Introduction

Advancements in healthcare and increasing life expectancy have led to a rising prevalence of multimorbidity, thereby imposing a substantial economic burden on society1. Multimorbidity, defined by the coexistence of two or more chronic conditions in a single individual, is particularly prevalent among the elderly population and is linked to increased healthcare utilization, diminished quality of life, and an elevated risk of mortality2. Cardiometabolic multimorbidity (CMM) refers to the concurrent presence of two or more cardiometabolic disorders (CMDs), including diabetes, cardiovascular diseases (CVDs), and stroke. This condition affects approximately 30% of the elderly population and is linked to increased mortality, diminished quality of life, and substantial healthcare expenditures2,3,4. Research has demonstrated that people with multiple CMDs have considerably elevated mortality rates compared to those with a single CMD4.

The rising prevalence of multimorbidity, particularly CMM, has become a critical public health concern in aging populations worldwide5. In the United States, the occurrence of CMM stands at 34.8% among individuals aged 60–79 years and increases to 42.9%6 among individuals aged 80 years and above7. In China, rapid urbanization and demographic aging have exacerbated this issue, with the incidence of CMM among elderly adults reaching 16.9%, a figure projected to rise as the population over 60 years old grows to 28% by 20408. While lifestyle factors and genetic predisposition contribute to CMM, environmental determinants, especially ambient air pollution, have gained recognition as modifiable risk factors9,10. These findings underscore the necessity of identifying potential risk factors for CMM in aging populations, particularly in the context of rapid demographic shifts and urbanization in China9,10.

Ambient air pollution, a pervasive consequence of industrialization and urbanization, is a leading environmental hazard in China11. In numerous Chinese cities, annual average concentrations of PM10 in many Chinese cities exceed the World Health Organization’s guideline of 20 µg/m3, with industrial emissions, coal combustion, and vehicular exhaust being primary contributors11. Ambient air pollution represents a substantial environmental determinant of CMM, including hypertension, obesity, and dyslipidemia10. These disorders contribute substantially to the health burden, particularly among middle-aged and elderly populations in China10. Previous research has established indoor air quality degradation as a significant risk factor for CMM and emphasized the potential role for a healthy lifestyle to mitigate the prevalence of CMM9. However, investigation on the prolonged impact of outdoor air pollution exposure on CMM remains limited. Projections from the Global Burden of Disease study indicate that mortality due to chronic diseases linked to ambient air pollution could rise by over 50% by 205012. Research on outdoor air pollution and CMM has predominantly concentrated on particulate matter (PM2.5) and ozone (O3)8. However, significant gaps persist in our comprehension of the roles played by pollutants like nitrogen dioxide (NO2), sulfur dioxide (SO2), and both fine (PM1) and coarse (PM10) particulate matter in the pathogenesis of CMM8.

To address the gaps in the current literature, our research seeks to examine the relationship between sustained exposure to outdoor air pollution, specifically PM1, PM2.5, PM10, O3, NO2, and SO2, and the development of CMM, utilizing data from the China Health and Retirement Longitudinal Study (CHARLS). The anticipated result of this research is to generate valuable insights and provide evidence-based recommendations regarding the risk factors for CMM, which could inform strategies for its prevention and management among elderly individuals in China.

Methods

Study design

Our investigation utilized data from the CHARLS, a nationally representative survey of Chinese adults in middle and later life13. The CHARLS study, which began in 2011, collects comprehensive data on health, socioeconomic status, and behavioral factors through biennial surveys. Comprehensive details regarding the study design, methodologies, and participant characteristics have been previously documented14. For our analysis, we selected participants aged 45 years or older from the 2015 CHARLS dataset (wave 3). The initial cohort comprised 21,095 individuals, from which we excluded those who engaged in less than 10 min of continuous exercise per week, as our analysis focused on the effects of chronic outdoor air pollution on CMM. Participants with insufficient weekly exercise duration were not considered to be in a state of chronic outdoor air pollution exposure. After applying these criteria, the final analytic sample consisted of 9,830 individuals from 28 provinces across the country (Fig. 1).The CHARLS study protocol received approval from the Biomedical Ethics Review Committee of Peking University (Beijing, China), and written informed consent was obtained from all participants prior to their involvement in the study.

