Diabetes mellitus (DM) is the most widely recognized metabolic illness with expanding morbidity among ongoing years. Its high incapacity rate and death rate badly affect individuals' quality of life. Increasing proofs backed the relationship between metal exposures with the risk of DM, but the methodological boundedness cannot clarify the complexity of the internal relationship of metal mixtures. We fitted the logistic regression model, weighted quantile sum regression model, and Bayesian kernel machine regression model to assess the relationship between the metal exposures with DM in adults who participated in the National Health and Nutrition Examination Survey 2013-2016. The metals (lead, cadmium, and copper) levels were significantly higher among diabetic compared to the healthy controls. In the logistic regression model established for each single metal, lead and manganese were associated with DM in both unadjusted and mutually adjusted models (highest vs. lowest concentration quartile). When considering all metal as a mixed exposure, we found a generally positive correlation between metal mixtures with DM (binary outcome) and glycohemoglobin (HbA1c) levels (continuous outcome). Exposure to metal mixtures was associated with an increased risk of DM and elevated levels of HbA1c.
Keywords: BKMR model; Blood metals; Diabetes mellitus; Logistic regression model; Serum metals; WQS model.
© 2022. The Author(s), under exclusive licence to Springer Nature B.V.