To examine the exposure-response relationship between mercury exposure and diabetes in adults, and to explore the possible effect modifications by selenium and omega-3 fatty acids.
Biomarker data (total blood mercury and blood methylmercury) from individuals ≥20 years of age were obtained from the 2005-2018 NHANES. Diabetes was defined through questionnaires, fasting plasma glucose, 2 -h plasma glucose and hemoglobin A levels. The exposure-response relationship between mercury exposure and diabetes was assessed with logistic regression and restricted cubic splines.
Comparing the highest to lowest quartile of exposure, the multivariable-adjusted odds ratio (95 % CI) of diabetes was 0.76 (0.63-0.92) with total blood mercury and 0.82 (0.66-1.00) with blood methylmercury. The inverse associations between total blood mercury [0.55 (0.40-0.77)] and blood methylmercury [0.61 (0.38-0.97)] and diabetes were observed among individuals having higher intakes of selenium (P<0.05). Trends toward lower odds of diabetes with mercury exposure were mainly confined to individuals having higher intakes of omega-3 fatty acid, but the interactions were not significant. The inverse associations between total blood mercury and blood methylmercury and diabetes remained in sensitivity analyses after excluding patients with hypertension that may change their dietary intake of fish. Exposure-response analyses showed an initial decrease in odds of diabetes followed by a platform or a weaker decrease beyond 3 μg/L of total blood mercury and methylmercury concentrations, respectively.
Total blood mercury and blood methylmercury concentrations were inversely associated with diabetes in adults, and the associations were modified by selenium.

Copyright © 2021 Elsevier GmbH. All rights reserved.