The risk of developing postmenopausal breast cancer from ambient exposures to selected volatile organic compounds

Abstract

ObjectivesTo estimate associations between incident postmenopausal breast cancer and environmental concentrations of five ambient volatile organic compounds (VOCs): benzene, hexane; ethylbenzene; n-decane; and 1,2,4-trimethylbenzene. MethodsWe conducted in Montreal a hospital-based case-control study between 1996 and 1997 and a population-based study between 2008 and 2011 of incident postmenopausal breast cancer. We measured five VOCs at about 130 locations during three sampling periods in 2005–2006 and we then used Bayesian geostatistical models to estimate concentrations of the VOCs at residential locations of participants in the two studies at the time of recruitment. We developed a directed acyclic graph to account for confounding but because of concerns with mis-specified models we used a series of unconditional logistic regression models to adjust for different combinations of personal risk factors, occupational exposures attributed by experts, air pollution data, and contextual variables. We estimated odds ratios (OR) for increases in concentrations across the interquartile ranges of the VOCs and associated 95% confidence intervals (CI). ResultsIn each study, we found that the ORs for each VOC were similar across the statistical models. In the hospital-based study, the ORs for the five VOCs varied between 0.77 and 1.04. In the population-based study, higher ORs were found for three VOCs: benzene, ORModel-7 = 1.12 (95%CI: 0.95, 1.31); hexane, ORModel-7 = 1.11 (95%CI: 0.97, 1.27); and 1,2,4-trimethyl-benzene, ORModel-7 = 1.07 (95%CI: 0.88, 1.31). We did not find important differences in sub-analyses by hormonal phenotype. ConclusionsIn the hospital-based study we did not find evidence of associations with these VOCs but in the population-based study there were indications of positive associations for environmental exposures to benzene and hexane.