Work

Nonlinear associations between serum vitamin mixtures and cardiovascular disease risk: insights from a national cross-sectional analysis

Year: 2026

Type: article

Abstract: OBJECTIVES: To evaluate the joint, nonlinear, and interactive associations between multiple serum vitamins and cardiovascular disease (CVD), addressing the limitations of prior single-nutrient studies. METHODS: Using nationally representative adults from the National Health and Nutrition Examination Survey (NHANES) 2005-2006 (n = 3055), we adjusted for potential confounders and applied complementary, advanced mixture-modeling techniques-multivariable logistic regression, adaptive elastic net with environmental risk score (AENET-ERS), quantile g-computation (Qgcomp), and Bayesian kernel machine regression (BKMR)-to assess component-specific and mixture effects, variable contributions, nonlinearity, and interactions. RESULTS: In the NHANES 2005-2006 adult sample, advanced mixture modeling was used to assess the association between vitamin mixtures and CVD. In fully adjusted logistic models, per log-unit increases in vitamins C, D, and E were associated with lower CVD risk, and quartile (Q4 vs Q1) analyses were directionally consistent; vitamins A and B6 were not significant, and B12 showed inconsistent associations. Qgcomp indicated an overall protective mixture effect (per one-quantile increase, logRR - 0.231; RR ≈ 0.79; 95% CI 0.68-0.92), with risk weights primarily from B6 (0.594) and A (0.406) and protective weights from D (- 0.375), C (- 0.359), and E (- 0.251); B12 contributed minimally (- 0.014). AENET-ERS retained E, D, C, B6, and A (β: - 0.219, - 0.204, - 0.204, + 0.195, + 0.179; B12 excluded), while BKMR posterior inclusion probabilities (PIPs) were highest for E (0.943) and D (0.936), followed by B6 (0.857), A (0.796), C (0.735), and B12 (0.437), revealing nonlinear overall protective effects and suggesting that vitamin D modifies the B6-CVD association. CONCLUSIONS: The vitamin mixture shows an overall protective effect associated with lower CVD risk; contributions are primarily driven by vitamins C, D, and E, with A and B6 contributing risk and B12 playing a minimal role, and there is exploratory evidence of a D × B6 interaction.

Source: European journal of medical research

DOI: 10.1186/s40001-025-03741-7