Assessment of Operative Risk in Patients
Undergoing Lung Resection*
Importance of Predicted Pulmonary Function
David]. Kearney, M.D.; Thomas H. Lee, M.D.; John]. Reilly, M.D.,
F.C.C.A. ; Malcolm M. DeCamp, M.D.; and David]. Sugarbaker, M.D.,
F.C.P.P.
Objective: To evaluate the ability of preoperative
variables to identify patients at increased risk for
complications after lung resection and the usefulness
of predicted postoperative FEV 1 as a marker of risk
for adverse outcomes.
Design: Prospective analysis of a cohort of patients
undergoing pulmonary resection. Complication rates
were analyzed according to preoperative pulmonary
variables, demographic variables, procedure per-
formed, and predicted postoperative FEV 1 Predicted
postoperative FEV 1 was calculated using a formula
estimating the decline in preoperative FEV 1 based on
the number of bronchopulmonary segments removed
during surgery.
Setting: A major teaching hospital and tertiary refer-
ral center.
Patients: A consecutive series of patients undergoing
pulmonary resection.
Measurements and main results: Medical complica-
tions were recorded as part of an ongoing clinical
database. The overall complication rate was low (17
percent rate of any complication, 1 percent death
rate). Univariate predictors of complications included
age ~ 60, male sex, history of smoking, a pneumon-
ectomy procedure, and a low predicted postoperative
~ng cancer is the most common cause of cancer
death in the United States. 1 An estimated 168,000
new cases will occur this year, 2 and the majority will
die within the first year after diagnosis.3 For patients
with non-small-cell carcinoma presenting as localized
For editorial comment see page 654
disease, surgical resection remains the only potentially
curative option 4 The high rate of coexistent medical
conditions in lung cancer patients complicates referral
of such patients to surgery. Smoking is highly associ-
ated with the development of lung carcinoma;5 the
incidence increases further in smokers with COPD.B-9
Additionally, smoking and chronic lung disease are
*From the Respiratory Division( Dr. Reilly) , Division of Tho-
racic Surgery (Drs. DeCamp and Sugarbaker), Department of
Medicine (Dr. Lee) , Brigham & Women's Hospital, and De-
partme nt of Medicine (Dr. Kearney) , Beth Israel Hospital ,
Boston.
Manuscript received March 24, 1993; revision accepted July 23.
Reprint requests: Dr. Srfgarbaker, Division of Thoracic Surgenj,
Briglwmand Wom en's Hospital, Boston 02115.
FEV 1 Hypercarbia (~ 45 mm Hg) on preoperative
arterial glood gas analysis, desaturation on exercise
oximetry(~ 90 percent), and a preoperative FEV 1 less
than 1 L were not predictive of complications. When
the effect of these variables was controlled for in a
multivariate analysis, a low predicted postoperative
FEV 1 remained the only significant independent pre-
dictor of complications. For each 0.2 L decrease in
predicted FEV 1, the odds ratio for complications was
1.46 (95 percent confidence interval [CI] 1.2 to 1.8).
Conclusions: A low predicted postoperative FEV 1
appears to be the best indicator of patients at high
risk for complications, and it was the only significant
correlate of complications when the effect of other
potential risk factors was controlled for in a multivari-
ate analysis. Pulmonary resection should not be de-
nied on the basis of traditionally cited preoperative
pulmonary variables, and a prediction of postopera-
tive pulmonary function by a technique of simple
calculation may be useful to identify patients at
increased risk for medical complications.
(Chest 1994; 105: 153-59)
ABG =arterial blood gas; MVV =maximum voluntary
ventilation; PPOFEV 1 =predicted postoperative FEV 1
associated with atherosclerotic vascular disease, mak-
ing these conditions prevalent in patients with lung
carcinoma. Thus, the clinician is often faced with the
problem of considering lung resection as the only
curative treatment in a patient with significant comorbid
disease and ventilatory impairment.
