Comments
Abstract
Rationale: Both genetic and environmental factors contribute to lung cancer, but the degree to which air pollution modifies the impact of genetic susceptibility on lung cancer remains unknown.
Objectives: To investigate whether air pollution and genetic factors jointly contribute to incident lung cancer.
Methods: We analyzed data from 455,974 participants (53% women) without previous cancer at baseline in the UK Biobank. The concentrations of particulate matter (PM) (PM ⩽2.5 μm in aerodynamic diameter [PM2.5], coarse PM between 2.5 μm and 10 μm in aerodynamic diameter [PMcoarse], and PM ⩽10 μm in aerodynamic diameter [PM10]), nitrogen dioxide (NO2), and nitrogen oxides (NOx) were estimated by using land-use regression models, and the association between air pollutants and incident lung cancer was investigated by using a Cox proportional hazard model. Furthermore, we constructed a polygenic risk score and evaluated whether air pollutants modified the effect of genetic susceptibility on the development of lung cancer.
Measurements and Main Results: The results showed significant associations between the risk of lung cancer and PM2.5 (hazard ratio [HR], 1.63; 95% confidence interval [CI], 1.33–2.01; per 5 μg/m3), PM10 (HR, 1.53; 95% CI, 1.20–1.96; per 10 μg/m3), NO2 (HR, 1.10; 95% CI, 1.05–1.15; per 10 μg/m3), and NOx (HR, 1.13; 95% CI, 1.07–1.18; per 20 μg/m3). There were additive interactions between air pollutants and the genetic risk. Compared with participants with low genetic risk and low air pollution exposure, those with high air pollution exposure and high genetic risk had the highest risk of lung cancer (PM2.5: HR, 1.71; 95% CI, 1.45–2.02; PM10: HR, 1.77; 95% CI, 1.50–2.10; NO2: HR, 1.77; 95% CI, 1.42–2.22; NOx: HR, 1.67; 95% CI, 1.43–1.95).
Conclusions: Long-term exposure to air pollution may increase the risk of lung cancer, especially in those with high genetic risk.
Both genetic factors and ambient air pollution contribute to lung cancer, but previous studies usually focused on the separate association of either air pollution or genetic susceptibility with the risk of lung cancer and largely ignored the combined effect of or interactions between genetic and environmental factors.
We revealed that residential exposure to air pollution was significantly associated with an elevated risk (∼63%) of lung cancer in the UK Biobank. In addition, the association was similar across all strata of genetic risk, and the greatest relative increase in the risk of lung cancer was observed among those with high genetic risk. This suggests that improving air quality will benefit the entire population, especially those with high levels of genetic risk. Furthermore, this study was the first to assess the possible additive interaction between air pollution and genetic factors with regard to lung cancer. The observed additive interaction could indicate that air pollution modifies the impact of genetic susceptibility on the risk of lung cancer.
Lung cancer is the most common cancer and the leading cause of cancer-related death worldwide (1). Smoking is the most well-established environmental risk factor, and the proportion of lung cancer cases attributable to smoking was found to be 81.7% (2). In addition, several studies have indicated that exposure to ambient air pollutants, including particulate matter (PM), nitrogen oxides (NOx), ozone, and sulfur dioxide, may be positively associated with lung cancer (3). The likely explanation is that long-term exposure to air pollution may increase lung cancer risk through oxidative damage, which occurs via inflammatory injury and the production of reactive oxygen species (4–6). However, the results of population studies remain controversial because of differences in the study designs and the number of participants (7–14). Therefore, it is necessary to evaluate the real effect of ambient air pollutants on lung cancer in a large, independent, prospective study.
Accumulating evidence, including from a Nordic twin study, has shown that genetic factors also play an important role in the development of lung cancer (15). Genome-wide association studies (GWASs) have been successful at identifying the genetic predisposition for many complex diseases, including lung cancer (16). Since 2008, GWASs have identified 51 susceptibility loci for lung cancer in different ethnicities (16, 17). Although each variant can only account for a small to moderate proportion of the heritable risk of lung cancer, a polygenic risk score (PRS) has been shown to be effective at measuring the cumulative effect of multiple risk-associated variants (17). Moreover, it has been widely accepted that the effects of genetic factors can be strongly modified by environmental factors, and the investigation of gene–environment interactions can provide novel insights into the precise prediction of and targeted interventions in human cancers (18–20).
Previous studies usually focused on the separate association of air pollution or genetic variants with the risk of lung cancer and have largely ignored their combined effect or interactions between genetic and environmental factors. Hence, we conducted a prospective study by using the UK Biobank to examine the hypothesis that exposure to residential ambient air pollution and genetic factors jointly contribute to incident lung cancer. We then assessed the extent to which air pollution exposure modified the impact of genetic susceptibility on lung cancer risk.
The UK Biobank study protocol is available online (https://www.ukbiobank.ac.uk/media/gnkeyh2q/study-rationale.pdf). In brief, the UK Biobank recruited more than 500,000 participants aged 40–69 years between 2006 and 2010. Participants visited 1 of 22 assessment centers in England, Wales, and Scotland, where they completed a self-reported touchscreen questionnaire, which collected comprehensive personal and exposure information; underwent physical measurements; and provided biological samples that were used for various types of assays (21). All of the participants provided informed written consent, and the UK Biobank study was granted ethics approval from the North West Multicenter Research Ethics Committee.