Fig. 1
figure 1

, Participant screening flowchart for this study. CHARLS, China Health and Retirement Longitudinal Study. CMM, cardiometabolic multimorbidity.

Full size image

Assessment of CMM

In line with prior research, the presence of CMM was assessed using three key questions, which addressed the history of diabetes or hyperglycemia, heart disease (including conditions such as myocardial infarction, angina, coronary artery disease, or heart failure), and stroke4. For study cohort stratification, participants manifesting concurrent presentation of any two aforementioned comorbidities were designated as CMM-positive cases.

Assessments of air pollution exposure

Ground-level concentrations of six primary pollutants (PM1, PM2.5, PM10, SO₂, NO₂, and O₃) were estimated using a hybrid spatiotemporal prediction framework from the CHAP (https://weijing-rs.github.io/product.html)15. The model integrates ground monitoring data from the China National Environmental Monitoring Center, satellite remote sensing products (e.g., aerosol optical depth, trace gas inversion data), atmospheric reanalysis fields (temperature, humidity, wind speed/direction, and planetary boundary layer height), land use parameters (road/traffic density, topography, and population distribution), and industrial and motor vehicle emission inventories16. Predictions were generated using the Spatio-Temporal Extreme Random Tree (STET) algorithm, a machine learning method demonstrating high predictive performance in estimating fine-scale pollution levels across China15,16,17. Detailed data have been described in previous studies15,16. A full summary of model performance, including R² and root mean square error (RMSE), is presented in Supplementary Table 1. We chose county-level exposure estimates instead of finer grid-level predictions because CHARLS participants’ residential information was available only at the county level. This approach minimized misclassification risk while ensuring compatibility between health and exposure data. Sensitivity analyses were performed using a two-year average concentration18.

Covariates

Meteorological factors, socio-demographic characteristics, socio-economic variables, health behaviors, and lifestyles were considered potential confounders in the analysis. Meteorological factors, such as temperature and relative humidity, were acquired from monitoring stations operated by the China Meteorological Administration (http://www.cma.gov.cn/). Socio-demographic characteristics encompassed age (years) and gender (male or female). Socioeconomic factors comprised place of residence (rural or urban), educational level (“no formal education or primary education” or “secondary education or higher”), marital status (“married and cohabiting with a spouse,” “married but separated from spouse” or “unmarried, divorced, and widowed”), and yearly household expenditure. Health behaviors and lifestyle factors comprised smoking behavior (“non-smoker” or “smoker”), Drinking status (“do not drink,” “drink < once a month,” or “drink > once a month”), physical activity (PA), and type of cooking fuel utilized (“solid fuel” or “clean fuel”)19. Smoking status was classified according to the individual’s history of using tobacco, which included chewing tobacco, smoking a pipe, self-rolling cigarettes, or the use of commercial cigarettes or cigars. Alcohol intake was evaluated based on self-reported drinking behavior over the past year, including the type of alcohol consumed (spirits, wine, or beer). Drinking status was assessed based on self-reported drinking behavior over the past year, including the type of alcohol consumed (spirits, wine, or beer). PA was quantified using the International PA Questionnaire (IPAQ). Time intervals within CHARLS were transformed into intermediate values as follows: “≥4 hours” to 240 min, “≥2 hours and < 4 hours” to 180 min, “≤30 minutes and < 2 hours” to 75 min, and “≥10 minutes and < 30 minutes.” PA score was calculated based on metabolic equivalents: PA score = 8.0 × total duration of vigorous activity per week + 4.0 × total duration of moderate activity per week + 3.3 × total duration of walking per week20,21. Indoor air pollution was evaluated based on the categories of cooking fuel employed. Solid fuels, including coal, crop residue, and wood, were classified as “solid fuels,” while natural gas, biogas, and liquefied petroleum gas were designated as “clean fuels”22.

