A carcinoma of the lung reported by Graham and. Predicted Pulmonary Function and Survival After Pneumonectomy for Primary Lung Carcinoma

Transcription

1 Predicted Pulmonary Function and Survival After Pneumonectomy for Primary Lung Carcinoma Joe B. Putnam, Jr, MD, David E. Lammermeier, MD, Rolando Colon, MD, Marion J. McMurtrey, MD, Mohammed K. Ali, MD, PhD, and Jack A. Roth, MD Department of Thoracic Surgery, The University of Texas MD Anderson Cancer Center, Houston, Texas Between 1982 and 1987, 139 patients with primary carcinoma of the lung were treated with pneumonectomy. Thirty-nine patients (28%) were in clinical stage I, 10 (7%) were in clinical stage 11, and 90 (65%) were in clinical stage 111. Overall actuarial 3-year survival was 33%. Actuarial 3-year survival for patients in clinical stage I was %; for those in clinical stage 11, 8%; and for those in clinical stage 111, 28%. Risk factors for operative mortality examined included preoperative forced vital capacity (FVC) of 2.13 L or less and forced expiratory volume in 1 second (FEV,) of 1.65 L or less, percent predicted FVC of 6% or less and FEV, of 65% or less, predicted postoperative FVC of 1.31 L or less and FEV, of 0.89 L or less, and predicted postoperative percent predicted FVC of 1% or less and FEV, of 3% or less. Operative deaths occurred only in clinical stage 111 patients (7/90 or 8%). Patients with compromised pulmo- nary function based on one or more of the examined risk factors were at increased risk for death (2/10) compared with patients with better pulmonary function (5/80 or 6.25%). Actuarial 3-year survival for high-risk clinical stage 111 patients ranged from 0% to 16% compared with 28% for other clinical stage 111 patients. Thirty-day mortality for pathological stage 111 patients was 6.3% (51791, and 3-year actuarial survival was 2%. No patient in pathological stage 111 who was at high risk survived beyond 3.1 years. Select individuals with adequate pulmonary function and stage 111 disease can achieve substantial long-term survival after pneumonectomy. Patients should not be excluded from pneumonectomy based on stage alone or on the results of any single pulmonary function study. (Ann Thorac Surg 2990;9:909-15) fter the successful one-stage pneumonectomy for A carcinoma of the lung reported by Graham and Singer [l] in 1933, pneumonectomy became the standard treatment of primary carcinoma of the lung for the next two decades. In 1950, Churchill and associates [2] presented improved results with lobectomy, which remains the procedure of choice in the majority of patients with carcinoma of the lung. Surgical resection including pneumonectomy continues as the key to potentially curative therapy for patients with primary lung carcinoma. Selection of patients who will optimally benefit from resection can be difficult. In patients with compromised cardiopulmonary function, pneumonectomy is associated with substantial risk and mortality [3-51. In past studies, operative mortality after pneumonectomy ranged from 6.2% to 12.% [3, 6-81 and 5-year survival, from 8% to 25% [3, 7, 91 for all stages of bronchogenic carcinoma. Historically, patients with stage I11 bronchogenic carcinoma treated by pneumonectomy achieve a 5-year survival rate of 10% to 13.8% [9, 101. Recently, our group [ll] identified specific criteria associated with perioperative morbidity and mortality after pneumonectomy for several types of cancer. In this current study, stage-specific survival was analyzed for all Presented at the Thirty-sixth Annual Meeting of the Southern Thoracic Surgical Association, Scottsdale, AZ, Nov 9-11, Address reprint requests to Dr Putnam, Department of Thoracic Surgery, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Box 109, Houston, TX patients with primary bronchogenic carcinoma who underwent pneumonectomy for cure to determine if patients with clinical or pathological stage 111 primary lung cancer benefit from surgical resection and whether preoperative pulmonary function studies can predict postoperative mortality and long-term survival. Material and Methods Between 1982 and 1987, 139 patients underwent pneumonectomy for primary carcinoma of the lung at The University of Texas MD Anderson Cancer Center. There were 99 men with an average age of 56.9 years (range, 38 to 80 years) and 0 women with an average age of 58.0 years (range, 39 to 72 years). All patients were evaluated by physicians from the Departments of Thoracic Surgery and Thoracic