Relationship between Air Pollution and Lung Cancer in Fujian Province: A Case-control study


 Background: Outdoor air pollutants, especially particulate matters, are defined as a type of carcinogen by the International Agency for Research on Cancer. Studies have shown that air pollution is associated with lung cancer morbidity or mortality . This study is aimed at exploring the relationship between air pollutants and primary lung cancer in Fujian Province, China. Methods: We conducted a hospital-based, retrospective, case–control epidemiological study on three different populations to assess the occurrence of lung cancer caused by exposure to various levels of air pollution. Results: The total study population comprised 885 lung cancer patients and 1,220 healthy controls. The following parameters were identified as risk factors for lung cancer among the total population: smoking; exposure to cooking oil fumes; passive smoking; medical history of lung disease; family history of lung cancer; and exposure to PM10, PM2.5, and O3. Fruit intake and physical exercise were protective against lung cancer. For smokers, medical history of lung disease, family history of lung cancer, and exposure to PM10, and PM2.5 were risk factors for lung cancer, while fruit intake and physical exercise were protective factors. Among non-smokers, exposure to cooking oil fumes; medical history of lung disease; family history of lung cancer; and exposure to PM10, PM2.5, and O3 were factors increasing the risk of lung cancer, while fruit intake, physical exercise, and tea drinking were protective. Conclusions: Long-term exposure to PM10, PM2.5, and O3 was found to be significantly associated with increased risk of lung cancer, with the risk being greater for non-smokers and persons exposed to cooking oil fumes.


Background
As per GLOBOCAN 2018, approximately 18.1 million new cancer cases are diagnosed worldwide each year. Lung cancer ranks first among the causes of new cancer cases, with about 2.09 million cases diagnosed annually, accounting for 11.6% of new cases of malignant tumors. The incidence of lung cancer in East Asia, particularly in China, is far higher than that worldwide (more than 40/100 000) 1 .
Lung cancer was the leading cause of cancer incidence and mortality in China for many years 2 . According to the 2017 Malignant Tumors report from Fujian Province, lung cancer ranked first among all types of cancers in males (62.61/100 000) and second in females (24.8/100 000). Clearly, lung cancer has become a serious threat to public health and places a heavy disease burden in Fujian Province and China as a whole.
Smoking is a known and confirmed risk factor for lung cancer. The increase in the number of female smokers in developed countries is one of the reasons for the increase in the incidence of lung cancer among women and decrease in the comparative incidence among men 3 . However, some studies have found that about 15% of men and 53% of women with lung cancer do not smoke. Moreover, about 25% of lung cancer cases worldwide cannot be attributed to tobacco usage 4 . This has led to an increased interest in the exploration of the causes of lung cancer other than smoking, particularly air pollution 5 , diet 6 , infection 7,8 , cooking fumes 9 , medical history of respiratory diseases 10 , and occupational exposure 11 .
For nearly 50 years, research has been underway to confirm the relationship between outdoor air pollution and lung cancer 12 . The International Agency for Research on Cancer (IARC) defines outdoor air pollutants, especially particulate matter, as a type of carcinogen. This classification is primarily based on evidence that long-term persistence of high average concentrations of PM 2.5 in outdoor air is associated with lung cancer morbidity or mortality 13 . However, the IARC report is a qualitative assessment of hazard identification, and it does not provide the relative risk or odds ratio (RR or OR) of lung cancer associated with outdoor air pollution.
The current research results still leave several key points unclear. First, the population groups that are most vulnerable to air pollution are yet to be identified [14][15][16][17] . Second, few studies have focused on the dose-response relationship between long-term exposure to outdoor air pollution and lung cancer 16,18,19 . In particular, very few studies from China have investigated the risk of lung cancer in relation to air pollution; among these studies, case-control studies and cohort studies are very few, and most of the existing studies focus on lung cancer mortality.
Given the limitations of epidemiological research methods, cohort studies on the causal relationship between outdoor air pollution and lung cancer would require a long time to yield conclusive results. In this study, we sought to conduct a hospitalbased, case-control study to investigate the relationship between air pollutants and lung cancer in Fujian Province, with special focus on the exposure levels of different population groups to air pollutants. We expect that the results of this study would provide a scientific basis for targeted improvement of the atmospheric environment for protection of public health.

