Synergistic effects of ABCG2 Q141K variant in combination with alcohol consumption and male sex on gout risk in a rare-event Taiwanese cohort.

Gout is influenced by genetic and lifestyle factors, with ABCG2 variants (rs2231142 Q141K and rs72552713 Q126X) implicated in urate transport and susceptibility. Standard statistical models often yield biased estimates in rare-event cohorts. This cross-sectional study (N = 324, 15 gout cases) used Firth-corrected logistic regression to mitigate small-sample bias and address complete separation in ABCG2 genotyping data. Model performance was evaluated via internal validation with 1000 bootstrap iterations. Q126X was nearly monomorphic (T allele 0.15%), while Q141K showed higher diversity (A allele 28%). Male gender was the strongest predictor (Firth OR 8.1-11.9, P < 0.001), followed by ABCG2 dysfunction (dose-dependent risk, P = 0.055 for severe dysfunction) and alcohol consumption (Firth OR 5.26 for infrequent drinking, P = 0.022). Firth regression successfully corrected upward ML bias (e.g., alcohol OR: 8.30→5.26). The integrated model achieved an apparent AUC of 0.857 and an optimism-corrected AUC of 0.818 via bootstrap validation, demonstrating synergistic effects of genetic, demographic, and lifestyle factors. Robustness was further confirmed by high E-values (e.g., 9.99 for alcohol). These findings illustrate the utility of Firth regression for bias-reduced risk quantification in rare-event studies.