Medicine after referral by the patient s primary physician. Clinical staging by the TNM classification [12] was performed by extensive preoperative evaluation and included plain chest roentgenograms and computed tomographic scans of the chest and abdomen. Mediastinal lymph nodes greater than 1 cm in any diameter were categorized as enlarged and defined as clinical N2 disease. Bone scans and computed tomographic scans of the brain were done if symptoms were present. The patient underwent pulmonary function tests including regional pulmonary ventilation and perfusion using xenon 133 gas. All pulmonary function tests were performed by the same laboratory for internal consistency. Details of the xenon by The Society of Thoracic Surgeons /90/$3.50

2 910 PUTNAM ET AL Ann Thorac Surg 133 pulmonary function test have been previously described [ Operative Technique The staff surgeon was the final arbiter of operative risk and potential resectability before surgical intervention. Central venous pressure, radial artery pressure, pulse oximetry, and urine output were monitored continuously throughout the operation. The operative technique consisted of a serratus anterior muscle-sparing posterolateral thoracotomy, and the extent of resection was determined by tumor size and local extension. Pneumonectomy was performed in all patients whose tumor was considered potentially completely resectable. Mediastinoscopy was not routinely performed. The necessity for pneumonectomy was determined at the time of operation by the surgeon. All patients underwent a mediastinal lymph node dissection. Removal of all identifiable nodes along the pulmonary ligament and in the subcarinal regions, tracheobronchial angle, and paratracheal regions was performed. Hilar and interlobar nodes were sampled during anatomical dissection [ 121, and additional lymph nodes were removed after removal of the lung. Survival Analysis and Risk Factors Actuarial survival analysis was performed by clinical stage, pathological stage, and risk factors. Risk factors examined for operative mortality after pneumonectomy included preoperative forced vital capacity (FVC) of 2.13 L or less, preoperative forced expiratory volume in 1 second (FEV,) of 1.65 L or less, percent predicted FVC of 6% or less, and percent predicted FEV, of 65% or less. Using the percent ventilation to the noninvolved lung from the xenon 133 ventilation/perfusion study, predicted postoperative pulmonary function was determined. Postoperative risk factors examined included predicted postoperative FVC of 1.31 L or less, predicted postoperative FEV, of 0.89 L or less, predicted postoperative percent predicted FVC of 1% or less, and predicted postoperative percent predicted FEV, of 3% or less. No patient was excluded from operation based on any individual test result. Operative death was defined as death within 30 days of pneumonectomy or death during hospitalization after pneumonectomy at any time. Any patient in one or the other category was counted as an operative death. Follow-up All patients were followed until death. Median follow-up is.1 years for surviving patients. Statistical Analysis Patient survival distributions were calculated by the method of Kaplan and Meier [16]. Differences in survival were calculated using the log-rank test. Differences in operative mortality based on pulmonary function were calculated using the x2 tests. All p values are two tailed. Results Sixty-five patients (7%) had squamous cell carcinoma; 50 patients (36%), adenocarcinoma; 9 patients (6.5%), ade- nosquamous carcinoma; 5 patients (%), large cell carcinoma; and 2 patients (1 %), bronchoalveolar carcinoma. Eight patients (6%) had unclassified primary carcinoma of the lung. Seventy-two patients underwent a right pneumonectomy (52%) and 67 (8%), a left pneumonectomy. Mortality and Morbidity The 30-day mortality after pneumonectomy was 5% (7 patients). All seven deaths occurred among patients who had a right pneumonectomy (operative mortality, 9.7% [7/72]). There were no operative deaths among patients in clinical stage I or 11. Operative mortality for clinical stage I11 patients was 7.8% (7190) (Table 1) and was precipitated by an acute cardiac-related event in patients (57%) with death eventually resulting from pneumonia. Two additional patients (29%) died of pneumonia and 1 patient (1%), of a bronchopleural fistula. Surgical complications included arrhythmias (21 patients, 15.1%), cardiac failure (1 patients, 10.1%), pneumonia (9 patients, 6.5%), hemorrhage (6 patients,.3%), pericarditis (5 patients, 3.6%), minor atelectasis (5 patients, 3.6%), and wound infection (2 patients, 1.