Ernst et al. claimed “acupuncture is neither a safe procedure nor superior to placebo.”

One of the last review articles on acupuncture research from Professor Edzard Ernst, prior to the announcement of his retirement, was met with broad dissatisfaction from the acupuncture research community, with concerns raised that he had presented a biased picture of the state of the science. A summary of the published letters rebutting Professor Ernst’s conclusions is presented below. We urge acupuncture researchers and practitioners to read Professor Ernst’s article in Pain (as well as the full complement of letters to the editor published in a subsequent issue of the journal (Witt et al.; Manheimer and Berman; Karst; Baumler; Usichenko; Henke 2011). For the sake of brevity, we summarize the letters to the editors.

A letter by Witt and colleagues found the article “highly misleading” and criticized the authors’ choice of evidence from different levels within the evidence hierarchy for different parts of the paper. In particular, the writers of the letter criticize the use of case studies to determine the safety of acupuncture and cite several larger-scale studies that had different results.

A letter by Manheimer and colleagues criticized Ernst et al.’s methods of data analysis and evaluation as neither systematic, reproducible, nor transparent and found Ernst et al.’s evaluation criteria for assessing the quality and results of systematic reviews to be poorly defined. The writers of the letter also suggest that Ernst et al. succumbed to presentation bias, presenting data in the Discussion section which were not included in the review.

A letter by Matthias Karst notes that the paper misinterprets one of the cited cases, and suggests that the paper’s other conclusions should be suspect, and also warns of a possible “efficacy paradox” which can lead to false positives when testing complex interventions by means of randomized controlled trials.

A letter by Baumler and colleagues pointed out several inconsistencies and accuses the authors of failing to incorporate basic scientific methodology. Specifically, the writers of the letter find the paper’s aim vague, study eligibility criteria absent, and (like Witt et al.) the choice of case reports in the assessment of acupuncture risk questionable. Like Manheimer et al., the letter-writers found Ernst et al.’s methods of determining the quality and outcome of each review unclear. The letter-writers determine that the paper is “not … based on a careful and scientific analysis of the literature.”

A letter by Usichenko and colleagues, like Witt et al. and Baumler et al., criticized the authors’ use of case studies to determine the adverse effects of acupuncture.

A letter by Christian Henke points out the relatively high instance of drug-related adverse events compared to acupuncture-related adverse events and criticizes Ernst et al.’s choice to include the conclusions of various acupuncture experiments in their discussion, despite the fact that conclusions of papers do not reflect statistical outcomes. The letter-writers also note that some important studies were overlooked.

Finally, a letter by James Mooney criticizes the authors’ addressing of the broad spectrum of acupuncture as little more than a single treatment. The letter-writers note that if only the highest-quality reviews are examined, the results are no longer condemning of acupuncture. Mooney also dismisses the authors’ conclusion that acupuncture should be questioned since it reduces pain in “some conditions while failing to work in many others,” noting that the same can be said about acetaminophen.

The authors’ response ran in the same issue of PAIN: Ernst E. Response from Author. Letters to the Editor / PAIN 152 (2011) 2184–2186. As of Oct 13, it is noted this is withdrawn at authors’ request.

Ernst E, Lee MS, Choi TY. Acupuncture: does it alleviate pain and are there serious risks? A review of reviews. Pain. 2011 Apr;152(4):755-64.

Karst M. Comment on ''Acupuncture: does it alleviate pain and are there serious risks? A review of reviews" Ernst et al. [Pain 2011;152:755-764]. Pain. 2011 Sep;152(9):2181

Bäumler P, Irnich D. Pain. 2011 Sep;152(9):2181-2

Henke C. Pain. 2011 Sep;152(9):2183-4.

Manheimer E, Berman BM. Pain. 2011 Sep;152(9):2179-80