Published ahead of print on May 3, 2007, doi:10.1164/rccm.200612-1792OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200612-1792OC
Chronic Obstructive Pulmonary Disease Is Associated with Lung Cancer Mortality in a Prospective Study of Never Smokers1 McLaughlin Center for Population Health Risk Assessment, Institute of Population Health, and 2 Department of Epidemiology and Community Medicine, Faculty of Medicine, University of Ottawa, Ottawa, Canada; and 3 Department of Epidemiology and Surveillance Research, American Cancer Society, Atlanta, Georgia Correspondence and requests for reprints should be addressed to Michelle C. Turner, M.Sc. McLaughlin Centre for Population Health Risk Assessment, Institute of Population Health, University of Ottawa, One Stewart Street, Room 318A, Ottawa, ON, Canada K1N 6N5. E-mail: mturner{at}uottawa.ca
Rationale: Several studies have suggested that previous lung disease may increase the risk of lung cancer. It is important to clarify the association between previous lung disease and lung cancer risk in the general population. Objectives: The association between self-reported physician-diagnosed chronic bronchitis and emphysema and lung cancer mortality was examined in a U.S. prospective study of 448,600 lifelong nonsmokers who were cancer-free at baseline. Methods: During the 20-year follow-up period from 1982 to 2002, 1,759 lung cancer deaths occurred. Cox proportional hazards models were used to obtain adjusted hazard ratios (HRs) for lung cancer mortality associated with chronic bronchitis and emphysema as well as for both of these diseases together. Measurements and Main Results: Lung cancer mortality was significantly associated with both emphysema (HR, 1.66; 95% confidence interval [CI], 1.06, 2.59) and with the combined endpoint of emphysema and chronic bronchitis (HR, 2.44; 95% CI, 1.22, 4.90) in analyses that combined men and women. No association was observed with chronic bronchitis alone (HR, 0.96; 95% CI, 0.72, 1.28) in the overall analysis, although the association was stronger in men (HR, 1.59; 95% CI, 0.95, 2.66) than women (HR, 0.82; 95% CI, 0.58, 1.16; p for interaction, 0.04). The association between emphysema and lung cancer was stronger in analyses that excluded early years of follow-up. Conclusions: This large prospective study strengthens the evidence that increased lung cancer risk is associated with nonmalignant pulmonary conditions, especially emphysema, even in lifelong nonsmokers.
Key Words: lung neoplasms • pulmonary disease, chronic obstructive • bronchitis, chronic • pulmonary emphysema • United States
Lung cancer is currently the leading cause of cancer death in the United States (1). In 2006, it is estimated that a total of 174,470 new cases and 162,460 lung cancer deaths occurred (1). Although cigarette smoking accounts for the great majority of lung cancer cases, there are many who smoke but who do not develop this disease; there are also nonsmoking lung cancer cases (2). A number of other factors, including a personal history of certain nonmalignant lung diseases, have been postulated to correlate with susceptibility for developing lung cancer (3). Several studies have suggested that chronic bronchitis and emphysema may increase the risk of lung cancer (4–11). Most of the lung cancer cases in published studies occurred in current or former cigarette smokers; thus, the observed associations may be biased by residual confounding from smoking. Furthermore, nearly all previous investigations are case-control studies and may be subject to biases in exposure assessment because patients with lung cancer may preferentially recall their experience of chronic lung diseases. Because lung cancer is highly fatal, many of these studies use a large proportion of proxy respondents (from 32 to 65%), or surviving cases only (4–6, 8, 10, 11). Patients with symptoms of lung cancer can be misdiagnosed as having other lung disease. Several, mostly small, prospective studies including current and former smokers also reported inverse relationships between lung function and lung cancer incidence or mortality (12). Studies in China have also reported similar findings; however, it has also been established that the high rates of both chronic obstructive pulmonary disease (COPD) and lung cancer found, including in studies of lifelong nonsmokers, are believed to result primarily from indoor air pollution due to coal burning and fumes from cooking oil (13–23). It is therefore unclear to what extent the COPD disease process may contribute to lung cancer risk, or whether both COPD and lung cancer are a consequence of the underlying exposure, or perhaps a combination of both. It is important to clarify the association between chronic bronchitis and emphysema and lung cancer in the general population. In this article, we examine the association between chronic bronchitis and emphysema and lung cancer mortality in a large population of lifelong nonsmokers in the United States using data from the Cancer Prevention Study II (CPS-II) cohort.