%). Survival and Clinical Stage Actuarial survival was determined for all patients (Fig 1) and by clinical stage (Fig 2). Overall survival was 23% at 5 years. Thirty-nine patients (28%) were in clinical stage I, 10 (7%) were in clinical stage 11, and 90 (65%) were in clinical stage 111. Patients with clinical stage I disease had a % 3-year survival (median survival, 2.66 years). Patients with clinical stage I1 disease had a 8% 3-year survival (median survival, 1.97 years). Patients with clinical stage I11 disease had a 28% 3-year survival (median survival, 1.27 years). Survival and Pathological Stage After operation, patients were restaged based on pathological examination of the tumor and lymph node mapping. There were 23 patients (16.5%) in pathological stage I, 36 patients (25.9%) in stage 11, 79 patients (56.8%) in stage 111, and 1 patient (0.7%) in stage IV. Survival was examined for each pathological stage (Fig 3). Patients with pathological stage I disease had a.3% operative mortality (1/23) and a 3-year survival of 5% (median survival, 1.88 years). Patients with pathological stage I1 disease had an operative mortality of 2.8% (1136) and a 3-year survival of 7% (median survival, 2.87 years). Patients with pathological stage 111 disease had an operative mortality of 6.3% (5179) and a 3-year survival of 2% (median survival, 1.17 years). Survival and Perioperative Pulmonary Function We examined the value of pulmonary function tests as a predictor of operative mortality and survival. Risk factors examined for operative mortality and survival after pneumonectomy included preoperative FVC of 2.13 L or less, preoperative FEV, of 1.65 L or less, percent predicted FVC of 6% or less, and percent predicted FEV, of 65% or less. In addition, predicted postoperative FVC of 1.31 L or less, predicted postoperative FEV, of 0.89 L or less, predicted postoperative percent predicted FVC of 1% or less, and

3 Ann Thorac Surg PUTNAh ET AL 911 h Z * N m w s N Y 5 H B w ; predicted postoperative percent predicted FEV, of 3% or less were examined. Xenon 133 ventilatiodperfusion scans were used to predict postoperative pulmonary function in the noninvolved lung. Actuarial patient survival was examined as a function of perioperative pulmonary function and clinical and pathological stage. Survival by Clinical Stage and Perioperative Pulmonary Function There were no operative deaths among patients in clinical stage I or 11. The operative mortality among clinical stage 111 patients was 7.8% (7/90), and death occurred in 20% (YlO) of patients at increased risk (one or more factors as previously described) compared with 6.25% (5/80) of other patients (see Table 1). Median survival for clinical stage 111 patients with a preoperative FVC of 2.13 L or less was 0.1 year (n = 6) compared with 1.32 years for those with a preoperative FVC of more than 2.13 L (n = 8) (p = 0.002) (Fig A). Median survival for clinical stage I11 patients with a preoperative FEV, of 1.65 L or less was 0.59 year compared with 1. years for other clinical stage 111 patients (p = 0.003) (Fig B). No patient in clinical stage 111 with a percent predicted FVC of 6% or less survived beyond 1.3 years compared with a 30% survival at 3 years for all other clinical stage 111 patients (p = ) (Fig 5). No clinical stage I11 patient with a predicted postoperative FVC of 1.31 L or less survived beyond 3.1 years compared with a 28% 3-year survival for other patients in clinical stage I11 (p = 0.07) (Fig 6). Finally, patients in clinical stage 111 with a predicted postoperative FEV, of 0.89 L or less had a 3-year survival of 10% (1 patient at risk) compared with a 28% 3-year survival for other clinical stage 111 patients ( p = not significant). There were no significant differences in survival after pneumonectomy between groups for the other risk factors examined in clinical stage 111 patients. Survival by Pathological Stage and Perioperative Pulmona y Function Pathological staging was accomplished at operation by lymph node mapping using established criteria [12]. Good-risk patients in pathological stage I11 had a 1.25% 30-day mortality and a 25% 5-year survival. Median survival by pathological stage was similar to that by clinical stage for groups with comparable pulmonary function. Median survival for pathological stage I11 patients with a