No abstract

Background and objective Ecological studies indicate ambient particulate matter ≤2.5mm (PM2.5) air pollution is associated with poorer COVID-19 outcomes. However, these studies cannot account for individual heterogeneity and often have imprecise estimates of PM2.5 exposure. We review evidence from studies using individual-level data to determine whether PM2.5 increases risk of COVID-19 infection, severe disease, and death. Methods Systematic review of case-control and cohort studies, searching Medline, Embase, and WHO COVID-19 up to 30 June 2022. Study quality was evaluated using the Newcastle-Ottawa Scale. Results were pooled with a random effects meta-analysis, with Egger′s regression, funnel plots, and leave-one-out and trim-and-fill analyses to adjust for publication bias. Results N=18 studies met inclusion criteria. A 10μg/m3 increase in PM2.5 exposure was associated with 66% (95% CI: 1.31-2.11) greater odds of COVID-19 infection (N=7) and 127% (95% CI: 1.41-3.66) increase in severe illness (hospitalisation or worse) (N=6). Pooled mortality results (N=5) were positive but non-significant (OR 1.40; 0.94 to 2.10). Most studies were rated "good" quality (14/18 studies), though there were numerous methodological issues; few used individual-level data to adjust for confounders like socioeconomic status (4/18 studies), instead using area-based indicators (12/18 studies) or not adjusting for it (3/18 studies). Most severity (9/10 studies) and mortality studies (5/6 studies) were based on people already diagnosed COVID-19, potentially introducing collider bias. Conclusion There is strong evidence that ambient PM2.5 increases the risk of COVID-19 infection, and weaker evidence of increases in severe disease and mortality.