Ernst E, Pittler MH. Efficacy of Homeopathic ArnicaA Systematic Review of Placebo-Controlled Clinical Trials. Arch Surg. 1998;133(11):1187-1190. doi:10.1001/archsurg.133.11.1187
Copyright 1998 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.1998
The efficacy of homeopathic remedies has remained controversial. The homeopathic remedy most frequently studied in placebo-controlled clinical trials is Arnica montana.
To systematically review the clinical efficacy of homeopathic arnica.
Materials and Methods
Computerized literature searches were performed to retrieve all placebo-controlled studies on the subject. The following databases were searched: MEDLINE, EMBASE, CISCOM, and the Cochrane Library. Data were extracted in a predefined, standardized fashion independently by both authors. There were no restrictions on the language of publications.
Eight trials fulfilled all inclusion criteria. Most related to conditions associated with tissue trauma. Most of these studies were burdened with severe methodological flaws. On balance, they do not suggest that homeopathic arnica is more efficacious than placebo.
The claim that homeopathic arnica is efficacious beyond a placebo effect is not supported by rigorous clinical trials.
HOMEOPATHY IS a system of medicine that was developed about 200 years ago and has remained highly controversial ever since.1 Essentially it is based on the "law of similars" and on the assumption that even nonmaterial dilutions ("potentiations") can be clinically effective.2 The law of similars claims that if a given remedy causes a certain symptom in a healthy person, the remedy should then be useful for treating that symptom in a patient who suffers from it. Homeopathic potentiations are prepared by serial dilutions and "succussions" (vigorous shaking) and often are so dilute that the likelihood of them containing a single molecule from the mother tincture is nil. Homeopaths believe that "potentizing" in this way will not reduce but rather increase the activity of the resulting remedy. It is in particular the use of highly diluted material that overtly flies in the face of science and has caused homeopathy to be regarded as placebo therapy at best and quackery at worst.
Thus, the efficacy of homeopathic remedies has always been a matter of bitter controversy.3 Recently the issue has been addressed in various ways. A recent systematic review found most homeopathic trials to come to positive conclusions.4 The authors, however, abstain from making a definitive conclusion as to the efficacy of homeopathic remedies. In their view, methodological problems with most clinical trials of homeopathy preclude a definitive judgement as to its efficacy.
Another approach to evaluate homeopathy is to perform a meta-analysis across all trials.5 Even though such a meta-analysis yielded a significantly positive result, with a combined odds ratio of 2.45 favoring homeopathic remedies over placebo, the conclusions that can be drawn from such a meta-analysis are limited and several caveats have been identified. There could be indefinable bias6 and the pooling of trials of vastly different remedies for vastly different conditions is of debatable legitimacy.7
A further approach is to systematically review homeopathic trials pertaining to a single disease or condition. We have chosen this method to assess homeopathy for postoperative ileus8 and delayed-onset muscle soreness,9 the 2 conditions most frequently submitted to controlled trials. In one case this resulted in a positive result8 and in the other in a negative overall result9 for homeopathy; eg, homeopathic remedies used to treat delayed-onset muscle soreness were not significantly better than placebo in alleviating symptoms. Yet again, several caveats preclude a firm conclusion.
Another approach could be to analyze homeopathic trial data by the type of remedy used; this was chosen for the present article. Its aim is to systematically review the homeopathic remedy most frequently submitted to controlled trials.
Computerized literature searches were performed to identify all placebo-controlled clinical trials of homeopathic Arnica montana. Databases included MEDLINE, EMBASE, CISCOM, and the Cochrane Library (all from their inception to October 1997). All 89 studies included in the above-mentioned systematic review4 and meta-analysis5 were also considered. One of us (E.E.) had been involved in a similar meta-analysis commissioned by the European Community.10 This exercise included extensive hand-searching of 28 specialized homeopathic journals, which resulted in the identification of more than 400 publications. The entire material was screened for this review. Furthermore, our own extensive files as well as review articles and books on homeopathy were searched for relevant publications. The bibliographies of the studies and reviews thus retrieved were searched for further trials. There were no restrictions regarding the language of publication.
Controlled clinical trials of homeopathic arnica (all potencies) vs placebo were included in this systematic review. Trials of one potency against another,11 trials with arnica as one of several remedies with no subanalysis on a pure arnica group,12 or studies in which arnica had been administered concomitantly with other remedies13 were excluded. Trials not published in the peer-reviewed literature were also excluded.
All studies were read in full by both of us. Data were extracted independently in a standardized, predefined fashion (Table 1). Methodological quality of the included trials was assessed using the score according to Jadad et al14 (Table 2). Discrepancies in the evaluation of individual trials were settled by discussion.
Eight studies fulfilled all of the aforementioned inclusion criteria and were admitted to this review.15- 21 Methodological details and outcomes of these trials are summarized in Table 1. Two studies15,17 yielded a statistically significant positive result (ie, arnica superior to placebo), 2 studies20 had a numerically positive result (ie, no formal test statistics were applied but an advantage of the arnica groups was apparent) and 4 studies16,18,19,21 showed a significantly negative result (ie, arnica not superior to placebo). Most trials included in this review are methodologically weak. Generally speaking, the more rigorous studies18,22 tended to be the ones that yielded negative findings. There is no obvious common denominator to differentiate between positive and negative studies, neither in terms of potency applied nor in terms of indication. For example, one trial on delayed-onset muscle soreness had positive results15 while another (methodologically superior) one had negative results.18 Similarly, one study of postsurgical complications had positive results17 and another had negative results.16 In addition, there are no data to suggest that one potency of arnica is superior to another.
Arnica is the classic homeopathic remedy for trauma of various kinds. "The word injury is constantly associated with the usefulness of arnica in trauma . . . this is especially true for soft tissue damage causing bruising, bleeding, and dislocation."22 The Homoeopathic Pharmacy lists as its first indication "trauma" and even recommends it as a first-aid treatment.23 The results of this systematic review unfortunately fail to lend support to this. On balance, the trial data do not support the notion that arnica is efficacious.
There are several ways of explaining this. The evidence could be scarce and a type II error could have produced a false-negative overall picture. This is not borne out by the data presented earlier. No other homeopathic remedy has been subject to more controlled clinical trials.
The existing studies could be severely flawed and therefore produce a misleading result. The trials certainly are burdened with a multitude of methodological limitations. Small sample size and lack of test statistics are frequent and obvious ones. However, such drawbacks would be likely to create a false-positive rather than a false-negative result.
Arnica could have been applied wrongly. Homeopaths do not treat a specific condition but rather the whole human being. Thus, it is not strictly according to the teaching of Hahnemann (the "inventor" of homeopathy) to use arnica for trauma much like an allopathic drug. There are, however, exceptions to this rule, and arnica is certainly one of them. The above quotations demonstrate that arnica is used for the conditions for which it was tested in the trials reviewed here.
Therefore, the hypothesis that homeopathic arnica is, in fact, not effective beyond a placebo effect must be considered. It is not possible to "prove a negative" with these data. It is, however, possible to comment on the most likely explanation of the overall result of this systematic review.
It is concluded that the hypothesis claiming that homeopathic arnica is clinically effective beyond a placebo effect is not based on methodologically sound placebo-controlled trials.
Reprints: E. Ernst, MD, PhD, FRCP(Edin), School of Postgraduate Medicine and Health Sciences, University of Exeter, 25 Victoria Park Rd, Exeter EX2 4NT, United Kingdom (e-mail: E.Ernst@exeter.ac.uk).