. 
5. Ernst E, Lee, M.S., Choi, T.Y., . Acupuncture: Does it alleviate pain and are there serious 
risks? A review of reviews. Pain 2011;152:755-64. 
6. Harland S, Bleakley, C., McDonough, S. M.,. Acupuncture in soft tissue injury 
management: a systematic review... Rehabilitation and Therapy Research Society 
Third Annual Conference: Collaborative research... making it a reality, held on 26-
27 April 2007 at University of Ulster. Physical Therapy Reviews 2008;13(2):121-22. 
7. Kaptchuk TJ. Acupuncture: theory, efficacy, and practice. Annals Of Internal Medicine 
2002;136(5):374-83. 
8. Furlan AD, van-Tulder, M., Cherkin, D., . Acupuncture and Dry-Needling for Low Back 
Pain: An Updated Systematic Review Within the Framework of the Cochrane 
Collaboration. Spine 2005;2005(8):944-63. 
9. Scottish Intercollegiate Guidelines Network. SIGN 50: Guideline Development. 
10. Hahne AJ, Ford, J. J., McMeeken, J. M., . Conservative management of lumbar disc 
herniation with associated radiculopathy: A systematic review. Spine 
2010;35(11):488-504. 
11. Zhu XM, Polus, B.,. A controlled trial on acupuncture for chronic neck pain. American 
Journal of Chinese Medicine 2002;30(1):13-28. 
12. Muller R, Giles, L. G. F.,. Long-term follow-up of a randomized clinical trial assessing 
the efficacy of medication, acupuncture, and spinal manipulation for chronic 
mechanical spinal pain syndromes. Journal of Manipulative & Physiological 
Therapeutics
 2005;28(1):3-11. 
13. Lynton GF, Giles, D.C., Muller, R.,. Chronic Spinal Pain: A Randomized Clinical Trial 
Comparing Medication, Acupuncture and Spinal Manipulation. Spine 
2003;28(14):1490-503. 
14. Cherkin DC, Sherman, K.J., Avins, A.L., Erro, J.H., Ichikawa, L., Barlow, W.E., 
Delaney, K., Hawkes, R., Hamilton, L., Pressman, A., Khalsa, P.S., Deyo, R.A., . A 
Randomised Trial Comparing Acupuncture and Usual Care for Chronic Low Back 
Pain. Archives of Internal Medicine 2009;169(9):838-66. 
15. Itoh K, Katsumi, Y., Hirota, S., Kitakoji, H. Effects of trigger point acupuncture on 
chronic low back pain in elderly patients -- a sham-controlled randomised trial. 
Acupuncture in Medicine 2006;24(1):5-12. 
16. Henderson H. Acupuncture: evidence for its use in chronic low back pain. British 
Journal of Nursing (BJN) 2002;11(21):1395-403. 
17. Yuan J, Purepong, N., Kerr, D.P., Park, J., Bradbury, I., McDonough, S., . Effectiveness 
of Acupuncture for Low Back Pain: A systematic review. Spine 2008;33(23):E887-
900. 
18. Brinkhaus B. WCM, Jena S., Linde K.,. Acupuncture in patients with chronic low back pain: a 
randomized controlled trial. . Archives of Internal Medicine 2006;166(4):450-57. 
19. Cherkin DC, Sherman, K. J., Hogeboom, C. J., Erro, J. H., Barlow, W. E., Deyo, R. A., 
Avins, A. L.,. Efficacy of acupuncture for chronic low back pain: protocol for a 
randomized controlled trial. Trials 2008;9:10-10. 
20. Itoh K, Katsumi, Y., Kitakoji, H. Trigger point acupuncture treatment of chronic low 
back pain in elderly patients -- a blinded RCT. Acupuncture in Medicine 
2004;22(4):170-77. 
21. Itoh K, Katsumi, Y., Hirota, S., Kitakoji, H. Randomised trial of trigger point 
acupuncture compared with other acupuncture for treatment of chronic neck pain. 
Complementary Therapies in Medicine 2007;15(3):172-79. 
22. Kennedy S, Baxter, G. D., Kerr, D. P., Bradbury, I., Park, J., McDonough, S. M. 
Acupuncture for acute non-specific low back pain: a pilot randomised non-
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page    16 

penetrating sham controlled trial. Complementary Therapies in Medicine 
2008;16(3):139-46. 
23. Fu L, Li, J., Wu, W.,. Randomized controlled trials of acupuncture for neck pain: 
systematic review and meta-analysis. Journal of Alternative & Complementary 
Medicine
 2009;15(2):133-45. 
24. van Tulder MW, Goosens, M., Hoving, J., . Non-surgical treatment of chronic neck 
pain In: Nachemson A, Jonsson, E., , editor. Neck and Back pain. Philadelphia: 
Lippincott, Williams and Wilkins, 2000. 
25. Trinh K, Graham, N., Gross, A., Goldsmith, C., Wang, E., Cameron, I., Kay, T., . 
Acupuncture for Neck Disorders. Spine 2007;32(2):236-43. 
26. He D, Hostmark, At., Veiersted, Kb., Medbo, Ji.,. Effect of intensive acupuncture on 
pain-related social and psychological variables for women with chronic neck and 
shoulder pain - an RCT with six month and three year follow up. Acupuncture in 
Medicine
 2005;23(2):52-61. 
27. White P, Lewith, G., Prescott, P., Conway, J.,. Acupuncture versus placebo for the 
treatment of chronic mechanical neck pain: a randomized, controlled trial. Annals 
of Internal Medicine
 2004;141(12):911-19. 
28. Grainger R. Physiotherapy and acupuncture treatment for Achilles tendinopathy in a 
high-level female rugby player. Journal of the Acupuncture Association of Chartered 
Physiotherapists
 2009:67-76. 
29. Knight RR. Integration of manual therapy, rehabilitation and acupuncture in the 
treatment of a 17-year-old male professional football player with chronic medial 
tibial stress syndrome. Journal of the Acupuncture Association of Chartered 
Physiotherapists
:81-87. 
30. Santha CC. Acupuncture treatment for bilateral heel pain caused by plantar fascitis. 
Journal of the Acupuncture Association of Chartered Physiotherapists:67-74. 
31. Green S, Buchbinder, R., Hetrick, S. E.,. Acupuncture for shoulder pain. Cochrane 
Database of Systematic Reviews 2005(2). 
32. Molsberger AF, Schneider T, Gotthardt H, Drabik A. German Randomized 
Acupuncture Trial for chronic shoulder pain (GRASP) - A pragmatic, controlled, 
patient-blinded, multi-centre trial in an outpatient care environment. Pain 
2010;151(1):146-54. 
33. Guerra de Hoyos JA, Andres Martin Mdel C, Bassas y Baena de Leon E, Vigara Lopez 
M, Molina Lopez T, Verdugo Morilla FA, et al. Randomised trial of long term effect 
of acupuncture for shoulder pain. Pain 2004;112(3):289-98. 
34. He D, Veiersted, Kb., Hostmark, At., Medbo, Ji.,. Effect of acupuncture treatment on 
chronic neck and shoulder pain in sedentary female workers: a 6-month and 3-year 
follow-up study. Pain 2004;109(3):299-307. 
35. Lathia AT, Jung, S. M., Chen, L. X.,. Efficacy of acupuncture as a treatment for chronic 
shoulder pain. Journal of Alternative & Complementary Medicine 2009;15(6):613-18. 
36. Favejee MM, Huisstede BM, Koes BW, Huisstede BMA. Frozen shoulder: the 
effectiveness of conservative and surgical interventions--systematic review. BJSM 
online
 2011;45(1):49-56. 
37. Nyberg A, Jonsson P, Sundelin G. Limited scientific evidence supports the use of 
conservative treatment interventions for pain and function in patients with 
subacromial impingement syndrome: Randomized control trials. Physical 
2010;15(6):436-52. 
38. Johansson K, Bergstrom A, Schroder K, Foldevi M, Johansson K, Bergstrom A, et al. 
Subacromial corticosteroid injection or acupuncture with home exercises when 
treating patients with subacromial impingement in primary care--a randomized 
clinical trial. Fam Pract 2011;28(4):355-65. 
39. Manheimer E, Cheng K, Linde K, Lao L, Yoo J, Wieland S, et al. Acupuncture for 
peripheral joint osteoarthritis. Cochrane Database Syst Rev 2010(1):CD001977. 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   17 

40. Kelly RB, Kelly RB. Acupuncture for pain. Am Fam Physician 2009;80(5):481-4. 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   18 

A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   19 



questionnaires at baseline, 8, 26, and 52 
pain than no acupuncture treatment in 
weeks after randomization.  
patients with chronic low back pain 
 
There was no significant differences between 
acupuncture and minimal acupuncture 
Primary outcome variable was the change in 
low back pain (VAS) intensity from baseline to 
the end of week 8  
2a. Lynton et al (2003) Chronic Spinal Pain: A  3 armed RCT (includes full spine) 
Earliest asymptomatic status: 
Randomized Clinical Trial Comparing 
Medication, Acupuncture and Spinal 
1- medication 
Manipulation (27%) 
Manipulation 
2–needle acupuncture 
Acupuncture (9.4%) 
Level of evidence 1+ 
3-spinal manipulation (chiropractic) 
Medication (5%) 
N=115  
 
Outcome measures at 0,2,5,9 weeks 
Best overall results from outcomes were for 
treatment 
manipulation 
 
