Objective To investigate the association of particulate matter and ozone with the prevalence of non-alcoholic fatty liver disease (NAFLD) in a district of Shanghai, and to provide epidemiological evidence for the further identification of early health hazards of air pollution and for the prevention and control of NAFLD.

Methods Based on Songjiang Sub-cohort of Shanghai Natural Population Cohort, a cross-sectional survey design was used to recruit participants from 2016 to 2017. Annual average exposure levels to air pollution from 2009 to 2017 were matched to the participant’s residential address using a high-resolution and high-quality ambient air pollutants dataset in China. NAFLD was diagnosed according to the “Guidelines for the prevention and treatment of metabolism⁃associated (non⁃alcoholic) fatty liver disease” by the Chinese Medical Association. Multivariate logistic regression models were employed to analyze the association between air pollution and the prevalence of NAFLD, and stratified analyses were used to compare differences by age, gender, obesity, and lifestyle habits within subgroups.

Results A total of 32 791 individuals were included in the study. The prevalence of NAFLD among community residents in suburban Shanghai was 38.88%. For every 1 μg·m^-3 increase in PM₁, PM_2.5, PM₁₀, or O₃, the risk of NAFLD increased correspongdinglt, with the odds ratios (95%CI) of 1.071 (1.043‒1.099), 1.065 (1.042‒1.089), 1.041 (1.027‒1.055), or 1.061 (1.032‒1.091), respectively. There were differences in effects across different gender, age, and obesity status subgroups.

Conclusion PM₁, PM_2.5, PM₁₀, and O₃ are positively associated with an increased risk of NAFLD. Stratified analyses reveal that individuals aged 65 years old or above exhibited greater susceptibility to PM₁, PM_2.5, and O₃, whereas those aged less than 65 years old are more vulnerable to PM₁₀. Males are more sensitive to PM₁ and O₃, and females are more susceptible to PM_2.5 and PM₁₀. The association between air pollutant exposure and NAFLD risk is more pronounced among obese participants compared to that in non-obese counterparts.