Air Pollution Exposure as a Relevant Risk Factor for COPD Exacerbations in Male and Female Patients

Chronic obstructive pulmonary disease (COPD) is a multifactorial lung inflammatory disease affecting 174 million people worldwide, with a recently reported increased incidence in female patients. Patients with COPD are especially vulnerable to the detrimental effects of environmental exposures, especially from air particulate and gaseous pollutants. Exposure to air pollution severely influences COPD outcomes, resulting in acute exacerbations, hospitalizations, and death. In the current study, we conducted a review of the literature addressing air pollution induced acute exacerbations of COPD (AECOPD) in order to determine whether air pollution affects COPD patients in a sex-specific manner. We found that while the majority of studies enrolled both male and female patients, only a few reported results disaggregated by sex. Most studies had a higher enrollment of male patients, only four compared AECOPD outcomes between sexes, and only one study identified sex differences in AECOPD, with females displaying higher rates. Overall, our analysis of the literature confirmed that air pollution exposure is a trigger for AECOPD hospitalizations and revealed a significant gap in our knowledge of sex-specific effects of air pollutants on COPD outcomes, highlighting the need for more studies considering sex as a biological variable.


COPD Definition
Chronic Obstructive Pulmonary Disease (COPD) is a lung inflammatory disease that includes emphysema and chronic bronchitis and is characterized by airflow blockage in the lungs [1]. The diagnosis of COPD includes spirometry values of less than 70% of predicted forced expiratory volume (FEV) that is incompletely reversible with the administration of an inhaled bronchodilator. Pathological features are observed in central airways, small airways, and alveolar spaces. The pathogenesis of COPD includes proteinaseantiproteinase imbalance, immunological mechanisms, oxidant-antioxidant balance, systemic inflammation, apoptosis, and ineffective repair, and accelerated decline in forced expiratory volume in 1 second (FEV1) and forced vital capacity (FVC) [1]. Airflow limitation in COPD is defined as a post-bronchodilator FEV1 to FVC ratio of 0.70 or lower. The diagnosis of COPD is also determined on the basis of symptoms and signs (e.g., exertional breathlessness, chronic cough, regular sputum production, frequent bronchitis, or wheeze, etc.) in people over 35 years of age who have a risk factor (e.g., smoking history), although these clinical findings have to be supported by spirometry, as defined by GOLD and NICE standards.
The development of COPD is multifactorial, and the risk factors include genetic, environmental, and sex and gender factors. While cigarette smoke is the most critical risk factor associated with COPD, occupational and other environmental exposures are known to cause approximately one in six cases [2]. Female sex and gender have also been charcoal, animal dung, and others used for cooking [27,28]. Due to traditional gender roles, these exposures have significantly contributed to COPD morbidity and mortality in women [29]. It is estimated that 50% of households worldwide (about 3 billion people) are exposed to smoke from biomass fuel combustion. These exposures contribute to about half of the deaths from COPD in developing countries, of which 75% are women [27,28].

