Risk of Lung Cancer following Nonmalignant Respiratory Conditions among Nonsmoking Women Living in Shenyang,Northeast China

Abstract

Background:

There has been conflicting evidence about possible associations between nonmalignant respiratory conditions (NMRCs) and subsequent risk of lung cancer. Determination of whether or not there is such an association has potential importance for managing NMRCs, for screening of lung cancer, and for understanding mechanisms of carcinogenesis.

Methods:

A hospital-based, case-control study involving interviews with 226 female nonsmoking lung cancer patients and 279 matching population controls was conducted in Shenyang, Northeast China, between January 2004 and December 2007. A standardized interview collected information on a variety of potential risk factors, including a history of physician-diagnosed NMRCs (pulmonary tuberculosis, chronic bronchitis, emphysema, asthma, and bronchiectasis), and age/year in which each condition was first diagnosed. Multivariate logistic regression analyses were applied to assess the associations between NMRCs and subsequent lung cancer risk.

Results:

Compared with those without, subjects with a history of NMRC experience greater risk of lung cancer (OR = 2.0, 95% CI 1.2-3.4), particularly following a diagnosis of pulmonary tuberculosis (OR = 4.7, 95% CI 1.6-13.2). The results from subgroup analysis, when limited to small cell lung cancer, showed a 6.2-fold increase in lung cancer risk among asthmatics (95% CI 1.5-25.8). However, there was no evidence of a significant association between chronic bronchitis and lung cancer.

Conclusions:

This study strengthens the evidence linking NMRCs, especially pulmonary tuberculosis, to lung cancer even in lifelong nonsmoking women.

Introduction

The epidemiological evidence accumulated over almost six decades since the reports of Doll and Hill^1,2 have established that tobacco smoke has a causal role in the tragic worldwide epidemic of lung cancer. Tobacco smoke alone, however, cannot fully explain the etiology of this disease.³ Among Chinese women particularly, although the prevalence of smoking is only about 3%,⁴ the lung cancer mortality rate of this population ranks among the highest in the world.^5,6 Furthermore, lung cancer in Chinese women has two characteristics different from those in men, with women more likely to have adenocarcinomas of the lung, and never smokers face a higher risk.⁷ Of the major risk factors for lung cancer, cigarette smoking was found to be strongly associated with squamous cell and small cell carcinoma.^8,9 Thus, nontobacco risk factors for lung cancer are required that influence an individual's susceptibility and affect disease course.

Several studies have suggested that prior nonmalignant respiratory conditions (NMRCs) may influence the risk of lung cancer,^{10

–23} although not all these studies reached unanimous conclusions. In addition, most of the lung cancer cases in published studies occurred in current or former cigarette smokers, and inadequate control for smoking could cause residual confounding. Consequently, doubt remains about the causal nature between prior NMRC and lung cancer, in particular if the timing of diagnosis of lung disease influences lung cancer risk. Determination of whether or not there is such an association has potential importance for managing NMRCs, for screening for lung cancer, and for understanding the mechanisms of carcinogenesis.

During January 2004 and December 2007, we initiated a comprehensive hospital-based case-control study of lung cancer in Shenyang, the capital of Liaoning Province, an area with a very high baseline risk of lung cancer for both sexes. A particular strength of this study is its large number of female patients, drawn from a never smoking population without residual confounding from active smoking, and who had detailed timing of diagnosis of both NMRC and lung cancer. In our report, we used data from this study to examine in detail the associations between previous NMRCs and subsequent risk of lung cancer.

