Long-term exposure to PM2.5 and depression in China: a systematic review

Long-term exposure to fine particulate matter (PM2.5) has been increasingly associated with adverse mental health outcomes, but evidence regarding depression among Chinese populations remains fragmented. This systematic review synthesized epidemiological evidence on the association between long-term ambient PM2.5 exposure and depression or depressive symptoms in China. PubMed and Scopus were searched for English-language studies published from January 2015 through December 2025 that examined human populations in China and assessed ambient PM2.5 exposure over at least one year in relation to depression or depressive symptoms. Studies of indoor or occupational exposure and short-term exposure were excluded. Sixteen studies met the eligibility criteria, including cross-sectional, longitudinal, cohort, and quasi-experimental studies, with sample sizes ranging from 1,833 to 30,712 participants. Most studies were conducted among middle-aged and older adults. Long-term PM2.5 exposure was generally positively associated with depressive symptoms or depression across different exposure assessment methods and study designs. Reported effect estimates varied substantially because of differences in exposure contrasts, exposure windows, outcome measures, and statistical models. Several studies additionally suggested that socioeconomic conditions, physical activity, and PM2.5 components may modify or mediate the observed associations. The evidence consistently indicates a positive association between long-term ambient PM2.5 exposure and depression or depressive symptoms among Chinese populations. Although the predominantly observational nature of the evidence limits causal interpretation, the findings consistently suggest that long-term PM2.5 exposure is associated with a higher likelihood of depression and depressive symptoms, particularly among middle-aged and older adults. Further longitudinal and quasi-experimental studies are needed to strengthen causal inference.