GENERAL THORACIC
ORIGINAL ARTICLES: GENERAL THORACIC
GENERAL THORACIC SURGERY:
The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org.
To take the CME activity related to this article, you must have either an STS member or an
individual non-member subscription to the journal.
Impact of Smoking Cessation Before Resection of
Lung Cancer: A Society of Thoracic Surgeons
General Thoracic Surgery Database Study
David P. Mason, MD, Sreekumar Subramanian, MD, Edward R. Nowicki, MD, MS,
Joshua D. Grab, MS, Sudish C. Murthy, MD, PhD, Thomas W. Rice, MD, and
Eugene H. Blackstone, MD
Department of Thoracic and Cardiovascular Surgery, Heart and Vascular Institute, and Department of Quantitative Health
Sciences, Research Institute, Cleveland Clinic, Cleveland, Ohio; and Duke Clinical Research Institute, Duke University, Durham,
North Carolina
Background. Smoking cessation is presumed to be
beneficial before resection of lung cancer. The effect of
smoking cessation on outcome was investigated.
Methods. From January 1999 to July 2007, in-hospital
outcomes for 7990 primary resections for lung cancer in
adults were reported to the Society of Thoracic Surgeons
General Thoracic Surgery Database. Risk of hospital
death and respiratory complications was assessed according to timing of smoking cessation, adjusted for clinical
confounders.
Results. Hospital mortality was 1.4% (n ⴝ 109), but
1.5% in patients who had smoked (105 of 6965) vs 0.39%
in those who had not (4 of 1025). Compared with the
latter, risk-adjusted odds ratios were 3.5 (p ⴝ 0.03), 4.6
(p ⴝ 0.03), 2.6 (p ⴝ 0.7), and 2.5 (p ⴝ 0.11) for those whose
timing of smoking cessation was categorized as current
smoker, quit from 14 days to 1 month, 1 to 12 months, or
more than 12 months preoperatively, respectively. Prevalence of major pulmonary complications was 5.7% (456
of 7965) overall, but 6.2% in patients who had smoked
(429 of 6941) vs 2.5%% in those who had not (27 of 1024).
Compared with the latter, risk-adjusted odds ratios were
1.80 (p ⴝ 0.03), 1.62 (p ⴝ 0.14), 1.51 (p ⴝ 0.20), and 1.29 (p ⴝ
0.3) for those whose timing of smoking cessation was
categorized as above.
Conclusions. Risks of hospital death and pulmonary
complications after lung cancer resection were increased
by smoking and mitigated slowly by preoperative cessation. No optimal interval of smoking cessation was
identifiable. Patients should be counseled to stop smoking irrespective of surgical timing.
M
patients who had never smoked vs those who quit
smoking at increasing time intervals before resection for
lung cancer.
ost patients who are candidates for pulmonary
resection are past or present smokers [1]. A common problem is counseling smokers before the operation
about smoking cessation and how it will affect their
operative risk. Some studies suggest that active smoking
increases operative risk, although to what extent remains
unclear [2– 4]. Should resection be delayed in favor of a
period of smoking cessation or performed immediately to
minimize cancer progression? To answer this question,
we queried the Society of Thoracic Surgeons (STS) General Thoracic Surgery Database and compared inhospital mortality and major pulmonary complications of
Accepted for publication April 1, 2009.
Presented at the Forty-fifth Annual Meeting of The Society of Thoracic
Surgeons, San Francisco, CA, Jan 26 –28, 2009.
Address correspondence to Dr Mason, Cleveland Clinic, Department of
Thoracic and Cardiovascular Surgery, 9500 Euclid Ave, Mail Stop J4-1,
Cleveland, OH 44195; e-mail: [email protected].
© 2009 by The Society of Thoracic Surgeons
Published by Elsevier Inc
(Ann Thorac Surg 2009;88:362–71)
© 2009 by The Society of Thoracic Surgeons
Patients and Methods
Data Source
The STS established the ongoing prospective General
Thoracic Surgery Database on January 1, 1999. Data are
submitted voluntarily for quality monitoring by multiple
hospitals, group practices, and surgeons throughout the
United States, which for this study included 79 centers.