A wide range of tests has been evaluated to aid the
identification of patients at high risk for developing
complications after lung resection. These tests have
included spirometry, arterial blood gas measurement
(ABC), exercise testing, and radionuclide lung scan-
ning. Previously reported studies have sought to de-
fine values for each of these modalities that would
identify patients at prohibitive risk. 10- 15 Demographic
variables such as male sex, advanced age, and smoking
status have been shown to identify groups at increased
risk. 6 814161i Reported complication rates after lung
resection have ranged from a 40 percent rate of death
or severe disability in the 1950s and 1960s to mortality
rates of9 to 27 percent in the 1970s} 2 18' 20 Advances in
operative technique and postoperative care have led
CHEST I 105 I 3 I MARCH, 1994 753
to a further decline in postoperative complication Table 1 - Baeline Characteri.tic of 331 Patient
rates, with mortality rates of 2 to 11 percent reported Undergoing Lung Reection
in recent studies.15 21 "26
No. of
This decline in risk for patients undergoing lung Characteristic Patients %
resection has led to a reappraisal of many of the
Sex
criteria that previously would have rendered a patient
Male 185 56
inoperable. The present study examines the relevance Female 146 44
of many of these traditionally cited risk factors in the Age
modem hospital setting and evaluates the use of a Mean SD 58 13.0
simple calculation of predicted postoperative pulmo- Range 17-82
Type of Operation
nary function as a predictor of postoperative complica- 1-3 wedge resection(s) 127 38
tions. <!: 4 wedge resections 9 3
Lobectomy 145 44
METHODS Bilobectomies 4 1
Pneumonectomy 46 14
Patients Smoking history
The patient population consists of 331 patients who underwent Past or present 243 73
single or multiple pulmonary wedge resections, lobectomies, or Still smoking 67 20
pneumonectomies. Subjects who underwent exploratory or staging Never smoked 88 27
thoracotomy procedures without resection of lung parenchyma Grade of dyspnea
were excluded from the study. This study population represents 86 preoperatively (n = 274)
percent (331) of 385 patients undergoing the above procedures Dyspnea with severe exertion only 153 56
during 24 months from August 1989, to August 1991, at the Dyspnea on hills or stairs 79 29
Brigham and Women's Hospital, Boston. The 54 patients (14 Can walk indefinitely on level ground 29 11
percent) meeting the above criteria not included in the analysis Dyspnea at 100 yds 10 4
were omitted because they were inadvertently not entered pro- Dyspnea with minimal activity 3 1
spectively in the clinical database. This occurred on a random basis Kamofsky score (n = 279)
without apparent selection bias. MeanSD 9.1 1.1
After surgery, all patients were cared for in a specialized Range 4-10
intermediate care unit by the thoracic surgery service. The purpose Pathology final diagnosis
of this unit is to provide specialized nursing care, respiratory Benign 65 20
therapy, pain control, and medical care to patients immediately Malignant 266 80
after thoracic surgery. Emphasis is placed on aggressive pulmonary
care, early ambulation, and pain control to minimize postoperative
pulmonary complications.
number of bronchopulmonary segments removed was recorded for
Data Collection each patient. Assuming that each of the 19 bronchopulmonary
All clinical data were recorded prospectively as part of an segments in the normal lung contributes equally to ventilatory
ongoing clinical database. Preoperative assessment included the function, each segment accounts for 5.26 percent of total lung
documentation of historic information (symptoms, coexistent medical function. The predicted postoperative FEY1 was then estimated by
conditions, and tobacco use) as well as functional status (Kamofsky the following formula: 17 predicted postoperative FEY1 = preoperative
score). Preoperative pulmonary evaluation included ABC analysis FEY1 x [1- (S x 5.26)/100), (S =number of bronchopulmonary
(recorded in 95 percent of patients) and spirometry (recorded in 98 segments removed).
percent of patients). Patients also underwent measurement of A right pneumonectomy was considered to cause a 55 percent
room air oxygen saturation by oximetry at rest and with ambulation decrement in preoperative FEY1 and a left pneumonectomy to
(recorded in 79 percent of patients). The ambulatory measure- cause a 45 percent decrement. In patients undergoing wedge
ments involved a 6 -min walk on a level surface followed by two resections, each wedge resection was assumed to account for one
flights of stairs. The lowest percent saturation during this course bronchopulmonary segment. The right and left lower lobes were
was examined for this study. considered to have five bronchopulmonary segments, the right
All patients were followed up prospectively after surgery and middle lobe to have two bronchopulmonary segments, the right
complications occurring during the patient's hospitalization that upper lobe to have three bronchopulmonary segments, and the left
were identified in physician progress notes were recorded. For this upper lobe to have four bronchopulmonary segments.
study, "complications" were defined to include respiratory compli-
cations (pneumonia, atelectasis, mechanical ventilation for longer Statistics
than 2 days, and respiratory failure), cardiac complications (myo- The relationships between potential predictors and postoperative
cardial infarction, supraventricular arrhythmias, ventricular complications were assessed in univariate and multivariate analyses.
arrhythmias), renal failure, and postoperative death. All statistical tests were two-tailed, with a significance level of 0.05.
