Declining FEy1 and Chronic Productive Cough in Cigarette ...cebp.aacrjournals.org/content/cebp/3/4/289.full.pdf · mulative age-adjusted lung cancer incidence in smokers with symptoms

Vol. 3, 289-298, june 1994 Cancer Epidemiology, Biomarkers & Prevention 289

Received 9/28/93; revised 12/21/93; accepted 12/22/93.1 To whom requests for reprints should be addressed.

Declining FEy1 and Chronic Productive Cough in CigaretteSmokers: A 25-Year Prospective Study of Lung Cancer

Incidence in Tecumseh, Michigan

Syed Shafiqul Islam and David Schottenfeld’

Department of Pediatrics, The Medical College of Georgia,Augusta, Georgia 30912-3710 15. S. 1.1, and Department of Epidemiology,School of Public Health, University of Michigan, Ann

Arbor, Michigan 48109-2029 ID. 5.1

Abstrad

A community-based study has been reviewed to assesswhether impaired forced expiratory volume in 1 s (FEV1)and/or the symptoms of chronic cough and sputumprodudion predid the incidence of lung cancer, aftercontrolling for age, sex, cigarette smoking history, andthe dietary intake of carotenoids and retinoids. A cohortof 2099 women and 1 857 men, 25 years of age or older,were first examined from 1 962 to 1 965. As of 1987,there were 60 validated lung cancers diagnosed in men(1 .83 per 1 000 person-years) and 1 7 in women (0.39 per1 000 person-years). The incidence density of lungcancer in current smokers at baseline, when comparedwith never smokers, was increased 5.34 (95%confidence interval, 1 .74, 1 6.38) times in women and4.1 1 (95% confidence interval, 1 .63, 10.34) times inmen. The risk of lung cancer increased in women andmen in relation to the average daily intensity ofexposure in current smokers and the duration ofsmoking history (<20 years, �20 years) in current andex-smokers. When stratified by cigarette smokingintensity, subjects with chronic cough and phlegmexperienced a future risk of lung cancer that was morethan 3 times higher than that in the nonsymptomaticsubgroup. Among the smoking women and men at entry,those in the lowest quartile of the percent predidedFEV1, after controlling for the average number ofcigarettes smoked per day, experienced a risk of lungcancer that was 2.7 times that of subjeds in the highestquartile. With each 1 0% decrease in percent predictedFEV1, the risk of lung cancer increased 1 .1 7 times (0.96,1 .42), after controlling for age, sex, and cigarettesmoking intensity at baseline. The average annualdecline in FEV1 as estimated between 1 962 and 1965and 1 967 and 1969 was a significant independentpredidor of future lung cancer incidence aftercontrolling for cigarette smoking history; the slope ofthe regression line indicated that with each decline inFEV1 of 100 mI/year, lung cancer incidence densityincreased 1.16 per 1000 person-years (95 % confidenceinterval, 0.30, 2.01). Controlling for potential

confounding by quartile distribution of calorie-adjusteddietary intake of vitamin A, beta-carotene, cholesterol,and fat did not weaken or alter the fundamentalrelationship with impaired pulmonary fundion. Rapidlydeclining ventilatory function in conjunction withpersistent symptoms of chronic bronchitis in currentsmokers is predidive of the increased risk of lung cancerand correlates with cumulative levels of exposure tocigarette smoke. In addition to the consideration ofindividual dosimetry, the biological implications ofchronic obstructive airways disease in conjunction withpulmonary inflammatory and proteolytic parenchymaldisease as antecedent events in the natural history oflung cancer may be hypothesized to be related toaltered bronchopulmonary clearance mechanisms andthe biochemical mediators of an inflammatory orreparative response.

Introduction

Cigarette smoking may result in chronic bronchitis and/oremphysema (COPD2) and/or lung cancer. In the early1960s, Passey (1 ) hypothesized that it was the irritatingproperties oftobacco smoke, resulting in chronic bronchitis

and inflammatory destruction of lung tissue, that was ofpathogenic significance in the causal pathway of lung

cancer, rather than any direct action by volatile and partic-ulate carcinogens in tobacco smoke. The experiments ofKuschner, however, suggested an alternative explanation,

i.e., that bronchial and bronchiolar inflammation, accom-panied by reactive proliferation, squamous metaplasia, anddysplasia in basal epithelial cells, provided a cocarcino-

genic mechanism for neoplastic cell transformation upon

exposure to polycyclic aromatic hydrocarbons (2). Riming-ton (3) determined the lung cancer incidence in relation to

cigarette smoking history and symptoms of chronic bron-chitis (cough with sputum production) in men participatingin a chest radiographic screening program. The 5-year Cu-mulative age-adjusted lung cancer incidence in smokerswith symptoms (9.09 per 1 000) exceeded that in smokerswithout symptoms (5.22 per 1000), from which we may

infer a relative risk of 1 .74 (0.94, 3.22). In the PhiladelphiaPulmonary Neoplasm Research Project, the cumulativeprobability of developing lung cancer over an interval of 10

years in men over age 45 years who were smoking one or

more packs of cigarettes per day was 6.23% in those with a

2 The abbreviations used are: �OPD, chronic obstructive pulmonary disease;

TcHS, Tecumseh Community Health Study; FEy,, forced expiratory volumein 1 5; %PFEV1, percent predicted FEV,; CI, confidence interval; FVC, forcedvital capacity.

on January 10, 2020. © 1994 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from

http://cebp.aacrjournals.org/

290 Declining FEy1 and Chronic Productive Cough

chronic cough, compared to 2.51% in those without a

cough (P< 0.05) (4).Although cigarette smoking is the predominant cause

of COPD, with an estimated attributable risk fraction ex-

ceeding 80% in smoking individuals, perhaps only 1 0-1 5%

of current smokers will eventually develop clinically signif-icant sequellae of productive cough, exertional dyspnea,

and cardiovascular disease. Individual variability in theexpression of risk may be demonstrated by the acceleratedrate of age-specific decline in measured expiratory airflow.The structural and functional correlates of COPD include:

chronic mucus hypersecretion accompanied by enlarge-ment and hyperplasia of tracheobronchial mucus glands;inflammation and narrowing of small bronchial and bron-

chiolar segments with progressive limitation and increasedresistance in expiratory airflow; and abnormal dilatation

and proteolytic destruction of terminal bronchioles and

alveolar ducts and sacs which may be associated with

impaired pulmonary gas exchange and reduced lung elasticrecoil (5).

The community-based historical cohort study that is tobe described will evaluate whether obstructive airway dis-ease is an independent predictor of lung cancer risk.

