Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2019 Sep;127(9):97001.
doi: 10.1289/EHP4913. Epub 2019 Sep 5.

Constituents of Household Air Pollution and Risk of Lung Cancer among Never-Smoking Women in Xuanwei and Fuyuan, China

Roel Vermeulen  1 , George S Downward  1 , Jinming Zhang  2 , Wei Hu  2 , Lützen Portengen  1 , Bryan A Bassig  2 , S Katharine Hammond  3 , Jason Y Y Wong  2 , Jihua Li  4 , Boris Reiss  5 , Jun He  4 , Linwei Tian  6 , Kaiyun Yang  7 , Wei Jie Seow  8   9 , Jun Xu  10 , Kim Anderson  11 , Bu-Tian Ji  2 , Debra Silverman  2 , Stephen Chanock  2 , Yunchao Huang  7 , Nathaniel Rothman  2 , Qing Lan  2
Affiliations

Constituents of Household Air Pollution and Risk of Lung Cancer among Never-Smoking Women in Xuanwei and Fuyuan, China

Roel Vermeulen et al. Environ Health Perspect. 2019 Sep.

Abstract

Background: Lung cancer rates among never-smoking women in Xuanwei and Fuyuan in China are among the highest in the world and have been attributed to the domestic use of smoky (bituminous) coal for heating and cooking. However, the key components of coal that drive lung cancer risk have not been identified.

Objectives: We aimed to investigate the relationship between lifelong exposure to the constituents of smoky coal (and other fuel types) and lung cancer.

Methods: Using a population-based case-control study of lung cancer among 1,015 never-smoking female cases and 485 controls, we examined the association between exposure to 43 household air pollutants and lung cancer. Pollutant predictions were derived from a comprehensive exposure assessment study, which included methylated polycyclic aromatic hydrocarbons (PAHs), which have never been directly evaluated in an epidemiological study of any cancer. Hierarchical clustering and penalized regression were applied in order to address high colinearity in exposure variables.

Results: The strongest association with lung cancer was for a cluster of 25 PAHs [odds ratio (OR): 2.21; 95% confidence interval (CI): 1.67, 2.87 per 1 standard deviation (SD) change], within which 5-methylchrysene (5-MC), a mutagenic and carcinogenic PAH, had the highest individual observed OR (5.42; 95% CI: 0.94, 27.5). A positive association with nitrogen dioxide ([Formula: see text]) was also observed (OR: 2.06; 95% CI: 1.19, 3.49). By contrast, neither benzo(a)pyrene (BaP) nor fine particulate matter with aerodynamic diameter [Formula: see text] ([Formula: see text]) were associated with lung cancer in the multipollutant models.

Conclusions: To our knowledge, this is the first study to comprehensively evaluate the association between lung cancer and household air pollution (HAP) constituents estimated over the entire life course. Given the global ubiquity of coal use domestically for indoor cooking and heating and commercially for electric power generation, our study suggests that more extensive monitoring of coal combustion products, including methylated PAHs, may be warranted to more accurately assess health risks and develop prevention strategies from this exposure. https://doi.org/10.1289/EHP4913.

