Chronic Exposure to Arsenic and Markers of Cardiometabolic Risk: A Cross-Sectional Study in Chihuahua, Mexico

Abstract

Background: Exposure to arsenic (As) concentrations in drinking water > 150 μg/L has been associated with risk of diabetes and cardiovascular disease, but little is known about the effects of lower exposures.

Objective: This study aimed to examine whether moderate As exposure, or indicators of individual As metabolism at these levels of exposure, are associated with cardiometabolic risk.

Methods: We analyzed cross-sectional associations between arsenic exposure and multiple markers of cardiometabolic risk using drinking-water As measurements and urinary As species data obtained from 1,160 adults in Chihuahua, Mexico, who were recruited in 2008-2013. Fasting blood glucose and lipid levels, the results of an oral glucose tolerance test, and blood pressure were used to characterize cardiometabolic risk. Multivariable logistic, multinomial, and linear regression were used to assess associations between cardiometabolic outcomes and water As or the sum of inorganic and methylated As species in urine.

Results: After multivariable adjustment, concentrations in the second quartile of water As (25.5 to < 47.9 μg/L) and concentrations of total speciated urinary As (< 55.8 μg/L) below the median were significantly associated with elevated triglycerides, high total cholesterol, and diabetes. However, moderate water and urinary As levels were also positively associated with HDL cholesterol. Associations between arsenic exposure and both dysglycemia and triglyceridemia were higher among individuals with higher proportions of dimethylarsenic in urine.

Conclusions: Moderate exposure to As may increase cardiometabolic risk, particularly in individuals with high proportions of urinary dimethylarsenic. In this cohort, As exposure was associated with several markers of increased cardiometabolic risk (diabetes, triglyceridemia, and cholesterolemia), but exposure was also associated with higher rather than lower HDL cholesterol.

Citation: Mendez MA, González-Horta C, Sánchez-Ramírez B, Ballinas-Casarrubias L, Hernández Cerón R, Viniegra Morales D, Baeza Terrazas FA, Ishida MC, Gutiérrez-Torres DS, Saunders RJ, Drobná Z, Fry RC, Buse JB, Loomis D, García-Vargas GG, Del Razo LM, Stýblo M. 2016. Chronic exposure to arsenic and markers of cardiometabolic risk: a cross-sectional study in Chihuahua, Mexico. Environ Health Perspect 124:104-111; http://dx.doi.org/10.1289/ehp.1408742.

Figure 1

Adjusted mean (95% CI) difference in fasting or 2-hr plasma glucose associated with As exposure among normoglycemic subjects. Adjusted mean (95% CI) difference in glucose measurement for increasing quartiles of water As or total speciated urinary As relative to individuals in the lowest quartile (< 25.5 μg/L for water and < 27.1 μg/L for urine). Estimated from linear regression models including age, sex, education, ethnicity, weight status, waist circumference, smoking status, alcohol consumption, recent seafood intake, and water source (well, treatment plant, or other). Urinary As models were additionally adjusted for urinary creatinine and ≥ median %DMAs, MAs, and iAs. Models excluded individuals with 2-hr plasma glucose > 140 mg/dL, fasting plasma glucose > 110 mg/dL, or diagnosed diabetes. Among normoglycemic subjects: n in each quartile of water As 1 = 186, 2 = 170, 3 = 174, 4 = 169; of urinary As: 1 = 195, 2 = 181, 3 = 186, 4 = 170.

Figure 2

Associations between urinary As metabolism indicators and cardiometabolic risk. Odds ratios (95% CIs) for elevated cardiometabolic risk associated with increasing quartiles of urinary iAs metabolism indicators from multinomial or logistic models adjusted for total speciated urinary arsenic, as well as age, sex, education, ethnicity, weight status, waist circumference, smoking status, alcohol consumption, recent seafood intake, and water source (well, treatment plant, or other). n = 1,090 Adults. Cardiometabolic markers: DM, diabetes mellitus; HDL, high-density lipoprotein cholesterol; HTN, hypertension; LDL, low-density lipoprotein cholesterol; TC, total cholesterol; TG, triglycerides. Urinary As indicators: DMAs, dimethylarsenic; iAs, inorganic As; MAs, methylarsenic. Cardiometabolic outcomes defined as follows: DM, fasting plasma glucose ≥ 126 mg/dL, 2-hr plasma glucose ≥ 200 mg/dL, or self-reported diabetes diagnosis or medication use; elevated TC ≥ 200 mg/dL; elevated TG ≥ 200 mg/dL; elevated LDL ≥ 130 mg/dL; low HDL = < 40 mg/dL; hypertension SBP > 140 mmHg, DBP > 90 mmHg or anti-hypertensive medication use. Quartile markers (1st = referent): 2nd, black diamond; 3rd, white square; 4th, blue circle. Quartile cutoffs for urinary As metabolism indicators defined as follows: (A) %DMAs = < 70.65, 70.65 to < 76.78, 76.78 to < 81.52, ≥ 81.52; (B) %MAs = < 10.90, 10.90 to < 14.0, 14.0 to < 17.66, ≥ 17.66; (C) DMAs/MAs = < 4.05, 4.05 to < 5.47, 5.49 to < 7.38, ≥ 7.38; (D) MAs/iAs = < 1.185, 1.185 to < 1.576, 1.576 to < 2.11, ≥ 2.11; (E) %iAs = < 6.389, 6.389 to < 8.873, 8.873 to < 12.270, ≥ 12.270.

Figure 3

Water As and odds of prevalent diabetes in subjects with vs. without elevated % urinary DMAs. ORs (95% CIs) for prevalent diabetes associated with household water As categories (< 25, 25 to < 50, 50 to < 100, and ≥ 100 μg/L) in subjects with proportions of DMAs defined as low versus high based on the median of 76.6%. The referent group for all ORs is subjects with %DMAs below the median in the lowest quartile of water As. The results are derived from multinomial models adjusted for age, sex, smoking status, alcohol consumption, BMI, elevated waist circumference, and main water source (well, treatment plant, or other). *p < 0.10 for additive interaction (relative excess risk for interaction) for the joint effect of water As and high %DMAs.