Materials and Methods

The Tecumseh Community Health Study began in 1959 asa comprehensive study of health and disease in a semi-rural

community located 45 miles southwest of Detroit, Michi-

gan. The population was relatively stable and almost en-tirely Caucasian. A population census was conducted in1957, which enumerated 9794 persons, among whom

8641 (88%) participated in the first round of examinationsin 1959-1960 (TCHS I). The second round of examinations

(TCHS II) took place between 1962 and 1965 and involved9226 subjects. Apart from routine demographic and socio-economic information, a medical history interview was

conducted which included questions about respiratory andcardiovascular symptoms and conditions. Participants werethen asked to attend a special clinic where they were ex-amined by physicians and received various laboratory pro-

cedures, including physical and anthropometnic measure-ments, x-rays of the chest and hands, resting and exercise

electrocardiograms, tests of ventilatory lung function, he-matological and biochemical testing, and urine analysis.

The third round of examinations (TCHS III) took place be-tween 1967 and 1969, and subsequently a chronic respi-

ratory disease survey was conducted between 1 978 and1979.

All individuals, aged 25 years and older (n = 4318)who had participated in TCHS II (1 962-1 965) were eligible

for the longitudinal COPD study. Subjects were excluded ifthey reported a history of any cancer prevalent at baseline,

except for nonmelanoma skin cancers, or if they developedcancer within 1 year of entering the study. At each exam-ination cycle, a history of any cancer was obtained and thenverified.

A retrospective cohort design was used to evaluate theincidence density of primary lung cancer. After the exclu-

sion of preexisting cancers (n 1 75), other than nonmela-noma skin cancers, 4143 individuals aged 25 years and

older were eligible for the longitudinal study that encom-passed follow-up until December, 1987. Vital status couldnot be determined for 187 participants who were consid-

ered lost to follow up. This resulted in a final sample of

3956 subjects, 2099 females and 1857 males for whom

complete follow-up data were available. The cohort wasconsidered fixed with no new entries.

All surviving participants of TCHS II (1 962-1 965) wereinvited to take part in TCHS Ill (1967-1969) and in the

special chronic respiratory disease survey (1978-1979).The incidence of lung cancer was determined in a compre-

hensive cancer survey of the participants and/or their nextof kin, which was conducted from 1 986 to 1 987. Membersof the cohort who had moved out of the Tecumseh com-munity received a mailed questionnaire or were inter-

viewed by telephone. The next of kin was contacted if thesubject was known to be deceased. Death certificates were

obtained for more than 99% of the decedents. The cancersurvey was conducted in approximately 95% of the TCHS II

participants. When the diagnosis of cancer was reported onthe mailed questionnaire or death certificate, permissionwas obtained to request an abstracted copy of each hospital

record which was reviewed to provide the date of diagnosis,histopathological classification, and other sources of diag-

nostic information for each registered cancer.The date of diagnosis of lung cancer was determined

from the hospital record, and the person-years were calcu-lated from the date of initial observation in 1962-1965 to

either the: (a) date of diagnosis of lung cancer; (b) dateof death; (c) last known date of contact; or (d) end of

follow-up (December 31, 1987).All but four of the reported 81 lung cancer cases were

verified histopathologically. Of the 77 histopathologically

confirmed lung cancer cases, 1 1 required additional clini-cal documentation before they were classified as primaryneoplasms of the bronchus and lung. Neither clinical

records nor histopathology records were available for fourreported lung cancer cases that were excluded from the

analysis. Information on morphology, or cell type, wasavailable and reviewed for 75 of the confirmed cases. Thecell type at diagnosis in males was squamous cell (1 7),

adenocarcinoma (14), small cell (13), and other carcino-mas, including large cell, undifferentiated, or not otherwise

specified (1 6); in females, the cell type consisted of squa-mous cell (8), adenocarcinoma (6), and other (3).

At TCHS II, a wedge spirometer (Med Science Elec-tronics, Model 1 70) and a two-channel recorder were used

to measure ventilatory lung function. All measures, includ-

ing FEV1, the volume of air exhaled during the first second

of the FVC, were based on maximal forced expiration aftermaximal inspiration.After at leasttwo satisfactorytracingswere obtained on each subject, the maximum FEV1 andmean FEV, were recorded. All measurements were ex-

pressed in liters and corrected to body temperature andambient air pressure, saturated with water vapor. Ninety-four percent of the participating cohort provided satisfac-tory pulmonary function testing. Several linear regression

models were tested, separately for females and males, inderiving the optimal predictive equation for FEV1 , indepen-

dent of cigarette smoking history. The most parsimoniousmodel incorporated age and standing height which ac-counted for 48-49% of the variation (coefficients of deter-mination) in women and men, respectively. Gender-spe-

cific regression equations for the predicted FEV1 were

derived on the basis of the subcohort of individuals whohad never smoked and did not have chronic respiratorysymptoms or skeletal or other structural thoracic deformities

at time of entry: FEV1 = a + b (height) + c (age). The%PFEV1 was calculated by dividing the observed maximum



Cancer Epidemiology, Biomarkers & Prevention 291

Table 1 Distribution by age groups at baseline and in relation to person-years of follow-up and incident lung cancer events in males and females of theTCHS cohort

Age group (yr)

Males Females

No.

Baseline

�P-Yrs

Lung cancer no.

(rate/i000 P-Yrs(

BaselineP-Yrs’

Lung cancer no.

(rate/i 000 P-Yrs)No. (%)

25-34

35-44

45-54

55-64

�65

Total

554

598

365

212

128

1,857

(29.8)

(32.2)

(19.7)

(11.4)

(6.9)

(100.0)

12,257

12,628

6,820

3,215

1,350

36,270

1 (0.08)

18 (1.43)

23 (3.37)

14 (4.35)

4 (2.96)

60 (1.65)

674

607

366

236

216

2,099

(32.1)

(28.9)

(17.4)

(11.2)

(10.3)

(100.0)

14,942

13,iio

7,677

4,452

2,726

42,907

7 (0.47)

4 (0.31)

4 (0.52)

1 (0.22)

1 (0.37)

17 (0.40)

Age-adjusted lung canc er rate/i 000 P-Yrs” 1 .83 0.39

a P-Yrs, person-years.b Adjusted to the combined baseline age distribution of the cohort of males and females.

FEV1 by the predicted FEy1, and multiplying the ratio by1 00. The observed FEV1 was also expressed as a percentageof the observed FVC; both ratios diminish with increasing

severity of obstructive disease of the proximal tracheobron-chial and distal small airways, the latter of 2 or 3 mm or lessin diameter.

