Association of Water Arsenic With Incident Diabetes in U.S. Adults: The Multi-Ethnic Study of Atherosclerosis and the Strong Heart Study

OBJECTIVE

We examined the association of arsenic in federally regulated community water systems (CWSs) and unregulated private wells with type 2 diabetes (T2D) incidence in the Strong Heart Family Study (SHFS), a prospective study of American Indian communities, and the Multi-Ethnic Study of Atherosclerosis (MESA), a prospective study of racially and ethnically diverse urban U.S. communities.

RESEARCH DESIGN AND METHODS

We evaluated 1,791 participants from SHFS and 5,777 participants from MESA who had water arsenic estimates available and were free of T2D at baseline (2001-2003 and 2000-2002, respectively). Participants were followed for incident T2D until 2010 (SHFS cohort) or 2019 (MESA cohort). We used Cox proportional hazards mixed-effects models to account for clustering by family and residential zip code, with adjustment for sex, baseline age, BMI, smoking status, and education.

RESULTS

T2D incidence was 24.4 cases per 1,000 person-years (mean follow-up, 5.6 years) in SHFS and 11.2 per 1,000 person-years (mean follow-up, 14.0 years) in MESA. In a meta-analysis across the SHFS and MESA cohorts, the hazard ratio (95% CI) per doubling in CWS arsenic was 1.10 (1.02, 1.18). The corresponding hazard ratio was 1.09 (0.95, 1.26) in the SHFS group and 1.10 (1.01, 1.20) in the MESA group. The corresponding hazard ratio (95% CI) for arsenic in private wells and incident T2D in SHFS was 1.05 (0.95, 1.16). We observed statistical interaction and larger magnitude hazard ratios for participants with BMI <25 kg/m2 and female participants.

CONCLUSIONS

Low to moderate water arsenic levels (<10 µg/L) were associated with T2D incidence in the SHFS and MESA cohorts.

Estimating Lithium Concentrations in Groundwater Used as Drinking Water for the Conterminous United States

Lithium (Li) concentrations in drinking-water supplies are not regulated in the United States; however, Li is included in the 2022 U.S. Environmental Protection Agency list of unregulated contaminants for monitoring by public water systems. Li is used pharmaceutically to treat bipolar disorder, and studies have linked its occurrence in drinking water to human-health outcomes. An extreme gradient boosting model was developed to estimate geogenic Li in drinking-water supply wells throughout the conterminous United States. The model was trained using Li measurements from ∼13,500 wells and predictor variables related to its natural occurrence in groundwater. The model predicts the probability of Li in four concentration classifications, ≤4 μg/L, >4 to ≤10 μg/L, >10 to ≤30 μg/L, and >30 μg/L. Model predictions were evaluated using wells held out from model training and with new data and have an accuracy of 47-65%. Important predictor variables include average annual precipitation, well depth, and soil geochemistry. Model predictions were mapped at a spatial resolution of 1 km2 and represent well depths associated with public- and private-supply wells. This model was developed by hydrologists and public-health researchers to estimate Li exposure from drinking water and compare to national-scale human-health data for a better understanding of dose-response to low (<30 μg/L) concentrations of Li.

Contribution of arsenic and uranium in private wells and community water systems to urinary biomarkers in US adults: The Strong Heart Study and the Multi-Ethnic Study of Atherosclerosis

BACKGROUND

Chronic exposure to inorganic arsenic (As) and uranium (U) in the United States (US) occurs from unregulated private wells and federally regulated community water systems (CWSs). The contribution of water to total exposure is assumed to be low when water As and U concentrations are low.

OBJECTIVE

We examined the contribution of water As and U to urinary biomarkers in the Strong Heart Family Study (SHFS), a prospective study of American Indian communities, and the Multi-Ethnic Study of Atherosclerosis (MESA), a prospective study of racially/ethnically diverse urban U.S. communities.

METHODS

We assigned residential zip code-level estimates in CWSs (µg/L) and private wells (90th percentile probability of As >10 µg/L) to up to 1485 and 6722 participants with dietary information and urinary biomarkers in the SHFS (2001-2003) and MESA (2000-2002; 2010-2011), respectively. Urine As was estimated as the sum of inorganic and methylated species, and urine U was total uranium. We used linear mixed-effects models to account for participant clustering and removed the effect of dietary sources via regression adjustment.

RESULTS

The median (interquartile range) urine As was 5.32 (3.29, 8.53) and 6.32 (3.34, 12.48) µg/L for SHFS and MESA, respectively, and urine U was 0.037 (0.014, 0.071) and 0.007 (0.003, 0.018) µg/L. In a meta-analysis across both studies, urine As was 11% (95% CI: 3, 20%) higher and urine U was 35% (5, 73%) higher per twofold higher CWS As and U, respectively. In the SHFS, zip-code level factors such as private well and CWS As contributed 46% of variation in urine As, while in MESA, zip-code level factors, e.g., CWS As and U, contribute 30 and 49% of variation in urine As and U, respectively.