Materials and Methods

Study subjects and data collection

Two hundred fifty-five women with newly diagnosed primary lung cancer, according to the International Classification of Diseases, Ninth Revision (ICD-9 code 162), were identified in 18 hospitals in Shenyang during January 2004–December 2007. All subjects, aged between 20 and 80 years and lifetime nonusers of tobacco, were interviewed in the hospital within 2 weeks of diagnosis. Every patient enrolled in this study was rediagnosed based on review of relevant medical records, chest x-ray, CT films, and cytological and histological slides (for the latter, in 68% of the cases) by senior pathologists or clinicians. To obtain at least one eligible control for each index case, 2 female controls per case matched for age (±5 years) were randomly selected from the general population located in the urban areas of Shenyang. The Shenyang Residential Registry, which includes all permanent adult residents of urban Shenyang, was used as the sampling frame. We randomly selected subjects from the personal identification cards retained by the Registry. Each card contained the name, address, date of birth, and gender of the subject. Using the age distribution of the patients, we randomly selected a set of starting positions in the card file and then selected the first two cards of women whose ages fell within predetermined 5-year intervals.

The study was approved by the Institutional Review Board of the China Medical University. After obtaining informed consent, face-to-face interviews were conducted for all cases and controls by trained interviewers using a structured questionnaire. Detailed information was collected on demographic characteristics, dietary and cooking practices, menstrual/reproductive history, physical activity habits, the type of fuel used, exposure to tobacco smoke, and history of cancer within the family (including lung cancer). Cases and controls were also asked if a physician had ever told them they had specific medical conditions (pulmonary tuberculosis, chronic bronchitis, emphysema, asthma, bronchiectasis, and other conditions) and, if so, the age and year in which each condition was first diagnosed. The diagnosis dates of the lung cancer cases were abstracted from medical records. The same interviewer interviewed both the cases and matched controls to reduce information bias. In light of the fact that interviewers were not blind to case or control status, possible observer bias was monitored by reviewing, at random, a sample of interviews.

Of the 255 female lung cancer patients, 23 were not eligible either because they were not Shenyang residents (n = 14) or they did not have lung cancer (n = 9). Of the 232 eligible cases, 226 women completed the interview. Reasons for nonresponse included subject refusal (n = 3) and moving out of Shenyang (n = 3). Of the 510 potential controls targeted for interview, 462 women completed the questionnaire. Of these, 443 were nonsmokers and healthy (without a prior history of cancer except nonmelanoma skin cancer). According to the age distribution of the patients, 279 controls for this study were randomly selected from the pool of eligible controls.

Statistical analysis

In addition to simple descriptive statistics to characterize the study populations, we assessed the association between lung cancer and each of the available preexisting lung conditions. Unconditional multivariate logistic regression models were used to estimate odds ratios (OR) and 95% confidence intervals (CI). In analyzing associations between preexisting medical conditions and risk of lung cancer, we included the following variables as potential confounders: demographic characteristics, exposure to second hand smoke (SHS), coal combustion, and fumes.

NMRCs identified close to the time of cancer diagnosis could conceivably be misdiagnoses resulting from early lung cancer symptoms, which could artificially increase the magnitude of any association between prior lung diseases and lung cancer risk. Hence, analyses were carried out between NMRCs and lung cancer stratifying on the interval between the respiratory condition and cancer diagnosis (1–5 years, 5+ years). Analyses were conducted using the Statistica software version 7.

Results

The sociodemographic characteristics of study subjects are presented in Table 1. Frequency matching resulted in a comparable distribution of cases and controls by age, weight, height, body mass index (BMI), ethnicity, and marital status. With respect to years of school attendance, however, patients reported a smaller number of years more frequently than did controls.

Table 1.

Selected Characteristics of Study Subjects from Case-Control Study of Lung Cancer among Nonsmoking Women, Shenyang, Northeast China, 2004–2007

Characteristic	Cases (n = 226)	Controls (n = 279)	p value
Mean age	53.9 ± 11.7	55.2 ± 11.2	0.19
Mean weight	60.2 ± 10.5	61.1 ± 9.0	0.32
Mean height	159.9 ± 5.2	160.5 ± 7.9	0.27
Mean BMI^a	23.6 ± 4.1	23.7 ± 3.4	0.63
Schooling, years
≤9	131 (58.0%)	111 (39.8%)	<0.001
10–12	60 (26.5%)	111 (39.8%)
≥13	35 (15.5%)	57 (20.4%)
Marital status
Married	207 (91.6%)	248 (88.9%)	0.31
Others	19 (8.4%)	31 (11.1%)
Ethnicity
Han	215 (95.1%)	264 (94.6%)	0.80
Others	11 (4.9%)	15 (5.4%)

BMI, body mass index: weight (kg)/height²(m²). To remove the influence of lung cancer, weight was of 5 years ago.

