Location Location Location. Acupuncture and chronic shoulder pain – CAM or Sham?

Having written a number of posts on acupuncture (see here, here, and here) I guess my particular biases are reasonably apparent. So imagine my surprise when a large RCT published in the journal “Pain” reports a significant and substantial effect of Chinese acupuncture in comparison with sham acupuncture or conventional orthopaedic therapy for chronic shoulder pain. First glance demonstrates a well performed trial that boasts power and rigour. It is randomised, allocation is concealed, the outcome measures are appropriate, the data analysis appears sound. Time for a reappraisal?

Perhaps, but this finding is at odds with the recent big trials of acupuncture for pain which consistently show little to no difference between real acupuncture and sham acupuncture, regardless of the type of sham used (see links in the previous posts).

Reading more closely it is always interesting to check how the authors got around the problematic issue of devising a reasonable sham. This study, like many recent others used shallow but penetrating needles at non-acupuncture points [points that are not considered to be beneficial in Traditional Chinese Medicine (TCM).] In terms of assessing the value of odd theories like meridians this is not unreasonable but it does leave the authors with some interpretive challenges. For example if you don’t find a difference between active and sham do you conclude against acupuncture, or just conclude that acupuncture (or rather needle penetration) works, but not due to the folklore of TCM. Such interpretations are popular these days, but controversial.

Still, there was a clear change between real and sham acupuncture in this study so that particular problem does not apply. So what could be going on? Here is a figure showing the location of the acupuncture points used in both groups.

Acupuncture points used in Research

Acupuncture points. Verum acupuncture: one to three locus dolendi (Ahshi) points; local and distal points according to the channel and the individual location of the pain: ventral – Lung 1, 2; ventrolateral – Large Intestine 4, 11, 14, 15; lateral – Sanjiao 5, 13, 14; dorsal – Small Intestine 3, 9, depth of needle insertion 1–2 cm. Sham acupuncture: 4 needles above the medial part of the left and right tibia, with depth of needle insertion less than 5 mm. Figure 1 from Molsberger AF, Schneider T, Gotthardt H, Drabik A. PAIN 2010 Oct; 151(1): 146-154. This figure has been reproduced with permission of the International Association for the Study of Pain® (IASP®). This may not be reproduced for any other purpose without permission.

Does anything particular jump out at you? If not, take a closer look. For real acupuncture (for shoulder pain, remember) the needles are spread around the shoulder and upper arm. In the sham group they are placed in the lower leg. Now put yourself in the patient’s position. If you are suffering with chronic shoulder pain which treatment would seem more credible to you? Treatment credibility is at the heart of the placebo effect and I think this represents a real issue. Unfortunately participants’ perceptions of treatment credibility were not measured (they often are these days) so we can’t quantify their effects in this study.

Now consider that the needles were inserted twice to four times as deep in the real acupuncture group, and twiddled by the acupuncturist throughout the treatment to elicit sensation (but not twiddled in the sham group) and it strikes me that you might have a much more convincing piece of therapeutic theatre. Then add the problem that it is clearly not possible to blind the therapists using this design and we can see that the real acupuncture group had a number of paths towards a bigger treatment effect that have nothing to do with the theories of Chinese Medicine.

There is the argument that the lasting improvements seen (after 3 months) and the improvement over conventional therapy add weight to the authors’ conclusions. But a convincing placebo might be enough to put patients on a path to lasting improvement and with most patients having had their pain for an average of around 10 months it is likely that many might have already tried and failed with conventional therapy, introducing a nocebo effect in this group. Participants in this study were acupuncture naïve, making this a novel therapy, and would have been signing up to take part for the chance of receiving acupuncture. We know that the kind of folk who volunteer for acupuncture trials seem to have generally high expectations of acupuncture to drive that placebo response. Unfortunately expectations were not measured but together these factors might add up to a heady brew and accentuate the differences between the groups.

Ultimately sham controls, if not indistinguishable (which would be the ideal) need to at least be as convincing as the real treatment. It is fair to acknowledge that these results could possibly be a reflection of the superiority of Chinese acupuncture over sham needling. Since the credibility of the sham, the integrity of patient blinding and the expectations of the participants were not measured we can’t say much for sure. But given the bulk of the existing evidence, the low probability that a pre-medieval argument from authority (the “ancient wisdom” plea)  is actually correct and the issues discussed here it would seem more likely to me that a different interpretation might be at hand.