Categoric variables (sex, smoking status, preoperative Pco2 45 mm
Predicted Postoperative FEV1 Hg, complications, and FEY1 > 1 L) were evaluated using a chi-
Calculation of the forced expiratory volume in Ls (FEY1) was square test or Fisher's exact test when the expected cell counts
performed using preoperative pulmonary function testing data and were less than five. Continuous variables (age, FEY 1 value) were
information on the number of bronchopulmonary segments re- evaluated using the Student's t test and a Wilcoxon rank sum test.
moved (which usually can be predicted on the basis of preoperative A multivariate logistic regression analysis was performed to identify
radiologic studies). All operative notes were reviewed, and the the independent correlates of complications.

754 Operative Risk In Patients Undergoing Lung Resection (Kearney eta/)

Table 2 -Complications Occurring in 331 Patienn Patients older than 60 years had a significantly higher
After Lung Resection rate of any complication. Thirty-seven (22 percent) of
167 patients aged 60 or older had complications vs 18
No. of
Complication Type Patients % (11 percent) of 159 patients less than 60 years of age
(p < 0.05). Male patients had a significantly higher
Any complication 56 17
overall complication rate. Complications occurred in
Respiratory
Pneumonia 8 2 40 (22 percent) male patients vs 16 (11 percent)
Atelectasis (~2 bronchoscopies req) 13 4 female patients (p < 0.05).
Atelectasis (>2 bronchoscopies req) 1 0.3 Smoking status also was evaluated in patients un-
Respiratory failure 6 2 dergoing surgery. Two hundred forty-three patients
Mechanical ventilation required for >2 days 4 1
had a history of cigarette use, and 67 patients were still
postoperatively
Cardiac smoking at the time of surgery (Table 3). Eighty-eight
Supraventricular tachycardia 37 11 patients denied any history of tobacco use. The num-
Ventricular arrhythmia 4 1 ber of adverse postoperative events was not signifi-
Arrhythmia req cardoiversion 1 0.3 cantly higher in patients still smoking at the time of
My?Cardial infarction 2 0.6
surgery. Complications occurred in 12 (18 percent) of
Renal failure 4 1
Postoperative death 3 0.9 67 patients still smoking vs 44 (17 percent) of 264
patients not using tobacco at the time of admission (p
value not significant). When patients with any history
RESULTS of tobacco use (current or at any point in the past)
were examined, a significantly higher rate of postop-
Study Population erative complications was found. Forty-eight (20 per-
The clinical characteristics of the 331 patients (mean cent) of 243 patients with any history of tobacco use
age 58; 56 percent men) who underwent wedge
Table 3 -Clinical Correlates of Poltoperatiofl
resections, lobectomies, or pneumonectomies are sum- Complications
marized in Table 1. The most common operation, a
lobectomy, was performed in 145 cases, while 46 No. No. (%)With Any
patients underwent pneumonectomy. Two hundred Parameter Patients Complication p Value
forty-three patients had a history of tobacco use. Only Age* <0.05
13 of the patients had dyspnea after walking 100 yards ~ 167 37 (22)
<60 159 18 (11)
or less before surgery, and the mean Karnofsky score
Sex <0.05
was 9.1. Two hundred sixty-six patients had malignant Male 185 40 (22)
neoplasms as their final diagnosis. Female 146 16 (11)
Smoker
Complication Rate Current NSf
Yes 67 12 (18)
Complications occurring postoperatively are dis- No 264 44 (17)
played in Table 2. Eight of 331 patients developed Ever smoked
pneumonia, and 14 patients developed atelectasis, one Yes 243 48 (20) <0.05
No 88 8 (9)
of whom required more than two bronchoscopies Preoperative FEV1 NS
postoperatively. Four patients required mechanical :2:1 L 307 51 (17)
ventilation for more than 2 days after surgery, and <lL 17 3 (18)
respiratory failure developed in six patients. Cardiac Preoperative Pco2 NS
:2:45 mm Hg 30 4 (13)
complications included supraventricular arrhythmias <45 mm Hg 285 50 (17)
that occurred in 37 patients and ventricular arrhythmias Exercise %Sa02 NS
that occurred in 4 patients. One patient required :2:90 185 28 (15)
cardioversion, and two patients developed postopera- <90 77 12 (16)
Procedure
tive myocardial infarctions. Renal failure occurred in Pneumonectomy <0.001
four patients, and three patients died during hospital- Yes 46 18 (39)
ization. No 285 38 (13)
Lobectomy NS
Predictors of Postoperative Complications Yes 145 28 (19)
No 186 28 (15)
Variables often considered predictors of postopera- 1-3 Wedge resections <0.001
tive complications are displayed in Table 3. These Yes 127 8 (6)
include advanced age, male sex, smoking status, pro- No 204 48 (24)
cedure performed, and preoperative pulmonary vari- *Data missing on 5 patients.