The annual average estimates of ambient air pollutants, including PM ⩽2.5 μm in aerodynamic diameter (PM2.5), coarse PM between 2.5 μm and 10 μm in aerodynamic diameter (PMcoarse), and PM ⩽10 μm in aerodynamic diameter (PM10); nitrogen dioxide (NO2); and NOx were calculated by the UK Biobank by using land-use regression (LUR) models developed by the ESCAPE (European Study of Cohorts for Air Pollution Effects) project (22, 23). On the basis of a range of predictive variables (such as traffic intensity, population, topography, and land use) derived from the geographic information system, LUR models were used to calculate the spatial variation in air pollutant concentrations at participants’ residential addresses provided at baseline. Leave-one-out cross-validation showed good model performance for PM2.5, PM10, NO2, and NOx (cross-validation R2 = 77%, 88%, 87%, and 88%, respectively) and a comparatively moderate performance for PMcoarse (cross-validation R2 = 57%) in the southeast England area (London/Oxford). Details on the development and validation of the ESCAPE LUR models have been described elsewhere (22, 23). The LUR estimates of PM were valid for 400 km from Greater London but not beyond, and as such, patients living in northern England and Scotland were eliminated from the PM analyses, resulting in 33,935 missing values.
In the UK Biobank, the annual concentrations of NO2 and PM10 were available for several years (2005–2007 and 2010 for NO2, 2007 and 2010 for PM10); as a result, we took the average of these values as the mean estimates. All other types of PM and NOx had a single year of exposure data (2010). The pollutant concentrations were categorized as low and high pollution on the basis of the World Health Organization yearly air quality guideline values (24) or the median.
The procedure for genotyping and imputation of the SNPs are described briefly in the online supplement. To construct the PRS, we included 18 SNPs reported in the largest available lung cancer GWAS of people of European descent conducted by the International Lung Cancer Consortium (Database of Genotypes and Phenotypes accession number phs000876) (25). All 18 SNPs needed for our analysis were available in the UK Biobank imputed database. Details on the generation of the PRS are shown in the online supplement. The PRS was categorized as low (lowest tertile), intermediate (second tertile), and high (highest tertile) genetic risk (tertiles were based on the distribution of the PRS among the noncases), as described previously (26).
The outcome for this study was incident lung cancer. By linking to national cancer registries in England, Wales, and Scotland, cancer cases in the UK Biobank cohort were identified. The complete dates of follow-up were March 31, 2016, for England and Wales, and October 31, 2015, for Scotland. All diseases were confirmed according to the 10th Revision of the International Classification of Diseases, and lung cancer was defined as 10th Revision of the International Classification of Diseases codes C33–C34 (27). Participants were followed up from the enrollment until the time of lung cancer diagnosis or censoring. Censoring was defined as the time of death, withdrawal from the study, or the end of follow-up, whichever came first.
A series of sociodemographic and behavioral confounders were identified on the basis of the current literature. Age was calculated by using the birth date and the date of baseline assessment. The body mass index (BMI) was derived from the measured height and weight. The five levels of the average total before-tax household income were classified as “less than” or “equal to or above” £31,000, which is closest to the UK median household income in October 2009 (£27,530) (28). The education level was defined as the “degree or professional education” or “other levels.” The smoking status variable divided participants into never-smokers and former or current smokers.
Multiple imputation was performed to impute missing covariate data, and the proportions of missing data were as follows: 15% for pack-years of smoking and household income; 2% for education level; and less than 1% for BMI and smoking status. Details on the procedure for multiple imputation are provided in the online supplement.
Figure 1 shows the process for the construction of the analytical cohort. After exclusions, the final samples sizes of the cohorts included in the lung cancer and air pollution analyses were 449,219 (NOx data set) and 417,422 (PM data set). In addition, when combined with genetic data, we included 435,733 and 404,750 participants in the NOx and PM data sets, respectively.
Figure 1. Study design and workflow. GWAS = genome-wide association study; NOx = nitrogen oxides; PM = particulate matter; PRS = polygenic risk score; WHO = World Health Organization.
[More] [Minimize]Cox proportional hazard models were used to assess the associations of ambient air pollution and genetic factors with incident lung cancer and to estimate hazards ratios (HRs) and 95% confidence intervals (CIs) with adjustment for age, sex, BMI, household income, education level, smoking status, pack-years of smoking, and the top 40 genetic principal components and genotyping batch, as well as pollutant concentrations (when appropriate). More details on the selection process of potential confounders can be found in the online supplement. Schoenfeld residuals were used to test the proportional hazard assumption. All associations are reported per 5-μg/m3 increase in PM2.5 and PMcoarse, per 10-μg/m3 increase in PM10 and NO2, and per 20- μg/m3 increase in NOx to directly compare with the ESCAPE project (8). The results are also reported per interquartile range (IQR) increase in pollutant effects in the UK Biobank population. The additive interaction was evaluated by using two indexes: the relative excess risk due to the interaction (RERI) and the attributable proportion due to the interaction (AP) (29). The 95% CIs of the RERI and AP were generated by drawing 5,000 bootstrap samples from the estimation data set (30). If there was no additive interaction, the CIs of the RERI and AP would include 0.
