Systematic reviews of complementary therapies - an annotated bibliography. Part 3: homeopathy.

BACKGROUND: Complementary therapies are widespread but controversial. We aim to provide a comprehensive collection and a summary of systematic reviews of clinical trials in three major complementary therapies (acupuncture, herbal medicine, homeopathy). This article is dealing with homeopathy. Potentially relevant reviews were searched through the register of the Cochrane Complementary Medicine Field, the Cochrane Library, Medline, and bibliographies of articles and books. To be included articles had to review prospective clinical trials of homeopathy; had to describe review methods explicitly; had to be published; and had to focus on treatment effects. Information on conditions, interventions, methods, results and conclusions was extracted using a pretested form and summarized descriptively.
RESULTS: Eighteen out of 22 potentially relevant reviews preselected in the screening process met the inclusion criteria. Six reviews addressed the question whether homeopathy is effective across conditions and interventions. The majority of available trials seem to report positive results but the evidence is not convincing. For isopathic nosodes for allergic conditions, oscillococcinum for influenza-like syndromes and galphimia for pollinosis the evidence is promising while in other areas reviewed the results are equivocal.
INTERPRETATION: Reviews on homeopathy often address general questions. While the evidence is promising for some topics the findings of the available reviews are unlikely to end the controversy on this therapy.

Introduction

In this third part of our series on systematic reviews in complementary therapies we report our findings on homeopathy. Homeopathy is one of the most widespread forms of complementary medicine worldwide. According to a recent survey 3.4% of Americans have used homeopathy in the past 12 months [1]. It is even more widespread in some European countries [2], some countries in South America, India or Pakistan [3]. This widespread use is in strong contrast with the position held by many in scientific medicine that homeopathy has no effect beyond placebo [4].

The basic principle of homeopathy is the principle of similars: A patient with a specific pattern of symptoms is best treated by a remedy which causes the same or a very similar pattern in healthy subjects. Homoeopathic remedies are often prescribed in high dilutions some of which are unlikely to contain any molecules of the originally diluted agents. In consequence, homoeopathic remedies – at least when applied in high dilutions – cannot act by pharmacological means. Theories for a potential mechanism of action, therefore, postulate the storage of information in the dilution process by physical means [5].

Methods

A detailed description of the methods used in this review of reviews is given in the first part of this series [6]. As a specific intervention-related inclusion criterion we required that reports reviewed prospective (not necessarily controlled) clinical trials of homoeopathic medicines in humans.

Results

From a total of 22 potentially relevant reviews identified in the literature screening, 18 reviews published in 19 papers met the inclusion criteria [7-25] (see table 1). Four papers were excluded as they were only subgroup or methodological analyses of previously published papers [26-29].

