Letters Section Editor: Robert M. Golub, MD, Senior Editor.
In Reply: Drs Young and Shaw argue that potential misclassification and residual confounding limit the interpretation of our study on migraine and risk of CVD in women. In this prospective cohort study of more than 27 000 middle-aged women, we found that women who self-reported active migraine with aura but not active migraine without aura had increased risk of overall and specifically ischemic CVD during a mean of 10 years of follow-up. Although misclassification is of concern in any observational study, the issue is whether potential misclassification is differentially influenced by the outcome of interest or not. In our study, migraine and aura information was ascertained at baseline, after which follow-up of incident CVD began. Since it is very unlikely that a future CVD event influenced a woman's response to the migraine or aura question at study entry, any potential misclassification would likely be random (or nondifferential) and would lead to an underestimation of the association between migraine and CVD.
Kurth T, Gaziano JM, Cook N, Logroscino G, Diener H, Buring JE. Association Between Migraine and Cardiovascular Disease in Women—Reply. JAMA. 2006;296(22):2677-2678. doi:10.1001/jama.296.22.2677-b