Breathing polluted air has long been tied to individual diseases of the heart, the metabolism and the liver, but a new prospective cohort study from China suggests the damage may not stop at single organ systems. Researchers analyzing thousands of middle-aged and older adults found that long-term exposure to common ambient pollutants was associated with the emergence of cardiovascular–liver–metabolic multimorbidity, a clustering of at least two chronic conditions spanning the heart, blood vessels, blood sugar regulation and fatty liver disease. The work, published in BMC Public Health, is among the first to treat this triple burden of disease as a single outcome and to evaluate the effects of pollutant mixtures rather than one chemical at a time.
The investigation drew on the China Health and Retirement Longitudinal Study, a nationally representative survey of adults aged 45 and older. The team followed 4,650 participants living in 119 cities, using 2011 as the baseline and 2015 as the follow-up wave. At each time point, participants underwent structured interviews, physical examinations and blood-based testing. The researchers defined incident multimorbidity as the new coexistence of at least two of three domains: cardiovascular disease, diabetes, and an operationally defined probable form of metabolic dysfunction-associated steatotic liver disease, the fatty liver condition now recognized as one of the most common chronic liver disorders worldwide. By combining these domains, the study captured a pattern of disease that clinicians increasingly see but that environmental epidemiology has rarely addressed as a unified phenomenon.
Exposure assessment relied on the China High Air Pollutants database, which provides city-level concentrations of seven pollutants: fine particulate matter known as PM2.5, coarser particulate matter known as PM10, nitrogen dioxide, ozone, sulfur dioxide, carbon monoxide and particles smaller than one micron. City-level averages were linked to each participant according to their place of residence, an approach that reflects the reality that people living in the same urban environment share a broadly similar pollution profile. The study period captured the substantial air quality challenges faced by Chinese cities in the early 2010s, when concentrations of particulate matter in many regions far exceeded guideline values set by international health agencies.
During the four-year follow-up window, 757 participants, or 16.3 percent of the cohort, developed the combined cardiovascular–liver–metabolic multimorbidity. In fully adjusted single-pollutant models, the associations were striking. Each interquartile-range increase in PM2.5 concentration was associated with 33 percent higher odds of developing multimorbidity, with an odds ratio of 1.33 and a 95 percent confidence interval of 1.10 to 1.61. PM10 showed an even stronger signal, with an odds ratio of 1.42 per interquartile-range increase. Carbon monoxide and nitrogen dioxide each carried an odds ratio of 1.30, indicating a 30 percent elevation in odds. These effect sizes, observed over just four years in a middle-aged and older population, point to air pollution as a potentially meaningful accelerant of chronic disease clustering.
The statistical architecture of the study deserves attention because it reflects a broader shift in environmental health science. Rather than relying solely on classical single-pollutant regression, the researchers fitted logistic generalized linear mixed models with a city-level random intercept, which accounts for unmeasured characteristics shared by residents of the same city. They then applied weighted quantile sum regression, a technique designed for mixture analysis that estimates the joint effect of correlated pollutants while assigning each pollutant a weight reflecting its relative contribution. Because pollutants in the real world travel together, emitted by the same traffic, industry and power generation sources, mixture methods are widely regarded as closer to actual human exposure than models that isolate one contaminant at a time.
The mixture results reinforced the single-pollutant findings. The positive weighted quantile sum index, representing the combined burden of the pollutant mixture, was associated with incident multimorbidity with an odds ratio of 1.25 and a 95 percent confidence interval of 1.09 to 1.42. However, the picture grew more complicated when pollutants were adjusted against one another. In co-pollutant models, only PM10 remained associated with multimorbidity after adjustment for carbon monoxide, and no single pollutant retained statistical significance in the fully adjusted three-pollutant models. This attenuation is a familiar pattern in mixture epidemiology: because pollutants share sources and correlate strongly, disentangling which component carries the independent risk is statistically demanding, and the total mixture effect may be more informative than any individual estimate.