Table 2 shows the ORs between lung cancer and some other potential related factors, such as fuel type; cooking fumes; coal stove for heating; exposure to SHS in childhood only, in adulthood only, or at both ages; family history of cancer in first-degree relatives; and two variables that we hypothesized, such as lifetime exercise habits and history of mental trauma in the last 20 years (e.g., failure of love affair or marriage, unemployment, death of relatives). The reasons for choosing mental trauma were mainly based on the observation that many lung cancer cases were of the mental trauma experience. Patients reported, more frequently than controls, an exposure to coal fuel ≥10 years for house heating or cooking, and the ORs for ever exposed vs. never exposed were 2.10 (95% CI 1.21-3.64) for cooking and heating, 2.08 (95% CI 1.26-3.43) for cooking or heating, 2.20 (95% CI 1.35-3.58) for cooking, and 1.13 (95% CI 0.75-1.70) for heating. Women who were exposed to cooking fumes ≥10 years, coal smoke ≥10 years, and SHS at both ages were more likely to be diagnosed with lung cancer. However, subjects who frequently engaged in physical exercise were significantly at lower risk of lung cancer.

Table 2.

Associations of Different Factors with Lung Cancer among Nonsmoking Women in Shenyang, Northeast China, 2004–2007

Variable	Cases (n = 226)	Controls (n = 279)	OR ^a (95% CI)
Passive smoking exposure^b
In childhood only
No	191 (84.5%)	226 (81.0%)	1.00 (Referent)
Yes	35 (15.5%)	53 (19.0%)	0.78 (0.49-1.25)
In adulthood only
No	174 (77.0%)	217 (77.8%)	1.00 (Referent)
Yes	52 (23.0%)	62 (22.2%)	1.05 (0.69-1.60)
At both ages
No	137 (60.6%)	192 (68.8%)	1.00 (Referent)
Yes	89 (39.4%)	87 (31.2%)	1.43 (1.00-2.07)
Coal fuel use ≥10 years
For cooking
No	168 (74.3%)	241 (86.4%)	1.00 (Referent)
Yes	58 (25.7%)	38 (13.6%)	2.20 (1.35-3.58)
For heating
No	152 (67.3%)	206 (73.8%)	1.00 (Referent)
Yes	74 (32.7%)	73 (26.2%)	1.13 (0.75-1.70)
For cooking or heating
No	136 (60.2%)	194 (69.5%)	1.00 (Referent)
Yes	90 (39.8%)	85 (30.5)	2.08 (1.26-3.43)
For both cooking and heating
No	184 (81.4%)	253 (90.7%)	1.00 (Referent)
Yes	42 (18.6%)	26 (9.3%)	2.10 (1.21-3.64)
Exposure to coal smoke ≥10 years
No	183 (81.0%)	251 (90.0%)	1.00 (Referent)
Yes	43 (19.0%)	28 (10.0%)	1.92 (1.13-3.28)
Exposure to cooking fumes ≥10 years
No	89 (39.4%)	156 (55.9%)	1.00 (Referent)
Yes	137 (60.6%)	123 (44.1%)	1.75 (1.21-2.52)
History of mental trauma in last 20 years
No	192 (85.0%)	248 (88.9%)	1.00 (Referent)
Yes	34 (15.0%)	31 (11.1%)	1.64 (0.93-2.86)
Lifetime exercise habits
Never	126 (56.5%)	71 (25.4%)	1.00
Occasionally	54 (24.2%)	92 (33.0%)	0.35 (0.22-0.55)
Frequently	43 (19.3%)	116 (41.6%)	0.22 (0.14-0.35)
Family cancer history in first-degree relatives
No	186 (82.3%)	213 (77.2%)	1.00 (Referent)
Yes	40 (17.7%)	63 (22.8%)	0.75 (0.48-1.19)

OR, odds ratios; CI, confidence interval. OR was adjusted for age, marital status, years of schooling, ethnicity, and 5 years ago BMI.