Study Population The CPS-II cohort is a prospective study of cancer mortality established by the American Cancer Society. Nearly 1.2 million study participants were enrolled by over 77,000 volunteers in 1982. Participants were recruited in all 50 states as well as the District of Columbia and Puerto Rico. Participants were at least 30 years of age at baseline. A self-administered questionnaire was completed at baseline that ascertained a variety of demographic, medical, and lifestyle data. No sickness-related exclusion criteria were applied for the baseline data collection. The Emory University School of Medicine Human Investigations Committee approved all aspects of the CPS II study. The vital status of study participants is determined every 2 years. The National Death Index has been used for computerized linkage and follow-up since 1989 (24). Previously, volunteers ascertained the vital status of participants they had enrolled, with confirmation by obtaining the corresponding death certificate. As of December 31, 2002, 385,245 participants had died (32.5%), 796,476 were alive (67.2%), and 2,840 (0.2%) had follow-up terminated in September of 1988 due to insufficient information to link to the National Death Index. Over 98% of deaths have been assigned a cause. Participants were excluded if, at baseline, they reported prevalent cancer (except nonmelanoma skin cancer) (82,340), were a current or former smoker (607,261), or if their smoking status was unknown (46,360). In total, 448,600 never smokers were retained for analysis, among which 1,759 lung cancer deaths occurred.
Ascertainment of Previous Lung Disease and Cancer Deaths Cancer deaths were classified by the underlying cause of death according to the International Classification of Diseases (ICD) (26, 27). Lung cancer deaths were defined by the following ICD codes: 162 (ICD 9 [9th revision]) and C33-C34 (ICD 10 [10th revision]).
Statistical Analysis
To examine potential biases in lung disease diagnosis, sensitivity analyses were conducted focusing on the joint effects of chronic bronchitis and emphysema on lung cancer mortality, and the effect of consecutively excluding deaths (events or censored) in the first 1 to 5 years of follow-up. In addition, interaction terms were entered into the multivariate models to examine whether the association between previous lung disease and lung cancer mortality was modified by sex, age at baseline (< 55 yr,
Previous physician-diagnosed chronic bronchitis and emphysema were reported by 2.7 and 0.5% of lifelong nonsmokers, respectively, at baseline. A total of 0.2% of nonsmokers reported having a diagnosis of both chronic bronchitis and emphysema. The prevalence of reported lung disease varied according to participant characteristics (Table 1). The prevalence of chronic bronchitis and emphysema diagnosis tended to increase with increasing age. Females and alcohol consumers were more likely to report chronic bronchitis compared with males and non–alcohol consumers. Chronic bronchitis also tended to increase with increasing exposure to environmental tobacco smoke. Participants reporting previous occupational exposures tended to report a greater prevalence of previous lung disease diagnosis at baseline.
Table 2 describes the relation between lung cancer mortality and chronic bronchitis and emphysema measured at baseline in never smokers. Lung cancer mortality was significantly associated with emphysema (hazard ratio [HR], 1.66; 95% confidence interval [CI], 1.06, 2.59) and with the combined endpoint of emphysema and chronic bronchitis (HR, 2.44; 95% CI, 1.22, 4.90) in analyses that combined men and women. No association was observed with chronic bronchitis alone (HR, 0.96; 95% CI, 0.72, 1.28).
Sensitivity analysis revealed the HR for emphysema alone without chronic bronchitis was 1.39 (95% CI, 0.78, 2.43). The HR for chronic bronchitis alone without emphysema was 0.86 (95% CI, 0.63, 1.18). The association between emphysema and lung cancer strengthened when analyses excluded early years of follow-up (Table 3). Few significant interactions were observed (Table 3); however, the association between chronic bronchitis and lung cancer was stronger in men (HR, 1.59; 95% CI, 0.95, 2.66) than women (HR, 0.82; 95% CI, 0.58, 1.16; p for interaction = 0.04). Because the association between lung disease and lung cancer mortality did not vary significantly by follow-up time, the proportional hazards assumption did not appear to be violated.