Study subjects
Between January 2010 and December 2015, we recruited from three sources:

Department of Thoracic Surgery and Respiratory Medicine of the First Affiliated
Hospital of Fujian Medical University, Union Hospital Affiliated to Fujian Medical University, and Fuzhou General Hospital of Nanjing Military Region. Patients were included if they had been newly diagnosed with primary lung cancer, as confirmed by fiberoptic bronchoscopy or surgical histopathology, and had resided in Fujian province of China for more than 10 years.
The population-based approach was applied to the recruit the control group, which comprised healthy community dwellers randomly selected from the resident records of Fujian province. The control group was frequency-matched to the cases by ethnicity, gender, and age. Every control subject was of ages ±2 years as compared to the matching cases. All the selected control individuals successfully met the inclusion criteria and completed the study, without any drop-out during the study.
In total, the case group comprised 885 patients with lung cancer, while the control group comprised 1,220 healthy individuals. All participants provided written, informed consent before undergoing the examination. The study protocol was approved by the Ethics Committee of Fujian Medical University.

Questionnaire and variables
All participants were surveyed using a standardized questionnaire, which was administered during a scheduled phone interview conducted via a study team member. The 18-page questionnaire had questions pertaining to the patients' demographic characteristics information, tumor characteristics, and data regarding the subject's living environment, dietary habits, smoking history, alcohol consumption, intensity of physical activity, occupational exposure to air pollutants, and exposure to domestic pollutants. The questionnaire also included items regarding medical history, family history, and lifestyle-related parameters. Data regarding reproductive health were collected for female participants.
Body mass index (BMI) was determined as the ratio of body weight (kg) and height[m] 2 . A positive smoking history was defined as a history of having smoked more than 100 cigarettes during his/her lifetime. Passive smoking history was defined by non-smoking history of inhaling cigarette smoke generated by others or exposure to exhaled smoke more than 15 minutes per day. A history of alcohol consumption was defined as drinking at least one alcoholic beverage per week for more than six months, irrespective of the type of alcoholic drink. Drinking tea was defined as consuming at least one cup of any kind of tea per week for more than six months. A family history of cancer was defined as the occurrence of a malignant tumor in first-degree or second-degree relatives. Occupational physical activity was rated as low, moderate, or high intensity, as defined by the Reference Standard of Labor Intensity recommended by the Chinese Nutrition Society in 2000. To check for exposure to cooking oil fumes, participants were enquired about whether their degree of exposure to fumes in the kitchen was none (no exposure), light, moderate, or heavy 20 .

Overview of the study area
Fujian Province is located on the southeast coast of China. The land area is 124 000 square kilometers in total. From north to south, the province measures about 530 kilometers and occupies about 480 kilometers from east to west. Fujian is located at a latitude of 23°30' and 28°22' north and at longitude of 115°50' and 120°40' east.
The terrain is higher in the northwest and lower in the southeast. About 90% of the total land area of the province is covered by mountains and hills. In addition, the coastal landform pattern is dominated by a tortuous coastline that includes several bays and peninsulas. The climate in Fujian Province is subtropical monsoon climate, which makes the area warm and humid. The annual average temperature is 15°C-22°C, which increases from the northwest to the southeast. The annual average precipitation is 800-1,900 mm, with the the precipitation being greatest between May and June every year, with several typhoons occurring at the turn of summer and autumn.

