Polygenic risk scores and current exogenous estrogen use on the risk of venous thromboembolism in the Heart and Vascular Health Study: a case-only approach

Venous thromboembolism (VTE) is a serious condition characterized by blood clot formation in veins. Exogenous hormone use is known to increase VTE risk, and understanding the interplay between genetic variants and hormone use can enhance our understanding of VTE development. Gene-environment interactions (GxE) can reveal how genetic factors modify the effect of environmental exposures, such as exogenous hormones on VTE risk. Polygenic risk scores (PRS) capture common genetic variation related to complex traits and diseases. Here, we aim to enhance our understanding of the joint contribution of a VTE PRS and exogenous hormones in the development of VTE. Methods: A series of logistic regression models were run in premenopausal and postmenopausal women in the Heart and Vascular Health Study using a case-only approach. Each group was divided and analyzed by their use of exogenous hormones, either oral contraceptives (OC) in premenopausal or hormone therapy (HT) in postmenopausal women. Using a VTE PRS as the predictor and hormone use as the outcome, models were created and adjusted for relevant covariates and additional adjustments were made for specific genetic variants (F5 rs6025 and F2 rs1799963) known to influence coagulation factors and thrombosis risk in women using exogenous hormones. Results: We found that among premenopausal women, the interaction of the PRS with OC use on the risk of VTE was primarily explained by the F5 rs6025 and F2 rs1799963 variants. We found that among postmenopausal women, the interaction of the PRS with HT use on the risk of VTE was unmasked by the inclusion of the F5 rs6025 and F2 rs1799963 variants. It is unclear that case-only model assumptions were met. Summary: The interaction of the VTE PRS with exogenous hormone use was not the same in premenopausal women using OCs and in postmenopausal women using HTs. Our findings need replication and better control of model assumptions.