Each center or surgeon completes a standardized form
that is keyed into certified software and harvested
annually for submission to the Duke Clinical Research
Institute, which is charged with maintaining and analyzing these data in compliance with the Health Insurance Portability and Accountability Act (HIPAA) of
1996. Variables collected include patient demograph0003-4975/09/$36.00
doi:10.1016/j.athoracsur.2009.04.035
Ann Thorac Surg
2009;88:362–71
Table 1. Characteristics of Patients and Their Therapy According to Smoking Status
Preoperative Smoking Cessation Interval
Current Smoker
(n ⫽ 1595)
Characteristic
a
⬎12 mo (n ⫽ 4026)
Never Smokeda
(n ⫽ 1025)
No. (%)b
Mean ⫾ SD
No. (%)b
Mean ⫾ SD
No. (%)b
Mean ⫾ SD
No. (%)b
Mean ⫾ SD
No. (%)b
Mean ⫾ SD
p Value
1595
62 ⫾ 10
404
62 ⫾ 10
940
63 ⫾ 10
4026
70 ⫾ 9.4
1025
64 ⫾ 15
⬍0.0001
⬍0.0001
26 ⫾ 6.3
192 (48)
212 (52)
385
26 ⫾ 5.8
473 (50)
467 (50)
878
26 ⫾ 5.9
2113 (52)
1913 (48)
3775
28 ⫾ 5.5
311 (30)
714 (70)
960
28 ⫾ 6.6
⬍0.0001
⬍0.0001
773 (48)
822 (52)
1514
702 (44)
893 (56)
143 (35)
261 (65)
398 (42)
542 (58)
1876 (47)
2150 (53)
514 (50)
511 (50)
⬍0.0001
1274
951
75 ⫾ 20
68 ⫾ 21
74 ⫾ 20
69 ⫾ 19
331
249
774
597
76 ⫾ 20
69 ⫾ 20
3133
2404
79 ⫾ 22
72 ⫾ 23
721
546
91 ⫾ 22
84 ⫾ 22
.9
1300 (82)
284 (18)
351 (87)
52 (13)
744 (81)
180 (19)
2987 (76)
966 (22)
902 (89)
107 (11)
1427 (90)
156 (9.9)
367 (92)
34 (8.5)
736 (79)
191 (21)
3422 (87)
527 (13)
878 (87)
131 (13)
1462 (92)
121 (7.6)
374 (93)
26 (6.5)
794 (86)
132 (14)
3566 (90)
382 (9.7)
932 (92)
79 (7.8)
114 (7.1)
1185 (74)
296 (19)
32 (7.9)
319 (79)
53 (13)
109 (12)
682 (73)
149 (16)
261 (6.5)
2895 (72)
870 (22)
51 (5.0)
731 (71)
243 (24)
.1
⬎0.9
⬍0.0001
Includes patients who smoked fewer than 100 cigarettes in their lifetime.
BMI ⫽ body mass index;
1–12 mo (n ⫽ 940)
CAD ⫽ coronary artery disease;
b
Number of patients with data available.
Dlco ⫽ diffusing capacity of lung for carbon monoxide;
FEV1 ⫽ forced expiratory volume in 1 second;
SD ⫽ standard deviation.
MASON ET AL
SMOKING CESSATION AND LUNG RESECTION
Demographics
Age, y
Gender
Male
Female
BMI, kg/m2
Zubrod score
Asymptomatic
Symptomatic
Spirometry
FEV1 (% predicted)
Dlco (% predicted)
CAD
No
Yes
Pre-op chemotherapy
No
Yes
Pre-op radiotherapy
No
Yes
Resection type
Pneumonectomy
Lobectomy or segmentectomy
Wedge
⬎14 d–1 mo (n ⫽ 404)
363
GENERAL THORACIC
GENERAL THORACIC
364
MASON ET AL
SMOKING CESSATION AND LUNG RESECTION
Ann Thorac Surg
2009;88:362–71
Group 2: Quit more than 14 days to 1 month before
operation
Group 3: Quit 1 to 12 months before operation
Group 4: Quit more than 12 months before operation
Group 5: Never smoked, or smoked fewer than 100
cigarettes in their lifetime
Fig 1. Interval of smoking cessation before resection for lung cancer
among 6941 patients with a history of smoking.
Among the 7990 patients, 6965 (87%) had a smoking
history, and 1025 (13%) had never smoked. Of the 6965
patients with a history of smoking, 4026 (58%) had quit
smoking more than 12 months preoperatively. Among
the remaining patients with a history of smoking, 1595
(23%) were current smokers, 404 (6%) had quit more than
14 days to 1 month before resection, and 940 (12%) had
quit between 1 and 12 months before resection (Fig 1).
End Points
ics, comorbidities, smoking history, pulmonary function testing, clinical and pathologic cancer staging,
surgical details, postoperative events, and hospital and
30-day mortality [5].
Patient Population
From January 1999 to July 2007, 7990 patients underwent
pulmonary resection for primary lung cancer whose data
were available on smoking and smoking cessation, hospital death, and postoperative pulmonary complications.
Excluded from analysis were patients aged younger than
18 years, those undergoing emergency operations, and
those with missing data on age, gender, or operation
date. Patient and cancer characteristics and details of
pulmonary resection, stratified by smoking status and
timing of smoking cessation, are reported in Table 1.
Pulmonary complications could not be assessed in 25
patients who died on the day of operation; thus, 7965
patients were available for analyzing occurrence of pulmonary complications.
Smoking Definitions
Patients were stratified into five groups by smoking
status and timing of smoking cessation according to STS
data collection fields:
Group 1: Active smoker or quit within 2 weeks of operation (current smoker)
HOSPITAL MORTALITY. Hospital mortality was defined as
death occurring within the hospitalization during which
the operation was performed. This was chosen over
either 30-day or operative mortality because of increased
completeness of data.
PULMONARY COMPLICATIONS. Pulmonary complications collected on the STS data form and believed to importantly
affect the postoperative course were chosen to form a
composite end point. These included prolonged ventilation (⬎ 48 hours postoperatively), need for reintubation,
atelectasis requiring bronchoscopy, tracheostomy, pneumonia, and development of acute respiratory distress
syndrome.
Data Analysis
A nonparsimonious mixed model with logit link was
used to generate risk-adjusted comparisons of hospital
mortality and pulmonary complications according to
smoking status and timing of smoking cessation. Individual institutions were incorporated as random effects.
