Preoperative Chemotherapy Does Not Increase Complications After Nonsmall Cell Lung Cancer Resection Emilie Perrot, MD, Benoit Guibert, MD, Pierre Mulsant, MD, Sonia Blandin, MD, Isabelle Arnaud, MD, Pascal Roy, MD, Laurence Geriniere, MD, and Pierre-Jean Souquet, MD Department of Thoracic Oncology and Chest Disease, Thoracic Surgery Unit, and Department of Biostatistics, Hospices Civils de Lyon, Centre Hospitalier Lyon Sud, Pierre-Benite, France Background. Neoadjuvant chemotherapy before resection of nonsmall cell lung cancer seems to increase survival, mainly in the early stage. Risks of postoperative complications after chemotherapy and surgery remain controversial. Here we review our experience with patients treated in one thoracic surgery center. Methods. undergoing resection for nonsmall cell lung cancer after induction chemotherapy between January 1993 and March 2002 were reviewed. Data collected included age, sex, preoperative forced expiratory volume in 1 second (FEV 1 ), hemoglobin, and arterial oxygen pressure tension (PaO 2 ), postoperative complications, and global survival.the main objectives were postoperative mortality and morbidity. Postoperative mortality and morbidity were defined as complications or deaths occurring within 30 days after surgery. Predictive morbidity factors were identified by univariate and multivariate analysis and overall survival by the Kaplan- Meier method. Results. In all, 114 patients were reviewed. Different induction chemotherapies were used, mainly cisplatin with vinorelbine or gemicitabine. Postoperative mortality was 2 of 114, 1 of 27 after pneumonectomy, and there were no deaths after lobectomy. Complications occurred in 29% of patients (33 of 114), usually infectious pneumonia and anemia requiring transfusion. Preoperative FEV 1, hemoglobin, and PaO 2 are not associated with morbidity in univariate or multivariate analysis. Conclusions. Preoperative chemotherapy does not increase postoperative mortality and morbidity after nonsmall cell lung cancer surgery, performed exclusively by thoracic surgeons. (Ann Thorac Surg 2005;80:423 7) 2005 by The Society of Thoracic Surgeons Preoperative chemotherapy seems to increase survival rates in stage IIIA nonsmall cell lung cancer (NSCLC), but there have been few randomized studies and these have had a small number of patients [1, 2]. In the randomized study by Depierre and colleagues [3], an increased survival with neoadjuvant chemotherapy was observed in the early stage patients (IB and II), mainly through reduced metastatic relapse. Preoperative chemotherapy has become the standard treatment for N2 patients. The influence of induction chemotherapy on postoperative outcomes remains controversial. Depierre and colleagues [3] and a few retrospective studies [4, 5] showed increased postoperative complications after neoadjuvant chemotherapy. Here we sought to determine the incidence of surgical morbidity and mortality after preoperative chemotherapy or chemoradiation in our thoracic surgery unit. The main objective was to analyze postoperative mortality and morbidity (complications occurring in the 30 days after surgery). A secondary Accepted for publication Feb 9, 2005. Address reprint requests to Dr Perrot, Department of Thoracic Oncology and Chest Disease, Hospices Civils de Lyon, Pneumologie 1A, Secteur JC, Chemin du Grand Revoyet, Centre Hospitalier Lyon Sud, Pierre Benite, 69495 Cedex France; e-mail: emilie.perrot@chu-lyon.fr. objective was to identify predictive factors of morbidity by univaried and multivaried analysis and overall survival by the Kaplan-Meier method. and Methods Subjects and Data Collected We made a retrospective study of all patients who received preoperative chemotherapy or chemoradiation before thoracotomy for resectable NSCLC from January 1993 to March 2002. All patients underwent thoracotomy exclusively by two