Genome-Wide Investigation of Exogenous Female Hormones, Genetic Variation, and Venous Thromboembolism Risk
Résumé
BACKGROUND: Increased risk of venous thromboembolism (VTE) is a life-threatening side effect for users of oral contraceptives (OCs) or hormone therapy (HT). OBJECTIVES: To investigate the potential for genetic predisposition to VTE in OC or HT users, we conducted a gene-by-environment (GxE) case-only meta-analysis of genome-wide association studies (GWAS). METHODS: Use or non-use of OCs (7 studies) or HT (8 studies) at the time of the VTE event was determined by pharmacy records or self-report. A synergy index (SI) was modeled for each variant in each study and estimated supra-multiplicative GxE interaction. The SI parameters were first meta-analyzed across OC and HT studies, and subsequently meta-analyzed to obtain an overall estimate. The primary analysis was agnostic GWAS and interrogated all imputed genotypes using a p-value threshold of <5.0x10(-8); secondary analyses were candidate-based. RESULTS: The VTE case-only OC meta-analysis included 2,895 OC users and 6,607 non-users; the case-only HT meta-analysis included 2,434 HT users and 12,793 non-users. In primary GWAS meta-analyses, no variant reached genome-wide significance, but the smallest p-value approached statistical significance: rs9386463 (p = 5.03x10(-8)). We tested associations for 138 candidate variants and identified 2 that exceeded statistical significance (0.05/138=3.62x10(-4)): F5 rs6025 (p = 1.87x10(-5), SI = 1.29: previously observed) and F11 rs2036914 (p = 2.0x10(-4), SI = 0.91; new observation). CONCLUSIONS: The candidate-variant approach to identify supra-multiplicative associations between genetic variation and OC-HT use identified a new association with common genetic variation in F11 while the agnostic interrogations did not yield new discoveries.