ables (FEV" Pco2 data, and exercise pulse oximetry). t NS =not significant.

CHEST I 105 I 3 I MARCH, 1994 755

 40
...
z (')
0
t=
c 30
0
::;
a.
:::E
0 20
0 II)
%
1-
i 10
0~
0
An y
TYPE OF COMPLICATION
FI<: l! HE I. Postoperati\e complication rate according to predicted postope ratiq FE\' ntlue.
had a postoperative complication vs 8 (9 percent) of 88
patients without any history (p < 0.05).
Patients with a preoperative FEV 1 of less than 1 L
did not have an increased overall complication rate.
Complications occurred in 3 (18 percent) of li
patients with a FEV 1 less than 1 L vs .51 ( 1i percent)
of 30i patients with a preoperative FEV 1 greater
than 1 L (p value not significant). Complication
rates also were evaluated according to preoperative
Pco., data on ABC analysis. No significant differ-
ence in the overall complication rate was noted for
patients with a Pco 2 greater than or equal to 4.5 mm
Hg preoperatively. Four ( 13 percent) of 30 patients
with hypercarbia had complications vs 50 (li per-
cent) of 285 patients without hypercarbia (p value
not significant). Preoperative exercise pulse oximetry
data also was assessed (as evaluated by recording
the lowest percent saturation during a 6-min walk
on a level surface followed by two flights of stairs) .
This was not associated with an increased occur-
rence of complications. Twelve ( 16 percent) of ii
patients with desaturation on exercise pulse oximetry
had complications vs 28 ( 1.5 percent) of 18.5 patients
without desaturation (p value not significant).
Patients were classified according to the extent of
lung resection performed, and differences in com-
plication rates were examined (Table 3). Patients
undergoing a pneumonectomy had a significantly
higher rate of complications . All deaths occurred in
the group undergoing pneumonectomies. Eighteen
(39 percent) of 46 patients undergoing a pneumon-
ectomy had a complication vs 38 (13 percent) of 28.5
patients not having this procedure (p < 0.001) . Pa-
tients undergoinig one to three wedge resections
756
Predicted postoperative FEV1 :
< 1.0 l n=47
0 1.01 .4L n=56
0 > 1.4 L n=221
p < 0.001 p < 0.05
C\1
"'
2 3
Cardiac Respiratory
1
had a significantly lower rate of any complication.
Eight (6 percent) of 12i patients having a wedge
resection had a complication vs 48 (24 percent) of
204 patients having a more extensive procedure
(p < 0.001). Patients who underwent lobectomies or
four or more wedge resections did not have signifi-
cantly different complication rates when compared
with patients not having these procedures.
Complication Rate by Predicted Postoperative
FEV 1 Values
Figure 1 shows the postoperative complication
rate stratified according to predicted postoperative
FEV 1 value. Complication rates were assessed for
other complications , cardiac complications, and res-
piratory complications . Forty-seven patients had a
predicted postoperative FEV 1 less than 1 L, and 16
of these (34 percent) had other complications, 9 ( 19
percent) had a cardiac complication, and i ( 15
percent ) had a respiratory complication. Fifty-six
patients had predicted postoperative FEV 1 of 1.01
to 1.40 L, and 14 (25 percent) of these had other
complications, 12 (21 percent) had a cardiac com-
plication, and .5 (9 percent) had a respiratory com-
plication. Two hundred twenty-one patients had a
predicted postoperative FEV 1 of greater than 1.4 L,
and 24 ( 11 percent) of these had other complica-
tions , 20 (9 percent) had cardiac complications, and
6 (3 percent) had respiratory complications. As the
predicted postoperative FEV 1 declined , a signifi-
cant association \Vas found for an increased occur-
rence of other complications (p < 0.001), cardiac
complications (p < 0.0.5), and respiratory complica-
tions (p < 0.001) .