Statistical analysis

Continuous variables in this research were presented as the mean ± standard deviation (SD), whereas categorical variables were presented as frequencies (percentage). To evaluate the effect of air pollution on CMM and explore potential interactions, we utilized generalized linear models (GLM). Effect estimates and corresponding 95% confidence intervals (95% CI) were reported as odds ratios (OR) per interquartile range (IQR) change in air pollutant concentrations. Initially, a crude model was applied, followed by an adjusted Model 1 (Model 2). These models incorporated meteorological variables (temperature and relative humidity), socio-demographic factors (age and gender), and socio-economic indicators (residence, education level, marital status, and yearly household expenditure)23. A further refinement in adjusted Model 2 (Model 3) adjusted for health behaviors and lifestyle factors, including smoking, alcohol consumption, PA, and cooking fuel use. Additionally, standardized air pollution concentrations were incorporated into the model to enable comparison of OR values across different pollutants, allowing for identification of the specific air pollution components most strongly associated with the increased prevalence of CMM.

To strengthen the reliability of our investigation, we conducted a series of sensitivity analyses. First, our research recalculated the GLM and interaction analyses using average air pollutant concentrations from 2013 to 2014. Second, we incorporated regional classifications (“East”, “Central” and “West”) as additional covariates. Third, participants who had relocated between 2011 and 2015 were excluded to reduce potential confounding effects related to geographic mobility. Finally, to avoid confounding about missing data in the CMM for this research, we removed missing data in the CMM and recalculated the GLM after using three years of average air pollution.

All statistical analyses were carried out utilizing R software (version 4.4.2), with missing data for several covariates imputed utilizing the “mice” package24. The correlation between air pollution and CMM was evaluated, with statistical significance set at a two-sided p-value of < 0.05.

Results

Descriptive statistics

The study cohort comprised 9,830 individuals aged 45 and above from 126 urban districts across 28 provinces in China. Figure 2 provides a spatial representation of the distribution of particulate matter (PM1, PM2.5, PM10), SO2, NO2, and O3 across the country in 2015. Table 1 summarizes the demographic characteristics of the individuals. Among the 9830 participants obtained after screening in this study, the average age was 60.1 years, with a gender distribution of 48.9% male participants. The prevalence of CMM was 13.1%, and the mean BMI was 24.1. Among the 1286 CMM patients, the average age was 64.5 years, with 43.4% of participants being male, and the mean BMI was 26.4.

Table 1 Basic characteristics of participants.
Full size table
Fig. 2
figure 2

, Distribution of air pollutants in China in 2015. (A) NO2 concentration distribution by city. (B) O3 concentration distribution by city. (C) PM1 concentration distribution by city. (D) PM2.5 concentration distribution by city. (E) PM10 concentration distribution by city. (F) SO2 concentration distribution by city.

Full size image

Association between air pollution and the prevalence of CMM

Our results had the following findings. In the crude model (Model 1), long-term exposure to PM1 (OR = 1.01, 95% CI: 1.00–1.02, P = 1.21 × 10− 3), PM2.5 (OR = 1.01, 95% CI: 1.00–1.01, P = 5.81 × 10− 6), PM10 (OR = 1.01, 95% CI: 1.00–1.02, P = 2.13 × 10− 9), SO2 (OR = 1.01, 95% CI: 1.01–1.01, P = 1.24 × 10− 6), NO2 (OR = 1.01, 95% CI: 1.00–1.02, P = 6.26 × 10− 4), and O3 (OR = 1.01, 95% CI: 1.00–1.02, P = 1.34 × 10− 2) all increased the prevalence of CMM (Fig. 3A). In adjusted Model 1 (Model 2), chronic exposure to PM2.5 (OR = 1.01, 95% CI: 1.00–1.01, P = 1.45 × 10− 3), PM10 (OR = 1.00, 95% CI: 1.00–1.01, P = 6.26 × 10− 6) and SO2 (OR = 1.01, 95% CI: 1.00–1.01, P = 1.08 × 10− 3) were linked to increased prevalence of CMM (Fig. 3A). In adjusted Model 2 (Model 3), chronic exposure to PM10 (OR = 1.01, 95% CI: 1.00–1.02, P = 3.95 × 10− 2) was related to an increased prevalence of CMM (Fig. 3A). In addition, we found an elevated OR of 1.33 for PM10 by including standardized air pollution concentrations in the model (Fig. 3B). These findings underscore the potential impact of various air pollutants, particularly PM10, on the incidence of CMM (Supplementary Table 2).