We performed several sensitivity analyses to examine the robustness of the results. First, we included additional smoking-related variables in the model, including the square term of the pack-years of smoking, smoking duration, and time since quitting smoking. Second, we restricted the analysis to participants with complete covariate data for comparison with the results based on multiple imputation (31). Third, we restricted the analysis to individuals who had lived at the same address throughout the follow-up to minimize the misclassification of long-term air pollution exposures. Fourth, we excluded participants in whom lung cancer was diagnosed within the first year of follow-up to evaluate the robustness of the association. Fifth, to investigate whether the association between air pollution and lung cancer risk was significant below a priori–defined thresholds (including below the European Union air quality limit values [32] for PM2.5 [25 μg/m3], PM10 [40 μg/m3], and NO2 [40 μg/m3]), we only included participants exposed to air pollution concentrations below those thresholds.
All P values were two-sided, and P < 0.05 was considered to indicate statistical significance. All analyses were performed by using R software, version 3.6.0 (R Foundation for Statistical Computing). Multiple imputation and Cox regression were performed by using R packages, including the mice and survival packages.
During a median follow-up of 7.1 years (IQR, 6.4–7.7 yr), 1,812 lung cancer events were observed in the PM data set, and 2,020 lung cancer events were observed in the NOx data set. Table 1 shows the distribution of the baseline characteristics of the participants. A comparison of characteristics in individuals with and without missing covariates is presented in Table E1 in the online supplement. The means of the annual average estimates for PM2.5, PMcoarse, PM10, NO2, and NOx were 10.00 μg/m3 (SD, 1.06), 19.33 μg/m3 (SD, 1.95), 6.43 μg/m3 (SD, 0.90), 29.14 μg/m3 (SD, 9.11), and 44.19 μg/m3 (SD, 15.57), respectively. NOx were strongly correlated with PM2.5 (r = 0.85) but were less strongly correlated with other types of PM. There was also a strong correlation between PM10 and NO2 (r = 0.80) (Table 2).
| Characteristic | PM Data Set (n = 417,422) | NOx Data Set (n = 449,219) |
|---|---|---|
| Sex, % | ||
| Male | 46.8 (n = 195,366) | 46.7 (n = 209,741) |
| Female | 53.2 (n = 222,056) | 53.3 (n = 239,478) |
| Age, y, mean (SD) | 56.2 (8.11) | 56.2 (8.10) |
| Age, % | ||
| <60 y | 58.1 (n = 242,548) | 58.3 (n = 261,827) |
| ⩾60 y | 41.9 (n = 174,874) | 41.7 (n = 187,392) |
| BMI, kg/m2, mean (SD) | 27.5 (4.80) | 27.4 (4.80) |
| BMI, % | ||
| Normal (<25 kg/m2) | 32.8 (n = 136,154) | 32.9 (n = 146,650) |
| Overweight (25–29.9 kg/m2) | 42.6 (n = 176,787) | 42.6 (n = 190,259) |
| Obesity (⩾30 kg/m2) | 24.6 (n = 101,869) | 24.5 (n = 109,573) |
| Missing value | 2,612 | 2,737 |
| Education level, % | ||
| Degree level or professional education | 47.3 (n = 193,422) | 47.6 (n = 209,563) |
| Other levels | 52.7 (n = 215,258) | 52.4 (n = 230,568) |
| Missing value | 8,742 | 9,088 |
| Household income, % | ||
| Less than £31,000 | 47.6 (n = 168,258) | 47.4 (n = 180,557) |
| £31,000 and above | 52.4 (n = 185,019) | 52.6 (n = 200,591) |
| Missing value | 64,145 | 68,071 |
| Smoking status, % | ||
| Never-smoker | 55.1 (n = 228,851) | 55.2 (n = 246,664) |
| Current or former smoker | 44.9 (n = 186,112) | 44.8 (n = 199,959) |
| Missing value | 2,459 | 2,596 |
| Pack-years of smoking, mean (SD) | 8.1 (15.57) | 8.1 (15.66) |
| Missing value | 65,091 | 69,399 |
| Lung cancer cases,% | 0.4 (n = 1,812) | 0.4 (n = 2,020) |
| Pollution | N | Mean (SD) (μg/m3) | Minimum (μg/m3) | Maximum (μg/m3) | IQR (μg/m3) | Pearson Correlation Coefficient | ||||
|---|---|---|---|---|---|---|---|---|---|---|
| PM2.5 | PM10 | PMcoarse | NO2 | NOx | ||||||
| PM2.5 | 417,422 | 10.00 (1.06) | 8.17 | 21.31 | 1.27 | 1 | 0.65 | 0.22 | 0.73 | 0.85 |
| PM10 | 417,422 | 19.33 (1.95) | 12.87 | 30.52 | 2.35 | — | 1 | 0.53 | 0.80 | 0.65 |
| PMcoarse | 417,422 | 6.43 (0.90) | 5.57 | 12.82 | 0.81 | — | — | 1 | 0.19 | 0.24 |
| NO2 | 449,219 | 29.14 (9.11) | 8.34 | 125.12 | 10.78 | — | — | — | 1 | 0.75 |
| NOx | 449,219 | 44.19 (15.57) | 19.74 | 265.94 | 16.42 | — | — | — | — | 1 |