Operative Risk in Patients Undergoing Lung Resection (Kearney et a/)
Multivariate Analysis
In a multivariate logistic-regression analysis adjust-
ing for age, sex, preoperative Pco 2 greater than or
equal to 45 mm Hg, exercise pulse oximetry saturation
of less than 90 percent, history of tobacco use, history
of coronary artery disease, and history of hyperten-
sion, postoperative FEV1 was a significant indepen-
dent predictor of complications. For each 0.2 L de-
crease in predicted FEV 1, the odds ratio for
complications was 1.46 (95 percent CI 1.2 to 1.8).
Only the predicted postoperative FEV1 subgroup re-
mained associated significantly with an increased rate
of other complications postoperatively when the effect
of the other variables was controlled for in the model.
DISCUSSIO:-;
This study examined the postoperative course of a
large cohort of patients undergoing lung resections
and evaluated the significance of a number of
preoperative variables as indicators of postoperative
risk. The postoperative complication rate was quite
low, and the best indicators of patients at high risk for
the development of complications were a low pre-
dicted postoperative FEV1 or a pneumonectomy as
the procedure performed. The predicted postopera-
tive FEV 1 remained the only significant predictor of
complications when the effect of other traditionally
cited markers of increased risk was controlled in a
multivariate analysis. Alow preoperative FEV 1did not
correlate with complications (although the small num-
ber of patients [17] with an FEV1 less than 1 L limits
the power to detect a significant difference). An
elevated Pco2 also did not correlate with a complicated
postoperative course in this study. These results sug-
gest that lung resection usually is tolerated well even
in patients with significant ventilatory impairment and
that estimation of the predicted postoperative FEV1
by a simple calculation may be useful in identifying
patients at high risk for a complicated course.
The postoperative complication rate (less than 1
percent mortality rate and 17 percent rate of any
medical complication) is somewhat lower than that
reported in previous studies. Large studies in the
1980s have reported mortality rates after lung resec-
tion as low as 2 to 5 percent with somewhat higher
mortality rates in older patients and in series examin-
ing pneumonectomies alone. 15 21 ' 26 2~ The low occur-
rence of adverse outcomes in this series may be due to
several factors . The method of recording complica-
tions by reviewing physician progress notes rather
than by examining objective measures of disease (ie,
reviewing chest x-ray films for evidence of pneumo-
nia) is likely to have contributed to the low rate of
postoperative complications. Treating patients after
surgery in a specialized intermediate care unit, with
emphasis on aggressive , preventive respiratory
modalities (chest physiotherapy, incentive broncho-
spirometry, adequate anesthesia, and early ambulation)
may also have led to a decreased occurrence of
respiratory complications. Furthermore, this study
was performed in a tertiary referral center with sur-
geons experienced in performing large numbers of
lung resections. Also, the patient population operated
on appears to have been relatively healthy. The small
number of patients in this study with preoperative
FEV1 values less than 1 L (17 patients) suggests that
selection of a patient population without severe venti-
latory impairment is likely to have contributed to a low
complication rate, and the relatively intact grades of
preoperative dyspnea and Kamofsky scores support
this hypothesis. Also, patients undergoing exploratory
procedures without resection of lung parenchyma
were excluded from the analysis. This may have
contributed to the low complication rate as well, with
the exclusion of patients with more advanced non-
resectable disease who may be more likely to have a
complicated course.
