clinical investigations

The Relationship Between Cigarette

Smoking and Quality of Life After Lung
Cancer Diagnosis*
Yolanda I. Garces, MD; Ping Yang, MD, PhD; Julia Parkinson, BS;
Xinghua Zhao, MS; Jason A. Wampfler, BS; Jon O. Ebbert, MD, MS; and
Jeff A. Sloan, PhD

Study objective: To describe the relationship between cigarette smoking and quality of life (QOL)
among lung cancer survivors as measured by the lung cancer symptom scale (LCSS).
Design and setting: The LCSS was mailed to eligible patients (1,506 patients) between 1999 and
2002. LCSS scores (total and individual QOL components) were compared among different
groups of cigarette smokers via univariate independent group testing and multivariate linear
models. The modeling process examined group differences adjusted for age, gender, stage, and
time of LCSS assessment. LCSS scores were transformed onto a scale of 0 to 100 points in which
higher LCSS scores corresponded to a lower QOL. A 10-point difference between groups was
defined a priori as being clinically significant.
Results: At the time of lung cancer diagnosis, 18% of the patients were never-smokers, 58% were
former smokers, and 24% were current smokers. Among survey respondents completing the LCSS at
follow-up assessment (1,028 respondents), the mean age was 65.2 years (SD, 10.8 years) and 45%
were women. Thirty percent of baseline current smokers continued to smoke at the time of the
follow-up assessment (ie, persistent smokers). The adjusted mean total LCSS scores for never-smokers
and persistent smokers were 17.6 (SD, 4.02) and 28.7 (SD, 5.09), respectively (p < 0.0001). Seven of
the individual LCSS QOL components (ie, appetite, fatigue, cough, shortness of breath, lung cancer
symptoms, illness affecting normal activities, and overall QOL) were clinically and statistically
(p < 0.001) different between never-smokers and persistent smokers. No clinically significant
differences were noted for pain or hemoptysis. Former smokers had intermediate LCSS scores. No
dose-response trends were observed between the number of packs of cigarettes smoked per day or
the total number of pack-years smoked and the adjusted scores.
Conclusion: The hypothesized relationship between smoking status and QOL was supported by this
correlational study. Our findings suggest that persistent cigarette smoking after a lung cancer
diagnosis negatively impacts QOL scores. (CHEST 2004; 126:1733–1741)

Key words: lung neoplasms; quality of life; non-small cell lung carcinoma; small cell carcinoma; smoking cessation; tobacco

Abbreviations: LCSS ⫽ lung cancer symptom scale; NSCLC ⫽ non-small cell lung cancer; PPD ⫽ packs per day;
QOL ⫽ quality of life; SCLC ⫽ small cell lung cancer

A ined
limited number of published studies have exam-
the relationship between tobacco use and
quality of life (QOL), despite the obvious intuitive
linkage. Investigations have focused on a general
patient population1,2 and on special populations such
as those undergoing percutaneous transluminal cor-
onary angioplasty, HIV-infected individuals, persons
with COPD, and elderly patients.3– 6 In general, the
QOL of a smoker is worse than that of a nonsmoker
or an ex-smoker, and an inverse relationship exists
between increasing tobacco use and QOL.2 Other
studies7,8 have evaluated QOL during smoking ces-
sation and have found that patients who are able to
achieve smoking abstinence had better QOL scores.
However, little is known about the impact of ciga-
rette smoking on QOL among lung cancer patients.
The available evidence suggests that smoking ab-
stinence at the time of cancer diagnosis improves

