Modifiable risk factors, blood proteins, and venous thromboembolism

Abstract

Venous thromboembolism (VTE) refers to blood clots in the veins, which is an underappreciated vascular disease that can cause disability and mortality. Although some triggers for VTE (e.g., surgery, fracture, infection, hospitalization, and cancer) have been established, the associations of modifiable risk factors and blood proteins with the risk of VTE remain uncertain. This PhD project aimed to 1) investigate the associations of obesity and lifestyle factors with VTE risk; 2) explore the associations of blood proteins with VTE risk; and 3) establish protein pathways linking modifiable risk factors to VTE development.

In Paper I, we explored the associations of overall and central obesity with the risk of VTE using both cohort and Mendelian randomization analyses. We found a potentially causal association between obesity and VTE risk. Waist circumference might be a preferable indicator linking obesity to VTE. Around 12.4% and 23.7% of VTE cases could be prevented if the population maintained a healthy body mass index and waist circumference, respectively. In Paper II, using the prospective cohort design, we investigated the associations of cigarette smoking, alcohol and coffee intake, physical activity, and diet with the risk of incident VTE. We found that high levels of physical activity and a healthy diet were associated with lower VTE risk in women and men. Cigarette smoking showed a positive association with VTE only in women. Alcohol and coffee intake was not associated with VTE. In Paper III, we explored the association between ultra-processed food intake and the risk of VTE using the prospective cohort design. A higher ultra-processed food intake was associated with a moderately increased risk of VTE. This association was not modified by age, sex, or body mass index. In Paper IV, we conducted a prospective cohort and Mendelian randomization study to estimate the associations of 257 blood proteins with VTE risk. The cohort analysis identified 21 blood proteins associated with incident VTE. Machine-learning analysis found that body mass index and von Willebrand factor shared an identical highest ranking concerning the contribution to the prediction model. Mendelian randomization analysis confirmed 7 protein-VTE associations. In Paper V, we performed a two-stage network Mendelian randomization analysis to decipher proteomic pathways underlying the associations of 15 modifiable risk factors with VTE. We found that several proteins, in particular annexin II and coagulation factor XI, mediated the associations of obesity, smoking, and insomnia with VTE. Proteome-wide Mendelian randomization analysis identified many VTE-associated proteins with druggable potentials.

In summary, the above five studies identified modifiable risk factors and blood proteins for VTE development and further revealed protein pathways underlying the associations between modifiable risk factors and VTE. These findings may deepen understanding of VTE pathogenesis and facilitate precision prevention and drug development for VTE.