Cigarette smoking history was measured at each ex-amination cycle in terms of average number of cigarettes

smoked per day. Smoking cessation was described withrespect to the categories: current; former; and never smok-

ens. The potential risk of environmental tobacco smoke willnot be addressed in this report.

Respiratory symptoms of cough, phlegm, and/or short-

ness of breath were evaluated at each examination cycle.The duration of cough and/or the production of phlegm ofless than 3 months was compared with symptoms lasting 3months or longer in the previous 1 year. The FEy1 in males

and females was significantly correlated with various yen-tilatory lung functions such as the FVC, and the maximum

mid-expiratory flow rate at 25-75% of the total forcedexpired volume. The analysis of airflow obstruction was

based upon the covaniates, %PFEV1, and average annualdecline of FEV1. Initially, the %PFEV1 distribution was di-chotomized using the conventional cut-point, >70% and

�70%. In addition, the %PFEV1 distribution for the TCHS IIcohort was stratified into quartiles, with the best functionallevel as the referent subgroup. The average decline of FEV1per year was determined by comparing in each subject the

FEV1 (1962-1965) with the FEy1 (1967-1969); two stratawere then selected to summarize the association with lung

cancer incidence, i.e., 0-<100 and �100 ml.At the time of the third examination cycle (TCHS Ill) in

1 967-i 969, a 24-h diary and food frequency dietary surveywas completed in 2664 of the TCHS II cohort subjects

(67%). For purposes of this analysis of selected nutrients,

i.e., the intake of vitamin A, beta-carotene, total fat, andcholesterol, the Tecumseh food frequency data were con-verted into the food frequency format that was compatible

with the software developed by Block et a!. (6). Mediumportion sizes as in the National Health and NutritionalExamination Survey II were assumed for all responses. Theindividual levelof intake for totalfat,cholesterol, vitamin A,or beta-carotene was standardized to the mean intake of

calories for the entire cohort.

Statistical Analysis The initial analysis reviewed the de-mographic structure and distribution of exposure variables

in the cohort. The 5-year age- and gender-specific person-

years matrix enabled the calculation of age-standardizedlung cancer incidence rates in male and female subjects. A

correlation matrix of exposure and demographic variablesidentified potential confounders that then required testing

by Mantel-Haenszel stratification analysis and multivaniateregression analysis. Kaplan-Meier survival curves and logrank statistics were employed using SAS PROC LIFETEST forinvestigating lung cancer incidence in relation to the van-

ous strata of age, sex, cigarette smoking intensity, and pul-monary ventilatory function. The Cox proportional hazards

method was used to examine the significance of each re-gression coefficient while simultaneously considering: age;

sex; cough and sputum production; occupational exposuresto dusts and fibers; cigarette smoking, either as a categorical(never, current, ex-smoker) on continuous (number of ciga-rettes per day in current smokers) covaniate; %PFEV1 ; de-

dining FEy1 (mi/year), and dietary intake of vitamin A,beta-carotene, and total fat. The Cox model explored p0-tential interactions among the covaniates, in particular cig-arette smoking and pulmonary function.

Results

Overthe 25-year interval offollowup, the age-adjusted lungcancer incidence rate (per 1 000 person-years) in the cohort

of males (1 .83) was 4.69 times (95% Cl = 2.74, 8.04) thatobserved in the females (0.39) (Table 1). The cumulativeincidence or lung cancer incidence proportion during the25-year period of observation was 3.2% in the males and0.8% in the females. At baseline (1962-1965), 55% of themales were current cigarette smokers and 21 % were formersmokers; among the females, 34% were current smokersand 7% were former smokers. Among those who had ever

smoked at baseline, 65% of the males and 40% of thefemales had smoked for 20 on more years. In the 1965

survey of adults conducted by the United States Bureau ofthe Census, the prevalence of current smokers was 53% inmales and 34% in females, and of former smokers it was21 % in males and 8% in females. In a cross-sectional surveyof surviving cohort participants from 1978 to 1979, theprevalence of current smokers was 38% in males and 30%in females.

In assessing the completeness of reporting of lung can-cer incidence in the Tecumseh cohort, we compared theobserved number of lung cancer cases with that expected,



292 Declining FEV1 and Chronic Productive Cough

Table 2 Lung cancer incidence rate per 1 000 person-years in relation to baseline history of cigarette smoking, males and females

Males Females

Cigarette smoking Rate Rate ratio Rate Rate ratioP-Yrs No. per 1000 P-Yns (95% CI) P-Yrs No. per i,ooo P-Yrs (95% CI)

Never smoker 8,872 5 0.56 1.00 24,659 4 0.16 1.00

Ex-smoker 7,385 9 1.22 2.16 3,116 0

(0.72, 6.44)

Current smoker 19,847 46 2.32 4.11 15,022 13 0.87 5.34

(cigarettes/day) (1.63, 10.34) (1.74, 16.38)

1-19 5,220 7 1.34 7,321 3 0.41

20-39 11,491 23 2.00 7 li-’ 9 1.26

�40 3,092 16 5.17 497 1 2.01

Duration (currentand ex-smokers)

<20 yrs 10,631 3 0.28 1.00 10,761 4” 0.37 1.00

�20 yrs 16,098 52 3.23 11.45 6,631 7” 1.06 2.84

(3.58, 36.63) (0.83, 9.70)

a P-Yrs, person-years.

b History of duration of smoking not available for two lung cancer cases.

Table 3 Lung cancer incidence rate per 1000 person-years in relation to baseline respiratory symptoms and average number of cigarettes smoked per day

1-19 cigarettes/day 20-39 cigarettes/day 40+ cigarettes/day

Symptoms Lung cancer Lung cancer Lung cancerP-Yrs’ P-Yrs P-Yrs

No. Rate/i 000 P-Yrs No. 1) Rate/i 000 P-Yrs No. Rate/i 000 P-Yrs

None 8,992 6 0.67 8,503 10 1.18 1,026 4 3.90

Phlegm orcough, 794 1 1.26 2,009 4 1.99 222 0

but <3 mos/yr

Phlegm or cough, 4,397 3 0.68 6,874 10 1.45 2,015 9 4.47

>3 mos/yr

Phlegm and cough, 289 0 1,176 5 4.25 325 4 12.31>3 mos/yr

a P-Yrs, person-years.1) History of respiratory symptoms not available for three lung cancer cases.

using the Connecticut tumor registry as the referent popu-lation. The Standardized Incidence Ratio and 95% confi-

dence interval for lung cancer was 0.79 (0.50, 1 .1 7) infemales and 1 .1 7 (0.94, 1 .43) in males. Of 60 incident lungcancer cases in males, 55 (92%) occurred in ever smokers.When compared with never smokers, the age-adjusted in-

cidence rate ratio was increased 4-fold in current smokersand 2-fold in former smokers. Of 1 7 incident lung cancercases in females, 1 3 (76%) occurred in ever smokers, andthe incidence in current smokers at baseline was increased

more than 5 times that in never smokers; there were no lungcancers recorded in the females who were former smokers

at baseline, after only 31 1 6 person-years of observation.The relative risk of lung cancer was increased significantlyin relation to the average intensity and history of duration ofsmoking (Table 2). The x2 test for linear trend by level of

daily intensity of cigarette smoking was, in the males, 8.03(P = 0.005) and, in the females, 3.30 (P = 0.069).

