In 2012, we published a systematic review of adverse effects of homeopathy. Here is its abstract:
Aim: The aim of this systematic review was to critically evaluate the evidence regarding the adverse effects (AEs) of homeopathy.
Method: Five electronic databases were searched to identify all relevant case reports and case series.
Results: In total, 38 primary reports met our inclusion criteria. Of those, 30 pertained to direct AEs of homeopathic remedies; and eight were related to AEs caused by the substitution of conventional medicine with homeopathy. The total number of patients who experienced AEs of homeopathy amounted to 1159. Overall, AEs ranged from mild-to-severe and included four fatalities. The most common AEs were allergic reactions and intoxications. Rhus toxidendron was the most frequently implicated homeopathic remedy.
Conclusion: Homeopathy has the potential to harm patients and consumers in both direct and indirect ways. Clinicians should be aware of its risks and advise their patients accordingly.
The paper prompted a number of angry reactions from proponents of homeopathy who claimed, for instance, that homeopathic remedies are highly diluted and thus safe. We responded that homeopaths can nevertheless be dangerous to patients through neglect and bad advice by homeopaths, and that not all homeopathic remedies are highly diluted, and that some might be toxic because of poor quality control of the manufacturing process.
Now, a different group of researchers have looked at the problem from a slightly different angle and with different methodologies. This systematic review and meta-analysis by researchers from NAFKAM focused on observational studies, as a substantial amount of the research base for homeopathy are observational.
Eight electronic databases, central webpages and journals were searched for eligible studies, and a total of 1,169 studies were identified, 41 were included in this review. Eighteen studies were included in a meta-analysis that made an overall comparison between homeopathy and control (conventional medicine and herbs).
Eighty-seven percent (n = 35) of the studies reported adverse effects. They were graded as CTCAE 1, 2 or 3 and equally distributed between the intervention and control groups. Homeopathic aggravations (homeopaths believe that, when the optimal remedy is given, patients will experience an aggravation of their presenting symptoms) were reported in 22,5% (n = 9) of the studies and graded as CTCAE 1 or 2. The frequency of adverse effects for control versus homeopathy was statistically significant (P < 0.0001). Analysis of sub-groups indicated that, compared to homeopathy, the number of adverse effects was significantly higher for conventional medicine (P = 0.0001), as well as other complementary therapies (P = 0.05).
The authors concluded that adverse effects of homeopathic remedies are consistently reported in observational studies, while homeopathic aggravations are less documented. This meta-analysis revealed that the proportion of patients experiencing adverse effects was significantly higher when receiving conventional medicine and herbs, compared to patients receiving homeopathy. Nonetheless, the development and implementation of a standardized reporting system of adverse effects in homeopathic studies is warranted in order to facilitate future risk assessments.
While these results are interesting, they have to be taken with a pinch of salt and beg a number of questions:
- Is there proof that aggravations exist at all?
- How can one differentiate them from adverse effects?
- As even placebos are known to cause adverse effects (nocebo effects), how can one be sure that the adverse effects of homeopathy are not nocebo effects?
- Is it a good reason to focus on largely inconclusive observational studies, because a substantial amount of the research base for homeopathy are observational?
- Can one produce conclusive results by meta-analysing inconclusive studies?
For me, the most impressive findings of this review is that in total 86 studies had to be excluded by the authors because they reported no adverse effects or aggravations. I think this renders the interpretation of the evidence from the 41 studies they did include even more flimsy. In fact, I don’t see how any meaningful conclusion can be drawn at all – except of course that many researchers of homeopathy violate the rules of research ethics by not reporting adverse effects in their studies.
As to aggravations, we clearly need to rely on placebo controlled studies, if we want to find out whether they exist at all. This we have done in our 2003 paper:
Homeopathic aggravations have often been described anecdotally. However, few attempts have been made to scientifically verify their existence. This systematic review aimed at comparing the frequency of homeopathic aggravations in the placebo and verum groups of double-blind, randomised clinical trials. Eight independent literature searches were carried out to identify all such trials mentioning either adverse effects or aggravations. All studies thus found were validated and data were extracted by both authors. Twenty-four trials could be included. The average number of aggravations was low. In total, 50 aggravations were attributed to patients treated with placebo and 63 to patients treated with homoeopathically diluted remedies. We conclude that this systematic review does not provide clear evidence that the phenomenon of homeopathic aggravations exists.
What is interesting, from my perspective, is the fact that the NAFKAM authors chose to ignore our 2012 paper completely (even though it is highly relevant to their paper and was not published in an obscure journal) and elected to completely misinterpret the findings of our 2003 paper (stating this about it: Grabia and Ernst reported a total of 103 cases of homeopathic aggravations in 3437 participants (3%), while, in fact, our paper demonstrated that aggravations are a homeopathic figment of imagination).
I wonder why.
In the past, NAFKAM did not have the reputation of doing research that was overtly biased towards homeopathy. Recently, the head of the team retired and was replaced by Miek C. Jong who is a co-author of the present review (plus head of CAMcancer, an organisation of which I am a founding member and which did, I think, some good work in the past). She happens to have a long history as a homeopath or homeopathic researcher and is co-author of many papers in this area. Here are three of her conclusions:
- The prognostic questionnaire proved a reliable tool to rank 11 homeopathic medicines by total scores, based on keynote symptoms. This PMS algorithm can be used for the treatment of PMS/PMDD in clinical practice.
- Both complex homeopathic products led to a comparable reduction of URTIs. In the CalSuli-4-02 group, significantly less URTI-related complaints and symptoms and higher treatment satisfaction and tolerability were detected. The observation that the use of antibiotics was reduced upon treatment with the complex homeopathic medications, without the occurrence of complications, is interesting and warrants further investigations on the potential of CalSuli-4-02 as an antibiotic sparing option.
- Our systematic evaluation demonstrated that the reporting rate of ADRs associated with anthroposophic and homeopathic solutions for injection is very low. Most reported ADRs were listed, and one quarter consisted of local reactions. These findings suggest a low risk profile for solutions for injection as therapeutically applied in anthroposophic medicine and homeopathy.
Could it be that, within NAFKAM, the attitude towards homeopathy has changed?