This article originally appeared in The Skeptic, Volume 2, Issue 5, from 1988.
There has recently been quite a controversy in the scientific press which involves this very question. An article by Dr Jaques Benveniste and his colleagues (Nature vol. 333, June 1988) claimed that solutions which statistically did not contain any antibodies (‘solute-free solutions’) still retained the activity of the antibody that had once been dissolved in them.
The claims surround a model which has been used to study allergic disease activity in vitro. This model involves a specialised type of white blood cell called the mast cell. In people with allergic disease, such as hay fever or certain types of asthma, this cell is coated with allergy-mediating antibodies called immunoglobulin E, or IgE for short. When two or more of these antibodies are cross-linked on the mast cell surface (usually by the substance, called the allergin, that the patient is allergic to) they cause the mast cell to release certain chemicals stored in granules inside the cell. Histamine is the most notable of these, causing many allergic symptoms such as watery eyes, sneezing, itching and so on. This is why allergic diseases are sometimes treated by anti-histamine therapy.
Basophils work in the same way as mast cells and can be easily isolated from blood, forming a suitable way to study allergy in the test tube. Instead of an allergen, antibodies directed against IgE (anti-IgE) can be used to cross-link surface IgE molecules and cause basophil granules to be released experimentally. In addition to measuring histamine release, which is technically quite difficult, granule release can also be measured by staining the cells. Cells which contain granules stain red, but those that have released their chemical packets are colourless, providing a simple means to quantify allergy (and its treatment) in vitro.
The controversy surrounding this affair stems from Benveniste’s attempts to answer the question ‘How little anti-IgE is necessary to cause basophil granule release?’, or rather the answer that these experiments provided. He obtained a solution of anti-IgE antibodies and diluted them in successive ten-fold dilutions. As one would expect, the more he diluted them the weaker their activity became, until finally they caused no more granule release than that observed when the cells were put in diluent only.
However, Benveniste continued to dilute the antibody solution still further, and to his surprise (and that of many others), the capacity to cause basophil degranulation recurred. As he diluted still further this capacity disappeared and then reappeared in cyclical fashion down to a dilution of 10140 when the last maxima of release was noted. This means he was observing the effect of the original antibody many millions of times after the last antibody should have been diluted out and lost. Put simply, a solution that statistically no longer contained antibody exhibited antibody activity.
This was not the only claim made in this extremely bold paper. He claimed that this effect could occur with other substances besides antibodies, that the effect depended on the nature of the diluent (it had to be polar) not the initial substance diluted, and, most important of all, the diluent had to be vortexed violently for at least 12 seconds for this effect to be transferred. This practice is commonly used by homeopaths when they dilute their substances, and they also claim it is fundamental to the efficacy of their treatments.
So astonished were the editors of Nature that they only agreed to publish these results if they could be ratified by other laboratories, and if the author would agree to an examination of the technique in practice by a team of ‘specialists’ of their own choosing (which turned out to include James Randi). The experiments were repeated by six other laboratories in four different countries and all confirmed the original findings, and so subject to an independent examination and with several disclaimers the article was published.
Approximately four weeks later, Nature printed the independent investigation results pronouncing the ‘High-dilution experiments a delusion’ and reported that the ‘phenomenon described’ was ‘not reproducible in the ordinary meaning of that word.’
The crux of their decision was that the trial had not been designed properly. First of all the experiments were not properly ‘blinded,’ and in experiments where adequate blinding had taken place the results were always negative (three times). In other words the experimenter who read the results (which are subjectively interpretative) knew in advance which tubes contained which dilution and could therefore have been influenced in her findings. When she did not know this information her findings were negative.
Secondly, Benveniste had chosen not to take account of those times that the experiment did not work. There were also claims that the sampling procedure was not adequate, leaving the experiment open to sampling error.
Finally, Benveniste’s results were ‘too perfect’. In his control experiments where cells were treated with water alone they should not degranulate except spontaneously. The published results reflected this with a release rate of 0-30%. However, as both the pre- and post-treatment cell numbers were counted by eye, one would expect that on some occasions more cells would be counted as staining red after the treatment than there were before – i.e., there should be some negative figures. The fact that there were not suggests that some ‘massaging’ of the data took place.
Nature obviously allowed Dr Benveniste the last say in the matter, and unfortunately his reply was acrimonious to say the least. He obviously felt that he had been the subject of a ‘witch-hunt’ where certain demands he had made had not been complied with, and that the principal worker had been forced to carry out an excessive workload in an unsatisfactory atmosphere of distrust which precluded her normal functioning. His argument maintains that this phenomenon is so impossible that it should never occur at all, and that if it only occurs one quarter of the time that this experiment is still worthy of further study. He further claims that if all unusual research is subjected to this torrent of abuse by the ‘orthodoxy’, then no original thought will ever come out of recognised scientific establishments.
Clearly, all of this leaves much to be desired. The most satisfactory course of events would have been to explain this phenomenon without recourse to experimental design or statistical methodology, either of which implies incompetence or dishonesty on the part of the experimenter. Jacques Benveniste is an accomplished scientist with an international reputation, and has nothing to gain by making spurious claims. However, if an experiment is poorly designed, subject to observer bias (however unintentional), and not repeatable then clearly there is no ‘result’ to explain.
The problem lies in that this does nothing to resolve the continuing conflict between homeopathic practitioners and ‘established’ medicine. Homeopaths will clearly feel that the major issues raised by this paper have not been addressed, ie that substances diluted far beyond a pharmacological dose can still exhibit specific activity and the retraction will be held as yet another example of a ‘hatchet job’ by the unenlightened. Furthermore, I, in common with many colleagues, consider that Nature had no right to publish a paper which they clearly believed to be untrue from the very start. This unfortunately appears to be an attempt to create sensationalism at Benveniste’s expense, and serves only to trivialise the issues that such research raises.
The original paper and the subsequent retraction have caused considerable reaction in both the scientific and lay press, and accusations, implications and conclusions have been flying wildly about. What is clear is that the story will not end here, and that Benveniste’s work neither proves nor disproves the principles behind homeopathy.
For my part, I would only say that this data is still too speculative and unproven to be used as validation (for which it will most certainly be quoted) of scientifically unproven claims.