www.chestjournal.org CHEST / 126 / 6 / DECEMBER, 2004 1733

local control and survival.9 –12 QOL measures can
guide treatment recommendations13 for interven-
tions with favorable risk/benefit ratios that improve
For editorial comment see page 1717
QOL. If smoking cessation after lung cancer diagno-
sis improves QOL among lung cancer patients, this
information could provide physicians with additional
motivation to help current smokers quit after they
receive a lung cancer diagnosis.
To our knowledge, no studies have attempted to
correlate the tobacco history of lung cancer patients
with their QOL in a detailed and systematic fashion.
The purpose of this study was to determine the
impact of the tobacco history of lung cancer survivors
on QOL. Our hypothesis was that current smokers
would have QOL scores that were indicative of a
lower QOL than people who had never smoked. We
further hypothesized that former smokers would
have QOL scores somewhere between the current
and never-smokers. We assessed cigarette-smoking
patterns at the time of lung cancer diagnosis and at
follow-up to determine the impact of cigarette smok-
ing on patients’ QOL, as measured by the lung
cancer symptom scale (LCSS).
Materials and Methods
Study Participants and Data Collection
Since January 1, 1997, patients who have received a pathologic
diagnosis of lung cancer have been approached to participate in
a prospective cohort study at Mayo Clinic (Rochester, MN),
called the Mayo Clinic Lung Cancer Cohort. As of December 31,
2002, 5,445 patients have been enrolled into this cohort with a
participation rate of ⬎ 95%. Of these patients, 5,198 (95%) have
non-small cell lung cancer (NSCLC) or small cell lung cancer
(SCLC). All patients have provided written informed consent
prior to their enrollment into the study. The patients who agreed
to participate have their medical chart reviewed by trained
personnel for disease stage,14 histology, and types of treatment.
Patients were excluded for the following reasons: (1) they did not
speak English, and/or (2) they were non-US citizens or residents
*From the Divisions of Radiation Oncology (Dr. Garces), Epide-
miology (Dr. Yang), and Biostatistics (Ms. Parkinson, Ms. Zhao,
Mr. Wampfler, and Dr. Sloan), and Community Internal Medi-
cine (Dr. Ebbert), Mayo Clinic College of Medicine, Rochester,
MN.
Dr. Garces was supported by National Institutes of Health grant
CA–90628 and by the Fraternal Order of the Eagles Cancer
Grant Fund. Dr. Yang was supported by National Institutes of
Health grants CA– 80127 and CA– 84354.
Manuscript received March 4, 2004; revision accepted June 7, 2004.
Reproduction of this article is prohibited without written permis-
sion from the American College of Chest Physicians (e-mail:
permissions@chestnet.org).
Correspondence to: Ping Yang, MD, PhD, Mayo Clinic, Charlton
6, 200 First St SW, Rochester, MN 55905; e-mail: yang.ping@
mayo.edu
1734
at the time of diagnosis. Beginning in 1999, the LCSS has been
implemented with the follow-up materials. We will refer to this as
the first LCSS follow-up assessment throughout this report. This
study has been approved by the Mayo Clinic Institutional Review
Board on an annual basis since its inception.
Follow-up
At the time of study enrollment, each patient was informed
about a six-step follow-up (ie, at 6 months and 1 year, and
annually thereafter) in the upcoming 5 years.15 These follow-up
assessments were mailed. A total of 1,506 living patients were
eligible for follow-up, 1,251 were sent the LCSS questionnaire in
the follow-up material, and 1,028 responded to their first re-
quested QOL questionnaire (response rate, 82%). The 1,028
respondents made up the study population in the LCSS analysis.
We defined the 478 patients on whom we did not have LCSS data
as nonrespondents. The first LCSS follow-up assessments were
made with patients at various times from their diagnosis date, as
follows: 6 months, 344 of 1,028 patients (33%); 1 year, 308 of
1,028 patients (30%); 2 years, 254 of 1,028 patients (25%); and 3
years, 122 of 1,028 patients (12%). A small number of patients
(81) who received their first follow-up with the LCSS at 4 or 5
years from the diagnosis date were excluded. Therefore, the
study design is cross-sectional with variable times from diagnosis
to the QOL assessment.
QOL Instrument
The QOL assessment for this study utilized the patient portion
of the LCSS, version 2. This scale was developed by investigators
at Memorial Sloan Kettering Cancer Center and has been shown
to be a psychometrically sound index for assessing lung cancer
patient QOL.16,17 The LCSS consists of nine individual items, and
the total score is the average sum of those nine individual items.
The first six items of the LCSS represent measures of the
specified lung cancer symptoms, including appetite, fatigue,
cough, shortness of breath, hemoptysis, and pain. The remaining
three items measure general lung cancer symptoms, how the
illness affects normal activities, and overall QOL. The items were
ranked on a visual analog scale of 0 to 100 mm, in which a check
mark at 0 mm would represent no symptoms and a check mark at
100 mm would represent the maximum number of symptoms.
The health professional portion of the LCSS was not utilized
because the follow-up assessments were mailed to patients.
Tobacco Use History
Definitions of Subcategories of Smokers: Information on
tobacco history is based on self-report. Smoking status was
unknown in nine patients who were excluded from the
analysis. Never-smokers were those indicating that they had
smoked ⬍ 100 cigarettes in their lifetime and were not current
smokers. Former smokers were those reporting that they had
previously smoked ⬎ 100 cigarettes in their lifetime and were
not currently smoking. Current smokers were those who
reported that they were smoking at the time of their cancer
diagnosis. In addition to these established definitions, we
added descriptors to fully characterize the smoking patterns of
patients at baseline and at the first LCSS follow-up assess-
