Outcome of Esophagectomy for Cancer in Elderly Patients

Outcome of Esophagectomy for Cancer in Elderly Patients Tanja M. Cijs, MD, Cees Verhoef, MD, PhD, Ewout W. Steyerberg, PhD, Linetta B. Koppert, MD, PhD, T. C. Khe Tran, MD, Bas P. L. Wijnhoven, MD, PhD, Hugo W. Tilanus, MD, PhD, and Jeroen de Jonge, MD, PhD Department of Surgery, Division of Gastrointestinal and Transplantation Surgery, Department of Surgery, Division of Surgical Oncology, and Department of Public Health, Center for Medical Decision Making, Erasmus Medical Center Rotterdam, Rotterdam, Netherlands Background. This study analyzes the outcome of esophageal resection in patients 70 or more years of age, compared with patients aged less than 70 years and identifies risk factors for worse outcome in the elderly. Methods. Comorbidity, postoperative morbidity, inhospital mortality and survival rates were compared between 811 patients aged less than 70 years and 250 patients aged 70 years or more who underwent esophagectomy for esophageal cancer in a single high-volume center from 1985 to 2005. Results. Groups were similar regarding surgical approach, resectability, and tumor stage. More patients aged 70 years or more had cardiovascular and respiratory concomitant disease. Among patients aged 70 years or more, the prevalence of adenocarcinoma and Barrett s transformation was higher (67% versus 53% for patients aged less than 70 years, and 22% versus 15%, respectively). There were no differences in surgical complications (20% versus 17%). Nonsurgical complications occurred more in patients aged 70 years or more (35% versus 27%) and operative mortality was higher among elderly patients (8.4 versus 3.8%), as was in-hospital mortality (11.6% versus 5.4%). The disease-specific 5-year survival was lower for patients aged 70 years or more (27% versus 34%). The 1-year survival, reflecting the impact of operative morbidity and mortality, was 58% for patients aged 70 years or more and 68% for the patients aged less than 70 years (p 0.002). Among patients aged 70 years or more, respiratory comorbidity and thoracoabdominal resection were risk factors for the occurrence of nonsurgical complications and respiratory comorbidity for in-hospital mortality. Conclusions. Older patients have increased operative and in-hospital mortality and decreased 5-year survival after esophageal resection for cancer. Our results indicate that especially thoracoabdominal resection for esophageal carcinoma should be carefully considered for patients older than 70 years who suffer from respiratory disease. (Ann Thorac Surg 2010;90:900 7) 2010 by The Society of Thoracic Surgeons Accepted for publication May 17, 2010. Address correspondence to Dr de Jonge, Erasmus MC Rotterdam, s-gravendijkwal 230, Rotterdam 3015 CE, Netherlands; e-mail: j.dejonge.1@erasmusmc.nl. Cancer of the esophagus and gastroesophageal junction shows a peak incidence among patients aged between 65 and 70 years. Some 20% of these carcinomas occur in patients more than 70 years of age [1, 2]. Surgery is still considered the treatment of choice in patients with carcinoma of the esophagus and gastroesophageal junction. However, with increasing life expectancy of the general population, the number of elderly patients referred for major surgical procedures continues to rise [3, 4]. The suitability of older patients to withstand an esophageal resection remains a controversial issue, as high morbidity and mortality rates after esophagectomy for cancer in the elderly have been reported. Several studies comparing the results of esophagectomy between elderly and younger patients showed both worse [5 9] and equal outcome [3, 4, 10 20] with regard to postoperative morbidity, mortality, and long term survival. However, most of these studies report on a rather small number of elderly patients. Furthermore, selection criteria and definitions of postoperative morbidity and mortality vary between the studies, which makes interpretation of the data difficult [21]. We studied short-term and long-term outcomes of esophagectomy in a large single-center group of patients aged more than 70 years. In addition, we identified in this study risk factors for worse outcome in the elderly patient. Patients and Methods Between January 1985 and January 2005, 1950 patients with carcinoma of the esophagus or gastroesophageal junction were evaluated for surgical treatment. Patient and tumor characteristics of these patients and individual outcome are filed prospectively in the database of the Rotterdam Esophageal Tumor Study Group. The Institutional Review Board of the Erasmus Medical Center was requested permission to perform this study, and necessity to obtain individual consent was waived as individuals are registered as entries and individual patients are not identified. 2010 by The Society of Thoracic Surgeons 0003-4975/$36.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2010.05.039