IMPACT STATEMENT

We found that water from unregulated private wells and regulated CWSs is a major contributor to urinary As and U (an estimated measure of internal dose) in both rural, American Indian populations and urban, racially/ethnically diverse populations nationwide, even at levels below the current regulatory standard. Our findings indicate that additional drinking water interventions, regulations, and policies can have a major impact on reducing total exposures to As and U, which are linked to adverse health effects even at low levels.

Cross-sectional associations between drinking water arsenic and urinary inorganic arsenic in the United States: NHANES 2003-2014

BACKGROUND

Inorganic arsenic is a potent carcinogen and toxicant associated with numerous adverse health outcomes. The contribution of drinking water from private wells and regulated community water systems (CWSs) to total inorganic arsenic exposure is not clear.

OBJECTIVES

To determine the association between drinking water arsenic estimates and urinary arsenic concentrations in the 2003-2014 National Health and Nutrition Examination Survey (NHANES).

METHODS

We evaluated 11,088 participants from the 2003-2014 NHANES cycles. For each participant, we assigned private well and CWS arsenic levels according to county of residence using estimates previously derived by the U.S. Environmental Protection Agency and U.S. Geological Survey. We used recalibrated urinary dimethylarsinate (rDMA) to reflect the internal dose of estimated water arsenic by applying a previously validated, residual-based method that removes the contribution of dietary arsenic sources. We compared the adjusted geometric mean ratios and corresponding percent change of urinary rDMA across tertiles of private well and CWS arsenic levels, with the lowest tertile as the reference. Comparisons were made overall and stratified by census region and race/ethnicity.

RESULTS

Overall, the geometric mean of urinary rDMA was 2.52 (2.30, 2.77) μg/L among private well users and 2.64 (2.57, 2.72) μg/L among CWS users. Urinary rDMA was highest among participants in the West and South, and among Mexican American, Other Hispanic, and Non-Hispanic Other participants. Urinary rDMA levels were 25% (95% confidence interval (CI): 17-34%) and 20% (95% CI: 12-29%) higher comparing the highest to the lowest tertile of CWS and private well arsenic, respectively. The strongest associations between water arsenic and urinary rDMA were observed among participants in the South, West, and among Mexican American and Non-Hispanic White and Black participants.

DISCUSSION

Both private wells and regulated CWSs are associated with inorganic arsenic internal dose as reflected in urine in the general U.S.

POPULATION

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

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.

Perfluoroalkyl and Polyfluoroalkyl Substances in Groundwater Used as a Source of Drinking Water in the Eastern United States

In 2019, 254 samples were collected from five aquifer systems to evaluate perfluoroalkyl and polyfluoroalkyl substance (PFAS) occurrence in groundwater used as a source of drinking water in the eastern United States. The samples were analyzed for 24 PFAS, major ions, nutrients, trace elements, dissolved organic carbon (DOC), volatile organic compounds (VOCs), pharmaceuticals, and tritium. Fourteen of the 24 PFAS were detected in groundwater, with 60 and 20% of public-supply and domestic wells, respectively, containing at least one PFAS detection. Concentrations of tritium, chloride, sulfate, DOC, and manganese + iron; percent urban land use within 500 m of the wells; and VOC and pharmaceutical detection frequencies were significantly higher in samples containing PFAS detections than in samples with no detections. Boosted regression tree models that consider 57 chemical and land-use variables show that tritium concentration, distance to the nearest fire-training area, percentage of urban land use, and DOC and VOC concentrations are the top five predictors of PFAS detections, consistent with the hydrologic position, geochemistry, and land use being important controls on PFAS occurrence in groundwater. Model results indicate that it may be possible to predict PFAS detections in groundwater using existing data sources.

Associations between private well water and community water supply arsenic concentrations in the conterminous United States

Geogenic arsenic contamination typically occurs in groundwater as opposed to surface water supplies. Groundwater is a major source for many community water systems (CWSs) in the United States (US). Although the US Environmental Protection Agency sets the maximum contaminant level (MCL enforceable since 2006: 10 μg/L) for arsenic in CWSs, private wells are not federally regulated. We evaluated county-level associations between modeled values of the probability of private well arsenic exceeding 10 μg/L and CWS arsenic concentrations for 2231 counties in the conterminous US, using time invariant private well arsenic estimates and CWS arsenic estimates for two time periods. Nationwide, county-level CWS arsenic concentrations increased by 8.4 μg/L per 100% increase in the probability of private well arsenic exceeding 10 μg/L for 2006-2008 (the initial compliance monitoring period after MCL implementation), and by 7.3 μg/L for 2009-2011 (the second monitoring period following MCL implementation) (1.1 μg/L mean decline over time). Regional differences in this temporal decline suggest that interventions to implement the MCL were more pronounced in regions served primarily by groundwater. The strong association between private well and CWS arsenic in Rural, American Indian, and Semi Urban, Hispanic counties suggests that future research and regulatory support are needed to reduce water arsenic exposures in these vulnerable subpopulations. This comparison of arsenic exposure values from major private and public drinking water sources nationwide is critical to future assessments of drinking water arsenic exposure and health outcomes.

