Melih Gaffar GÖZÜKARA, Servet Birgin İRİTAŞ
Ankara Medical Journal - 2026;26(3):381-405
Objectives: Environmental trace element exposure has been proposed as a modifiable risk factor for type 2 diabetes mellitus (DM), but findings are inconsistent. We examined associations between urinary and blood trace elements and DM, and their sensitivity to model specification. Materials and Methods: We analysed 1,486 adults aged 20 years or older from the 2013-2014 NHANES cycle. DM was defined by American Diabetes Association criteria. Seventeen urinary and five blood elements were measured by mass spectrometry. Groups were compared using Mann-Whitney U and chi-square tests with Benjamini-Hochberg correction. Log-transformed concentrations were entered into logistic regression models adjusted for age, sex, body mass index, race/ethnicity, socioeconomic position, smoking, hypertension, and kidney disease, fitted with and without complex survey weights. Odds ratios (OR) are reported per interquartile-range increase. Results: DM prevalence was 13.7%. In adjusted models (n=1,388), urinary arsenous acid was inversely associated with DM (OR 0.55, 95% CI 0.36-0.86, p=0.008); this persisted under the survey design (OR 0.47, 0.30-0.72, p<0.001) and after creatinine adjustment. Urinary cadmium, strongly associated when age was omitted (OR 1.56, 1.17-2.07, p=0.002), showed no association after adjustment for age (OR 0.89, 0.65-1.23, p=0.481). Blood selenium was associated in the unweighted analysis (OR 1.26, p=0.020) but not in the weighted analysis; urinary tin and molybdenum were only associated under the survey design. Conclusion: The cadmium-DM association was fully attributable to age confounding, and arsenous acid showed an inverse rather than positive association. Such associations are highly sensitive to model specification; adequate age adjustment and use of the complex survey design are essential in NHANES-based analyses.