ment. A former smoker is described as a former smoker who
was not smoking at diagnosis or at follow-up. A relapsed
former smoker is described as someone who was not smoking
at diagnosis but had relapsed to smoking cigarettes at the time
of follow-up. A persistent smoker describes a current smoker
at diagnosis and at follow-up. And, an abstinent smoker
Clinical Investigations
describes a patient who was a smoker at diagnosis but became
abstinent from smoking at the time of follow-up.
Definitions of Dose-Response Categories
The cigarette dose-response categories were defined as self-
reported packs per day (PPDs) as well as the total number of
pack-years of smoking history. For the persistent, abstinent, and
former smokers, PPD categories were created for those who
smoked ⬎ 0 to ⱕ 0.5 PPD, ⬎ 0.5 to ⱕ 1 PPD, and ⬎ 1 PPD. We
calculated the total number of pack-years by multiplying the
self-reported number of PPDs by the number of years of regular
cigarette smoking.
Statistical Analysis
Baseline demographics were summarized utilizing mean, SD,
and range for continuous variables, and frequency or percent for
categoric variables. The association between the LCSS and
covariates was determined utilizing the Spearman correlation. A
clinically significant correlation was defined as one with an
observed coefficient of ⱖ 0.30.18 A two-sample t test was utilized
to look for associations between gender and the LCSS. To
compare the differences in the LCSS items for the different
groups of cigarette users as well as for the dose-responses (ie,
PPDs and total number of pack-years), a multivariate regression
model was developed. Although no clinically significant correla-
tions were found among the covariates and the LCSS scores, an
adjusted multivariate model was utilized due to the likelihood
that the covariates could be associated with the LCSS assessment
and because of the uneven distribution of some of these factors
among the different groups. The model was adjusted for age at
diagnosis, gender, disease stage, and time of follow-up assess-
ment since diagnosis (ie, 6 months, 1, 2, or 3 years).
The total LCSS score and the individual items ranged from 0
(no symptoms) up to 100 (maximal amount of symptoms) with a
10-point difference between groups being considered a clinically
significant difference.19 Because multiple comparisons were
made, we thought that a conservative approach would be to
define statistical significance as two-sided p values of ⱕ 0.001.
We did not adjust or interpolate for missing data or for missing
individual items, as the number of missing items was inconse-
quentially small (fewer than five). A statistical software package
(SAS, version 8.0; SAS Institute; Cary, NC) was utilized for the
analysis.
Results
Patient Characteristics
Respondents to the first follow-up assessment
including the LCSS (1,028 respondents) were more
likely to have earlier stage NSCLC compared to the
nonrespondents (respondents, 50.7%; nonrespon-
dents, 33.5%) and to be never-smokers (respon-
dents, 17.7%; nonrespondents, 14.7%). Respondents
reported a higher percentage of current alcohol use
(51.5% vs 45.5%, respectively). Nonrespondents
(478 nonrespondents) were younger with 16% of the
nonrespondents and 9% of the respondents ⬍ 50
years of age at diagnosis. However, a similar percent-
age of patients were ⱖ 80 years of age (respondents,
7%; nonrespondents, 6%). No significant differences
www.chestjournal.org
were noted between respondents and nonrespon-
dents for SCLC stage, gender, race, number of PPDs
or total number of pack-years, and other current or
past cigar, pipe, or smokeless tobacco use.
Among respondents, the mean age was 65.2 years
(SD, 10.8 years), and 45% were women. The major-
ity of respondents had NSCLC (92%), and the
remainder had SCLC. Among those patients with
NSCLC, 50.8%, 11.8%, 23.6%, and 13.7%, respec-
tively, had stage I, II, III, or IV disease, and 69.2%
of the SCLC patients had limited-stage disease
(Table 1).
Former smokers were the oldest group (Table 1).
Persistent smokers were younger than former smok-
ers but were similar in age to the never-smokers.
Furthermore, there was a higher percentage of
women among the never-smokers compared to the
percentage among former and relapsed former
smokers. Among smokers (abstinent and persistent),
there were about equal numbers of men and women.
Among the respondents with NSCLC, more ad-
vanced stage disease was noted in the persistent
smokers compared to the other categories of smok-
ers. Never-smoker respondents were the largest
percentage of respondents with stage IA disease.
Tobacco Use Status at Baseline and Follow-up
At the time of diagnosis, 18% of the patients (n ⫽
180) were never-smokers, 58% (n ⫽ 591) were
former smokers, and 24% (n ⫽ 248) were current
smokers. None of the never-smokers reported smok-
ing at the time of the first follow-up assessment.
Among former smokers, 95% (n ⫽ 562) remained
abstinent at the first follow-up; however, 5% (n ⫽
29) had relapsed to smoking at their first follow-up
(ie, relapsed former smokers). Among current smok-
ers, 70% (n ⫽ 173) had quit smoking at the time of
the follow-up (ie, abstinent smokers). However, 30%
(n ⫽ 75) continued to smoke at the time of their
follow-up (persistent smokers) [Table 2].
Approximately 1% of the patients described them-
selves as current cigar or pipe smokers, or smokeless
tobacco users (Table 2). The patients who were
utilizing other types of tobacco were not excluded
from the analysis because they were only a small
subset of the population, and there were no correla-
tions between other types of tobacco use and the
LCSS. Current alcohol use was reported by 51.5% of
the population.
Correlations Between Potential Covariates and the
Unadjusted LCSS Scores
Age, race, marital status, baseline and follow-up
treatment types (ie, surgery, radiation, and chemo-
therapy), the timing of treatment in relation to
CHEST / 126 / 6 / DECEMBER, 2004 1735
 Table 1—Demographic Characteristics and Cigarette Smoking Status*

