From the National Cancer Institute, Division of Cancer Epidemiology and Genetics, Bethesda, MD.
Supported by the Intramural Research Program of the NIH (Division of Cancer Epidemiology and Genetics, NCI).
Address correspondence to: Aaron Blair, National Cancer Institute, Division of Cancer Epidemiology and Genetics, EPS Room 8008, Bethesda, MD 20892. E-mail: firstname.lastname@example.org.
Although lung cancer deaths among nonsmokers are often viewed as a rare event, there are 17,000 to 26,000 such deaths annually in the United States.1 This would rank them in the top 10 for deaths from cancer if they were considered a separate cancer site.2 The recent death of Dana Reeve, wife of actor Christopher Reeve and a lifelong nonsmoker, from lung cancer has generated considerable scientific and public interest in the causes of this disease among nonsmokers.3 The article by Zeka and colleagues4 in this issue focuses on the identification of occupational risk factors for lung cancer among nonsmokers. This timely article raises a number of scientific issues regarding the development of lung cancer in nonsmokers, including what we know about its etiology—specifically, the contribution of occupational factors and how to identify additional risk factors.
What are the Risk Factors for Lung Cancer Among Nonsmokers and do They Differ From Those Among Smokers?
Although the literature on the etiology of lung cancer among nonsmokers is small compared with that among smokers, it appears that the factors affecting the development of lung cancer in smokers and nonsmokers are quite similar. Overall, approximately 25% of the nonsmoking cases can be explained by established risk factors.5 These factors include indoor air pollution,6 outdoor air pollution,7 diet,8 previous lung disease,9 radon in the home,10 environmental tobacco smoke,11 physical inactivity,12 and occupation.4 Host factors such as reported family history13 and genetic polymorphisms14 are also linked to lung cancer among nonsmokers as they are among smokers.
Lung cancer rates among nonsmokers provide important information regarding the etiology of this disease. Findings from the American Cancer Society Cancer Prevention Cohorts indicate that mortality rates for lung cancer among nonsmokers were slightly higher among men than women (17.1 vs 14.7 per 100,000) and appear to have been relatively stable over the last 4 decades.1 This small difference does not suggest any huge disparity in exposure to nontobacco risk factors between men and women. It is roughly in the range that might be due to differences between men and women in occupational exposures that have been linked with lung cancer. The relatively stable rates over time suggest no large changes over the past few decades in the overall level of exposure to environmental risk factors associated with lung cancer. Individual exposures could, of course, change without affecting the overall lung cancer mortality rate if there were a balance between increases and decreases in various risk factors.
The similarity in risk factors, other than active smoking, for lung cancer among smokers and nonsmokers suggests that risk factors for lung cancer identified among smokers should be also considered risk factors among nonsmokers. If this assumption is correct, then it would not appear necessary to replicate each finding among smokers with a study among nonsmokers before making recommendations and undertaking preventive actions among nonsmokers.
Are Occupational Exposures Important?
Zeka et al4 conclude that occupational exposures play a limited role in lung cancer risk in their study of nonsmokers. Although they do not quantify “limited role,” it seems not inconsistent with Axelson's15 attributable risk estimates of 15% to 25% for lung cancer from occupational exposures when smoking individuals are also included. This might be viewed, however, as a more important contribution to lung cancer among nonsmokers when recognizing that only approximately 25% can be explained so far.5 It is important to note that currently no established nontobacco risk factor explains a large percentage of lung cancer cases.
How can we Best Identify Lung Cancer Risk Factors Among Nonsmokers?
Studies restricted to nonsmokers such as the one by Zeka et al4 are designed to eliminate active smoking as a possible confounder in the evaluation of risk factors for lung cancer. This approach assumes that studies among nonsmokers provide the best evidence that a nontobacco factor causes a tobacco-related disease. This is a view held by many epidemiologists despite the fact that most evidence regarding the role of established nontobacco risk factors for lung cancer comes from studies that include smokers.
Although it seems like a good idea to restrict studies to nonsmokers, there are several limitations with this approach. A major appeal of studies among nonsmokers is the belief that this may be the only way to confidently eliminate the possibility of confounding from active smoking. It is not clear, however, that confounding in occupational studies (and we suspect for nonoccupational studies also) by active smoking is a very common event.16 Even if confounding by active smoking is a problem, it would not necessarily be eliminated by studying individuals who reported they were nonsmokers, because self-report of smoking status is subject to error. Studies find that 15% to 20% of self-reported nonsmokers have cotinine evels in body fluids in a range typically associated with active smoking.17–20 This is a fairly sizable level of misclassification of smoking status, and one can imagine that misclassification of historical smoking status would be even worse.