The prevalence of chronic recurring cough and phlegmwas significantly associated with current cigarette smoking

at entry (odds ratio, 3.74). Chronic respiratory Symptoms ofcough and phlegm were independent risk factors for lungcancer, after controlling for the number of cigarettes

smoked pen day (Table 3). Within the cohort subgroup whosmoked 20-39 cigarettes per day, the incidence rate ratio

for those with symptoms was 3.60 (95% CI = 1 .23, 10.52)when compared with the subgroup without symptoms; forthe subgroup who smoked 40 or more cigarettes pen day,the rate ratio was 3.1 6 (95% CI 0.79, 1 2.64). Shortness ofbreath was not an independent predictor of lung cancerrisk, after adjusting for cough and phlegm, cigarette smok-ing history, age, and sex.

In assessing the risk factor of impaired expiratory

airflow, the initial analysis dichotomized %PFEV1 into�70% and >70%, and compared lung cancer incidencedensity in the subcohort of never smokers with that ofever smokers (Table 4). In the subcohort of never smokerswith impaired expinatory airflow (%PFEV �70%), therewere two lung cancer cases (1 .45 pen 1 000 person-years)

and the risk of lung cancer was increased significantlywhen compared with that in the subcohort of neversmokers with %PFEV >70% [rate ratio = 6.30 (1.31,30.33)1. The combined risk factors of smoking and im-

paired expiratory airflow increased the incidence densityof lung cancer 13 times (95% Cl 4.92, 33.99) thatrecorded in the subcohort with relatively better pulmo-




Table 4 Lung cancer incid ence rate pe r 1000 pe rson- years in relation to baseline ventilatory lung function and cigarett e smoking status

Smoking status

Percent predicted forced expiratory volume, 1 .0

�70% >70#{176}!,,

P-Yrs’ No. Rate Rate ratio P-Yrs No. Rate Rate ratio

Never smoker

Ever smoker

i 379

3,342

2

iO

i .45

2.99

6.30 30, 1 02 7

13.00 39,640 56

0.23

1.41

1 .00”

6.13

a P-Yrs, person-years.b Reference subgroup.

Table 5 Association of estimated average decline in FEy, (mI/year) in

1 962-1 965 and 1 967-i 969 with subsequent lung cancer incidence per1 000 person-years, stratified by history of cigarette smoking at entry�’

Average annual decline in FEy,Smoking status (mI/year)(cigarettes/day)

<ioo

<20 Incidence per 1 000 P-Yrs 0.34 i .40

Rateratio 1.0’ 4.i2

20-39 Incidence per 1000 P-Yrs 1 .iO 2.08

Rate ratio 3.24 6.12

�40 Incidence per 1000 P-Yrs 3.80 iO.05

Rate ratio il.i8 29.56

Person-years 27,420 (86%) 4,309 (i4%)

a P-Yrs, person-years.

b Rate ratios (95% CI) within smoking strata: <20 = 4.1 2 (0.80, 2i .24);

20-39 = 1 .89 (0.52, 6.86); �40 2.64 (1 .24, 5.64).

C Reference subgroup.

nary function who were never smokers. After controlling

for smoking status, the Mantel-Haenszel estimation ofrelative risk of lung cancer in relation to impaired expi-natory airflow was 2.47 (95% Cl = 1 .49, 4.09), and the x2with 1 degree of freedom was 1 2.34 (P < 0.001). Among

the smoking women and men at entry, those in the lowestquartile of the percent predicted FEV1, after controlling

for the average number of cigarettes smoked per day,experienced a risk of lung cancer that was 2.7 times thatof subjects in the highest quartile. In summary, for the 75incident lung cancer cases presented in Table 4, approx-

imately 75% (56 cases) were composed of ever smokerswithout markedly impaired ventilatory function; 1 3% (10

cases) with prior exposures to active smoking and im-pained FEV1 ; 9% (7 cases) in the absence of exposure to

active smoking and impaired pulmonary function; and2-3% (2 cases) with impaired FEy1 in the absence of

active smoking.Among the subcohort of currently smoking individuals,

the estimated average annual decline in FEV, , as measuredbetween 1 962 and 1 965 and 1 967 and 1 969, was deter-

mined to be an independent predictor of lung cancer events(Table 5). Within each stratum of cigarette smoking inten-sity, the demonstration of an average annual decline in

FEy, of 1 00 ml on more increased the risk 1 .89-4.1 2 timeswhen compared with a lessen rate of decline. The slope of

the regression line indicated that with each decline in FEy1of 1 00 mI/year, the lung cancer incidence density increased1.16 pen 1,000 person-years (95% Cl 0.30, 2.01). The

Cox proportional hazards model was consistent with theinference, that even after controlling for age, sex, and cig-

arette smoking intensity, that various measures of impairedexpiratory airflow were independently significant explana-tory covaniates of lung cancer incidence in the Tecumseh

Table 6 Cox proportional hazards model of the association ofventilatory lung function and cigarette smoking with

lung cancer incidence

impaired

. RegressionCovariate . . SE x

coefficientP-value

Age 0.0708 0.Oi 46.69 <0.Oi

Sex -i.2200 0.30 13.83 <0.Oi

%PFEV, -0.0156 0.01 5.72 0.017

Ex-smoker 0.9610 0.50 3.65 0.056

Current smoker 1.8700 0.40 21.40 <0.01

%PFEV, x current smoker - 1 .i 1 70 0.82 1 .85 0.1 7

cohort. There were no apparent interactions between to-bacco smoking and pulmonary function risk factors whenincorporated into the Cox model (Table 6). In a Cox model

which was not presented, the covaniates of age and sexwere combined with the number of cigarettes pen day incurrent smokers entered as a continuous variable. The neg-

ative �3 coefficient for each unit percent increase in FEV,(%PFEV1 ) was -0.01 32 (SE = 0.0072). We may infer fromthe regression model that with each 1 0% decrease in per-cent predicted FEV1, in current smokers, the risk of lung

cancer increased 1 .1 7 times (0.96, 1 .42), after controllingfor age, sex, and cigarette smoking intensity.