Never- Former Relapsed Abstinent Persistent

Smoker Smoker Former Smoker Smoker Smoker Total
Characteristics (n ⫽ 180) (n ⫽ 562) (n ⫽ 29) (n ⫽ 173) (n ⫽ 75) (n ⫽ 1,019)

Age, yr
Mean (SD) 62.7 (14.0) 67.3 (9.5) 63.5 (10.3) 62.4 (9.5) 61.9 (10.2) 65.2 (10.8)
Range 17.0–87.0 34.0–90.0 37.0–78.0 38.0–82.0 36.0–82.0 17.0–90.0
Gender
Female 121 (67.2) 208 (37) 9 (31) 86 (49.7) 37 (49.3) 461 (45.2)
Male 59 (32.8) 354 (63) 20 (69) 87 (50.3) 38 (50.7) 558 (54.8)
Race
Missing 27 63 7 36 10 143
Alaskan/ 2 (1.3) 9 (1.8) 1 (4.5) 2 (1.5) 2 (3.1) 16 (1.8)
Indian
Black 0 (0) 2 (0.4) 0 (0) 4 (2.9) 0 (0) 6 (0.7)
White 147 (96.1) 485 (97.2) 21 (95.5) 131 (95.6) 63 (96.9) 847 (96.7)
Other 4 (2.6) 3 (0.6) 0 (0) 0 (0) 0 (0) 7 (0.8)
NSCLC stage
IA 70 (40.2) 158 (30.6) 4 (16) 48 (31) 19 (30.2) 299 (32)
IB 23 (13.2) 108 (20.9) 10 (40) 28 (18.1) 7 (11.1) 176 (18.8)
IIA 7 (4) 18 (3.5) 0 (0) 3 (1.9) 3 (4.8) 31 (3.3)
IIB 8 (4.6) 44 (8.5) 4 (16) 19 (12.3) 4 (6.3) 79 (8.5)
IIIA 27 (15.5) 84 (16.2) 1 (4) 23 (14.8) 9 (14.3) 144 (15.4)
IIIB 12 (6.9) 41 (7.9) 1 (4) 16 (10.3) 7 (11.1) 77 (8.2)
IV 27 (15.5) 64 (12.4) 5 (20) 18 (11.6) 14 (22.2) 128 (13.7)
SCLC stage
Limited 4 (100) 25 (61) 0 (0) 15 (88.2) 10 (83.3) 54 (69.2)
Extensive 0 (0) 16 (39) 4 (100) 2 (11.8) 2 (16.7) 24 (30.8)
*Values given as No. (%), unless otherwise indicated.

follow-up assessment (ie, before, during, or after),
the number of pack-years they had smoked, cancer
stage, cigar smoking, pipe smoking, smokeless to-
bacco use, and alcohol use were all assessed for
correlations with the LCSS score. We found no
clinically significant associations (Spearman correla-
tion coefficients all showed weak correlations). Fe-
male gender was associated with better LCSS scores
(p ⬍ 0.0001 [two-sample t test]). However, the mean
difference was 4.4 points on the LCSS. Although this
is less than a clinically significant difference (ie, 10
points), we did keep gender in the model along with
age at diagnosis, stage, and time of assessment. We
chose not to adjust for active treatment during the
QOL assessment as determined by a sensitivity
analysis (described below).
LCSS
Total Adjusted LCSS Scores: The adjusted mean
total LCSS scores for the never-smokers and
persistent smokers were 17.6 (SD, 4.02) and 28.7
(SD, 5.09), respectively (p ⬍ 0.0001), while
former smokers (abstinent and relapsed) had
adjusted mean total LCSS scores that were similar
to those of never-smokers (former smokers, 19.8;
SD, 4.12; never-smokers, 20.0; SD, 4.9). The
abstinent smokers had an adjusted mean total
LCSS score of 22.1 (SD, 4.03) [Fig 1, Table 3].
The relapsed former smokers were not included in
subsequent analyses due to the small number of
patients (29 patients) in this category, so the
adjusted score for this group should be interpreted
with caution. The lower QOL scores correspond to
a better QOL among never-smokers compared to
persistent smokers who had the highest scores and
the worst QOL.
Adjusted LCSS Scores for Individual Items: Sim-
ilar trends and findings as stated above for the total
LCSS score can be noted for all nine individual
LCSS items among the different categories of never-
smokers, former smokers, persistent smokers, and
abstinent smokers (Fig 1). Seven of the nine individ-
ual items (ie, appetite, fatigue, cough, shortness of
breath, lung cancer symptoms, illness affecting nor-
mal activities, and overall QOL) were clinically and
statistically significant. Although statistically signifi-
cant differences were noted for hemoptysis and pain,
these differences were not considered to be clinically
significant (criterion, a ⬎ 10-point difference). A test
for an increasing trend across the four groups was
performed. A statistically significant increasing trend
was observed among the groups of smokers (ie,
never-smokers, former smokers, and abstinent to
persistent smokers) [F ⫽ 12.34; p ⫽ 0.0003] as

