Repeated assessments of physical activity and risk of incident venous thromboembolism

Essentials It is debated whether physical activity influences the risk of venous thromboembolism. The association was explored accounting for fluctuations in physical activity over time. Overall and in the elderly, physical activity was associated with 23% and 30% lower risk. A moderate proportion of the association (14-36%) was mediated via body mass index. SUMMARY: Background Whether physical activity influences the risk of incident venous thromboembolism (VTE) remains controversial, potentially because of methodological challenges, such as regression dilution bias. Objectives To investigate whether physical activity was associated with VTE risk, and explore the role of body mass index (BMI) as a mediator in a population-based cohort with repeated assessments of physical activity. Methods Participants (n = 30 002) attending one or more surveys of the Tromsø Study 4-6 (1994-1995, 2001-2002, and 2007-2008) were included and categorized on the basis of weekly physical activity. Incident VTE was registered until 31 December 2016. Hazard ratios (HRs) were calculated by the use of time-varying Cox regression models. The Aalen additive hazard model was used to quantify the total, direct and indirect effects of physical activity. Results There were 531 incident VTEs during follow-up. Physical activity (≥ 1 per week) was associated with a lower risk of VTE (HR 0.77, 95% confidence interval [CI] 0.64-0.92) than being inactive. The effect was most pronounced for those aged ≥ 65 years (HR 0.70, 95% CI 0.55-0.88) and for provoked events (HR 0.66, 95% CI 0.50-0.89). The differences in absolute risk between active and inactive individuals were - 0.42 (95% CI - 0.73 to - 0.14) and - 1.59 (95% CI - 2.74 to - 0.52) events annually per 1000 individuals in the total and elderly populations, respectively. A moderate proportion of the association (14-36%) was mediated via BMI. Conclusion Our findings suggest that regular physical activity is associated with a lower risk of VTE, particularly in the elderly. The association occurred at a low weekly amount of physical activity, and was only partly mediated by BMI.