Am Fam Physician. 2020;101(6):325-326
Original Article: Acupuncture for Pain
Issue Date: July 15, 2019
See additional reader comments at: https://www.aafp.org/afp/2019/0715/p89.html
To the Editor: This article does not meet American Family Physician's standard of rigorously evidence-based continuing medical education. The authors acknowledge the difficulties of performing good research on acupuncture, but then cite demonstrably low-quality studies in which acupuncture is compared with usual care or no care (a wait-list group) or another treatment is compared with the same treatment plus acupuncture (A vs. A+B). Such studies are likely to make acupuncture appear more effective than it really is. In the Strength-of-Recommendation Taxonomy (SORT) table, they give undeserved “A” evidence ratings to these types of studies, which were also not consistent with other studies.
Two crucial studies are missing from their discussion. First, a cogently reasoned and extensively referenced analysis pointed out flaws and inconsistencies in the research and concluded that “the benefits of acupuncture are likely nonexistent, or at best are too small and too transient to be of any clinical significance. It seems that acupuncture is little or no more than a theatrical placebo.”1 Second, a systematic review of 57 systematic reviews of acupuncture for pain found that there were only four conditions for which more than one systematic review reached the same conclusion: they agreed that it did not work in three cases and that it did work in only one case (for neck pain).2 If the research on a pain pill showed that it worked for neck pain but not for pain elsewhere in the body, we would conclude that it does not work and that the neck pain studies represent false positives.
The final section of the American Family Physician article correctly identifies acupuncture as a placebo but still recommends its use. There is no justification for the use of placebos outside the setting of clinical trials.3 Prescribing placebos involves deception and precludes informed consent.
Additionally, there have been at least 95 published reports of serious complications of acupuncture, including five deaths.2 The risks are admittedly small, but no risk is acceptable for a placebo treatment.
In Reply: I agree with Dr. Hall that acupuncture has a large placebo response, as we emphasized in our article. This effect is not unique to acupuncture and, in fact, the degree to which meaning and context contribute to overall therapeutic responses is probably underestimated for many commonly used medical and surgical interventions.1
An obviously important question is whether there are acupuncture effects that exceed meaning and context effects. We were careful to answer this question as accurately as possible using available evidence from meta-analyses and systematic reviews, many of which were published more recently than those cited by Dr. Hall. We provided quantitative outcome comparisons so that readers can draw their own conclusions regarding the clinical importance of the differences between verum and sham acupuncture or between acupuncture and other available treatments.
Our SORT table indicated an evidence rating of A for clinical recommendations for which there are consistent findings from multiple systematic reviews of randomized controlled trials, and a rating of B when the evidence for a clinical recommendation is inconsistent or limited-quality. From a safety perspective, there have been rare instances of serious adverse effects of acupuncture, despite a very good record overall. The reference Dr. Hall cites in this regard proposed that many of these uncommon but serious complications are not intrinsic to acupuncture but rather are caused by malpractice of acupuncturists, and recommends that all acupuncturists receive adequate training to reduce risk of complications.2 I agree with this recommendation.