The cutoff between childhood and adulthood was 22 years of age.

Table 3 describes the lifetime prevalence of each NMRC by case/control status. Among the control groups, the reported lifetime prevalence of the NMRCs varied as follows: chronic bronchitis, 4.3; asthma, 2.2; pulmonary tuberculosis, 1.8; bronchiectasis, 0.4; emphysema, 0.4. Over 10% of the controls reported they had experienced one or other NMRCs. The prevalence in the present study was of a similar order of magnitude as those found in other Chinese populations.²⁴ Table 3 also shows the temporal relationships between diagnosis of the various NMRCs and lung cancer among cases. For any NMRC alone or in all cases, the mean interval measured in decades was long enough that there is little likelihood of the NMRC representing a misdiagnosis of early symptoms of lung cancer.

Table 3.

Reported Lifetime Prevalence of NMRCs by Case/Control Status and Mean Age at Diagnosis of NMRCs and Lung Cancer among Cases, Shenyang, Northeast China, 2004–2007

			Cases
NMRCs	Cases n = 226 (%)	Controls n = 279 (%)	Mean age at NMRC ^a	Mean age at lung cancer ^b
Any NMRC	20.8	11.5	30.8	55.4
Pulmonary tuberculosis	7.1	1.8	28.6	56.3
Chronic bronchitis	7.1	4.3	36.8	75.0
Emphysema	1.3	0.4	52.8	63.5
Asthma	4.4	2.2	20.4	50.9
Bronchiectasis	0.4	0.4	29.0	59.0

Self-reported.

Medical record abstraction.

Table 4 summarizes the ORs between the previous NMRCs and lung cancer. Compared with those without, subjects with a history of any preexisting NMRC had a significant increase in lung cancer risk (OR 2.0, 95% CI 1.2-3.4). Results on prior pulmonary tuberculosis demonstrated an OR of 4.5 (95% CI 1.6-12.7). However, in overall analysis, asthma and chronic bronchitis did not provide evidence of an association with lung cancer. The prevalence of exposure was quite low for emphysema and bronchiectasis. As a result, the CIs were very wide (imprecise); furthermore, the relationships with lung cancer could not be fully explored.

Table 4.

ORs between Preexisting NMRCs and Lung Cancer among Nonsmoking Women in Shenyang, Northeast China, 2004–2007

NMRCs	Cases (n = 226)	Controls (n = 279)	OR ^a (95% CI)
None	179	247	1.0 (Referent)
Pulmonary tuberculosis^b	16	5	4.7 (1.6-13.2)
Chronic bronchitis^b	16	12	1.7 (0.8-3.7)
Emphysema^b	3	1	3.7 (0.4-36.2)
Asthma^b	10	6	1.9 (0.6-5.3)
Bronchiectasis^b	1	1	1.0 (0.1-16.4)
Any previous NMRC	47	32	2.0 (1.2-3.4)

OR was adjusted for age, marital status, years of schooling, ethnicity, 5 years ago BMI, passive smoking exposure, coal use, exposure to coal smoke and cooking fumes.

For each specific NMRC, OR was estimated with the adjustment for other NMRCs besides the characteristics shown in^a.

Previous NMRCs were significantly related to both adenocarcinomas (OR = 2.1) and squamous cell types (OR = 2.7) (Table 5). For prior pulmonary tuberculosis, the association with adenocarcinoma was stronger than that with squamous cell tumors. There was little evidence that the risks associated with other specific NMRCs differed by histological type of lung cancer. Asthma, however, appeared significantly related to small cell types (OR = 6.2) rather than adenocarcinoma or squamous cell types, whereas no overall risk was noted (OR = 1.8, 95% CI 0.6-5.1). Ideally, an assessment of the impact of any specific NMRC should be conducted by the available histological types. There were, however, too few subjects with emphysema and bronchiectasis to support analysis for each histological type separately. Thus, these two NMRCs subsequently were dropped.