Exacerbation triggers
Exacerbations of COPD are episodes of worsening of symptoms, leading to substantial morbidity and mortality [30]. COPD exacerbations are associated with increased airway and systemic inflammation and physiological changes, including hyperinflation. These are triggered mainly by respiratory viruses and bacteria, which infect the lower airway and increase airway inflammation. Some patients are particularly susceptible to exacerbations and show worse health status and faster disease progression than those who have infrequent exacerbations [31].
The mechanisms of COPD exacerbations are complex. While respiratory viruses (in particular rhinoviruses) and bacteria play a major role in their causative etiology [32], in some patients, noninfective environmental factors also contribute to their development. Data recently published from a large observational study identified a phenotype of patients that are more susceptible to frequent exacerbations from environmental exposures [33]. Other quantitative studies indicated that anxiety and depression could lead to a statistically significant increase in the likelihood of COPD patients being hospitalized [34]. Although more than 80% of exacerbations are managed on an outpatient basis, hospitalization is all too common and associated with considerable health care costs and mortality. In this regard, noninvasive ventilation has greatly decreased the mortality in exacerbations that require ventilatory support. However, across the range of exacerbation severity, treatment failure and relapses are frequent [35].
Among individuals with COPD, exposure to outdoor air pollutants is associated with loss of lung function and increased respiratory symptoms, leading to exacerbations and increased mortality [36]. Some studies suggest that temperature may modify the effect of air pollution exposure, although their results are not conclusive [37]. For example, Yan et al. explored the environmental effect of two different geographical places on COPD exacerbations (Beijing in summer, Sanya in winter) and found that poorer air quality index (AQI) and higher temperatures in Beijing were associated with lower FEV1, higher dyspnea, and a twice higher relative risk of exacerbations than in patients in Sanya [38]. The authors also reported that ambient air pollution was strongly associated with COPD exacerbations by triggering apoptosis in airway epithelial cells [38].
Although adequate evidence for a direct relationship between ambient air pollution components and the development of COPD is lacking, higher mortality rates from respiratory and cardiovascular diseases have been reported among patients exposed to air pollution for a very long time [19,39]. Several reports have also pointed out the possibility that acute exacerbations of COPD can be caused by short-term exposures to air pollutants [25,40,41], as well as secondhand tobacco smoke [42].
Regarding sex differences in COPD exacerbations, the available literature indicates that outdoor air pollution affects lung function and triggers exacerbations in both male and female patients, but nonsmoker women may be more affected than men [11,43,44]. This indicates that air pollution may result in differential COPD exacerbation rates and outcomes in men vs. women. Data from multi-center studies have also shown that air pollution concentrations in the ambient are associated with declined lung function and increased risks for hospitalization and mortality in COPD patients. Because sex differences in AECOPD is an understudied area, in the current review, we investigated the association between exposure to gaseous and particulate pollutants and hospitalizations for COPD exacerbations, paying particular attention to differences between males and females. Other systematic reviews and meta-analyses have found that short-term exposures to air pollutants significantly increase the burden of risk of COPD acute exacerbations [4,45]. In the current study, we focus on the association of air pollution exposure and hospitalizations for COPD exacerbations with an emphasis on sex differences. Therefore, we selected studies that included both male and female participants, including those that did or did not analyze outcomes by sex.

Literature Search, Databases and Key Terms Searched
We used PubMed and Google Scholar to search for articles related to our study's focus, using the following search terms: "air pollution", "COPD", "COPD exacerbation", "hospital admission", and "sex". The search was limited to epidemiological studies from 2000 to 2020, although we also included articles prior to 1990 if they contained relevant information. We focused on articles that pooled results on a global scale, reported analytical pooled estimates, were written in English or with an English abstract and studied associations between air pollution and hospitalization for COPD exacerbation as well as respiratory response to shorter-term exposure of air pollution.

Inclusion Criteria
The literature search was limited to human epidemiological studies on (1) Hospitalization due to acute exacerbation of COPD, as identified by the International Statistical Classification of Diseases, 10th Revision (ICD-10) codes J40-J44; (2) A diagnosis of COPD; and presentation for treatment of acute exacerbations of COPD (AECOPD), as defined by increasing shortness of breath, worsening cough, or change in sputum production at presentation; (3) Research-data based; (4) From adult patients (age > 18 years); and (5) published in English language.

Search Process and Study Selection
PubMed and Google Scholar were the main databases utilized. Records were de-duplicated using built-in mechanisms of university library services (Covidence software) and further completed manually. Articles were then screened by their titles and abstracts for inclusion or exclusion. Final selections were determined after full reading of articles.

Data Extraction and Analysis
We extracted information on the association between daily mean concentrations of particulate matter of a diameter of less than 10μm (PM10) or 2.5μm (PM2.5) as well as other gas pollutants (O3, CO, NO2, SO2) with hospital admissions, analyzing the sex variable, based on daily measurements reported in each study or other data that could be aggregated into daily mean values. Thus, results are presented as associations of 24-hour average air pollutant concentrations and daily hospital admissions for AECOPD.

