Sep 03 2013

The 1996 WHO Acupuncture Report

One interesting effects of social media is that news items can have a second life – someone posts an item from years ago on their Facebook page, for example, and it makes the rounds as if it’s something new. Those of us who spend time analyzing science news reports and correcting misreporting or misleading information often refers to such items as zombies. They keep rising from the dead and have to be staked all over again.

The World Health Organization (WHO) 1996 report on acupuncture has recently risen from the grave and is haunting online acupuncture discussions. The fact that the data on which this report is based is over 17 years out of date does not seem to be a problem for those referencing it. The allure seems to be the apparent authority of the WHO.

The WHO report is generally positive toward acupuncture, reviewing clinical evidence and listing many conditions for which acupuncture is apparently effective. The report, however, is also deeply flawed.

The underlying problem with the report is that it was put in the hands of highly biased acupuncture proponents. The acknowledgements are very revealing:

Special thanks are due to Dr Zhu-Fan Xie, Honorary Director of the Institute of Integrated Medicines, First Hospital of Beijing Medical University, China, who drafted, revised and updated this report. Further, Dr Xie made numerous Chinese language documents available in English. We also thank Dr Hongguang Dong, Geneva University Hospital, Switzerland for providing additional information.

The report itself reads like a piece of propaganda, not a critical scientific review. This begins right from the first line, which is nothing but an appeal to antiquity:

Over its 2500 years of development, a wealth of experience has accumulated in the practice of acupuncture, attesting to the wide range of diseases and conditions that can be effectively treated with this approach.

Completely absent from the report is anything critical. The section on safety does not even mention potential adverse effects (including serious effects, such as infection or pneumothorax).  There are howlers throughout the report, including the insistence that acupuncture is as effective as morphine in the treatment of pain.

The fact that the architects of this review are all Chinese and clearly relied heavily upon Chinese research is relevant because of the documented bias in the Chinese literature. A 1998 review found that 99% of papers published on Acupuncture in China were positive. This defies the odds, even for a highly effective treatment, which we know acupuncture is not. The only explanation for such results is extreme researcher and publication bias. Any systematic review that relies on Chinese acupuncture studies without taking this into consideration is therefore fatally flawed.

The WHO review lists conditions for which acupuncture has been proven, those for which there is evidence of efficacy but further research is needed, those for which there is some evidence, and those for which it is reasonable for an experienced practitioner to try acupuncture. Missing is a list of conditions for which acupuncture has been shown not to work – a curious omission, but consistent with the pseudoscientific advocacy approach of this report.

Under the proven category (meaning no further research is necessary, in the opinion of the authors) there are many pain conditions but also stroke. The authors concluded that the evidence in 1996 was sufficient to conclude that acupuncture is effective for stroke. Meanwhile, a 2005 Cochrane review found:

Five trials (368 patients) met the inclusion criteria. Methodological quality was considered inadequate in all trials.

And concluded:

Currently there is no clear evidence on the effects of acupuncture on subacute or chronic stroke. Large, methodologically-sound trials are required.

And another 2005 review for acute stroke:

Acupuncture appeared to be safe but without clear evidence of benefit. The number of patients is too small to be certain whether acupuncture is effective for treatment of acute ischaemic or haemorrhagic stroke. Larger, methodologically-sound trials are required.

Apparently the authors of the WHO study believe that a small number of methodolically flawed studies with mixed results is sufficient to conclude that a treatment works, without even the need for further research.

The same is true of all the other alleged indications – the research at the time was methodologically poor, and those studies that were better designed tended to be negative. Now we have 17 years more of research, often with more rigorous methodology, and that is what we find.

In a recent analysis by Colquhoun and myself of acupuncture for pain we argued (convincingly, in my opinion) that the evidence is consistent with the conclusion that the pain relieving effects of acupuncture are insignificant to nonexistent.

What the scientific evidence shows is that acupuncture is highly implausible and it does not work – this despite a documented favorable bias in the literature.


The 1996 WHO report on acupuncture is a highly flawed pro-acupuncture piece of propaganda, not a scientific review of evidence. It was worthless in 1996, and now it has the added burden of being outdated.

Systematic reviews of acupuncture (themselves highly variable in technique) tend to show that acupuncture does not work or that there is simply insufficient evidence to conclude that there is any effect.

A best evidence analysis, however, as reflected in the paper by myself and Colquhoun, shows that the most rigorous studies in acupuncture consistently show a lack of specific benefit from needle placement or insertion. This pattern of evidence is consistent with an intervention with no specific effect surrounded by the usual noise of placebo effects, researcher bias, and publication bias.


6 responses so far