Long-term exposure to air pollution and severe COVID-19 in Catalonia: a population-based cohort study

Ambient air pollution significantly contributes to the global disease burden, particularly cardiovascular and respiratory illnesses. While the effects of short- and long-term exposure to air pollution on chronic respiratory diseases are well-documented, research on the impact of long-term exposure on the incidence and severity of acute respiratory infections is limited. COVID-19, primarily an acute respiratory infection caused by SARS-CoV-2, has several identified risk factors for severe disease and mortality, including age, sex, and pre-existing conditions. Air pollutants can impair lung defenses against infections and may upregulate the expression of SARS-CoV-2 receptors in the lungs. Ecological studies have suggested a link between air pollution and increased COVID-19 hospitalization and death risk, but individual-level cohort studies are needed to address methodological limitations of ecological studies. Existing individual-level studies have shown varying associations between long-term air pollutant exposure and COVID-19 severity, highlighting the need for a larger, more comprehensive study. This research addresses these gaps by analyzing a large population-based cohort in Catalonia to investigate the associations between PM2.5, NO2, O3, and BC and various COVID-19 severity outcomes.

The existing literature on the relationship between long-term exposure to ambient air pollutants and severe COVID-19 outcomes is mixed. Some studies have found positive associations between exposure to fine particulate matter (PM2.5) and increased risk of hospitalization or death due to COVID-19, while the association with nitrogen dioxide (NO2) is less consistent. Many studies have focused on cohorts of individuals with confirmed COVID-19 infections or specific populations, limiting generalizability. Methodological heterogeneity in outcome definitions and sample sizes have also contributed to the inconsistent findings. This study aims to address these limitations by employing a large population-based cohort and detailed exposure assessment, allowing for robust evaluation of associations while controlling for multiple potential confounders.

This population-based cohort study used data from the COVAIR-CAT study, encompassing the adult population of Catalonia, Spain. Data were obtained through record linkage of public health administration databases, including sociodemographic information, healthcare utilization (primary care, urgent care, hospital discharge), and COVID-19 testing results. The study included 4,660,502 adults alive and residing in Catalonia on March 1, 2020, after exclusions for loss to follow-up, inconsistent data, missing residential addresses, and missing air pollution exposure data. Exposure to PM2.5, NO2, O3, and black carbon (BC) was assessed using annual average values from the COVAIR-CAT exposure assessment models, which were created using meteorological and air pollution data at a high spatial resolution (250m). In a sensitivity analysis, estimates from land-use regression models (ELAPSE project) from 2010 were used for comparison. Cox proportional hazards models were used to analyze the association between air pollutant exposure and COVID-19-related hospitalization, ICU admission, death, and hospital length of stay. The main model adjusted for age (penalized spline), sex, smoking status, income, health risk group, socioeconomic index, proportion of non-Spanish nationals, distance to nearest primary care unit, urbanicity, weekly test-positive proportion, and health region. Several sensitivity analyses were performed, including adjustments for potential mediating factors, different socioeconomic indicators, multiple imputation of missing data, use of laboratory-confirmed cases only, inclusion of nursing home residents, and restriction to individuals with stable residences. Additional sensitivity analyses explored cause-specific hospitalizations and assessment of the effects during the different waves of the pandemic. Negative binomial regression models were used for hospital length of stay. The proportional hazards assumption was checked visually.

The study found a positive association between long-term exposure to PM2.5 and NO2 and severe COVID-19 outcomes. Specifically, higher annual average exposure to PM2.5 and NO2 was associated with increased hazard of COVID-19-related hospitalization, ICU admission, and death. For PM2.5, a one interquartile range (IQR) increase was associated with a 19% increase in hospitalizations (95% CI, 16–21%), a 16% increase in ICU admissions (95% CI, 1.09-1.24), and a 13% increase in deaths (95% CI, 1.07-1.19). For NO2, a one IQR increase was associated with a 25% increase in hospitalizations (95% CI, 1.22-1.29), a 42% increase in ICU admissions (95% CI, 1.30-1.55), and an 18% increase in deaths (95% CI, 1.10-1.27). Black carbon (BC) showed a positive association with hospitalizations (HR 1.19, 95% CI, 1.16–1.22), ICU admissions (HR 1.19, 95% CI, 1.10–1.28), and deaths (HR 1.06, 95% CI, 1.00–1.13). The associations remained largely consistent across sensitivity analyses using different models, outcome definitions, and exposure assessment methods. There were notable differences between the first and second COVID-19 waves with larger effects observed in the first wave, this could be due to unmeasured contextual factors or higher susceptibility in the earlier phase of the pandemic. O3 showed positive associations with severe outcomes when adjusted for NO2.

This large population-based study provides strong evidence for a positive association between long-term exposure to PM2.5 and NO2 and an increased risk of severe COVID-19 outcomes. These findings align with some previous research but differ from others in the magnitude of the effect and consistency of associations across various pollutants. This variation may be due to differences in study design (population-based versus COVID-19 cases only), exposure assessment methods, confounder adjustments, and outcome definitions. The study's large sample size and detailed exposure assessment provided greater statistical power than previous studies, potentially accounting for some discrepancies. The stronger effect estimates observed during the first wave compared to the second wave suggest differences in overall susceptibility to severe COVID-19 during the different phases of the pandemic or the influence of contextual factors on healthcare systems. Potential biological mechanisms for this association include air pollution's effect on lung defenses, exacerbation of underlying conditions, and influence on SARS-CoV-2 receptor expression. The study's findings have important public health implications, highlighting the need for strategies to reduce air pollution levels to mitigate the risk of severe COVID-19 and improve overall population health.

This large population-based study robustly demonstrates the association between long-term air pollution exposure and increased risk of severe COVID-19. The findings highlight the importance of reducing air pollution levels to improve public health and mitigate the impact of respiratory infections. Future research should focus on elucidating the specific biological mechanisms linking air pollution exposure to COVID-19 severity and investigating the long-term effects of air pollution on various respiratory and non-respiratory diseases. Further research should also explore the impact of different air pollution interventions and policies on COVID-19 outcomes. More detailed investigation of the effect modification by COVID-19 wave and potential causal mediation analyses are needed to gain a deeper understanding of the complex relationships at play.

While this study is comprehensive and provides robust evidence, certain limitations should be acknowledged. Data on race/ethnicity, migration status, physical activity, and occupation were unavailable, potentially leading to residual confounding. The outcome definition, using a 30-day window after COVID-19 diagnosis, might have included some unrelated hospitalizations. The majority of hospitalizations were related to COVID-19, but not all. Although sensitivity analyses addressed some of these limitations, residual bias might persist. Finally, the study focuses on the first year of the pandemic, which might not completely reflect the effects of air pollution during subsequent periods with vaccinations and new variants. The study's population is from Catalonia which limits the generalizability of the findings.