Patient characteristics (including spirometry and packyears of smoking), pathologic cancer classifications, prior
cancer therapy, comorbidities, clinical status at time of
operation, and details of the pulmonary resection were
selected to be included in the models for risk adjustment
according to their clinical relevance (Appendix). Categories of some ordinal variables were collapsed because of
Table 2. Outcome After Resection for Lung Cancer According to Smoking Status
Category
Overall
Hospital
Mortality
Overall
Pulmonary
Complications
No. (% of 7990)
No. (%)
No. (% of 7965)a
No. (%)
1595 (20)
24 (1.5)
1590 (20)
110 (6.9)
404 (5.1)
940 (12)
4026 (50)
1025 (13)
7990 (100)
7 (1.7)
12 (1.3)
62 (1.5)
4 (0.39)
109 (1.4)
402 (5.0)
938 (12)
4011 (50)
1024 (13)
7965 (100)
25 (6.2)
60 (6.4)
234 (5.8)
27 (2.6)
456 (5.7)
Current smoker
Pre-op smoking cessation interval
⬎14 d–1 mo
1–12 mo
⬎12 mo
Never smokedb
Total
a
Excludes 25 patients who died on the day of operation.
b
Includes patients who smoked fewer than 100 cigarettes in their lifetime.
MASON ET AL
SMOKING CESSATION AND LUNG RESECTION
365
Table 3. Multivariable Logistic Model of Hospital Mortality After Resection for Lung Cancer
Coefficient ⫾ SE
Variable
a
Current smoker
Pre-op smoking cessation interval
⬎14 d–1 moa
1–12 moa
⬎12 moa
Age
Body mass index
Pack-years
FEV1 (% of predicted)
Female
Zubrod score
ASA risk class
Hypertension
Steroids
Heart failure
Coronary artery disease
Peripheral arterial disease
Rx-treated diabetes
Renal insufficiency
Pre-op chemo and/or radiotherapy
Cancer stage pT
Cancer stage pN
Lobectomy
Pneumonectomy
a
Est OR (95% CL)
1.2 ⫾ 0.58
3.5 (1.1, 11)
1.5 ⫾ 0.70
0.96 ⫾ 0.71
0.91 ⫾ 0.57
0.054 ⫾ 0.015
⫺0.013 ⫾ 0.027
⫺0.0017 ⫾ 0.0054
⫺0.012 ⫾ 0.0049
0.016 ⫾ 0.20
⫺0.021 ⫾ 0.19
0.34 ⫾ 0.17
0.16 ⫾ 0.21
0.66 ⫾ 0.42
0.53 ⫾ 0.34
0.35 ⫾ 0.24
0.74 ⫾ 0.22
⫺1.3 ⫾ 0.70
0.74 ⫾ 0.31
0.49 ⫾ 0.28
0.099 ⫾ 0.17
0.26 ⫾ 0.14
0.37 ⫾ 0.32
1.5 ⫾ 0.40
4.6 (1.2, 18)
2.6 (0.65, 11)
2.5 (0.82, 7.6)
1.1 (1.03, 1.09)
0.99 (0.94, 1.0)
1.00 (0.99, 1.01)
0.99 (0.98, 1.00)
1.02 (0.68, 1.5)
0.98 (0.67, 1.43)
1.4 (1.01, 1.9)
1.2 (0.78, 1.8)
1.9 (0.85, 4.4)
1.7 (0.87, 3.3)
1.4 (0.88, 2.3)
2.1 (1.4, 3.2)
0.27 (0.06, 1.2)
2.1 (1.2, 3.8)
1.6 (0.94, 2.8)
1.1 (0.79, 1.5)
1.3 (0.97, 1.7)
1.4 (0.77, 2.7)
4.3 (2.0, 9.5)
p Value
0.03
0.03
0.2
0.1
0.0002
0.6
0.8
0.01
0.9
0.9
0.04
0.4
0.11
0.12
0.15
0.0006
0.08
0.02
0.08
0.6
0.08
0.2
0.0002
Versus never smoked.
ASA ⫽ American Society of Anesthesiologists;
CLs ⫽ confidence limits;
ratio;
Rx ⫽ pharmacologically;
SE ⫽ standard error.
small numbers (Appendix). Missing values for variables
were handled in several ways, including removal of
patients from the analysis (rare), informative imputation,
and 10-fold multiple imputation (Appendix) [6]. Reported model-based statistics are based on the aggregated results for the multiple imputation.
FEV1 ⫽ forced expiratory volume in 1 second;
OR ⫽ odds
Results
Hospital Mortality
Although overall hospital mortality was low at 1.4% (109
of 7990), it was 1.5% among current or past smokers (105
of 6965) compared with 0.3% among patients who had
never smoked (4 of 1025, Table 2). Mortality was lower
among patients with longer intervals of smoking cessation before resection (Table 3, Fig 2); however, no sharp
transition to low risk was identified. Cause of death was
not available in the STS database.
Pulmonary Complications
Major pulmonary complications were infrequent (5.7% of
patients; 456 of 7965); however, they occurred more often in
current or past smokers (6.2%; 429 of 6941) than in those
who had never smoked (2.6%; 27 of 1024). The risk of
pulmonary complications steadily decreased as the interval
between smoking cessation and the operation increased
(Table 4, Fig 3), but there was no sharp transition suggesting
optimal timing of smoking cessation before operation.
Fig 2. Forest plot shows odds ratios and 95% confidence intervals
for hospital mortality according to interval of smoking cessation
compared with patients who have never smoked (odds ratio of 1.0).
