Severe Acute Respiratory Syndrome and Particulate Matter Exposure: A Systematic Review

Background: Particulate matter (PM) exposure is responsible for seven million deaths annually and has been implicated in the pathogenesis of respiratory infections such as severe acute respiratory syndrome (SARS). Understanding modifiable risk factors of high mortality, resource burdensome C19 and exposure risks such as PM is key to mitigating their devastating effects. This systematic review focuses on the literature available, identifying the spatial and temporal variation in the role of quantified PM exposure in SARS disease outcome and planning our future experimental studies. Methods: The systematic review utilized keywords adhered to the PRISMA guidelines. We included original human research studies in English. Results: Initial search yielded N = 906, application of eligibility criteria yielded N = 46. Upon analysis of risk of bias N = 41 demonstrated high risk. Studies found a positive association between elevated PM2.5, PM10 and SARS-related outcomes. A geographic and temporal variation in both PM and C19’s role was observed. Conclusion: C19 is a high mortality and resource intensive disease which devastated the globe. PM exposure is also a global health crisis. Our systematic review focuses on the intersection of this impactful disease-exposure dyad and understanding the role of PM is important in the development of interventions to prevent future spread of viral infections.


Introduction
Coronaviruses (CoV) are a common cause of respiratory disease. However, at least two novel CoVs have plagued humanity [1,2]. In 2003, the severe acute respiratory syndrome-CoV-1 (SARS-CoV-1) virus caused SARS, which affected over 8000 people worldwide and caused the death of over 700. In 2019, the latest novel CoV was identified in Wuhan, China, and was named SARS-CoV-2 [1]. By early 2020 the spread of SARS-CoV-2 was declared a pandemic [3]. Coronavirus disease 2019 (COVID-19; C19) was the official name given by the World Health Organization (WHO) to the disease caused by SARS-CoV-2 [3]. In addition, to the clinical signs and symptoms of cough and fever, radiographic findings in severe cases include lung infiltrates that require hospitalization. The COVID-19 pandemic is the third leading cause of death since 2020, and continues to threaten the health and well-being of humanity [4]. Therefore, it is imperative that we further evaluate exacerbating factors such as particulate matter (PM) that may allow us to mitigate morbidity and mortality.
Elevated PM exposure is associated with cancer, obstructive airway disease, ischemic heart disease, stroke, and respiratory infections resulting in 7-million deaths annually [5][6][7]. PM-induced pulmonary inflammation causes acute exacerbation of cardiovascular disease Figure 1. Study Design. Per Preferred reporting Items for systematic reviews and Meta-Analyses (PRISMA) Guidelines. PRISMA is an evidence-based minimum set of items for reporting in systematic reviews and meta-analyses.

Search Terms
A PUBMED Medical Subject Headings (MeSH) search was performed and the following entry terms were identified: ( We then searched for articles that addressed how quantifiable particulate matter exposure is associated with the risk, severity and mortality due to SARS infection.
For the purposes of this review we define PM as a mixture of solid particles and liquid droplets found in the air [27]. Severe acute respiratory syndrome is a viral respiratory illness caused by coronaviruses first detected in 2003. This review focuses on both SARS-CoV-1 and SARS-CoV-2.
Articles were selected based on the following inclusion criteria: (1) adult population; (2) articles written in English; (3) articles should include the concentration of the PM exposure in association with incidence, prevalence, severity and mortality due to SARS (SARS-CoV-1 and SARS-CoV-2); (4) studies after November 2002.
*Consider, if feasible to do so, reporting the number of records identified from each database or register searched (rather than the total number across all databases/registers). **If automation tools were used, indicate how many records were excluded by a human and how many were excluded by automation tools.

PRISMA 2020 flow diagram for new systematic reviews which included searches of databases, registers and other sources
Commented [M1]: Please confirm if this r in figure 1 is the ref. [26] in the reference s Commented [PS2R1]: Yes, this is correct Figure 1. Study Design. Per Preferred reporting Items for systematic reviews and Meta-Analyses (PRISMA) Guidelines. PRISMA is an evidence-based minimum set of items for reporting in systematic reviews and meta-analyses [26]. http://www.prisma-statement.org/, accessed on 7 January 2023.
Articles were selected based on the following inclusion criteria: (1) adult population; (2) articles written in English; (3) articles should include the concentration of the PM exposure in association with incidence, prevalence, severity and mortality due to SARS (SARS-CoV-1 and SARS-CoV-2); (4) studies after November 2002.
Articles were excluded if they: (1) were not in English language; (2) did not quantify the concentration of PM exposure; (3) involved any non-human subjects/in vitro work/cell studies/immunohistochemistry; (4) were conducted on pediatric population; (5) focused on gaseous pollutants; or (6) were not original research. Two independent researchers conducted the literature search and determined studies that met the inclusion/exclusion criteria. A third investigator resolved disagreements.