Air pollution data collection
For this study, we aimed to obtain a reasonable estimate of each subject's exposure to atmospheric pollutants by using a longer average annual concentration. The levels of sulfur dioxide (SO 2 ), nitrogen dioxide (NO 2 ) and suspended particulates smaller than 10 μm in aerodynamic diameter (PM 10 ) in Fujian Province were obtained from the complete monitoring data of national control points, which has In addition, the concentrations of O 3 , CO, and PM 2.5 were measured from 2013 to the time of enrollment of the subjects. According to the detailed home address (specific to the district, street, and house number), the spatial coordinates (latitude and longitude) of each subject was accurately determined by GPSspgxGeo coding software and Tencent map software, which can be further mapped to the predicting model and used to generate the estimating pollution data for each subject.

Statistical analysis
The Chi-square test was used to compare the general characteristics of cases and controls. Stratified analysis for different populations was performed using a t test for the exposure levels of air pollution indexes. Unconditional logistic regression modeling and identification of possible confounding factors were performed by calculating odds ratios (ORs) and 95% confidence intervals (CIs) for air pollution indexes and lung cancer risk. All analysis was performed using the SPSS 25.0 software package (IBM Corporation, Armonk, New York, USA) and ArcGIS 10.3 (ESRI Inc, USA). All p-values were based on a two-sided test with an α of 0.05.

Demographic characteristics of the study subjects
The current study comprised 2,105 subjects, including 885 cases and 1,220 controls. Subjects included 1,354 males and 751 females, with ages between 23 and 90 years of age and average age of 58.94±10.73 years. The case and control groups did not show any significant difference in the distribution of gender, age, ethnicity, and marital status (P>0.05). However, significant intergroup differences were noted in the educational level, occupation, and BMI (P<0.05).
Among the smokers, no significant difference was noted between the case group and the control group in terms of the distribution of gender, ethnicity, and marital status (P>0.05). However, statistically significant intergroup differences were noted in the distribution of education, occupation, and BMI (P<0.05). Among the nonsmokers, the case group and control groups did not differ significantly in the distribution of age, ethnicity, and marital status (P>0.05), but showed significant differences in gender, education level, and occupational exposure (P<0.05).

The risk factors of lung cancer
In this study, we explored the potential factors for lung cancer in the three groups: general population, smokers, and non-smokers (Table 2). Of the total population, after adjusting for BMI, education, and occupation, we found that subjects with a family history of lung cancer or medical history of lung disease, drinking alcohol, smoking, passive smoking, or exposure to cooking oil fumes were susceptible to lung cancer. We found that the risk of lung cancer was low for subjects who performed physical exercise and had a regular intake of fruit (more than 3 times per week). Among the smokers, after adjustment for age, BMI, education, and occupation, the risk factors for lung cancer were family history of lung cancer, medical history of lung disease, and exposure to cooking oil fumes, while the protective factors were physical exercise and fruit intake. Similarly, for nonsmokers, the risk factors for lung cancer were family history of lung cancer, history of lung diseases, passive smoking, and exposure to cooking oil fumes, after adjustment for gender, education, and occupation, whereas the protective factors were drinking tea, physical exercise, and fruit intake.