Previous studies have yielded conflicting results
regarding the use of preoperative pulmonary function
tests as predictors of postoperative complications. The
original work by Gaensler et al in 1955, 18 was based on
the outcome of patients undergoing surgery for tuber-
culosis. He found that patients with a preoperative
FEV1 less than 70 percent predicted or with a maxi-
mum voluntary ventilation ( MW) less than 50 percent
predicted had a 40 percent mortality rate. Subsequent
investigators noted in the 1960s and 1970s that pa-
tients with an MW less than 50 percent or a depressed
FEV 1 ( < 1.2 L or < 2 L and age > 60) had a high
mortality rate when undergoing surgery for lung can-
cer.1119.21 Other investigators found no significant dif-
ference in FEV1 between survivors or nonsurvivors in
patients undergoing pneumonectomy or lobectomy. 2028
Olsen et al 29 identified patients considered at high risk
(FEV1 <2 Lor MW <50 percent predicted) and found
that these patients could undergo resection with ac-
ceptable mortality rates if they did not have pulmo-
nary hypertension or hypoxia on pulmonary occlusion
studies, or had a predicted postoperative FEV1 of
greater than 800 ml by quantitative perfusion scan-
ning.29 Interestingly, they note in this study that the
predicted postoperative FEV 1 value of 800 ml as a
cutoff point was not derived through studies of pa-
tients undergoing lung resection but through observa-
tion of the degree of ventilatory impairment usually
tolerated by patients with severe obstructive lung
disease. A postoperative FEV1value of less than 800 to
1,000 ml has been cited widely as a prohibitive level of
pulmonary function since that time. One study since
has shown that patients with marked impairment of
pulmonary function (preoperative FEV1< 1Lor MW
<35-40 percent predicted) tolerated limited resec-
CHEST I 105 I 3 I MARCH, 1994 757
 tions well. 30 The present study appears to confirm
these findings . Limited resections (single or multiple
wedge resections) were performed in 12 of the 17
patients with a preoperative FEV1ofless than 1 L, and
the remaining 5 patients underwent lobectomies. These
patients with markedly impaired preoperative pulmo-
nary function did not have an increased incidence of
complications.
A predicted postoperative FEV1of less than 1 L was
the best predictor of postoperative complications in
this study (p < 0.001), whereas a preoperative FEV1of
less than 1 L was not predictive of complications.
While this appears to confirm the criteria of Olsen et
al29 that patients with a low predicted postoperative
FEV1 are at an increased risk of complications, the
degree of risk does not appear to be prohibitive. Only
two deaths occurred in this group (4.3 percent of the
47 patients), and two thirds of the patients had an
uncomplicated postoperative course. These results
also suggest that the magnitude of change in pulmo-
nary function may be an important indicator of risk for
postoperative complications. The group with a pre-
dicted postoperative FEV1ofless than 1 L necessarily
includes all patients with a preoperative FEV1 of less
than 1 L, and this low preoperative FEV1 value alone
did not confer increased risk. When patients with a
preoperative FEV1 of greater than 1 L and a postop-
erative FEV1ofless than 1 L, however, are analyzed as
a group including the patients with both preoperative
and postoperative FEV1 values less than 1 L, a pre-
dicted postoperative FEV1 less than 1 L is the most
significant predictor of complications. This suggests
that patients with preoperative FEV1 values greater
than 1 L who have a predicted decline in FEV1 to less
than 1 L after surgery are at higher risk for complica-
tions. The higher rate of complications among patients
undergoing pneumonectomies supports this possibil-
ity that a large decline in ventilatory function rather
than a low preoperative FEV1 value alone may be an
important marker of increased risk.
The use of the technique of estimating predicted
postoperative pulmonary function by simple calcula-
tion is supported by previous studies comparing quan-
titative nuclear medicine scans, simple calculation,
and actual measured postoperative spirometric val-
ues. 152731 These studies have shown simple calculation
to be as accurate as quantitative perfusion scans for
this purpose. Estimation of predicted postoperative
pulmonary function by simple calculation, however, is
based on the assumption that each bronchopulmonary
segment contributes equally to ventilatory function,
and patients with severe atelectasis, hilar disease, or
endobronchial involvement may have significantly
unequal ventilation or perfusion. In cases where this is
suspected, quantitative ventilation/perfusion scanning
may be more useful in predicting postoperative pul-
758
monary function. Previous studies also have shown
that low predicted postoperative FEV 1 values (using
quantitative perfusion lung scanning) correlate with
complications. 1521 Markos et aP 5 examined the out-
come of 55 patients undergoing lung resections, and
they found a significantly increased mortality rate
among patients with predicted FEV1 less than 40
percent predicted. In a larger study, Wahi and col-
leagues21 analyzed 197 patients undergoing pneumon-
ectomies, and they found a significantly elevated
mortality rate among patients whose predicted post-
operative FEV1 was less than 1.65 L. Another study
(involving only 22 patients) has shown no significant
correlation between predicted postoperative pulmo-
nary function values and complications rates.3 1 The
present study is the first to our knowledge to examine
the use of estimating the predicted postoperative
FEV1 by the technique of simple calculation alone.
Our study indicates that with modem surgical tech-
niques and improved postoperative care, the level of
prohibitive pulmonary function as defined by calcula-
tion of predicted postoperative FEV 1 may be signifi-
cantly lower than indicated by previous studies using
this method. Although the rate of complications sig-
nificantly was higher in our study among patients with
a predicted postoperative FEV1 less than 1 L, the
magnitude of risk did not appear to be prohibitive.