Comment
Most patients undergoing pulmonary resection for bronchogenic carcinoma have some smoking history, and
GENERAL THORACIC
Ann Thorac Surg
2009;88:362–71
GENERAL THORACIC
366
MASON ET AL
SMOKING CESSATION AND LUNG RESECTION
Ann Thorac Surg
2009;88:362–71
Table 4. Multivariable Logistic Model of Pulmonary Complications After Resection for Lung Cancer
Coefficient ⫾ SE
Variable
a
Current smoker
Pre-op smoking cessation interval
⬎14 d–1 moa
1–12 moa
⬎12 moa
Age
Body mass index
Pack-years
FEV1 (% of predicted)
Female
Zubrod score
ASA risk class
Hypertension
Steroids
Heart failure
Coronary artery disease
Peripheral arterial disease
Rx-treated diabetes
Renal insufficiency
Pre-op chemo and/or radiotherapy
Cancer stage pT
Cancer stage pN
Lobectomy
Pneumonectomy
a
0.59 ⫾ 0.27
0.48 ⫾ 0.33
0.41 ⫾ 0.32
0.26 ⫾ 0.26
0.026 ⫾ 0.0051
⫺0.015 ⫾ 0.0098
0.0033 ⫾ 0.0023
⫺0.0096 ⫾ 0.0024
⫺0.12 ⫾ 0.10
0.14 ⫾ 0.11
0.20 ⫾ 0.090
0.20 ⫾ 0.095
0.61 ⫾ 0.19
0.58 ⫾ 0.19
0.23 ⫾ 0.11
0.16 ⫾ 0.12
⫺0.011 ⫾ 0.26
0.19 ⫾ 0.24
0.53 ⫾ 0.11
0.044 ⫾ 0.085
0.092 ⫾ 0.074
0.802 ⫾ 0.14
1.3 ⫾ 0.19
Est OR (95% CL)
p Value
1.8 (1.05, 3.1)
0.03
1.6 (0.85, 3.1)
1.5 (0.81, 2.8)
1.3 (0.77, 2.2)
1.03 (1.02, 1.04)
0.99 (0.97, 1.00)
1.00 (1.00, 1.01)
0.99 (0.99, 1.00)
0.89 (0.72, 1.09)
1.2 (0.92, 1.4)
1.2 (1.02, 1.5)
1.2 (1.01, 1.5)
1.8 (1.3, 2.7)
1.8 (1.2, 2.6)
1.3 (1.01, 1.6)
1.2 (0.92, 1.5)
0.99 (0.57, 1.7)
1.2 (0.76, 1.9)
1.7 (1.4, 2.1)
1.05 (0.88, 1.2)
1.1 (0.95, 1.3)
2.2 (1.7, 2.9)
3.6 (2.5, 5.2)
0.14
0.2
0.3
⬍0.0001
0.13
0.14
⬍0.0001
0.2
0.2
0.02
0.04
0.001
0.002
0.04
0.2
⬎0.9
0.4
⬍0.0001
0.6
0.2
⬍0.0001
⬍0.0001
Versus never smoked.
ASA ⫽ American Society of Anesthesiologists;
standard error.
CL ⫽ confidence limits;
many are active smokers [3]; however, the optimal timing
of smoking cessation before pulmonary resection remains uncertain. In delaying the operation, thoracic
surgeons must balance the risk of local tumor growth
(tumor doubling) [7], risk of metastasis, and patient
anxiety [8] against the benefit of reduced operative risk
[9]. This clinical dilemma arises frequently: Almost 30%
FEV1 ⫽ forced expiratory volume in 1 second;
OR ⫽ odds ratio;
SE ⫽
of patients in this STS database study were active smokers or had quit smoking within the previous month.
This study evaluated a large cohort of patients undergoing resection for lung cancer using registry data collected from multiple institutions, with a focus on determining the optimal timing of smoking cessation before
resection. When this study was conceived, the expectation was that smoking cessation would lead to rapid
lowering of risk, likely within weeks, such that risk would
be indistinguishable from that of patients who had never
smoked. Surprisingly, this return toward normal was
prolonged and an optimal time frame likely far greater
than would be clinically recommended before cancer
resection. Although any smoking, past or current, was
clearly associated with increased hospital mortality and
more pulmonary complications, overall mortality was
impressively low and major pulmonary complications
surprisingly few.
Principal Findings
Fig 3. Forest plot shows odds ratios and 95% confidence intervals
for pulmonary complications according to interval of smoking cessation compared with patients who have never smoked (odds ratio of
1.0).
MORTALITY. The low mortality after pulmonary resection
noted in this study was similar to recently published
multi-institutional studies of pulmonary resection for
malignancy performed by experienced surgeons [10, 11].
However, we found that any history of smoking, past or
current, was associated with a negative impact on mor-
tality after pulmonary resection, with large increases in
relative risk. When evaluating the timing of operation
related to smoking cessation, risk appeared to steadily
decrease as the interval of smoking cessation increased.
Reasons for this finding are likely multifactorial. Smoking cessation improves pulmonary function, both objectively and subjectively [12–14]. However, improvements
in spirometry, ciliary clearance, and immune mechanisms occur over prolonged periods [15, 16]. In addition,
a decrease in sputum production occurs weeks to months
after cessation [13, 17–19]. The combined, deleterious
effects of smoking likely contribute to mortality, but are
less with longer interval of cessation.