Selected studies
A flow chart of the literature search is shown in Figure 1. The search string returned 8,302 potentially relevant article citations. After systematically reviewing all the abstracts, 7,014 irrelevant studies and 1,083 duplicates were removed. The two authors independently reviewed the remaining 205 full articles for inclusion. After full-text revision, 40 articles were included for systematic analysis and are summarized in Table 1. Combined, these articles reported a total of 2,329,320 hospital admissions for AECOPD, with an average of 58,233 hospitalizations per study ranging from 40 to 578,006 and a standard deviation of 134,419. Hospitalizations for AECOPD in the selected studies spanned four different continents, and the statistics per continent are shown in Figure 2.   Table 1 summarizes the effects of air pollution of exposure in AECOPD in 40 studies. Regardless of geographical location, most studies identified a significant association between particulate pollution exposure and AECOPD. The incremental increases in concentrations of PM2.5 and PM10 were significantly associated with increased risk of hospitalization of AECOPD [40,46], but also stroke and myocardial infarction. However, the adverse influences of PM2.5 on these diseases were generally more robust than those of PM10 [47]. In the US mid-Atlantic states, PM2.5 exposure was associated with all COPD hospital admissions, with a relative risk increase of 1.83 for every 10μg/m 3 increase in PM2.5 [46]. In Central and Eastern Europe, increases in hospital admissions were reported as 3.3% and 2.8% for PM10 and PM2.5, respectively [48].

Individual and combined air pollutant concentrations and their association to daily hospital admissions for AECOPD
When assessing the effects of gaseous air pollutants on AECOPD, it was found that SO2 increases of 10μg/m 3 were related to a 6% increase in hospital admissions for chronic bronchitis, with a two-day lag [49]. Comparably, an independent air pollution modeling study found that when modeled jointly with other pollutants, only SO2 remained significantly associated with AECOPD (hazard ratio 1.038), although the five pollutants assessed in this study were highly correlated (r = 0.89) [48]. In addition, short-term exposures to SO2 were associated with an increase in COPD exacerbation risk, with an odds ratio (OR) of 2.45 per 1 ppb increase in SO2 levels, after adjustment for PM2.5 in a region with a relatively low AQI (central Massachusetts, USA) [41]. Regarding NO2 and CO, both were significantly associated with AECOPD hospitalizations [50]. Tellingly, the magnitude of effects was expanded slightly with increasing days of exposure, with a relative risk of 1.11 and 1.08 for NO2 and CO, respectively, for a 7-day exposure average [50]. Likewise, a study in South Korea found that each 10μg/m 3 increase in CO was associated with a 2% increase in the odds of admission for AECOPD [51].
Finally, other environmental factors have been found to contribute to AECOPD in the studies analyzed. For example, the COPD-related emergency room admissions for all age groups were significantly associated with previous-day BS levels, and lag 0-2 (1.60% and 2.26% increase per 10μg/m 3 , respectively) in a study conducted in Serbia [55]. Similarly, a study in Guangzhou, China, found that haze (at lag1) and air pollution (NO2 at lag 5 and SO2 at lag 3 combined presented more drastic effects on patients aged 19-64, especially in females [56]. Increases in NO2 were associated with the highest risk of hospital admissions for total and respiratory diseases in both single-and multi-pollutant models, and a relative risk of 1.94 in ER at lag 0 for COPD patients [56]. Relative risks at lag0 ranged from 1·018 to 1·036 for each interquartile range increase in air pollution concentration. These increased risks became non-significant by lag4 [56]. The concentrations of 6 monitored pollutants and AECOPD hospitalizations showed statistically significant spatial clustering. After adjusting for potential confounders, residential SO2, NO2 and O3 concentrations were significantly associated with increased AECOPD hospitalizations. Ambient air pollution was spatially correlated with AECOPD hospitalizations. Air pollution increased the rate of hospitalization for AECOPD. The risk of hospitalization for AECOPD in the age ≥65 group was greater than age < 65 group for all day lags. The risk of male and female hospitalizations for AECOPD after lag3-lag5 was higher than that after lag0-lag2, and the strongest risk of hospitalizations for both was with lag3. The incremental increased concentrations of PM2.5 and PM10 were significantly associated with increased risk of hospitalization of AECOPD, stroke, and MI, and the adverse influences of PM2.5 on these diseases were generally stronger than that of PM10 in Jinan, China. In this large multicenter analysis, daily average concentrations of CO and NO2 exhibited the most consistent associations with ED visits for cardiac conditions, while O3 exhibited the most consistent associations with visits for respiratory conditions.