The level of prohibitive risk may lie lower than 1 L,
possibly in the 60 to 800 ml range.
Patients with hypercarbia on preoperative ABC
measurement have been classified traditionally as
being at very high risk for postoperative respiratory
complications (as noted in a recent position paper by
the American College of Physicians on pulmonary
function testing). 32 The categorization of these pa-
tients as very high risk candidates for pulmonary
resection is based on anecdotal reports in the litera-
ture and has never been studied formally in a large
group of patients. The 30 patients in this study with a
preoperative Pco2 greater than or equal to 45 mm Hg
did not have an increased occurrence of postoperative
complications, and it appears that this should not be
considered a contraindication to surgery. Previous
investigators also have cited desaturation on pulse
oximetry as indicating a high risk population for lung
resection, 15 but in this study desaturation to less than
90 percent did not correlate with postoperative com-
plications. Patients still smoking until the time of
surgery did not have significantly more complications
compared with those not still smoking. Prior studies of
patients undergoing elective coronary artery bypass
grafting have indicated that patients who choose to
stop smoking at least 2 months before surgery have a
lower incidence of complications compared with pa-
tients who stop smoking closer to their date of sur-
gery.33 Unfortunately, it is not clear whether it is
Operative Risk in Patients Undergoing Lung Resection (Kesmey et sl)
appropriate to extrapolate these data to patients un-
dergoing lung resection, and the time interval from
smoking cessation to surgery is not recorded in our
database. Patients without any smoking history had
significantly fewer overall complications.
These data suggest that lung resection can be
performed in patients with significant ventilatory im-
pairment with a low rate of morbidity and mortality
and that surgery should not be denied on the basis of
preoperative variables such as a low FEV" hypercarbia
on blood gas analysis, or exercise desaturation on pulse
oximetry. Additionally, patients undergoing pneumon-
ectomies or with a low predicted postoperative FEV 1
as estimated by simple calculation appear to be at
increased risk for complications, but they may still
undergo surgery with acceptable mortality rates.
ACKNOWLEDGMENT: The authors thank Mary Sullivan Visciano
for her editorial assistance.