Although in this study relative risk appears higher for
patients who quit smoking more than 14 days to 1 month
before operation than for current smokers, confidence
intervals are wide and overlapping, and clinical importance, if any, is unclear. Contrary to the notion that a
short period of smoking cessation will result in reduced
surgical risk, these data demonstrate that even after a
year of smoking cessation, risk-adjusted mortality remains elevated compared with lifetime nonsmokers, suggesting that adverse effects never completely disappear.
This may be due to cardiovascular risk factors that,
although controlled for in our multivariable analysis, still
negatively affect operative outcome in smokers despite a
long period of smoking respite.
There is a paucity of information about the impact of
smoking cessation on hospital mortality after lung cancer
resection. The few studies that have been performed
have focused on long-term survival rather than inhospital death. Smoking cessation has been shown to be
associated with increased long-term survival, and active
smoking has the highest risk of cancer-related and overall death [20 –22]. This study demonstrates that although
hospital mortality is higher for smokers after pulmonary
resection, the risk appears to be modified by cessation
and improved with a longer interval of cessation. Therefore, surgeons should counsel smokers that risk remains
elevated regardless of timing of cessation, but that quitting holds a benefit that improves over time. Although
the relative risk of current smokers or recent quitters is
substantial, unduly delaying the operation does not seem
justified because of the low overall hospital mortality
noted in this study and the long period during which risk
remains elevated.
PULMONARY COMPLICATIONS. Occurrence of pulmonary complications in this study was considerably lower than in
other studies focusing on pulmonary complications after
pulmonary resection; in some studies, they occurred in
more than 50% of patients [23–25]. However, this may be
partly accounted for by definitions of pulmonary complications that differ widely from study to study. These
include the more subtle findings of atelectasis to the most
severe pulmonary complications such as acute respiratory distress syndrome. When defining pulmonary complications in this study, we chose only those we believed
importantly affected a patient’s postoperative course and
were potentially life threatening.
MASON ET AL
SMOKING CESSATION AND LUNG RESECTION
367
This study found that smoking had less of an effect on
pulmonary complications than on mortality. Risk of pulmonary complications steadily decreased among patients
with longer interval of smoking cessation, and risk began
to converge to that of patients who had never smoked.
Explanation for this finding likely includes the time
required for the deleterious impact on lung function to
recede on a macroscopic, microscopic, and functional
level.
That smoking increases the risk of pulmonary complications in many fields of surgery was noted decades ago
[26 –29]. However, the usefulness and optimal timing of
smoking cessation before operation to decrease pulmonary complications remain poorly defined. Two of the
largest studies were performed in cardiac surgery patients and evaluated the relationship between development of pulmonary complications and interval of smoking abstinence [30, 31]. The conclusion was that at least 8
weeks of smoking abstinence was necessary before any
reduction in pulmonary complications could be realized.
In fact, one study suggested a paradoxic increase in the
risk of pulmonary complications if the operation was
performed with less than 2 months of smoking abstinence [31], but confidence limits, which were not provided, undoubtedly widely overlapped.
In thoracic surgery, findings of increased risk of pulmonary complications related to smoking have been
demonstrated in several studies [23–25], although the
timing and effect of smoking cessation on pulmonary
complications after resection have been addressed in
only a few small studies:
●
●
●
Nakagawa and colleagues [32] studied 288 patients undergoing pulmonary resection and concluded that at least 4 weeks of smoking abstinence
was necessary for a discernible reduction in pulmonary complications. However, small sample
size and grouping together of all patients who
stopped smoking more than 4 weeks before the
operation limit the strength of this conclusion.
Barrera and colleagues [33] studied smoking cessation in 300 patients undergoing pulmonary resection and noted no differences in development
of pulmonary complications in patients who were
active smokers compared with those who had less
than 2 months of smoking cessation and those
who had more than 2 months of smoking cessation. Small sample size was also a limiting factor
in this study.
Vaporciyan and colleagues [34] evaluated pulmonary complications after pneumonectomy and
concluded that patients who smoked within 1
month of operation had the highest risk. Their
focus on pneumonectomy prevents generalization
of recommendations to lesser extent of pulmonary
resection.
The consistent decrease noted in postoperative pulmonary complications as interval of smoking cessation increased in this study suggests that surgeons can safely
counsel patients about the benefits of smoking cessation
GENERAL THORACIC
Ann Thorac Surg
2009;88:362–71
GENERAL THORACIC
368
MASON ET AL
SMOKING CESSATION AND LUNG RESECTION
preoperatively, regardless of the interval. Although the
relative risk of active smokers or recent quitters is substantial, unduly delaying the operation does not seem
justified due to the low overall risk of pulmonary complications and the long period during which risk remains
elevated.
Ann Thorac Surg
2009;88:362–71
4.
5.
Strengths and Limitations
This study has the advantage of being able to compare
short-term postoperative survival and pulmonary complications in a large number of patients from multiple
institutions for whom a wide range of patient, surgical,
and pathologic variables was available. Its limitations
include the self-reported and voluntary data collection
system that lacks formal auditing and may predispose to
underreporting of complications. The STS General Thoracic Surgery Database is relatively new and has not yet
been mined extensively for research such as this. In
addition, it does not currently have the penetration of the
STS Adult Cardiac Database, so the results presented
may not be representative of general thoracic surgery
outcomes in the United States.
No objective verification of smoking status was performed by nicotine metabolite testing; rather, reliance
was placed on patient self-reporting, which may be
systematically biased toward claiming a longer interval of
cessation than actual. In addition, coarse grouping of
interval of smoking cessation on the STS data collection
form precluded evaluating timing of cessation as a continuous variable.