Long-term exposure to air pollutants significantly increased the risk of lung cancer (Table 3). The concentrations of PM2.5 (HR, 1.63; 95% CI, 1.33–2.01; per 5 μg/m3) and PM10 (HR, 1.53; 95% CI, 1.20–1.96; per 10 μg/m3) were significantly associated with incident lung cancer. However, we did not observe an association between PMcoarse and lung cancer (HR, 1.01; 95% CI, 0.78–1.30; per 5 μg/m3). Accounting for the fact that PM10 is composed of PMcoarse and PM2.5, we further adjusted for the PM2.5 concentration and found that the association between PM10 and lung cancer was weakened to the point of becoming nonsignificant (HR, 1.11; 95% CI, 0.81–1.52; per 10 μg/m3). Furthermore, we observed that the risk of lung cancer also increased with increasing NO2 (HR, 1.10; 95% CI, 1.05–1.15; per 10 μg/m3) and NOx (HR, 1.13; 95% CI, 1.07–1.18; per 20 μg/m3) concentrations. Similar associations were also observed in the analysis performed per IQR increase in air pollutant concentrations (see Table E2 in the online supplement). The risk of lung cancer increased significantly from quartile 1 to quartile 4 for PM2.5, PM10, NO2, and NOx (P for trend [Ptrend] < 0.001; Table E3 and Figure E1). Stratified analyses showed that the effects of ambient air pollutant exposure on lung cancer were generally similar across different subgroups (Table E4).
| Pollution | Increase | Number of Cases/Person-Years | Model 1* | Model 2† | ||
|---|---|---|---|---|---|---|
| HR (95% CI) | P Value | HR (95% CI) | P Value | |||
| PM2.5 | 5 μg/m3 | 1812/2932721 | 2.68 (2.66–2.70) | <2 × 10−16 | 1.63 (1.33–2.01) | 3.68 × 10−6 |
| PM10 | 10 μg/m3 | 1812/2932721 | 2.14 (1.69–2.71) | 2.82 × 10−10 | 1.53 (1.20–1.96) | 0.001 |
| PMcoarse | 5 μg/m3 | 1812/2932721 | 1.19 (0.93–1.53) | 0.159 | 1.01 (0.78–1.30) | 0.648 |
| NO2 | 10 μg/m3 | 2020/3176143 | 1.19 (1.14–1.25) | 2.40 × 10−14 | 1.10 (1.05–1.15) | 1.55 × 10−4 |
| NOx | 20 μg/m3 | 2020/3176143 | 1.24 (1.18–1.29) | <2 × 10−16 | 1.13 (1.07–1.18) | 3.02 × 10−6 |
In the sensitivity analyses, further adjustment for smoking-related covariates did not significantly alter the associations with lung cancer (Table E5). We noted no significant differences in the HRs for lung cancer before and after the exclusion of participants with incomplete covariate data (Table E6). Restriction of the analysis to participants who had a fixed home address throughout follow-up yielded comparatively stronger associations with incident lung cancer (Table E7). The observed associations between air pollution and lung cancer risk did not change significantly after the exclusion of participants who received a lung cancer diagnosis within the first year of follow-up (Table E8). Restriction of the participants to those exposed to air pollution below several predefined thresholds for PM2.5, PM10, and NO2 concentrations also yielded similar HRs (Table E9).
The 18 SNPs used to construct the PRS are shown in Table E10. Participants with incident lung cancer tended to have a higher PRS (genetic risk) than unaffected participants in the analysis data sets for PM and NOx (Figure E2). We observed a significantly increase in the risk for lung cancer across the deciles of the PRS (Ptrend < 0.001; Figure E2), which did not change with additional adjustment for pollutant concentrations (Ptrend < 0.001; Table E11). Participants in the high genetic risk category (top tertile) had a 54% (PM data set: HR, 1.54; 95% CI, 1.37–1.74) and 50% (NOx data set: HR, 1.50; 95% CI, 1.34–1.67) higher risk of lung cancer than those in the low genetic risk category, respectively. These results did not significantly change after additional adjustment for pollutant concentrations (Table E12).
We observed a joint effect of genetic factors and air pollution on lung cancer risk that behaved in a dose–response manner; that is, the overall risk of lung cancer increased as both genetic risk and air pollution exposure increased (Figure 2). Specifically, compared with participants with a low genetic risk and low air pollution exposure, participants with a high genetic risk and high air pollution exposure had the highest risk of incident lung cancer (PM2.5: HR, 1.71; 95% CI, 1.45–2.02; PM10: HR, 1.77; 95% CI, 1.50–2.10; NO2: HR, 1.77; 95% CI, 1.42–2.22; NOx: HR, 1.67; 95% CI, 1.43–1.95).
