Maya Spaur, Lauren M Hurwitz, Danielle N Medgyesi, Alexander P Keil, Laura E Beane Freeman, Jared A Fisher, Jessica M Madrigal, Samantha Ammons, Emma S Spielfogel, Komal Bangia, Paul S Albert, Tiffany R Sanchez, James V Lacey, Rena R Jones, Mary H Ward
Novel associations between drinking water contaminants and ovarian cancer risk at levels below regulatory limits warrant further investigation.
BACKGROUND: Several drinking water contaminants are known or suspected carcinogens; however, there are only a few investigations of drinking water exposure and ovarian cancer. We evaluated associations between regulated contaminants in community water systems (CWS) and ovarian cancer risk in the California Teachers Study, a prospective cohort of female California educators.
METHODS: Participants were cancer-free, without bilateral oophorectomy, living in California at baseline (1995-1996) with geocoded addresses linked to a CWS (N = 91,127, 92%), with follow-up through 2020 (mean = 19.0 years). Among participants with a residential duration at enrollment of at least 10 years, we computed 15-year (1990-2005) averages of log2-transformed arsenic, nitrate, total trihalomethanes (TTHM) (N = 59,881), and uranium concentrations (N = 56,314). We estimated hazard ratios (HRs, 95% CIs) for all epithelial ovarian cancers (n = 413) and the high-grade serous histotype (n = 199), using Cox proportional hazards regression, adjusting for age, body mass index, menopause status, oral contraceptive use, and parity. We evaluated the mixture effect (per IQR in log2 concentrations) using quantile-based g-computation.
RESULTS: Almost all women (>99%) had average exposures below regulatory limits for all contaminants. In single contaminant analyses, a doubling in average uranium concentrations was associated with all ovarian cancer (HRperlog2 = 1.07, CI 1.00-1.15), whereas a doubling in nitrate was associated with the high-grade serous histotype (HRperlog2 = 1.09, 95% CI 1.01, 1.17). Findings were similar in models adjusted for other contaminants. We observed positive but imprecise associations for arsenic and TTHM in single-contaminant and contaminant-adjusted analyses. HRs per increase in the mixture were 1.34 (0.95, 1.88) and 1.77 (1.09, 2.87) for all ovarian cancer and the high-grade serous histotype, respectively. Uranium was the largest contributor (56%) to the mixture effect for all ovarian cancers, and nitrate was the largest contributor (48%) for the high-grade serous histotype.
CONCLUSIONS: Novel associations between drinking water contaminants and ovarian cancer risk at levels below regulatory limits warrant further investigation.