Ann Thorac Surg CIJS ET AL 2010;90:900 7 OUTCOME OF ESOPHAGECTOMY IN THE ELDERLY 901 In all patients, preoperative workup and staging were performed including esophagoscopy, computed tomography scan (since 1986) of the thorax and abdomen, and chest radiography. For carcinomas of the upper intrathoracic esophagus, bronchoscopy and cervical ultrasonography were performed. Pulmonary function tests and arterial blood gas analyses were performed in patients with concomitant respiratory disease. An electrocardiogram was obtained for all patients, whereas a dobutamine stress test was performed on indication. Since 1993, a routine endosonography was performed, which resulted in a more accurate assessment of gastrointestinal regional tumor stage [22]. Patients were considered eligible for surgical exploration if there was no evidence of metastasized disease and they had sufficient cardiorespiratory reserve. Eventually, 1,277 patients underwent surgical exploration. The surgical technique was based on a combined right thoracotomy and separate laparotomy for mid and distal esophageal tumors or by transhiatal resection for carcinomas of the gastroesophageal junction [23, 24]. The anastomosis between the remnant esophagus and stomach tube was created in the neck. After the operation, all patients were transferred to the intensive care unit. Postoperative analgesia was provided by continuous infusion of a local anesthetic mixture (marcaine/fentanyl) through an epidural catheter or by repetitive intravenous injections of morphine 0.1 mg/kg. Patients received nutrition by a routinely placed feeding jejunostomy or nasojejunal feeding tube (Flocare Bengmark Tube; Nutricia, Schiphol, Netherlands). Contrast swallow radiography was carried out on the seventh postoperative day to assure anastomotic integrity before the start of oral feeding. Postoperative complications were predefined in the database and divided into surgical complications and nonsurgical complications. Surgical complications were defined as bleeding, anastomotic leak, necrosis of the gastric tube, chylothorax, and leakage of the feeding jejunostomy. Nonsurgical complications were defined as pneumonia, adult respiratory distress syndrome, acute myocardial infarction, sepsis, and thrombosis or embolism. Other complications such as wound infection, urinary tract infection, and persistent dysphagia were registered but not considered as severe. Operative mortality was defined as any death occurring within 30 days of the operation; in-hospital mortality was defined as any death occurring during hospital admittance. Patients were seen in the outpatient clinics every 3 months for the first year, every 6 months for the next 2 years and yearly thereafter, until 5 years after surgery. Survival time was calculated from the date of surgery until death or end of follow-up. Statistics Patients aged 70 years or older were compared with patients aged younger than 70 years with regard to their preoperative morbidity, tumor characteristics, type of surgery, prevalence of postoperative complications, operative mortality, and 5-year survival. Values are given as mean and standard error (SE) or as median and range. Proportions were compared with Fisher s exact test. Nonparametric tests were used to make comparisons for nonnormally distributed variables. Univariate and multivariate analysis of risk factors for morbidity and mortality was performed using binary logistic regression analysis. Only variables that showed statistic significance (p 0.05) in the univariate analysis were included in the multivariate model. Predicted 30- day mortality was calculated using the formula of Steyerberg and coworkers [25] for every 5-year interval time period. In this scoring system, four variables are considered: age, comorbidity, neoadjuvant therapy, and hospital volume. The sum of these factors as a score is entered into a formula to calculate the 30-day mortality risk as percentage (Tables 1 and 2). This simple scoring system has a better discriminative power than the O-POSSUM (Oesophagogastric Surgery Physiological and Operative Severity Score for the Enumeration of Mortality and Morbidity) score and equal discrimination compared with the P-POSSUM (Portsmouth Predictor Equation-POSSUM) scoring system (area under the receiver operating characteristic curve is 0.70, 0.68, and 0.61, respectively). Survival rates were calculated with the method of Kaplan and Meier. Differences in survival were compared with the Mantel-Cox log rank test. Multivariate analysis of risk factors for disease-specific 5-year survival was performed using Cox regression analysis. In a subanalysis of the patients more than 70 years of age, risk factors for operative mortality and postoperative morbidity were identified separately. To indicate relative sur- Table 1. Steyerberg Risk Score for 30-Day Mortality Characteristics Score Age 50 years 1 60 years 0 85 years 1 Comorbidity Respiratory 1 Cardiovascular 1 Diabetes mellitus 1 Hepatic 1 Renal 1 Neoadjuvant therapy Radiotherapy 1.5 Chemoradiotherapy 1 Hospital volume number of esophagectomies per year Low, 1 0 Intermediate, 1.1 to 2.5 0.5 High, 2.6 1.5 Very high, 50 2 Sum score is obtained by adding scores. Intermediate scores for age can be approximated by linear interpolation. For example, age 72 corresponds to a score of 0.5. The formula to calculate predicted 30-day mortality 1/[1 exp (2.4 0.32 score)]. The Erasmus Medical Center is a very high volume tertiary referral center.