Figure 2. Risk of incident lung cancer according to air pollution and genetic categories. The hazard ratios for lung cancer according to (A) PM2.5, (B) PM10, (C) NO2, (D) NOx, and polygenic risk score categories were estimated by using Cox proportional hazard models with adjustment for age, sex, body mass index, household income, education level, smoking status, pack-years of smoking, and the top 40 principal components of ancestry and genotyping batch. CI = confidence interval; NO2 = nitrogen dioxide; NOx = nitrogen oxides; PM2.5 = particulate matter ⩽2.5 μm in aerodynamic diameter; PM10 = particulate matter with diameter ⩽10 μm in aerodynamic diameter.
[More] [Minimize]The RERI and AP were significant, which indicated positive additive interactions of air pollutants, including PM2.5, NO2, and NOx, with the PRS (Table 4). Specifically, for high PM2.5 exposure with a high PRS, the RERI was 0.37 (95% CI, 0.24–0.49), which suggested that there would be a 0.37 relative excess risk because of the additive interaction, accounting for 21% (14–29%) of the risk of lung cancer in participants exposed to both high genetic risk and high PM2.5 exposure.
| Pollution Category | PRS* | |||
|---|---|---|---|---|
| Intermediate | High | |||
| RERI† (95% CI) | AP† (95% CI) | RERI (95% CI) | AP (95% CI) | |
| PM2.5‡ | ||||
| High pollution | 0.36 (0.25 to 0.48) | 0.26 (0.18 to 0.34) | 0.37 (0.24 to 0.49) | 0.21 (0.14 to 0.29) |
| PM10§ | ||||
| High pollution | −0.03 (−0.17 to 0.11) | −0.02 (−0.13 to 0.08) | 0.11 (−0.04 to 0.26) | 0.06 (−0.03 to 0.15) |
| NO2‖ | ||||
| High pollution | 0.07 (−0.13 to 0.27) | 0.05 (−0.11 to 0.20) | 0.26 (0.06 to 0.51) | 0.15 (0.03 to 0.27) |
| NOx¶ | ||||
| High pollution | 0.33 (0.23 to 0.43) | 0.26 (0.18 to 0.34) | 0.53 (0.43 to 0.64) | 0.32 (0.26 to 0.38) |
In this study, we observed that residential exposure to PM and NOx pollution at enrollment was significantly associated with an increased (∼63%) risk of lung cancer in the UK Biobank. High genetic risk was associated with an approximately 50% higher risk of lung cancer. Furthermore, when examining the joint effects of genetic risk and air pollution, we found that the greatest relative increase in risk was observed among those with high air pollution exposure levels and high genetic risk. Our study also provides quantitative data about the effect of the additive interaction between genetic factors and air pollution on lung cancer.
In 2013, the International Agency for Research on Cancer classified PM pollution as a group 1 carcinogen (33), and several epidemiologic studies have confirmed the relationship between air pollution and lung cancer–related mortality (14, 34–38). However, the conclusions of different population-based studies about incident lung cancer were not consistent. Studies in the United States (10) (HR, 1.43; 95% CI, 1.11–1.84; per 10 μg/m3) and Canada (9) (HR, 1.34; 95% CI, 1.10–1.65; per 10 μg/m3) found significant associations between PM2.5 and incident lung cancer, whereas the NHS (Nurses’ Health Study) cohort (39) in the United States and the NLCS (Netherlands Cohort Study on Diet and Cancer) (7, 12) did not observe significant associations between PM and lung cancer risk, and the ESCAPE study (8) only found an association with PM10 (HR, 1.22; 95% CI, 1.03–1.45; per 10 μg/m3). In this large prospective study with 0.5 million people, our results confirmed the associations of different PM exposure components, including PM2.5 and PM10, with incident lung cancer (PM2.5: HR, 1.63; 95% CI, 1.33–2.01; per 5 μg/m3; PM10: HR, 1.53; 95% CI, 1.20–1.96; per 10 μg/m3). These varied effect sizes might be due, at least in part, to the method used to assess air pollutant concentrations. Furthermore, we observed that the association between PM10 and lung cancer risk was not significant after additional adjustment for the PM2.5 concentration, which suggests that PM2.5 is more relevant to the development of lung cancer. Compared with PM10, PM2.5 has a greater total surface area and a porous surface, so it can adsorb and retain more compounds, such as carcinogenic organic species and heavy metals, which are more toxic than sole particles (40). Furthermore, PM2.5 can be deposited deeper in the lung, and the molecules adsorbed on the surface can be released into the alveoli and be retained for longer in the lung parenchyma (41).
A significant increase in lung cancer risk was also associated with exposure to NO2 (HR, 1.06; 95% CI, 1.01–1.11; per 10 μg/m3) in the present study, which is similar to the results from both the NLCS (12) (HR, 1.29; 95% CI, 1.08–1.54; per 30 μg/m3) and a meta-analysis including 20 studies around the world (42) (HR, 1.04; 95% CI, 1.01–1.08; per 10 μg/m3). However, we need to interpret this finding more carefully because NOx are only an indicator of exposure to motor vehicle exhaust and the prior evidence for them being carcinogenic is weak. Moreover, given the strong correlation between NOx and PM2.5 in this study, it is hard to distinguish the effect of NOx from the effect of PM2.5, and more studies are needed to further explore the exact effect of NOx on lung carcinogenesis.