REFERENCES
Centers for Disease Control. Trends in lung cancer incidence
and mortality-United States,1980-1987. MMWR 1990; 39:875-
81
2 Boring CC, Squires TS, Tong T. Cancer statistics, 1992. CA
1992; 42:19-43
3 American Cancer Society. Cancer Manual. 8th ed. Boston:
American Cancer Society. 1990
4 Flehinger BJ. Kimmel M, Melamad MR. The effect of surgical
treatment on survival in early lung cancer: implications for
screening. Chest 1992; 101:1013-18
5 US Department of Health and Human Services. Reducing the
health consequences of smoking: 25 years of progress-a report
of the Surgeon General. Rockville, Md: US Department of
Health and Human Services, Centers for Disease Control,
Center for Chronic Disease Prevention and Health Promotion,
Office on Smoking and Health, Public Health Service, DHHS
publication No. (CDC) 89-8411, 1989:43-169
6 Nomura A, Stemmermann GN. Chyou P, Marcus GB, Buist AS.
Prospective study of pulmonary function and lung cancer. Am
Rev Respir Dis 1991; 144:307-11
7 Large P, Nyboe J. Appleyard M, Jensen G, Schnohr P. Ventila-
tory function and chronic mucus hypersecretion as predictors
of death from lung cancer. Am Rev Respir Dis 1990; 141:613-
16
8 Skillrud DM, Offord KP, Miller RD. Higher risk oflung cancer
in chronic obstructive pulmonary disease. Ann Intern Med
1986; 10.5:503-7
9 Tockman MS. Anthonisen NR. Wright EC, Donithan MG.
Ainvay obstruction and the risk oflung cancer. Ann Intern Med
1987; 106:512-8
10 Zibrak JD, O'Donnell CR. Marton K. Indications for pulmo-
nary function testing. Ann Intern Med 1990; 112:763-71
11 Boushy SF, Billing DM, North LB, Helgason AH. Clinical
course related to preoperative and postoperative pulmonary
function in patients with bronchogenic carcinoma. Chest 1971;
59:383-91
12 Boysen PC, Block AJ, Moulder PV. Relationship between
preoperative pulmonary function tests and complications after
thoracotomy. Surg Gynecol Obstet 1981; 152:813-15
13 Bechard D, Wetstein L. Assessment of exercise oxygen con-
sumption as preoperative criterion for lung resection. Ann
Thorac Surg 1987; 44:344-49
14 Smith TP, Kinasewitz GT, Tucker WY, Spillers WP, George
RB. Exercise capacity as a predictor of post-thoracotomy
morbidity. Am Rev Respir Dis 1984; 129:730-34
15 Markos J, Mullan BP, Hillman DR, Musk AW, Centico VF,
Lovegrove FT. Preoperative assessment as a predictor of
mortality and morbidity after lung resection. Am Rev Respir
Dis 1989; 139:902-10
16 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 compli-
cations: a blinded prospective study of coronary artery bypass
patients. Mayo Clin Proc 1989; 64:609-16
17 Garibaldi RA, Britt MR, Coleman ML, Reading JC, Pace NL.
Risk factors for postoperative pneumonia. Am J Med 1981;
70:677-80
18 Gaensler EA, Cugell DW, Lindgren I, Verstraeten JM, Smith
SS, Strieder JW. The role of pulmonary insufficiency in mortal-
ity and invalidism following surgery for pulmonary tuberculo-
sis. J Thorac Surg 1955; 29:163-85
19 Mittman C. Assessment of operative risk in thoracic surgery.
Am Rev Respir Dis 1961; 84:197-207
20 Didolkar MS, Moore RH, Takita H. Evaluation of the risk in
pulmonary resection for bronchogenic carcinoma. Am J Surg
1974; 127:700-03
21 Wahi R, McMurtrey MJ, DeCaro LF, Mountain CF, Ali MK,
Smith TL. Determinants of perioperative morbidity and mor-
tality after pneumonectomy. Ann Thorac Surg 1989; 48:33-7
22 Ginsberg RJ, Hill LD, Eagan RT, Thomas P. Mountain CF.
Deslauriers J. Modem thirty-day operative mortality for surgi-
cal resections in lung cancer. J Thorac Cardiovasc Surg 1983;
86:654-58
23 Krowka MJ, Pairolero PC, Trastek VF, Payne WS, Bematz PE.
Cardiac dysrhythmia following pneumonectomy: clinical corre-
lates and prognostic significance. Chest 1987; 91 :490-95
24 Kohman LJ, Meyer JA, Ikins PM, Oates RP. Random versus
predictable risks of mortality after thoracotomy for lung cancer.
J Thorac Cardiovasc Surg 1986; 91:551-54
25 Keagy BA, Lores ME, Starek PJK, Murray GF, Lucas CL,
Wilcox BR. Elective pulmonary lobectomy: factors associated
with morbidity and operative mortality. Ann Thorac Surg 1985;
40:349-52
26 Nagasaki F, Flehinger BJ, Martini N. Complications of surgery
in the treatment of carcinoma of the lung. Chest 1982; 82:25-
9
27 Juhl B, Frost N. Acomparison between measured and calcu-
lated changes in the lung function after operation for pulmo-
nary cancer. Acta Anaesth Scand 1975; 57 (suppl):39-45
28 Keagy BA, Schorlemmer GR, Murray GF, Starek PJK, Wilcox
BR. Correlation of preoperative pulmonary function testing
with clinical course in patients after pneumonectomy. Ann
Thorac Surg 1983; 36:253-57
29 Olsen GN, Block AJ, Swenson EW, Castle JR, Wynne JW.
Pulmonary function evaluation of the lung resection candidate:
a prospective study. Am Rev Respir Dis 1975; 111:379-87
30 Miller JI, Hatcher CR. Limited resection of bronchogenic
carcinoma in the patient with marked impairment of pulmo-
nary function . Ann Thorac Surg 1987; 44:340-43
31 Wemly JA, DeMeester TR, Kirchner PT. Myerowitz PD,
Oxford DE, Golomb HM. Clinical value of quantitative venti-
lation-perfusion scans in the surgical management of
bronchogenic carcinoma. J Thorac Cardiovasc Surg 1980;
80:535-43
32 American College of Physicians. Preoperative pulmonary func-
tion testing. Ann Intern Med 1990; 112:793-94
33 Warner MA, Divertie MB, Tinker JH. Preoperative cessation of
smoking and pulmonary complications in coronary artery by-
pass patients. Anesthesiology 1984; 60:380-83
CHEST I 105 I 3 I MARCH, 1994 759