Acupuncture research: more wishful thinking than good science

This new RCT by researchers from the National Institute of Complementary Medicine in Sydney, Australia was aimed at ‘examining the effect of changing treatment timing and the use of manual, electro acupuncture on the symptoms of primary dysmenorrhea’. It had four arms:

low frequency manual acupuncture (LF-MA), high frequency manual acupuncture (HF-MA), low frequency electro acupuncture (LF-EA) and high frequency electro acupuncture (HF-EA).

A total of 74 women were given 12 treatments over three menstrual cycles, either once per week (LF groups) or three times in the week prior to menses (HF groups). All groups received a treatment in the first 48 hours of menses. The primary outcome was the reduction in peak menstrual pain at 12 months from trial entry.

During the treatment period and 9 month follow-up all groups showed statistically significant reductions in peak and average menstrual pain compared to baseline. However, there were no differences between groups. Health related quality of life increased significantly in 6 domains in groups having high frequency of treatment compared to two domains in low frequency groups. Manual acupuncture groups required less analgesic medication than electro-acupuncture groups. HF-MA was most effective in reducing secondary menstrual symptoms compared to both–EA groups.

The authors concluded that acupuncture treatment reduced menstrual pain intensity and duration after three months of treatment and this was sustained for up to one year after trial entry. The effect of changing mode of stimulation or frequency of treatment on menstrual pain was not significant. This may be due to a lack of power. The role of acupuncture stimulation on menstrual pain needs to be investigated in appropriately powered randomised controlled trials.

If I were not used to reading rubbish research of alternative medicine in general and acupuncture in particular, this RCT would amaze me – not so much because of its design, execution, or write-up, but primarily because of its conclusion (why, oh why, I ask myself, did PLOS ONE publish this paper?). They are, I think, utterly barmy.

Let me explain:

“acupuncture treatment reduced menstrual pain intensity” – oh no, it didn’t; at least this is not what the study proves; the fact that pain was perceived as less could be due to a host of factors, for instance regression towards the mean, or social desirability; as there was no proper control group, nobody can tell;

the lack of difference between treatments “may be due to a lack of power”. Yes, but more likely it is due to the fact that all versions of a placebo therapy generate similar outcomes.

“acupuncture stimulation on menstrual pain needs to be investigated in appropriately powered randomised controlled trials”. Why? Because the authors have a quasi-religious belief in acupuncture? And if they have, why did they not design their study ‘appropriately’?

The best conclusion I can suggest for this daft trial is this: IN THIS STUDY, THE PRIMARY ENDPOINT SHOWED NO DIFFERENCE BETWEEN THE 4 TREATMENT GROUPS. THE RESULTS ARE THEREFORE FULLY COMPATIBLE WITH THE NOTION THAT ACUPUNCTURE IS A PLACEBO THERAPY.

Something along these lines would, in my view, have been honest and scientific. Sadly, in acupuncture research, we very rarely get such honest science and the ‘National Institute of Complementary Medicine in Sydney, Australia’ has no track record of being the laudable exception to this rule.