preoperative FVC of 2.13 L or less was 0.38 year compared with 1.25 years for those with a preoperative FVC of more than 2.13 L (p = not significant). Median survival for pathological stage 111 patients with a preoperative FEV, of 1.65 L or less was 0.70 year compared with 1.27 years for other pathological stage I11 patients (p = not significant). There were no survivors beyond 3.1 years for pathological stage 111 patients with an FVC of 2.13 L or less or an FEV, of 1.65 L or less. The percent predicted FVC of 6% or less (preoperative) was associated with decreased survival after pneumonectomy. No patient in pathological stage I11 with a percent predicted FVC of 6% or less survived beyond 1.3 years

4 912 PUTNAM ET AL Ann Thorac Surg 1.o I FOR ALL STAGES PSI A PSI PS g 0.6 -I z I- = 0 $ 0. = n 0.3 J I- = K n Fig 1. Overall actuarial survival of patients undergoing pneumonectomy for primary carcinoma of the lung. compared with a 26% survival at 3 years for the other pathological stage I11 patients (p = 0.009). No patient in pathological stage I11 with a predicted postoperative FVC of 1.31 L or less or a predicted postoperative FEV, of 0.89 L or less survived beyond 3.1 years (p = not significant). 1.o g I z I- LL: = n 0.- Oa2I CSI A 10 8 CSII CS L residual function in the noninvolved lung, and pneu- 0.0 Fig 2. Actuarial survival of patients with clinical stage (CS) I, 11, or 111 prima y carcinoma of the lung. Fig 3. Actuarial survival of patients with pathological stage (PS) I, 11, or Ill primary carcinoma of the lung (based on histological examination of the resected specimen and mediastinal lymph node dissection). One patient (not shown) was found to have pathological stage IV disease and survived for 259 days after pneumonectomy. Comment Pneumonectomy for the treatment of lung cancer carries the highest mortality rate among pulmonary resections. The recognized value of this operation as a curative procedure for select patients with bronchogenic carcinoma has led to the continued development of better methods to evaluate these patients preoperatively to identify those patients who would benefit from pneumonectomy and those patients who would be at increased risk from operation. Appropriate screening of patients before pneumonectomy included a thorough evaluation for metastatic disease, pulmonary function assessment by spirometry, and cardiac evaluation for patients with potential cardiac dysfunction. If pulmonary function by spirometry was marginal, xenon 133 ventilation/perfusion scans were used to predict postoperative lung function in the noninvolved lung after pneumonectomy. Patients with compromised ventilation to the involved lung might have sufficient monectomy could be performed with hope of success. There has been considerable debate over the value of pulmonary resection for clinical stage I11 bronchogenic carcinoma [ Often the presence of enlarged mediastinal nodes on computed tomographic scans of the chest has prompted a diagnosis of unresectable cancer and eliminated consideration of surgical intervention without further evaluation. Clinical staging is often unreliable, and therapeutic decisions should not be based on clinical

5 Ann Thorac Surg 199O;9: PUTNAM ET AL FVC <=2.13 A 8 61 FVC > FVC PERC <=6% A FVC PERC > 6% A 1.o* FEVl <= 1.65 A FEVl > 1.65 Fig 5. Actuarial survival of patients with clinical stage 111 lung cancer based on percent predicted forced vital capacity (FVC PERC) of the noninvolved lung (p = 0,0001). The percentage of FVC was based on standard values. ative mortality was low, and some patients achieved long-term survival. A recent study from our center [ll] identified a group of factors influencing perioperative morbidity and mortality POST-OP NC <= 1.31 na A POST-OP FVC > B 0.0 O*l * Fig. (A) Actuarial survival of patients with clinical stage 111 lung cancer based on preoperative forced vital capacity (FVC) in liters (p = 0.023). (B) Actuarial survival of patients with clinical stage Ill lung cancer based on preoperative forced expiratory volume in 1 second in liters (FEV,) (p = 0.03) E Oe3I I stage alone. We have shown that patients with primary bronchogenic carcinoma, clinical stage I11 and pathological stage I11 disease, and good pulmonary function will have an actuarial 3-year survival greater than 25%. Oper Fig 6. Actuarial survival of patients with clinical stage III lung cancer based on predicted postoperative forced vital capacity (POST-OP FVC) of the noninvolved lung (p = 0.07).