902 CIJS ET AL Ann Thorac Surg OUTCOME OF ESOPHAGECTOMY IN THE ELDERLY 2010;90:900 7 Table 2. Changes in Operative (30-Day) Mortality During Study Period 1985 1989 n 259 1990 1994 n 299 1995 1999 n 238 2000 2004 n 266 Predicted 70 years old 6.1% 5.3% 4.6% 4.6% Observed 70 years old 13/199 (6.5%) 9/235 (3.8%) 5/178 (2.8%) 4/199 (2.0%) Predicted 70 years old 9.0% 7.6% 6.3% 6.5% Observed 70 years old 11/60 (18%) 3/64 (6.3%) 3/60 (5.0%) 3/67 (4.7%) Predicted 30-day mortality calculated with the Steyerberg risk index. vival in the patients 70 years of age or older, diseasespecific survival was compared with the expected survival for a corresponding group of people in the general population with the referral life expectancy tables from the Human Mortality Database (University of California, Berkeley, CA; www.mortality.org). All statistical analyses were performed using the statistical package SPSS 11.0 (SPSS, Chicago, IL). A p value less than 0.05 was considered statistically significant. Results Of the initial 1950 patients evaluated for surgery, 1386 patients were less than 70 years and 564 patients were aged 70 years or older. After preoperative work-up and staging, esophageal resection seemed feasible for 1,277 patients: 976 patients aged less than 70 years (70%) and 301 patients aged 70 or older (53%; p 0.001). Eventually, in 1,061 of these 1,277 patients (83%) scheduled for surgical exploration, an esophageal resection could be performed. Of these patients, 811 (76%) were less than 70 years of age and 250 patients (24%) were more than 70 years. Demographic details and age distributions are shown in Table 3. Resectability (defined as the proportion of patients who underwent a resection among those who were scheduled for surgery) was not significantly different between the two groups (83% in both groups, p 1.0). Preexisting respiratory diseases Table 3. Patient and Tumor Characteristics of Patients With Esophageal Resection Demographics 70 Years (n 811) 70 Years (n 250) Sex Male (%) 618 (76) 177 (71) 0.10 Female (%) 193 (24) 73 (29) Age, range (median) 31 69 (59) 70 86 (74) Octogenarians (%) 21 (8.4) Resectability (%) 811/976 (83) 250/301 (83) 1.0 Comorbidity Respiratory (%) 102 (13) 61 (24) 0.001 Cardiovascular (%) 149 (18) 70 (28) 0.001 Diabetes mellitus (%) 37 (5) 15 (6) 0.40 Neoadjuvant therapy Radiotherapy (%) 244 (30) 58 (23) Chemotherapy (%) 85 (10) 12 (5) Chemoradiotherapy (%) 10 (1) 1 (0) 0.001 Surgical approach Transhiatal (%) 493 (61) 166 (66) Thoraco/laparotomy (%) 318 (39) 84 (34) 0.12 Tumor type Adenocarcinoma (%) 306 (38) 111 (44) Adenocarcinoma in Barrett (%) 123 (15) 56 (22) Squamous cell carcinoma (%) 318 (39) 71 (28) No malignancy after neoadjuvant therapy (%) 40 (5) 7 (3) 0.016 Tumor stage I (T1 N0 M0) (%) 13 (2) 4 (2) IIa (T2,3 N0 M0) (%) 106 (13) 35 (14) IIb (T1,2 N1 M0) (%) 330 (41) 102 (41) III (T3 N1 M0) (%) 231 (29) 75 (30) (T4 anyn M0) IV (anyt anyn M1) (%) 89 (11) 27 (11) 0.89

Ann Thorac Surg CIJS ET AL 2010;90:900 7 OUTCOME OF ESOPHAGECTOMY IN THE ELDERLY 903 Table 4. Postoperative Complications and Outcome 70 Years (n 811) 70 Years (n 250) Complications Surgical (%) 138 (17) 51 (20) 0.271 Nonsurgical (%) 220 (27) 88 (35) 0.017 Operative mortality (%) 31 (3.8) 20 (8.0) 0.011 In-hospital mortality (%) 44 (5.4) 28 (11) 0.002 One-year survival (%) 68 59 0.003 Five-year survival (%) a 34 26 Median months (95% CI) 24.9 (21.1 28.6) 20.2 (17.2 23.2) 0.030 a Disease-specific survival. CI confidence interval. (chronic obstructive airway disease, asthma, emphysema) and cardiovascular diseases (ischemic heart disease, hypertension, myocardial infarction) were