After > 200 years of existence, homeopathy still remains unproven – in fact, most rational thinkers would call it disproven. Today only homeopaths doubt this statement; they work hard to find a water-tight proof that might show the doubters to be wrong.

What is better suited for this purpose than a few rigorous animal experiments?

Engystol® is a popular homeopathic product promoted as an anti-viral agent manufactured by Heel GmbH, Baden-Baden, Germany. In several in vivo and in vitro studies, it apparently affected an immune response. This new study was to “evaluate the innate and adaptive immuno-modulatory effects of oral Engystol® (1 or 10 tablets/L water consumed), prior to and post antigenic challenge in a mouse model with a well-characterized and clinically measureable immune system.”

The investigators first evaluated the murine immune response when oral Engystol® was given alone for 28 days. to mice. The animals were then challenged with an antigen-specific H5N1 HA vaccine while on Engystol® for an additional 33 days. Serum and supernatants from cultured splenic lymphocytes were collected and screened with a 32-cytokine panel. Serum vaccine epitope-specific IgG titers plus T cell and B cell phenotypes from splenic tissue were also evaluated.

The results showed that Engystol® alone did not alter immunity. However, upon vaccine challenge, Engystol® decreased CD4+/CD8+ ratios, altered select cytokines/chemokines, and anti-H5N1 HA IgG titers were increased in the group of mice receiving 10 tablet/L.

The authors concluded that “these data suggest that Engystol® can modulate immunity upon antigenic challenge.”

Engystol is being advertised as “a homeopathic preparation which has been scientifically proven to significantly reduce the duration and severity of symptoms during an acute viral infection and help protect from subsequent infections.” I was unable find good evidence for this claim and therefore have to assume that it is bogus. The only human trial I was able to locate was this one:


To compare the effects of a complex homeopathic preparation (Engystol; Heel GmbH, Baden-Baden, Germany) with those of conventional therapies with antihistamines, antitussives, and nonsteroidal antiinflammatory drugs on upper respiratory symptoms of the common cold in a setting closely related to everyday clinical practice.


Nonrandomized, observational study over a treatment period of maximally two weeks.


Eighty-five general and homeopathic practices in Germany.


Three hundred ninety-seven patients with upper respiratory symptoms of the common cold.


Engystol-based therapy or common over-the-counter treatments for the common cold. Patients receiving this homeopathic treatment were allowed other short-term medications, but long-term use of analgesics, antibiotics, and antiinflammatory agents was not permitted. Patients were allowed nonpharmacological therapies such as vitamins, thermotherapies, and others.


The effects of treatment were evaluated on the variables fatigue, sensation of illness, chill/tremor, aching joints, overall severity of illness, sum of all clinical variables, temperature, and time to symptomatic improvement.


Both treatment regimens provided significant symptomatic relief, and this homeopathic treatment was noninferior in a noninferiority analysis. Significantly more patients (P < .05) using Engystol-based therapy reported improvement within 3 days (77.1% vs 61.7% for the control group). No adverse events were reported in any of the treatment groups.


This homeopathic treatment may be a useful component of an integrated symptomatic therapy for the common cold in patients and practitioners choosing an integrative approach to medical care.

Let me comment on the human study first. It is an excellent example of the bias that can be introduced by non-randomization. The patients in the homeopathic group obviously were those who chose to be treated homeopathically. Consequently they had high expectations in this therapy. Consequently they reported better results than the control group. In other words the reported outcomes have nothing to do with the homeopathic remedy.

But what about the animal study? Animals, we hear so often, do not exhibit a placebo response. Does that render this investigation any more reliable?

The answer, I am afraid is no.

The animal study in question had no control group at all. Therefore a myriad of factors could have caused the observed result. This study is very far from a poof of homeopathy!

But even if the findings of the two studies had not been the result of bias and confounding, I would be more than cautious about viewing them as anything near conclusive. The reason lies in the nature of this particular homeopathic remedy.