Finally, spirometry data were missing for about onethird of patients, which, despite modern imputation methods, degrades the value of the contribution of these
variables to the analysis. These variables, in fact, distinguish nonsmokers from smokers.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Conclusions
Smoking is associated with increased hospital death and
risk of pulmonary complications after pulmonary resection for lung cancer, but risk declines with a longer
interval of smoking cessation. However, no optimal interval of smoking cessation was identified. Patients
should be counseled to stop smoking irrespective of
surgical timing and advised that their operative risk,
although elevated compared with nonsmokers, can be
favorably modified by cessation.
We thank Tess Parry for editorial assistance. This study was
supported in part by the Kenneth Gee and Paula Shaw, PhD,
Chair in Heart Research (Dr Blackstone) and the Daniel and
Karen Lee Endowed Chair in Thoracic Surgery (Dr Rice).
16.
17.
18.
19.
20.
21.
References
1. Alberg AJ, Samet JM. Epidemiology of lung cancer. Chest
2003;123(Suppl 1):21– 49S.
2. Smetana GW. Preoperative pulmonary evaluation. N Engl
J Med 1999;340:937– 44.
3. Sardari Nia P, Weyler J, Colpaert C, Vermeulen P, Van
Marck E, Van Schil P. Prognostic value of smoking status in
22.
23.
operated non-small cell lung cancer. Lung Cancer 2005;47:
351–9.
Wright CD, Gaissert HA, Grab JD, O’Brien SM, Peterson ED,
Allen MS. Predictors of prolonged length of stay after
lobectomy for lung cancer: a Society of Thoracic Surgeons
General Thoracic Surgery Database risk-adjustment model.
Ann Thorac Surg 2008;85:1857– 65; discussion 65.
Cardiothoracic Surgery Network. The Society of Thoracic Surgeons General Thoracic Surgery Database Data Collection
Form version 207. http:www.ctsnet.org/file/ThoracicDCFV2_
07_Nonannotated.pdf.
Rubin DB. Multiple imputation for non-response in surveys.
New York: Wiley; 1997.
Detterbeck FC, Gibson CJ. Turning gray: the natural history
of lung cancer over time. J Thorac Oncol 2008;3:781–92.
Byrne MM, Weissfeld J, Roberts MS. Anxiety, fear of cancer,
and perceived risk of cancer following lung cancer screening. Med Decis Making 2008;28:917–25.
Myrdal G, Lambe M, Hillerdal G, Lamberg K, Agustsson T,
Stahle E. Effect of delays on prognosis in patients with
non-small cell lung cancer. Thorax 2004;59:45–9.
Allen MS, Darling GE, Pechet TT, et al. Morbidity and
mortality of major pulmonary resections in patients with
early-stage lung cancer: initial results of the randomized,
prospective ACOSOG Z0030 trial. Ann Thorac Surg 2006;81:
1013–9; discussion 9 –20.
Meguid RA, Brooke BS, Chang DC, Sherwood JT, Brock MV,
Yang SC. Are surgical outcomes for lung cancer resections
improved at teaching hospitals? Ann Thorac Surg 2008;85:
1015–24; discussion 24 –5.
Buist AS, Sexton GJ, Nagy JM, Ross BB. The effect of smoking
cessation and modification on lung function. Am Rev Respir
Dis 1976;114:115–22.
Verbanck S, Schuermans D, Paiva M, Meysman M, Vincken
W. Small airway function improvement after smoking cessation in smokers without airway obstruction. Am J Respir
Crit Care Med 2006;174:853–7.
Bode FR, Dosman J, Martin RR, Macklem PT. Reversibility of
pulmonary function abnormalities in smokers. A prospective study of early diagnostic tests of small airways disease.
Am J Med 1975;59:43–52.
Kotani N, Kushikata T, Hashimoto H, Sessler DI, Muraoka
M, Matsuki A. Recovery of intraoperative microbicidal and
inflammatory functions of alveolar immune cells after a
tobacco smoke-free period. Anesthesiology 2001;94:999 –
1006.
Pearce AC, Jones RM. Smoking and anesthesia: preoperative
abstinence and perioperative morbidity. Anesthesiology
1984;61:576 – 84.
Camner P, Philipson K, Arvidsson T. Withdrawal of cigarette
smoking: a study on tracheobronchial clearance. Arch Environ Health 1973;26:90 –2.
Arcavi L, Benowitz NL. Cigarette smoking and infection.
Arch Intern Med 2004;164:2206 –16.
Yamashita S, Yamaguchi H, Sakaguchi M, et al. Effect of
smoking on intraoperative sputum and postoperative pulmonary complication in minor surgical patients. Respir Med
2004;98:760 – 6.
Sobue T, Suzuki T, Fujimoto I, Doi O, Tateishi R, Sato T.
Prognostic factors for surgically treated lung adenocarcinoma patients, with special reference to smoking habit. Jpn
J Cancer Res 1991;82:33–9.
Halpern MT, Gillespie BW, Warner KE. Patterns of absolute
risk of lung cancer mortality in former smokers. J Natl
Cancer Inst 1993;85:457– 64.
Tammemagi CM, Neslund-Dudas C, Simoff M, Kvale P.
Smoking and lung cancer survival: the role of comorbidity
and treatment. Chest 2004;125:27–37.
Stephan F, Boucheseiche S, Hollande J, et al. Pulmonary
complications following lung resection: a comprehensive
analysis of incidence and possible risk factors. Chest 2000;
118:1263–70.
