Statistical approaches to analyse the combined effect of seven air pollutants and breast cancer risk: a case–control study nested in the French E3N-Generations cohort

Air pollution is a complex mixture of closely correlated pollutants, making it challenging to assess both the overall mixture effect and to isolate the individual impact of each pollutant on breast cancer (BC) risk. This study assessed the effect of exposure to a mixture of seven correlated air pollutants (benzo[a]pyrene, cadmium, dioxins, polychlorinated biphenyl 153 (PCB153), nitrogen dioxide (NO 2 ), particulate matter (PM 2.5 and PM 10 )) on BC risk. The study was based on a case–control study nested within the French E3N-Generations cohort (5222 incident BC cases/5222 matched controls). Annual average concentrations of each pollutant were estimated using the CHIMERE chemistry-transport model, based on participants’ residential addresses from 1990 to the index date. Bayesian kernel machine regression (BKMR) and quantile G-computation (QGC) were used to evaluate the joint effect of the pollutant mixture, individual pollutant contributions, and potential interactions. In all women, the BKMR model showed an increasing trend in BC risk associated with a joint increase in exposure to the seven pollutants. Among individual pollutants, NO₂, PCB153, and PM showed the strongest positive dose–response associations. The QGC model also found a significant association between the pollutant mixture and BC risk (odds ratio (OR) = 1.12; 95% confidence interval (CI) = 1.02–1.24) per quartile increase in the mixture. This study provides evidence of a positive association between exposure to a mixture of seven air pollutants and the risk of BC for the two statistical approaches. NO 2 contributed most significantly to the overall effect, followed by PCB153 and PM. These findings underscore the necessity of evaluating combined pollutant mixtures in risk assessment, identifying high-risk subpopulations, and designing targeted preventive strategies.