Patients with chronic spinal pain results in 
greatest short term improvement.  Data are 
not strong 
2b. Muller et al (2005) Long-term follow-up of 
Extended follow-up (>1 year) of RCT  
Comparisons of initial and extended follow-up 
a randomized clinical trial assessing the 
questionnaires to assess absolute efficacy 
efficacy of medication, acupuncture, and 
N=62/69  
showed that only the application of spinal 
spinal manipulation for chronic mechanical 
N=40/62 patients who had received 
manipulation revealed broad-based long-term 
spinal pain syndromes 
exclusively the randomly allocated treatment 
benefit 
Level of evidence 1+ 
for the whole observation period since 
 
randomization 
In patients with chronic spinal pain 
syndromes, spinal manipulation, if not 
contraindicated, may be the only treatment 
modality of the assessed regimens that 
provides broad and significant long-term 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
21 

benefit 
3a. Cherkin et al (2008) Efficacy of 
4 arm RCT, n=640 
Protocol detail only 
acupuncture for chronic low back pain: 
protocol for a randomized controlled trial 
3b. Cherkin et al (2009) A Randomised Trial 
4 arm RCT, n=638 
At 8 weeks grps 1,2,3 improvement in function 
Comparing Acupuncture and Usual Care for 
Chronic Low Back Pain 
1-Individualised acupuncture 
At 1 year follow up grps 1,2,3 improved 
function but not symptoms 
Level of evidence 1- 
2-Standardised acupuncture 
 
3-Simulated acupuncture 
Site and depth of penetration appear 
4-Usual care 
unimportant in eliciting therapeutic benefit. 
Outcome measures at 0,8,26,52 weeks post 
Raises question about physiological effect, 
treatment onset 
may represent placebo or non-specific effects 
4. Hahne et al (2010) Conservative 
Systematic review of randomized controlled 
Search returned no studies including 
management of lumbar disc herniation with 
trials for specific diagnosis of LDHR 
acupuncture 
associated radiculopathy: A systematic review  radiologically confirmed 
Level of evidence 1++ 
5. Henderson (2002) Acupuncture: evidence 
Systematic review on Western countries (11 
No conclusive evidence to support or refute 
for its use in chronic low back pain 
articles; 3 case studies, 5 randomized 
the use of acupuncture in Low back pain 
controlled trials & 2 cross-over trials) 
Level of evidence 1+/2++ 
 
 
Increasing number of patients suffering from 
back pain seeking complementary therapies  
to supplement traditional medical treatments 
6. Itoh et al (2006) Effects of trigger point 
RCT 
At the end of the first treatment phase,  
acupuncture on chronic low back pain in 
elderly patients -- a sham-controlled 
N=26 randomised to two groups 
 
randomised trial 
Each group received one phase of trigger 
Group A receiving trigger point acupuncture 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
22 

Level of evidence 1+ 
point acupuncture and one of sham 
had significantly less pain than the sham 
acupuncture with a three week washout 
control group 
period between them, over 12 weeks 
 
 
 
Group A (n = 13) trigger point acupuncture in 
first phase & sham acupuncture in the second  There were significant within-group reductions 
in pain in both groups during the trigger point 
 
acupuncture phase but not in the sham 
treatment phase 
Group B (n = 13) received the same 
interventions in the reverse order  
 
Beneficial effects were not sustained 
 
These results suggest that trigger point 
acupuncture may have greater short term 
effects on low back pain in elderly patients 
than sham acupuncture 
7. Itoh et al (2004) Trigger point acupuncture 
Double blind crossover RCT 
Deep resulted in less pain intensity and 
treatment of chronic low back pain in elderly 
improved QoL compared to standard 
patients -- a blinded RCT 
N=35 were randomised to 1 of 3 groups over 
acupuncture or superficial needling to trigger 
12 weeks 
points 
Level of evidence 1+ 
Each group received 2 phases of acupuncture   
treatment with an interval between them 
Reduction in pain intensity between the 
1. Standard acupuncture group received 
treatment & interval in the group that received 
treatment at traditional acupuncture points for 
deep needling (not the case in standard 
low back pain 
acupuncture or superficial needling to trigger 
 
points 
2. Superficial treatment on trigger points 
 
Deep needling to trigger points may be more 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
23 

 
effective in the treatment of low back pain in 
elderly patients than standard acupuncture or 
3. Deep treatment on trigger points 
superficial needling to trigger points 
8. Kennedy et al (2008) Acupuncture for acute  A pilot patient and assessor blinded 
For pain, the only statistically significant 
non-specific low back pain: a pilot randomised  randomized controlled trial 
difference was at the 3 months follow up 
non-penetrating sham controlled trial 
N=48, 12 weeks treatment. 
 
Level of evidence 1- 
1. Placebo group with sham needle 
At the end of treatment; verum acupuncture 
group were taking significantly fewer tablets of 
2. Verum acupuncture 
pain control medication  
Outcome measures at baseline, end of 
 
treatment & 3 months follow up 
This study has demonstrated 120 participants 
would be required in a fully powered trial.  
The placebo needle used in this study proved 
to be a credible form of control 
9. Furlan et al (2005) Acupuncture and Dry-
Systematic review of RCTs (1996-2003) 
Insufficient evidence to support efficacy of 
Needling for Low Back Pain: An Updated 
acupuncture or dry needling in acute LBP 
Systematic Review Within the Framework of 
Acupuncture for (sub) acute & chronic non-
the Cochrane Collaboration 
specific LBP 
 
Level of evidence 1++ 
Dry needling for myofascial trigger points, 
For chronic LBP Acupuncture more effective 
compared to; 
than no treatment or sham treatment up to 3 
months.   
- No treatment 
For chronic LBP acupuncture is more effective 
- Sham therapy 
than no treatment for improving function in the 
- Other therapy 
short term  
- Addition of acupuncture to other therapy 
As an adjunct to other conventional therapies 
acupuncture relieves pain and improves 
 
function better than conventional therapies 
alone 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
24 



3. Acupuncture versus inactive treatment 
3. Moderate evidence that acupuncture is 
more effective than inactive treatment for pain 
4. Acupuncture versus wait list control 
relief post treatment and at short term follow 
up 
4. Moderate evidence that patients receiving 
acupuncture report less pain than those on a 
wait list control at short term follow up 
8. White et al (2004) Acupuncture versus 
Randomized, single-blind, placebo-controlled,  Both groups improved statistically from 
placebo for the treatment of chronic 
parallel-arm trial with 1-year follow-up 
baseline 
mechanical neck pain: a randomized, 
controlled trial 
n=135, 4 weeks, 8 treatments 
 
Level of evidence 1- 
1. acupuncture 
Acupuncture was more effective than mock 
treatment for pain relief at short term follow up 
2. Mock transcutaneous electrical stimulation 
of acupuncture points using a 
 
decommissioned electroacupuncture 
stimulation unit 
However, this difference was not clinically 
significant 
 
Limitations All treatments were provided by 1 
practitioner, control did not mimic the process 
of needling, non-intervention group was not 
present  
 
Acupuncture reduced neck pain and produced 
a statistically, but not clinically, significant 
effect compared with placebo. The beneficial 
effects of acupuncture for pain may be due to 
both nonspecific and specific 
9. Zhu et al (2002) A controlled trial on 
Chinese medicine (CM) acupuncture for 
Significant reduction in subjective pain 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
26 

acupuncture for chronic neck pain 
chronic neck pain (CNP)  
intensity (VAS), pain hours per day, analgesic 
pill consumption & increased activity level 
Level of evidence 1+ 
Single blind, controlled, crossover, clinical trial  following 9 session real CM acupuncture 
n=29 
 
2 groups received two phases of treatment 
with a washout period between the two 
The same for sham but to a lesser degree 
phases 
Sham acupuncture has a therapeutic effect  
Group A - CM acupuncture, washout, sham 
acupuncture Group B – Sham, washout, CM 
Acupuncture may be a suitable intervention 
acupuncture 
for neck pain – not applicable to those with 
neurological or psychosocial signs present 
 
 
9 sessions over 3 weeks 
Acupuncture and sham treatment have a long 
 
term effect of neck pain lasting at least 16 
weeks 
Manual twisting of the needle was applied on 
all points plus strong electrical stimulation of 
 
distal points in CM acupuncture. Sham 
acupoints (lateral to the real) and sham 
Neither Sham or real CM acupuncture had 
(weak) electrical stimulation was used in the 
any significant effect on objective measures 
control group.  
Outcome measures at baseline, after each 
phase of treatment, after washout, & at 16 
week follow-up 
He et al 2005 Effect of intensive acupuncture 
This study examines whether intensive 
The ‘pain-related activity impairment at work’ 
on pain-related social and psychological 
acupuncture treatment can improve several 
was significantly less in Acupuncture group 
variables for women with chronic neck and 
social and psychological variables for women 
than sham (control) by the end of treatment  
shoulder pain - an RCT with six month and 
with chronic pain in the neck and shoulders, 
three year follow up 
and whether possible effects are long-lasting 
There were significant differences between 
the groups for; quality of sleep, anxiety, 
Level of evidence 1- 
 
depression & satisfaction with life  
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
27 

N=24 female office workers  
 
Acupuncture was applied 10 times during 3-4 
At 6 month & 3 year follow ups the 
weeks  
acupuncture group showed further 
improvements in most variables and was 
 
again significantly different from the control 
1. Acupuncture points  
group  
2. Sham points (control group) 
 
 
Intensive acupuncture treatment may improve; 
activity at work and social & psychological 
In addition, acupressure was given to patients  variables for women with chronic pain in the 
between treatments, at either real or sham 
neck and shoulders 
points. Questionnaires for social and 
psychological variables were completed 
The effect may last for at least 3 years 
before treatment, just after the course, 6 
months & 3 years follow up 
He et al (2004) Effect of acupuncture 
Randomized single blind controlled trial 
The intensity & frequency of pain decreased 
treatment on chronic neck and shoulder pain 
more for TG than CG during treatment period 
in sedentary female workers: a 6-month and 
N= 24 female office workers  
3-year follow-up study 
 
randomly assigned to 
Level of evidence 1+ 
At 3 year follow up, TG reported less pain 
 
than pre treatment   
1. Test Group (TG) - anti-pain acupoints 
 
 
Headache decreased during treatment period 
for both groups, but more for TG than for CG  
2. Control Group (CG) - placebo-points 
 