In a subset of the original cohort members, food fre-

quency questionnaires were administered in 1 967-1 969 to1 284 (69%) males and 1 380 (66%) females. Comparison of

those who participated in the dietary survey with those whodid not (33%) did not reveal systematic differences in smok-ing and pulmonary function characteristics. After control-ling for cigarette smoking intensity, age, and sex, the quar-

tile distribution of the level of intake of total fat, saturated

fat, or cholesterol was not associated with the level of lungcancer incidence density. The level of intake of each of

these macronutnients was not associated with age-sex-tobacco-standard ized measu nes of pulmonary ventilatoryfunction.

Within each of the 3 cigarette smoking intensity expo-

sure levels, there was a suggestive pattern of enhancementof lung cancer incidence density by the subgroup in thelowest quartile of intake of vitamin A or of beta-carotene,when compared with the subgroup in the highest quartile.

One major limiting factor, however, was the number of lungcancer events and the distribution of person-years allocatedfor each cell and the resulting wide confidence intervals for

the point estimates of risk. For example, within the stratum20-39 cigarettes pen day, the incidence density rate ratiocomparing the lowest to the highest quartile of beta-caro-tene was 2.1 8 (95% Cl = 0.42-1 1 .24). The x2 test for trendwas not statistically significant for either vitamin A on beta-

carotene. In the Cox proportional hazards analysis, control-



Do

.�-0_ 0

�C 1) 0)

o �

o � .9� �

E o�” ,�

0 0� �

a .c 0 �12 �C U � -C � -

0)00 owU C U U

0.

(0

DoC

C��Lt�

a,

U N

C0)

E0

C

‘0U

0)UC(0U

DOC

0z

C

Nm

:g’t(0

0L1�

#{149}1�

UO�

0)0).

>‘E#{176}

-� A)

V) �

C

N.

N.

h��0.

OSO�u a, C

.,(0�

U �)- Ca, .� _o U

1111�CE�

.� �0� 0)

��

� �>- �t C

‘0�O0 o..�;>

� #{149}

� �0C�C’ �0DO_

x0�’0 0)0)CC

�C��C(0 �

#{149}_-0�� a,-0Cr-U �0)�_

� �

�

C0

Do �

Do��0)� .E E

� Do’0� C � C

C

> �1� .� ��E�2

LU .C >-U .0

0.

0

.20

(00)>,

N

N

.�-

to

11!:2

CN

E �

fl�� L(�

� .5N.N

0)UC(0UDCC

C(0

‘00)

(0C0E

0.

0)>

U

.00

U

C0

.0

U

0

0.00U

0

C

0

(00.

E0

U

.�

.0

(0C0)EE0

U

�

.�

0

‘C0)

�0C

(00.

.5 0

(0’.-,, 0

>-�0.

0)

SC

a,0QJC

0’0

>-

0.0

U2#{176} �

�

� ‘-S. �� 1’E �

V .000 �0. � Do�� #{176}Ec� .OV>’ N C

� � �0) C’”0).

� .0

E � �-� � �.C .0 0)

0. v�

II;>,� �

.�0) .0

�ODO Do

C a,C C�00)� �

DoE�(D �

� 0� .0.C .00. vi

� CN

C(0’-

b�Ca,� .0E�

�

�U��’.- 0.

� ‘0�Um�>’

Do��’( .2C�0. -

� 0. 0

.2 0

� � . � &�0)� C� �Ln.� t�C � �� 2’- . .a � 0.N.’.- �CU�

� � �

� � ��

(0 �- � � � .0 C U 0. � � � Do � �D � C

�“-�a,0) Do � Q� CCm C .0.0”0��

m

0�,-

�.;�

0)

-0)

0.

.

p0-.,-�

0),-C-(ON’

E��SC0�

U�0

�

-�‘D‘-C0),-CZ0)N.�;�o

.2�C�0)�

294 Declining FEy1 and Chronic Productive Cough



.�


g.� .v�’

E �� 2E Q0� -0U �00 ‘- �C � �0):t(0 C(0�0 � 0)U CC 0�>. .�

0)� .0 0.�o :� �E�vi EQ�

.� � � �

C U �.0 � �00a) ‘- � E0�,. C ‘� ‘- E� C�0) �g .2�E “E#{176}��

.� � �ua, OOC(0 E�#{176}’

� <

II a, 0) RI � II-. (0 ,.� . a) .0NLn� � 0) -�- N � DoN�C

:� -� .� H 0�-��Lf� �

#{128} � U� ‘:i’#{176}�r�i� ��iI�E

> �r�E � � � �LU_,� L-.C.�C Ox #{149}.� E #{244}r’�-�Do2U�Lf� � E � �C � Do E Do,_ �

�1� N (0 > C � �. C � C � � .0 U 0)

,- .00)5 � u�V V� �:.�-‘-: 0.E� .OCO .�Do D�’ U0._. �- � tN.�

� C�0.2 - � �i; � �

0__� Sat> �.2.2 .� CN �C ,.�UN 0’ �

�r;s �-�- � F’� � c�e��r’� �UL..S �=a, 0.�U� �r�r’�2

L� 0. U � U- U

�- � � Do

�E-� � .� � .5

2.2�0� C� � t � 0)C � �o � U > (0 a, 0) � C = >. C �.- - N N. (0 Do .- .0 �- E � t .0 a,

.�0. .,� ._0j� CO 0)Do 0)DoO (0 O�C

0)E A) Ca,C :o . � 0)E �� r0 � 2 -� ‘#{176} (0 .0 .0 .0 � .C D’ � .0�>

.00) 0) � � ‘“0) .0.0 � U”’”� � ES � >

� U_� 0 Do , C � � =

� �U�E :�-� � 2.CO� ‘U� � � ��-0�EN L�

� ;:� EE-#{176} �‘EU� ‘�

);:O-O�-:2 0.u0.�. 2m’� � � 0.�5LU .0 .0U- 0�� U u- U 0� I

, 0)

