Abstracts

RATIONALE: Analysis of concordance of epilepsy subtypes in families has been used to assess shared and distinct genetic influences on generalized and localization-related epilepsy (LRE). However, it is unclear how the results should be interpreted in relation to specific genetic hypotheses. METHODS: We examined concordance of generalized and LRE in 49 families containing multiple individuals with idiopathic or cryptogenic epilepsy in a genetic linkage study. Diagnoses were based on semistructured diagnostic interviews with consensus review by senior epileptologists, blind to family history. We developed novel methods for statistical analysis that correct for (1) concordance expected by chance, and (2) the methods used to ascertain families. RESULTS: Overall, 31 out of 49 (63%) families were concordant for generalized (N=13) or LRE (N=18). We formulated two non-mutually exclusive null hypotheses. H01 states that no genetic influence raises risk for only generalized or only LRE, while H02 states that no genetic influence raises risk for both generalized and LRE. To test H01, we computed the expected number of concordant families based on the distribution of epilepsy subtypes in the data set as a whole. Observed concordance was significantly greater than expected; hence H01 was rejected. To test H02, we selected families that would have met the study inclusion criteria based on only one type of epilepsy, and examined the incidence of the other type of epilepsy in the remaining relatives. In the generalized group, only 1/135 (0.7%) individuals had LRE, and in the LRE group, only 1/133 (0.8%) individuals had generalized epilepsy. These numbers are not greater than expected from population incidence; hence H02 was not rejected. CONCLUSIONS: These findings provide evidence for distinct genetic influences on generalized and LRE, but not for shared genetic influences on generalized and LRE. These findings contrast with the results of previous familial aggregation studies. Large sample sizes, ascertainment correction, and population-based designs will be needed in future studies to explore reasons for this discrepancy.