Impacts of Outdoor Particulate Matter Exposure on the Incidence of Lung Cancer and Mortality

Background and objectives: Long-term exposure to air pollution has been associated with lung cancer. This study aimed to evaluate the relative risk (RR) and hazard ratio (HR) of lung cancers and the prognostic implication of outdoor particulate matter (PM) pollution using a meta-analysis. Materials and Methods: We performed the meta-analysis using 19 eligible studies and evaluated the PMs, dividing into PM smaller than 2.5 µm (PM2.5) and PM smaller than 10 µm (PM10). In addition, subgroup analyses, based on the increment of PM exposure, location, sex, smoking history, and tumor histology, were performed. Results: Lung cancer was significantly increased by exposure to PM2.5 (RR 1.172, 95% confidence interval (CI) 1.002–1.371), but not PM10 exposure. However, there was no significant correlation between PM10 exposure and the incidence of lung cancers (RR 1.062, 95% CI 0.932–1.210). The all-cause and lung-cancer-specific mortalities were significantly increased by PM2.5 exposure (HR 1.1.43, 95% CI 1.011–1.291 and HR 1.144, 95% CI 1.002–1.307, respectively). However, PM10 exposure significantly increased the all-cause mortality, but not the lung-cancer-specific mortality. The lung-cancer-specific mortality was significantly increased by PM10 per 12.1 μg/m3 increment and in the Europe area. Conclusions: PM2.5 significantly increased lung cancer and the all-cause and lung-cancer-specific mortalities, whereas PM10 did not increase lung cancer or lung-cancer-specific mortality. However, PM10 increased the all-cause mortality and the PM10 per 12.1 μg/m3 increment and PM10 in the Europe area may increase the lung-cancer-specific mortality.


Introduction
Lung cancer is one of the most frequently diagnosed cancers and the leading cause of cancer-related mortality worldwide [1]. Nonsmoking-related lung cancers, suggesting other environmental factors, such as occupational exposure and residential radon, have been implicated as causes of lung cancer [2]. In addition, long-term exposure to air pollution has been associated with lung cancer [2]. The International Agency for Research on Cancer classifies outdoor particulate matter (PM) exposure as carcinogenic to humans [3]. Because lung cancers are associated with various causes, including PM exposure, it is difficult to determine the individual effect of PM exposure on the incidence of lung cancers. In addition, regional differences may be important in studies of exposure to PM [3,4]. In various regions, long-term exposure to fine PM has been found to increase the risk of lung cancer [5][6][7][8][9][10][11]. PM is classified by particle size, specifically PM 2.5 (fine particles with an aerodynamic diameter ≤ 2.5 µm) and PM 10 (inhalable particles with a diameter ≤ 10 µm), and also comprises a mixture of PM 2.5 and PM 10 , which are prominent components. Because particle size affects the degree of penetration into the lung, the harmful effects of PM can be different for different PM sizes [12]. Previous studies have reported that PM 2.5 included a higher proportion of mutagenic species and was more carcinogenic than PM 10 [13][14][15][16][17]. In addition, PM 10 mainly contains minerals and biological materials [17]. A comparison of carcinogenicity between PM 2.5 and PM 10 may be needed to elucidate the impact of particle size on the incidence of lung cancers.
In the present study, we aimed to evaluate the impact on the incidence of lung cancer and the prognostic implication of outdoor PM using a meta-analysis. We performed subgroup analyses based on the degree of increment, location, sex, smoking history, and tumor histology.

Published Study Search and Selection Criteria
The literature search was performed using the PubMed database on 15 May 2022. The search was performed using the following keywords: "particulate matter" and "lung" and "carcinoma". The titles and abstracts of searched articles were primarily screened for exclusion. Literature including original research and systematic review articles were also screened to identify additional eligible studies. The inclusion and exclusion criteria were as follows: (1) studies on the incidence of lung cancers and the all-cause and/or lung-cancerspecific mortality from PM exposure in humans were included and (2) non-original articles, such as case reports or review articles, were excluded.