Table 5.

ORs ^a between Preexisting NMRCs and Selected Histological Types of Lung Cancer among Nonsmoking Women in Shenyang, Northeast China, 2004–2007^b

	Adenocarcinoma		Squamous cell		Small cell		Others
NMRCs	Cases/controls	OR (95% CI)	Cases/controls	OR (95% CI)	Cases/controls	OR (95% CI)	Cases/controls	OR (95% CI)
None	117/247	1.00 (Referent)	23/247	1.00 (Referent)	23/247	1.00 (Referent)	16/247	1.00 (Referent)
Any NMRC	32/32	2.1 (1.2-3.7)	7/32	2.7 (1.1-7.2)	5/32	1.9 (0.6-5.9)	3/32	1.1 (0.3-4.7)
Tuberculosis	11/5	7.8 (1.5-41.3)	3/5	4.5 (1.5-13.4)	1/5	2.4 (0.2-23.1)	1/5	4.1 (0.3-51.8)
Bronchitis	10/12	2.7 (0.6-11.0)	3/12	1.4 (0.6-3.5)	2/12	1.5 (0.3-7.9)	1/12	1.2 (0.1-11.4)
Asthma	4/6	1.0 (0.1-9.5)	1/6	1.1 (0.3-4.1)	4/6	6.2 (1.5-25.8)	1/6	1.6 (0.1-20.9)

OR was adjusted for age, marital status, years of schooling, ethnicity, 5 years ago BMI, passive smoking exposure, coal use, exposure to coal smoke and cooking fumes.

Two hundred twenty-six cases of lung cancer, of which 149 (66%) were adenocarcinoma, 30 (13%) were squamous cell carcinoma, 28 (12%) were small cell carcinoma, and 19 (11%) were other types.

To examine the time course of the associations, we computed OR by the number of years between the reference age (age at diagnosis of lung cancer for cases or age at interview for controls) and age when first diagnosed with the individual NMRC (Table 6). Lung cancer mortality is largely equivalent to lung cancer incidence in China (with the 5-year relative survival rate <7%).²⁵ Therefore, subsequent analyses considered the risk of lung cancer only according to time since NMRCs (1–5 vs. 5+ years). The associations with tuberculosis and any prior NMRC were substantially stronger 5+ years after NMRC diagnosis than at 1–5 years. There was no clear pattern of risk differing according to the length of the interval between lung cancer and asthma, however, because of sparse data.

Table 6.

ORs ^a between Lung Cancer and Previous NMRCs by Years since the Diagnosis of Respiratory Condition, Shenyang, Northeast China, 2004–2007

Condition/years since diagnosis	Cases/controls	OR (95% CI)	p value ^b
None	179/247	1.0 (Referent)
Any previous NMRC
1–5 years	10/8	1.8 (0.7-4.9)
5+ years	37/24	2.1 (1.2-3.6)	0.015
Pulmonary tuberculosis
1–5 years	1/1	1.3 (0.1-27.7)
5+ years	15/4	5.4 (1.8-16.9)	0.009
Chronic bronchitis
1–5 years	2/1	2.8 (0.3-31.4)
5+ years	14/11	1.8 (0.8-4.2)	0.724
Asthma
1–5 years	3/0	NE^c
5+ years	7/6	1.1 (0.3-3.4)	NE

OR was adjusted for age, marital status, education, ethnicity, 5 years ago BMI, passive smoking exposure, coal use, exposure to coal smoke and cooking fumes.

Test of homogeneity of the OR across 1–5, 5+ year intervals.

NE, not estimable because there were no exposed controls.