 
At 3 year follow up TG still had decrease in 
Acupuncture was applied 10 times during 3-4 
headaches 
weeks  
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
28 

 
 
Outcome measures; pain threshold (PPT) in 
In CG headache returned to pre-treatment 
the neck and shoulders with algometry before  level 
first treatment, after the last treatment & at 6 
month follow up. Questionnaires on muscle 
 
pain and headache were answered at the 
same time points & at 3 years follow up  
PPT of some muscles increased during the 
treatment period for TG & remained higher 6 
months post treatment 
 
Acupuncture treatment may have long term 
effect in reducing chronic pain in neck & 
shoulders & related headache 
Acupuncture more effective than sham 
 
Sham acupuncture may have immediate pain 
relieving effect on chronic neck & shoulder 
pain 
 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
29 

7. Appendix 3: Evidence tables for shoulder, knee and pain  
SHOULDERS 
Reference and study 
Studies 
Intervention/comparison
Outcome measure 
Results/effect size 
Conclusions 
design 
Favejee MM, 
N = 5 Cochrane 
Interventions: oral 
Pain 
Cheing 2008 (n=70) 
In the short term, 
Huisstede BM, Koes 
reviews & 18 RCTs [1 
medications, injection 
moderate evidence 
BW, Huisstede BMA.  
Cochrane review and 1  therapy, physiotherapy, 
 
• electroacupuncture from one small study 
RCT for acupuncture] 
acupuncture, 
vs. interferential 
was found for the 
 
Function 
arthrographic distension & 
electrotherapy 
effectiveness of 
 
suprascapular nerve 
vs.placebo.  
acupuncture and 
Frozen shoulder: the 
 
block 
• Significant 
exercises with respect 
effectiveness of 
Total number of 
 
differences were 
to shoulder function 
conservative and 
patients in the studies: 
 
found between 
[Sun 2001] 
surgical interventions--
not reported 
Quality scores: 
both treatment 
systematic review.  
Length of treatment: 
groups and the 
 
 
variable 
Cheing 2008, 33% 
control group, on 
 
[low] 
pain and function 
Limited evidence for 
Inclusion criteria: 
 
(all p<0.001) at 4 
effectiveness of 
BJSM online 
patients with frozen 
Sun 2001, 55% [high] 
weeks. 
electroacupuncture 
2011;45(1):49-56. 
shoulder, not caused 
Comparison (placebo): 
 
compared to placebo 
by acute trauma or 
variable 
Lin 1994, 36% [low] 
 
on pain and function at 
systemic disease; an 
Sun 2001 (n=35)  
 
Yuan 1995, 36% [low] 
4 weeks [Cheing 2008] 
Netherlands 
intervention for treating 
frozen shoulder; pain, 
• acupuncture 

Co-interventions: variable 
exercises vs. 
 
function or recovery 
outcomes were 
exercises alone 
 
Included studies 

reported; in English, 
 significant 
looking at acupuncture: 
difference in favour 
French, German or 
 
Cheing 2008, Sun 
of acupuncture + 
Dutch. 
2001, Lin 1994, Yuan 
exercises on 
shoulder function 
1995. 
 
at 20 weeks 
Exclusion criteria: none 
•  9.40 WMD; 95% 
reported 
CI 0.52 to 18.28 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
30 

 
Lin 1994 (n=100)  
Databases used: 
• suprascapular 
Cochrane library, 
nerve blocks 
PubMed, EMBASE, 
(SSNB) vs. 
CINAHL, PeDro 
acupuncture  
• significant 
 
differences in 
favour of SSNB on 
Description of the 
pain and ROM 30 
methodological 
min after treatment 
assessment of studies: 
• WMD 
(pain) 
1.33; 
score adapted from 
95% CI 1.22 to 
Cochrane review 
1.44)  
handbook 
• WMD 
(flexion) 
−7.00; −11.17 to 
 
−2.83)  
 
No meta-analysis 
Yuan 1995 
 
• significant 
Qualitative (‘best-
difference in favour 
evidence’) analysis 
of acupuncture 
according to Jing 
Luo over traditional 
acupuncture on 
recovery 
•  RR 1.50; 95% CI 
1.08 to 2.09 
•  follow-up time not 
reported 
Study type: Systematic review with qualitative analysis 
 
Quality: SIGN 1+ 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
31 

Comments: Wide range of interventions; good search and methodology appraisal; qualitative analysis appropriate; heterogeneity not formally reported; some 
reporting not sufficient enough (due to inability to access online supplementary appendices) 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
32 

 
 
Reference and study 
Participants 
Intervention/comparison
Outcome measure 
Results/effect size 
Conclusions 
design 
Molsberger AF, 
n=424 participants 
1. Acupuncture: 
15 
Pain (VAS) 
Primary end-point: 
In people with chronic 
Schneider T, Gotthardt 
treatments (1-3 per 
shoulder pain, ‘true’ 
H, Drabik A.  
•  135 ‘sham’ group 
week, lasting 20 
 
 
acupuncture reduced 
• 154 
acupuncture 
mins) 
pain and improved 
 
group 
 
[‘Responder’ = 
‘Responders’ at 3 
mobility significantly 
•  135 ‘COT’ group 
reduction of pain by 
months: 
more than ‘sham’ 
German Randomized 
 
2. ‘Sham” 
acupuncture:  ≥50% on VAS from 
acupuncture or 
Acupuncture Trial for 
as above 
initial score] 
1. 64.9% 
conventional therapy at 
chronic shoulder pain 
Inclusions: one-sided 
 
2. 23.7% 
6 weeks and 3 months. 
(GRASP) - A 
3. 37.0% 
shoulder pain ≥6 weeks 
 
pragmatic, controlled, 
 
and up to two years; an  3. ‘COT’: 
conventional 
Shoulder mobility (Jobe 
patient-blinded, multi-
average pain score of 
orthopaedic therapy 
test; degree of 
1 vs. 2  p<0.01 
centre trial in an 
with 50mg diclofenac 
≥50 mm on a VAS in 
abduction; % full 
outpatient care 
daily and 15 
the past week; age 
elevation of arm 
1 vs. 3  p<0.01 
environment.  
treatment sessions 
between 25 and 65 
individually selected 
possible) 
years; the ability to 
 
 
from physiotherapy, 
communicate 
physical exercise, 
OR (1 vs. 2 ) = 5.96 
Pain 2010;151(1):146-
heat/cold therapy, 
in German 
[95%CI: 3.45-10.35] 
54. 
ultrasound and TENS 
 
 
 
 
Exclusions: injections 
Length of treatment: 6 
OR (1 vs. 3) = 3.15 
Germany 
or cortisone of any kind;  weeks 
[95%CI: 1.90-5.23] 
neurological disorders 
 
causing shoulder pain; 
 
 
Multicentre study 
referred pain from the 
Selection of acupuncture 
cervical spine; OA of 
points:  
Secondary end-point: 
 
the 
1. Acupuncture: 
 
gleno-humeral joint or 
consensus agreement 
systemic bone and joint 
from experts to use 
‘Responders’ 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
33 

disorder (e.g. 
particular points ± 
immediately after 
rheumatoid arthritis); 
others (5-10 needles) 
treatment ended: 
history of shoulder 
2.  ‘Sham’: 8 needles at 
surgery; other 
defined non-
1. 68.1% 
acupuncture points 
2. 39.3% 
current therapy 
near both tibia 
3. 28.1% 
involving analgesics; 
 
overt psychiatric illness; 
pregnancy; incapacity 
1 vs. 2  p<0.001 
for work >3 months 
1 vs. 3  p<0.001 
preceding 
 
the trial, and pending 
compensation 
OR (1 vs. 2 ) = 2.30 
procedure 
[95%CI: 1.40-3.78] 
 
 
Randomisation & 
OR (1 vs. 3) = 3.77 
allocation concealment 
reported 
[95%CI: 2.24-6.41] 
Blinding:  
 
•  patients blinded to 
Post hoc analyses of 
whether in 
shoulder mobility: 
acupuncture or 
sham group but 
Acupuncture group all 
not to COT group 
significantly improved 
• statisticians 
at 6 weeks & 3 months 
blinded to 
compared to sham or 
allocation group 
COT (see full text for 
• observers 
not 
details) 
blinded 
• those 
 
administering 
 
treatment not 
blinded 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
34 

 
Dropouts at 3 months: 
• 61/135 
(45%) 
‘sham’ group 
• 26/154 
(17%) 
acupuncture 
group 
• 29/135 
(22%) 
‘COT’ group 
• overall 
drop-out 
rate ~ 27% 
 
Follow-up: at end of 
treatment and at 3 
months after 
 
Characteristics:  
Mean age: 51 (sham); 
50 (acupuncture); 51 
(COT) years  
% Male: 33; 43; 33% 
Duration: 12; 11; 10 
months 
 
No significant 
differences between 
groups in any 
characteristic reported 
i.e. affected shoulder, 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
35 

pain intensity 
  
Secondary care 
(outpatients) 
 
Power calculation 
 
ITT analysis 
 
Mixed diagnoses: 40% 
bursitis subacromialis, 
29.4% bursitis calcarea, 
3.9% frozen shoulder & 
2.5% biceps tendinitis 
 
 
 
 
 
