In a recent study posted to the medRxiv* preprint server, researchers assessed the association between particulate matter ≤2.5mm (PM2.5) exposure and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and severity outcomes such as hospitalizations and deaths.
Studies have reported a positive association between the severity of air pollution due to PM2.5 and coronavirus disease 2019 (COVID-19). However, the study findings were based on geographical unit comparisons that did not consider differences at the individual level and, therefore, often misclassified PM2.5 exposures with inaccurate estimates.
Consequently, the reported associations between ambient PM2.5 air pollution and SARS-CoV-2 infections could have been spurious and confounded by differences in factors such as socioeconomic status that can influence air pollution exposure and the risk of SARS-CoV-2 infection and severity outcomes.
About the study
In the present systematic review, researchers reviewed existing literature on the association between PM2.5 air pollutant exposure and the risk of SARS-CoV-2 infection and severity outcomes.
The study comprised original research studies with cohort and case-control study designs that were peer-reviewed, published in English, used individual-level information, and were obtained from databases such as Embase, WHO (World Health Organization) COVID-19, and Medline until June 30, 2022.
Additionally, previously published systematic reviews and reference lists of grey literature with topic similarity were searched for eligible studies.
Studies were excluded if they utilized case-series, cross-sectional, ecological, in vitro, or animal-based study designs. Review, hypothesis, commentary, opinion articles, editorials, preprints, or conference presentations were excluded from the analysis. Studies that did not utilize PM2.5 as the study exposure or only investigated tobacco-generated smoke as an air pollutant or air pollution only in indoor settings were excluded. The quality of the included studies was assessed based on NOS (Newcastle-Ottawa scale) scores transformed to AHRQ (agency for health research and quality) standards.
Random effects modeling was used for the meta-analysis of the pooled data with adjustments to reduce publication bias. Two reviewers independently screened the studies, and another reviewer solved disputes. Data on sample population characteristics, PM2.5 pollutant measurement operationalization, and SARS-CoV-2 infection outcomes were extracted, and odds ratios (OR) and relative risks (RR) were calculated. Only studies of good or fair quality were included in the final analysis. However, sensitivity analysis conducted to evaluate the robustness of the study findings included studies irrespective of their quality.
Initially, 1,442 studies were identified, of which 509 studies were excluded due to non-case-control or non-cohort study design, 323 studies were excluded due to non-PM2.5 exposure or outcomes other than SARS-CoV-2 infections, and 257 records were excluded since they were editorials or commentaries.
As a result, only 18 studies with cohort designs were considered for the final analysis. The included studies used data from North America (especially the United States, Canada, and Mexico), Europe (the United Kingdom, Italy, Spain, and Poland), and China.
The team found that a 10.0 µg/m3 in PM2.5 exposure increased the odds of SARS-CoV-2 infections and COVID-19 severity outcomes by 66% and 127%, respectively. Pooled death reports data, although positive, showed non-significant associations (OR 1.4). Most studies (n=14) were of good quality, although several issues related to study methodology were present; only a few studies utilized individual-level information for confounding variables adjustments such as socioeconomic status (n=4) and region-based indicators (n=12).
All except one study assessing COVID-19 severity and mortality were restricted to cohorts diagnosed with SARS-CoV-2 infections, while three were limited to patients hospitalized due to COVID-19. Most studies that assessed COVID-19 severity outcomes (n=9) and associated deaths (n=5) analyzed individuals who were already diagnosed with COVID-19 and, therefore, had the possibility of collider bias introduction.
In addition, three studies only included individuals with a COVID-19 test report, whereas the other four infection studies either used cohorts irrespective of SARS-CoV-2 testing reports or analyzed the entire study cohort. Further, three studies used multiple predictors within one model. The period of measuring PM2.5 exposure considerably varied, from just seven days before inclusion or recruitment to almost 20 years.
United Kingdom-based studies used only United Kingdom Biobank data. Furthermore, three studies utilized PM2.5 levels measured in 2010, and all studies used participant residential addresses documented between 2006 and 2010.
The model considering only one pollutant and accounting for socioeconomic status found that SARS-CoV-2 infections increased by four percent for every 1.0µg/m3 in PM2.5 exposure. The one-pollutant model omitting socioeconomic status showed a comparable but slightly bigger impact (RR: 1.1), which considerably increased when the model included other substances that pollute the air (NO: 1. the 1, O3: 1.1, NO2: 1.4).
The findings showed multi-collinearity among air-polluting substances, which could bias the link between PM2.5 exposure and SARS-CoV-2 infections. Sensitivity analysis marginally attenuated the study results (OR: 2.0) and the association between PM2.5 exposure and SARS-CoV-2 infections remained significant. The studies showed no evidence of publication bias.
Overall, the study findings showed strong evidence indicating that ambient PM2.5 increases the risk of SARS-CoV-2 infections and weaker evidence associating PM2.5 with an increase in the severity outcomes of COVID-19.
medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.