Venous Thromboembolism in an Industrial North American City: Temporal Distribution and Association with Particulate Matter Air Pollution

BACKGROUND: Emerging evidence, mainly from Europe and Asia, indicates that venous thromboembolism (VTE) occurs most often in winter. Factors implicated in such seasonality are low temperature-mediated exacerbation of coagulation and high levels of particulate matter (PM) air pollution. However, in contrast to most European and Asian cities, particulate matter pollution peaks in the summer in many North American cities. OBJECTIVES: We aimed to exploit this geographical difference and examine the temporal distribution of VTE in a cold-weather, North American city, Detroit, with a summer PM peak. Our goal was thereby to resolve the influence of temperature and PM levels on VTE. METHODS: Our retrospective, analytical semi-ecological study used chart review to confirm 1,907 acute, ambulatory VTE cases, divided them by location (Detroit versus suburban), and plotted monthly VTE frequency distributions. We used Environmental Protection Agency data to determine the temporal distribution of PM pollution components in Detroit. Suburban PM air pollution is presumed negligible and therefore not monitored. RESULTS: Acute VTE cases in Detroit (1,490) exhibited a summer peak (June 24(th)) and differed from both a uniform distribution (P<0.01) and also that of 1,123 no-VTE cases (P<0.02). Levels of 10 µm diameter PM and coarse particle (2.5 to 10 µm) PM also exhibited summer peaks versus a winter peak for 2.5 µm diameter PM. Contrary to their urban counterparts, suburban cases of acute VTE (417) showed no monthly variation. CONCLUSIONS: The summer peak of acute VTE in Detroit indicates that low temperature is not a major factor in VTE pathogenesis. In contrast, concordance of the 10 µm diameter PM, coarse particle, and the Detroit VTE monthly distributions, combined with no monthly suburban VTE variation, is consistent with a role for PM pollution. Furthermore, divergence of the VTE and 2.5 µm PM distributions suggests that particle size may play a role.