Quality Assessment and Risk of Bias (RoB)
The overall RoB of the Cohort studies included in this review was determined with the approach described by Lee et al., 2020 (Figure 2A,B) [28]. We assessed three key domains of interest in the studies:

Assessment of Outcomes
Studies that performed Nucleic Acid Amplification Test (NAAT) using reversetranscription polymerase chain reaction (RT-PCR) to detect SARS-CoV-2 RNA from the upper respiratory tract, physician diagnosis or other clinical tests, were categorized as low risk for detection bias. For studies with unknown methods of diagnosis, we categorized them as unclear risk of detection bias.

Adjustment for Confounding
Studies that adjusted for age, gender, individual levels of exposure or any other relevant covariates were categorized as low risk for this domain. Studies that did not adjust results for at least one covariate were categorized as high risk.

Control/Dose-Response Comparator Was Used for Comparative Analysis
Studies that included a control group were categorized as low risk for this domain, whereas those that did not were categorized as high risk. The three key domains were assessed for overall risk of bias judgment. Studies were categorized as low overall risk of bias if it was at low risk for all key domains, and high if any of the domains were high. For the time series studies only two domains, i.e., assessment of outcomes and adjustment for confounding were considered to analyze the risk of bias.

Data Management/Extraction
Based on the inclusion and exclusion criteria, we screened and selected manuscripts (EndNote™ 20.1). Each article was screened for study design, patient characteristics, sample size, tools used, incidence, severity and mortality of SARS in association with quantifiable PM exposure. Results from each database search were filtered for human subjects, English language, publication date (after November 2002) and imported into EndNote. The references were then screened for duplicates. Only original research papers were then reviewed for title, abstract and full text to ascertain eligibility. The references cited in the relevant articles were also examined. All results were screened by SP and MSF and further independently evaluated by AN. Disagreements were resolved by consensus (see Supplementary Tables S1-S3).
2.4. Data Synthesis (GraphPad Prism 9; Ver 9.2.0) Data was generated from sources using our review PEO question and summarized into tables and plots ( Figure 3). Qualitative data synthesis was performed for studies, using thematic analysis that included three stages: (i). identifying information about the selected studies' methodology and findings; (ii). organizing them into subheadings and descriptive categories; and (iii). developing these categories into analytic themes [29].

Literature Search
A total of 906 studies (334 PubMed and 572 Embase) were identified after filtering for relevant studies (Figure 1, Supplementary Tables S1 and S2). After removing duplicates, N = 732 were assessed for inclusion (abstract and title review). Finally, 46 original research articles were considered eligible [25,  Articles were excluded if they: (1) were not in English language; (2) did not quantify the concentration of PM exposure; (3) involved any non-human subjects/in vitro work/cell studies/immunohistochemistry; (4) were conducted on pediatric population; (5) focused on gaseous pollutants; or (6) were not original research. Two independent researchers conducted the literature search and determined studies that met the inclusion/exclusion criteria. A third investigator resolved disagreements.

Quality Assessment and Risk of Bias (RoB)
The overall RoB of the Cohort studies included in this review was determined with the approach described by Lee et al., 2020 (Figure 2A, 2B) [28]. We assessed three key domains of interest in the studies: Risk of bias assessment. Time series studies, which were evaluated in two domains, i.e., outcome assessment, risk of confounding. Studies were color coded red or green for high vs. low risk of bias. Studies were categorized as low overall risk of bias if they were at low risk (green) for all key domains and high if any of the domains were high (red). RoB of * Meo [59]; † Meo [60]; ‡ Meo [61]; § Meo [68]; ‖ Meo [45]; ¶ Meo [62].   (B) Risk of bias assessment. Time series studies, which were evaluated in two domains, i.e., outcome assessment, risk of confounding. Studies were color coded red or green for high vs. low risk of bias. Studies were categorized as low overall risk of bias if they were at low risk (green) for all key domains and high if any of the domains were high (red). RoB of * Meo [59]; † Meo [60]; ‡ Meo [61]; § Meo [68]; Meo [45]; ¶ Meo [62].