Table 1

Systematic reviews of clinical trials of homoeopathy

				Features
Author Year	Indication	Homoeopathy/	Studies	1 / 2 / 3 /	Results	Conclusion
		Control		4 / 5
All homoeopathy in all conditions
Cucherat	all	all/placebo	17 RCT	y / y / y /	Combined p value for an effect	There is some evidence that
2000 [7]				n /y	over placebo p = 0.000036, for	homoeopathy is more than placebo.
					best trials only p = 0.08	Studies of high quality more likely to be
						negative
Linde 97 [8]	all	all/placebo	89 RCT	y / y / y /	OR of all trials over placebo 2.45	Results not compatbile with the
				y /y	(95%CI 2.05; 2.93), in better trials	hypothesis that all homeo-pathy is
					1.66 (1.33; 2.08)	placebo. No firm evidence for any
						single condition
Walach 97	all	all/placebo,	41 RCT	y / p / y /	Random effect size g = 0.259	The effects of homoeopathy are not
[9]		conventional		y / y	(95%CI -0.319; 0.837), fixed	different from placebo on a statistical
					effects 0.295 (0.223; 0.366)	level
Lutz 93 [10]	all	all/placebo,	21 RCT/CCT	? / n / y /	Results of available studies	No clear conclusions drawn.
		conventional		y / p	contradictory	(Comment: thesis mainly discussing
						problems of meta-analysis)
Kleijnen 91	all	all/placebo,	107 CCT	y / p / y /	81 trials reported positive results.	Available evidence positive but not
[11]		conventional		y / n	Most trials low quality but many	sufficient to draw definitive conclusions
					exceptions
Hill 90 [12]	all	all/placebo,	40 RCT	n / p / y /	The authors of half of the studies	The results do not provide acceptable
		conventional		y / n	concluded that homoeopathy was	evidence that homoeopathy treatments
					effective, further 7 promising	are effective
Individualized homoeopathy in all conditions
Ernst 99	all	individualized/	3 RCT, 3 CCT	y / p / n /	All trials were burdened with	The relative efficacy of individualized
[13]		conventional		y / n	serious methodological flaws.	homoeopathy compared to
					Results non-uniform	conventional treatments is not known
Linde 98	all	individualized/	32 RCT	y / y / y /	Responder RR vs. placebo 1.62	Available evidence suggests effects
[14]		placebo, convent.		y / y	(95%CI 1.17; 2.23), in better	over placebo. Evidence not convincing
					quality trials 1.12 (0.87; 1.44)	due to shortcomings and
						inconsistencies
Various homoeopathic treatments in a single condition/area
Barnes 97	postoperative	various/placebo	4 RCT, 2 CCT	y / y / y /	Time to first flatus in homoeopathy	Available evidence positive but several
[15]	ileus			y / y	significantly shorter. Best trial	caveats preclude definitive conclusions
					negative
Ernst 98	delayed-	various/placebo	8 double-blind	y / y / y /	Most trials with severe flaws. The	Published evidence does not support
[16]	onset muscle		trials (3	y / n	3 RCT showed no significant	the hypothesis that homoeopathic
	soreness		explicitly RCT)		effects over placebo	remedies are effective for muscle
						soreness
Jacobs 91*	rheumatic	various/placebo	4 CCT	p / y / n /	3 of 4 trials positive. Quality poor	No specific conclusion on
[17]	diseases			y / n		homoeopathy (generally: no convincing
						evidence for alternative therapies in
						rheumat.)
Linde 98	asthma	various/placebo	3 RCT	y / y / y /	Trials highly heterogeneous. Two	Currently available evidence insufficient
[18]				y / n	report statistically significant	to assess the possible role of
					effects	homoeopathy in the treatment of
						asthma
Arnica in various conditions (mainly various tissue traumata)
Lüdtke 99	all	arnica/placebo,	23 RCT, 14	y / y / y /	Quality often low. 13 of 35 studies	Available evidence suggests that arnica
[19]		no treatment	CCT	n / n	vs. placebo with significant results,	can be efficacious. Further rigorous
					10 with trend	trials needed
Ernst 98	all (mainly	arnica/placebo,	4 RCT, 4 CCT	y / y / y /	2 trials positive, 2 trials positive	Claims that homoeopathic arnica is
[20]	trauma)	conventional		y / n	trend. Most studies with severe	efficacious are not supported by
					flaws	rigorous trials
Similar homoeopathic treatments in one condition/a group of conditions
Taylor	allergic	isopathic	4 RCT	n / n / n /	Pooled analysis of 100 mm visual	Isopathic nosodes were different from
2000** [21]	conditions	nosodes/placebo		y / y	analogue scores 9.8 (95%CI	placebo on both subjective and
					4.2;15.4) mm better with isopathy	objective measures
Vickers	influenza-like	oscillococcinum/	7 RCT	y / y / y /	No evidence for preventative	Oscillococcinum probably reduces the
2000 [22]	syndrome	placebo		y / y	effect (3 trials) but reduction of	duration of influenza-like syndromes.
					length of illness in treatment trials	Further trials needed
Ernst 99	headache	individualized/	4 RCT	y / p / y /	one trial positive, one partially	The trial data do not suggest an effect
[23]	prophylaxis	placebo		y / n	positive, 2 negative	over placebo in the prophylaxis of
						migraine or headache
Wiesenauer	pollinosis	galphimia/placebo	8 RCT, 1 CS,	p / n / n /	Responder RR galphimia vs.	Galphimia is significantly more effective
96** [24,25]			2 UCS	y / y	placebo from 7 trials 1.25 (95%CI	than placebo
					1.09; 1.43)