6 91 PUTNAM ET AL Ann Thorac Surg in patients undergoing pneumonectomy. Factors associated with higher risk included extent of resection, right pneumonectomy, and impaired pulmonary function. The risks of pneumonectomy in clinical stage 111 patients can be assessed before operation by various pulmonary function studies. Simple spirometry combined with xenon ventilation/perfusion scans can be an effective tool to assist the surgeon in assessing operative risk. Clinical stage I11 patients with a preoperative FVC of greater than 2.13 L, an FEV, of greater than 1.65 L, a percent predicted FVC of more than 6%, a predicted postoperative FVC of greater than 1.31 L, or a predicted postoperative FEV, of more than 0.89 L have less risk associated with pneumonectomy. Associated operative mortality is lessened and long-term survival enhanced in patients with these findings. Patients in pathological stage 111 with a preoperative FVC of 2.13 L or less, a preoperative FEV, of 1.65 L or less, a percent predicted FVC of 6% or less, a predicted postoperative FVC of 1.31 L or less, or a predicted postoperative FEV, of 0.89 L or less did not survive beyond 3.1 years. Patients with clinical stage I11 disease will benefit from resection if a complete resection can be accomplished. Patients who are in clinical stage 111 and are considered a good risk based on preoperative and postoperative predicted pulmonary function have a significant improvement in overall survival compared with patients at high risk based on these criteria. No deaths occurred among patients in clinical stage I or 11. Seven deaths occurred in patients who were in clinical stage 111. Those patients considered at increased risk had an operative mortality of 20% compared with 6% for other patients. Clinical stage 111 patients with an FVC greater than 2.13 L and an FEV, greater than 1.65 L had an overall operative mortality (5/80, 6.25%) similar to that reported for patients having pneumonectomy by the Lung Cancer Study Group [3] (/569, 7.7%). If we consider the differences in operative mortality and long-term survival between these two groups of patients within the same clinical stage, it is evident that a group of patients can be identified whose overall survival can be improved by surgical intervention. Exclusion of patients considered to have clinical stage I11 lung cancer should be based on associated risk factors including diminished pulmonary function, extent of resection, and overall cardiac status. Although we cannot recommend stringent criteria for operability, these guidelines may assist the surgeon in his or her evaluation of the patient's suitability for pneumonectomy. In conclusion, patients with clinical stage I11 disease undergoing pneumonectomy have an actuarial 3-year survival of 28%. Those patients with clinical stage 111 disease and good pulmonary function had better survival (23% at 5 years) than clinical stage I11 patients with poor pulmonary function (no survivors beyond 3.1 years). No single variable enables the clinician to predict outcome, particularly death, after pneumonectomy, and patients should not be excluded from pneumonectomy on the basis of any single criterion. Rather, the use of preoperative studies including predicted postoperative pulmonary function by xenon 133 might assist the surgeon in better selecting patients who will potentially benefit from pneumonectomy. References 1. Graham EA, Singer JJ. Successful removal of an entire lung for carcinoma of the bronchus. JAMA 1933;101: Churchill ED, Sweet RH, Soutter L, et al. The surgical management of carcinoma of the lung. J Thorac Surg 1950; 20: Ginsberg RJ, Hill LD, Eagan RT, et al. Modern thirty-day operative mortality for surgical resections in lung cancer. J Thorac Cardiovasc Surg 1983;86:658.. Krowka MJ, Pairolero PC, Trastek VF, Payne WS, Bernatz PE. Cardiac dysrhythmia following pneumonectomy: clinical correlates and prognostic significance. Chest 1987;91: Sherman S, Guidot CE. The feasibility of thoracotomy for lung cancer in the elderly. JAMA 1987;259: 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: 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;36:25%7. 