significantly more prevalent among patients more than 70 years (13% versus 24%, p 0.001, and 18% versus 28%, p 0.002, respectively; Table 3). There was no difference in diabetes mellitus (5% versus 6%, p 0.46). Patients less than 70 years of age received significantly more often neoadjuvant treatment than patients aged 70 years or older (41% versus 28%, p 0.001). The surgical approach and histopathology staging are also shown in Table 3. Four hundred and three patients (38%) underwent a combined thoracotomy and laparotomy, and 659 patients (62%) underwent a transhiatal resection. There were no significant differences between groups regarding surgical approach. Tumor localization, ptnm classification, and tumor stage were comparable between the two groups (Table 3). Histopathology examination showed significant differences in tumor type. Patients less than 70 years old had adenocarcinoma in 53% and squamous cell carcinoma in 39%. Among patients 70 years or older, adenocarcinoma was present in 67% and squamous cell carcinoma in 28% (p 0.001). Adenocarcinoma in Barrett s transformed esophagus was present in 15% of the patients less than 70 years old and in 22% of the patients aged 70 years or more (p 0.028). Predefined postoperative complications occurred in 39% of all patients. There was no significant difference in the number of surgical complications between patients younger than 70 yeas (138 of 811; 17%) and patients aged 70 years or more (51 of 251; 20%; p 0.271; Table 4). Nonsurgical complications, however, occurred significantly less frequently among patients less than 70 years of age (220 of 811; 27%), compared with patients more than 70 years of age (88 of 251; 35%; p 0.017; Table 4). Overall operative (30-day) mortality was 3.8% for patients younger than 70 years and 8.0% for patients aged 70 years or older (p 0.011). In Tables 1 and 2, the Steyerberg score of expected 30-day mortality was applied to the study population (Table 1) and compared with observed mortality per 5-year time intervals (Table 2). In both groups, mortality decreased significantly over time. Excessive mortality was identified in the patients aged 70 years or older in the first 5-year time interval. In-hospital mortality was also significantly higher for patients more than 70 years old: 5.4% for the group less than 70 years of age and 11.6% for the group more than 70 years old (p 0.001). Hospital stay itself was equal in both groups: 25.2 0.7 days for patients less than 70 and 26.0 1.2 days for patients more than 70 years of age (values are means SE; p 0.47). The 1-year survival, reflecting the impact of operative morbidity and mortality, was 68% for patients aged less Fig 1. Overall and cancer specific 5-year survival for patients less than 70 years of age and patients more than 70 years of age. The expected survival of the normal Dutch population with the same age and sex distribution at the time of surgery is indicated as reference. Numbers are patients at risk. The difference in cancer-specific survival between patients younger than 70 years and patients older than 70 years of age is statistically significant (p 0.03). The life loss is represented by the difference between the expected survival and actuarial survival for patients aged 70 years or older. For patients aged less than 70 years, disease-specific survival (solid squares) was 33.7% (median 24.9), and overall survival (open squares) was 29.3% (median 21.5 months). For patients aged 70 years or more, disease-specific survival (solid triangles) was 26.8% (median 20.2 months), and overall survival (open triangles) was 21.7% (median 16.9 months).