Study type: multi-centre pragmatic RCT 
 
Quality: 1+ 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
36 

Comments: Well conducted pragmatic, three-armed, patient-blinded, multi-centre RCT. Not observer blinded for acupuncture or sham & not blinded for COT 
therefore possibility of bias present.  At 3 months ~27% participants dropped out but ITT analysis i.e. drop-outs considered ‘non-responders’. ‘Mixed’ 
population. 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
37 

 
Reference and study 
Participants 
Intervention/comparison
Outcome measure 
Results/effect size 
Conclusions 
design 
Guerra de Hoyos JA, 
n=130 
Treatment: “standardised”  Primary outcome:  
Mean difference VAS 
“All results consistently 
Andres Martin Mdel C, 
electro-acupuncture i.e. 
(95%CI): 
suggested that real 
Bassas y Baena de 
•  65 in acupuncture  all patients had same 4 
Pain (VAS) 
acupuncture is more 
Leon E, Vigara Lopez 
arm 
acupuncture points used 
 
effective than placebo-
 
M, Molina Lopez T, 
•  65 in placebo 
acupuncture to treat 
7 weeks: 
Verdugo Morilla FA, et 
(“sham”) arm 
 
Secondary outcomes: 
pain and disability in 
al.  
 
patients with shoulder 
Length of treatment: 8 
• 1.5 
(0.8-2.3) 
Lattinen index (pain) 
p<0.0005 
pain from different 
 
Inclusions: Clinical 
weeks 
causes, mainly rotator 
diagnosis of soft tissue 
 
ROM 
cuff disease and 
Randomised trial of 
shoulder lesions; no 
 
3 months: 
capsulitis.” 
long term effect of 
swelling signs; no 
SPADI (pain & 
acupuncture for 
recent trauma (previous  Comparison: “sham” 
disability) 
• 1.5 
(0.6-2.5) 
shoulder pain.  
3 months); no previous 
acupuncture with needles 
p<0.0005 
acupuncture 
not penetrating skin and 
COOP/WONCA (quality   
 
treatments; age of 18 or  no electrical current 
of life) 
older 
6 months: 
Pain 2004;112(3):289-
 
 
98. 
 
Co-interventions: 

Adverse effects 
 2.0 
(1.2-2.9) 
 
p<0.0005 
Exclusions: critical 
diclofenac 50mg every 8 
 
 
physical or mental 
hours, if needed and 
Spain 
condition, febrile 
famotidine 20mg every 12 
Similar results for all 
condition, systemic 
hors if needed for 
secondary outcomes 
dermatological 
dyspepsia 
(see table 3 below) 
conditions, neoplasms, 
 
allergy to diclofenac, 
 
referred pain from neck 
ITT analysis 
or thorax, rupture of 
Adverse events: 
tendons or bone 
Intervention group: 
fractures, pregnancy, 
litigation, no intention to 
2 fainted during 
participate or follow 
treatment; 3 reported 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
38 

instructions. 
dizziness; 5 bruising at 
puncture site 
 
 
Blinding: patient and 
evaluators blinded to 
5 reported dyspepsia (1 
allocation 
intervention , 4 in 
control group) 
 
 
Dropouts: 10 in both 
groups i.e. 15% at 6 
3 reported anxiety 
months 
reaction (1 intervention, 
2 in control group) 
 
Follow-up: weekly for 
the 7 weeks of 
treatment, then 3 and 6 
months 
 
Characteristics 
(treatment/placebo):  
Mean age: 60/ 59yrs  
% Female: 49 /48% 
Duration of symptoms: 
5.7/6.8 months 
Additional data: marital 
status, education, 
working, exercise, 
diagnosis, location 
pain… 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
39 

Setting: primary care 
Study type: RCT 
 
Quality: SIGN 1+ 
 
Comments: Well conducted RCT. Randomisation method and allocation concealment good. Power calculation done. 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
40 


A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
41 

 
Reference and study 
Participants 
Intervention/comparison
Outcome measure 
Results/effect size 
Conclusions 
design 
Lathia AT, Jung SM, 
n= 31 
1.  Traditional 
SPADI (Shoulder Pain  Change from baseline 
“Acupuncture may be 
Chen LX.  
acupuncture: 
& Disability Index) 
SPADI score (see 
an alternative and 
• 11 
traditional 
individualized 
Table 2 below): 
adjunctive treatment 
 
acupuncture 
acupuncture 
reported that after 6 
• 9 
standardised 
treatment according 
weeks treatment the 
to help improve pain 
Efficacy of 
acupuncture 
to the approaches 
traditional and 
and function in 
acupuncture as a 
•  11 sham acupuncture 
established by TCM; 
standard groups 
patients with chronic, 
treatment for chronic 
 
at  each session, the 
showed at clinically 
non-rheumatologic 
shoulder pain.  
patient was 
shoulder pain.” 
Inclusion: ≥18 years old; 
significant* change in 
evaluated, and 
 
SPADI score ≥30; shoulder 
SPADI scores from 
different treatment 
 
pain ≥8 weeks; 
baseline  
points were chosen 
J Altern Complement 
acupuncture naïve; either 
according to the 
Reviewer’s conclusion: 
 
Med 2009;15(6):613-8. 
no previous treatment or 
patient’s symptoms; 
Statistically and 
 
failed conventional 
the points used varied 
Treatment Effect (see 
treatment ≥1 month prior to 
between patients and 
clinically significant 
Table 3 below)
reduction in SPADI 
USA 
enrolment 
between treatment 
sessions for each 
Difference in mean 
score after 6 weeks 
 
patient. 
SPADI score (95%CI) 
treatment for both the 
 
from sham 
traditional acupuncture 
Exclusion: inflammatory or 
acupuncture group: 
group and standard 
infectious arthritis; shoulder  2.  Standardised 
acupuncture group 
fracture; stroke; pregnancy; 
acupuncture: 
 
compared to sham 
any corticosteroid 
treatment based on 
acupuncture. The 
injections in last 3 months 
fixed, standard point 
Pain 
effect size was similar 
protocols.; 7 
for both the traditional 
 
acupuncture points 
1.  -16.2 (-2.7, -29.7) 
acupuncture group 
relevant to shoulder 
p=0.021 
Blinding: subjects blind to 
and standard 
pain were used and 
2.  -17.2 (-4.9, -29.6) 
intervention; SPADI 
acupuncture groups. 
remained the same 
p=0.009 
questionnaire investigator 
for each session. 
 
blind to allocation; 
 
acupuncturists not blinded 
Disability 
3.  Sham acupuncture: 
1.  -11.6 (-2.5, -20.6) 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
42 

to allocation 
carried out with sham 
p=0.015 
acupuncture needles 
2.  -10.6 (1.1, -22.3) 
 
and the same points 
p=0.073 
as the standard point 
 
Allocation concealment: 
acupuncture group. 
not reported 
 
Total SPADI 
 
Subjects in each group 
1.  -13.8 (-3.0, -24.7) 
p=0.015 
Drop-outs: 3 in sham group  received the relevant 
2.  -13.9 (-3.3, -24.5) 
failed to complete 
acupuncture treatment 
p=0.013 
intervention because of 
twice per week for 6 
 
time constraints (2) or 
weeks.  
increased pain (1) i.e. 10%   
Results from the 6 
drop-out rate; only 8 
month follow-up were 
participants filled out 
For the acupuncture, 
only available foe 8 
SPADI questionnaire at 6 
subjects, of which, 
months i.e. 74% ‘drop-out’ 
8–16 single-use, 
only 2 reported 
rate 
disposable, sterile 36-
improvement in pain 
gauge needles were 
and disability since last 
 
used and were left in 
treatment (no figures 
Follow-up: 6 weeks (end of  place for 20 minutes. 
reported) 
treatment) and 6 months 
Each session lasted 
 
 
about 30 minutes. 
*Clinically significant 
Characteristics 
 
worsening in shoulder 
(traditional/standard/sham):
pain and function is an 
Co-interventions: any 
increase of ≥10 points. 
Mean age: 62/65/59 yrs 
medications were to be 
continued and not 
Men: 100/100/73%*  
changed for 3 months 
Duration pain: 48/28/51 
prior and during study 
months 
 
Diabetes: 18/51/30% 
Unilateral pain: 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
43 

100/78/70% 
 
Setting: Secondary care 
 
*[p=0.05] 
Study type: RCT 
 
Quality: SIGN 1- 
 
Comments: Small study in mainly male veterans with no power calculation. Long-term follow-up severely limited. Cannot entirely rule out bias from non-
blinding of acupuncturists to allocation group. The significance of a similar effect size for both the traditional and standard acupuncture groups is unclear. 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
44 