thoracic surgeons in the thoracic surgery unit of the Lyon Sud Hospital Center. The majority of patients did not have mediastinoscopy, and the preoperative stage was determined mainly through a thoracic computer scan. Preoperative data collected included age, sex, stage, forced expiratory volume in 1 second (FEV 1 ), hemoglobin, and arterial oxygen pressure tension (PaO 2 ), form and number of chemotherapy sessions, and whether the patient had preoperative radiotherapy. Details of the surgical procedure were recorded. Postoperative data collected included the date and form of resection, the final pathologic stage, classified according to the 1997 2005 by The Society of Thoracic Surgeons 0003-4975/05/$30.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2005.02.043
424 PERROT ET AL Ann Thorac Surg CHEMOTHERAPY BEFORE NSCLC RESECTION 2005;80:423 7 Table 1. Preoperative and Postoperative Stages Stage ctnm ptnm IA 0 10 IB 9 19 Total 9/114 (8%) 29/114 (25%) IIA 0 3 IIB 14 24 Total 14/114 (12%) 27/114 (23%) IIIA 54 28 IIIB 37 23 Total 91/114 (79%) 51/114 (44%) IV 0 7 Total 0 (0%) 7/114 (6%) International System for Staging Lung Cancer [6], length of hospital stay and thoracic drainage, and postoperative chemotherapy or radiotherapy. Postoperative complications occurring within the 30 days after surgery were recorded. Pneumonia, extrarespiratory infection, bronchopleural fistulas, pleural or parietal bleeding, anemia requiring transfusion, pulmonary embolism, respiratory failure and adult respiratory distress syndrome, prolonged drainage, cardiovascular complications, other complications, and death were noted. Statistical Methods Univariate analysis of factors associated with postoperative mortality and morbidity was performed fitting the unconditional logistic regression models of Breslow and Day [7]. Predictive complication factors reviewed were age, stage, preoperative FEV 1, PaO 2, and hemoglobin. Factors with a p value smaller than 0.20 (likelihood ratio test) were included in the multivariate analysis. Odds ratios were presented with corresponding 95% confidence intervals. Survival was estimated using Kaplan-Meier estimators. Survival curves were compared by log-rank tests. A multivariate analysis was performed fitting a Cox proportional hazard model [8]. Statistical tests (two-tailed) were considered as significant at the 0.05 level. Results Demographics We reviewed 114 patients, 100 men and 14 women, with a median age of 61 years (range, 25 to 74). Histologic tumor analysis detected 64 squamous cell carcinomas, 29 adenocarcinomas, 17 large cell carcinomas, and 4 large cell neuroendocrine carcinomas. The majority of patients had reached clinical stage III before induction therapy, and mediastinoscopy was not usually carried out. Neoadjuvant Treatments All patients received preoperative chemotherapy, cisplatine-vinorelbine in the case of 56 patients, cisplatinegemcitabine for 12 patients, carboplatine-vinorelbine for 5 patients, and other regimens. Twenty patients also Table 2. Postoperative Complications Complication No. of Comments Pneumonia 7 Extrapulmonary infection 0 Bronchopleural fistula 0 Pleural or parietal bleeding 5 Requiring transfusion Anemia requiring 6 Without hemorrhage transfusion Pulmonary embolism 0 Respiratory distress 3 No deaths Prolonged drainage 4 Cardiovascular 5 Auricular fibrillations, 5 complications Other 1 Urine retention Deaths 2 Hemorrhage and left pneumonia received preoperative radiation, usually for apical tumors, and received a dose of radiation of 40 Gy. Surgical Resection Forty-one lobectomies, 4 bilobectomies, 55 pneumonectomies (27 right pneumonectomies and 28 left pneumonectomies), 5 wedge resections, and 9 exploratory thoracotomies were performed. Pneumonectomy was the most common operation (48%). Complete mediastinal lymphadectomy was perfomed in all patients. The 9 patients who