The associations between urinary metals and metal mixtures and kidney function in Chinese community-dwelling older adults with diabetes mellitus

Abstract

Background: Previous studies have found associations between single toxic metals, such as arsenic and cadmium, and kidney function in adults with diabetes. However, studies with regards to other metals and metal mixtures are still limited.

Objective: Our study aimed to investigate the associations between urinary concentrations of 5 selected metals and metal mixtures and kidney function using a sample of older adults with diabetes mellitus in Chinese communities.

Methods: In a sample of older adults (n = 5186), 592 eligible subjects were included in this study. Urinary concentrations of 5 metals, i.e., arsenic (As), cadmium (Cd), vanadium (V), cobalt (Co), and thallium (Tl), were measured by inductively coupled plasma mass spectrometer (ICP-MS). Estimated glomerular filtration rate (eGFR) was calculated and dichotomized into indicator of chronic kidney disease (CKD). Logistic analysis and Bayesian kernel machine regression (BKMR) were used to explore the associations between single metals and metal mixtures and CKD, respectively.

Results: Urinary levels of As and V were positively correlated with CKD (OR=2.37, 95% CI: 1.31-4.30 for As; OR=2.24, 95% CI: 1.25-4.03 for V), when compared the 4th quartile with the 1st quartile. After adjustment for potential confounders, the significant association between As and CKD still existed (OR=2.73, 95% CI: 1.23-6.07). BKMR analyses showed strong linear positive associations between As and V and CKD. Higher urinary levels of the mixture were significantly associated with higher odds of CKD in a dose-response pattern. As and V showed the highest posterior inclusion probabilities.

Conclusion: Urine As and V were positively associated with CKD in older adults with diabetes mellitus, separately and in a mixture. The metals mixture showed a linear dose-response association with the odds of CKD. The analyses of mixtures, rather than of single metals, may provide a real-world perspective on the relationship between metals and kidney function.

Keywords: Bayesian kernel machine regression (BKMR); Chronic kidney disease; Diabetes mellitus; Metal mixtures; Older adults.