Abstracts

Rationale: Many epilepsy genetic association studies have been published in the past decade, but no convincing susceptibility gene has been identified. We have created the Epilepsy Genetic Association Database (epiGAD; at www.epiGAD.org). This is an open-access, web-based database of genetic association studies, which functions as a tool for collating data and for future collaborative genetic meta-analyses, using published and unpublished data. Harnessing epiGAD database structure, we present data from published epilepsy genetic association studies, 1997-2007. Methods: We identified published genetic association studies using 4 search engines (Pubmed, Google Scholar, HuGE Navigator, ISI Web), and standardized search terms. Data was extracted from each study using a standard form, capturing information about phenotype, gene, allele, number of cases and controls, source of controls, and country of origin. Data was then uploaded to epiGAD and converted into an online, open-access, searchable database. Results: A total of 145 studies examining 85 genes were identified; 121 studies examined 71 epilepsy susceptibility genes, 24 examined 14 genes related to pharmacogenomics. These studies were generally small in patient numbers (number of cases 155±169, controls 200±171, expressed as mean±SD); the majority (81%) recruited <200 cases. The majority (78%) were published in 2003 or later; these studies recruited more cases compared to those published before 2003 (mean number of cases:164 vs 112, p<0.01). Most studies (90%) used population-based controls; 4% used family-based controls, 6% used both types of controls. Four studies were in non-English language journals; 2 were not found on Pubmed. The phenotype examined in the 121 studies involving putative susceptibility genes was: generalized epilepsy (50%), focal epilepsy (24%), febrile seizures (22%) and all types of epilepsy (4%). Of the 24 pharmacogenomic studies, the majority (54%) examined the relationship between the ABCB1 gene and refractory epilepsy. Twenty-nine studies reported an allelic association with p≤0.01; all were followed by studies that failed to replicate the initial association. Six gene-disease associations were examined by 4 or more independent studies; these may be potential candidates for meta-analyses. Conclusions: Most published epilepsy genetic association studies are small and underpowered, despite attempts by more recent studies to increase sample size. Despite multiple studies examining varying gene-disease associations, no association has been convincingly replicated by 2 or more subsequent studies. There is a pressing need for research groups to collaborate to increase sample sizes - either by prospectively combining cohorts, or collaborative meta-analyses. We believe epiGAD provides an online platform for facilitating such collaboration.