In addition, limiting investigations of lung cancer to nonsmokers typically leads to a much reduced sample size and many fewer exposed cases. This is a particular problem for studies of occupational exposures because smoking rates among blue collar workers are higher than in the general population. Therefore, exclusion of smokers limits not only the number of cases in the study, but a disproportionate number of blue collar cases who are most likely to have exposures of interest. This is shown by the relatively small number of cases in any specific occupational exposure category in the fairly large study described by Zeka et al.4 Small numbers limit interpretations and make conclusions difficult.
Finally, exclusion of smokers also does not allow evaluation of effect modification or interaction from active smoking as has been shown to occur for asbestos and a few other exposures.21 For example, in one study, the relative risk for lung cancer was 5.17 from asbestos exposure, 10.85 from smoking, and 53.24 from interaction of these factors.21,22 Focusing on either factor independently could seriously underestimate the overall impact of that factor on lung cancer development. Presentation of analyses focused exclusively on smokers or nonsmokers can often be useful, but inclusion of results on both smokers and nonsmokers in the same paper might provide the most information regarding nontobacco causes of lung cancer.
Lung cancer among nonsmokers is a moderately common disease, but established risk factors can only explain approximately 25% of its occurrence. Risk factors among nonsmokers appear to be similar to those among smokers, although interaction with smoking also occurs. Studies that include smokers and nonsmokers would seem to offer the best opportunity to identify additional causes of lung cancer among nonsmokers and to further identify tobacco and other risk factor interactions.
1. Thun MJ, Henley SJ, Burns D, et al. J Natl Cancer Inst
2. American Cancer Society. Cancer Facts and Figures
. Atlanta: American Cancer Society; 2006.
3. Vastag B. Attention turns to lung cancer in nonsmokers. J Natl Cancer Inst
4. Zeka A, ‘t Mannetje A, Zaridze D, et al. Lung cancer and occupation in nonsmokers: a multi-center case–control study in Europe. Epidemiology
5. Brownson RC, Alavanja MCR, Caporaso N, et al. Epidemiology and prevention of lung cancer in nonsmokers. Epidemiol Rev
6. Lan Q, Chen W, Chen H, et al. Risk factors for lung cancer in non-smokers in Xaunwei County of China. Biomed Environ Sci
7. Laden F, Schwartz J, Speizer FE, et al. Reduction in fine particulate air pollution and mortality—extended follow-up of the Harvard six cities study. Am J Respir Crit Care Med
8. Brennan P, Fortes C, Butler J, et al. Multicenter case–control study of diet and lung cancer among non-smokers. Cancer Causes Control.
9. Wu Ah, Fontham ETH, Greenburg RS, et al. Previous lung disease and risk of lung cancer among lifetime nonsmoking women in the United States. Am J Epidemiol.
10. Darby S, Hill D, Auvinen DH, et al. Residential radon and lung cancer—detailed results of a collaborative analysis of individual data on 7148 persons with lung cancer and 14208 persons with lung cancer from 13 epidemiologic studies in Europe. Scand J Work Environ Health
. 2006;32(suppl 1).
11. Taylor R, Cumming R, Woodward A, et al. Passive smoking and lung cancer a cumulative meta-analysis. Aust N Z J Public Health
12. Tardon A, Lee WJ, Delgado-Rodriguez M, et al. Leisure-time physical activity and lung cancer; a meta-analysis. Cancer Causes Control
13. Wu AH, Fontham ETH, Reynolds P, et al. Family history of cancer and risk of lung cancer among lifetime nonsmoking women in the United States. Am J Epidemiol
14. Malats N, Camus-Radon A-M, Nyberg F, et al. Lung cancer risk in nonsmokers and GSTM1 and GSTT1 genetic polymorphism. Cancer Epidemiol Biomarkers Prev
15. Axelson O. Alternative for estimating the burden of lung cancer from occupational exposures—some calculations based on data from Swedish men. Scand J Work Environ Health
16. Blair A, Stewart P, Lubin JH, et al. Methodologic issues regarding confounding and exposure misclassification in epidemiologic studies of occupational cancer. Am J Ind Med.
17. Boyd NR, Windsor RA, Perkins LL, et al. Quality of measurement of smoking status by self-report and saliva cotinine among pregnant women. Matern Child Health J
18. Lewis SJ, Cherry NM, Niven RM, et al. Cotinine levels and self-reported smoking status inpatients attending a bronchoscopy clinic. Biomarkers
19. Martinez ME, Reid M, Jiang RY, et al. Accuracy of self-reported smoking status among participants in a chemoprevention trial. Prev Med
20. Parna K, Rahu M, Youngman LD, et al. Self-reported and serum-cotinine-validated smoking in pregnant women in Estonia. Matern Child Health J
21. Alberg AJ, Samet JM. Epidemiology of lung cancer. Chest
22. Hammond EC, Selikoff IJ, Seidman H. Asbestos exposure, cigarette smoking and death rates. Ann NY Acad Sci
Figure. No caption available.
© 2006 Lippincott Williams & Wilkins, Inc.