Error in Figure
In the original publication, there was an omission in Figure 3 as published. Figu are missing coordinates for correlation coefficients and odds ratios. The corrected appears below. The authors state that the scientific conclusions are unaffecte correction was approved by the Academic Editor. The original publication has a updated [1].    Risk of bias assessment was performed for outcome, confounders and control group assessment. Of the three domains assessed for cohort studies, N = 2 studies were high risk for outcome assessment, N = 24 were high risk due to lack of adjustment for confounders and N = 39 were high risk due to lack of a control group in their studies. Overall, N = 3 studies had low risk of bias, whereas N = 40 had high risk of bias For the time series studies of the two domains assessed (outcome and confounders), N = 3 were low risk for outcome assessment and N= 1 was considered high risk for due to lack of adjustment for confounders. Overall, N = 2 were low risk for bias and N = 1 had high risk of bias.

Study Characteristics
As the C19 pandemic swept the globe from 31 December 2019, understanding the phenotype of both the disease and associated risk factors of disease severity has been challenging. Cohort studies were the predominant type (N = 43), while N = 3 were timeseries studies [31,33,43]. The association of PM in the context of C19 surges, geographic location, and type of SARS infection are also of great interest and were further examined. In the context of these categories we will also discuss how PM 2.5 and PM 10 have played a role in SARS severity and spread.

Coronaviruses Have Been the Cause of Several Outbreaks
SARS-1 originated in Guangdong, China in 2003, and in six months had spread to more than two dozen countries resulting in at least 774 deaths [75]. Due to limited transmission, there are few studies that focus on this pathogen. Only one study that analyzed and noted positive association between PM and SARS-1 infection was noted by Kan et al., who found that for every 10 µg/m 3 increase in PM 10 the Relative risk (RR) of daily SARS mortality was 1.06 (1.00-1.12) [30]. There were few variants or recurrence of SARS-1 [76]. In contrast, the SARS-CoV-2 virus has several variants and lineages, and been responsible for at least 6 million deaths worldwide [77,78].

Temporal Relationship of PM and SARS
A decline in the incidence, mortality, and hospitalization was observed during the later pandemic period, from approximately late April-June 2020. A temporal analysis in Beijing from 25 April-31 May 2020 showed a declining trend in daily mortality count [30]. While this could be attributed to the implementation of more stringent mitigation measures, there are several other factors that may be relevant [79]. To understand the role of PM in the temporal variegation of outcomes, investigators have examined the impact of PM during the early and later phase of the pandemic. Dragone et al. noted that PM levels exceeded the daily limit during two early pandemic periods (16)(17)(18)(19)(20)(21)(22)(23)(24)(25) February and 17-20 March 2020) in Italy. During this period, areas with the highest levels of ambient PM also had the highest number of infected populations [40]. Similarly, Li et al. noted a positive association between C19 cases and PM 2.5 through a risks study using days with the highest and lowest incidence numbers in February [31]. Analysis investigating C19 cases in January-April 2020 in India showed that a 10 µg/m 3 increase in PM 2.5 and PM 10 resulted in 2.21% (95% CI:1.13-3.29), 2.67% (95% CI: 0.33-5.01), increase in daily counts of C19 infected cases, respectively [33]. The early pandemic was the focus of 14/46 studies. PM was positively associated with C19 for a number of studies in the following aspects: incidence (N = 6); prevalence (N = 2); morbidity (N = 1) and mortality (N = 6). Negative association was observed with mortality (N = 1) and incidence (N = 1), and equivocal results were reported by N = 1 [34,52,63]. Of the 32 studies from the later pandemic period, PM was positively associated with C19 based on: incidence (N = 17); prevalence (N = 4); morbidity (N = 6) and mortality (N = 18). Negative association with incidence was observed in N = 2. Equivocal results reported by N = 4 [72].