Statistical Analyses
In the present meta-analysis, all data were analyzed and obtained using the Comprehensive Meta-Analysis software package Ver. 2 (Biostat, Englewood, NJ, USA). The RRs and HRs after PM exposure were determined and used in the meta-analysis. We performed subgroup analysis based on the increment of PM exposure, study location, sex, smoking history, and tumor histology. In this meta-analysis, the interpretation of the fixed and random effect models used the values of a random-effects model. The heterogeneity between eligible studies was assessed using Q and I 2 statistics and presented using p-values. In addition, sensitivity analysis was conducted to assess the heterogeneity of eligible studies and the impact of each study on the combined effect. To assess the publication bias, Egger's test was used. If significant publication bias was found, the fail-safe N and trim-fill tests were performed to confirm the degree of publication bias. A p-value < 0.05 was considered significant using a two-sided analysis.

Selection and Characteristics of Studies
A total of 375 studies were identified in the PubMed database search for the metaanalysis. Through the primary screening, 338 studies were excluded. Full-text reviews were undertaken for the remaining 37 studies. Finally, 19 studies were selected according to the inclusion and exclusion criteria. Among excluded studies, 170 studies were excluded due to a lack of sufficient information, such as the incidence of lung cancers and the allcause and/or lung-cancer-specific mortality from PM exposure. In addition, 96 studies were excluded due to non-human studies. The remaining reports were excluded for the following reasons: non-original articles (n = 68), focusing on other diseases (n = 11), articles in a language other than English (n = 7), and duplicate articles (n = 4) ( Figure 1). The characteristics of the eligible studies are shown in Table 1. for the following reasons: non-original articles (n = 68), focusing on other diseases (n = 11), articles in a language other than English (n = 7), and duplicate articles (n = 4) ( Figure 1). The characteristics of the eligible studies are shown in Table 1.

The Incidence of Lung Cancers by PM Exposure
The RRs of lung cancers due to exposure to PM 2.5 and PM 10 were 1.081 (95% CI 0.939-1.245) and 0.972 (95% CI 0.914-1.034), respectively ( Table 2). There were no significant differences in the RRs of lung cancers due to exposure to PM 2.5 and PM 10 . The incidence of lung cancer was non-significantly increased by the exposure of PM 2.5 (per 10 µg/m 3 increment; RR 1.081, 95% CI 0.939-1.245) and PM 10 (per 10 µg/m 3 increment; RR 1.062, 95% CI 0.932-1.210). In the PM 2.5 subgroup, the RRs in Asia were significantly increased by the exposure to PM (RR 1.061, 95% CI 1.044-1.078), but not in North America (RR 1.082, 95% CI 0.853-1.372). In subgroup analysis based on smoking history, RR of lung cancers was significantly higher in the former smoker subgroup but not in the never or current smoker subgroups. The RR of lung cancers was 1.650 (95% CI 1.040-2.619) in small cell carcinomas. Among the histologic subgroup of non-small cell carcinomas, squamous cell carcinoma was significantly correlated with increased RR by the exposure of PM 2.5 (RR 1.151, 95% CI 1.107-1.198). In the exposure of PM 10 , the RRs of lung cancers were not significantly different by smoking history and tumor histology.

The Mortality by PM Exposure
The mortalities due to all causes and lung cancers were marginally increased by the exposure to PM 2.5 (HR 1.143, 95% CI 1.011-1.291 and HR 1.144, 95% CI 1.002-1.307, respectively; Table 3). In Europe, the mortalities due to all causes and lung cancers were significantly increased by the exposure to PM 2.5 (HR 1.010, 95% CI 1.000-1.020 and HR 1.144, 95% CI 1.002-1.307, respectively). However, in Asia, the mortality by all causes, but not lung cancer, was significantly increased. There was no significant difference in the North America subgroup. The mortality due to exposure to PM 10 was significantly increased in the all causes subgroup but not in the lung cancer subgroup (HR 1.091, 95% CI 1.023-1.162 and HR 1.168, 95% CI 0.962-1.419, respectively). In the PM 10 subgroup, the mortality due to lung cancers was significantly increased per 12.1 µg/m 3 increment and in Europe (HR 1.270, 95% CI 1.250-1.290 and HR 1.930, 95% CI 1.294-2.879, respectively). However, in the PM 10 subgroup of North America, the mortality by all causes, but not lung cancer, was significantly increased. Table 3. The estimated hazard ratios for risk of the all-cause and lung-cancer-specific mortalities according to the increasing particulate matter concentration and geographical locations.