Influence of sex and age variables in the effects of short-term exposure to air pollution on AECOPD
Of the forty studies identified in this review, twenty-one reported the sex of the study participants, including one study enrolling only female patients [48], and one including all male patients [79]. In addition, seven studies reported AECOPD results disaggregated by sex [47,53,58,59,63,67,68], even though only four of these included the total number of male and female patients enrolled [47,53,58,68]. Overall, all studies found that there were significant associations between exposure to air pollutants and hospital admissions due to AECOPD.
A total of 426,630 hospital admissions for COPD were recorded in all 7 studies combined [47,53,58,59,63,67,68]. On average, there were approximately 409 admission counts per day, with males accounting for 72% (296 admissions) and females for 28% (113 admissions). After adjusting for potential confounders, SO 2 , NO 2 , and O 3 concentrations were significantly associated with increases in AECOPD hospitalizations in both sexes. Additionally, the relative risks (95%CIs) of AECOPD hospitalization in association with an inter-quartile range increase in air pollutants for 10 mg/m 3 increases in PM 10 , SO 2 and NO 2 , respectively were analyzed in single model in two studies [58,67]. In these, it was found that the relative risks of exposure to these pollutants were lower for males than for females, except for PM10 exposure. Table 2 summarizes the descriptive statistics on the average AECOPD daily hospitalizations and the daily levels of the six environmental risk factors from the only 7 studies identified that compared male and female patient outcomes [47,53,58,59,63,67,68]. Six of these studies were conducted in China, and one in Brazil. Overall, all studies identified more male than female patients with AECOPD (42.3 males vs. 16.1 females on average) in the total population analyzed, although all studies also enrolled more male patients than female patients (Table 2). In addition, while reporting results of AECOPD cases by sex, 3 of these studies failed to report the total number of male and female total patients enrolled [59,63,67].
In the only 4 studies reporting the number of male and female patients enrolled [47,53,58,68], the percentage of patients that developed AECOPD was similar for both sexes in all but one study, where the hospitalizations for female patients were twice as high as those for males (0.39% vs. 0.18%, respectively, Table 2) [47]. Interestingly, this study reported some of the higher concentration averages for PM2.5, PM10, and SO2 (60, 102, and 52 μg/m 3 , respectively), as well as maximum values, when compared to the rest of the studies that also reported sex-disaggregated data ( Table 2).   Abbreviations: AECOPD: acute exacerbation chronic obstructive pulmonary disease; PM10-particulate matter of less than 10 microns in aerodynamic diameter; PM2.5-particulate matter less than 2.5 microns in aerodynamic diameter; Min: minimum; Max: maximum; NR: not reported; N/A: data not available.
Regarding age, most studies enrolled patients over 18 years of age, except one study that enrolled patients over 40 [53] and two studies enrolling patients over 65 [47,68]. Combined, these studies revealed that the relative risk for AECOPD for patients aged ≤65 years is lower than that of patients aged ≥65 years (Table 2). In addition, Tao et al. reported that the relative risk for COPD exacerbations was higher in elder females than males with increases in PM 10 , NO 2 , and SO2 concentrations at lag 1-4 [58]. This concurs with results from previous studies suggesting that females and the elderly are some of the most vulnerable groups to outdoor air pollution [3,[82][83][84][85].

Weather and geographic influences in air pollution effects on AECOPD
Studies conducted in different countries independently identified significant associations of temperature, humidity, and various air pollutants with hospital admissions in COPD patients. In a study conducted in Spain, de Miguel-Díez et al. found that COPD was negatively affected by colder climatological factors and exposure to O3, CO, PM10 and NO2 [77]. In a multipollutant model in Hong Kong, SO2, NO2, PM10, and O3 were also shown to display a greater effect on AECOPD admissions in the cold season (December to March) than in the warm season [52]. On the other hand, a study in Taiwan showed that during the warmer season, COPD exacerbations occurred more frequently on days of temperature increases than on other days [86]. Stieb et al. also found that associations tended to be of greater magnitude during the warm season (April -September) in seven Canadian cities during the 1990s and early 2000s [81]. Another study in Romania reported that the adverse effect of PM exposure on chronic bronchitis was reduced by higher humidity, and that dry air aggravated the adverse effects of PM exposure in COPD patients [49]. Finally, Du et al. found that O 3 was the most closely spatially correlated with AECOPD hospitalizations at sites located in the northwest region of Jinhua, China, likely due to many industrial complexes in this region [68].

