Plasma proteomics enhances genetic risk prediction and biological understanding of venous thromboembolism

Venous thromboembolism (VTE) is a leading cause of morbidity and the third most common vascular cause of death globally. While clinical and genetic risk factors have improved VTE risk stratification, they do not fully capture underlying disease biology. We hypothesized that integrating large-scale plasma proteomics with a validated polygenic risk score (PRS_VTE) would reveal novel biological understanding and improve risk stratification.

In 44,138 UK Biobank participants, we examined associations between PRS_VTE and 2,911 plasma proteins using multivariable linear regressions. We evaluated proteins as predictors of incident VTE using Cox regression models. We employed bidirectional Mendelian randomization (MR) using cis-acting protein quantitative trait loci (cis-pQTLs) instruments and genome-wide association summary statistics to infer potential causality. A protein-based risk score (ProteinRS) was developed with Least Absolute Shrinkage and Selection Operator (LASSO) logistic regression and compared against a clinical risk model.

PRS_VTE associated with 265 proteins after false-discovery-rate correction (FDR < 0.01). Pathway enrichment analysis of these proteins highlighted cell adhesion, cell-junction organization, and transmembrane receptor signaling. Time-to-event analyses identified 354 proteins that predicted incident VTE (FDR < 0.01). MR implicated 14 proteins in potential causal pathways to VTE, while genetic liability to VTE affected the levels of 44 proteins. A 60-protein ProteinRS improved discrimination beyond clinical factors (area under the curve (AUC) 0.664 vs 0.647; P = 0.024). Participants above the 80th percentile of ProteinRS had a 49% higher hazard of VTE (hazard ratio (HR) = 1.49, 95% confidence interval (CI): 1.17-1.90) and those below the 20th percentile had a 48% lower hazard (HR = 0.52, 95% CI: 0.34-0.79) relative to the 20-80th percentile.

Integrating genetic and proteomics data uncovered molecular pathways that link genetic susceptibility to VTE, highlighted actionable protein targets, and modestly enhanced risk prediction. Plasma proteomics therefore holds promise for precision prevention of VTE.