Association between blood cadmium and prevalent coronary heart disease in NHANES 2013 to 2014: A cross-sectional study with machine-learning analyses.

Cadmium is a toxic environmental metal, but its association with prevalent coronary heart disease (CHD) remains uncertain. We examined whether blood cadmium concentrations were associated with prevalent CHD in US adults and explored machine-learning model performance. This cross-sectional study used data from the National Health and Nutrition Examination Survey 2013 to 2014. Adults aged ≥20 years with complete data on self-reported CHD, blood cadmium, and covariates were included. Blood cadmium was measured by inductively coupled plasma mass spectrometry. Survey-weighted logistic regression was used to assess associations between blood cadmium and prevalent CHD. Sensitivity analyses additionally adjusted for serum cotinine and were repeated in never smokers. Ten machine-learning algorithms were developed to assess discrimination, calibration, and feature importance. Among 2219 participants, higher blood cadmium concentrations were associated with higher odds of prevalent CHD. In the fully adjusted model, each 1-unit increase in ln-transformed blood cadmium was associated with prevalent CHD (odds ratio, 1.38; 95% confidence interval, 1.05-1.81). Compared with the lowest quartile, the highest quartile was associated with higher odds of prevalent CHD (odds ratio, 1.95; 95% confidence interval, 1.25-3.04). The association remained after additional adjustment for serum cotinine but was attenuated in never smokers. The multilayer perceptron showed the best overall discrimination and calibration; age ranked first and blood cadmium ranked second by mean absolute Shapley Additive Explanations value. Higher blood cadmium concentrations were associated with higher odds of prevalent CHD in this cross-sectional analysis. These findings should be interpreted cautiously because residual smoking-related confounding remains possible and the CHD outcome was self-reported.