 
Reference and study 
Studies 
Intervention/comparison
Outcome measure 
Results/effect size 
Conclusions 
design 
Green S, Buchbinder R,  N=9 ( 
Intervention: ‘Traditional’ 
See Table below 
See Table below for 
“Due to a small number 
Hetrick S.  
or ‘classic’ acupuncture 
summary of results 
of clinical and 
 
methodologically 
 
 
diverse trials, little can 
Inclusion: All RCTs or 
be concluded from this 
Acupuncture for 
quasi-randomised 
Length of treatment: 
review. There is little 
shoulder pain.  
controlled trials; adults 
variable 
evidence to support or 
>16yrs; shoulder pain 
 
refute the use of 
or disorder >3 weeks 
 
acupuncture for 
Cochrane Database 
 
Comparison: Placebo 
shoulder pain although 
Syst Rev 
(Berry 1980; Kleinhenz 
there may be short-
2005(2):CD005319. 
Exclusion criteria: a 
1999; Moore 1976); 
term benefit with 
history of significant 
ultrasound & steroid 
respect to pain and 
 
trauma or systemic 
injection (Berry 1980); 
function.” 
Australia 
inflammatory conditions  nerve block (Lin 1994); 
such as rheumatoid 
mobilisation (Romali 
 
 
arthritis, polymyalgia 
2000); exercise (Sun 
Reviewer’s conclusion: 
rheumatica and 
2001); Trager (Dyson-
Included studies: Berry 
fracture, hemiplegic 
Hudson 2001) 
One small study (n=35) 
1980 (n=60); 
shoulders, 
showed that exercise 
Ceccherelli 2001 
postoperative and peri-
 
and acupuncture 
(n=44); Dyson-Hudson 
operative shoulder pain 
together was more 
2001(n=20); Kleinhenz 
and pain in the 
NB: Ceccherelli 2001 
efficacious than 
1999 (n=52); Lin 1994 
shoulder region as part 
compared deep with 
exercise alone for the 
(n=150); Moore 1976 
of a complex myofacial 
shallow acupuncture, and 
treatment of adhesive 
(n=42); Romoli 2000 
neck/shoulder/arm pain  Yuan 1995 compared 
capsulitis both post-
(n=24); Sun 2001 
acupuncture with sites 
intervention and at 20 
(n=35); Yuan 1995 
 
determined by TCM 
weeks. 
(n=98) 
compared to the 
Databases: MEDLINE, 
distribution of Jing-Luo 
The results from the 
EMBASE, CINAHL, 
rest of the studies are 
Science Citation Index 
 
conflicting or mixed, for 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
45 

 
Co-interventions: see 
example, One study 
summary table below 
(n=52) found that 
Methodological 
acupuncture was more 
assessment: 
efficacious than 
descriptive (including 
placebo in improving 
appropriate 
the Constant-Murley 
randomisation, 
score for rotator cuff 
allocation concealment, 
disease at 4 weeks and 
blinding, number lost to 
4 months. This is in 
follow up and intention 
contrast with another 
to treat analysis), 
study (n=60) that found 
quantitative scoring for 
that acupuncture was 
allocation concealment 
less efficacious than 
only 
placebo for rotator cuff 
disease when 
 
measuring treatment 
No meta-analysis due 
‘success’. However, 
to clinical heterogeneity 
because these two 
studies used different 
 
outcomes, they ability 
to directly compare 
Fixed effects model 
them is limited, at least. 
 
Study type: Systematic review 
 
Quality: SIGN 1++ 
 
Comments: Well conducted SR with narrative synthesis. The paper by Moore (1976) was not used to construct forest plot. Nine studies of varying 
methodological quality, most with small numbers of participants. Heterogeneity of populations, interventions, comparators and outcomes. 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
46 



Active flexion  MD = -13.13 (-39.79, 13.53)  
favours mobilisation 
Active abduction  MD = -14.37 (-49.94, 21.20) 
favours mobilisation 
Sun 2001  
Acupuncture vs. exercise 
Constant12 (post-intervention):  MD = 9.20 (0.54, 17.86)  
acupuncture 
(adhesive capsulitis) 
Constant (20 weeks):  MD = 9.40 (0.52, 18.28) 
acupuncture 
Dyson-Hudson 2001 
Acupuncture vs. Trager 
Wheelchair index13 (post- MD = 1.70 (-21.91, 25.31)  
favours Trager 
(general shoulder pain) 
intervention):  MD = 16.00 (-9.03, 41.03) 
favours Trager 
Wheelchair index (5 weeks): 
Ceccherelli 2001 
Deep vs. shallow 
McGill Pain14 (post-intervention):  MD = -10.31 (-15.44, -5.18)  
deep 
(general shoulder pain) 
acupuncture 
McGill Pain (3 months):  MD = -8.00 (-12.20, -3.80) 
deep 
Yuan 1995  
Traditional vs. Jing Luo 
Recovery:  RR = 1.50 (1.08, 2.09) 
Jing Luo 
acupuncture 
(peri-arthritis) 
 
                                                
12 Constant-Murley Score (measure of shoulder function) 
13 Wheelchair Users Shoulder Pain Index (WUSPI) 
14 McGill Pain Questionnaire 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
48 

 
Reference and study 
Participants 
Intervention/comparison
Outcome measure 
Results/effect size 
Conclusions 
design 
Johansson K, 
n=85  
Intervention: 
Constant-Murley Score  Individual score 
“The results suggest 
Adolfsson L, Foldevi M.  
‘standardised’  
changes not reported 
that acupuncture is 
•  44 in acupuncture 
acupuncture at 4 points 
 
more efficacious than 
 
group 
(10 sessions)* + home 
 
ultrasound in patients 
•  41 in ultrasound 
Adolfsson-Lysholm 
exercise programme 
with impingement 
Effects of acupuncture 
group 
Shoulder score 
Combined score 
syndrome.” 
versus ultrasound in 
 
 
showed larger change 
patients with 
 
(p=0.045) at all 4 time 
 
impingement 
Inclusions: 30 - 60yrs of 
Length of treatment: 5 
points for acupuncture 
UCLA End-Result 
syndrome: randomized  age; ‘typical’ history of 
weeks  
Reviewer’s conclusion: 
Score 
clinical trial.  
shoulder impingement; 
 
positive Neer 
 
 
Acupuncture no more 
 
No differences were 
impingement test; ≥2 
effective than 
Comparison: 
found across the 4 time 
months duration; 3 of 4 of 
Combined Score of all 
ultrasound on ITT 
Physical Therapy 
standardised ultrasound 
points when ITT 
analysis. 
2005;85(6):490-501. 
Hawkins-Kennedy 
above scales 
analysis 
impingement sign, Jobe 
(10 sessions)* + home 
 
supraspinatus test, Neer 
exercise programme 
impingement sign or 
Sweden 
painfull arc between 60 
  
and 120° active 
abduction 
Co-interventions: unclear, 
but “additional” pain 
 
medication reported 
Exclusions: X-ray 
 
findings of malignancy, 
G-H joint OA, bony 
*twice weekly for 5 weeks 
spurs/osteophytes 
decreasing subacromial 
space; polyarthritis, 
rheumatoid arthritis, 
fibromyalgia; history of 
surgery, fracture or 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
49 

dislocation in shoulder; 
history/present instability 
any shoulder joint; 
suspicion of frozen 
shoulder; cervical spine 
problems; previous 
ultrasound or 
acupuncture for same 
problem; steroid injection; 
ruptured rotator cuff 
clinically; acute 
subacromial bursitis; 
communication difficulty 
 
Dropouts: none post-
treatment; 3.5% (2 
acupuncture group/1 
ultrasound group) at 3 
months; 5.9% (0/2) at 6 
months; 12.9% (2/4) at 
12 months 
 
Follow-up: immediately 
post-intervention; 3, 6 & 
12 months 
 
Blinding: observer 
blinded 
 
Characteristics 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
50 

(acupuncture/ultrasound):
Mean age: 49/49 yrs 
% Male: 27/34%  
No significant differeence 
in duration, occupation, 
sick leave taken, 
analgesic use, exercise 
frequency or smoking 
status 
 
Setting: Primary care 
Study type: RCT 
 
Quality: SIGN 1- 
 
Comments: Smallish study with inappropriate analysis showing a (barely) significant result. Complicated selection criteria. Reported “concealed” 
randomisation but only observers blinded. Comparator of dubious therapeutic value. Power calculation done. ITT analysis. 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
51 

 
Reference and study 
Participants 
Intervention/comparison
Outcome measure 
Results/effect size 
Conclusions 
design 
(see fig 1 below) 
He D, Veiersted K, 
N=24  
Intervention: 
Pain (intensity, 
Pain intensity (mean 
“The main finding in 
Hostmark A, Medbo J. 
electroacupuncture, 
frequency) 
intervention vs. mean 
this study was that 
•  14 in acupuncture  acupuncture, & ear 
control group) 
adequate acupuncture 
 
group 
acupressure of 
 
treatment reduced the 
•  10 in control 
standardised points* 
 
intensity and frequency 
Effect of acupuncture 
group 
Pain threshold 
of muscle pain, the 
treatment on chronic 
 
 
At end of treatment: 
degree of headaches, 
neck and shoulder pain 
 
and a number of trigger 
in sedentary female 
Inclusions: Women 
Length of treatment: 3 
15 vs. 36 units; p=0.02 
points became less 
workers: a 6-month and  office workers with 
treatments per week with 
Headache 
 
tender.” 
3-year follow-up study.   chronic (≥3 months in 
a total of 10 treatments 
 
previous year) pain in 
over 2-4 weeks; each 
At 6 months: 
 
 
the shoulder and neck 
treatment lasted 45 min 
Blood variables 
region; 20-50yrs of age; 
24 vs. 36; p=0.15 
 
Pain 2004;109(3):299-
 
 
307. 
pain was severe 
enough to interfere with 
 
Reviewer’s conclusion: 
Comparison: 
 
 
work/spare time 
electroacupuncture 
At three years: 
Some statistically 
activities; 
without any voltage 
 
significant differences 
Norway 
19 vs. 44; p<0.04 
 
applied, acupuncture 10-
in outcomes in favour 
40mm distal to actual 
 
 
of intervention at 
Exclusions: diabetes, 
points, & ear acupressure   
differing timepoints. 
neurological, 
4-6mm below actual 
Frequency of pain: 
However, due to this 
rheumatological or 
points  
being a very small 
other diseases; 
(intervention vs. control  study and questions 
pregnancy, breast-
 
group) 
about validity of 
feeding 
outcome measures and 
Co-interventions: none 
 
variation in results, the 
 
 
reviewer cannot 
At end of treatment: 
exclude that the results 
Dropouts: none 
*16 body acu-points, 6 
not reported 
seen are due largely to 
ear acu-points 
bias. 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
52 

 
 