8. Weiss W. Operative mortality and five-year survival rates in patients with bronchogenic carcinoma. Chest 197;66: Page A, Nakhle G, Mercier C, et al. Surgical treatment of bronchogenic carcinoma: the importance of staging in evaluating late survival. Cancer J Surg 1987;30: McGovern EM, Trastek VF, Pairolero PC, Payne WS. Completion pneumonectomy: indications, complications, and results. Ann Thorac Surg 1988;6: Wahi R, McMurtrey MJ, DeCaro LF, et al. Determinants of perioperative morbidity and mortality after pneumonectomy. Ann Thorac Surg 1989;8: Mountain CF. A new international staging system for lung cancer. Chest 1986;89(Suppl):225S33S. 13. Ali MK, Ewer MS, Atallah MR, et al. Regional and overall pulmonary function changes in lung cancer: correlations with tumor stage, extent of pulmonary resection, and patient survival. J Thorac Cardiovasc Surg 1983;86: Ali MK, Mountain CF, Ewer MS, Johnston D, Haynie TP. Predicting loss of pulmonary function after pulmonary resection for bronchogenic carcinoma. Chest 1980;77: Bria WF, Kanarek DJ, Kazemi H. Prediction of postoperative pulmonary function following thoracic operations: value of ventilation-perfusion scanning. J Thorac Cardiovasc Surg 1983;86: Kaplan E, Meier P. Non-parametric estimation from incomplete observations. J Am Stat Assoc 1958;53: Mountain CF. The biological operability of stage 111 nonsmall cell lung cancer. Ann Thorac Surg 1985;0: Pearson FG, Delarue NC, Ilves R, Todd TRJ, Cooper JD. Significance of positive superior mediastinal nodes identified at mediastinoscopy in patients with resectable cancer of the lung. J Thorac Cardiovasc Surg 1982;83: Martini N, Flehinger BJ, Zaman MB, Beattie EJ Jr. Results of resection in non-oat cell carcinoma of the lung with mediastinal lymph node metastases. Ann Surg 1983;198:

7 Ann Thorac Surg I'UTNAM ET AL 915 FNEUMONECTOMY IN LUNG CARCINOMA DISCUSSION DR JOHN P. CLARKE (Virginia Beach, VA): My associates and I have been using the split-function lung predicted values for a good while and trying to correlate them. I enjoyed your paper; I think it was well done, and I think it was valuable. However, one self-evident factor is that if patients smoke right up to the time that these predicted values are obtained, you really can expect those patients to have a postoperative course that is quite a bit better than the patients with chronic obstructive pulmonary disease who have not smoked for 5 years and have the same values. For instance, a person who has a predicted postoperative FEV, of 0.8 L and is a heavy smoker right up to the time of operation is more likely to have long-term survival and tolerance if he or she stops smoking than the chronically ill patient who has stopped smoking and has the same FEV,. MR MARK V. BRAIMBRIDGE (London, United Kingdom): 1 was very interested in the last statement in your conclusion that the predicted values might help the surgeon. When I was working at the Brompton Hospital, my colleagues and I used the stair test in this respect. If the patient reached the first floor, he could survive a segmental resection; if he reached the second floor, a lobectomy; and if he reached the third floor, he could tolerate a pneumonectomy. At St. Thomas' Hospital, this seemed a little crude, and we did a lot of the tests that you recommend, particularly to arrive at the predicted FVC. After trying to calculate this many times, because it obviously is quite different when the lung to be removed is collapsed compared with when it is fully aerated, we found that all our predictive tests were no better than the old-fashioned Brompton three-floor test. How do you arrive at the predicted FVC? As I see it, that clearly is