Multipollutant, longitudinal analysis of the association between urinary tungsten and incident diabetes in a rural population

BACKGROUND: Cross-sectional studies suggest tungsten (W) exposure may be associated with diabetes. We assessed longitudinal associations between urinary W and fasting glucose, 2-hour glucose, insulin resistance (HOMA-IR), β-cell function (HOMA-β), and incident type 2 diabetes. METHODS: We used data from 1,609 Hispanic and non-Hispanic White adults with 20 to 74 years of age residing in rural Colorado and participating in the San Luis Valley Diabetes Study. Urinary metal exposure values were measured at baseline and natural log-transformed. We assessed longitudinal associations between urinary W and continuous outcome measures using linear-mixed effect models and associations with incident diabetes using Fine and Gray competing risks regression models (competing event = all-cause mortality). The main adjustment set of covariates included: age, sex, ethnicity, education, smoking status, hypertension, body mass index, caloric intake, alcohol intake, and urinary creatinine levels. Secondary models were further adjusted for arsenic, cadmium, and lead exposures. We assessed whether sex or ethnicity were effect modifiers. RESULTS: At baseline, the median W concentration was 0.22 μg/L (interquartile range = 0.20, 0.59). In the main cross-sectional analyses, lnW levels were significantly associated with 3% higher lnHOMA-IR (95% CI = 1 to 5). In the main longitudinal models, lnW was significantly associated with 1% higher natural log-transformed fasting glucose (95% CI = <1 to 1), 3% higher natural log-transformed HOMA-IR (95% CI = 2 to 5), and 28% higher incident diabetes (subdistribution hazard ratio=1.28, 95% CI = 1.09 to 1.50). Results remained significant when further adjusting for other metals. We observed evidence for effect modification by sex and ethnicity. CONCLUSION: Urinary W was longitudinally associated with adverse metabolic health indicators.