Blinding: Participants 
At 6 months: 
and examiner blinded to 
allocation; 
24 vs. 31; p=0.18 
acupuncturist not 
 
blinded 
At three years: 
 
19 vs. 46; p=0.003 
Follow-up: 6 months, 3 
yrs 
 
 
Pain threshold (PPT) 
Characteristics 
used algometry on 
(acupuncture/control):  
particular trigger points 
(13); unclear but 
Mean age: 49/45 yrs 
reported “ several 
Sex: all women 
improvements but no 
impairments in the PPT 
Pain duration: 12/12 yrs 
for the [treatment 
group] during the 
Total days pain: 4.3/4.5 
study.” The control 
days per week   
group showed no 
improvements. 
All other variables 
similar 
 
 
Headache 
Setting: secondary care 
no significant difference 
at end of treatment or 6 
months; significant 
difference at 3 years 
 
Blood variables 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
53 

blood platelet 
concentration 
increased by 15% just 
after treatment in 
intervention group; no 
change in control 
group; no change in 
any other measured 
blood variable 
 
Study type: RCT 
 
Quality: 1- 
 
Comments: Small study in women office workers from Norway. Acupuncture carried out by one of the authors. Unsure of validity of questionnaires although 
VAS well accepted. Complex acupuncture intervention. Utility of trigger point pain threshold uncertain. Three subjects (21%) in the intervention group and 5 
(50%) in the control group had other treatments during the 3 year follow-up period. 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
54 



 
Reference and study 
Participants 
Intervention/comparison
Outcome measure 
Results/effect size 
Conclusions 
design 
He D, Hostmark A, 
N=24  
Intervention: 
Pain-related activity 
Pain-related activity 
“Intensive acupuncture 
Veiersted K, Medbo J.  
electroacupuncture, 
impairment at home 
impairment 
treatment may improve 
•  14 in acupuncture  acupuncture, & ear 
and work 
activity at work and 
 
group 
acupressure of 
Work: significant 
several relevant social 
•  10 in control 
standardised points* 
 
difference after 6th & 
and psychological 
Effect of intensive 
group 
10th treatment [p 
variables for women 
acupuncture on pain-
 
 
Quality of sleep 
values not reported], 
with chronic pain in the 
related social and 
and at 3 years [p=0.04] 
neck and shoulders. 
psychological variables  Inclusions: Women 
Length of treatment: 3 
 
The effect may last for 
for women with chronic 
office workers with 
treatments per week with 
 
Degree of irritability & 
at least three years.” 
neck and shoulder pain  chronic (≥3 months in 
a total of 10 treatments 
anxiety 
- an RCT with six 
previous year) pain in 
over 2-4 weeks; each 
Home: significant 
 
month and three year 
the shoulder and neck 
treatment lasted 45 min 
differeence at 3 years 
 
follow up.  
region; 20-50yrs of age; 
[p=0.03] 
 
pain was severe 
 
Degree of satisfaction 
 
 
enough to interfere with 
with life 
Reviewer’s conclusion: 
Comparison: 
work/spare time 
Acupuncture in 
electroacupuncture 
Quality of sleep 
activities; 
 
Some statistically 
Medicine 
without any voltage 
Significant difference 
significant differences 
2005;23(2):52-61. 
 
applied, acupuncture 10-
Frequency of 
after 9th treatment and 
in outcomes in favour 
40mm distal to actual 
depression 
of intervention at 
 
6 months and 3 years 
Exclusions: diabetes, 
points, & ear acupressure 
[p<0.01; p<0.03; 
differing timepoints. 
4-6mm below actual 
Norway 
neurological, 
p<0.03] 
However, due to this 
rheumatological or 
points  
being a very small 
other diseases; 
 
study and questions 
pregnancy, breast-
 
about validity of 
feeding 
Degree of irritability & 
outcome measures and 
Co-interventions: none 
anxiety 
variation in results, the 
 
 
reviewer cannot 
significant difference 
exclude that the results 
Dropouts: none 
between groups after 
*16 body acu-points, 6 
seen are due largely to 
6th treatment and at 6 
 
ear acu-points 
bias. 
months and 3 years 
follow-up [p<0.02; 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
56 

Blinding: Participants 
p=0.02; p=0.02] 
and examiner blinded to 
allocation; 
 
acupuncturist not 
blinded 
Degree of satisfaction 
with life 

 
significant difference 
Follow-up: 6 months, 3 
between groups after 
yrs 
the 8th treatment [p 
value not reported]  and 
 
at 6 months [p<0.01] 
and 3 years [p value 
Characteristics 
not reported] follow-up  
(acupuncture/control):  
 
Mean age: 49/45 yrs 
Frequency of 
Sex: all women 
depression 
Pain duration: 12/12 yrs 
significant difference 
between groups after 
Total days pain: 4.3/4.5 
the 5th-9th treatments 
days per week   
and at 6 months and 3 
All other variables 
years follow-up [all 
similar 
p=0.04] 
 
 
Setting: secondary care 
 
Study type: RCT 
 
Quality: 1- 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
57 

Comments: as He 2004 above 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
58 

 
Reference and study 
Participants 
Intervention/comparison
Outcome measure 
Results/effect size 
Conclusions 
design 
Johansson K, 
n=117 
Intervention: injection 
Pain & shoulder 
No significant 
“Neither treatment was 
Bergstrom A, Schroder 
methylprednisolone + 
function (Adolfsson-
differences between 
superior in decreasing 
K, Foldevi M. 
• 65 
in 
local anaesthetic (if 
Lysholm shoulder 
two groups with respect  pain and improving 
corticosteroid 
requested they could get 
assessment score) 
to pain and function as 
shoulder function” 
 
group 
another injection) 
measured by the 
•  58 in acupuncture 
 
Adolfsson-Lysholm 
 
Subacromial 
group 
 
shoulder assessment 
corticosteroid injection 
 
Health-related quality 
score 
Reviewers’ conclusion: 
or acupuncture with 
Comparator: manual 
of life (EQ-5D) 
Neither treatment was 
home exercises when 
Inclusions: 30-65 yrs 
acupuncture 
 
superior, however, 
treating patients with 
old; presented at one of  (standardised acu-points) 
 
cannot exclude 
subacromial 
5 primary health care 
+ home exercise 
No significant 
Patients’ global 
selection bias and/or 
impingement in primary  centres with shoulder 
programme 
differences between 
assessment of change 
performance bias i.e. 
care--a randomized 
pain and a ‘typical’ 
two groups with respect  baseline characteristics 
clinical trial.  
history of shoulder 
 
to other secondary 
dissimilar, no blinding 
impingement; positive 
outcomes (QoL; global 
to allocation 
 
Neer impingement test; 
Length of treatment: 
assessment) 
≥2 months duration; 3 
acupuncture treatment 
Fam Pract 
was 2x weekly for 5 
 
2011;28(4):355-65. 
of 4 of Hawkins-
Kennedy impingement 
weeks (30 min session);  
 
 
sign, Jobe 
 
supraspinatus test, 
Sweden 
Neer impingement sign 
Co-interventions: none 
or painfull arc between 
reported 
60 and 120° active 
abduction 
 
 
Acupuncture administered 
by 3 physiotherapists; 
Exclusions: X-ray 
corticosteroid injection by 
findings of malignancy, 
3 GPS 
G-H joint OA, bony 
spurs/osteophytes 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
59 

decreasing subacromial 
space; polyarthritis, 
rheumatoid arthritis, 
fibromyalgia; history of 
surgery, fracture or 
dislocation in shoulder; 
history/present 
instability any shoulder 
joint; suspicion of 
frozen shoulder; 
cervical spine 
problems; previous 
ultrasound or 
acupuncture for same 
problem; steroid 
injection; ruptured 
rotator cuff clinically; 
acute subacromial 
bursitis; communication 
difficulty 
 
Dropouts: 123 
participants randomised 
but 6 developed frozen 
shoulder and were 
excluded, therefore 117 
participants; 26 (22%) 
of theses 117 were lost 
to follow-up 
 
Follow-up: 12 months 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
60 

Blinding: treatment 
practitioners (3 GPs 
and 3 physiotherapists) 
 
ITT analysis: those who 
changed treatment 
groups but still 
continued assessments 
as per protocol 
 
Relevant characteristics 
(steroid vs. 
acupuncture): 
Mean age: 50 vs. 51 
yrs 
% women: 27 vs. 26% 
Duration 2-3 months: 
24 vs. 48%   
 
Setting: Primary care 
Study type: RCT 
 
Quality: 1- 
 
Comments: Multi-centre pragmatic RCT; patients and treatment providers not blinded to allocation; 22% lost to follow-up; ITT analysis included those who 
had changed treatment groups but not other ‘drop-outs’; sample size estimation done; 8 participants changed treatment groups (6 from steroid group; 2 from 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
61 

acupuncture group) 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
62 

 
Reference and study 
Studies 
Intervention/comparison
Outcome measure 
Results/effect size 
Conclusions 
design 
Nyberg A, Jonsson P, 
N = 20 studies 
Interventions: 
Pain 
Kleinhenz 1999 
“The result of this 
Sundelin G.  
acupuncture, 
(n=52) 
systematic review 
 
electrotherapy modalities,   
indicates contradictory 
 
exercises, mixed 
• acupuncture 
vs. 
scientific evidence to 
Total number of 
Function 
modalities, changing 
placebo 
support the use of 
Limited scientific 
patients in the studies: 
posture, functional brace 
acupuncture 
acupuncture for pain 
evidence supports the 
not reported 
 