Table 5. Risk Factors for Surgical and Nonsurgical Morbidity, Operative Mortality, and Disease-Specific 5-Year Survival Risk Factor Surgical Complication Nonsurgical Complication Operative Mortality Disease-Specific 5-Year Survival Univariate Multivariate Univariate Multivariate Univariate Multivariate Univariate Multivariate Disease-Specific 5-Year Survival Age 70 years a 1.25 (0.88 1.79) 1.46 (1.08 1.97) 1.44 (1.06 1.97) 2.19 (1.22 3.91) 2.83 (1.49 5.38) 26% versus 34% 1.21 (1.00 1.46) 0.22 0.014 0.021 0.008 0.001 0.030 0.049 Age continuous, per decade 1.21 (1.03 1.42) 1.10 (0.97 1.26) 1.56 (1.14 2.12) 1.15 (1.06 1.25) 0.020 0.15 0.005 0.001 Study period, per 0.86 (0.63 1.18) 0.87 (0.67 1.14) 0.44 (0.24 0.82) 0.58 (0.24 1.41) 30% versus 33% decade b 0.35 0.32 0.010 0.23 0.46 Respiratory 1.45 (0.97 2.18) 1.68 (1.18 2.37) 1.68 (1.17 2.40) 2.25 (1.18 4.27) 2.32 (1.16 4.61) 25% versus 33% comorbidity a 0.07 0.004 0.005 0.013 0.017 0.06 Cardiovascular 1.17 (0.80 1.70) 1.04 (0.75 1.44) 1.65 (0.88 3.07) 23% versus 34% 1.22 (1.00 1.48) comorbidity a 0.43 0.81 0.12 0.009 0.06 Diabetes mellitus 0.97 (0.46 2.02) 0.72 (0.37 1.40) 2.86 (1.16 7.05) 2.97 (1.04 8.50) 23% versus 32% comorbidity a 0.93 0.33 0.023 0.043 0.11 Neoadjuvant 1.01 (0.71 1.42) 1.54 (1.16 2.04) 1.05 (0.73 1.51) 2.62 (1.49 4.61) 2.02 (0.80 5.06) 33% versus 32% radiotherapy a 0.97 0.003 0.80 0.001 0.14 0.69 Neoadjuvant 1.14 (0.69 1.89) 1.34 (0.88 2.05) 0.75 (0.26 2.12) 38% versus 31% chemotherapy a 0.61 0.17 0.58 0.32 Operative 1.06 (0.77 1.47) 1.92 (1.47 2.52) 1.96 (1.39 2.76) 2.25 (1.28 3.98) 1.15 (0.52 2.55) 32% versus 33% technique c 0.71 0.001 0.001 0.005 0.73 0.89 Tumor type 0.85 (0.61 1.18) 0.89 (0.68 1.17) 0.47 (0.26 0.86) 0.52 (0.27 0.99) 28% versus 36% histology d 0.32 0.41 0.014 0.045 0.21 Tumor stage e 1.10 (0.80 1.52) 0.83 (0.63 1.09) 0.71 (0.38 1.32) 12% versus 45% 2.80 (2.37 3.30) 0.56 0.19 0.28 0.001 0.001 Reference categories: a present (versus absent); b 1996 2005 (versus 1985 1995); c thoracoabdominal (versus transhiatal); d adenocarcinoma versus squamous cell carcinoma; and e stage III and IV (versus stage I and II). CI confidence interval. 904 CIJS ET AL Ann Thorac Surg OUTCOME OF ESOPHAGECTOMY IN THE ELDERLY 2010;90:900 7

Ann Thorac Surg CIJS ET AL 2010;90:900 7 OUTCOME OF ESOPHAGECTOMY IN THE ELDERLY 905 Table 6. Subgroup Analysis of Patients More Than 70 Years of Age: Risk Factors for Surgical and Nonsurgical Morbidity and Operative Mortality Surgical Complication Nonsurgical Complication In-Hospital Mortality Univariate Multivariate Univariate Multivariate Univariate Multivariate Risk Factor Age, continuous 0.95 (0.86 1.05) 1.05 (0.97 1.14) 0.99 (0.88 1.12) 0.34 0.23 0.90 Study period, per 1.01 (0.55 1.87) 0.72 (0.43 1.21) 0.59 (0.27 1.32) decade a 0.98 0.21 0.20 Respiratory 1.76 (0.90 3.45) 2.19 (1.21 3.94) 2.18 (1.20 3.97) 3.14 (1.40 7.05) 3.40 (1.48 7.80) comorbidity b 0.10 0.009 0.011 0.006 0.004 Cardiovascular 0.85 (0.42 1.72) 0.95 (0.53 1.69) 2.12 (0.95 4.75) comorbidity b 0.66 0.85 0.07 Diabetes mellitus b 0.58 (0.13 2.67) 0.65 (0.20 2.12) 2.10 (0.56 7.95) 0.49 0.48 0.28 Neoadjuvant 0.75 (0.38 1.60) 1.49 (0.82 2.72) 2.79 (1.24 6.31) 2.65 (0.92 7.6) radiotherapy b 0.45 0.19 0.013 0.07 Neoadjuvant 10.4 (3.07 35.5) 1.16 (0.37 3.66) 1.48 (0.31 7.03) chemotherapy b 