Understanding Geographic Epidemiology Based on Region-Based Outcomes
Few, if any, areas of the globe have been left unaffected by the C19 pandemic. Meo et al. studied 17 countries across the globe and noted a significant positive association between PM and C19 incidence [45]. Certain areas like Malawi and Indonesia have been disproportionately impacted, and reported the highest case fatality rates on 8/26/22 [80].
Six studies with a focus on the Middle East were identified [66][67][68][69][70][71]. In a study conducted in Riyadh, Jeddah and Makkah, PM 10 positively correlated to daily cases of C19 (Pearson correlation coefficients were 0.68, 0.54, 0.38, respectively) [66]. Similar observations were made in three Iranian cities where exposure to PM 2.5 for several days showed significant association to confirmed cases [67]. Increase in PM 2.5 due to a sandstorm in Saudi Arabia was associated with a significant increase in the number of SARS-CoV-2 cases (Spearman's correlation coefficient ρ = 0.944 (<0.0001)) [68].
In the U.S, data from seven New York City (NYC) hospitals concluded that higher and long-term exposure to PM 2.5 was associated with an increased risk of mortality (RR 1.11, 95%CI: 1.02-1.21) and ICU admission (RR 1.13, 95%CI: 1.00-1.28) per 1-µg/m 3 increase in PM 2.5 [55]. Similarly, a study from five regions noted that the number of cases significantly augmented with a rise in the levels of PM 2.5 (ρ = 0.176, p < 0.001). PM was positively associated with C19: incidence rate (N = 7); mortality (N = 4) and prevalence (N = 1). A negative association with mortality was observed by N = 2 [63,64].
Similarly, as in 2/3 Latin American countries, PM and C19 incidence and mortality had a positive association. In contrast, one study reported equivocal results [72]. Specifically, an increase of 1 µg/m 3 in PM 2.5 increased the mortality risk by approximately 7.4% in Mexico City metropolitan area in October 2020 [73]. PM is a heterogeneous mixture of solid particles and liquid droplets found in the air. It is commonly grouped by diameter into fine PM 2.5 (<2.5 mm) and coarse PM 10 (<10 mm). PM 2.5 is more likely to travel and deposit deeper in the lungs like the alveolus, whereas PM 10 can deposit on the surfaces of larger airways inducing inflammation. Ambient air pollutants are risk factors for cardiopulmonary diseases and responsible for over 6 million annual deaths.
PM 10 was evaluated in eight studies, and it was positively associated with C19 incidence (N = 5); prevalence (N = 1); and mortality (N = 1). Finally, PM 10 and PM 2.5 were evaluated by N = 19 studies. A positive association with C19 incidence (N = 9); mortality (N = 5) was seen. Negative association was seen in N = 2 and equivocal results were identified in N = 1.

Discussion
Our systematic review identified the role of PM to be important in the incidence, mortality and morbidity due to SARS infection. These studies had significant differences in the populations, methods, and outcomes that were studied (Table 1). We identified three themes, temporality, PM size/dose, and spatial, which define the relationship of C19 with PM exposure. Evaluating the heterogeneous characteristics of the disease across different territories and phases of the pandemic is important to implement measures to contain spread.
Longer durations and higher levels of PM increased the risk of ICU admissions and deaths due to C19. Several mechanisms have been hypothesized. Oxidative stress and Life 2023, 13, 538 15 of 21 disruptive immune and/or neuroendocrine function can result in increased severity of viral pulmonary infections [81,82].
PM has been associated with enhanced infection with RNA viruses such as SARS [83]. PM concentration and virus dissemination were positively correlated in the spread of measles in several studies [12]. A 10 µg/m 3 increase in PM 2.5 was associated with increased measles incidence. A similar observation was also noted with respiratory syncytial virus, which causes bronchiolitis and pneumonia [15]. A study in Kuala Lumpur collected PM 2.5 in four study sites and found the highest levels of SARS-CoV-2 RNA on PM 2.5 [84]. PM exposure in murine models was associated with upregulated Angiotensin converting enzyme 2 (ACE2) and Transmembrane protease serine 2 (TMPRSS2), receptors required for SARS entry into host cells [85,86]. Exposure to PM induced Renin Angiotensin-aldosterone (RAAS) and Kallikrein-kinin systems (KKS), involved in cardiovascular and lung diseases. PM-induced damage to lung cells increases the inflammatory state which can increase the mortality and severity of C19 [87,88]. Therefore, it may be important to implement measures to reduce PM emissions in the atmosphere. Studies in this review further highlighted the importance of measures such as lockdown and movement restrictions, public awareness regarding pollution via media tools and professional programs and strengthening rural infrastructure that may limit the infectivity of SARS [69,72].