It is noteworthy that, below most predefined threshold values, the HRs for lung cancer in participants with or without restrictions were similar, which is consistent with the findings of the ESCAPE study (8). Furthermore, below the thresholds, an apparent linear relationship was observed between air pollution exposure and lung cancer risk. These findings suggest that exposure to air pollution, even at concentrations below the existing European Union air quality thresholds, might still increase lung cancer risk, which indicates that the formulation of more rigorous environmental health policies to prevent air pollution may lead to a reduction in incident lung cancer. In addition, we found that the associations between the risk of lung cancer and air pollution were similar across all strata of genetic risk, supporting the benefit of improving air quality for the entire population.
Although previous studies explored the interaction between air pollution and genetic factors on lung cancer (43), to our knowledge, the present study is the first to assess possible additive interactions. The results revealed that high genetic risk and high air pollution exposure synergistically increased the risk of lung cancer. Moreover, 15–32% of lung cancer risk could be attributed to the additive interactions, suggesting that the interactive effect of high genetic risk and high air pollution exposure was greater than the sum of the two individual effects. Given these results, it can be speculated that air pollution modifies the impact of genetic susceptibility on lung cancer. Furthermore, the observed additive interactions between genetic factors and air pollution have public health implications because they can be used to identify people who are more likely to benefit from targeted interventions designed to reduce air pollution (44, 45). Specifically, in this study, people with high genetic risk should pay more attention to air pollution. Further studies will be needed to confirm our findings.
Strengths of our study included the collection of a large sample, use of uniform data collection protocols, inclusion of information about potential confounders, and use of standardized individual exposure assessments. In addition, we assessed the contribution of genetic factors to the association between air pollution and lung cancer, enabling us to precisely determine the effects of air pollutants on groups with varying levels of susceptibility. However, the present study has several limitations. First, air pollution is a dynamic and complex mixture of substances and contains many anthropogenic and natural pollutants with carcinogenic potential (3). It is difficult to disentangle the effects of individual components. Second, a single measurement of air pollution at baseline was available in the UK Biobank, which did not take into account changes in air pollution before and after enrollment. Future cohort studies with multiple measurements of air pollution at different time are needed to explore the effect of air pollution changes on lung cancer. Third, occupational exposure is an additional potential confounder, but it was not assessed in the UK Biobank, which could result in potential residual confounding. Fourth, smoking status was oversimplified, but adjustment for other smoking-related variables did not result in significant difference. Fifth, biomarkers of oxidative damage were not measured, and we could not confirm our findings mechanistically. Finally, multiple imputation was used to impute missing covariates, which might have led to the partial deviation of our estimates from the true values, even though we observed no significant difference after the exclusion of participants with incomplete covariate data.
In conclusion, this large prospective cohort study demonstrated that exposure to residential ambient air pollution plays an important role in the development of lung cancer and can enhance the deleterious effect of predisposing genetic factors. It is important to formulate policies aimed at improving air quality and decreasing exposure to air pollution, especially for those with a high genetic risk.
| 1. | Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2018;68:394–424. |
| 2. | Islami F, Goding Sauer A, Miller KD, Siegel RL, Fedewa SA, Jacobs EJ, et al. Proportion and number of cancer cases and deaths attributable to potentially modifiable risk factors in the United States. CA Cancer J Clin 2018;68:31–54. |
| 3. | Kurt OK, Zhang J, Pinkerton KE. Pulmonary health effects of air pollution. Curr Opin Pulm Med 2016;22:138–143. |
| 4. | DeMarini DM. Genotoxicity biomarkers associated with exposure to traffic and near-road atmospheres: a review. Mutagenesis 2013;28:485–505. |
| 5. | Møller P, Danielsen PH, Karottki DG, Jantzen K, Roursgaard M, Klingberg H, et al. Oxidative stress and inflammation generated DNA damage by exposure to air pollution particles. Mutat Res Rev Mutat Res 2014;762:133–166. |
