Long-Term Exposure to Arsenic in Community Water Supplies and Risk of Cardiovascular Disease among Women in the California Teachers Study

Abstract

Background: Inorganic arsenic in drinking water (wAs) is linked to atherosclerosis and cardiovascular disease. However, risk is uncertain at lower levels present in US community water supplies (CWS), currently regulated at the federal maximum contaminant level of $10\mu g/L$.

Objectives: We evaluated the relationship between long-term wAs exposure from CWS and cardiovascular disease in the California Teachers Study cohort.

Methods: Using statewide health care administrative records from enrollment through follow-up (1995-2018), we identified fatal and nonfatal cases of ischemic heart disease (IHD) and cardiovascular disease (CVD). Participants' residential addresses were linked to a network of CWS boundaries and annual wAs concentrations (1990-2020). Most participants resided in areas served by a CWS (92%). Exposure was calculated as a time-varying, 10-year moving average up to a participant's event, death, or end of follow-up. Using Cox models, we estimated hazard ratios (HRs) and 95% confidence intervals (95% CIs) for the risk of IHD or CVD. We evaluated wAs exposure categorized by concentration thresholds relevant to regulation standards ($<1.00$, 1.00-2.99, 3.00-4.99, 5.00-9.99, $\ge 10\mu g/L$) and continuously using a log2-transformation (i.e., per doubling). Models were adjusted for baseline age, neighborhood socioeconomic status, race/ethnicity, body mass index (BMI), and smoking status. We also stratified analyses by age, BMI, and smoking status.

Results: Our analysis included 98,250 participants, 6,119 IHD cases, and 9,936 CVD cases. The HRs for IHD at concentration thresholds (reference, $<1\mu g/L$) were 1.06 (95% CI: 1.00, 1.12), 1.05 (95% CI: 0.94, 1.17), 1.20 (95% CI: 1.02, 1.41), and 1.42 (95% CI: 1.10, 1.84) for $1.00-2.99\mu g/L$, $3.00-4.99\mu g/L$, $5.00-9.99\mu g/L$, and $\ge 10\mu g/L$, respectively. HRs for every doubling of wAs exposure were 1.04 (95% CI: 1.02, 1.06) for IHD and 1.02 (95% CI: 1.01, 1.04) for CVD. We observed statistically stronger risk among those $\le 55$ vs. $>55$ years of age at enrollment ($p_{\text{interaction}}=0.006$ and 0.012 for IHD and CVD, respectively).

Discussion: Long-term wAs exposure from CWS, at and below the regulatory limit, may increase cardiovascular disease risk, particularly IHD. https://doi.org/10.1289/EHP14410.

Figure 1.

Map of California community water supply (CWS) boundaries outlined in black ($n=\mathrm{2,933}$). A zoomed-in map of Los Angeles County (left) shows areas served by CWS (colored gray) in higher resolution. CWS boundaries were obtained from the publicly accessible “Water Boundary Tool” project developed by the California Environmental Health Tracking Program and further revised by the California Office of Environmental Health Hazard Assessment. Underlying map shows the county-level population density of California Teachers Study participants, based on their residential address at enrollment ($n=\mathrm{118,692}$; see flowchart, Figure S1, for inclusion criteria). County boundaries are from the US Census Bureau 2016 TIGER shapefile for California. Map was created by the authors of this paper using R version 4.3.3 in the CTS Researcher Platform.

Figure 2.

Penalized splines of hazard ratios (HRs) and 95% confidence intervals (95% CIs), shaded in gray, for ischemic heart disease (top) and cardiovascular disease (bottom) by time-varying 3-, 5-, or 10-year moving average water arsenic exposure ($\mathrm{\mu }\mathrm{g}/\mathrm{L}$) up to the time of event or end of follow-up. Analysis conducted in the California Teachers Study cohort from 1995 to 2018 ($n=\mathrm{98,250}$). Estimated with Cox models adjusted for continuous age, continuous BMI, smoking status, race/ethnicity and quartiles of neighborhood socioeconomic status. HRs (95% CIs) are provided at 5 and $10\mathrm{\mu }\mathrm{g}/\mathrm{L}$, marked by vertical dashed lines, and are in reference to $1\mathrm{\mu }\mathrm{g}/\mathrm{L}$, indicated by the vertical gray line. Note: BMI, body mass index.

Figure 3.

Stratified analyses by age, body mass index (BMI), and smoking status at enrollment in the California Teachers Study from 1995 to 2018 ($n=\mathrm{98,250}$). Hazard ratios [95% confidence intervals (95% CI)] for ischemic heart disease (left) and cardiovascular disease (right) by time varying, 10-year moving average water arsenic exposure ($\mathrm{\mu }\mathrm{g}/\mathrm{L}$), categorized into concentration thresholds in reference to $<1\mathrm{\mu }\mathrm{g}/\mathrm{L}$. Cox models adjusted for continuous age, continuous BMI, smoking status (except for stratified model by smoking status), race/ethnicity, and quartiles of neighborhood socioeconomic status. Bottom-right corner shows the $p$-value for interaction (p-int.). See Tables S9–S11 for corresponding numeric data.