1736 Clinical Investigations

 Table 2—Tobacco Use Characteristics and Cigarette Smoking Status*

Never- Former Abstinent Persistent

Smoker Smoker Relapsed Former Smoker Smoker Total
Variables (n ⫽ 180) (n ⫽ 562) Smoker (n ⫽ 29) (n ⫽ 173) (n ⫽ 75) (n ⫽ 1,019)

Smoking status at diagnosis

Never 180 (100) 0 (0) 0 (0) 0 (0) 0 (0) 180 (17.7)
Former 0 (0) 562 (100) 29 (100) 0 (0) 0 (0) 591 (58)
Current 0 (0) 0 (0) 0 (0) 173 (100) 75 (100) 248 (24.3)
PPDs
Missing 0 13 1 2 1 17
0 180 (100) 0 (0) 0 (0) 0 (0) 0 (0) 180 (18)
⬎ 0 to ⱕ 0.5 0 (0) 70 (12.8) 5 (17.9) 17 (9.9) 2 (2.7) 94 (9.4)
⬎ 0.5 to ⱕ 1 0 (0) 217 (39.5) 11 (39.3) 76 (44.4) 30 (40.5) 334 (33.3)
⬎1 0 (0) 262 (47.7) 12 (42.9) 78 (45.6) 42 (56.8) 394 (39.3)
Total pack-years
Missing 180 21 1 2 1 205
1–20 0 (0) 127 (23.5) 6 (21.4) 15 (8.8) 3 (4.1) 151 (18.6)
21–40 0 (0) 145 (26.8) 7 (25) 44 (25.7) 10 (13.5) 206 (25.3)
41–60 0 (0) 122 (22.6) 8 (28.6) 60 (35.1) 28 (37.8) 218 (26.8)
ⱖ 61 0 (0) 147 (27.2) 7 (25) 52 (30.4) 33 (44.6) 239 (29.4)
Smoked cigars
No 152 (97.4) 378 (86.1) 22 (81.5) 124 (89.9) 57 (95) 733 (89.4)
Yes
Current† 0 (0) 6 (1.4) 1 (3.7) 1 (0.7) 1 (1.7) 9 (1.1)
In the past 4 (2.6) 55 (12.5) 4 (14.8) 13 (9.4) 2 (3.3) 78 (9.5)
Smoked pipes
No 154 (98.7) 378 (86.1) 23 (85.2) 125 (90.6) 57 (95) 737 (89.9)
Yes
Current† 0 (0) 7 (1.6) 1 (3.7) 1 (0.7) 0 (0) 9 (1.1)
In the past 2 (1.3) 54 (12.3) 3 (11.1) 12 (8.7) 3 (5) 74 (9)
Used smokeless tobacco
No 154 (99.4) 414 (94.7) 24 (88.9) 132 (95.7) 57 (95) 781 (95.6)
Yes
Current† 0 (0) 5 (1.1) 1 (3.7) 1 (0.7) 1 (1.7) 8 (1)
In the past 1 (0.6) 18 (4.1) 2 (7.4) 5 (3.6) 2 (3.3) 28 (3.4)
Current† alcohol use
No 99 (60.4) 239 (45) 15 (57.7) 76 (45.8) 36 (50) 465 (48.5)
Yes 65 (39.6) 292 (55) 11 (42.3) 90 (54.2) 36 (50) 494 (51.5)
*Values given as No. (%).
†Current use is reported at baseline only.

shown in Figure 1 with increasing (worse) scores across
the groups.
Sensitivity Analysis for Treatment During Follow-
up Assessment
There were 11 patients who completed the follow-up
assessment while they were receiving chemotherapy
and/or radiation. There were also 167 patients who had
reported receiving chemotherapy, but no end date was
provided. Therefore, it was possible that these patients
were receiving chemotherapy during their assessment
and as a result might have higher (worse) LCSS scores.
We developed a second multivariate regression model
that, in addition to the other covariates (ie, age, gender,
stage, and time of assessment), also adjusted for any
active treatment received during the follow-up assess-
ment. The model was run under two different scenarios
with an estimated minimal length of chemotherapy (3
months) and maximal (7 months) length of chemother-
apy (with 1 month of recovery from chemotherapy).
We found no differences in the model under either
scenario compared to our original model (data not
shown).
Dose-Response Data Across All Times
We evaluated the dose-response trend of the
LCSS scores and PPDs, as well as of the total
number of pack-years of tobacco use among the
persistent smokers, abstinent smokers, and former
smokers. The hypothesis was that for an increasing
number of PPDs smoked per day as well as for an
increasing total number of pack-years smoked that
the persistent smokers would have a lower QOL.
The abstinent smokers and former smokers also were
assessed. This portion of the project was viewed as
exploratory and hypothesis-generating due to the

www.chestjournal.org CHEST / 126 / 6 / DECEMBER, 2004 1737

 Figure 1. Adjusted mean LCSS scores. Increasing trend across smoker groups from never-smokers to
persistent smokers, p ⫽ 0.0003; for each individual item score and the total LCSS score for
never-smokers vs persistent smokers, p ⱕ 0.0001. All differences were clinically significant except for
hemoptysis and pain.

