Associations between air pollution and cardiorespiratory function and metabolic indicators: New insights into mixture...

Air pollution is a major global health risk that adversely affects cardiovascular, respiratory, and metabolic systems. Previous studies have predominantly focused on the health effects of individual pollutants to identify pollutants with greater impacts on health. However, under real-world mixture-exposure scenarios, it remains unclear whether pollutants that show non-significant associations in single-pollutant models may still affect health. Using data from the Chronic Obstructive Pulmonary Disease in Beijing (COPDB) study, we analyzed 112 older adults who contributed 410 follow-up visits. We continuously measured fine particulate matter (PM_2.5), ultrafine and accumulation-mode particles (UFPs and Acc), gaseous pollutants (SO₂, NO, NO₂, and O₃), and PM_2.5 components (carbonaceous species, metals, and water-soluble ions). Fasting blood glucose (FBG), heart rate (HR), and peak expiratory flow (PEF) were selected as health indicators in this study. Using a Bayesian kernel machine regression (BKMR) model, we first identified pollutants associated with each health indicator in single-pollutant models. The results indicated that SO₂ and UFPs were significantly associated with increased FBG; Ca^2 + in PM_2.5 was significantly associated with increased HR; and O₃ and UFPs were significantly associated with decreased PEF. After excluding the pollutants associated with each health indicator, we further evaluated the overall associations between the remaining pollutants and health indicators, and then assessed pairwise interaction associations among all pollutants. The results showed that the overall associations became stronger and more statistically significant, particularly for the metal and gaseous pollutant groups. For example, over the 14-day exposure window, the gaseous pollutant group was significantly associated with a 19.1% increase in FBG (10.8%, 27.4%) and a 16.0% decrease in PEF (8.7%, 23.3%). Additionally, significant associations suggested potential synergistic interactions, mainly involving the metal group. Evaluating health impacts from a mixture-exposure perspective is essential. Even pollutants that do not show significant associations in single-pollutant models may still affect health under mixture exposure.