� U

.� .2 U� 0 a, U>,�‘0)�: � �-

vi.> �0 �- �

C0)� ,� 0.0)Do ‘“.0��cm>. ,- �

0.IE�Lfl. 0� r’,-0� 0.�C� .020)N. � ‘-:�o��

0 Do � ro .2 o ‘0 ‘0 .0 0- (0 - ,‘“ - 0)-� ,- 0) -0 DO ‘� 0 (1� U = a, �0.- I E0).E ‘.- �C.00)(’� .5� 0)5 � .0)DoN. (0 Do.- �....C (0

�‘�.Za � ,�.U(0

0)U-O�� 0)�_ N..S� �r-a,02

� ‘� �

.� �

110) 00 � = -

C .0 0) � ‘� m C � (0 � ,� .� �

:; � � �� 0.,�C0) �“”C

(ODo0 .�CE E (01 �O0) �C� � �

“-;�----5� �CtU �� a,4)aC �

� 2�0)% E’#{176}�IIa,0.0)��02C>. ��;2 O��CC

� Uo 0 �C �..- 0.0)‘0’ . � � .� ‘� �O m E - 00 � (0 0

0)O�-0) �‘U � �N.O� OCCN. �CC� �Lr�.CO -��#{149}>o� N.0)�

.�o � �c - (0 a, ‘�. ‘�‘� 5. N. � U c C - � C ._

U 0. � .0

(00)>-

. C

#{163}� - � �-..:.-� -..N. (0� ‘-,-

(O� (O� .-.‘“ 0)�

� �6 0)� (0’-

a� O� OO� a� a,

.� � � �

_5� �- �- O�� -J > z



Inflammatory injury oflung parenchyma:

#{149}Release of proteolytic enzymes

#{149}Production ofoxygen radicals

#{149}Macrophage secretion of polypeptide

growth factors

Fig. 1. Enhancement of the causal path-way of cigarette smoke and lung cancerby the association with chronic obstruc-

tive pulmonary disease.

a Promoters (e.g., aldehydes, peroxides)

Biochemical autocrine tumor growth factors:

#{149}insulin-like growth factor�#{149}neuroendocnne peptides

#{149}transforming growth factor - a

296 Declining FEy, and Chronic Productive Cough

[ Cigarette Smoke

Duration, Intensity, Depth of Inhalation,

Genetic Susceptibility

Obstruction oftracheobronchial andbronchiolar airways:

S Mucus hypersecretion

. Impaired mucociliary clearance

. Deposition and retention of particulates

. Increased airway reactivity (“asthmatic

bronchitis”)L , -t

Molecular genetic events I #{149}Genotoxic initiators (e.g., n-nitrosamines,c.myc amplification, k-ras point mutation, ) polycyclic aromatic hydrocarbons)c-erB.2 altered transcription, loss oftumorsuppressor genes (Rb. p53. 3p deletion) � #{149}Co-carcinogens (e.g., catechols)

LUNG CANCER

ling for potential confounding by calorie-adjusted intake ofvitamin A or beta-carotene did not weaken on alter the

fundamental relationships with age, sex, cigarette smoking,

and impaired pulmonary function.

Discussion

The Tecumseh community-based cohort study enabled anevaluation in women as well as in men of the independent

significance of chronic recurring symptoms of cough andphlegm, in association with baseline impaired and longitu-dinally declining FEV1 in predicting lung cancer incidencedensity; the relevance of these relationships was assessedwhile controlling for age, cigarette smoking history, anddietary intake of calorie-adjusted total fat,cholesterol, vi-tamin A, and beta-carotene. In the population-based studyconducted by Lange et a!. (7), the elevated Cox regressioncoefficients for chronic phlegm, and percent predicted FEV1and FEV1/FVC were similar in women and men. Other

cohort studies have controlled for age and smoking on thebasis of surveys of healthy working men in urban or rural

areas (8-1 3), male participants in a lung cancer screeningtrial on clinic patients receiving treatment for COPD (14),and male participants at increased risk for coronary heartdisease in a randomized Multiple Risk Factor InterventionTrial (1 5) (Table 7). Although baseline health status, degreeof abnormality in ventilatory function and length of fol-lowup interval varied considerably in its distribution in the

various cohorts, there was notable consistency in the rangesof estimations of relative risk for the same indices of airflow

obstruction.In the cohort subgroup who were smoking at entry,

about 1 4% experienced a markedly accelerated rate of

decline in FEV1 (�100 mi/year) after age 25 years, whichwas at least 3 to 4 times the expected average rate of declinein never smokers (25-35 mI/year) (1 6); both the baselinelevel of percent predicted FEV1 and the rate of decline in

FEV1 were independent predictors of lung cancer mci-

dence. In the subgroup characterized by smoking two or

more packs of cigarettes per day accompanied by a decline

in FEy1 of 1 00 mI/year or greater, their long-term risk of

lung cancer was almost 30 times that in the subgroup who

had smoked less than 1 pack pen day and had experienced

a lesser rate of decline in FEy1. In addition to the cumula-

tive level of exposure to cigarette smoke, i.e., the exposurelevel derived from the product of average intensity and

duration, susceptibility to the toxic effects of smoking andthe pathological sequellae of COPD and/on lung cancer

would appear to be correlated with the accelerated rate of

decline in FEy, with age.

Cigarette smokers manifest a substantially increasedrate of loss in age-height-standardized FEV1 and FVC

relative to never and former smokers. The rate of decline

is correlated with baseline pulmonary function and the

current daily intensity of exposure and increased airway

reactivityto tobacco smoke (17-26). Fletcher et a!. (27) in

an 8-year longitudinal study reported that the mean an-nual loss in FEV1 was 36 ml for never smokers, 45 ml for

up to 1 5 cigarettes pen day, and 54 ml for 1 5 or morecigarettes per day (27). Those who stop smoking cease to

lose FEV1 at an accelerated rate, with the benefits ofcessation enhanced among the youngest smokers who

have experienced the shortest mean duration and pack-years of exposure (28-30).