small number of patients in each category, as de-
scribed in Table 2. The LCSS scores were adjusted
for age, gender, stage, and time of follow-up evalu-
ation.
Among the persistent smokers, we found no
clinically or statistically significant dose responses
for increasing (worse) LCSS scores and increased
amount of tobacco use when they were described
as PPDs or total number of pack-years (data not
shown). However, a pattern of increasing LCSS
scores was noted for increasing PPDs smoked, but
not for increasing total number of pack-years. The
total number of smokers was small (74 smokers),
and the number of smokers using ⱕ 0.5 PPD (2
smokers) and with a history of 1 to 20 pack-years
(3 smokers) was very small, so the results should
be interpreted with caution. Similarly, among the
abstinent smokers we found no clinically or statis-
tically significant differences for worsening LCSS
scores between any of the PPD groups or total
pack-year groups. Among the former smokers,
there were statistically significant differences;
however, no clinically significant differences (the
majority were differences of only 2 to 3 points)
were found for the PPD groups or for the total
pack-year groups (data not shown).

Table 3—Adjusted Mean LCSS Scores for Never-Smokers vs Persistent Cigarette Smokers

Never- Persistent
Smokers Smokers Difference in Means
Between Persistent
LCSS Item Mean (SD) Mean (SD) and Never-Smokers*

Appetite 16.1 (4.88) 26.1 (5.57) 10

Fatigue 31.0 (5.14) 44.4 (6.24) 13.4
Coughing 17.5 (4.53) 34.5 (5.66) 17
Dyspnea 23.4 (4.50) 37.0 (5.96) 13.6
Hemoptysis 1.6 (0.81) 5.2 (1.27) 3.5†
Pain 8.8 (2.07) 16.2 (2.53) 7.4†
Symptomatic distress 8.8 (3.46) 19.1 (5.34) 10.3
Effect on activities 14.7 (5.53) 26.6 (7.84) 11.9
Overall QOL 17.4 (6.35) 31.1 (8.03) 13.7
LCSS total score 17.6 (4.02) 28.7 (5.09) 11.1
*All differences shown are statistically significant. Rank sum p values all ⬍ 0.001.
†Difference not clinically significant.
1738 Clinical Investigations