Symptoms in the respiratory response to shorter-term exposure to air pollution
Regarding COPD exacerbation symptoms, most studies showed that COPD symptoms, but not lung function, were mainly associated with raises in air pollution levels. Of these, dyspnea was significantly associated with PM10 with an increase in odds for an interquartile range change in pollutant of 13% (95% CI 4% to 23%) which is one common approach to presenting multi-pollutant health effect estimates, and this association remained significant after adjustment for other pollutant exposures [79]. In addition, short-term exposure to traffic pollution was shown to prevent the beneficial cardiopulmonary effects of walking in individuals with COPD [70].

Discussion
Chronic Obstructive Pulmonary Disease (COPD) is an inflammatory lung disease involving chronic bronchitis and emphysema. Patients with COPD are particularly vulnerable to the detrimental effects of environmental exposures, especially from air particulate and gaseous pollutants. While sex and gender differences in COPD prevalence and severity have been previously reported, sex-specific effects of air pollution exposure on COPD exacerbations and hospitalizations have not been studied in detail. The available evidence indicates that outdoor air pollution exposure affects lung function and triggers exacerbations in both male and female COPD patients. However, in reviewing the literature, we found that most studies conducted in this area have not accounted for sex in their analyses.
Our review of the literature identified 40 studies measuring associations of air pollution exposures and AECOPD. In these, it was widely reported that increases in environmental particulate and gaseous pollution concentrations were associated with increased risk of hospitalization for AECOPD, with varying effects depending on air quality composition, pollutant concentration, and time of exposure. We found that the majority of these studies enrolled mostly male subjects, and some enrolled men exclusively. This was a surprising finding considering that the incidence of COPD among women has increased in the past few decades [3]. Potential factors that may contribute to this bias are the historical (although not current) higher incidence of tobacco use in men, occupational exposures, and the previously described gender bias in COPD diagnosis [10,11,[87][88][89].
This study has several limitations. First, the number of studies identified by the selection criteria was limited and overrepresented in European and Asian countries, and the studies including or reporting participant data disaggregated by sex was markedly low, severely limiting the implications of our findings and our ability to conduct an analysis beyond descriptive. Second, our literature search was based on only two databases and including only studies in English, which could have omitted work available in other databases or languages, leading to selection bias. Third, using hospitalization rates as a comparison measure could also lead to bias, since hospitalization criteria may vary among countries and health systems, and since mortality associated with hospitalization for AECOPD does not always occur in the hospital.
This study has also several strenghts. First, it is the first review of the literature available assessing sex differences in an important outcome of the COPD pathogenesis and its relationship with air quality (i.e., hospitalization and mortality). Second, this study revealed a major gap in the research conducted to date in the area of COPD associations with air pollution in men and women, highlighting the importance of research design strategies that will identify sex-and gender-specific factors. Third, our review of the literature identified multiple studies where associations of air quality measures and AECOPD hospitalizations were reported, highlighting the importance of more research in these areas in order to design better preventative measures for COPD patients who live in geographical locations with poor air quality.
In the past few decades, the number of studies assessing the effects of air pollution exposure on lung disease has considerably increased [90]. However, studies considering sex (a biologic factor), or gender (a social construct, often used to refer to sex in publications) have been limited. Likewise, sex-specific disaggregation of data in the Global Burden of Diseases study has revealed that there are substantial differences between men and women that are frequently overlooked due to limitations in study designs [91]. This is highlighted by our findings in which only seven studies reported sex-disaggregated results, and only four studies had sufficient information to compare outcomes between male and female patients. Therefore, future studies should consider incorporating sex and gender variables at the design stage, and perform sex and gender disaggregated results reporting and analysis.

Conclusions
In conclusion, the available literature indicates that air pollution exposure is a relevant risk factor for AECOPD hospitalizations, although there is a significant absence of studies assessing sex-specific effects in this area. This review emphasizes the need of more studies designed to address sex-and gender-specific effects of air pollution exposure, as well as studies including women, a vulnerable population.