Geographic Epidemiology
Variation in C19 outcomes in different regions could be attributed to social determinants of health such as poverty, access to health care facilities and health literacy. Populations with limited resources also have a high prevalence of chronic health conditions [89,90]. Urban areas with industries had elevated levels of PM 2.5 . Spatial variation in the concentration of PM 2.5 in some areas such as California's central valley and Italy's Po Valley can be contributed to geographical features with climate inversion events that trap these pollutants. The air trapping in these regions can also lead to chronic exposure to these particulates increasing the risk for respiratory and cardiovascular diseases which further enhances the risk [91].

Temporal Association
Worsening asthma and COPD leading to hospitalization has been noted with shortterm (up to 24 h) exposures to PM 10 [92]. Long-term PM 2.5 is known to increase the risk of COPD, a known risk factor for severe C19 infection [93].
An NYC hospital-based study noted that mortality rates dropped from 25.6% in March to 7.6% in August 2020 [94]. This reduction in the number of severe C19 cases that was observed during the later pandemic phase could be attributed to multiple reasons, including the development of immunity due to availability of vaccinations, and treatment modalities including corticosteroids, targeted antiviral therapy, and anti-cytokine treatments. The quarantine restrictions and mask policies enforced by many countries also could have reduced exposure to ambient PM [95][96][97].
Strengths of Systematic Review. This review focuses on the environmental inducers of infectious diseases, a global health issue. It incorporates the variation in PM and the risk of Coronaviruses geographically. Manuscripts from across the globe were reviewed, which made this study more generalizable. Each article was screened for study design and was further subcategorized to understand temporal, spatial variability.
Our systematic review has several limitations. Since assessing the risk of bias inherently has some level of subjectivity, we categorized high-vs. low-risk studies using a determined set of criteria. While many of the studies that we assessed had a high risk of bias, future studies would benefit from assessing their data in the context of potential confounders including age, gender, and comorbidities. Several variants have been identified since the 2019; however, our manuscript does not discuss the disparity in disease outcomes for different variants [77]. The manuscripts that have been included in our review have not determined the variants that were present in their communities during their data acquisition [30,72,73]. There is also a variation in the PM concentration across countries and regions that could add to the bias. However, only N = 6 studies analyzed the spatial variation [32,38,40,55,62,65]. Also, additional studies not identified in the two large databases could have caused selection bias. There is a limitation in data available on SARS-1, which could be due to selection of manuscripts in the English language. Finally, while meta-analysis in the context of a systematic review may provide a more accurate effects estimate, for this to occur we would require source data availability and methodologic similarity. We therefore reviewed all 46 studies for available supplemental data and for similar methodologies and outcomes. Studies were grouped according to the statistical outcomes measured, i.e., relative risk vs. odd's ratio. Seven out of 46 evaluated relative risks. Four out of these seven had supplementary data available. Out of the three studies that evaluated the odds ratio, one had supplementary data available. Unfortunately, only two studies performed the Generalized additive models (GAM) to analyze the association between PM and C19 outcomes; however, the C19 outcomes were different (Incidence vs. morbidity), which limits our ability to perform meta-analysis.

Conclusions
In conclusion, these studies have expanded our knowledge of PM exposure and its association with SARS infection. The review highlights the clinical impact of PM and the need for implementing measures to combat climate change and dangerous levels of environmental toxin. There was a spatial and temporal variation in the characteristics of the disease. Overall, it was seen that exposure to quantified PM was associated with increased incidence, mortality and morbidity to C19. Measures can be taken on both global and a personal level, such as improving air ventilation design and systems in enclosed spaces and buildings, restricting wildlife trading and deforestation, and training our healthcare professionals to educate masses on taking personal steps to ensure less production and exposure to pollution, such as using facemasks, and walking/cycling instead of motorized transport.

Future Plans
Future experimental studies will include developing our understanding of the role of PM in accentuating the response to pathogens such as C19, understanding the effects estimate for chronic vs. short-term exposure to PM, and in furthering our management of PM exposure to limit severity of viral infections. These projects will focus on quantifying the association of PM concentrations (by zip code and/or geocoding) and the incidence of C19 related morbidity and mortality.