Engystol® contains Vincetoxicum hirundinaria (D6), Vincetoxicum hirundinaria (D10), Vincetoxicum hirundinaria (D30), sulphur (D4) and sulphur (D10). In other words, it is one of those combination remedies which are not sufficiently dilute to be devoid of active molecules. Sulphur D4, for instance, means that the remedy contains one part of sulphur in 10 000 parts of diluent. It is conceivable, even likely that such a concentration might affect certain immune parameters, I think.

And my conclusion from all this?

The proof of homeopathy – if it ever came – would need to be based on investigations that are more rigorous than these two rather pathetic studies.

18 Responses to At last! The PROOF that homeopathy works… no, not really

  • Even if ONE homeopathic wingwang treatment were to be shown to work, what would that prove? Just ONE, after more than two hundred years of attempts, involving how many different potions? Would we have to wait another two hundred years for the second success? If genuine medicine had such a feeble success rate, we would have A. Given up, or B. Come to the conclusion that these results were false. And this in an area where its proponents claim that the truth is right there in front of us, if only we would open our eyes? It’s as though a football team loses by 2,000 goals to 1, and the only goal they scored is ruled out later by a technicality/evidence of bribery, but its small group of fanatical fans said ‘ Ah but our results can’t be judged by the normal rules of football’.

      • @Alex Smithers

        As you have posted links without comment, no reader can know what you think they’re intended to show. But in the context of this thread, I’ll assume you consider these two papers prove a positive effect of homeopathy products. They don’t.

        The first paper purports to show that 2 homeopathy products, Engystol and Gripp-Heel, show antiviral effects in vitro and stimulate type 1 IFN release from cells in culture. Small problem: there’s no control product tested (ideally it would have been the excipients for Engystol and Gripp-Heel). This renders the paper and its conclusions worthless.

        The second paper is a review of studies on the effects of homeopathy products in a range of animal experiments. From the abstract… “Some experimental lines…give support to a real effect of homeopathic high dilutions in animals, but often these data are of preliminary nature and have not been independently replicated.” “Despite a few encouraging observational studies, the effectiveness of the homeopathic prevention or therapy of infections in veterinary medicine is not sufficiently supported by randomized and controlled trials.” Say no more.

        May I PLEAD with those who post links in this blog to published material to do two things. (1) Inform readers why you think the publication is impressive/important/relevant. (2) Try to think critically about the flaws in the publication(s). If the work can’t easily be picked apart it’s possibly sound, but if you’ve merely skimmed it with your confirmation bias set to full it’s unlikely to help your case.

  • There is far too little information in the mouse study abstract. Do you have the full paper?
    They compared at least 32 cytokines and several other tests. Did they correct their ‘significance’ for multiple comparisons?
    If they had no controls, what did they compare against?
    It is odd to me that they used cultured cells for the cytokines measurements. That introduces further possibilities for inadvertent bias due to lack of blinding.
    These combination remedies often seem to have measurable amounts of medicine. D6 is 1 part per million. I’ve taken real medicines that are that diluted, and as you say, D4 is 1 part in 10,000.

    • The paper is accessible behind a paywall, but my institution subscribes via the ‘Elsevier Freedom’ arrangement and I’ve had a look at it. It’s a total LOL shambles that should never have made it into print if reviewed by halfway competent referees.

      The controls (in the portion of the paper that claims significant effects for the Engystol) appear to be the mice that were given no Engystol tablets (in 1L of drinking water) vs 1 tablet/L vs 10 tablets/L. In homeopathic terms, surely the 10 tablets/L should be less effective than 1 tablet/L, but the opposite seems to be the case. I keep using the words’appear’ and ‘seem’ because it’s far from clear. The quality of the cytokine data can be judged by the fact that they are expressed to 5-digit precision (and sometimes more), e.g. 3422.6 ± 1647.2 (example chosen totally at random), where the second figure is an SEM yet, not even an SD!

