Credibility of genetic predictors for antiepileptic drug resistance: A systematic Bayesian reappraisal of published meta-analyses

We systematically reappraised meta-analyses of pharmacogenetic studies to evaluate the credibility of association between gene polymorphisms and resistance to anti-epileptic drugs (AEDs). A systematic search was performed in PubMed, Web of Knowledge, Cochrane Library and OpenGrey up to April 2025. The methodological quality of the included systematic meta-analyses was evaluated with the AMSTAR-2 tool, and the credibility of the genetic comparison results was determined by the Venice criteria and two Bayesian analytic approaches, false positive reporting probability (FPRP), and Bayesian false discovery probability (BFDP). Of the 33 studies identified, 32 were systematic meta-analyses, all of which were rated as critically low quality by AMSTAR-2. Our reassessment indicated seven single nucleotide polymorphisms (SNPs) of four genes—ABCB1 (rs1045642, rs2032582), ABCC2 (rs717620, rs3740066), GABRG2 (rs211037) and SCN1A (rs2298771, rs10167228)—which could be regarded as potential determinants of response to AED. Among these, only ABCB1 rs2032582 (G vs. A and GG vs. GA + AA) was found to be noteworthy in Caucasian epilepsy patients under FPRP or BFDP at the pre-specified probability level of 0.001. However, the application of the Venice criteria to such relationships identified as weak the strength of the cumulative evidence for epidemiological relationship because of a potential publication bias. Our findings, illustrating the poor yield of genetic predictors from meta-analyses of candidate gene studies, underscore the need for large-scale genome-wide association studies (GWAS) and subsequent replication studies for identification of robust predictors of resistance to AEDs.