• Constant-Murley  and function in SAIS 
use of conservative 
 
 
score significantly 
patients.” 
treatment interventions 
 
improved in 
for pain and function in 
Length of treatment: 
Quality scores: 
treatment group 
patients with 
Inclusion criteria: RCTs  variable 
post-intervention 
subacromial 
of patients diagnosed 
Kleinhenz 1999, 9/10 
[p<0.014]; pain 
impingement 
with subacromial 
 
[high] 
intensity 
syndrome: 
impingement syndrome 
significantly higher 
Randomized control 
(SAIS) and/or 
Comparison (placebo): 
Vas 2008, 8/10 [high] 
in placebo group at 
trials.  
established signs and 
variable 
3 months follow-up 
symptoms consistent 
Johansson 2005, 8/10 
[p<0.05] 
 
with SAIS; 
 
[high] 
 
conservative 
Physical Therapy 
treatment* (alone or in 
Co-interventions: variable 
 
Reviews 
combination) vs. 
2010;15(6):436-52. 
 
any/placebo/no 
Vas 2008 (n=425) 
intervention;  
 
 
• single 
point 
 
acupuncture + 
Sweden 
physiotherapy vs. 
 
mock-TENS + 
 
physiotherapy 
*interventions other 
Studies included that 
• Constant-Murley 
than surgery, 
were investigating 
score significantly 
pharmacological 
acupuncture: Kleinhenz 
improved 
treatment and steroid 
et al. (1999) , Vas et al. 
[p<0.001] in 
injections 
(2008), Johansson et 
treatment group 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
63 

al. (2005), 
 
post-intervention 
and at 3 month 
Exclusion criteria: 
follow-up 
shoulder diagnoses 
 
other than SAIS; 
multiple diagnoses 
 
 
Johansson 2005 
(n=85) 
Databases used: 
Cochrane library, 
• acupuncture 

PubMed, CINAHL; 
home exercise 
English only 
programme vs. 
ultrasound + home 
 
exercise 
programme 
Description of the 
•  no significant 
methodological 
differences 
assessment of studies: 
between groups on 
as per PEDro scale 
ITT analysis; per 
 
protocol analysis 
acupuncture group 
No meta-analysis 
better [p=0.045] 
 
Qualitative (‘best-
evidence’) analysis 
Study type: Systematic review with qualitative analysis 
 
Quality: SIGN 1++ 
 
Comments: Wide range of interventions; good search and methodology appraisal; qualitative analysis appropriate; heterogeneity not formally reported 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
64 

KNEES 
 
Reference and study 
Studies  
Intervention/comparison
Outcome measure 
Results/effect size 
Notes 
design 
Manheimer, E., K. 
N =16 studies (n = 
Treatment/procedure: 
Pain 
Acupuncture vs. 
Sham-controlled trials 
Cheng, et al. (2010). 
3498 subjects) 
traditional (needle) 
Sham acupuncture 
show statistically 
"Acupuncture for 
acupuncture 
Function 
(all joints): 
significant benefits; 
peripheral joint 
 
however, these 
Symptom severity 
osteoarthritis." 
 
 
benefits are small, do 
Cochrane Database of 
Inclusion: RCTs in any 
 
not meet our pre-
Systematic Reviews(1):  language of at least 6 
Length of treatment:  
Pain 
defined thresholds for 
CD001977. 
weeks observation; 
clinical relevance, and 
people with 
Short-term = 8 weeks 
Short-term: 
are probably due at 
 
osteoarthritis (OA) of 1 
Long-term = 26 weeks 
Standardized mean 
least partially to 
or more peripheral 
placebo effects from 
US, UK, Netherlands, 
difference (SMD)  
joints i.e. knee, hip, or 
 
incomplete blinding.  
China, Korea, Germany  hand; traditional 
= -0.28  
acupuncture compared  Description of comparison 
 
 
to a sham, other active 
(placebo): sham, other 
95% confidence 
Waiting list-controlled 
Included studies: 
treatment or waiting list  active treatment or 
interval (95%CI): -0.45 
control group  
waiting list control group 
to -0.11 
trials of acupuncture for 
16 RCTs (Christensen 
peripheral joint 
1992; Molsberger 1994;   
 
9 trials; 1773 subjects 
osteoarthritis suggest 
statistically significant 
Takeda 1994; Berman 
Exclusions: only OA of 
Co-interventions: 
I2 = 64% 
and clinically relevant 
1999; Fink 2001; 
spine; dry 
diclofenac or placebo 
 
benefits, much of which 
Haslam 2001; Sangdee  needling/trigger point 
tablet in one study 
may be due to 
therapy; laser or 
Long-term (6 months) 
expectation or placebo 
2002; Berman 2004; 
electro-acupuncture 
effects. 
Stener-Victorin 2004; 
with no needle 
SMD = -0.10 
Tukmachi 2004; 
insertion; comparison 
of one form of 
95%CI: -0.21 to 0.01 
Vas 2004; Witt 2005; 
acupuncture with 
Scharf 2006; Witt 2006;  another 
4 trials; 1399 subjects 
Foster 2007; 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
65 

Williamson 2007)  
 
I2 = 0% 
Databases: Cochrane 
 
Central Register of 
Controlled trials, 
Function 
MEDLINE, and 
Short-term: 
EMBASE 
 
SMD = -0.28 
95%CI:  -0.46 to -0.09 
Methodological 
assessment: used the 
9 trials; 1829 subjects 
following criteria 
(adequate sequence 
I2 = 69% 
generation, allocation 
concealment, blinding, 
 
incomplete outcome 
Long-term: 
data addressed, free of 
selective reporting) 
SMD = -0.11 
plus prognostic factors 
similar at baseline, co-
95%CI: -0.22 to 0.00 
interventions avoided 
or similar, compliance 
4 trials; 1398 subjects  
acceptable in all 
I2 = 6% 
groups, timing of 
outcome assessment 
Symptom severity 
similar, and intention to 
treat analysis. 
Short-term: 
 
SMD = -0.29  
Random effects model 
95%CI: -0.50 to -0.09 
9 trials; 1767 subjects  
I2 = 74% 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
66 

Long-term: 
SMD = -0.11  
95%CI: -0.22 to 0.00  
4 trials; 1398 subjects 
I2 = 2% 
 
Acupuncture vs. 
Sham acupuncture 
(Knee OA only)
 
Pain 
Short-term: 
SMD = -0.29 
95%CI: -0.48 to -0.10  
8 trials; 1773 subjects 
 
Long-term: 
SMD = -0.10 
95%CI: -0.21 to 0.01 
4 trials; 1399 subjects 
 
Function 
Short-term: 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
67 

SMD = -0.29 
95%CI:  -0.49 to -0.08 
8 trials; 1767 subjects 
 
Long-term: 
SMD = -0.11 
95%CI: -0.21 to 0.00  
4 trials; 1398 subjects 
 
Symptom severity 
Short-term: 
SMD = -0.29 
95%CI: -0.50 to -0.09 
8 trials; 1767 subjects 
 
Long-term: 
SMD = not estimable 
4 trials; 1398 subjects 
Study type: systematic review with meta-analysis 
 
Quality: SIGN 1++ 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
68 







Total score 

1083 
-0.38 [-0.62, -0.15] 
•  Acupuncture vs. supervised OA education 

250 
-0.46 [-0.71, -0.20] 
•  Acupuncture + physiotherapy vs. physiotherapy 

209 
-0.12 [-0.39, 0.15] 
•  Acupuncture vs. consultation (exercise based 

624 
-0.52 [-0.68, -0.36] 
physiotherapy as a co-intervention) 
 
 
Reference and study 
Participants 
Intervention/comparison
Outcome measure 
Results/effect size 
Conclusions 
design 
Miller E, Maimon Y, 
55 participants 
Intervention: acupuncture 
Knee Society Score 
8 weeks: 
“Adjunctive 
Rosenblatt Y, Mendler 
needles placed in points 
(KSS) [acupuncture vs. 
acupuncture seems to 
A, Hasner A, Barad A, 
• 
treatment arm = 
selected by team of TCM 
sham] 
61.6 vs. 56.8; p=0.15 
provide some added 
et al. Delayed Effect of 
28 
practitioners (see full 
improvement to 
23.7 vs. 24.4; p=0.7 
Acupuncture Treatment 
• 
control arm = 27 
paper for details); needles 
• total 
score 
standard care in elderly 
in OA of the Knee: A 
 
in place for 20mins and 
• pain 
score 
65 vs. 59.7; p=0.23 
patients with OA of the 
Blinded, Randomized, 
• function 
score 
knee.”  
Inclusions: ≥45 yrs; 
manually manipulated 
Controlled Trial. Evid 
 
 
diagnosis of OA knee 
every 5mins 
Based Complement 
 
≥6 months; moderate-
 
Alternat Med 2009. 
 
