Acupuncture Help harm or placebo

The most recent meta-analysis appearing in Fertility and Sterility on acupuncture was reevaluated in view of the marked heterogeneity of interventions, controls, data analysis, and timing of interventions.

David R. Meldrum, M.D., Andrew R. Fisher, B.S.E., Samantha F. Butts, M.D., M.S.C.E., H. Irene Su, M.D., M.S.C.E., Mary D. Sammel, Sc.D.

Volume 99, Issue 7, Pages 1821-1824, June 2013


The most recent meta-analysis appearing in this journal on acupuncture was reevaluated in view of the marked heterogeneity of interventions, controls, data analysis and timing of interventions in the trials that were included. After removing some of the trials and data based on more rigorous standards for a high quality meta-analysis, a significant benefit of the intervention could no longer be shown. When studies with and without placebo controls were analyzed separately, a placebo effect was suggested. Individual trials with a confidence limit below unity emphasized the potential for a detrimental impact on outcomes, which should be considered both in using acupuncture clinically as an adjunct for IVF and in design of future trials. Much more data that includes a placebo control will be required before a conclusion can be made that acupuncture has a true treatment effect on IVF outcomes. However, unless the timing and method of the acupuncture are standardized, practitioners will still have difficulty being sure that their particular method will help beyond the apparent benefit provided by a placebo.

  • María José Vidal

    It’s very rewarding to share information and experiences at this level because currently our team at the University Hospital La Fe of Valencia in Spain is also conducting a research in this field. Specialists in Human Reproduction and Traditional Chinese Medicine know that designing a study, where both medical philosophies coexist, is a big challenge. Randomized studies of acupuncture & IVF tend to
    reduce the variables in order to simplify the statistical analysis of the data. This limits their designs completely, often resulting in meaningless conclusions. I consider that the meta-analysis of Zeng
    et al. tries to advise on the most important keys which
    should be assessed before creating new designs for future researches. I absolutely agree with you, Dr. Magarelli, when you explain the limitation that involves “USING ACTIVE PLACEBOS”, such as needle Streitberger or Park, since their studies of validation give a sense of “deqi”, which is equal to present a therapeutic effect in acupuncture. Thank you Dr. Magarelli for sharing the high quality of your researches.
    Other important limitations that are presented in Acupuncture & IVF
    researches would be:
    1st; “USING FIXED PROTOCOLS WITHOUT REFERRING THEM TO A DIFFERENTIATION OF SYNDROMES”. The most correct way would be to use a fixed-points protocol which includes some basic points and a protocol of complementary points depending on the diagnostic pattern that the patient presents. (Kidney Yang deficiency, Kidney Yin deficiency, Qi and blood deficiency, retention of cold in the uterus, phlegm and blood stagnation).
    2; “INSUFFICIENT DOSAGE OF ACUPUNCTURE ADMINISTERED “, I agree with Zeng et al. in administering at least a minimum of 4
    sessions, even though, to administer acupuncture sessions before IVF cycle, would bring greater benefits to the follicular response by acting at the level of the folliculogenesis process.
    3; “USING CONTRAINDICATED ACUPUNCTURE POINTS AFTER EMBRYO TRANSFER “, by excessively mobilizing Qi (energy) can trigger uterine contractions.
    See (Belinda Anderson et al. “Acupuncture and in vitro fertilization:
    Critique of the evidence and application to clinical practice”).
    I also agree with the comments and recommendations of Zeng et al. and Meldrum et al., in relation to the stress that patients suffer when they receive real acupuncture versus placebo. In their article “Acupuncture – help, harm, or placebo”,conducting a multicenter studies is a solution that the authors propose to this problem , which I consider an excellent idea; bearing in mind that it would be very
    important to achieve an accurate design, as far as possible, taking into account of all these limitations. Taking part in an investigation of this nature could be an exciting challenge.

  • Paul Magarelli

    The key flaw in any of these discussions is the notion that there is a “thing” such as Acupuncture medicine… there are Acupuncture treatments which vary based on symptoms and signs. Be that as it may, the saying “Garbage in Garbage out” related to Micah’s comment below… unless and until we have powered studies with NO PLACEBO EFFECT (see red light versus laser in Fraterelli’s study)… given that there are no placebo acupuncture needled points and studies that attempt to discern an effect are conflicting any conclusions are suspect. Fraterelli’s study WAS powered correctly (see F&S) and demonstrated impact of an acupuncture protocol. One should not assume that an abstract presented by a well known research group (RMA Richard Scotts program) would be suspect. It may serve no purpose to publish the full paper or the author went on to other things.

    I certainly understand RCT are the gold standard of drug trials and other experimental interentions… however propective cohort studies may well represent the best method for Acupuncture treatment testing given the very nature of the intervention. If that is the case then there are numerous studies with those criteria that have robust data sets with excellent comparison groups… see (magarelli et al)

    It remains exciting to see continued interest in a “standard of care” for IVF patients with the desire to add to their conventional treatments… just call any IVF center and listen to the hold message for advertisements about their TCM program

    • Micah Hill

      Thank you Dr Magarelli for your expertise and thoughts. Frattarelli was my mentor in residency and in my experience his work is always of very high quality! My questions was not a critique of his specific study, but rather a methodological one when conducting a meta-analysis. From an abstract its hard to assess studies for quality issues that can have significant impacts on bias (quasi-randomization, concealment, allocation, drop out, ect). I think such assessments are important in interpreting the meta-analysis.

  • Micah Hill

    I very much appreciated this review of the Zheng meta-analysis on acupuncture. As this review demonstrates thoroughly, there was such marked heterogeneity in how acupuncture studies have been performed, it is challenging to statistically combine the results into meaningful conclusions. I think this also shows that more data for a meta-analysis is not always a good thing, if that data itself isn’t of high quality.

    Im curious about the author’s opinion on including trials in a meta-analysis which never made it to full peer review publication? For example, the Frattarelli trial of 1000 patients was a RCT presented at ASRM 5 years ago now, but to my knowledge was never published. It seems you could argue inclusion of these sorts of abstracts and grey literature might help reduce the impact of publication bias in a meta-analysis. But conversely there may be significant reasons why such trials never made it to full publication and could represent lower quality data.

Translate »