Long-term exposure to ambient air pollution mixtures and incident cardiovascular–liver–metabolic multimorbidity: a prospective cohort study

Abstract Background Ambient air pollution is associated with cardiovascular, metabolic, and liver diseases, but evidence for cardiovascular–liver–metabolic multimorbidity (CLMM) and pollutant mixtures is limited. We examined these associations in middle-aged and older adults in China. Methods We analyzed 4,650 adults from 119 cities in the China Health and Retirement Longitudinal Study at the 2011 baseline and 2015 follow-up. Incident CLMM was defined as the new coexistence of at least two domains: cardiovascular disease, diabetes, and operationally defined probable metabolic dysfunction-associated steatotic liver disease. City-level concentrations of seven pollutants were obtained from the China High Air Pollutants database. Logistic generalized linear mixed models with a city-level random intercept were supplemented by weighted quantile sum (WQS) regression, interaction analyses, and city-grouped machine-learning models. Results During follow-up, 757 participants (16.3%) developed CLMM. In fully adjusted single-pollutant models, the odds ratios per interquartile-range increase were 1.33 (95% CI 1.10–1.61) for PM₂.₅, 1.42 (1.13–1.77) for PM₁₀, 1.30 (1.08–1.56) for CO, and 1.30 (1.06–1.58) for nitrogen dioxide. Estimates were attenuated after co-pollutant adjustment: only PM₁₀ remained associated with CLMM after adjustment for CO; and no association remained statistically significant in the three-pollutant models. The positive WQS index was associated with CLMM (OR 1.25, 95% CI 1.09–1.42). No multiplicative interaction was detected by glycemic or blood-pressure status. Odds-ratio-based additive interaction signals were observed only for PM₂.₅ and CO with elevated blood pressure. In the selected random-forest model, the mean absolute Shapley additive explanations (SHAP) value was largest for CO (0.0128), followed by PM₁₀ (0.0110) and O₃ (0.0090). Conclusions Long-term exposure to PM₂.₅, PM₁₀, CO, nitrogen dioxide and the positive pollutant-mixture index was associated with incident CLMM. The findings support further evaluation of air pollution as a potentially modifiable environmental correlate of multimorbidity. Studies with finer exposure assessment and external validation are needed, particularly for adults with elevated blood pressure.