Sensitivity analysis

The GLM was recalculated using average air pollutant concentrations from 2013 to 2014, with results consistent with the primary analysis (OR = 1.01, 95% CI: 1.00–1.02, P = 4.32 × 10− 2) (Supplementary Fig. 1A). Both PM10 (OR = 1.01, 95% CI: 1.00–1.02, P = 3.38 × 10− 2) and O3 (OR = 1.05, 95% CI: 1.01–1.09, P = 2.54 × 10− 2) were found to be linked to increased CMM prevalence when regional categorization (“East,” “Central,” and “West”) was included as an additional covariate (Supplementary Fig. 1B). In addition, after excluding potential confounding effects related to geographic migration, our analysis of 6,926 participants (out of a total of 2,904 participants who geographically migrated) showed that only PM10 (OR = 1.01, 95% CI: 1.00–1.02, P = 4.19 × 10− 2) remained significantly linked to increased CMM risk (Supplementary Fig. 1C). Finally, our repeated calculations for 8,048 participants (a total of 1,782 participants with CMM deletion) showed that PM10 (OR = 1.01, 95% CI: 1.00–1.02, P = 3.48 × 10− 2) remained a possible underlying risk factor for CMM (Supplementary Fig. 1D). In summary, these sensitivity analyses further elucidate the potential impact of various air pollutants, particularly PM10, on CMM incidence rates (Supplementary Table 3).

Fig. 3
figure 3

, Forest plots for the two analyses. (A) Forest plots for major analyses. (B) Forest plot of the main analyses after adding normalized concentrations of air pollutants.

Full size image

Discussion

Our research, involving a substantial cohort of Chinese adults aged 45 and above, provides a comprehensive assessment of the impact of choric exposure to air pollution on the prevalence of CMM. Our findings suggest that chronic exposure to air pollution, particularly PM10, is related to an increased prevalence of CMM (OR = 1.01, 95% CI: 1.00–1.02, P = 3.95 × 10− 2). These observations align with prior research that has emphasized the harmful impact of air pollution on cardiometabolic health, particularly among aging populations5,8,25,26,27. For instance, Su et al. demonstrated that chronic exposure to PM2.5 and O3 was significantly linked to CMM among the elderly in China, further supporting the notion that particulate matter is a critical environmental risk factor for CMM8. Similarly, Luo et al. found that chronic exposure to ambient air pollution was a contributing factor for the progression of CMM in the UK Biobank cohort, emphasizing the global relevance of our findings28.

Our study expands upon previous research by examining a broader range of air pollutants, namely PM1, PM2.5, PM10, SO2, NO2, and O3, and their associations with CMM. While earlier studies have primarily focused on PM2.5 and O3, our findings highlight the significant role of PM10 in the development of CMM, which has been less explored in the literature9,13,15. This is particularly pertinent in the context of China, where PM10 levels remain elevated in many regions due to ongoing industrial activities and rapid urbanization29. Furthermore, our findings align with those of Chen et al., who found household air pollution, particularly from solid fuels, was linked to an increased risk of CMM, suggesting that both indoor and outdoor air pollution contributes to cardiometabolic health risks9.

The mechanisms underlying the association between PM10 and CMM are multifaceted. Studies have shown that exposure to PM10 triggers systemic inflammation and oxidative stress through inhalation of particulate-bound reactive oxygen species (ROS) and pro-inflammatory components (such as transition metals, and polycyclic aromatic hydrocarbons). These particles penetrate alveolar tissue, activate alveolar macrophages and epithelial cells, and release pro-inflammatory cytokines (such as IL-6, TNF-α) and chemokines into the systemic circulation30. Chronic low-grade inflammation disrupts insulin signaling pathways, promotes endothelial dysfunction, and accelerates atherosclerosis, which collectively increases susceptibility to CMM. Oxidative stress further exacerbates mitochondrial dysfunction in metabolic tissues, impairing glucose homeostasis and lipid metabolism31. Similarly, certain gaseous pollutants like CO can impair oxygen delivery and exacerbate cardiovascular strain, while volatile organic compounds such as benzene and formaldehyde have been linked to metabolic dysfunction and endothelial damage through oxidative stress pathways32,33,34.