Discussion

The present study builds on a small but growing body of literature suggesting that women with a history of NMRCs experience greater risk of lung cancer, most particularly following the diagnosis of pulmonary tuberculosis, even among those who had never been active smokers. The ORs for asthma and chronic bronchitis also were elevated but did not always reach the traditional level of statistical significance. Hence, results must be interpreted more cautiously. For bronchiectasis and emphysema, the analyses were largely uninformative because of sparse data.

The results from our study add to those from prior reports,^{16,20,25

–28} demonstrating an association between lung cancer and tuberculosis, and this association was related to both squamous cell carcinoma and adenocarcinoma. Four studies,^14,24,25,29 also conducted in China, reported ORs/hazard ratio for lung cancer ranging from 1.3 to 6.1. An increased risk for lung cancer after a diagnosis of tuberculosis is biologically plausible. First, infection with Mycobacterium tuberculosis could induce substantial pulmonary inflammation, with production of tumor necrosis factor (TNF) linked to prolonged fever and wasting.³⁰ Reactive oxygen and nitrogen species produced by activated leukocytes participating in the inflammatory response can bind to DNA, leading to genomic alterations.³¹ Additional evidence from pathological studies shows that lung cancers frequently arise in proximity to scar tissue caused by tuberculosis.³² Tissue repair is associated with cellular proliferation during which errors in chromosomal replication might lead to further DNA mutations.

Although no association was seen between lung cancer and asthma alone in the overall analysis, there was some suggestion of a histological subtype difference, with asthma being significantly associated with small cell lung cancer (OR = 6.2, 95% CI 1.5-25.8). This evidence, to some extent, supports the results from a meta-analysis that found a stronger association with nonadenocarcinoma (RR = 2.5, 95% CI 1.5-4.1).²² Asthma may increase the risk of lung cancer by reducing the clearance of toxins and carcinogens in the bronchoalveolar epithelium, by chronic inflammation of the airways, or by the immune system.^12,13,16 Another underlying biological mechanism between asthma and lung cancer involves finding a positive association between Chlamydia pneumoniae infection, which is a known risk factor for chronic bronchitis and asthma, and lung cancer.^33
–35 C. pneumoniae may induce lung cancer through mediators of inflammation (e.g., nitric oxide and free radicals) that can activate NF-κB and other inflammation-related genes, which has been postulated in stomach carcinogenesis.³⁶

It is conceivable that early symptoms of lung cancer might be misdiagnosed or confused by the patient as signifying an NMRC. This misclassification could artificially increase the magnitude of the association between NMRCs and lung cancer. To determine if the timing of diagnosis of NMRCs influences lung cancer risk, prior studies computed OR by number of years between the reference age (age at diagnosis of lung cancer for cases or age at interview for controls) and age when first diagnosed with the individual NMRC.^24,29,37 Earlier studies, however, are inconsistent on the possible patterns of lung cancer risk stratified according to the interval between the two diagnoses¹¹; furthermore, for none of these results is previous evidence considered to be convincing. Findings from the present study can be interpreted in two ways. First, the increased lung cancer risk appeared greater 5+ years after NMRC diagnosis than within the first 5 years. Second, no significant associations with any previous NMRC were seen when limited to the 1–5-years interval. As the 5-year relative survival rate for lung cancer is low, <7% in China,²⁵ it seems unlikely that the observed association was entirely due to misdiagnosis of lung cancer.

Our study relied on subject recall for the diagnosis of NMRCs. Therefore, two innate defects, which could have partly affected both our study and prior case-control studies, deserve mention. First is the so-called Berkson's bias, whereby patients with an index diagnosis are more likely than patients without the index diagnosis to have the other disease diagnosed.³⁸ Because of nonuniversal access to quality medical care in China, this bias could have arisen in the studies under consideration, as patients with NMRCs compared with unaffected individuals may have been more likely to be diagnosed with lung cancer through the use of chest x-rays obtained in evaluation of the infection. A second bias, reverse causality, would have occurred if early stage lung cancer caused a weakening of the immune system, leading to reactivation of latent NMRCs. The NMRCs would then appear to precede the cancer when the cancer actually preceded the NMRCs. Both these biases would be reflected in strong associations between NMRCs and cancer over short time intervals. However, the associations observed over much longer periods (persisting for >5 years in our study) are unlikely to be explained by such artifacts.

