Abstracts

Rationale: Two different approaches may be used to obtain information about seizure disorders in families. First, in the “family history” approach, information on seizure occurrence in family members is obtained indirectly, by interviewing patients or other relatives. Second, in the “family study” approach, each relative is interviewed directly to determine his or her seizure history. For many disorders, the family history approach is insensitive - many truly affected family members are missed. However, the accuracy of family history data on seizure disorders is uncertain. Methods: This study used data from the Genetic Epidemiology of Seizure Disorders in Rochester study (GESDR), a population-based investigation using Rochester Epidemiology Project resources. GESDR included all 910 Rochester residents with incidence of unprovoked seizures from 1935-1994 (case probands) and their families. For each case proband, we selected a control proband matched by age, gender, and length of Rochester residency. Occurrence of seizure disorders in the first-degree relatives of case and control probands was ascertained by review of the medical record of each individual, rather than by interviewing the probands. Independently of the medical record review, surviving case and control probands were interviewed by telephone about seizure disorders in themselves and each of their first-degree relatives, using an 8-question screening questionnaire. Interviewers were blinded to case-control status. Results: Sensitivity was defined as the proportion of subjects who screened positive among those with medical record-documented epilepsy or isolated unprovoked seizure; specificity was defined as the proportion who screened negative among subjects determined to be seizure-free through record review. When probands reported on their own seizure histories, sensitivity was 94% (157/168) for epilepsy and 85% (46/54) for an isolated unprovoked seizure; specificity was 94% (203/217). Family history reports were obtained for 68 relatives with medical record-documented unprovoked seizures and 2856 relatives without unprovoked seizures. Sensitivity of the family history was 62% for epilepsy (32/52) and 38% for isolated unprovoked seizure (6/16); specificity was 98%. Sensitivity of family history reports for epilepsy did not differ significantly between case and control probands (55% vs. 74%, p=0.17), but was higher for female vs. male probands (74% vs. 43%, p=0.02) and for probands reporting on their offspring (90%) vs. siblings (76%) vs. parents (33%) (p=0.002). Conclusions: These data suggest that family history reports are insensitive for epilepsy, as for other disorders. Our screening interview is highly sensitive for identifying people with epilepsy when administered directly, but has lower sensitivity for identifying affected family members when administered indirectly. Only 62% of first-degree relatives with epilepsy were reported to be affected in proband interviews; the remaining 38% were missed. Genetic studies should screen each family member directly to ascertain seizure disorders rather than relying on family history reports. (Supported by NIH R01 NS043472)