| 6. | Wang Y, Zhang M, Li Z, Yue J, Xu M, Zhang Y, et al. Fine particulate matter induces mitochondrial dysfunction and oxidative stress in human SH-SY5Y cells. Chemosphere 2019;218:577–588. |
| 7. | Beelen R, Hoek G, van den Brandt PA, Goldbohm RA, Fischer P, Schouten LJ, et al. Long-term exposure to traffic-related air pollution and lung cancer risk. Epidemiology 2008;19:702–710. |
| 8. | Raaschou-Nielsen O, Andersen ZJ, Beelen R, Samoli E, Stafoggia M, Weinmayr G, et al. Air pollution and lung cancer incidence in 17 European cohorts: prospective analyses from the European Study of Cohorts for Air Pollution Effects (ESCAPE). Lancet Oncol 2013;14:813–822. |
| 9. | Tomczak A, Miller AB, Weichenthal SA, To T, Wall C, van Donkelaar A, et al. Long-term exposure to fine particulate matter air pollution and the risk of lung cancer among participants of the Canadian National Breast Screening Study. Int J Cancer 2016;139:1958–1966. |
| 10. | Gharibvand L, Shavlik D, Ghamsary M, Beeson WL, Soret S, Knutsen R, et al. The association between ambient fine particulate air pollution and lung cancer incidence: results from the AHSMOG-2 study. Environ Health Perspect 2017;125:378–384. |
| 11. | Hystad P, Demers PA, Johnson KC, Carpiano RM, Brauer M. Long-term residential exposure to air pollution and lung cancer risk. Epidemiology 2013;24:762–772. |
| 12. | Hart JE, Spiegelman D, Beelen R, Hoek G, Brunekreef B, Schouten LJ, et al. Long-term ambient residential traffic-related exposures and measurement error-adjusted risk of incident lung cancer in the Netherlands Cohort Study on Diet and Cancer. Environ Health Perspect 2015;123:860–866. |
| 13. | Nyberg F, Gustavsson P, Järup L, Bellander T, Berglind N, Jakobsson R, et al. Urban air pollution and lung cancer in Stockholm. Epidemiology 2000;11:487–495. |
| 14. | Pope CA III, Burnett RT, Thun MJ, Calle EE, Krewski D, Ito K, et al. Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA 2002;287:1132–1141. |
| 15. | Lichtenstein P, Holm NV, Verkasalo PK, Iliadou A, Kaprio J, Koskenvuo M, et al. Environmental and heritable factors in the causation of cancer: analyses of cohorts of twins from Sweden, Denmark, and Finland. N Engl J Med 2000;343:78–85. |
| 16. | Bossé Y, Amos CI. A decade of GWAS results in lung cancer. Cancer Epidemiol Biomarkers Prev 2018;27:363–379. |
| 17. | Dai J, Lv J, Zhu M, Wang Y, Qin N, Ma H, et al. Identification of risk loci and a polygenic risk score for lung cancer: a large-scale prospective cohort study in Chinese populations. Lancet Respir Med 2019;7:881–891. |
| 18. | Wu C, Kraft P, Zhai K, Chang J, Wang Z, Li Y, et al. Genome-wide association analyses of esophageal squamous cell carcinoma in Chinese identify multiple susceptibility loci and gene-environment interactions. Nat Genet 2012;44:1090–1097. |
| 19. | Nan H, Hutter CM, Lin Y, Jacobs EJ, Ulrich CM, White E, et al.; CCFR; GECCO. Association of aspirin and NSAID use with risk of colorectal cancer according to genetic variants. JAMA 2015;313:1133–1142. |
| 20. | Chen K-Y, Hsiao C-F, Chang G-C, Tsai Y-H, Su W-C, Chen Y-M, et al.; GELAC Study Group. Estrogen receptor gene polymorphisms and lung adenocarcinoma risk in never-smoking women. J Thorac Oncol 2015;10:1413–1420. |
| 21. | Collins R. What makes UK Biobank special? Lancet 2012;379:1173–1174. |
| 22. | Eeftens M, Beelen R, de Hoogh K, Bellander T, Cesaroni G, Cirach M, et al. Development of land use regression models for PM2.5, PM2.5 absorbance, PM10 and PMcoarse in 20 European study areas; results of the ESCAPE project. Environ Sci Technol 2012;46:11195–11205. |
| 23. | Beelen R, Hoek G, Vienneau D, Eeftens M, Dimakopoulou K, Pedeli X, et al. Development of NO2 and NOx land use regression models for estimating air pollution exposure in 36 study areas in Europe: the ESCAPE project. Atmos Environ 2013;72:10–23. |
| 24. | World Health Organization Regional Office for Europe. Air quality guidelines: global update 2005. Particulate matter, ozone, nitrogen dioxide and sulfur dioxide. Geneva: World Health Organization; 2006. |
| 25. | McKay JD, Hung RJ, Han Y, Zong X, Carreras-Torres R, Christiani DC, et al.; SpiroMeta Consortium. Large-scale association analysis identifies new lung cancer susceptibility loci and heterogeneity in genetic susceptibility across histological subtypes. Nat Genet 2017;49:1126–1132. |
| 26. | Arthur RS, Wang T, Xue X, Kamensky V, Rohan TE. Genetic factors, adherence to healthy lifestyle behavior, and risk of invasive breast cancer among women in the UK Biobank. J Natl Cancer Inst 2020;112:893–901. |
| 27. | World Health Organization. ICD-10: international statistical classification of diseases and related health problems. Geneva: World Health Organization; 2004. |