PubMed Disclaimer

Figures

Figure 1A plots cluster of pollutants, namely, WB 7, S O 2, P A H 2, P A H 7, P A H 25, and N O 2, and 1B plots cluster of pollutants, namely, retene, WB 2, P A H 5 plus, P A H 2, and P A H 29 plus, on the y-axis, across odds ratios (95 percent confidence intervals) 0.1, 0.2, 0.5, 1, 2, 5, and 10 on the x-axis.
Figure 1.
Odds ratios (ORs) per 1 standard deviation increase in exposure for different exposure clusters in the full population (A) and the smoky coal subpopulation (B). Exposure clusters were polycyclic aromatic hydrocarbon (PAH) 25 [consisting of 25 PAHs, including benzo(a)pyrene (BaP) and 5-methylchrysene (5-MC)], WB7 [consisting of seven wood burning–associated exposures, including fine particulate matter with aerodynamic diameter ≤2.5μm (PM2.5)], PAH7, PAH2, and two clusters consisting of the single exposures, nitrogen dioxide (NO2) and sulfur dioxide (SO2), respectively (See also Table S2 for complete listing of compounds in each cluster). Note: CI, confidence interval.
Figure 2 is a forest plot showing odds ratios (95 percent confidence intervals) 0.002, 0.02, 0.1, 0.5, 2, 5, 20, 100, and 500 (x-axis) in exposure for individual pollutants (y-axis) belonging to the following clusters: WB 7, S O 2, P A H 2, P A H 7, P A H 25, and N O 2.
Figure 2.
Odds ratios per 1 standard deviation increase in exposure for individual pollutants and lung cancer in the full population.
Figure 3 is a forest plot showing odds ratios (95 percent confidence intervals) 0.002, 0.02, 0.1, 0.5, 2, 5, 20, 100, and 500 (x-axis) in exposure for individual pollutants (y-axis) belonging to the following clusters: retene, WB 2, P A H 5 positive, P A H 2, P A H 29 positive.
Figure 3.
Odds ratios for individual pollutants and lung cancer in the smoky coal subpopulation. Estimates were rescaled to reflect the estimated effect for a - standard deviation increase in exposure in the full population to allow direct comparison with estimated effects in Figure 2. The (rescaled) point estimate for retene was off the scale (indicated by the “<” symbol).
Figure 4A is a line graph plotting values for sensitivity (y-axis) from 0.0 to 1.0 in increments of 0.2 across values for specificity (x-axis) from 1.0 to 0.0 in intervals of negative 0.2 for most frequently used deposit (AUC equals 72 percent), exposures (AUC equals 79 percent), and exposures plus deposit (AUC equals 81 percent). The trend of the line is undulating but increasing. Figure 4B is a scatter plot with a regression line plotting prediction from model with exposures plus deposit information (y-axis) from negative 2 to 4 in increments of 2 across predictions from exposures-only model (x-axis) from negative 2 to 4 in increments of 2. The trend of the line is increasing. Figure 4C is line graph plotting values for sensitivity (y-axis) from 0.0 to 1.0 in increments of 0.2 across values for specificity (x-axis) from 1.0 to 0.0 in intervals of negative 0.2 for most frequently used deposit (AUC equals 57 percent), exposures (AUC equals 72 percent), and exposures plus deposit (AUC equals 72 percent). The trend of the line is undulating but increasing. Figure 4D is a scatter plot with a regression line plotting prediction from model with exposures plus deposit information (y-axis) from negative 2 to 6 in increments of 2 across predictions from exposures-only model (x-axis) from negative 2 to 6 in increments of 2. The trend of the line is increasing.
Figure 4.
Receiver operating characteristic (ROC) curve analyses and model prediction quality of logistic models for household air pollutant (HAP) exposures vs. coal deposit layers and risk of lung cancer. Models were fitted to data from the full study population (A,B) or the smoky coal subpopulation (C,D). (A,C) show a comparison of the area under the curve (AUC) for lung cancer case and control predictions from a model including only the exposure estimates, a model including only information on the most frequently used deposit, and a model including both. ROC curves and corresponding AUC were estimated using cross validation. (B,D) show a comparison of model predictions for predictions based on an exposures-only model vs. a model with exposures+deposit information. The location of the referenced deposits is indicated in Figure S1. Predictions for deposit information is derived from logistic regression–adjusted forage and food sufficiency before marriage or 20 y of age and reported by Wong et al. (2019).
Figure 4A is a line graph plotting values for sensitivity (y-axis) from 0.0 to 1.0 in increments of 0.2 across values for specificity (x-axis) from 1.0 to 0.0 in intervals of negative 0.2 for most frequently used deposit (AUC equals 72 percent), exposures (AUC equals 79 percent), and exposures plus deposit (AUC equals 81 percent). The trend of the line is undulating but increasing. Figure 4B is a scatter plot with a regression line plotting prediction from model with exposures plus deposit information (y-axis) from negative 2 to 4 in increments of 2 across predictions from exposures-only model (x-axis) from negative 2 to 4 in increments of 2. The trend of the line is increasing. Figure 4C is line graph plotting values for sensitivity (y-axis) from 0.0 to 1.0 in increments of 0.2 across values for specificity (x-axis) from 1.0 to 0.0 in intervals of negative 0.2 for most frequently used deposit (AUC equals 57 percent), exposures (AUC equals 72 percent), and exposures plus deposit (AUC equals 72 percent). The trend of the line is undulating but increasing. Figure 4D is a scatter plot with a regression line plotting prediction from model with exposures plus deposit information (y-axis) from negative 2 to 6 in increments of 2 across predictions from exposures-only model (x-axis) from negative 2 to 6 in increments of 2. The trend of the line is increasing.
Figure 4.
Receiver operating characteristic (ROC) curve analyses and model prediction quality of logistic models for household air pollutant (HAP) exposures vs. coal deposit layers and risk of lung cancer. Models were fitted to data from the full study population (A,B) or the smoky coal subpopulation (C,D). (A,C) show a comparison of the area under the curve (AUC) for lung cancer case and control predictions from a model including only the exposure estimates, a model including only information on the most frequently used deposit, and a model including both. ROC curves and corresponding AUC were estimated using cross validation. (B,D) show a comparison of model predictions for predictions based on an exposures-only model vs. a model with exposures+deposit information. The location of the referenced deposits is indicated in Figure S1. Predictions for deposit information is derived from logistic regression–adjusted forage and food sufficiency before marriage or 20 y of age and reported by Wong et al. (2019).
See this image and copyright information in PMC

Comment in

References

    1. Agier L, Portengen L, Chadeau-Hyam M, Basagaña X, Giorgis-Allemand L, Siroux V, et al. . 2016. A systematic comparison of linear regression-based statistical methods to assess exposome-health associations. Environ Health Perspect 124(12):1848–1856, PMID: 27219331, 10.1289/EHP172. - DOI - PMC - PubMed
    1. Barone-Adesi F, Chapman RS, Silverman DT, He X, Hu W, Vermeulen R, et al. . 2012. Risk of lung cancer associated with domestic use of coal in Xuanwei, China: retrospective cohort study. BMJ 345:e5414, PMID: 22936785, 10.1136/bmj.e5414. - DOI - PMC - PubMed
    1. Betancourt M. 2018. A Conceptual Introduction to Hamiltonian Monte Carlo. https://arxiv.org/pdf/1701.02434.pdf, [accessed 15 December 2018].
    1. Box GEP, Cox DR. 1964. An analysis of transformations. J Roy Stat Soc B 26:211–243, 10.1111/j.2517-6161.1964.tb00553.x. - DOI
    1. Brooks SP, Gelman A. 1998. General methods for monitoring convergence of iterative simulations. J Comput Graph Stat 7(4):434–455, 10.2307/1390675. - DOI

Publication types

LinkOut - more resources