Discussion
Outdoor PM has been implicated as a carcinogen [15]. The impact of outdoor PM exposure can differ between location, exposure period, and outdoor activity. However, the impact of outdoor PM exposure according to other various factors is unclear. To the best of our knowledge, the present study is the first study using a meta-analysis to elucidate the factors that impact PM exposure on the incidence of lung cancers and all-cause and lung-cancer-specific mortality.
In existing studies, PM in outdoor air is generally classified into PM 2.5 and PM 10 . Larger particles can be filtered out by the clearance system of the upper respiratory tract. Thus, smaller particles can penetrate more deeply. Because the PM size affects the degree of penetration into the lung, the harmful effects of different PM sizes can differ [12]. In addition, PM 2.5 includes a higher proportion of mutagenic species [13][14][15][16][17]. A comparison between particle size groups will be needed to elucidate the impact of PM exposure. In addition, the concentration and exposure to PM can differ by location; thus, subgroup analysis based on locations is needed.
Previous studies have reported correlations between PM 2.5 exposure and the incidence and mortality of lung cancers [2,5,[18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34]. In a previous meta-analysis, lung cancer risk was found to significantly increase, by 9%, according to 18 studies [15]. The previous meta-analysis collected data from worldwide studies from the 1970s to the 2000s. However, the means of incidence and mortality are unclear in this previous study [15]. We evaluated the impact of PM exposure on the incidence of lung cancers. In the present study, the incidences of lung cancers due to PM exposure were evaluated. The RRs of lung cancers due to exposure to PM 2.5 and PM 10 were 1.081 (95% CI 0.939-1.245) and 0.972 (95% CI 0.914-1.034), respectively. Although PM exposure was associated with a slight increase in the incidence of lung cancers, there was no statistical significance between the incidence of lung cancers and PM exposure.
Hamra et al. reported the relative risks of lung cancers due to outdoor PM exposure using a meta-analysis and assessed lung cancer risk for the combined incidence and mortality of lung cancers [15]. They reported that the relative risks of lung cancer were 1.09 (95% CI 1.04-1.14) and 1.08 (95% CI 1.00-1.17) due to PM 2.5 and PM 10 exposure, respectively. Similar to the previous meta-analysis, we analyzed subgroups based on the degree of PM exposure (per 10 µg/m 3 increment). Lung cancer was slightly increased by the exposure of PM 2.5 (per 10 µg/m 3 increment; RR 1.081, 95% CI 0.939-1.245) and PM 10 (per 10 µg/m 3 increment; RR 1.062, 95% CI 0.932-1.210). However, statistical significance was found. In contrast to our study, Hamra et al. obtained and converted the raw data from the authors of the original studies [15]. This factor may be the cause of the different results between the present and previous meta-analyses.
In the present study, subgroup analyses based on various factors, including the increment of PM exposure, location, sex, smoking history, and tumor histology, were performed. In a previous meta-analysis, the RRs of lung cancers were evaluated per 10 µg/m 3 increment [15]. We assessed the change in mortality according to the increment of PM exposure. In addition, a comparison of mortality between all causes and lung cancers after PM exposure was performed. Statistical significance was found in some increment subgroups. In the all causes subgroup, the HRs due to the exposure to PM 2.5 were 1.194 and 1.101 per 5.3 and 10 µg/m 3 increment, respectively. The HRs due to the exposure to PM 10 were 1.043, 1.190, 1.070, and 1.150 per 6, 7, 10, and 29.5 µg/m 3 increment, respectively. The increasing values were slightly lower than those from lung cancers.
