Acupuncture versus clomiphene controversy

Stimulated by Gao et al 2019.[1]

Much as I would have liked to let this one slide by unannounced, I feel obliged to comment in view my previous blog (Segmental EA falls short in ovulation induction) on one of the included trials.[2]

We have a lot of new systematic reviews (SRs) coming to the journal (Acupuncture in Medicine) that include, or are entirely made up of, papers written in Chinese. The data in these papers have previously been inaccessible to most reviewers, so that is good. But we then have to balance this data against the methodological weakness of the predominantly open acupuncture trials from this region of the world.

This review nicely illustrates the divide as it contains one huge sham controlled trial (n=1000) published in English, along with 8 smaller open trials (n=87, n=124, n=66, n=100, n=124, n=132, n=120, n=240) published in Chinese. So, in total we have 993 participants from the open studies and 1000 from the single sham controlled study. All of these trials were performed in China.

Seven of the eight open studies compare acupuncture head to head with clomiphene citrate.

Clomiphene citrate (CC) interacts with oestrogen receptor (OR)-containing tissues. It is thought to compete with oestrogen at OR binding sites, and thus interfere with the negative feedback of oestrogen on gonadotrophin hormone release (including gonadotrophin releasing hormone – GnRH, follicle-stimulating hormone – FSH, and luteinising hormone – LH). The result is follicle stimulation in the ovary and increased circulating oestradiol.

RxList

Live birth rate following clomiphene is about 25%

Live birth rate (LBR) following CC is about 25% (based on the biggest data I could find online: RxList), although the Cochrane review on anti oestrogens for ovulation induction in polycystic ovarian syndrome (PCOS) seems less certain (no data found on LBR for CC).[3] These are likely to be different populations. The Cochrane review on acupuncture for PCOS is equally uncertain, but this was restricted to looking at efficacy over sham.[4]

The pooled estimate in this SR found acupuncture to be twice as good as CC – OR 2.34 (odds ratio). This estimate included 7 trials and 883 women. There was no statistical heterogeneity.

Acupuncture was twice a good as clomiphene…

The two remaining trials (including the big sham controlled one with 1000 women) compared acupuncture plus CC against CC alone. Well one of the trials performed this comparison, but the other (the big one) compared acupuncture plus CC with minimal sham acupuncture plus CC (there were two other arms in this study with the same acupuncture groups but placebo CC). This analysis was of course dominated by the big sham controlled study and showed no effect of adding acupuncture to CC.

How can acupuncture be twice as good as clomiphene, but do nothing when added to it?

That is a difficult set of results to assimilate. How can acupuncture be twice as good as CC but do nothing when added to CC? Bias is the default explanation in evidence-based medicine (EBM), and there is a higher risk of bias in the open trials. But can you think yourself pregnant? Or at least alter any relevant factors that result in pregnancy by the power of your mind, and hence bias the results in favour of acupuncture over CC? Well relaxation or anxiolysis might well help,[5,6] and this may explain the similarity of real and sham acupuncture performed around the time of embryo transfer within IVF.[7]

In the big sham controlled 4 arm study (Wu et al 2017) that I discussed before on this blog, adding real CC to acupuncture (whether segmental EA or minimal non-segmental needling) nearly doubled the LBR, with no difference between the acupuncture groups.[2] Whilst the LBR is the most externally valid outcome, none of the open trials reported this, so pregnancy rate is used in the meta-analysis. Looking at the raw figures from the forest plots, the pregnancy rate with CC is about 30% in Wu et al, and about 40% in the open trials. The pregnancy rate was 60% in the acupuncture groups of the open trials, and 15% in the acupuncture groups of Wu et al.

The pregnancy rate was 60% in the acupuncture groups of the open trials, and 15% in the acupuncture groups of Wu et al.

I was curious about the difference in the CC groups. Could this all be put down to biased methods? I did a bit more thinking and digging… The populations may have been different! The 30% rate comes from PCOS patients who had been trying to conceive for 2 years, yet the inclusion criteria for this SR was women who were unable to conceive after 12 months of regular unprotected intercourse. Perhaps this explains the better results? The open trials may have been performed in groups with a better overall prognosis. But no, this does not explain it. Where it was recorded, the length of infertility ranged from 2-5 years in the open trials.

So, apart from bias, I am only left with one plausible explanation for the disparity in results between Wu et al (the big rigorous sham controlled 4 arm double dummy trial published in JAMA) and all the smaller open trials (published in Chinese language journals). The context effects of acupuncture trumps the real effects of CC. In Wu et al the context effects occur equally in all 4 arms, and in the open trials they are only in the acupuncture arm of the trials.

That does leave us with the rather embarrassing situation that this context effect can vary from 15% to 60% in terms of pregnancy rates. I’ll leave the reader to judge the plausibility of this.

Pregnancy rates in acupuncture arms of RCTs
range from 15% to 60%

References

1         Gao R, Guo B, Bai J, et al. Acupuncture and clomiphene citrate for anovulatory infertility: a systematic review and meta-analysis. Acupunct Med Published Online First: 3 October 2019. doi:10.1136/acupmed-2017-011629

2         Wu X-K, Stener-Victorin E, Kuang H-Y, et al. Effect of Acupuncture and Clomiphene in Chinese Women With Polycystic Ovary Syndrome: A Randomized Clinical Trial. JAMA 2017;317:2502–14. doi:10.1001/jama.2017.7217

3         Brown J, Farquhar C. Clomiphene and other antioestrogens for ovulation induction in polycystic ovarian syndrome. Cochrane Database Syst Rev 2016;12:CD002249. doi:10.1002/14651858.CD002249.pub5

4         Lim CED, Ng RWC, Cheng NCL, et al. Acupuncture for polycystic ovarian syndrome. Cochrane Database Syst Rev 2019;7:CD007689. doi:10.1002/14651858.CD007689.pub4

5         Palomba S, Daolio J, Romeo S, et al. Lifestyle and fertility: the influence of stress and quality of life on female fertility. Reprod Biol Endocrinol 2018;16:113. doi:10.1186/s12958-018-0434-y

6         Joseph D, Whirledge S. Stress and the HPA Axis: Balancing Homeostasis and Fertility. Int J Mol Sci 2017;18:2224. doi:10.3390/ijms18102224

7         Smith CA, Armour M, Shewamene Z, et al. Acupuncture performed around the time of embryo transfer: a systematic review and meta-analysis. Reprod Biomed Online 2019;38:364–79. doi:10.1016/j.rbmo.2018.12.038

---

Declaration of interests MC