We would like to thank Dr Shaheen for his interest in our recent editorial (Thorax 2007;62:745–6). Shaheen brings up several important points, which we were unable to discuss because of the space limitations. Many of these points have been discussed in our recently published commentary in another journal.1 We agree with one of the points that Shaheen makes—namely, we certainly acknowledge that there is some evidence for a contrary hypothesis that Wjst and Dold put forth, as we have discussed.1 However, there are other points where we disagree.

Firstly, Shaheen states that the Wjst–Dold vitamin D hypothesis was a “result of fortification of foods such as margarine, which may have contributed to the rise in atopy in the West”. Despite food fortification, multiple studies have now shown that vitamin D deficiency is highly prevalent2 even in sun-replete areas of the world3 and that vitamin D supplementation and fortification of foods in current doses are inadequate to prevent deficiency.4 This suggests to us that behaviours that have led to decreased sun exposure have had a larger effect than diet on vitamin D status of human populations.

The two greatest sources of vitamin D are sun exposure and supplements, not diet, except in rare circumstances. Therefore, we would argue that fortification of foods may have improved the vitamin D status of populations just enough to prevent widespread rickets (although not entirely eradicated in Western society5) but did not have sufficient impact on the development of asthma and allergic disorders. Secondly, he suggests that before clinical trials of supplementation are performed, “we need more convincing data to support their hypothesis, and greater confidence that such an intervention would be safe”. A query of the US clinical trials database (www.clinicaltrials.gov) shows that there are at least 218 registered vitamin D studies for various diseases, with at least seven trials investigating vitamin D supplementation in pregnancy, lactation and low birth weight infants.

These studies are ongoing and preliminary data have shown no adverse effects of vitamin D supplementation in pregnancy with doses up to 6000 IU/day.6

Thirdly, Shaheen suggests that we need more observational epidemiology to assess whether the effect observed in the Project Viva and SEATON cohort studies are real. We disagree. More observational epidemiology will never answer this question definitively; what is needed is a randomised controlled trial of vitamin D supplementation during pregnancy, the time period where we observed the large beneficial effect on asthma outcomes in two separate and independent birth cohorts with widely different exposure levels.

Finally, Shaheen suggests that study of vitamin D receptor polymorphisms would strengthen causal inference. While it is not entirely clear how it would do this, we do agree that this is of some relevance. However, genetic polymorphisms in VDR and its binding protein will only explain a small percentage of phenotypic variation in vitamin D levels and will be much less important than environmental factors such as dietary supplements and sun exposure behaviours in determining vitamin D levels, which are primarily a measure of recent rather than chronic exposure.

We regret that the limits of the editorial did not allow a complete review of this interesting and important topic.