      The quality of the paper is enhanced by a footnote to the tables with cytokine data “Cytokine/chemokine levels numerically shifted by >/= 15% from control are shown in bold”, when in fact they’re asterisked. Authors who get excited by 15% shifts (up or down) in mean cytokine data are obviously inexperienced with this kind of work!

      Two final points of note.

      1) The Engystol tablets were analysed for metals (which included the well known metals, phosphorus and sulphur). No sulphur was detected, though it’s the main active ingredient listed in Engystol. On the other hand, 22 metals were present at detectable levels, with sodium (2442 ppm), magnesium (447 ppm), calcium (388 ppm), aluminium (200 ppm) and phosphorus (103 ppm) found in the highest concentrations. This list appears to support the suggestion expressed in a comment some time ago (by Guy Chapman?) that homeopathic dilutions will result in products containing more traces of materials leached from the container during succussion than any active ingredient.

      2) As stated in the original post, Engystol contains (from the manufacturer’s website) Sulfur D4 37.5 mg; Sulfur D10 37.5 mg; Vincetoxicum hirundinaria D6 75 mg; Vincetoxicum hirundinaria D10 75 mg; Vincetoxicum hirundinaria D30 75 mg. What bizarre homeopathic logic requires (theoretically less active) D4 sulphur in combination with (more dilute therefore more potent) D10 sulphur? Ditto for three ‘potencies’ of Vincetoxicum hirundinaria. Come on, chaps, make up your minds. Either something is more potent the more it’s diluted or it’s not. Perhaps one of the regular homeopathists who post so copiously on this blog can amuse us by explaining the ‘science’ behind this curious combination therapy.

      • I see Engystol is manufactured by Heel GmbH and the second author was ‘supported’ by Heel GmbH.

        Note also that it wasn’t randomised and:

        The choice of treatment was a joint decision of the practitioner and the patient. … To reflect everyday treatment practices, in both groups, the doses were not stipulated in the protocol but were decided for each individual patient for a maximum of two weeks. No limit was set to the number of different therapies in the control group. In the homeopathic treatment group, patients were allowed other short-term medications, but the long-term use of analgesics, antibiotics, and antiinflammatory agents was not permitted.

        The protocol allowed for the use of additional therapies not included among the defined study medications, and both groups made use of such remedies. In the homeopathic treatment group, menthol- or camomile-based inhalations were used by 57.7% of patients, vitamins by 37.7%, sympathomimetic decongestants by 27.4%, and antipyretics/analgesics by 23.4% of patients. The most common nonstudy therapies in the control group were cough remedies (antitussives/expectorants) (59.0%), menthol- or camomile-based inhalations (53.5%), vitamins (34.7%), and decongestants (24.3%).

        So, although they say the protocol allowed for those in the ‘verum’ arm to take conventional medications in the ‘short-term’ there is no information (that I can see) that specifies what ‘short-term’ means nor any compliance data.

  • Edzard wrote

    ‘In other words, it is one of those combination remedies which are not sufficiently dilute to be devoid of active molecules. Sulphur D4, for instance, means that the remedy contains one part of sulphur in 10 000 parts of diluent. It is conceivable, even likely that such a concentration might affect certain immune parameters, I think.’

    First of all the spelling used by Heel for Engystol® from I have seen online is Sulfur not Sulphur.
    Check out on Wikipedia that an egg contains about 200µg of Sulfur. If each tablet of Engystol® contains 37.5mg of sulfur D4 then each tablet contains 3.75µg of Sulfur. Take the max daily dose and you dont even get 50µg of Sulfur per day. You get around 25% Sulfur that you would get from eating an egg.
    Therefore I would suggest that it is unlikely that any immune parameters are affected by Engystol®.
    Any affects are then therefore obviously due to magic. Please bring in Dr Rawlings for an explanation!