| 28. | The effects of taxes and benefits on household income, disposable income estimate: 2019. London, UK: Office for National Statistics; 2020 [created 2020 Mar 5; accessed 2020 Aug 2]. Available from: https://www.ons.gov.uk/peoplepopulationandcommunity/personalandhouseholdfinances/incomeandwealth/datasets/householddisposableincomeandinequality. |
| 29. | Li R, Chambless L. Test for additive interaction in proportional hazards models. Ann Epidemiol 2007;17:227–236. |
| 30. | Assmann SF, Hosmer DW, Lemeshow S, Mundt KA. Confidence intervals for measures of interaction. Epidemiology 1996;7: 286–290. |
| 31. | Sterne JAC, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ 2009;338:b2393. |
| 32. | Brunekreef B, Künzli N, Pekkanen J, Annesi-Maesano I, Forsberg B, Sigsgaard T, et al. Clean air in Europe: beyond the horizon? Eur Respir J 2015;45:7–10. |
| 33. | Loomis D, Grosse Y, Lauby-Secretan B, El Ghissassi F, Bouvard V, Benbrahim-Tallaa L, et al.; International Agency for Research on Cancer Monograph Working Group IARC. The carcinogenicity of outdoor air pollution. Lancet Oncol 2013;14:1262–1263. |
| 34. | Turner MC, Andersen ZJ, Baccarelli A, Diver WR, Gapstur SM, Pope CA III, et al. Outdoor air pollution and cancer: an overview of the current evidence and public health recommendations. CA Cancer J Clin 2020;70:460–479. |
| 35. | Jerrett M, Burnett RT, Beckerman BS, Turner MC, Krewski D, Thurston G, et al. Spatial analysis of air pollution and mortality in California. Am J Respir Crit Care Med 2013;188:593–599. |
| 36. | Turner MC, Krewski D, Pope CA III, Chen Y, Gapstur SM, Thun MJ. Long-term ambient fine particulate matter air pollution and lung cancer in a large cohort of never-smokers. Am J Respir Crit Care Med 2011;184:1374–1381. |
| 37. | Dockery DW, Pope CA III, Xu X, Spengler JD, Ware JH, Fay ME, et al. An association between air pollution and mortality in six U.S. cities. N Engl J Med 1993;329:1753–1759. |
| 38. | Li J, Lu X, Liu F, Liang F, Huang K, Yang X, et al. Chronic effects of high fine particulate matter exposure on lung cancer in China. Am J Respir Crit Care Med 2020;202:1551–1559. |
| 39. | Puett RC, Hart JE, Yanosky JD, Spiegelman D, Wang M, Fisher JA, et al. Particulate matter air pollution exposure, distance to road, and incident lung cancer in the nurses’ health study cohort. Environ Health Perspect 2014;122:926–932. |
| 40. | Losacco C, Perillo A. Particulate matter air pollution and respiratory impact on humans and animals. Environ Sci Pollut Res Int 2018;25:33901–33910. |
| 41. | Falcon-Rodriguez CI, Osornio-Vargas AR, Sada-Ovalle I, Segura-Medina P. Aeroparticles, composition, and lung diseases. Front Immunol 2016;7:3. |
| 42. | Hamra GB, Laden F, Cohen AJ, Raaschou-Nielsen O, Brauer M, Loomis D. Lung cancer and exposure to nitrogen dioxide and traffic: a systematic review and meta-analysis. Environ Health Perspect 2015;123:1107–1112. |
| 43. | Hosgood HD III, Song M, Hsiung CA, Yin Z, Shu X-O, Wang Z, et al. Interactions between household air pollution and GWAS-identified lung cancer susceptibility markers in the Female Lung Cancer Consortium in Asia (FLCCA). Hum Genet 2015;134:333–341. |
| 44. | VanderWeele TJ, Mirjam JK. A tutorial on interaction. Epidemiol Methods 2014;3:37–72. |
| 45. | Joshi AD, Lindström S, Hüsing A, Barrdahl M, VanderWeele TJ, Campa D, et al.; Breast and Prostate Cancer Cohort Consortium (BPC3). Additive interactions between susceptibility single-nucleotide polymorphisms identified in genome-wide association studies and breast cancer risk factors in the Breast and Prostate Cancer Cohort Consortium. Am J Epidemiol 2014;180:1018–1027. |
*These authors contributed equally to this work
Supported by the National Natural Science Foundation of China Integration Project (91943301); the Natural Science Foundation of Jiangsu Province (BK20180675); the National Natural Science Foundation of China (81922061, 81803306, 81973123, and 81820108028); the Research Unit of Prospective Cohort of Cardiovascular Diseases and Cancers, Chinese Academy of Medical Sciences (2019RU038); and the National Science Foundation for Post-doctoral Scientists of China (2018M640466).
Author Contributions: H.S. and H.M. supervised the entire project and designed the work. Y.H. and M.Z. contributed to the data analysis, data interpretation, and writing of the report. M.J., J.F., J. Xie, X.W., X.J., J. Xu, L.C., R.Y., Y.W., J.D., G.J., L.X., and Z.H. contributed to the discussion and data interpretation and revised the manuscript. All authors reviewed or revised the manuscript and approved the final draft for submission.
This is a corrected version of the article; it was updated when the May 15, 2022, issue was posted online. See erratum: Am J Respir Care Med 2022;205:1254; https://www.atsjournals.org/doi/full/10.1164/rccm.v205erratum2.
This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org.
Originally Published in Press as DOI: 10.1164/rccm.202011-4063OC on July 12, 2021
Author disclosures are available with the text of this article at www.atsjournals.org.