Another major limitation of our study is that the data are very sparse for some NMRCs (e.g., bronchiectasis and emphysema). As a result, the CIs are very wide, and some ORs cannot even be calculated (Tables 4, 5, and 6). Findings about bronchiectasis, emphysema, and asthma should be interpreted with caution.

Although some of the comparisons were based on small numbers, the findings of the present study are informative and are specific to nonsmoking Chinese women. The current study of lifelong nonsmoking women lends further credence to the possibility of an association between prior NMRCs and lung cancer, especially for pulmonary tuberculosis. The potential contribution points to the continuing need to develop and implement improved tuberculosis prevention and treatment programs, particularly in the developing world, such as China. Our findings also raise the possibility that periodic lung cancer screening of patients with a history of NMRCs may be an effective strategy for early detection.

Footnotes

Acknowledgments

This study was supported by grant No. 2008S232 from Liaoning Provincial Department of Education and grant No. 00726 from China Medical Board. The authors are most grateful to all the participants in this study.

Disclosure Statement

The authors have no conflicts of interest to report.

References

Doll

, Hill

. Smoking and carcinoma of the lung: Preliminary report. BMJ, 1950; 2:739–748.

Doll

, Hill

. The mortality of doctors in relation to their smoking habits: A preliminary report. BMJ, 1952; 1:1451–1455.

Peter

, Tarek

, Alejandro

. Is lung cancer in the nonsmoker a different disease? Chest, 2004; 126:326–329.

Yang

. The epidemiologic investigation of the smoking behavior among Chinese population in 2002. Chin Smoking Health, 2004; 62:7–18.

, Blot

, Fraumani

. Geographic variation in female lung cancer in China. Am J Public Health, 1986; 76:1249–1250.

Jin

, Devesa

, Zheng

et al. Cancer incidence trends in urban Shanghai, 1972–1989. Int J Cancer, 1993; 53:764–770.

Zang

, Wynder

. Differences in lung cancer risk between men and women: Examination of the evidence. J Natl Cancer Inst, 1996; 88:183–192.

Brownson

, Chang

, Davis

. Gender and histologic type variations in smoking related risk of lung cancer. Epidemiology, 1992; 3:61–64.

Lubin

, Blot

. Assessment of lung cancer risk factors by histologic category. J Natl Cancer Inst, 1984; 73:383–389.

10.

Osann

, Lowery

, Schell

. Small cell lung cancer in women: Risk associated with smoking, prior respiratory disease, and occupation. Lung Cancer, 2000; 28:1–10.

11.

Brownson

, Alavanja

. Previous lung disease and lung cancer risk among women (United States) Cancer Causes Control, 2000; 11:853–858.

12.

Mayne

, Buenconsejo

, Janerich

. Previous lung disease and risk of lung cancer among men and women nonsmokers. Am J Epidemiol, 1999; 149:13–20.

13.

, Fontham

, Reynolds

et al. Previous lung disease and risk of lung cancer among lifetime nonsmoking women in the United States. Am J Epidemiol, 1995; 141:1023–1032.

14.

Wu-Williams

, Dai

, Blot

et al. Lung cancer among women in north-east China. Br J Cancer, 1990; 62:982–987.

15.

Samet

, Humble

, Pathak

. Personal and family history of respiratory disease and lung cancer risk. Am Rev Respir Dis, 1986; 134:466–470.

16.

Alavanja

, Brownson

, Boice

et al. Preexisting lung disease and lung cancer among nonsmoking women. Am J Epidemiol, 1992; 136:623–632.

17.

Islam

, Schottenfeld

. Declining FEV₁ and chronic productive cough in cigarette smokers: A 25-year prospective study of lung cancer incidence in Tecumseh, Michigan. Cancer Epidemiol Biomarkers Prev, 1994; 3:289–298.