www.chestjournal.org
Study/Year
General patient populations:
Stewart et al8/1995
Tillmann and Silcock1/
1997
Wilson et al2/1999
Specific patient populations:
Maxwell and Hirdes6/
1993
Sippel et al5/1999
Taira et al3/2000
Turner et al4/2001
*Modified ⫽ includes only 20 items; SF-36 ⫽ Medical Outcomes Study 36-item short-form general health survey; AQLQ ⫽ asthma quality of life questionnaire; PTCA ⫽ percutaneous transluminal
coronary angioplasty; MOS-HIV ⫽ modified outcomes survey scale adapted for patients with HIV.
†The term health status was defined differently on all three surveys.
CHEST / 126 / 6 / DECEMBER, 2004
1739
Table 4 —Select Prior Studies on Tobacco and QOL*
Patients,
Target Population No. Design Instrument QOL Findings Comments
Adults, smoking 350 Observational, Modified SF-36 Those who quit smoking had improved 70% response rate; QOL and smoking
cessation trial in longitudinal psychological well-being, depression, status assessments obtained separately;
United States anxiety, energy and current health utilized only a subset of the SF-36
perceptions compared to those who scale
relapsed
Adults in Scotland 1,665 Cross-sectional SF-36 and EuroQol Current smokers had lower QOL in 59–64% response rate; current smokers
survey the general health, vitality, and also reported more cough, phlegm,
mental health dimensions compared shortness of breath on exertion, and
to ex-smokers wheezes than ex-smokers
Adults in South 3,010 Two cross- SF-36 Smokers had a lower QOL in all 8 72–74% response rate
Australia sectional dimensions compared to never-
surveys smokers with a dose response for
heavier vs lighter smokers
Elderly (ⱖ 65 yr) in 9,023 Three cross- Health status† Current male and female smokers 79–95% response rate; one survey
Canada sectional reported more respiratory problems, allowed proxy responses; men used
surveys and men also reported trouble more analgesic medications and
walking and climbing stairs women utilized more CNS and GI
medications
Asthmatics in United 619 Cross-sectional, AQLQ and SF-36 Current smokers had a depressed Clinical significant difference not stated,
States longitudinal mood and more breathlessness on but ⱖ 10-point difference for physical
the AQLQ, and worse physical functioning, vitality, pain, and general
functioning, mental health, vitality, health
pain, and general health in the SF-
36 than the never-smokers
PTCA patients in 1,432 Cross-sectional, SF-36 Quitters had larger improvements in 80% response rate
United States longitudinal QOL following PTCA than did
nonsmokers and persistent smokers
HIV patients in 548 Cross-sectional, MOS-HIV Current smokers had worse physical 71% response rate; trend for those with
United States longitudinal functioning, pain, energy, role symptoms to have low QOL scores
functioning, and cognitive
functioning
 Discussion
Our hypothesis regarding the relationship be-
tween the smoking status of a lung cancer survivor
and their QOL was supported by our data. Contin-
ued cigarette smoking is related to a lung cancer
survivor experiencing a relative deficit in their QOL.
Thirty percent of the patients continued to smoke
despite being diagnosed with lung cancer. Persistent
smokers had worse appetite, fatigue, coughing, dys-
pnea, symptomatic distress, effect on activities, and
overall QOL compared to never-smokers, with for-
mer smokers and abstinent smokers having interme-
diate adjusted LCSS scores. No dose-response trend
was found between the number of PPDs or the total
number of pack-years utilized and the adjusted
LCSS scores, but the small number of patients in
each of the PPD and total pack-year categories limits
these inferences.
Our findings are comparable to those found in the
literature regarding how smoking effects QOL in
both the general patient population and in specific
subgroups (Table 4). In general, these selected
studies1– 6,8 found that patients who continued to
smoke had worse assessments of their own QOL. For
example, smokers thought that their general health
was worse, and they had lower assessments of vitality
and mental health across the different studies. Some
reported that these patients also reported more
difficulty with physical functioning,4,5 and correlated
the worse QOL findings with respiratory problems5
and trouble walking and climbing stairs.6 Additional
studies7,20 –24 with similar findings were not included
in Table 4 for simplicity. Other studies have ad-
dressed QOL in lung cancer patients, but smoking
history was not a major focus of these reports. One
report25 found that smoking status was not predictive
of QOL; however, this study was performed only in
long-term survivors and had a small number of
subjects, and the authors suggested that further
study in this area is needed. On the other hand,
current smokers undergoing surgery for lung cancer
had lower mental health dimension and vitality
scores than did former smokers and never-smokers.
While a cross-sectional study cannot prove causa-
tion, we can hypothesize that either smokers per-
ceive themselves as having a worse QOL and choose
to continue smoking or that the smoking itself causes
them to have a worse QOL. However, we observed
that the smokers who were abstinent from tobacco
use at the time of their first LCSS assessment had a
better QOL, suggesting that a tobacco intervention
may lead to improvements in QOL in lung cancer
patients. This information may provide oncologists
with additional evidence for recommending that all
of their lung cancer patients stop smoking.
1740
There are limitations to this study. First, selection
bias due to self-referral or physician-referral and
survival bias for being able to answer LCSS ques-
tions may have influenced the strength of our results.
These potential biases are reflected by the differ-
ences between the respondents and nonrespondents
to our study follow-up (eg, age at diagnosis, disease
stage, and smoking status). To correct for this, our
models were adjusted for age and stage of disease at
diagnosis, and were stratified by smoking status.
Second, the race and ethnicity of the study sample
was predominantly white, so the generalizability of
the study might be compromised by the limited
sample of underrepresented minorities. Third, the
patients’ smoking status may have fluctuated be-
tween diagnosis and follow-up, and biochemical
confirmation of tobacco use status was not per-
formed. It is known that a small percentage of