    • @SulfuricAcid

      Please don’t let the different spellings of a word in UK and US English confuse you. The formulation of Engystol is a far more confusing issue. Your conclusion that it’s unlikely any immune parameters are affected by Engystol is undoubtedly correct. Your statement “Any affects are then therefore obviously due to magic” is also correct, except for your spelling of the noun ‘affects’, which should be ‘effects’ in both US and UK English.

      Homeopathic dilutions are always prepared from a ‘mother tincture’, not from a pure substance. In the case of sulfur (let’s stay with US spelling) the mother tincture consists of a saturated solution of sulfur in ethanol ( Since sulfur is virtually insoluble in ethanol, we don’t know if this description is correct, or if a homeopath just lobs some sulfur into ethanol and works with the resulting suspension. Bottom line: we don’t know the concentration of sulfur the manufacturer starts from.

      I have hunted on line and can find no indication of how we should even interpret e.g. “37.5 mg of sulfur D4”. Since D4 sulfur is a fluid, I presumed that the solution was mixed with however much lactose forms the base of the tablet so there’s 37.5 mg (or 37.5 µL) of the dilution in the final product.

      So we have two of us reading this blog who require enlightenment. Can a homeopathist please explain precisely what “1 tablet contains … Sulfur D4 37.5 mg” really means?!

      • Wouldn’t that be a D4 solution dropped into a glas of sugar pillules and left to dry?
        Also, is not the 1/10 000 “potentised” solution made from a diluted base solution so the concentration is even lower in the end product?

        We know from the FDA’s report of their inspection in 2012 at Nelson’s homeopathic production facilities, how haphazard this manufacturing process can be. The machines missed every sixth vial and the “active” ingredient was not uniformly distributed as the top pillules received more of it than the rest. Nevertheless, no one ever complained of lack activity, or of overactivity as far as is known 😀
        Assuming (wrongly?)that the drop or few drops of 1/10 000+ Sulphur solution does saturate all the pills evenly, the dose of sulphur per pill would be further divided by the number of pillules in the vial, right?

        • “Wouldn’t that be a D4 solution dropped into a glas of sugar pillules and left to dry?”

          That’s what I assumed, too. (“I presumed that the solution was mixed with however much lactose forms the base of the tablet so there’s 37.5 mg (or 37.5 µL) of the dilution in the final product.” I should have added “on average”.) But why do a few homeopathic manufacturers insist on specifying quantities of (ahem) “active” ingredients to 3-digit precisions, just like normal pharmaceuticals? Most of the homeopathic products I see just say “contains X 30C” on the box. Are the 37.5 mg people complying with some national regulation?

          I think we should be told.

      • You are correct about effects/affects. The spelling of Sulphur was changed to Sulfur in all Pharmacopoeia in 2014- or was it 2013?

        A bit of internet searching can show that Sulfur D4 is manufactured according to the German Homoeopathic Pharmacopoeia by adding 1 part of Sulfur to 10000 parts of 86% Ethanol. Therefore it is perfectly clear what 37.5mg of Sulfur D4 is regarding a liquid formulation.
        Regarding a tablet formulation I note that Sulfur can be manufactured under method 6 of the German Homoeopathic Pharmacopoeia. This on search corresponds to a Ph Eur method where Sulfur can be triturated to D4. Therfore it is perfectly clear what 37.5mg of Sulfur D4 is regarding a tablet formulation.
        It should all be clear now.

        • Thanks a lot for the enlightenment!

        • @SulfuricAcid, pardon my ignorance – this may be perfectly clear to you, but I am still struggling. Please help me understand.

          “… Sulfur D4 is manufactured according to the German Homoeopathic Pharmacopoeia by adding 1 part of Sulfur to 10000 parts of 86% Ethanol. Therefore it is perfectly clear what 37.5mg of Sulfur D4 is regarding a liquid formulation.”

          So does this mean:
          (a) 37.5 mg of sulfur was added to 375 g (or perhaps Liters?) of 86% Ethanol?

