Arsenic in private well water and birth outcomes in the United States

Abstract

Background: Prenatal exposure to drinking water with arsenic concentrations >50 μg/L is associated with adverse birth outcomes, with inconclusive evidence for concentrations ≤50 μg/L. In a collaborative effort by public health experts, hydrologists, and geologists, we used published machine learning model estimates to characterize arsenic concentrations in private wells-federally unregulated for drinking water contaminants-and evaluated associations with birth outcomes throughout the conterminous U.S.

Methods: Using several machine learning models, including boosted regression trees (BRT) and random forest classification (RFC), developed from measured groundwater arsenic concentrations of ∼20,000 private wells, we characterized the probability that arsenic concentrations occurred within specific ranges in groundwater. Probabilistic model estimates and private well usage data were linked by county to all live birth certificates from 2016 (n = 3.6 million). We evaluated associations with gestational age and term birth weight using mixed-effects models, adjusted for potential confounders and incorporated random intercepts for spatial clustering.

Results: We generally observed inverse associations with term birth weight. For instance, when using BRT estimates, a 10-percentage point increase in the probability that private well arsenic concentrations exceeded 5 μg/L was associated with a -1.83 g (95% CI: -3.30, -0.38) lower term birth weight after adjusting for covariates. Similarly, a 10-percentage point increase in the probability that private well arsenic concentrations exceeded 10 μg/L was associated with a -2.79 g (95% CI: -4.99, -0.58) lower term birth weight. Associations with gestational age were null.

Conclusion: In this largest epidemiologic study of arsenic and birth outcomes to date, we did not observe associations of modeled arsenic estimates in private wells with gestational age and found modest inverse associations with term birth weight. Study limitations may have obscured true associations, including measurement error stemming from a lack of individual-level information on primary water sources, water arsenic concentrations, and water consumption patterns.

Keywords: Arsenic; Birth outcomes; Epidemiology; Private wells; Water contamination.

Figure 1.

Machine Learning Model-Predicted Arsenic Probabilities at 1 km² and Aggregated to the County-Level. BRT1 refers to the probability that arsenic concentrations exceeded 1 μg/L as predicted by a boosted regression tree; BRT5 refers to the probability that arsenic concentrations exceeded 5 μg/L as predicted by a boosted regression tree; BRT10 refers to the probability that arsenic concentrations exceeded 10 μg/L as predicted by a boosted regression tree; RFC2 refers to the probability that arsenic concentrations fell between >5 to ≤10 μg/L whereas RFC3 refers to the probability that arsenic concentrations exceeded 10 μg/L, with both probabilities predicted by a single random forest classification model.

Figure 2.

Sex-Stratified Associations Between the Private Well Arsenic Probabilities and Term Birth Weight

Figure 3.

Region-Specific Associations Between the Probability that Arsenic Concentrations Exceeded 5 μg/L and Gestational Age, Stratified by County-Level Rates of Private Well Use. Results have been organized by U.S. Geological Survey Ground Water Atlas Regions as follows: B (California, Nevada), C (Arizona, Colorado, New Mexico, Utah), D (Kansas, Missouri, Nebraska), E (Oklahoma, Texas), F (Arkansas, Louisiana, Mississippi), G (Alabama, Florida, Georgia, South Carolina), H (Idaho, Oregon, Washington), I (Montana, North Dakota, South Dakota, Wyoming), J (Iowa, Michigan, Minnesota, Wisconsin), K (Illinois, Indiana, Kentucky, Ohio, Tennessee), L (Delaware, Maryland, New Jersey, North Carolina, Pennsylvania, Virginia, West Virginia), and M (Connecticut, Maine, Massachusetts, New Hampshire, New York, Rhode Island, Vermont). Ground Water Atlas Regions typically share hydrogeologic and hydrologic conditions across the major aquifers in each regional area (U.S. Geological Survey, 2016). BRT-estimated Pr(As >5 μg/L) refers to the probability that private well arsenic concentrations exceeded 5 μg/L as estimated by a boosted regression tree.

Figure 4.

Region-Specific Associations Between the Probability that Arsenic Concentrations Exceeded 5 μg/L and Term Birth Weight, Stratified by County-Level Rates of Private Well Use. Results have been organized by U.S. Geological Survey Ground Water Atlas Regions as follows: B (California, Nevada), C (Arizona, Colorado, New Mexico, Utah), D (Kansas, Missouri, Nebraska), E (Oklahoma, Texas), F (Arkansas, Louisiana, Mississippi), G (Alabama, Florida, Georgia, South Carolina), H (Idaho, Oregon, Washington), I (Montana, North Dakota, South Dakota, Wyoming), J (Iowa, Michigan, Minnesota, Wisconsin), K (Illinois, Indiana, Kentucky, Ohio, Tennessee), L (Delaware, Maryland, New Jersey, North Carolina, Pennsylvania, Virginia, West Virginia), and M (Connecticut, Maine, Massachusetts, New Hampshire, New York, Rhode Island, Vermont). Ground Water Atlas Regions typically share hydrogeologic and hydrologic conditions across the major aquifers in each regional area (U.S. Geological Survey, 2016). BRT-estimated Pr(As >5 μg/L) refers to the probability that private well arsenic concentrations exceeded 5 μg/L as estimated by a boosted regression tree.