subjects intentionally report incorrect smoking sta-
tus, especially among adults.26 The Society for Re-
search on Nicotine and Tobacco Subcommittee on
Biochemical Verification27 has reported that bio-
chemical verification might affect outcome among
pregnant women with a history of alcohol use or
depression as well as among patients in Veterans
Affairs hospitals after surgery. However, verifica-
tion did not affect the outcome among cancer
patients, drug dependent patients, or primary care
patients, thus supporting our contention that it is
likely that a small percentage of patients falsely
reported their smoking status.28 Finally, the num-
bers of patients in the dose-response categories
are relatively small, so the data for them should be
interpreted with caution.
Despite these limitations, the study has several
strengths. To our knowledge, our study is among the
first published reports to specifically address how
cigarette smoking affects QOL among lung cancer
survivors. The large number of patients who com-
pleted the assessments and provided smoking status
enabled us to model the LCSS scores, and to adjust
for age, gender, disease stage, and time of assess-
ment. Changes over time or time trends will be
forthcoming as we gather additional data from the
patients. We hope to be able to provide data on the
relapsed former smokers in the future. Survival-
related outcomes and treatment-related outcomes,
as they relate to smoking status, will be assessed as
future aims with the entire cohort.
ACKNOWLEDGMENT: We thank Paul Novotny, Brent Wil-
liams, and Julian Molina for their helpful suggestions during our
research ideas meetings, and Jessica Gardner for her assistance
with manuscript production.
Clinical Investigations
 References
1 Tillmann M, Silcock J. A comparison of smokers’ and ex-
smokers’ health-related quality of life. J Public Health Med
1997; 19:268 –273
2 Wilson D, Parsons J, Wakefield M. The health-related qual-
ity-of-life of never smokers, ex-smokers, and light, moderate,
and heavy smokers. Prev Med 1999; 29:139 –144
3 Taira DA, Seto TB, Ho KK, et al. Impact of smoking on
health-related quality of life after percutaneous coronary
revascularization. Circulation 2000; 102:1369 –1374
4 Turner J, Page-Shafer K, Chin DP, et al. Adverse impact of
cigarette smoking on dimensions of health-related quality of
life in persons with HIV infection. AIDS Patient Care STDs
2001; 15:615– 624
5 Sippel JM, Pedula KL, Vollmer WM, et al. Associations of
smoking with hospital-based care and quality of life in
patients with obstructive airway disease. Chest 1999; 115:
691– 696
6 Maxwell CJ, Hirdes JP. The prevalence of smoking and
implications for quality of life among the community-based
elderly. Am J Prev Med 1993; 9:338 –345
7 Shaw JW, Coons SJ, Foster SA, et al. Responsiveness of the
Smoking Cessation Quality of Life (SCQoL) questionnaire.
Clin Ther 2001; 23:957–969
8 Stewart AL, King AC, Killen JD, et al. Does smoking
cessation improve health-related quality-of-life? Ann Behav
Med 1995; 17:331–338
9 Browman GP, Wong G, Hodson I, et al. Influence of cigarette
smoking on the efficacy of radiation therapy in head and neck
cancer. N Engl J Med 1993; 328:159 –163
10 Browman GP, Mohide EA, Willan A, et al. Association
between smoking during radiotherapy and prognosis in head
and neck cancer: a follow-up study. Head Neck 2002; 24:
1031–1037
11 Videtic GM, Stitt LW, Dar AR, et al. Continued cigarette
smoking by patients receiving concurrent chemoradiotherapy
for limited-stage small-cell lung cancer is associated with
decreased survival. J Clin Oncol 2003; 21:1544 –1549
12 Tammemagi CM, Neslund-Dudas C, Simoff M, et al. Smok-
ing and lung cancer survival: the role of comorbidity and
treatment. Chest 2004; 125:27–37
13 O’Shaughnessy JA, Wittes RE, Burke G, et al. Commentary
concerning demonstration of safety and efficacy of investiga-
tional anticancer agents in clinical trials. J Clin Oncol 1991;
9:2225–2232
14 American Joint Committee on Cancer. AJCC cancer staging
www.chestjournal.org
manual. 6th ed. Philadelphia, PA: Lippincott-Raven Publish-
ers, 2002
15 Yang P, Wentzlaff KA, Katzmann JA, et al. Alpha1-antitrypsin
deficiency allele carriers among lung cancer patients. Cancer
Epidemiol Biomarkers Prev 1999; 8:461– 465
16 Hollen PJ, Gralla RJ, Kris MG, et al. Quality of life assess-
ment in individuals with lung cancer: testing the Lung Cancer
Symptom Scale (LCSS). Eur J Cancer 1993; 29A(suppl):S51–
S58
17 Hollen PJ, Gralla RJ, Kris MG, et al. Measurement of quality
of life in patients with lung cancer in multicenter trials of new
therapies: psychometric assessment of the Lung Cancer
Symptom Scale. Cancer 1994; 73:2087–2098
18 Cohen J. Statistical power analysis for the behavioral sciences.
2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates, 1988
19 Sloan JA, Symonds T, Vargas-Chanes D, et al. Practical
guidelines for assessing the clinical significance of health
related quality of life changes within clinical trials. Drug Inf J
2003; 37:23–31
20 Manning MR, Osland JS, Osland A. Work-related conse-
quences of smoking cessation. Acad Management J 1989;
32:606 – 621
21 Olufade AO, Shaw JW, Foster SA, et al. Development of the
Smoking Cessation Quality of Life questionnaire. Clin Ther
1999; 21:2113–2130
22 Woolf SH, Rothemich SF, Johnson RE, et al. The functional
status of inner-city primary care patients: diminished function
in a family practice population and its potential determinants.
J Fam Pract 1998; 47:312–315
23 Hirdes JP, Maxwell CJ. Smoking cessation and quality of life
outcomes among older adults in the Campbell’s Survey on
Well-Being. Can J Public Health 1994; 85:99 –102
24 Lyons RA, Lo SV, Littlepage BNC. Perception of health
amongst ever-smokers and never-smokers: a comparison us-
ing the SF-36 Health Survey Questionnaire. Tobacco Control
1994; 3:213–215
25 Sarna L, Padilla G, Holmes C, et al. Quality of life of
long-term survivors of non-small-cell lung cancer. J Clin
Oncol 2002; 20:2920 –2929
26 Patrick DL, Cheadle A, Thompson DC, et al. The validity of
self-reported smoking: a review and meta-analysis. Am J
Public Health 1994; 84:1086 –1093
27 SRNT Subcommittee on Biochemical Verification. Biochem-
ical verificaiton of tobacco use and cessation. Nicotine To-
bacco Res 2002; 4:149 –159
28 de Andrade M, Ebbert JO, Wampfler J, et al. Environmental
tobacco smoke exposure in women with lung cancer. Lung
Cancer 2004; 43:127–134
CHEST / 126 / 6 / DECEMBER, 2004 1741