Prediction of Venous Thromboembolism Based on Clinical and Genetic Factors

BACKGROUNDBoth clinical and genetic factors drive the risk of venous thromboembolism. However, whether clinically recorded risk factors and genetic variants can be combined into a clinically applicable predictive score remains unknown. METHODSUsing Cox proportional-hazard models, we analyzed the association of risk factors with the likelihood of venous thromboembolism in U.K. Biobank, a large prospective cohort. We created a novel ten point clinical score using seven established clinical risk factors for venous thromboembolism. We also generated a polygenic risk score of 21 single nucleotide polymorphisms to quantify genetic risk. The genetic score was categorized into high risk (top two deciles of scores), intermediate risk (deciles three to eight), and low risk (lowest two deciles). The discrete clinical score led to the following approximate decile categorizations: high risk (5 to 10 points), intermediate risk (3 to 4 points), and low risk (0 to 2 points). RESULTSAmongst the 502,536 participants in the U.K. Biobank, there were 4,843 events of venous thromboembolism. Analyses of established clinical risk factors and the most commonly used medications revealed that participants were at decreased risk of venous thromboembolism if they had ever used oral contraceptive pills (hazard ratio, 0.88; 95% confidence interval [CI], 0.79 to 0.99) or if they currently used bendroflumethiazide (hazard ratio, 0.84; 95% CI, 0.74 to 0.95), cod liver oil capsules (hazard ratio, 0.87; 95% CI, 0.77 to 0.99), or atenolol (hazard ratio, 0.79; 95% CI, 0.68 to 0.91). Participants were at significantly increased risk of venous thromboembolism if they were at high clinical risk (hazard ratio, 5.98; 95% CI, 5.43 to 6.59) or high genetic risk (hazard ratio, 2.28; 95% CI, 2.07 to 2.51) relative to participants at low clinical or genetic risk, respectively. Combining clinical risk factors with genetic risk factors produced a model that better predicted risk of venous thromboembolism than either model alone (P<0.001). Participants at high clinical and genetic risk in the combined score had over an eightfold increased risk of venous thromboembolism relative to participants at low risk (hazard ratio, 8.27; 95% CI 7.59 to 9.00). CONCLUSIONSBy assessing venous thromboembolic events in over 500,000 participants, we identified several known and novel associations between risk factors and venous thromboembolism. Participants in the high risk group of a combined score, consisting of clinical and genetic factors, were over eight times more likely to experience venous thromboembolism than participants in the low risk group.