TRIALS INVOLVING MIGRAINE AND FOOD ALLERGY

If foods caused symptoms when they were reintroduced, they were withdrawn again, and the children continued until they had tested all commonly eaten foods. They were then tested again with one of the foods that seemed to cause a bad reaction, but this time the tests were done double-blind – neither the child, the parents nor the experimenter knew which foods were being eaten. To this end, the foods were disguised in a strongly-flavoured puree and supplied in tins to be given to the children at home. The tins were identified by a code number, and it was only when the code was broken at the end of the experiment that it became clear which tin contained the suspect food, and which tin was the placebo – unadulterated puree, that should have caused no symptoms.

For various reasons, some children could not be tested double-blind, but 40 were tested in this way.

The results of this trial were a surprise to many doctors, not least to those involved. Of the 88 children who completed the diet, 78 recovered completely on one or other of the simple diets that were tried during the first stage of the experiment. Another four children ‘improved greatly”, and only six showed no improvement whatever. The response, in other words, was 93 per cent, a staggeringly high figure by any standards. Of the children who improved, eight remained well even when foods were reintroduced, and they continued to be healthy on a normal diet. That left 74 children for whom particular foods could be identified as a cause of migraine in open trial. Of the 40 who were then retested double-blind, 35 were made ill by the tin containing the suspect food, but not by the placebo tin. Three reacted to neither tin, and two were made ill by the placebo. Given the vagaries of food reactions generally, 35 out of 40 is an impressive response. The results would have been more convincing if the children could have been tested for all the foods to which they reacted on open trial, and if they could have been tested more than once, but the practical difficulties of doing this in such a large-scale trial are obvious.

*112\180\8*