Moreover, the role of PM10 in the development of CMM may also be mediated through its impact on autonomic nervous system dysfunction. Exposure to particulate matter has been shown to alter heart rate variability (HRV), a marker of autonomic balance, which is closely linked to cardiovascular health35. Additionally, PM10 exposure has been associated with increased blood pressure and arterial stiffness, both of which are key risk factors for cardiovascular diseases36. These pathways collectively contribute to the increased prevalence of CMM observed in our research.

Although PM2.5 is generally considered more harmful to human health due to its larger surface area, our findings reveal a potentially significant association between PM10 and CMM30,35,37. This apparent inconsistency may be attributable to several factors. First, while PM10 is primarily composed of coarse particles deposited in the upper respiratory tract, these particles can still induce substantial systemic effects through inflammatory responses and oxidative stress triggered in the airway epithelium31. Second, the chemical composition of PM10 in many Chinese regions is complex and often enriched with toxic heavy metals and polycyclic aromatic hydrocarbons derived from coal combustion and industrial emissions29, which may amplify its cardiometabolic toxicity beyond what would be expected based solely on particle size. Third, our data-driven statistical approach captures population-level associations that may reflect both exposure differences (higher regional levels of PM10 compared with PM2.5 in some provinces) and heterogeneous vulnerability among elderly individuals. Thus, although PM2.5 is widely recognized as more harmful in toxicological studies, our epidemiological findings emphasize that PM10 remains a critical environmental risk factor in China’s specific context.

The weaker or inconsistent associations observed for PM2.5, SO2, and other pollutants are noteworthy. Several factors may contribute to these discrepancies. First, pollutant composition varies by region and season29. Second, exposure misclassification could attenuate associations, as pollutant estimates at county-level may not fully capture intra-urban variability, particularly for traffic-related pollutants such as NO216. Third, biological mechanisms may differ across pollutants; SO2, for instance, is highly water-soluble and primarily affects the upper airways, which may explain its weaker systemic effects compared to particulate matter35,37,38,.

This investigation is subject to several limitations that warrant consideration. Primarily, the observational design restricts the capacity to establish causal relationships. While a robust association between PM10 and CMM was observed, longitudinal studies are necessary to validate these results and further elucidate potential causal pathways. Second, residual confounding due to unmeasured variables (such as genetic predisposition or dietary habits) may still influence the findings, despite our adjustment for a range of confounders. Third, certain covariates, including PA and smoking status, were self-reported, which could introduce measurement bias. Future research should incorporate objective measures of PA and smoking status to reduce potential bias. Specifically, self-reported physical activity may be subject to recall bias or social desirability bias, potentially leading to overestimation of activity levels. Similarly, smoking status may be underreported due to stigma or recall inaccuracies. Such misclassification could attenuate the observed associations between air pollution and CMM. Future studies should incorporate objective measurement methods to assess PA levels and smoking status, thereby minimizing potential bias and enhancing the accuracy of exposure assessment.

Despite these limitations, our results hold potential implications for public health. The observed association between PM10 and CMM suggests that reducing ambient air pollution, especially particulate matter, may contribute to the prevention and management of CMM in aging populations. However, given the cross-sectional nature of this study, these findings should be interpreted with caution and do not directly support specific public health interventions or policy decisions. Further longitudinal studies are needed to confirm these associations and elucidate causal pathways before concrete recommendations can be made.

Conclusion

This cross-sectional study provides evidence that long-term exposure to air pollution, particularly PM10, is associated with a higher prevalence of CMM among middle-aged and older adults in China. These findings highlight the need for further longitudinal research to establish causality and explore underlying mechanisms. They also underscore the potential importance of air quality management in public health strategies aimed at reducing the burden of cardiometabolic diseases in aging populations.