The attributable fraction is the estimated proportion ofexposed cases that is due to the risk factor. In the context ofthe incidence densities or hazard rates derived in theTecumseh population, the attributable fraction in the ex-posed lung cancer cases was: male current smokers, 75.9%;

female current smokers, 81 .6%; and with combined expo-sunes, i.e., impaired FEy1 and ever active smokers, 92.3%.Of potential significance is the increased lung cancer mci-

dence density observed in the never smokers with markedly




impaired ventilatory function. However, the increased rateratio, 6.30 (1 .31 , 30.33), was based on only two cases andfewer than 1 500 person-years of observation. Van Den

Eeden and Friedman (31) in their study of members of theNorthern California Kaiser Penmanente Medical Care Pro-gram did not observe any trend or variation in lung cancer

incidence among never smokers by qumntile of FEV1.Chronic cigarette smoking retards mucociliary clean-

ance of foreign particulates and respiratory tract secretions,

evokes an inflammatory response (i.e., alveolar macro-phages and neutrophils) accompanied by fibrosis and thick-ening in the membranous and respiratory bronchioles, and

causes mucus gland hypertrophy, hyperplasia and dysplasiain the proximal airways. The defensive role of pulmonarymacrophages can be compromised by cigarette smoke and,furthermore, subverted to release excessive amounts of ly-

sosomal proteases that cause structural damage in the lung.Excessive phagocytic activation and oxides of nitrogen gen-erated by tobacco combustion elevate tissue levels of reac-

tive oxygen-free radical species that degrade mucus glyco-proteins and render lung tissue more susceptible to

proteolytic injury (32-38).

The manifestations of COPD signal the extent of bron-

chopulmonary structural and functional damage arisingfrom the interaction of sustained exposure to toxic productsof tobacco combustion and host susceptibility. In this con-text, COPD is both a biomanker of dosimetry and tissuesusceptibility. A more controversial interpretation would be

that of how COPD impacts the development of lung cancer.A conceptual model is proposed that incorporates the po-

tential cocancinogenic effects of chronic obstructive pulmo-nary disease in the causal pathway of cigarette smoke and

lung cancer (Fig. 1). Cocarcinogenesis implies an augmen-tation of neoplastic transformation brought about by non-

genotoxic factors or mechanisms operating in conjunctionwith an initiating agent. The molecular events in the natural

history of lung cancer comprise multiple genetic mutationsthat are determinants of neoplastic transformation and tu-mon progression, and the elaboration of autocnine growthfactors that influence the clonal behavior and morpholog-

ical features of neoplastic cells (39-41). Chronic inflam-mation in the distal small airways is an important cause ofobstructive symptoms and provides the dynamic setting foroxidative stress and the formation of free radicals that ac-company the reparative proliferative response (42-44). In-creased proliferation kinetics and the interaction of hy-

droxyl radicals with DNA augment the likelihood of DNAstructural and transcriptional errors.

The collective results of at least 1 0 cohort studies cur-rently provide a compelling scientific rationale for targeting

preventive interventions (e.g., smoking cessation, introduc-tion of a chemopreventive agent (45), and use of candidate

biomarkers for screening and early detection (46) in a sus-

ceptible subgroup of cigarette smokers.

References

1 . Passey, R. D. Some problems of lung cancer. Lancet, 2: i 07-i 1 2, 1962.

2. Kuschner, M. The J. Burns Amberson Lecture: the causes of lung cancer.

Am. Rev. Respir. Dis., 98: 573-590, 1968.

3. Rimington, J. Smoking, chronic bronchitis, and lung cancer. Br. Med. j.,

2:373-375, 1971.

4. Boucot, K. R., Seidman, H., and Weiss, W. The Philadelphia Pulmonary

Neoplasm Research Project: the risk of lung cancer in relation to symptomsand roentgenographic abnormalities. Environ. Res., 13: 451-469, 1977.

5. U.S. Department of Health and Human Services. The health conse-

quences of smoking: chronic obstructive lung disease. A report of the Sur-

geon General. U.S. Department of Health and Human Services, Office onSmoking and Health. DHHS Pub. No. 84-50205, 1984.

6. Block, G., Hartman, A. M., and Naughton, D. A reduced dietary ques-

tionnaire: development and validation. Epidemiology, 1: 58-64, 1990.

7. Lange, P., Nyboe, I., Appleyard, M., Jensen, G., and Schnohr, P. Venti-latory function and chronic mucus hypersecretion as predictors ofdeath from

lung cancer. Am. Rev. Respir. Dis., 141:613-617, 1990.

8. Peto, R., Speizer, F. E., Cochrane, A. L., Moore, F., Fletcher, C. M., Tinker,

C. M., Higgins, I. T. T., Gray, R. G., Richards, S. M., Gilliland, I., andNorman-Smith, B. The relevance in adults of air-flow obstruction, but not ofmucus hypersecretion, to mortality from chronic lung disease. Results from

20 years of prospective observation. Am. Rev. Respir. Dis., 128: 491-500,

1983.

9. Skillrud, D. M., Offord, K. P., and Miller, D. Higher risk of lung cancer in

chronic obstructive pulmonary disease. Ann. Intern. Med., 105: 503-507,1986.

10. Tenkanen, L., Hakulinen, T., and Teppo, L. The joint effect of smoking

and respiratory symptoms on risk of lung cancer. Int. J. Epidemiol., 16:509-5i5, i987.

1 1 . Tenkanen, L., Teppo, L., and Hakulinen, T. Smoking and cardiac symp-

toms as predictors oflung cancer. j. Chron. Dis., 40: 1121-1128, i987.

i 2. Vestbo, J., Knudsen, K. M., and Rasmussen, F. V. Are respiratory symp-

toms and chronic airflow limitation really associated with an increased riskof respiratory cancer? Int. J. Epidemiol., 20: 375-378, 1991.

13. Nomura, A., Stemmermann, G. N., Chyou, P-H., Marcus, E. B., andBuist, A. S. Prospective study of pulmonary function and lung cancer.

Am. Rev. Respir. Dis., 144: 307-31 1, 1991.

14. Tockman, M. S., Anthonisen, N. R., Wright, E. C., and Donithan, M. G.

Airways obstruction and the risk for lung cancer. Ann. Intern. Med., 106:512-518, 1987.

15. Kuller, L. H., Ockene, I., Meilahn, E., and Svendsen, K. H. Relationof forced expiratory volume in one second )FEV, ) to lung cancer mortality inthe multiple risk factor intervention trial (MRFIT). Am. I. Epidemiol., 132:

265-274, 1990.

1 6. Ferguson, G. T., and Cherniack, R. M. Management of chronic obstruc-

tive pulmonary disease. N. EngI. J. Med., 328: 1017-1022, 1993.

1 7. Beaty, T. H., Menkes, H. A., Cohen, B. H., and Newill, C. A. Risk factors

associated with longitudinal change in pulmonary function. Am. Rev. Respir.

Dis., 129: 660-667, i984.

18. Habib, M. P., Klink, M. E., and Knudson, D. E., et al. Physiologicalcharacteristics of subjects exhibiting accelerated deterioration of ventilatoryfunction. Am. Rev. Respir. Dis., 136:638-645, 1987.

