Veterinary homeopathy: meta-analysis of randomised placebo-controlled trials

Background: Meta-analysis of randomised controlled trials (RCTs) of veterinary homeopathy has not previously been undertaken. For all medical conditions and species collectively, we tested the hypothesis that the outcome of homeopathic intervention (treatment and/or prophylaxis, individualised and/or non-individualised) is distinguishable from corresponding intervention using placebos. Methods: All facets of the review, including literature search strategy, study eligibility, data extraction and assessment of risk of bias, were described in an earlier paper. A trial was judged to comprise reliable evidence if its risk of bias was low or was unclear in specific domains of assessment. Effect size was reported as odds ratio (OR). A trial was judged free of vested interest if it was not funded by a homeopathic pharmacy. Metaanalysis was conducted using the random-effects model, with hypothesis-driven sensitivity analysis based on risk of bias. Results: Nine of 15 trials with extractable data displayed high risk of bias; low or unclear risk of bias was attributed to each of the remaining six trials, only two of which comprised reliable evidence without overt vested interest. For all N=15 trials, pooled OR = 1.69 [95% confidence interval (Cl), 1.12 to 2.56]; P=0.01. For the N=2 trials with suitably reliable evidence, pooled OR = 2.62 [95% Cl, 1.13 to 6.05]; P=0.02). Conclusions: Meta-analysis provides some very limited evidence that clinical intervention in animals using homeopathic medicines is distinguishable from corresponding intervention using placebos. The low number and quality of the trials hinders a more decisive conclusion.