Overview
Cross-sectional analysis of 3,514 US adults from NHANES 2011-2016 examining seven metals (lead, cadmium, mercury, selenium, manganese in blood; copper and zinc in serum) in relation to chronic kidney disease risk, using logistic regression, restricted cubic splines, WQS, BKMR, and XGBoost machine learning.
Key numbers
n=3,514 US adults, NHANES 2011-2016.
Mixture findings:
- WQS: metal mixture exposure positively associated with CKD (OR=1.58, 95% CI: 1.30-1.94); lead and cadmium showed the highest contribution.
- BKMR confirmed cumulative and interactive effects of the metal mixture on CKD risk.
- XGBoost model: AUC=0.801, indicating good predictive performance for CKD using metal profiles.
- Single-metal: higher lead, cadmium, and copper concentrations linked to increased CKD risk; zinc showed protective effects.
- Selenium and manganese showed negative (protective) associations with CKD in some models, partially offsetting lead and cadmium effects.
The OR of 1.58 for the overall metal mixture is a mixture-level estimate after accounting for the opposing directions of different metals; this estimate would be attenuated relative to what lead and cadmium alone would yield if protective metals were excluded.
Evidence Fitness
Context only. Cross-sectional design. Blood metals reflect mixed recent and cumulative exposure by metal half-life. The mixture OR of 1.58 characterizes the full seven-metal mixture, including protective metals. Lead and cadmium, the two metals with positive single-metal associations, are the dominant contributors. No food concentration data.
Methods (brief)
NHANES 2011-2016 cycles (2011-2012, 2013-2014, 2015-2016). Blood metals by ICP-MS; serum metals by ICP-MS. CKD defined by CKD-EPI eGFR below 60 mL/min/1.73 m2 or ACR at least 30 mg/g. Metal concentrations below LOD replaced with LOD divided by square root of two. WQS, BKMR, and XGBoost applied.
Implications
Certification: WQS OR of 1.58 for the full blood metal mixture with lead and cadmium as dominant contributors is consistent with HMT&C analyte prioritization.
Courses: Useful for demonstrating how XGBoost can complement traditional epidemiological methods in mixture analysis.
App: No food concentration data.
Microbiome: Not applicable.
Related evidence
- Lead
- Cadmium
- health/renal-toxicity-metals
- Single-metal regulatory limits systematically underprotect against dietary heavy-metal exposure
Update history
No substantive edit history is available in this build. The full commit record is available in git.