Machine Learning Models of Arsenic in Private Wells Throughout the Conterminous United States As a Tool for Exposure Assessment in Human Health Studies

Arsenic from geologic sources is widespread in groundwater within the United States (U.S.). In several areas, groundwater arsenic concentrations exceed the U.S. Environmental Protection Agency maximum contaminant level of 10 μg per liter (μg/L). However, this standard applies only to public-supply drinking water and not to private-supply, which is not federally regulated and is rarely monitored. As a result, arsenic exposure from private wells is a potentially substantial, but largely hidden, public health concern. Machine learning models using boosted regression trees (BRT) and random forest classification (RFC) techniques were developed to estimate probabilities and concentration ranges of arsenic in private wells throughout the conterminous U.S. Three BRT models were fit separately to estimate the probability of private well arsenic concentrations exceeding 1, 5, or 10 μg/L whereas the RFC model estimates the most probable category (≤5, >5 to ≤10, or >10 μg/L). Overall, the models perform best at identifying areas with low concentrations of arsenic in private wells. The BRT 10 μg/L model estimates for testing data have an overall accuracy of 91.2%, sensitivity of 33.9%, and specificity of 98.2%. Influential variables identified across all models included average annual precipitation and soil geochemistry. Models were developed in collaboration with public health experts to support U.S.-based studies focused on health effects from arsenic exposure.

Assessing the Impact of Drought on Arsenic Exposure from Private Domestic Wells in the Conterminous United States

This study assesses the potential impact of drought on arsenic exposure from private domestic wells by using a previously developed statistical model that predicts the probability of elevated arsenic concentrations (>10 μg per liter) in water from domestic wells located in the conterminous United States (CONUS). The application of the model to simulate drought conditions used systematically reduced precipitation and recharge values. The drought conditions resulted in higher probabilities of elevated arsenic throughout most of the CONUS. While the increase in the probability of elevated arsenic was generally less than 10% at any one location, when considered over the entire CONUS, the increase has considerable public health implications. The population exposed to elevated arsenic from domestic wells was estimated to increase from approximately 2.7 million to 4.1 million people during drought. The model was also run using total annual precipitation and groundwater recharge values from the year 2012 when drought existed over a large extent of the CONUS. This simulation provided a method for comparing the duration of drought to changes in the predicted probability of high arsenic in domestic wells. These results suggest that the probability of exposure to arsenic concentrations greater than 10 μg per liter increases with increasing duration of drought. These findings indicate that drought has a potentially adverse impact on the arsenic hazard from domestic wells throughout the CONUS.

A multi-model approach toward understanding iron fouling at rock-fill drainage sites along roadways in New Hampshire, USA

Factors affecting iron fouling in wet areas adjacent to roadways were investigated by collecting field rock cut and aqueous physicochemical data; developing exploratory predictive models; and developing geochemical models. Basic data included the identification of iron fouling from aerial imagery and field visits at 374 New Hampshire rock cut locations, and their associated rock-fill sites. Based on field water quality measurements from wet areas at 36 of the rock-fill sites, the occurrence of iron fouling was associated with higher values of specific conductance, lower concentrations of dissolved oxygen and lower pH compared to areas without iron fouling. A statistical model, using boosted regression trees, was developed to predict the occurrence of iron fouling in wet areas adjacent to roadways where rock-fill from nearby rock cuts was used in roadway construction. The model was used to develop a continuous iron fouling probability map for the state of New Hampshire that can be used to better understand the occurrence of iron fouling. Geochemical models illustrate how iron fouling of waters increases along roadways built with fill from sulfidic rock cuts as a result of acid generation from pyrite dissolution and ferrous iron (Fe2+) oxidation and increases in areas with greater specific conductance from deicing runoff caused by cation exchange. More iron is precipitated as goethite in simulations that include pyrite, and in simulations with deicing salts added, indicating that rock-fill sites with rocks that contain pyrite and water with greater salt content could have enhanced iron fouling.