MASON ET AL
SMOKING CESSATION AND LUNG RESECTION
24. Patel RL, Townsend ER, Fountain SW. Elective pneumonectomy: factors associated with morbidity and operative mortality. Ann Thorac Surg 1992;54:84 – 8.
25. Abolhoda A, Liu D, Brooks A, Burt M. Prolonged air leak
following radical upper lobectomy: an analysis of incidence
and possible risk factors. Chest 1998;113:1507–10.
26. Wightman JA. A prospective survey of the incidence of
postoperative pulmonary complications. Br J Surg 1968;55:
85–91.
27. Bluman LG, Mosca L, Newman N, Simon DG. Preoperative
smoking habits and postoperative pulmonary complications.
Chest 1998;113:883–9.
28. Theadom A, Cropley M. Effects of preoperative smoking
cessation on the incidence and risk of intraoperative and
postoperative complications in adult smokers: a systematic
review. Tob Control 2006;15:352– 8.
29. Moores LK. Smoking and postoperative pulmonary complications. An evidence-based review of the recent literature.
Clin Chest Med 2000;21:139 – 46, ix–x.
369
30. Warner MA, Divertie MB, Tinker JH. Preoperative cessation
of smoking and pulmonary complications in coronary artery
bypass patients. Anesthesiology 1984;60:380 –3.
31. Warner MA, Offord KP, Warner ME, Lennon RL, Conover
MA, Jansson-Schumacher U. Role of preoperative cessation
of smoking and other factors in postoperative pulmonary
complications: a blinded prospective study of coronary artery bypass patients. Mayo Clin Proc 1989;64:609 –16.
32. Nakagawa M, Tanaka H, Tsukuma H, Kishi Y. Relationship
between the duration of the preoperative smoke-free period
and the incidence of postoperative pulmonary complications
after pulmonary surgery. Chest 2001;120:705–10.
33. Barrera R, Shi W, Amar D, et al. Smoking and timing of
cessation: impact on pulmonary complications after thoracotomy. Chest 2005;127:1977– 83.
34. Vaporciyan AA, Merriman KW, Ece F, et al. Incidence of
major pulmonary morbidity after pneumonectomy: association with timing of smoking cessation. Ann Thorac Surg
2002;73:420 –5; discussion 5– 6.
Appendix
Variables Included in Logistic Regression Modeling
Levelsa
Variable
Smoking cessation
Age
Gender
Pack-years
Race
Body mass index, kg/m2
Current smoker or quit ⱕ2 weeks pre-op
Quit ⬎14 days to 1 month pre-op
Quit 1 to 12 months pre-op
Quit more than 12 months pre-op
Never smoked (or smoked ⬍ 100 cigarettes in
lifetime)
Continuous
Female
Continuous
African-American
Other (not African-American or Caucasian)
Caucasian
Continuous
Zubrod score
Symptoms but full ambulatory normal activity
No symptoms
ASA risk class
IV/V combined into single group III
I/II combined into single group
Hypertension
Steroids
Heart failure
Peripheral arterial disease
Cerebrovascular history
Yes/No
Yes/No
Yes/No
Yes/No
Any reversible or irreversible event
None
Yes/No
Serum creatinine ⱖ2.0 mg/dL and/or dialysis
None
Rx-treated diabetes
Renal insufficiency
Pre-op chemo and/or radiotherapy
FEV1 (% of predicted)
FVC (% of predicted)
Dlco (% of predicted)
Yes/No
Continuous
Continuous
Continuous
Missing Data Handling
Record excluded from analysis
Record excluded from analysis
Record excluded from analysis
Set to zero for nonsmokers;
multiple imputation for
smokers
Set to “other” (0.5% of records)
Either multiple imputation or set
to gender-specific median
within study population
Records with other Zubrod scores
or missing data excluded from
analysis
Set missing to I/II (conservative
approach)
Multiple
Multiple
Multiple
Multiple
Multiple
imputation
imputation
imputation
imputation
imputation
Multiple imputation
Set missing (⬍0.5%) to none
before multiple imputation
process
Multiple imputation
Multiple imputation (⬇20%)
Multiple imputation (⬇20%)
Multiple imputation (⬇40%)
Continued
GENERAL THORACIC
Ann Thorac Surg
2009;88:362–71
GENERAL THORACIC
370
MASON ET AL
SMOKING CESSATION AND LUNG RESECTION
Ann Thorac Surg
2009;88:362–71
Appendix Continued
Levelsa
Variable
Cancer stage pT
Cancer stage pN
Cancer stage pM
Surgery type
3 and 4 combined into single group 2
1, 0, and in situ as single group (treat as linear in
model, coding as 0 ⫽ 0/1/in situ, 1 ⫽ path T of 2,
2 ⫽ path T of 3 or 4)
2 and 3 combined into single group (code as 2)
1
0
(treat as linear variable)
Not used
Pneumonectomy
Lobectomy (includes segmentectomy)
Wedge resection (single/multiple)
Random effect parameter (to account for
clustering of similar patients within
STS participating sites)
a
Missing Data Handling
Multiple imputation for value of
p or missing
Multiple imputation for value of
p or missing
—
No missing data allowed here by
study design
No missing data allowed here by
study design
Reference level listed last or not at all.