12 weeks: 
severe pain most days 
Reviewer’s conclusion: 
 
in last month for which 
Comparison: sham 
 
63.54 vs. 53.6; p=0.036  Acupuncture added to 
analgesics were used 
acupuncture (no insertion 
“standard care” may 
Israel 
for at least 1 month 
into skin) at same points 
 
24.0 vs. 21.1; p=0.31 
improve total knee 
as in treatment group at 
score and knee function 
 
 
67.4 vs. 54.7; p=0.01 
same frequency 
at 12 weeks after 
Patient satisfaction 
commencement of 
Exclusions: intra-
 
 
[acupuncture vs. sham] 
therapy but not at 8 
articular steroid 
 
weeks. However, 
injection into knee(s) 
Length of treatment: twice   
cannot rule out bias or 
within 4 weeks; severe 
weekly for 8 weeks 
4.87 vs. 3.75; p=0.005 
the effect of co-
unstable chronic illness 
 
intervention, as this is 
e.g. CHF, CRF, cancer 
 
 
poorly reported. 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
72 

 
Co-interventions: 
Adverse effects  
 
“standard therapy (e.g. 
Dropouts: 25% 
NSAIDs) 
 
10 during treatment 
None reported 
(18%) 
[4 in acupuncture group 
& 6 in control] 
4 lost during follow-up 
(7.7%) 
[3 in acupuncture group 
& 1 in control] 
 
Follow-up: 12 weeks 
 
Blinding: reported as 
being “applied 
successfully” 
 
Relevant 
characteristics: Mean 
age: 70.3yrs 
[acupuncture], 72.2yrs 
[control] 
Sex: 75% women 
[acupuncture], 63% 
[control] Stage of 
disease: not reported 
Co-morbidity: not 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
73 

reported 
Secondary care 
Study type: RCT 
 
Quality: SIGN 1- 
 
Comments: Small study; possible confounders not reported; co-intervention not reported clearly; no controls of therapist behaviour e.g. time spent, 
communication; high drop-out rate; drop-outs may be different from completers i.e. KSS function score lower 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
74 

PAIN 
Reference and study 
Studies 
Intervention/comparison
Outcome measure 
Results/effect size 
Conclusions 
design 
Hopton A, MacPherson  N= 8 systematic 
Intervention: Acupuncture  Pain 
Knee pain only 
“The meta-analyses of 
H. Acupuncture for 
reviews 
all recent systematic 
chronic pain: is 
 
 
 
reviews of acupuncture 
acupuncture more than 
 
for the most commonly 
Comparison: “Sham” 
 
an effective placebo? A 
Kwon 2006 
occurring chronic pain 
See table below for 
acupuncture (variable 
systematic review of 
conditions show that 
more details 
methods15
• Short-term 
 
pooled data from meta-
•  N=2 studies (264 
there is consistent 
analyses. Pain pract 
 
 
subjects) 
evidence that 
2010;10(2):94-102. 
•  SMD = 0.24, 95%  acupuncture is more 
Inclusion: systematic 
Length of treatment: for 
CI: 0.01 to 0.47 
effective than sham 
 
reviews of acupuncture  this review, defined as: 
 
acupuncture for chronic 
and chronic pain (knee, 
osteoarthritis of the 
UK 
back, head) in English 
Short-term = <3 months 
Bjordal 2007 
knee and headache in 
 
with meta-analyses and 
both the short term and 
Long-term = ≥3 months 
statistically pooled data 
All short-term  
longer term.” 
Studies included: 
 
Manual acupuncture 
Furlan 2005; 
 
 
Manheimer 2005; Kwon 
No co-interventions 
Exclusion: reviews of 
•  N=4 studies (746 
“However, the results 
2006; Bjordal 2007; 
reported 
shoulder, neck, elbow 
subjects) 
for back pain are 
Manheimer 2007; White  or leg pain, myofascial 
•  WMD = 1.3, 95% 
mixed.” 
2007; Davis 2008; Sun 
 
trigger point pain, 
CI: -2.7 to 4.7 
2008 
chronic pain from RA, 
Electroacupuncture 
 
circulatory disorders, 
•  N=3 studies (242 
Reviewer’s conclusion: 
cancer or other 
subjects) 
terminal illness; 
•  WMD = 21.9, 
Overall, the evidence 
injection of substances 
95% CI: 17.3 to 
from 4 good quality 
alone e.g. bee venom 
25.3 
systematic reviews 
show that acupuncture 
 
 
reduces pain compared 
                                                
15 including superficial insertion of needles at inappropriate sites and the use of blunt devices to apply pressure without penetration of skin 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
75 

Databases: Medline, 
Manheimer 2007 
to ‘sham’ acupuncture 
Allied & 
both in the short- and 
Complementary 
Short-term  
long-term [NB: except 
Medicine database, 

for the manual 
 N=6 
(1636 
Cochrane library, Web 
acupuncture ‘arm’ of 
subjects) 
of Science, authors’ 

the Bjordal study]. 
  SMD = 0.35, 95% 
database and 
CI: 0.15 to 0.55 
reference list (2005-
 
Long-term  
2008). 
However, the effect 
• N=3 
(1304 
 
size is small with lower 
subjects) 
confidence limits near 
Methodological 
•  SMD = 0.13, 95%  zero. 
assessment: based on 
CI: 0.01 to 0.24 
14 questions derived 
 
 
from the Oxman and 
White 2007 
 
Guyatt index and the 
AMSTAR tool 
Short-term  
 
• N=5 
(1334 
subjects) 
 
•  WMD = 1.54, 
No pooling of results 
95% CI 0.49 to 
from individual 
2.60 
systematic reviews 
Long-term  
• N=3 
(1178 
subjects) 
•  WMD = 0.54, 
95% CI 0.05 to 
1.04 
Study type: Systematic review of systematic reviews 
 
Quality: 1++ 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
76 


Comments: Selected reviews that contained pooled data for meta-analyses from high-quality trials that compare sham and true acupuncture for specific, 
common pain conditions. Well conducted search and methodological assessment. Qualitative analysis of results appropriate. Each SR formally assessed the 
internal validity of each study, applied strict inclusion & exclusion criteria, and tested for heterogeneity. Three of the 4 knee studies conducted a sensitivity 
analysis and considered publication bias.  
 
 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
77 

 
 
Reference and study 
Studies 
Intervention/comparison
Outcome measure 
Results/effect size 
Conclusions 
design 
Madsen MV, Gotzsche 
N=13 studies (3025 
Intervention: acupuncture 
Pain 
Pooled results 
“We found a small 
PC, Hrobjartsson A. 
subjects) 
analgesic effect of 
Acupuncture treatment 
 
 
acupuncture that 
for pain: systematic 
 
seems to lack clinical 
Acupuncture vs. 
review of randomised 
Length of treatment: 1 
relevance and cannot 
placebo acupuncture 
clinical trials with 
Inclusion criteria: all 
day to 12 weeks 
be clearly distinguished 
(see fig 1 below): 
acupuncture, placebo 
trials labelled 
 
from bias. Whether 
acupuncture, and no 
“acupuncture”; any 
 
needling at 
acupuncture groups. 
placebo interventions 
Comparison: both a 
acupuncture points, or 
Bmj 2009;338:a3115. 
used by authors e.g. 
SMD = -0.17  
at any site, reduces 
non-penetrating 
placebo acupuncture and 
a no acupuncture control 
pain independently of 
 
needles, insertion into 
(95%CI: -0.26 to -0.08) 
the psychological 
non-acupuncture 
group 
impact of the treatment 
Denmark 
points; pain measured 
 
13 trials*; 3025 
ritual is unclear.” 
by VAS or another 
subjects 
 
scale; two control 
Co-interventions: all 
 
groups (placebo and no 
I2 = 36% 
Included studies: 
patients were supplied 
acupuncture) 
Reviewer’s 
Melchart 2005; Linde 
with standard care which 
Funnel plot: 
conclusions: Both 
2005; Lin 2002; Sprott 
was analgesics in 13 trials 
 
symmetrical with clear 
meta-analyses show a 
1993; Fanti 2003; 
and physiotherapy in 5 
peak (data not 
statistically significant 
Wang 1997; Witt 2005; 
Exclusion criteria: 
reported) 
benefit with regards to 
Scharf 2006; Foster 
TENS, manual 
pain and moderate 
2007; Molsberger 2002;  acupressure; different 
 
degree of 
Brinkhaus 2006; 
co-interventions in each 
*one trial excluded as 
heterogeneity. Whether 
Leibing 2002; Kotani 
group 
an outlier [Kotani 2001]  this effect is clinically 
2001 
 
i.e. % weight = 0 
significant is debatable, 
 
however, if it reflects a 
Databases: Cochrane 
 
true effect then it is 
Relevant to report: Witt 
library, Medline, 
small. 
Placebo acupuncture 
2005 (OA knee); Scarf 
EMBASE, Biological 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
78 

2006 (OA knee); Foster  Abstracts, and PsycLit 
vs. no acupuncture 
 
2007 (OA knee) 
(see fig 2 below) 
 
 
 
Assessment of bias: 
adequate allocation 
SMD = -0.42  
concealment; patients 
were blinded; drop-outs 
(95%CI: -0.60 to -0.23) 
<15% [if all 3 present 
12 trials; 3025 subjects 
then low risk of bias]; 
funnel plot to assess 
I2 = 66% 
small sample size bias 
Funnel plot: broad peak 
 
as large trials reported 
both large and small 
Meta-analysis 
effects of placebo; 
 
small trials tended to 
report small effects 
Fixed or variable 
effects: “used a random 
 
effects model if 
Individual results (All 
heterogeneity 
OA knee)  
existed (P<0.10) and a 
Acupuncture vs. 
fixed effect model 
placebo acupuncture 
otherwise.” 
 
Witt 2005 
-0.52 (-0.80 to -0.23) 
 
Scharf 2006 
-0.13 (-0.28 to 0.02) 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
79 

 
Foster 2007 
-1.66 (-2.34 to -0.98) 
 
Placebo acupuncture 
vs. no acupuncture 
 
Witt 2005 
-0.68 (-1.02 to -0.34) 
 
Scharf 2006 
-0.42 (-0.58 to -0.27) 
 
Foster 2007 
-0.21 (-0.47 to 0.06) 
 
Study type: Systematic review with meta-analysis 
 
Quality: SIGN 1- 
 
Comments: Due to moderate levels of statistical heterogeneity i.e. I2 = 25-75%, probably not appropriate to conduct a meta-analysis. In addition, considerable 
heterogeneity is present in the populations, treatments and outcome measures. The results need to be interpreted in this light.  
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
80 


 
 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
81 


 
 
 
 
 
A c c i d e n t   C o m p e n s a t i o n   C o r p o r a t i o n                                                                                                       Page   
82