Univariate analysis of levels of exposure to atmospheric pollutants in different populations
The annual average of SO 2 , NO 2 , and PM 10 concentrations for the period between  (Table 3). For the total population, after adjustment for education, occupation, BMI, family history of lung cancer, and medical history of lung disease, alcohol consumption, smoking, passive smoking, exposure to cooking oil fumes, physical exercise, and fruit intake, the results showed that the risk of developing Next, we compared the results for smokers, after adjusting for age, education, occupation, BMI, family history of lung cancer, and medical history of lung disease, exposure to cooking oil fumes, physical exercise, and fruit intake. Further, for nonsmokers, the results were compared after adjusting for gender, education, occupation, family history of lung cancer, medical history of lung disease, passive smoking, exposure to cooking oil fumes, drinking tea, physical exercise, and fruit intake. The analysis indicated that for both groups, exposure to PM 10 and PM 2.5 was a risk factors for lung cancer. Moreover, exposure to PM 2.5 had a greater impact on non-smokers than on smokers, as shown by the following results for both groups: OR was 61.431 (95% CI: 18.041-209.181) and 24.545 (95% CI: 8.12-74.265) (in nonsmokers and smokers, respectively, for exposure to PM 2.5 concentrations of 0.020-0.027 (mg/m 3 )) and OR was 11.814 (95% CI: 3.622-38.540) and 2.630 (95% CI: 0.910-7.595) (for non-smokers and smokers, respectively, for exposure to PM 2.5 concentrations of 0.027-0.035 (mg/m 3 ). Since none of the participants were exposed to O 3 concentrations of less than 0.047 mg/m 3 in the control group, the effect of O 3 exposure on the smokers could not be assessed. Among the nonsmokers, exposure to O 3 was identified as a risk factor for lung cancer. However, SO 2 and lung cancer did not show any associated, both in the general population and in the group of smokers or non-smokers. Similarly, none of the participants in the control group were exposed to a low concentration of CO, and therefore, the association of CO with lung cancer could not be evaluated (Table 3).
We also analyzed the relationship between air pollution and lung cancer stratified by passive smoking and exposure to cooking oil fumes, and the results were consistent with those of smoking. The detailed results are shown in Additional Table   1 to 3 (for passive smoking) and Additional Table 4 to 6 (for cooking oil fume exposure).

Multivariate analysis of levels of exposure to atmospheric pollutants in different populations
All the variables that were found to have a significant impact on the development of lung cancer in the previous analysis were further subjected to multi-factor unconditional logistic regression analysis using the backward stepwise method. We used P≤0.05 as the inclusion criterion and P≥0.10 as the exclusion criterion. The results of the analyses indicated that the following factors increased the risk of lung cancer among the general population: smoking; exposure to cooking oil fumes; passive smoking; medical history of lung disease; family history of lung cancer; and exposure to O 3 , PM 10 , and PM 2.5 . On the other hand, fruit intake and physical exercise were found to be protective factors against risk of lung cancer (Table 4).
For smokers, medical history of lung disease; family history of lung cancer; and exposure to PM 10 , and PM 2.5 were factors that increased the risk of lung cancer, whereas fruit intake and physical exercise were factors that reduce the risk of lung cancer (Table 5). Similarly, among the non-smokers, cooking oil fumes; medical history of lung disease; family history of lung cancer; exposure to PM 10 , PM 2.5 , and O 3 may increase the risk of lung cancer, while fruit intake, physical exercise, and drinking tea were found to protect against the risk of lung cancer ( Table 6).
The results of the analysis in different populations are shown in Additional Table 7-9 for passive smoking and in Additional Table 10-12 for exposure to cooking oil fumes.