About Neil

Neil O’Connell is a researcher in the Centre for Research in Rehabilitation, Brunel University, West London, UK. He divides his time between research and training new physiotherapists and previously worked extensively as a musculoskeletal physiotherapist. He also tweets! @NeilOConnell

Neil is currently fighting his way through a PhD investigating chronic low back pain and cortically directed treatment approaches. He is particularly interested in low back pain, pain generally and the rigorous testing of treatments. He also tends to get all geeky over controlled trials.


Molsberger AF, Schneider T, Gotthardt H, & Drabik A (2010). German Randomized Acupuncture Trial for chronic shoulder pain (GRASP) – a pragmatic, controlled, patient-blinded, multi-centre trial in an outpatient care environment. Pain, 151 (1), 146-54 PMID: 20655660

O’Connell NE, Wand BM, & Goldacre B (2009). Interpretive bias in acupuncture research?: A case study. Evaluation & the health professions, 32 (4), 393-409 PMID: 19942631

Linde K, Witt CM, Streng A, Weidenhammer W, Wagenpfeil S, Brinkhaus B, Willich SN, & Melchart D (2007). The impact of patient expectations on outcomes in four randomized controlled trials of acupuncture in patients with chronic pain. Pain, 128 (3), 264-71 PMID: 17257756


  1. I find no value in this paper. There is simply no understanding that points distal to the shoulder such as ST 38, GB 34, TH 5, and LI 4 can be effective in treating shoulder pain, nor of the concept the six meridians pass through the shoulder apparatus.
    Furthermore, one of the sham leg points is close to ST 38, which is a major point for shoulder pain. So this confuses the sham protocol.
    The real issue is that the paradigms of Chinese medicine and acupuncture are so completely different and disctinct from those of western allopathic medicine that these kinds of studies may not be possible. And so it goes with homeopathy and perhaps other fields of medicine.

  2. Thought this study might interest you guys.

    Evaluating the efficacy of graded in vivo exposure for the treatment of fear in patients with chronic back pain: a randomized controlled clinical trial.
    Woods MP, Asmundson GJ.
    Pain. 2008 Jun;136(3):271-80. Epub 2007 Aug 22.

    Department of Psychology, University of Regina, Regina, SK, S4S 0A2, Canada. marcwoods@hotmail.com

    Psychological treatments for chronic pain, particularly those based upon cognitive behavioural principles, have generally been shown to be efficacious. Recently, a treatment has been developed based upon the fear-avoidance model of chronic musculoskeletal pain, which suggests chronic pain can be relieved by exposing the individual to movements and tasks that have been avoided due to fear of (re)injury. This graded in vivo exposure treatment has been found to be beneficial in case studies. The present investigation utilized a randomized controlled trial method to assess the effectiveness of graded in vivo exposure relative to other conditions. Forty-four chronic low back pain patients were randomly assigned to graded in vivo exposure, graded activity, or a wait-list condition. While only trend differences were observed for pain-related disability, patients in the graded in vivo exposure condition demonstrated (a) significantly greater improvements on measures of fear of pain/movement, fear avoidance beliefs, pain-related anxiety, and pain self-efficacy when compared to those in the graded activity condition, and (b) significantly greater improvements on measures of fear-avoidance beliefs, fear of pain/movement, pain-related anxiety, pain catastrophising, pain experience, and anxiety and depression when compared to those in the wait-list control condition. Additionally, patients in the graded in vivo exposure condition maintained improvements in these areas at one month follow-up. Implications of these findings for the treatment of individuals with chronic low back and other pain conditions are discussed.

  3. Perhaps I misunderstood, Ian. It seemed like you described the ScienceBasedMedicine.org article as “negative and dismissive,” and that is what I responded to.

  4. ian stevens says

    Paul ,thanks for your comments.
    I am not dismissive of Neil’s perceptive commentary and review of the acupuncture literature (which probably could be applied to almost any paper on the topic as far as I am aware) and as you say much of ‘folk’ medicine .
    I am not defending acupuncture, merely pointing out that it is perhaps the setting , therapeutic surroundings/relationship as discussed by Oakely that is worth considering.
    What does interest me is how many people in the medical and therapy professions seem to elevate acupuncture ( in terms of the evidence base) as a modality and why culturally acupuncture has rode a wave of popularity …I enjoyed reading this review of acupuncture therapy and I feel that the overall tone and message is worthy of consideration.http://www.acampbell.ukfsn.org/acupuncture/articles/touch.html

  5. ian stevens says

    why keep doing trials ??
    Maybe due to research careers being dependent on producing more studies?
    This is a fairly negative and dismissive view of acupuncture .
    However this review http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1892509/ of Ann Oakley’s Fracture – (a personal narrative by a Sociology professor of recovery and rehabilitation) may offer a better account of why people benefit and visit acupuncturists and other therapists?

