key: cord-0755462-t6uxb1er
authors: Stieb, David M.; Evans, Greg J.; To, Teresa M.; Brook, Jeffrey R.; Burnett, Richard T.
title: Response to Goldberg and Villeneuve re: An ecological analysis of long-term exposure to PM(2.5) and incidence of COVID-19 in Canadian health regions
date: 2020-12-25
journal: Environ Res
DOI: 10.1016/j.envres.2020.110623
sha: 9a20efe91f4931e404fc1ce479d7bda6725934f1
doc_id: 755462
cord_uid: t6uxb1er

nan

Response to Goldberg and Villeneuve re: An ecological analysis of long-term exposure to PM 2.5 and incidence of COVID-19 in Canadian health regions Dear Editor.

We would like to thank our colleagues, Drs. Goldberg and Villeneuve, for their interest in our recent paper (Stieb et al., 2020) . We agree with the methodological shortcomings affecting ecological studies of air pollution and COVID-19 that they have enumerated in their recent review, as well as their concerns about the release of findings from these analyses prior to peer review (Villeneuve and Goldberg, 2020a) . In fact, we were aware of their views based on their earlier critique (Villeneuve and Goldberg, 2020b) and attempted to address them as much as possible, strengthening our analysis.

While some of the views expressed in their letter pertain directly to our study, and we address these below, much of the letter reiterates what they have already articulated in their published critiques, or is not applicable. We make no inferences in our paper, for example, about the role of air pollution in explaining differences in the COVID-19 experiences of South Korea or New Zealand. Moreover, unlike several of the papers cited in their review, our paper was subjected to this journal's standard peer review and editorial procedures prior to its public release. To our knowledge the review was not expedited nor did it circumvent standards of peer review. We have also been fully transparent in our paper about the limitations of ecological studies of air pollution and COVID-19, cautioning that our results "should be interpreted with caution and require further examination using study designs less prone to bias". In particular, we acknowledged in our paper that COVID-19 incidence, PM 2.5 concentrations and the prevalence of other risk factors can vary over a smaller scale than health region. While some of the health regions which served as the unit of analysis in our study were very large, observed associations were in fact larger in magnitude in sensitivity analyses excluding provinces that reported cases only for aggregated health regions, excluding health regions with less than median population density, and restricted to Quebec and Ontario, the two most populous provinces which also experienced the highest incidence of COVID-19.

Drs. Goldberg and Villeneuve point to evidence that there is a greater likelihood of identification of COVID-19 cases in densely populated urban areas, which also experience higher air pollution exposures, potentially leading to spurious associations between air pollution and COVID-19. We addressed this issue in our analysis in several ways: our results were not sensitive to inclusion of population density as a covariate; we observed a larger magnitude association in a sensitivity analysis restricted to health regions with greater than the median population density; and we observed that deaths were very strongly correlated with case counts by health region (R 2 = 0.95), suggesting that differences in case numbers by health region were not driven by factors such as greater testing capacity or public awareness in dense urban areas, which would be expected to introduce more scatter in the relationship between deaths and cases.

More generally, we disagree with the conclusion presented in Drs. Goldberg's and Villeneuve's review that "we cannot envision any additional value that these studies provide that further public health" (Villeneuve and Goldberg, 2020a) . We take the view that identification of putative risk factors has value, at the very least in generating hypotheses for further evaluation in subsequent studies. Our analysis was intended to serve as a preliminary examination of whether reported ecological associations between air pollution and COVID-19 in the United States (Liang et al., 2020; Wu et al., 2020) could be detected in Canada, given the lower COVID-19 incidence, narrower range of PM 2.5 exposure and access to universal health care in Canada. These types of studies can be readily replicated in diverse settings to further examine consistency of associations and the direction and magnitude of potential biases. While we believe that parallels in the observed associations are noteworthy as preliminary evidence, we did acknowledge in our paper that the similarity in findings could also be attributable to biases affecting all studies employing the same design. We note that our view on the potential value of these early analyses is shared by our colleague, Dr. Francesca Dominici of Harvard University, who has co-authored a similar area level analyses in the United States (Wu et al., 2020) . In response to Drs. Goldberg's and Villeneuve's review, Dr. Dominici countered, "Are these studies valuable to pave the way for future research on this topic? Yes." (Nicole, 2020) .

Putting this work in historical perspective, early studies examining the association between air pollution and health employed relatively crude data and methods. These were by no means definitive but served to spawn interest and led to successive generations of research using more refined methods such that there is now a vast evidence base from multiple lines of research persuasively linking air pollution to adverse health effects even at low levels of exposure. Drs. Goldberg and Villeneuve hypothesize that, "The possible contribution of air pollution on increasing incidence of COVID-19, put in this context, is at best a drop in a very large bucket." Nowhere in our paper do we suggest that air pollution is likely to predominate over other factors mediating the spread of COVID-19. However, even based on evidence available prior to the COVID-19 pandemic, air pollution is likely to incrementally increase the risk of COVID-19, as it does other respiratory infections (Ciencewicki and Jaspers, 2007; Domingo and Rovira, 2020) . Finally, we take particular issue with Drs. Goldberg's and Villeneuve's assertion that somehow, "Studies of air pollution and COVID-19 may detract from the needed implementation of public health measures needed to control the spread of COVID-19". This assertion is at best speculative; they have provided no evidence to support this contention, nor even a plausible explanation for how this could possibly occur.

Sincerely, David M. Stieb, Greg J. Evans, Teresa M. To, Jeffrey R. 

Air pollution and respiratory viral infection

Effects of air pollutants on the transmission and severity of respiratory viral infections

Urban air pollution may enhance COVID-19 casefatality and mortality rates in the United States

Air of uncertainty: can we study pollution and COVID-19 in the midst of a pandemic? Environ. Health Perspect. 128, 114005

An ecological analysis of long-term exposure to PM2.5 and incidence of COVID-19 in Canadian health regions

Methodological considerations for epidemiological studies of air pollution and the SARS and COVID-19 coronavirus outbreaks

Air Pollution, COVID-19 and Death: the Perils of Bypassing Peer Review [WWW Document

Air pollution and COVID-19 mortality in the United States: strengths and limitations of an ecological regression analysis

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.