Discussion
This hospital-based case-control study was designed to evaluate the relationship between atmospheric concentrations of air pollutants and the occurrence of lung cancer in Fujian Province. We found that the overall risk factors for lung cancer were smoking; exposure to cooking oil fumes; passive smoking; medical history of lung disease; family history of lung cancer; and exposure to PM 10 , PM 2.5, and O 3 .
Fruit intake and physical exercise were identified as protective factors. Our results on the risk factors are consistent with those of several previous reports. We found that long-term exposure to PM 10 , PM 2.5 , and O 3 was significantly associated with an increased risk of lung cancer, and the risk appears to be greater in non-smokers than in smokers or the general population. Further, this difference appears to be even greater for people exposed to cooking oil fumes than among those without such an exposure.
In the recent past, extensive research has been conducted in China on the shortterm health effects of air pollution 22,23 . However, studies exploring the long-term effects of air pollution on health, particularly those employing more reliable methods such as cohort studies, have been scarce. Most of the studies conducted thus far are related to mortality 24-27 . In recent years, China has been making continuous improvement in air pollution monitoring system. However, China is a vast country. Therefore, further investigation is necessary to evaluate the exposure characteristics, dose-risk models, and long-term health risks of single or multiple pollutants in different regions.
At present, there is a strong research focus on the relationship between air pollution and lung cancer. Studies 28,29 have reported that long-term exposure to PM 2.5 , NO 2 , NO x, and SO 2 is significantly associated with an increased risk of lung cancer. Such associations have been shown to hold true for both smokers and nonsmokers as well as men and women, with the impact being the same for all groups.
For non-smokers, exposure to outdoor PM 2.5 is greater for patients with lung cancer and exhibits a linear dose-response relationship 30 . The results of the ACS CPS-II study 31 and the European Air Pollution Impact Cohort Study (ESCAPE) 5,32 have also demonstrated a positive correlation between exposure to outdoor PM 2.5 and the occurrence of lung cancer. Our results had a wide 95% CI range, which might be attributed to the small sample size in our study, but our results also suggested that PM 10 and PM 2.5 were risk factors for lung cancer. Similar to our study, the study by Yang et al. 28 also revealed that the association of PM 2.5 with lung cancer (RR = 1.18) was more pronounced in non-smokers than in the general population (RR = 1.07). However, because of the limited number of studies included (n = 3) and a wide overlap between the general population and people who never were smokers, the controversy still remains regarding whether there is a connection between lung cancer and air pollution. Currently, there is still no report on the effect of exposure to cooking oil fumes on individuals.
Gerard Hoek et al. 33 investigated the risk of lung cancer in the general population associated with exposure to ambient air pollution. They used the population attributable risk fraction (PAF) indicator, which describes the fraction of lung cancer incidence in the general population that can be prevented by eliminating PM 2.5 exposure. The proportion of the population with lung cancer was found to be between 28.6% and 86.7% when RR = 1.5 and the concentration of PM 2.5 was reduced by 10-60 μg/m 3 . The RR value of ambient air pollution was found to be much smaller than that for active smoking; however, air pollution affects the entire population. Thus, a reduction in the concentration of PM 2.5 will result in a substantial reduction in the incidence of lung cancer in the overall population.
Study results on the association between long-term exposure to O 3  The current study has a few limitations. First, the study was conducted with a hospital case-control design and, therefore, it has flaws that are inherent in the method itself. To overcome this, various control measures, such as selecting cases from multiple hospitals and choosing objective indicators, were employed in this study. However, most of the controls were from Fuzhou, and the difference in the regional distribution of cases and controls may have influenced the results. Second, although the locations of the atmospheric sampling points were established scientifically and were reasonable, the air quality in Fujian Province that was analyzed in this study may not completely reflect the air quality of the entire Fujian Province because of the limited number of sampling points. Third, the levels of SO 2 , NO 2 , and PM 10 were measured since 2005, while those of O 3 , CO, and PM 2.5 were obtained since 2013. The difference in the data collection period for the two set of parameters may have influenced the results. Fourth, the concentrations of the air pollutants measured in the indices of this study were not obtained by individual portable monitoring; therefore, there is a possibility of an estimate bias that resulted in a wider CI. Future research that is based on a greater sample size and forward-looking research design is necessary to provide in-depth insight into these points and strengthen environmental and public health monitoring.

Conclusion
This hospital-based case-control study revealed a few factors that influence the onset of lung cancer, including smoking, passive smoking, exposure to cooking oil fumes, family history of lung cancer, and medical history of lung disease. Increasing fruit intake, physical exercise, and drinking tea were found to be protective against lung cancer. More importantly, long-term exposure to PM 10, PM 2.5 , and O 3 is significantly associated with increased risk of lung cancer, and the risk appears to be greater in non-smokers than in smokers or the general population; moreover, this difference appears to be greater in people exposed to cooking oil fumes than those without.

Declarations
Ethics approval and consent to participate

Consent for publication
Not applicable Availability of data and material The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.    Unconditional logistic regression with the backward stepwise method Unconditional logistic regression with the backward stepwise method  Additional Tables .docx