We examined detailed analyses of mortality due to PM exposure, dividing mortality into all causes and lung cancers. Following exposure to PM 2.5 and PM 10 , mortality due to all causes was significantly increased. However, the mortality due to lung cancers increased considerably in the PM 2.5 subgroup but not in the PM 10 subgroup. As described above, the mutagenic species were more included in PM 2.5 than in PM 10 and were more carcinogenic than PM 10 [13][14][15][16][17]. In the subgroup analysis, there was some difference in mortality between locations. In the all causes subgroup, mortalities were significantly increased in Asia and Europe. However, in North America, there was a significant increase in mortality due to PM 2.5 exposure, but not due to PM 10 exposure. In North America, the change in mortality due to lung cancer was similar to that due to all causes. In addition, in the PM 10 exposure subgroup, there was a significant correlation between PM exposure and mortality by lung cancer in Europe but not in North America.
Previous studies reported the correlation between air pollution and lung cancer in never-smokers [35,36]. In addition, no significant differences were found in the incidence of lung cancers between sex and between smoking history. Per 10 µg/m 3 increment of PM 2.5 , the RRs of North America and Europe subgroups were 1.11 (95% CI 1.05-1.16) and 1.03 (95% CI 0.89-1.20), respectively. Per 10 µg/m 3 increment of PM 10 , the RRs of North America and Europe subgroups were 1.02 (95% CI 0.96-1.09) and 1.27 (95% CI 0.96-1.68), respectively. Statistical significance was found only in North America with a per 10 µg/m 3 increment of PM 2.5 .
In Yang's report, the incidence of overall lung cancers was significantly increased due to PM 2.5 exposure (6.0%, 95% CI 4.3-7.7%) [5]. In addition, squamous cell carcinomas and adenocarcinomas were significantly increased due to PM 2.5 exposure (14.8%, 95% CI 10.3-19.4%, and 6.5%, 95% CI 3.3-9.8%, respectively) [5]. Hamra et al. reported that adenocarcinomas were significantly correlated with PM exposure based on a meta-analysis [15]. However, there was no significant correlation between the incidence of squamous cell carcinomas and PM exposure [15]. Moon et al. reported that an increasing incidence of lung cancers was identified in adenocarcinomas of current smokers [2]. However, the increasing incidence of other histology types was not determined [2]. In our study, the incidences of non-small cell and small cell lung cancers were compared in contrast to previous studies. The exposure of PM 2.5 had a significant impact on increased RR in squamous cell and small cell carcinomas. However, in the PM 10 exposure group, there were no statistical differences in various histologic subtypes.
Some limitations in the current meta-analysis exist. First, a comparison of the mortality between our study and previous studies in tumor histology could not be performed due to insufficient information. Second, changes in the incidence of lung cancers due to the degree of PM exposure other than 10 µg/m 3 increment could not be obtained from eligible studies. The present study only performed an analysis for the per 10 µg/m 3 increment. Third, the impact of the mixture of PM 2.5 and PM 10 could not be evaluated due to a lack of information in eligible studies. Fourth, the analysis according to previous cancer history could not be included in the present meta-analysis due to insufficient information of eligible studies. Fifth, a detailed evaluation according to the pack years of cigarette consumption could not be performed due to no information on eligible studies. Six, we could not evaluate the effects of concentration range. Instead of concentration range, the impacts of the increment of PM were investigated and evaluated.

Conclusions
In conclusion, exposure to PM was significantly correlated with mortality by all causes. However, the incidence and mortality of lung cancer were significantly increased by PM 2.5 exposure, but not PM 10 exposure. Although various factors associated with PM exposure affect the incidence of lung cancers and mortality, careful interpretation is needed, such as the size and exposure of PM and location.

Conflicts of Interest:
The authors declare no conflict of interest.