Long-term exposure to black carbon in outdoor air pollution is associated with cancer incidence

Abstract

PDS 76: Source specific outdoor air pollution studies, Exhibition Hall (PDS), Ground floor, August 26, 2019, 4:30 PM - 5:30 PM Background/Aim: There is growing evidence that outdoor fine particles (PM2.5) cause cancer, but the individual effects of their many components remain to be understood. We investigated the relationships between black carbon (a by-product of combustion mainly from anthropogenic sources) and cancer incidence in Gazel, a large French population-based cohort. Methods: Gazel included 19,349 participants followed up until 2015, among which we identified 4,042 cases of incident all-site cancer and 342 for lung cancer. We estimated exposure to black carbon and PM2.5 using a land-use regression developed for 2010, which we extrapolated yearly to each participant address between 1990 and 2015. Our main exposure metric was the log-transformed cumulative exposure to black carbon with a 10-year lag; we did the same for PM2.5 for subsequent adjustment. We used extended Cox models with attained age as underlying time-scale and time-dependent exposures. The main model included black carbon and adjusted for relevant confounders (including sex, tobacco, alcohol consumption, socio-economic status, calendar time). As secondary model, since black carbon is a component of PM2.5, we used the black carbon:PM2.5 ratio plus an interaction with PM2.5. Results: We found positive associations with all-site and lung cancer incidence, with respective hazard ratios of 1.17 (95% confidence interval 1.06-1.29) and 1.41 (1.03-1.94) per increase in log-transformed cumulative black carbon exposure. When adjusted for PM2.5, we found hazard ratios of 1.51 (1.12-2.02) and 2.22 (0.74-6.71) for a 10% increase in the black carbon:PM2.5 ratio for all-site and lung cancer incidence, respectively. Conclusions: These findings highlight black carbon as a potential candidate to better understand the effects of air pollution on cancer incidence.