only had exploratory surgery had nonresectable tumors. Postoperative thoracic drainage lasted from 2 to 25 days, averaging 5 days; hospitalization lasted from 3 to 27 days, with an average of 11 days. Clinical and pathological tumor extensions are presented in Table 1. Postoperative Two of the 114 patients died. The first death, due to pneumonia, occurred 5 days after a right pneumonectomy, and postpneumonectomy mortality rate was 1 of 55. The second death, by hemoptysis, occurred 3 days after explorative thoracotomy. Because of the small number of deaths, we could not analyze predictive factors of mortality. Postoperative Morbidity Overall morbidity was 28.9%; 33 of the 114 patients had postoperative complications. The 2 deaths were included in the 33 because they had complications before death (pneumonia and massive hemoptysis). Complications are detailed in Table 2. The most common complications were pneumonia, anemia, and hemorrhage. Cardiovascular complications were usually auricular fibrillation. No bronchopleural fistulas were included in our series. The relationship between morbidity and the type of surgical resection is shown in Table 3. Twenty patients received preoperative radiotherapy and chemotherapy. Among those patients, 7 had postoperative complications. These patients complications were not statistically different compared with patients
Ann Thorac Surg PERROT ET AL 425 2005;80:423 7 CHEMOTHERAPY BEFORE NSCLC RESECTION Table 3. Postoperative and Morbidity by Type of Surgery Morbidity Type of Surgery No. of % No. of % Right pneumonectomy (n 27) 1 3.7 8 29.6 Left pneumonectomy (n 28) 0 0 9 32.1 Total after pneumonectomy (n 55) 1 1.8 17 30.9 Lobectomy (n 45) 0 0 11 24.4 Bilobectomy (n 4) 0 0 1 25 Exploratory thoracotomy (n 9) 1 11 2 22.2 Wedge resection (n 5) 0 0 2 40 Total (n 114) 2 1.8 33 28.9 without preoperative radiotherapy: 7 of 20 (35%) and 26 of 94 (27.6%), respectively (p 0.51). We analyzed age, stage, preoperative FEV 1, PaO 2, and hemoglobin as predictors of complications. The medians were 61 years for age, 80% for FEV 1, 11.8 kpa for PaO 2, and 11.5 for hemoglobin. The thresholds used were 60 years for age, I II versus III IV for stage, 80% of the theoretical value for FEV 1, 10 kpa for PaO 2, and 12 g/dl for hemoglobin. The univariate analysis demonstrated that postoperative morbidity was not higher among older ( 60 years) patients (odds ratio 1.55, confidence interval 0.67 to 3.55; p 0.301), more advanced stage patients (odds ratio 1.35, confidence interval 0.60 to 3.06; p 0.460), patients with lower FEV 1 (odds ratio 1.61, confidence interval 0.68 to 3.82; p 0.274), patients with lower PaO 2 (odds ratio 0.978, confidence interval 0.31 to 3.04; p 0.969), or patients with lower hemoglobin (odds ratio 0.907, confidence interval 0.39 to 2.06; p 0.815). None of these factors was included in multivariate analysis because all the p values were higher than 0.20. Fig 1. Overall survival by the Kaplan-Meier method. Survival Overall survival rates, analyzed by the Kaplan-Meier method, are shown in Figure 1. One-year and 5-year survivals were 76.5% and 35.5%, respectively. We analyzed the influence of the preceding factors on survival by univariate and multivariate analysis. The stage and preoperative PaO 2 were the two factors that influenced survival among age, stage, preoperative FEV 1, PaO 2, and hemoglobin in univariate and multivariate analysis (p 0.003 and 0.002, respectively). Comment Postoperative complications in relation to preoperative chemotherapy for NSCLC in randomized studies are summarized in Table 4 (postoperative morbidity was not specified in most studies). Postoperative mortality varies from 0% to 7.8%, and the study by Depierre and associates [3], which includes the largest number of patients, shows increased postoperative mortality in the group given chemotherapy. As is this study, other