The team also probed whether pre-existing metabolic vulnerability modified the pollution effect, testing interactions with glycemic status and blood pressure. No multiplicative interaction was detected for either factor, meaning the relative association between pollution and multimorbidity did not differ measurably between people with and without elevated blood sugar. Yet odds-ratio-based additive interaction signals did emerge for PM2.5 and carbon monoxide among participants with elevated blood pressure, hinting that people with hypertension may experience a disproportionately larger absolute increase in risk. The authors emphasize that these additive signals warrant confirmation, particularly in studies with finer individual-level exposure assessment, but the finding is clinically relevant because elevated blood pressure is extraordinarily common in aging populations.
Machine learning added a complementary perspective. The researchers compared several city-grouped algorithms, including random forests, gradient boosting decision trees, extreme gradient boosting, k-nearest neighbors and multilayer perceptron models, and selected the best-performing approach for interpretation. Within the chosen random-forest model, Shapley additive explanations, a technique that quantifies each feature’s contribution to individual predictions, ranked carbon monoxide as the most influential pollutant with a mean absolute SHAP value of 0.0128, followed by PM10 at 0.0110 and ozone at 0.0090. The convergence of this data-driven ranking with the regression-based findings, particularly for carbon monoxide and particulate matter, strengthens the case that combustion-related pollutants deserve close scrutiny in multimorbidity research.
Why would polluted air promote disease across three organ systems simultaneously? The leading biological hypotheses center on chronic systemic inflammation and oxidative stress. Inhaled fine particles and gaseous pollutants can trigger inflammatory signaling that extends beyond the lungs, promoting atherosclerotic plaque development in blood vessels, impairing insulin sensitivity in metabolic tissues and encouraging fat accumulation and fibrogenesis in the liver. Shared upstream pathways, including autonomic nervous system activation, endocrine disruption and alterations in lipid metabolism, could plausibly drive the co-occurrence of cardiovascular disease, diabetes and steatotic liver disease rather than each condition arising independently. The multimorbidity framework used in this study is well suited to detecting such shared mechanisms, because it asks whether exposures push people toward clustered disease rather than isolated diagnoses.
The authors are careful about the limits of their evidence. Exposure was measured at the city level rather than at individual addresses, which introduces the possibility of exposure misclassification, and the operational definition of probable steatotic liver disease relied on surrogate markers rather than imaging confirmation. The observational design cannot exclude residual confounding by lifestyle, socioeconomic factors or healthcare access, and the four-year follow-up captures a relatively short slice of lifelong exposure. The researchers call for studies with finer exposure assessment and external validation, with particular attention to adults with elevated blood pressure. Even so, the findings carry a clear public health message: air pollution should be considered a potentially modifiable environmental correlate of multimorbidity, and reducing the combined burden of particulate matter, carbon monoxide and nitrogen dioxide may help slow the accumulation of chronic disease in aging populations, where multimorbidity drives much of the cost and complexity of care.
Subject of Research: Long-term ambient air pollution exposure and incident cardiovascular–liver–metabolic multimorbidity in middle-aged and older Chinese adults
Article Title: Long-term exposure to ambient air pollution mixtures and incident cardiovascular–liver–metabolic multimorbidity: a prospective cohort study
Article References: Wang, M., Wu, F., Zhang, J., & Yang, J. (2026). Long-term exposure to ambient air pollution mixtures and incident cardiovascular–liver–metabolic multimorbidity: a prospective cohort study. BMC Public Health. https://doi.org/10.1186/s12889-026-29891-9
Image Credits: AI Generated
DOI: 10.1186/s12889-026-29891-9
Keywords: air pollution, multimorbidity, PM2.5, PM10, carbon monoxide, nitrogen dioxide, steatotic liver disease, diabetes, cardiovascular disease, weighted quantile sum regression, CHARLS, machine learning
News Source: Russell Cooper. (October 13, 2026). Air Pollution Mixtures Linked to Heart, Liver and Metabolic Disease Clusters in Older Adults. Scienmag.