          “Regarding a tablet formulation I note that Sulfur can be manufactured under method 6 of the German Homoeopathic Pharmacopoeia. This on search corresponds to a Ph Eur method where Sulfur can be triturated to D4. Therfore it is perfectly clear what 37.5mg of Sulfur D4 is regarding a tablet formulation.
          It should all be clear now.”

          So does this mean:
          (b) 37.5 mg of sulfur was added to 375 g of lactose and then compressed into a tablet?

  • This paper is a terrible failure of peer review. I will make some comments here which I hope will point out some of the flaws that should have been picked up by peer reviewers. This paper should never have been published.
    My principal objection is that, in the final “vaccination” phase there are 4 animals per group and bloods are only taken once. I’ve just returned from the PIVAC conference where yet again there was some discussion about the variability in responses to vaccine even in congenic mouse strains. Intuitively 4 mice per group would not be enough to demonstrate effects of vaccine. There should have been a power calculation to determine the number of mice and that should have been based on previous literature. None of that was done. I would add that this whole mouse experiment was done only once. There is no validation of the original data with a second experiment.
    The vaccine itself is a mystery to me. It was “generously donated” by Dr Biou He. That in itself is confusing since Dr He is one of the authors. Where is the evidence that the vaccine has been characterised elsewhere? There is no reference to any prior literature on the vaccine. That is not acceptable. Likewise I am clueless about the route of injection. That it went into the hind leg does not tell me whether it is intramuscular, subcutaneous or intradermal. Route of injection is important in vaccination and should have been clearly stated.
    A great deal of the paper relies on flow cytometry data (a methodology that uses fluorescently labelled antibodies to detect the expression of cell surface markers). The methodology lacks gold standard controls (isotype controls) without which it is inconceivable that the authors could have picked out different lymphocyte populations with reliability. There are also no examples of their gating (an arbitrary decision made by the experimenter using flow cytometry but which is based on understanding of the controls in the experiment). Without example data I am forced to disbelieve any of the highly manipulated data. The data generated by flow cytometry appears to be presented as percentages. Of what we are never told. There must be a baseline cell population which the authors assume will not change but without knowing what that is we cannot make sense of the percentages. Data in the first table is presented as mean +/- SEM. Given that these data represent 2 mice I cannot think what a mean would tell us and how an SEM was calculated I can’t begin to guess. Altogether I don’t find that I can accept any of the flow cytometry data
    Overall the reliance on SEM as the key statistical tool hides the essential variability in the data. The authors should have used SD. I blame the journal for this since they should have insisted on proper use of stats. In figure 1 the authors present ELISA data but fail to explain their error bars. I also cannot find the key statistical comparison for the combination (homeopathy + vaccine) compared to vaccine alone.
    I could go on but I haven’t the energy. I gave up on this paper at this point since nothing thus far is credible. If you want to believe in homeopathy look elsewhere for your proof because this paper doesn’t look like it proves anything at all.

    • Thanks for your assessment John. Overall it feels like cargo cult science, as coined by Feynman. Lots of science-like things going on (cytokine panels, IgG titers plus T cell and B cell phenotypes) but without understanding or anything real to analyse.

    • Agree with everything you wrote, John, with two clarifications (Disclaimer: I haven’t read the paper since I have no access to it):

      1. It is possible to obviate the need for Isotype controls by using FMO (Fluorescence Minus One) controls, in which fluorescent antibodies for all surface antigens except for the cognate antigen (in that set) are added. This can help produce sharper gating. (No clue if this is what the authors did, though.)

      2. The percentages in the Flow Cytometry data are most likely generated by the serial or parallel gating of the cell population based on fluorescence channels. This technique is prone to errors, and therefore, all gates must be cautiously and justifiably applied. (Again, no clue how the authors handled this.)

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

If you want to be able to edit your comment for five minutes after you first submit it, you will need to tick the box: “Save my name, email, and website in this browser for the next time I comment.”
Recent Comments

Note that comments can be edited for up to five minutes after they are first submitted.

Click here for a comprehensive list of recent comments.