    Paul Ingraham Reply:

    Ian: Most of the folk medicine still being practiced in the world richly deserves negativity and dismissal. If you want to defend acupuncture, you will have to attempt to address the evidence, and not merely snipe at the tone of those who are already doing so.

  6. Christopher, while I will grant you that it always theoretically possible that a different/better intervention might yield better results than what is done in the carefully controlled and simplified conditions of a trial, that also happens to be the inevitable defensive response when a popular treatment is under scientific siege. No matter how many studies are done, proponents simply move the goal post and claim that an even better testing method is required to reveal the elusive therapeutic effect. It becomes ridiculous after a couple of decades. Many modern trials of acupuncture have been more than good enough that they ought to at least show some clear benefit, even if the intervention isn’t ideal. Or is acupuncture so unreliable and flaky that its benefits consistently vanish completely if it isn’t delivered under perfect conditions?

    Either way, acupuncture fails to impress.

  7. Neil O'Connell says

    Hi Christopher,

    Thanks for your comments. I agree that a non-penetrating sham is a better option and I think sham laser acupuncture with all of its bells and whistles to enhance the placebo effect is promising.

    There are a number of problems with incorporating traditional Chinese Medicine principles:

    1. The authors of this study would contend that they have – which illustrates an important point – there is little consensus on what those principles are in practice. This is neatly illustrated by these neat studies by Barker Bausell : http://www.ncbi.nlm.nih.gov/pubmed/14727501

    There was little agreement between different Chinese medicine practitioners st all on diagnosis and treatment over 2 cohorts of patients.

    2. TCM is a system of medicine built on rather prescientific foundations. For its basic tenets to hold true we would have to reappraise our understanding not just of the human body but of the laws of physics etc etc. The last coulple of hundred years have seen stratospheric strides in our scientific understanding of health and disease. It seems unlikely that this body of knowledge can be true whilst simultaneously a completely separate and contradictory explanation is true.

    3. Finally it is worth reflecting that across trials of acupuncture for all conditions there has been little value of utilising TCM points and meridian theory as demonstrated in this review: http://www.ncbi.nlm.nih.gov/pubmed/18538996

  8. Christopher Lo says

    Physiotherapist from Hong Kong who has 1 year Chinese Acupuncture training and two years practicing experience.

    I read certain number of acupuncture article in physiotherapy journals. The main problem I find is the authors did use Chinese acpuncture approach with Chinese Medical theory. However, none of them concern about using Chinese medical diagnosis but just a broad term of Chronic shoulder pain / chronic low back pain.

    Few points I can comment on the results of these researches:
    For placebo group – sham acupuncture, high expectation, passive treatment, distraction from pain in other sites do act on central inhibitory mechanism to decrease chronic pain. I wounder sham acupuncture is really a good ‘placebo’ intervention?

    One study design was using shame lazer on acu-point, which I think is the best placebo I have seen.

    For the intervention group:
    As mention above, no clear classification was made whether the conditions are mechanical pain or inflammatory conditions. Heat diseases or cold diseases (Chinese medical classification) etc.

    The acupuncture points were used as a recipe protocol, when you look into acupuncture studies conducted in mainland China (of course I know their methodology is ……….). More flexible in the choice of acupuncture points , better the treatment effect is.

    At least I want to suggest the researchers if possible, investigate a bit into the Chinese medical classification and think about adding it in inclusion criteria.

    Christopher Lo
    Msc in Muscuoskeletal and Sports Physiotherapy

  9. Well done! A significant criticism of a study that certainly seems sound at first glance. But a sham has to be convincing to be compared to a treatment, and this sham clearly was not! Um, dude, that’s my LEG … I may not know much about acupuncture, but I know that sticking needles in my leg probably isn’t a treatment for my shoulder! D’oh.

    Danged if I can remember which one, but I’m positive I noticed the same problem with another acupuncture study.

    Neil O'Connell Reply:

    Thanks Paul, Lots of studies use the shallow non “TCM” point sham but few quite so far from the area of pain! It’s an odd choice of sham in my view and ideally a non-penetrating sham over the same points is a better option, although not without its own difficulties. But as I’ve said here before, we have lots of data on acupuncture for pain and it is resolutely unconvincing. Why keep doing trials?