studies providing details on postoperative mortality and morbidity after preoperative chemotherapy are retrospective. The most important study is that by Martin and coworkers [4], who retrospectively analyzed 470 patients who had undergone preoperative chemotherapy before ablation of NSCLC. Postoperative mortality and morbidity were, respectively, 3.8% and 38%, and right pneumonectomy was a predictive complication factor in multivariate analysis. Roberts and associates [5] analyzed 34 patients operated on after preoperative chemotherapy; there were no deaths after surgery, and postoperative major complications were 47%. This study concludes that neoadjuvant chemotherapy increases perioperative complications. Bernard and colleagues [10] analyzed 64 patients who received preoperative chemotherapy: postoperative mortality and morbidity were, respectively, 2% and 24%, and neoadjuvant chemotherapy remained independently associated with postoperative complications in multivariate analysis (p 0.02). The results from studies analyzing postoperative mortality and morbidity outside all neoadjuvant treatment are summarized in Table 5. All these studies are retrospective. Postoperative mortality rates varied from 1.3%
426 PERROT ET AL Ann Thorac Surg CHEMOTHERAPY BEFORE NSCLC RESECTION 2005;80:423 7 Table 4. Postoperative Rates in Randomized Trials Author, Reference, Year Number of Postoperative in Surgical Group in Neoajuvant Chemotherapy Group Pass et al [9] 1992 27 Stage IIIA 0% 0% 0% Rosell et al [1] 1994 60 Stage IIIA 7% 7% 7% Roth et al [2] 1994 60 Stage IIIA 3% 6% 0% Depierre et al [3] 373 Stage IB, II, IIIA 7.8% 4.5% 6.7% to 6.6% for overall mortality, from 1.6% to 13.4% for death after pneumonectomy, and from 0.6% to 4% for death after lobectomy. The most frequently reported causes of death were respiratory distress, pneumonia fistulas, and pleural empyema. Table 5 shows clearly that both postoperative mortality and morbidity are higher after pneumonectomy. In our study, overall mortality was 1.8%, namely, 2 deaths among 114 patients, the one from massive hemoptysis 3 days after an exploratory thoracotomy and the other from pneumonia 5 days after a right pneumonectomy. It was noted that overall mortality and mortality rates after pneumonectomies are identical (1.8%). On the other hand, mortality after right pneumonectomies was higher at 3.7% (1 of 27) as well as after exploratory thoracotomies, which reached 11% (1 of 9). These percentages are of little value, however, owing to the small number of deaths. The rate of postoperative complications in our study reached 28.9% (33 of 114 patients), the most common complications being pneumonia, anemia with transfusion, and hemorrhage. The majority of patients had advanced stage cancers, as 79% were in preoperative clinical stage III. Postoperative mortality and morbidity rates in our study (1.8% and 28.9%, respectively) are identical to, or even less than, those in the majority of retrospective studies with or without preoperative chemotherapy. These figures are also less than those observed for patients operated on in the same unit without neoadjuvant treatment: of 2,238 NSCLC ablations, there were 57 deaths, or a total mortality rate of 2.7%. In addition, our results are very close to the postoperative mortality rate of patients who had resection of a NSCLC in the same thoracic surgery unit, without preoperative chemotherapy: of 2,210 patients operated on, there were 57 deaths, or a mortality rate of 2.7%. Several elements can perhaps explain the limited number of postoperative complications in our sample. First, all the patients in our study were operated on exclusively by thoracic surgeons. It was, in any case, noted that systematic ganglion curettage carried out by the surgeons in our study did not raise the mortality rate. The retrospective study by Silvestri and coworkers [21] carried out on 1,583 patients showed a higher postoperative mortality rate for patients operated on by general surgeons than for those operated on by specialist thoracic surgeons (5.3% against 3%, p 0.05). In addition, all patients in the study were operated on in a unit with a high number of thoracotomies performed (about 150 bronchial cancers operated on in a year). The study by Bach and colleagues [22], which included 2,118 patients operated on in 76 different hospitals, concluded that that there were lower complication rates in large centers (3% mortality, 20% morbidity) than in hospitals handling fewer procedures (6% mortality, 44% morbidity). One reason for the higher postoperative mortality in the study by Depierre and associates [1] is probably that it was a multicenter study, which included centers with variable inclusion levels, corresponding to teams with more or less training. In our study, thanks to the high number of surgical interventions per year, the teams of anesthetists, nurses, and physiotherapists were used to performing bronchial cancer surgery. Moreover, patients in our study did not go into the intensive care unit in the operating theaters and bene- Table 5. Postoperative and Morbidity Rates in 10 Retrospective Studies Without Neoadjuvant Therapy Author, Reference Year Number of Stage Overall After Pneumonectomy After Lobectomy Morbidity Ginsberg et al [11] 1983 2,200 All stages 3.7% 6.2% 2.9% Not specified Deslaurier et al [12] 1994 783 All stages 3.8% 27% Shah et al [13] 1996 313 Stage I and II 3.2% 1.6% 4.03% 6.7% Duque et al [14] 1997 605 All stages 6.6% 13.4% 4.4% 32.4% Wada et al [15] 1998 7,099 All stages 1.3% 3.2% 1.2% Not specified Joo et al [16] 2001 105 Only pneumonectomy 10.5% Myrdal et al [17] 2001 616 All stages 2.9% 5.7% 0.6% 30.8% Bernard et al [18] 2001 639 Only pneumonectomy 7% 38.3% Alexiou et al [19] 2001 206 Only pneumonectomy 6.8% 39% Licker et al [20] 2002 193 Only pneumonectomy 9.3% 47%
Ann Thorac Surg PERROT ET AL 2005;80:423 7 CHEMOTHERAPY BEFORE NSCLC RESECTION fited from intensive respiratory physiotherapy in the surgical unit. That would appear to be an important point for explaining the low incidence of complications: artificial ventilation time is shorter, the risk of postoperative nosocomial infections is probably less, and the patient s treatment is less invasive. The selection of patients probably also has a beneficial role: the majority of operating questions were asked during a multidisciplinary meeting of lung specialists, thoracic surgeons, radiologists, and pathologists. In addition, patients who benefited from preoperative chemotherapy were often young and generally in better health, whereas older patients or those with a performance status score of 2 or more were more likely to be operated on straightaway. Finally, the preoperative FEV 1, PaO 2, and hemoglobin values show that most patients had good respiratory function and were not anemic before the operation. In our study, none of the data collected preoperatively was found to be predictive of complication factors. who had received preoperative radiotherapy seemed to have more postoperative complications than others (35% against 28.9%, respectively), but this difference was not statistically significant. The thresholds that we chose (80% for FEV 1, 10 kpa for PaO 2, and 12 g/dl for hemoglobin) corresponded to the lowest normal limits for each value. Our analysis of the survival of 114 patients by the Kaplan-Meier method showed overall survival rates of 76% at 1 year and 35% after 5 years, rates that correspond more or less to survival at the IIB stage of the International System for Staging Lung Cancer [6], although our study contains 23% stage II and 44% stage III patients. Factors that influenced survival in our study are the stage (p 0.003) and preoperative PaO 2 with a limit of 10 kpa (p 0.002). The fact that stage influences survival confirms known data. On the other hand, the results concerning PaO 2 indicate increased vigilance with patients showing preoperative hypoxia. References 1. Rosell R, Gomez-Condina J, Camps C, et al. A randomized trial comparing chemotherapy plus surgery with surgery alone in patients with non-small cell lung cancer. N Engl J Med 1994;330:153 8. 