A recent RCT in the BMJ reported a reduction in migraine in patients receiving acupuncture compared to those receiving sham acupuncture.[1]

The trial was small, at less than 100 participants, and the improvement modest. The applicants make a claim that the participants could not distinguish between sham and real acupuncture, because when asked there was no significant difference in detection between the two groups.

However, neither the applicants, nor the writer of an accompanying editorial,[2] point out the obvious fallacy in such a conclusion. The statistical comparison is woefully under powered to detect a difference in detection rates that might be sufficient to skew the results. In fact, the direction of the effect was such that four people were not fully taken in by the sham procedure.

Many acupuncture trials have a three-way comparison: control, sham and full acupuncture. They seem to show a pattern of increasing effect across these three groups. A plausible explanation, for at least part of the effect, is that a proportion of people in the sham group suspect that this is the case, thereby reducing the placebo effect, which is not reduced in the full acupuncture group. It would be interesting to carry out a multi-indication systematic review across all acupuncture trials.

Richard Lilford, ARC WM Director

References:

1. Xu S, et al. Manual acupuncture versus sham acupuncture and usual care for prophylaxis of episodic migraine without aura: multicentre, randomised clinical trial. BMJ. 2020; 368: m697.
2. Angus-Leppan H. Manual acupuncture for migraine. BMJ. 2020; 368 :m1096