New review of acupuncture: a breath of fresh air in a morass of dodgy research

Polycystic ovarian syndrome (PCOS) is a common condition characterised by oligo-amenorrhoea, infertility and hirsutism. Conventional treatment of PCOS includes a range of oral pharmacological agents, lifestyle changes and surgical modalities. Some studies have suggested that acupuncture might be helpful but the evidence is often flawed and the results are mixed. What is needed in such a situation is, of course, a systematic review.

The aim of this new Cochrane review was to assess the effectiveness and safety of acupuncture treatment of oligo/anovulatory women with polycystic ovarian syndrome (PCOS). The authors identified relevant studies from databases including the Cochrane Central Register of Controlled Trials (CENTRAL), Ovid MEDLINE, EMBASE, PsycINFO, CNKI and trial registries. The data are current to 19 October 2015.

They included randomised controlled trials (RCTs) that studied the efficacy of acupuncture treatment for oligo/anovulatory women with PCOS. We excluded quasi- or pseudo-RCTs. Primary outcomes were live birth and ovulation (primary outcomes), and secondary outcomes were clinical pregnancy, restoration of menstruation, multiple pregnancy, miscarriage and adverse events. We assessed the quality of the evidence using GRADE methods.

Two review authors independently selected the studies, extracted data and assessed risk of bias. They calculated Mantel-Haenszel odds ratios (ORs) and mean difference (MD) and 95% confidence intervals (CIs).

Five RCTs with 413 women were included. They compared true acupuncture versus sham acupuncture (two RCTs), true acupuncture versus relaxation (one RCT), true acupuncture versus clomiphene (one RCT) and electroacupuncture versus physical exercise (one RCT). Four of the studies were at high risk of bias in at least one domain. No study reported live birth rate. Two studies reported clinical pregnancy and found no evidence of a difference between true acupuncture and sham acupuncture (OR 2.72, 95% CI 0.69 to 10.77, two RCTs, 191 women, very low quality evidence). Three studies reported ovulation. One RCT reported number of women who had three ovulations during three months of treatment but not ovulation rate. One RCT found no evidence of a difference in mean ovulation rate between true and sham acupuncture (MD -0.03, 95% CI -0.14 to 0.08, one RCT, 84 women, very low quality evidence). However, one other RCT reported very low quality evidence to suggest that true acupuncture might be associated with higher ovulation frequency than relaxation (MD 0.35, 95% CI 0.14 to 0.56, one RCT, 28 women). Two studies reported menstrual frequency. One RCT reported true acupuncture reduced days between menstruation more than sham acupuncture (MD 220.35, 95% CI 252.85 to 187.85, 146 women). One RCT reported electroacupuncture increased menstrual frequency more than no intervention (0.37, 95% CI 0.21 to 0.53, 31 women). There was no evidence of a difference between the groups in adverse events. Evidence was very low quality with very wide CIs and very low event rates. Overall evidence was low or very low quality. The main limitations were failure to report important clinical outcomes and very serious imprecision.

The authors concluded that, thus far, only a limited number of RCTs have been reported. At present, there is insufficient evidence to support the use of acupuncture for treatment of ovulation disorders in women with PCOS.

This is, in my view, a rigorous assessment of the evidence leading to a clear conclusion. Foremost, I applaud the authors from the Faculty of Science, University of Technology Sydney for using such clear language. Such clarity seems to be getting a rare event in reviews of alternative medicine. To demonstrated this point, here are the most recent 5 systematic reviews which came up on my screen when I searched today Medline for ‘complementary alternative medicine, systematic review’.

The combination of TGP and LEF in treatment of RA presented the characteristics of notably decreasing the levels of laboratory indexes and higher safety in terms of liver function. However, this conclusion should be further investigated based on a larger sample size.

Compared to control groups, both MA and EA were more effective in improving AHI and mean SaO2. In addition, MA could further improve apnea index and hypopnea index compared to control.

CHM as an adjunctive therapy is associated with a decreased risk of in-hospital mortality compared with WT in patients with AKI. Further studies with high quality and large sample size are needed to verify our conclusions.

clinicians may consider Tai Chi as a viable complementary and alternative medicine for chronic pain conditions.

As an important supplementary treatment, TCM may provide benefits in repair of injured spinal cord. With a general consensus that future clinical approaches will be diversified and a combination of multiple strategies, TCM is likely to attract greater attention in SCI treatment.

I think the phenomenon is fairly obvious: authors of such papers are far too often not able or willing to express the bottom line of their work openly. As systematic reviews are supposed to be the ultimate type of evidence, this trend is very worrying, I think. In my view, such conclusions merely display the bias of the authors. If the evidence is not convincingly positive (which it very rarely is), authors have an ethical obligation to clearly say so.

If they don’t do it, journal editors have the duty to correct the error. If neither of these actions happen, funding agencies should make sure that such teams get no further research money until they can demonstrate that they have learnt the lesson.

This may sound a bit drastic but I think such steps would be both necessary and urgent. The problem is now extremely common, and if we do not quickly implement some effective preventative measures, our scientific literature will become contaminated to the point of becoming useless. This surely would be a disaster that affects us all.

There can, of course, be several reasons for the evidence being not positive:

• there can be a paucity of data
• the results might be contradictory
• the trials might be open to bias
• some of the primary data might look suspicious

In all of these cases, the evidence would be not convincingly positive, and it would be wrong and unhelpful not to be frank about it. Beating about the bush, like so many authors nowadays do, is misleading, unhelpful, unethical and borderline fraudulent. Therefore it constitutes a disservice to everyone concerned.