Authors' replyBMJ 1995; 311 doi: https://doi.org/10.1136/bmj.311.7014.1227a (Published 04 November 1995) Cite this as: BMJ 1995;311:1227
- Michael Bjorn Russell,
- Jes Olesen
- Research fellow Professor Department of Neurology, Glostrup Hospital, University of Copenhagen, DK-2600 Glostrup, Denmark
EDITOR,--Since migraine without aura aggregates in families of probands with migraine without aura, it is expected that migraine without aura is less common in families of probands who have never had migraine. We calculated the population relative risk, adjusting for sex and age, since migraine without aura depends on these factors. If the population relative risk significantly exceeds one an increased family aggregation is implied, whereas a result significantly below one implies a decreased family aggregation.
W E Waters states that the risk of migraine without aura among the first degree relatives of probands who had never had migraine was expected to be considerably lower than the figure we found. This is not correct. We calculated that the expected population relative risk was 0.87 (95% confidence interval 0.62 to 1.12) and found that it was 1.11 (0.83 to 1.39). These results are not significantly different. Thus random variation may explain our result. The calculation of the expected population relative risk was based on the increased risk of migraine without aura in first degree relatives of probands with migraine without aura, the overall prevalence of migraine without aura, and a similar number of first degree relatives of probands who had never had migraine to that in our study.
Waters suggests that the disclosure of the objective of the family study could have biased the study, since the overall prevalence of migraine was higher in the family study than in the general population.1 Firstly, these prevalences are not comparable, since migraine aggregates in some families and the prevalence of migraine in first degree relatives depends on the relative proportion of probands with migraine and probands who have never had migraine. Secondly, responders and non-responders were not significantly different regarding migraine. A minority of probands did not allow their families to be interviewed, but this is not likely to have caused bias in the family study since migraine as assessed by probands' report was not significantly different in these families and those included in the family study. Probands who had never had migraine had a slightly higher average number of first degree relatives than probands with migraine, which is probably due to chance.
Although Waters pioneered epidemiological studies of migraine, the methods section in his study that we cited is not completely clear. It states that “relatives were visited and a short administered questionnaire on headache was completed.”2 This can easily be taken as indicating that the relatives, rather than the trained interviewer, filled in the questionnaire. Our results indicate the importance of genetic factors in migraine without aura and migraine with aura. This is supported by previous twin studies.3 Studies of unselected twins interviewed blind by a physician will, however, be important, and the final confirmation will be identification of the gene(s).