Estimating domestic well locations and populations served in the contiguous U.S. for years 2000 and 2010

Domestic wells provide drinking water supply for approximately 40 million people in the United States. Knowing the location of these wells, and the populations they serve, is important for identifying heavily used aquifers, locations susceptible to contamination, and populations potentially impacted by poor-quality groundwater. The 1990 census was the last nationally consistent survey of a home's source of water, and has not been surveyed since. This paper presents a method for projecting the population dependent on domestic wells for years after 1990, using information from the 1990 census along with population data from subsequent censuses. The method is based on the "domestic ratio" at the census block-group level, defined here as the number of households dependent on domestic wells divided by the total population. Analysis of 1990 data (>220,000 block-groups) indicates that the domestic ratio is a function of the household density. As household density increases, the domestic ratio decreases, once a household density threshold is met. The 1990 data were used to develop a relationship between household density and the domestic ratio. The fitted model, along with household density data from 2000 and 2010, was used to estimate domestic ratios for each decadal year. In turn, the number of households dependent on domestic wells was estimated at the block-group level for 2000 and 2010. High-resolution census-block population data were used to refine the spatial distribution of domestic-well usage and to convert the data into population numbers. The results are presented in two downloadable raster datasets for each decadal year. It is estimated that the total population using domestic-well water in the contiguous U.S. increased 1.5% from 1990 to 2000 to a total of 37.25 million people and increased slightly from 2000 to 2010 to 37.29 million people.

Synthetic matrix metalloproteinase inhibitors and tissue inhibitor of metalloproteinase (TIMP)-2, but not TIMP-1, inhibit shedding of tumor necrosis factor-alpha receptors in a human colon adenocarcinoma (Colo 205) cell line

The solubilization of plasma membrane receptors through proteolytic cleavage of the ligand binding domain at the cell surface is an important mechanism for regulating cytokine function and receptor signaling. The inhibition of the shedding of a variety of receptors by synthetic inhibitors of the matrix metalloproteinases (MMPs) implicates metalloproteinases in this regulatory event. We examined the effects of two naturally occurring tissue inhibitors of metalloproteinases, TIMP-1 and TIMP-2, and several synthetic MMP inhibitors (MMPIs) on the shedding of both tumor necrosis factor alpha receptor type I (TNFalpha-RI; Mr 55,000) and TNFalpha-RII (Mr 75,000) by the Colo 205 human colon adenocarcinoma cell line. Culture of Colo 205 cells for 48 h resulted in the shedding of both TNFalpha-RI and TNFalpha-RII, as determined by ELISA. The shedding of TNFalpha receptors was not affected by TIMP-1 or protease inhibitors aprotinin, pepstatin, or leupeptin but was inhibited in a dose-dependent manner by the following synthetic MMPIs: batimastat and marimastat (BB-94 and BB-2516, respectively, British Biotech, Inc.); CT1418 (Celltech Therapeutics); CGS27023A (Novartis Pharmaceuticals); and RO31-9790 (Roche), with IC50s ranging from 3.2 to 38.0 microM. Similarly, TIMP-2 from two different sources reproducibly inhibited the shedding of both TNFalpha-RI and TNFalpha-RII in a dose-dependent manner (IC50 = 286 +/- 33 nM for TNFalpha-RI shedding and 462 +/- 52 nM for shedding of TNFalpha-RII). The inhibition of TNFalpha-RI shedding was confirmed in the SW626 human ovarian adenocarcinoma cell line. The synthetic MMPIs and TIMP-2, but not TIMP-1, also caused a dose-dependent increase in the number of TNFalpha receptors retained on the surface of Colo 205 cells, as determined by flow cytometry. Inhibition of TNFalpha receptor shedding with TIMP-2 occurs at molar concentrations 10-100 times less than those required with low molecular weight, synthetic MMPIs but at concentrations greater than those required to inhibit collagen degradation. Modulation of TNFalpha receptor shedding by TIMP-2 could have important implications for the pleiotropic effects of TNFalpha in both normal and malignant cells and for the pharmacological activity of synthetic MMPIs.

Potent and sustained satiety actions of a cholecystokinin octapeptide analogue

The relative ability of a norleucine substituted cholecystokinin (CCK) analogue, U-67827E, to interact with CCK receptors and to inhibit food intake was examined across a variety of paradigms. U-67827E and CCK had identical in vitro potencies as demonstrated by their ability to induce pyloric contractions or competitively inhibit [125I]CCK-8 binding to type A and B CCK receptors. However, the in vivo potency of U-67827E was significantly greater than that of CCK-8. In rats, U-67827E inhibited food intake with 10-100 times the potency of CCK. In rhesus monkeys, U-67827E produced significantly greater inhibitions of daily food intake and did so in a dose-dependent manner with no evidence of compensation or tolerance. U-67827E also inhibited gastric emptying for significantly longer durations than CCK. Together these results demonstrate that CCK analogues with increased in vivo bioavailability can affect food intake beyond a single meal.