19. Camilli, A. E., Burrows, B., Knudsen, R. I., Lyle, S. K., and Lebowitz,

M. D. Longitudinal changes in forced expiratory volume in one second inadults. Effects of smoking and smoking cessation. Am. Rev. Respir. Dis., 135:794-799, 1987.

20. Burrows, B., Knudson, R. J., Camilli, A. E., Lyle, S. K., and Lebowit.z,M. D. The “horse-racing effect” and predicting decline in forced expiratory

volume in one second from screening spirometry. Am. Rev. Resp. Dis., 135:788-793, 1987.

21 . Dockery, D. W., Speizer, F. E., Ferris, B. G., Jr., Ware, J. H., Louis, T. A.,and Spiro, A., Ill. Cumulative and reversible effects of lifetime smoking ontests of lung function in adults. Am. Rev. Respir. Dis., 137: 286-292, 1988.

22. Sparrow, D., O’Connor, G., and Weiss, S. T. The relation of airwaysresponsiveness and atopy to the development of chronic obstructive lung

disease. Epidemiol. Rev., 10: 29-47, 1988.

23. Ware, J. H., Dockery, D. W., Louis, T. A., Xu, X., Ferris, B. G., lr., and

Speizer, F. E. Longitudinal and cross-sectional estimates of pulmonary func-tion and decline in never-smoking adults. Am. 1. Epidemiol., 132: 685-700,1990.

24. Sherman, C. B., Xiping, X., Speizer, F. E., Ferris, B. G., Jr., Weiss, S. T.,and Dockery, D. W. Longitudinal lung function decline in subjects withrespiratory symptoms. Am. Rev. Respir. Dis., 146: 855-859, 1992.

25. Xu, x., Dockery, D. W., Ware, J. H., Speizer, F. E., and Ferris, B. G., Jr.Effects of cigarette smoking on rate of loss of pulmonary function in adults:a longitudinal assessment. Am. Rev. Respir. Dis., 146: 1345-1348, 1992.

26. Cauberghs, M., Clement, J., and Van de Woestijne, K. P. Functional

alterations accompanying a rapid decline in ventilatory function. Am. Rev.Respir. Dis., 147: 379-384, 1993.

27. Fletcher, C., Peto, R., Tinker, C., and Speizer, F. E. The Natural Historyof Chronic Bronchitis and Emphysema. An Eight-Year Study of Early ChronicObstructive Lung Disease in Working Men in London. New York: Oxford

University Press, 1976.

28. U.S. Department of Health and Human Services. The health benefits ofsmoking cessation. A report of the Surgeon General. U.S. Department of



298 Declining FEV1 and Chronic Productive Cough

Health and Human Services, Office on Smoking and Health. DHHS PubI. 37. Snider, G. L. Emphysema: the first two centuries - and beyond. ANo. 90-8416, i990. historical overview, with suggestions for future research: Part I. Am. Rev.

29. Townsend, M. C., DuChene, A. G., Morgan, J., and Browner, W. S. Respir. Dis., 146: 1334-1344, 1992.Pulmonary function in relation to cigarette smoking and smoking cessation. 38. Ryrfeldt, A., Bannenberg, C., and Moldeus, P. Free radicals and lungPrey. Med., 20:621-637, 199i.

disease. Br. Med. Bull., 49: 588-603, 1993.30. Krzyzanowski, M., Robbins, D. R., and Lebowitz, M. D. Smoking ces-sation and changes in respiratory symptoms in two populations followed for 39 Carbone, D. P., and Minna, J. D. The molecular genetics of lung cancer.13 years. Int. J. Epidemiol., 22: 666-673, i993. Adv. nt. Med., 37: i53-17i, i992.

3i . Van Den Eeden, S. K. and Friedman, G. 0. Forced expiratory volume 40. Carney, D. N. Biology of small-cell lung cancer. Lancet, 339: 843-846,(i second) and lung cancer incidence and mortality. Epidemiology, 3: 1992.253-257, i992.

32. Matthys, H., Vastag, E., Kohler, D., Daikler, G., and Fischer, I. 4i . Richardson, G. E., and Johnson, B. E. The biology of lung cancer. Sem.Mucociliary clearance in patients with chronic bronchitis and bronchial Oncol., 20: 105-i 27, 1993.carcinoma. Respiration, 44: 329-337, i983. 42. Brain, I. D. Toxicological aspects of alterations of pulmonary macro-

33. Cross, C. E., Halliwell, B., Borish, E. T., Pryor, W. A., Ames, B. N., Saul, phage function. Ann. Rev. Pharmacol. Toxicol., 26: 547-565, 1986.R. L., McCord, J. M., and Harman, D. Oxygen radicals and human disease.

43. Farley, J. M. Inhaled toxicants and airway hyperresponsiveness. Ann.Ann. tnt. Med., 107: 526-545, i987.

Rev. Pharmacol. Toxicol., 32: 67-88, 1992.34. Samet, I. M. The relationship of smoking to COPD. in: N. S. Cherniack(ed), Chronic Obstructive Pulmonary Disease, pp. 249-258. Philadelphia: 44. Guyton, K. Z., and Kensler, T. W. Oxidative mechanisms in carcino-W.B. Saunders, i99i . genesis. Br. Med. Bull., 49: 523-544, 1993.

35. Rich, E. A. The role of immune abnormalities and inflammation in 45. Huber, M. H., Lee, J. S., and Hong, W. K. Chemoprevention of lungcausing and perpetuating COPD. In: N.S. Cherniack (ed), Chronic Obstnuc- cancer. semin. Oncol., 20: 1 28-141 , 1993.tive Pulmonary Disease, pp. 235-248. Philadelphia: W.B. Saunders, 1 991.

36. Wright, J. L., Cagle, P., Churg, A., Colby, T. V., and Myers, J. Diseases 46. Mulshine, I. L., and Tockman, M. S. Early Detection of Lung Cancer:

ofthe small airways. Am. Rev. Respir. Dis., 146: 240-262, 1992. Where Are We? Philadelphia: J.B. Lippincott, 1991.



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

Updated version

http://cebp.aacrjournals.org/content/3/4/289

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cebp.aacrjournals.org/content/3/4/289To request permission to re-use all or part of this article, use this link



http://cebp.aacrjournals.org/cgi/alerts

mailto:[email protected]



Documents

Declining FEy1 and Chronic Productive Cough in Cigarette ...cebp.aacrjournals.org/content/cebp/3/4/289.full.pdf · mulative age-adjusted lung cancer incidence in smokers with symptoms