Carcinogenic industrial air pollution and postmenopausal breast cancer risk in the National Institutes of Health AARP Diet and Health Study.

BACKGROUND

Chemicals emitted from industrial facilities include known or suspected mammary carcinogens and endocrine disruptors, but epidemiologic studies are limited. We evaluated associations between air emissions of multiple carcinogenic chemicals and postmenopausal breast cancer risk in a large prospective U.S.

METHODS

We used the U.S. Environmental Protection Agency's Toxics Release Inventory to estimate historical airborne emissions (1987-1995) of 19 known and probable carcinogens for participants enrolled (1995-1996) in the NIH-AARP Diet and Health Study. Among 170,402 women, 15,124 breast cancers were diagnosed through 2018. We constructed inverse distance- and wind-weighted average emissions metrics within 1, 2, 5, and 10 km of the enrollment address for each chemical. We estimated multivariable adjusted HRs and 95 % CIs for categories (quartiles, tertiles, medians) of each chemical in association with breast cancer overall and separately by type (invasive, ductal carcinoma in situ) and estrogen receptor (ER) status.

RESULTS

We observed an association between benzene emissions and breast cancer risk that was strongest at 1 km (HRQ4 vs. non-exposed = 2.06, 95 %CI: 1.34-3.17; p-trend = 0.001). The magnitude of the association weakened with increasing distance (2 km HRQ4 vs. non-exposed = 1.17, 95 %CI=0.92-1.49; p-trend = 0.19; 5 km HRQ4 vs. non-exposed = 1.05, 95 %CI=0.94-1.16; p-trend = 0.37; 10 km HRQ4 vs. non-exposed = 0.95, 95 %CI=0.89-1.02; p-trend = 0.19) and appeared to be most relevant for invasive rather than intraductal disease. Overall risk was also elevated for vinyl chloride at 5 km (HR≥median vs. non-exposed = 1.20, 95 %CI=1.01-1.43; p-trend = 0.04), but not 2 km or 10 km. We observed suggestive associations for asbestos, trichloroethylene, and styrene in different subgroup analyses, but risk patterns were not clear across distances. Associations with other chemicals were generally null, with limited evidence of heterogeneity by disease type or ER status.

CONCLUSIONS

An increased risk of breast cancer associated with relatively high levels of industrial benzene emissions warrants additional study, particularly among participants with diverse sociodemographic characteristics that live in areas with higher density of industrial facilities.