the critical criterion. DR LEWIS WETSTEIN (Manalapan, NJ): I congratulate you, Dr Putnam, on your results. My associates and I have also been extremely interested in defining preoperative variables that would decrease operative morbidity and mortality among patients having lung resection. In fact, 3 years ago at this meeting, we looked at exercise oxygen consumption [I]. Dr Bechard and I found that despite adequate preoperative static pulmonary function studies in patients who would equate to your non-high-risk group, there remains a population of patients with major operative morbidity and mortality. We demonstrated that patients who could not generate oxygen consumption of more than 10 ml/kg/min have increased morbidity and mortality, again despite normal static pulmonary function criteria. Therefore, as a follow-up to Mr Braimbridge, my question is, did you look at an exercise modality? I hypothesize that this would explain why your non-high-risk patients did not do well. DR WILLIAM L. COX, JR (Las Vegas, NV): Over the last few years in my practice, I have observed that hiatal hernias and reflux have developed in patients with left pneumonectomy. This has been demonstrated in almost 80% of the patients by the cineesophagogram and water test. Some of these symptoms postoperatively I have found to be due to the reflux, which I have recently been repairing when I do a left pneumonectomy. I would certainly appreciate knowing if you had any of these problems in your patients who had a left pneumonectomy. DR PUTNAM: Dr Clarke, our patients are instructed to stop smoking entirely for 2 weeks before operation to optimize their pulmonary function and minimize their convalescence. Ninetyeight percent of patients comply with this request when they are counseled as to the value of not smoking before operation. If patients who smoke have marginal pulmonary function while smoking, then the pulmonary function tests are reevaluated after these 2 weeks after cessation of smoking. When nonsmoking is coupled with a program of progressive walking and incentive spirometry, operative morbidity is dramatically reduced. Mr Braimbridge, I agree that walking a patient up three flights of stairs does evaluate adequate pulmonary function before pneumonectomy. Although it is inexpensive, is easy to perform, and provides valuable information, it is somewhat subjective and difficult to quantitate. The pulmonary function studies performed on our patients attempted to quantitate preoperative and predicted postoperative pulmonary function. The best single test for predicting operative morbidity and mortality has not been determined and may be a combination of studies. The predicted FVC was obtained by a xenon 133 ventilation/ perfusion scan. The predicted FVC was calculated by multiplying the percentage of ventilation in the noninvolved lung times the preoperative FVC. This is the "predicted' postoperative FVC. Dr Wetstein, exercise oxygen consumption is an accurate measure of physiological cardiopulmonary function and can appropriately define operative risk with high-risk patients having less than 10 ml/kg/min oxygen consumption. This test was not available for the entire study period. Currently we have a prospective study ongoing to evaluate the use of oxygen consumption as an indicator of risk in the thoracic surgical patient. Dr Cox, symptomatic gastroesophageal reflux with aspiration has not been a problem in our patients after pneumonectomy. In patients with a history of reflux, esophageal manometric and ph testing should be performed before operation and the reflux problem should be corrected. If the left chest is to be explored for pulmonary resection, a floppy Nissen fundoplication can be performed readily with Collis gastroplasty if required. Reference 1. Bechard D, Wetstein L. Assessment of exercise oxygen consumption as preoperative criterion for lung resection. Ann Thorac Surg 1987;3-9.