*Disease-focused review on a variety of complementary medicine interventions including homoeopathy; **Meta-analytic overviews of researchers of their own trials on the topic Features: 1 = comprehensive search, 2 = explicit inclusion criteria, 3 = formal quality assessment, 4 = summary of results for each included study, 5 = meta-analysis; y = yes, p = partly, n = no, - = not applicable, ? = unclear RCT = randomized controlled trials, CCT = non-randomized controlled trials, CS = cohort study, UCS = uncontrolled study; OR = odds ratio, RR = rate ratio

Three quantitative meta-analyses addressed the general question whether homeopathy is different from placebo by pooling highly heterogeneous study samples [7-9]. Study samples and meta-analytic methods differed considerably (total number of trials covered 97). While two reviews reported significant effects of homeopathy [7,8] a third found no effect over placebo in the main analysis [9]. Several years before the publication of these studies a meta-analytic approach had already been tried in a thesis [10]. However, this review is mainly dealing with the problems encountered when trying to pool the data and cannot be interpreted meaningfully with respect to the effectiveness of homeopathy. Two older reviews included both placebo-controlled trials and comparisons with standard treatment [11,12] (total number of trials covered 107). Results were classified in a vote count as positive and negative. The majority of the studies had reported positive results. The conclusions were positive with reservations in one review and ambiguous in the other.

Two reviews focused on individualized homeopathy but were not restricted in terms of conditions investigated. A review of comparisons of individualized homoeopathic and conventional treatment found that only few trials of low quality exist [13]. The other review included mainly placebo-controlled trials [14]. Overall, the results suggested that individualized homeopathy is superior to placebo but when the analysis was limited to studies of better quality the difference was no longer significant.

Four reviews focused on a single condition or a group of conditions but included a variety of homoeopathic treatments [15-18]. Positive results have been reported for the treatment of postoperative ilues and asthma but definitive conclusions are not possible.

Arnica is the most often investigated homoeopathic remedy. Typically it is used in conditions involving tissue trauma. Two reviews with slightly different inclusion criteria have been published [19,20] (total number of trials covered 37). While the results of the available trials seem to be contradictory the more comprehensive of the two reviews had slightly more favorable conclusions.

Systematic reviews addressing more focused questions are available for the use of isopathic nosodes (diluted allergens) in allergic conditions, Oscillococcinum for influenza-like syndromes, individualized homeopathy for headache and galphimia for pollinosis [21-25]. Significant differences over placebo were reported for all but the headache review.

Discussion

Systematic reviews on homeopathy address, more often than in other areas of complementary medicine, general questions such as "is it more than placebo?" or "is it effective?" This is probably due to the fact that any effect of homeopathy over placebo is considered scientifically implausible. In consequence, the discussion does not primarily focus on specific clinical problems but on whether there is a real effect at all. While many overviews report that the majority of trial results are positive conclusions of reviewers are contradictory.

With few exceptions such as arnica for trauma or individualized homeopathy for headache, the reviews (and probably the primary research) do not cover conditions and treatment approaches which are relevant in homoeopathic practice. Self-medication with Oscillococcinum for influenza-like syndromes is popular in several countries but cannot be considered representative practice.

We want to emphasize again that it was not our primary objective to assess the effectiveness of homeopathy and the other therapies included in our series but to provide an annotated bibliography of the available systematic reviews. This provides an overall picture of the evidence but for an in-depth review readers must go back to the original reviews.

In conclusion, the available systematic reviews on homeopathy provide little guidance for patients and doctors. They rather reflect the ongoing fundamental controversy on this therapy and strengthen the perception that, on one side, positive evidence from clinical trials will not convince skeptics, and that on the other side negative results from trials not representing actual practice will not have any impact on homoeopaths.

Competing interest

KL, AV, GtR and DM have been involved in some of the reviews analyzed. These were extracted and assessed by other members of the team.

Pre-publication history

The pre-publication history for this paper can be accessed here: