Abstracts

Rationale: The comorbidity of migraine and epilepsy has long been suggested, since these disorders of abnormal neuronal excitability share similar clinical symptoms. Previous studies of this association have produced conflicting results, which may be attributable to methodological issues such as variation in diagnostic criteria, source of study subjects, or methods for analysis. Few population-based studies have addressed this question. Methods: This study assessed the comorbidity of migraine and epilepsy in the Genetic Epidemiology of Seizure Disorders in Rochester (GESDR) study, a population-based study that includes all 910 individuals born in 1920 or later who had incidence of epilepsy or isolated unprovoked seizure while residing in Rochester, MN from 1935 to 1994, as well as 941 controls without unprovoked seizures matched by age, sex, and time period of residency in Rochester. Surviving cases and controls were interviewed by telephone using a validated questionnaire for diagnosis of any migraine, migraine with aura (MA), and migraine without aura (MO), based on the International Classification of Headache Disorders, 2nd edition (ICHD-2). Among all 1851 cases and controls, 529 could not be interviewed because they were deceased, could not be located, or unable to participate due to illness or disability; 34% (N=449) of the remaining subjects were interviewed. Results: Interviewed subjects averaged 53 years of age; 58% were women; and 49% were college graduates. Prevalence of a lifetime history of migraine was higher among females than males for any migraine (21% vs. 10%, p=0.001), MO (11% vs. 4%, p=0.012), and MA (10% vs. 5%, p=0.070). Prevalence of a history of any migraine also declined with advancing age at interview (≤40 years: 22%, 41-60: 17%, >60: 13%, p=0.057, X2 test for trend), and was lower in college graduates than in other subjects (13% vs. 20%, p=0.048). Cases and controls did not differ in age or gender, but fewer cases than controls were college graduates (44% vs. 54%, p=0.032); hence we used logistic regression to compute odds ratios (ORs) for migraine in cases vs. controls, including education in the model to control for potential confounding. Overall, cases and controls did not differ significantly with respect to prevalence of a history of any migraine (OR=1.44 [95% CI 0.87-2.40]), MO (OR=1.69 [0.84-3.42], or MA (OR=1.08 [0.54-2.15]). However, among individuals aged 40 or younger when interviewed, cases were more likely than controls to have had MO (OR=8.68 [1.02-73.63]) but not MA. In analyses restricted to cases, neither seizure disorder (epilepsy vs. isolated unprovoked seizure), age at onset, etiology of epilepsy, seizure type (generalized vs. focal), nor broad epilepsy syndrome was predictive of migraine status. Conclusions: The identification of comorbidity between MO and unprovoked seizure has implications for the diagnosis, management, and treatment of both disorders. The age-specific effect may be attributable to underreporting of migraine by older individuals. The results suggest a link between migraine and any type of unprovoked seizure, rather than epilepsy alone.