Migraine and breast cancer risk: a prospective cohort study and meta-analysis

Abstract

Background: The evidence for an association between migraine and breast cancer risk is inconclusive. While female sex hormones have been proposed as one underlying mechanism, data on sex hormone levels in migraineurs are sparse.

Methods: We prospectively evaluated the association between migraine and breast cancer risk among 115378 Nurses' Health Study II participants using Cox proportional hazards models. Differences in endogenous sex hormone levels according to migraine status were assessed among 2034 premenopausal women using linear regression. We performed a meta-analysis of studies investigating the association between migraine and invasive breast cancer risk published through October 2013. All statistical tests were two-sided.

Results: Seventeen thousand six hundred ninety-six women (15.3%) reported a physician's diagnosis of migraine at baseline. Over 20 years of follow-up, 833 in situ and 3091 invasive breast malignancies occurred. Migraine was not associated with total (hazard ratio [HR] = 0.96, 95% confidence interval [CI] = 0.88 to 1.04), in situ (HR = 0.97, 95% CI = 0.82 to 1.15), or invasive breast cancer risk (HR = 0.95, 95% CI = 0.87 to 1.04). Endogenous sex hormone levels did not differ according to migraine status. In the meta-analysis, migraine was associated with a lower risk of breast cancer overall (pooled risk ratio [RR] = 0.84, 95% CI = 0.73 to 0.98). However, this inverse association was apparent only among case-control studies (pooled RR = 0.72, 95% CI = 0.66 to 0.79), and not among cohort studies (pooled RR = 0.98, 95% CI = 0.87 to 1.10).

Conclusion: In this large cohort study, migraine was not associated with breast cancer risk or differences in endogenous sex hormone levels. While case-control studies suggest an inverse association between migraine and breast cancer risk, prospective cohort studies do not support an association in pooled analyses.

Figure 1.

Forest plot of the association between migraine and invasive breast cancer in five studies stratified by study design. Random effect models were used. Additional information on study characteristics is summarized in Supplementary Table 2. CI = confidence interval; CARE = Women’s Contraceptive and Reproductive Experience Study; NHSII = Nurses’ Health Study II; RR = risk ratio; WHI = Women’s Health Initiative; WHS = Women’s Health Study.