ASA ⫽ American Society of Anesthesiologists;
BMI ⫽ body mass index;
Dlco ⫽ diffusing capacity of lung for carbon monoxide;
FEV1 ⫽
forced expiratory volume in 1 second;
FVC ⫽ forced vital capacity;
Rx ⫽ pharmacologically treated;
STS ⫽ Society of Thoracic Surgeons.
DISCUSSION
DR JOSEPH B. SHRAGER (Stanford, CA): That is very interesting. Wasn’t there a large VA [Veterans Affairs] study that
showed that there is actually increased mortality and complications if you quit within about 1 month of surgery and supposedly
due to a hypersecretory response after surgery? I wonder if you
can tell us in relation to that what the causes of mortality were in
those patients who had an elevated mortality if they quit within
15 days of the operation. Were those not pulmonary mortalities?
actually quit, they quit because they were too sick to be smoking
right before their surgery. I hate to say that, but that is actually
what we found.
DR SUBRAMANIAN: We do not have specific data on the cause
of hospital mortality for these patients. There is a large VA study
showing a paradoxically elevated risk. In our analysis of this
paper and others, we recognize that those are not homogeneous
groups of patients, and the confidence intervals of these data
appear to be very wide. So 2 people could look at the same data,
take into account the confidence intervals, and then make
different conclusions. The confidence intervals in our paper are
also very wide, leading to considerable overlap between the
groups.
DR DAVID TOM COOKE (Sacramento, CA): That was a very
good presentation. Is there a statistically significant difference
between the smoking groups amongst each other or are the
differences just trends?
DR SHRAGER: But there is overlap on the complication chart as
well as on the mortality chart.
DR SUBRAMANIAN: That’s right.
DR SHRAGER: So I am not sure why you are emphasizing
necessarily the pulmonary complication finding when the mortality finding was higher in patients who quit close to the time of
surgery.
DR DAVID H. HARPOLE (Durham, NC): I actually have the VA
data to clarify that. What we found was that there was an
association between people who stopped smoking and the
increased rates of mortality and morbidity, but it actually covaried with the comorbidities of the patients, and what we found in
our veteran population where smoking is endemic, the ones who
DR ROBERT CERFOLIO (Birmingham, Alabama): Dave, here is
an idea: If they are too sick to smoke, don’t operate on them.
DR HARPOLE: Exactly.
DR SUBRAMANIAN: That is a very good question. There are
no significant differences between groups.
DR COOKE: So looking at that data, if I have a stubborn patient
who really can’t get off the cigarettes, should I just look at that
data and say, well, statistically there is no difference between
that patient and someone who quit smoking 6 months prior, so
I should go ahead and resect, regardless of smoking status? The
second thing is, do you think there is any relevance in terms of
how much someone smokes? Do they smoke a pack a day or are
they down to 1 to 2 cigarettes a day?
DR SUBRAMANIAN: Those are excellent questions. With regard to your first question about the timing of operation, we
believe that you should continue with the preoperative evaluation of the patient if he or she refuses to quit smoking. We
advocate putting the patient in a smoking cessation program
and aggressively working toward getting the patient to quit
smoking. However, if the patient cannot quit, then continue with
your evaluation and schedule the operation.
For the second question, our data looked at pack-year smoking history as a confounding variable. However, we do not have
data on the interplay of various factors. For example, if you have
a significant pack-year smoking history and severely reduced
lung function, is there a benefit for you to stop smoking over
patients with more normal lung function?
DR DANIEL L. MILLER (Atlanta, GA): I think a very important
pulmonary complication which is not in your list is the requirement for supplemental oxygen therapy at dismissal. We have
seen retrospectively in the patients who smoked within that
2-week period that about one third of those patients required
oxygen therapy upon discharge. Do you have any data at all in
regards to that?
DR SUBRAMANIAN: No. That is not a variable listed on the
STS [Society of Thoracic Surgeons] database. I think on a larger
note, the point you bring up is that the database should be
expanded to collect additional variables. For example, as you
noticed in our group, we don’t have a separation of patients who
were active smokers and those who quit, say, a week ago vs 2
weeks ago. Some of the heterogeneity in the literature on the
effects of smoking cessation on outcomes is based on different
complications that have been included.
DR CERFOLIO: Dan, I have seen a lot of your patients. I think
they all go home on oxygen.
MASON ET AL
SMOKING CESSATION AND LUNG RESECTION
371
DR RICHARD R. O’REILLY (Bakersfield, CA): I was discouraged to see that even after a year of not smoking, they still had
increased complications. Somewhere I had heard that 8 weeks
was a good time, but most of my patients wouldn’t wait for 8
weeks until you operate on them, and if you operate immediately, then there is this bronchorrhea of cessation of smoking. So
some way or another, 3 weeks seemed to be optimum in my
practice in the old days, although you seem to show that 15 to 30
days is worse than 1 to 15 days.
DR SUBRAMANIAN: I would like to clarify. Although results
appear worse, outcomes of patients who quit smoking within 15
to 30 days of operation were not statistically significantly worse
than those of either patients who had quit between 0 and 14 days
or were active smokers.
DR CERFOLIO: One final question. How do we know the
accuracy of this? This is by patient report. I think if you really
talk to the patients, they tell you they stopped a month ago, but
if you really talk to them, they had a cigarette right before
surgery.
DR SUBRAMANIAN: That is an excellent point. Not only is
self-reporting of smoking at issue, but also pack-year smoking
history, because patients tend to underreport how much they
have smoked.
GENERAL THORACIC
Ann Thorac Surg
2009;88:362–71