18.

Sasco

, Merrill

, Dari

et al. A case-control study of lung cancer in Casablanca, Morocco. Cancer Causes Control, 2002; 13:609–616.

19.

Ger

, Hsu

, Chen

et al. Risk factors of lung cancer by histological category in Taiwan. Anticancer Res, 1993; 13:1491–1500.

20.

, Lee

, Chen

et al. Risk factors for primary lung cancer among non-smoking women in Taiwan. Int J Epidemiol, 1997; 26:24–31.

21.

Luo

, Wu

, Yi

et al. Indoor burning coal air pollution and lung cancer—Case-control study in Fuzhou, China. Lung Cancer, 1996; 14,Suppl 1:S113–S119.

22.

Santillan

, Camargo

, Colditz

. A meta-analysis of asthma and risk of lung cancer (United States) Cancer Causes Control, 2003; 14:327–334.

23.

Littman

, Jackson

, Vaughan

. Chlamydia pneumoniae and lung cancer: Epidemiologic evidence. Cancer Epidemiol Biomarkers Prev, 2005; 14:773–778.

24.

Brenner

, Wang

, Kleinerman

et al. Previous pulmonary diseases and risk of lung cancer in Gansu province, China. Int J Epidemiol, 2001; 30:118–124.

25.

Zheng

, Blot

, Liao

et al. Lung cancer and prior tuberculosis infection in Shanghai. Br J Cancer, 1987; 56:501–504.

26.

Campbell

, Guilfoyle

. Pulmonary tuberculosis, isoniazid and cancer. Br J Dis Chest, 1970; 64:141–149.

27.

Howe

, Lindsay

, Coppock

et al. Isoniazid exposure in relation to cancer incidence and mortality in a cohort of tuberculosis patients. Int J Epidemiol, 1979; 8:305–312.

28.

Davis

, Boice

, Hrubec

et al. Cancer mortality in a radiation-exposed cohort of Massachusetts tuberculosis patients. Cancer Res, 1989; 49:6130–6136.

29.

Engels

, Shen

, Chapman

et al. Tuberculosis and subsequent risk of lung cancer in Xuanwei, China. Int J Cancer, 2009; 124:1183–1187.

30.

Dheda

, Booth

, Huggett

et al. Lung remodeling in pulmonary tuberculosis. J Infect Dis, 2005; 192:1201–1209.

31.

Weitzman

, Gordon

. Inflammation and cancer: Role of phagocyte-generated oxidants in carcinogenesis. Blood, 1990; 76:655–663.

32.

Auerbach

, Garfinkel

, Parks

. Scar cancer of the lung: Increase over a 21 year period. Cancer, 1979; 43:636–642.

33.

Laurila

, Anitila

, Samet

et al. Serological evidence of an association between Chlamydia pneumoniae infection and lung cancer. Int J Cancer, 1997; 74:31–34.

34.

Hahn

, Dodge

, Golubiatnikov

. Association of Chlamydia pneumoniae (strain TWAR) infection with wheezing, asthmatic bronchitis, and adult-onset asthma. JAMA, 1991; 266:225–230.

35.

Von

Hertzen L

, Leinonen

, Surcel

et al. Measurement of sputum antibodies in the diagnosis of acute and chronic respiratory infections associated with C. pneumoniae. Clin Diagn Lab Immunol, 1995; 2:454–457.

36.

Ohshima

, Bartsch

. Chronic infections and inflammatory processes as cancer risk factors: Possible role of nitric oxide in carcinogenesis. Mutat Res, 1994; 305:253–264.

37.

Ramanakumar

, Parentb

, Menziesc

et al. Risk of lung cancer following nonmalignant respiratory conditions: Evidence from two case-control studies in Montreal, Canada. Lung Cancer, 2006; 53:5–12.

38.

Berkson

. Limitations of application of 4-fold tables to hospital data. Biomet Bull, 1946; 2:47–53.