Long-term exposure to ambient air pollution and COVID-19 incidence in Italy

Abstract

Background and Aim The SARS-CoV-2 pandemic gave rise to a large number of epidemiological studies linking ambient air pollution (AP) to COVID-19 incidence. However, to date, the role of chronic exposure to AP in explaining COVID-19 occurrence is still not clarified. Methods Within a national study (EpiCovAir) promoted by National Institute of Health (ISS) and Italian Network of Environmental Protection (SNPA), we identified all COVID-19 cases in Italy from February 2020 to June 2021. Chronic exposure to particulate matter < 10 micron (PM10), < 2.5 micron (PM2.5) and nitrogen dioxide (NO2) were assigned at municipality level (n=7800) using a national satellite-based AP exposure model (1-km2 spatial resolution) for the period 2016-2019. We applied principal component analysis (PCA) and generalized propensity score (GPS) approaches to summarize information of about 50 area-level covariates to account for major determinants of the spatial distribution of COVID-19 cases (municipality characteristics, population density, mobility, population health, socio-economic status). Finally, we applied generalized negative-binomial regression models matched on GPS, age, sex, province and month to study the association between AP and COVID-19 incidence. Results 3,995,202 COVID-19 cases occurred in Italy in the entire period (incidence rate=67x1000). Overall, incidence rates increase by 0.6% (95% CI: 0.5%, 0.7%), 0.5% (95% CI: 0.4%, 0.6%) and 0.9% (95% CI: 0.9%, 1.0%) per 1 g/m3 increment in PM2.5, PM10 and NO2, respectively. Associations were consistent to different sensitivity analyses, particularly for NO2. Increments were higher among elderly subjects, and during the second pandemic waves (Sept. 2020-Dec. 2020). In two-pollutant models, NO2 estimates were robust to PM adjustment, while PM showed no associations after adjustment for NO2. Conclusions We found evidence of association between long-term exposure to ambient AP (particularly NO2) and the occurrence of 4 million COVID-19 cases in Italy, suggesting a role of AP in increasing the incidence of COVID-19 disease.