2. Roth JA, Fossella F, Komaki R, et al. A randomized trial comparing preoperative chemotherapy and surgery with surgery alone in resectable stage IIIa non-small cell lung cancer. J Natl Cancer Inst 1994;86:673 80. 3. Depierre A, Milleron B, Morot-Sibilot D, et al. From the French Thoracic Cooperative Group. Preoperative chemotherapy followed by surgery compared with primary surgery 427 in resectable stage I (except T1N0), II, and IIIa non-small cell lung cancer. J Clin Oncol 2001;20:247 53. 4. Martin J, Ginsberg RJ, Abolhoda A, et al. Morbidity and mortality after neoadjuvant therapy for lung cancer: the risks of right pneumonectomy. Ann Thorac Surg 2001;72:1149 54. 5. Roberts JR, Eustis C, Devore R, et al. Induction chemotherapy increases perioperative complications in patients undergoing resections for non-small cell lung cancer. Ann Thorac Surg 2001;72:885 8. 6. Brambilla E, Travis WD, Colby TV, Corrin B, Shimosato Y. The New World Health Organization classification of lung tumours. Eur Respir J 2001;18:1059 68. 7. Breslow NE, Day NE. Statistical methods in cancer research. Volume I. The analysis of case-control studies. IARC Sci Publ, No 32, 1980. 8. Cox DR. Regression models and life tables. J R Stat Soc 1972;34:187 220. 9. Pass HI, Pogrebniak HW, Steinberg SM, et al. Randomized trial of neoadjuvant therapy for lung cancer: interim analysis. Ann Thorac Surg 1992;53:992 8. 10. Bernard A, Ferrand L, Hagry O, et al. Identification of pronostic factors determining risk groups for lung resection. Ann Thorac Surg 2000;70:1161 7. 11. Ginsberg RJ, Hill LD, Eagan RT, et al. Modern thirty day operative mortality for surgical resections in lung cancer. J Thorac Cardiovasc Surg 1983;86:654 8. 12. Deslaurier J, Ginsberg RJ, Piantadosi S, et al. Prospective assessment of 30-day operative morbidity for surgical resections in lung cancer. Chest 1994;106(Suppl):329 30S. 13. Shah R, Sabanathan S, Richardson J, et al. Results of surgical treatement of stage I and II lung cancer. J Cardiovasc Surg 1996;37:169 72. 14. Duque JL, Ramos G, Castrodeza J, et al. Early complications in surgical treatment of lung cancer: a prospective, multicenter study. Ann Thorac Surg 1997;63:944 50. 15. Wada H, Nakamura T, Nakamoto K, et al. Thirty day operative mortality for thoracotomy in lung cancer. J Thorac Cardiovasc Surg 1998;115:70 3. 16. Joo JB, DeBord JR, Montgomery CE, et al. Perioperative factors as predictors of operative mortality and morbidity in pneumonectomy. Am Surg 2001;67:318 22. 17. Myrdal G, Gustafsson G, Lambe M, et al. Outcome after lung cancer surgery. Factors predicting early mortality and major morbidity. Eur J Cardiothorac Surg 2001;20:694 9. 18. Bernard A, Deschamps C, Allen MS, et al. Pneumonectomy for malignant disease: factors affecting early morbidity and mortality. J Thorac Cardiovasc Surg 2001;121:1076 82. 19. Alexiou C, Beggs D, Rogers ML, et al. Pneumonectomy for non-small-cell lung cancer: predictors of operative mortality and survival. Eur J Cardiothorac Surg 2001;20:476 80. 20. Licker M, Spiliopoulos A, Frey JG, et al. Risks factors for early mortality and major complications following pneumonectomy for non-small-cell lung cancer. Chest 2002;121: 1890 7. 21. Silvestri GA, Handy J, Lackland D, et al. Specialists achieve better outcomes than generalists for lung cancer surgery. Chest 1998;114:675 80. 22. Bach PB, Cramer LD, Downey RJ, et al. The influence of hospital volume on survival after resection for lung cancer. N Engl J Med 2001 19;345:181 8.