Background: 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 risk. Advanced statistical approaches are needed to evaluate the joint effect of multiple pollutants and their individual contributions.
Objectives: To assess the effect of exposure to a mixture of air pollutants on breast cancer risk using advanced statistical mixture methods.Methods: This study was based on a case-control study nested within the French E3N-Generations cohort, including 5,222 incident BC cases and 5,222 matched controls. For each participant, annual average concentrations of eight pollutants (benzo[a]pyrene, cadmium, dioxins, polychlorinated biphenyl153 [PCB153], nitrogen dioxide [NO?], ozone, particulate matter [PM2.5 and PM10]) were estimated using the CHIMERE chemistry-transport model based on residential addresses from 1990 to the index date. We applied three complementary statistical approaches: Bayesian Profile Regression (BPR) to identify exposure-risk clusters, Bayesian Kernel Machine Regression (BKMR) to assess the joint effect and potential interactions, and Quantile G-Computation (QGC) to estimate the joint effect of simultaneous increases in pollutant exposures. Models were adjusted for matching factors and relevant confounders.
Results: Among the 21 clusters identified by the BPR model, the cluster characterised by low exposures to all pollutants, except ozone, was taken as reference. A consistent increase in breast cancer risk compared to the reference cluster was observed for three clusters: cluster 9 (odds ratio [OR] = 1.61; 95% credible interval [CrI] = 1.13–2.26), cluster 16 (OR = 1.59; 95% CrI = 1.10–2.30) and cluster 15 (OR = 1.38; 95% CrI = 1.00–1.88), characterised by high levels of NO?, PMs and PCB153. The BKMR model showed an increasing trend in breast cancer risk associated with a joint increase in exposure to the seven pollutants. Among individual pollutants, NO?, PCB153, and PM showed the strongest positive dose-response associations. QGC analyses confirmed a significant association between the pollutant mixture and breast cancer risk (OR per quartile increase = 1.12; 95% CI = 1.02–1.24), with positive weights for the same key pollutants.
Conclusions/Implications: Across three complementary statistical approaches, our findings provide consistent evidence that exposure to a mixture of air pollutants is positively associated with breast cancer risk. NO?, PCB153 and PM contributed most significantly to the overall effect. These results underscore the importance of mixture-based analyses for environmental risk assessment and targeted preventive strategies.