The principal finding in this large prospective study is that increased lung cancer mortality was associated with a history of emphysema, even among persons who had never been active smokers. The association was stronger among those who reported both emphysema and chronic bronchitis, and increased in analyses that excluded early years of follow-up, consistent with a causal relationship. Although no association was seen between lung cancer and chronic bronchitis in the overall analysis, there was some suggestion of a sex difference, with chronic bronchitis possibly being more strongly associated with lung cancer in men than women. Most (4–6, 9, 10) but not all (8) of the case-control studies conducted in the United States have reported stronger associations between lung cancer and emphysema than with chronic bronchitis. Self-reported emphysema was strongly associated (odds ratio, 2.87; 95% CI, 2.20, 3.76) with lung cancer in the largest, hospital-based case-control study of both male and female cases from Texas (9), whereas no such association was found for bronchitis. Four population-based studies of women (4, 5, 10, 11) have reported relative-risk estimates ranging from 1.9 to 2.7 for emphysema, and from 0.9 to 1.7 for chronic bronchitis. All but one (11) of these studies included current or former smokers, as well as never smokers, and controlled for the effect of smoking in multivariate analyses. No clear patterns have been observed in analyses by histologic subtype of lung cancer (4, 6, 11), although the number of cases within specific subtypes is small. The current study using CPS-II data permitted an evaluation of the association for the first time among a large cohort of lifelong never smokers, thereby avoiding complex interrelationships with smoking, which may obscure any causal inferences relating to lung disease and lung cancer risk. Although changes in smoking status were not evaluated for the full CPS-II cohort, they were assessed among a subset of participants enrolled in the Nutrition Cohort in 1992/1993 (2, 30). The prospective follow-up also aided in eliminating many of the potential limitations of previous studies, including differential recall bias and the use of large numbers of proxy respondents. Several investigations also reported a positive association between asthma and lung cancer risk (31) but these are subject to many of the limitations described above. In our previous analysis of the CPS-II cohort, we reported a modest association (HR, 1.11; 95% CI, 0.79, 1.56) between a history of asthma and lung cancer mortality in never smokers (32). The main limitation of this study, and of other studies of this type, is that we are unable to distinguish whether COPD is in the causal pathway for lung cancer or whether both COPD and lung cancer are related to an underlying exposure, or some combination of both, with or without inherited familial predisposition. Nonetheless, the findings suggest that lung dysfunction among lifelong nonsmokers is an important population health issue. Physician-diagnosed lung disease measured at baseline was ascertained by self-report and may be associated with a certain degree of misclassification. The basis of the physician diagnosis is also unknown. The prevalence of previous lung disease in the current study appears to be similar to, although slightly lower than, that reported in other studies in the United States (11, 33, 34). The prevalence of self-reported physician-diagnosed chronic bronchitis or emphysema in lifelong nonsmokers in the U.S. National Health and Nutrition Examination Survey (NHANES) I and II, and the Hispanic Health and Nutrition Examination Survey overall, was 3.7% in men and 5.1% in women (33). The prevalence of self-reported physician-diagnosed chronic bronchitis in the control population of a United States case-control study of lung cancer in lifetime nonsmoking women was 7.0 and 0.9% for emphysema (11). Using spirometric criteria in nonsmoking NHANES III participants, Behrendt (35) reported that approximately 4.7% of nonsmokers had mild COPD and 1.9% had moderate to severe COPD. Few studies have attempted to validate self-reported diagnoses of chronic bronchitis or emphysema (25, 36). A validation of self-reported physician-diagnosed COPD in the Nurses' Health Study found high rates of lung disease (emphysema, chronic bronchitis, and COPD) confirmed by medical records (36). However, the findings may not be generalizable to the overall population because health professionals may report their medical histories more accurately (36). Another study validating data from a United States cohort in Tucson showed that low rates of self-reported physician-diagnosed chronic bronchitis met the symptom criteria (25). It was suggested that the term "chronic bronchitis," a popular term widely used by both the public and health professionals, may often be misapplied (25). Misclassification of self-reported lung disease in the current study, however, would likely be nondifferential and result in biases of the HRs toward the null. If present, this bias could partially explain the lack of association found between chronic bronchitis and lung cancer mortality. It has also been suggested that there may exist a large proportion of undiagnosed COPD cases in the general population and in lifelong never smokers, which could also bias study results toward the null hypothesis of no association (36, 37). Detailed information on disease history, including age at diagnosis or those diagnosed during the follow-up period, was not collected in this study. Although age at baseline and attained age were examined as potential modifying factors, there were no significant findings in this regard (2, 25, 36). A significant interaction with sex was observed among those with chronic bronchitis, in which men with the disease tended to experience an elevated, but not significant, risk of lung cancer death compared with women. Although the reasons for this finding are not clear, it is possible that there may exist sex differences in physician diagnosis of COPD (25, 36, 38, 39). The subgroup analyses were limited, however, by low numbers of lung cancer cases with previous lung disease among the different strata. Last, because histologic information was not available, we were unable to examine the association between previous lung disease and lung cancer mortality by histologic subtype. In conclusion, the current, large, prospective study of lifelong nonsmokers provides further evidence for an association between COPD, especially emphysema, and lung cancer mortality. Additional prospective studies of never smokers, particularly with validated information on lung disease status, would help to confirm the findings of the current study, as well provide a better understanding of the nature of COPD in nonsmokers. Finally, further examination of the potential underlying biological mechanisms may be warranted to better understand the basis for an association between COPD and lung cancer.
Originally Published in Press as DOI: 10.1164/rccm.200612-1792OC on May 3, 2007 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form December 11, 2006; accepted in final form May 3, 2007
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
55 yr), or attained age (< 70 yr, 70–79 yr, 