0.001 0.80 0.63 Operative 0.99 (0.51 1.89) 2.07 (1.20 3.56) 2.07 (1.19 3.59) 2.17 (1.00 4.80) 1.26 (0.45 3.52) technique c 0.96 0.009 0.010 0.05 0.66 Tumor type 0.73 (0.38 1.42) 0.85 (0.48 1.50) 1.12 (0.47 2.67) histology d 0.36 0.56 0.80 Tumor stage e 1.17 (0.63 2.18) 0.71 (0.42 1.22) 0.45 (0.18 1.10) 0.62 0.22 0.08 Reference categories: a 1996 2005 (versus 1985 1995); b present (versus absent); c thoracoabdominal (versus transhiatal); d adenocarcinoma versus squamous cell carcinoma; and e stage III and IV (versus stage I and II). CI confidence interval. than 70 and 59% for patients aged 70 years or older (p 0.003). The overall 5-year survival including postoperative mortality and all causes of death was 27% for all patients (median 20.1 months, 95% confidence interval [CI]: 18.2 to 22.0). The 5-year cancer-specific survival was still significantly higher for patients less than 70 years (34%, median 24.9 months, 95% CI: 21.1 to 28.6), compared with patients 70 years of age or older (26%, median 20.2 months, 95% CI: 17.2 to 23.2; p 0.030; Fig 1). To have a visual indication about the loss of life in the group of patients aged 70 years or older, in Figure 1, the life expectancy of the normal Dutch population with the same age and sex distribution as our study population is indicated. In the normal population, life expectancy was still 9.0 years for male patients and 11.4 years for female patients at the time of surgery. After 5 years, 27% would have died in the normal population, indicating a reduction of life expectancy of about 45%. In the multivariate analysis, age more than 70 years was identified as independent risk factor for nonsurgical complications (odds ratio [OR] 1.4; p 0.021; Table 5). Other independent risk factors for nonsurgical complications were respiratory comorbidity (OR 1.7; p 0.005) and combined thoracoabdominal resection (OR 2.0; p 0.001). Risk factors for operative mortality in the multivariate analysis were age more than 70 years (OR 2.8; p 0.001), respiratory comorbidity (OR 2.3; p 0.017, diabetes mellitus (OR 3.0; p 0.043), and squamous cell carcinoma compared with adenocarcinoma (OR 1.9; p 0.045). Age more than 70 years (OR 1.2; p 0.049) and tumor stage (OR 2.8; p 0.001) were the two independent risk factors for 5-year survival in the Cox analysis. In a separate analysis of the group of patients more than 70 years of age, respiratory comorbidity (OR 2.2; p 0.011) and the use of a combined thoracoabdominal resection technique (OR 2.1; p 0.010) were independent risk factors for nonsurgical complications (Table 6). Respiratory comorbidity (OR 3.4; p 0.004) was the only identified risk factor for in-hospital mortality in the group of patients aged 70 years or older. Comment We investigated the outcome after esophagectomy for carcinoma of the esophagus and gastroesophageal junction in patients younger and older than 70 years of age and identified risk factors for postoperative morbidity and mortality in these patients. During the study period in this single-center study population there were no significant changes in operative technique, and the same surgeon supervised all the resections. Still, changes in anesthesiology management and postoperative care in-

906 CIJS ET AL Ann Thorac Surg OUTCOME OF ESOPHAGECTOMY IN THE ELDERLY 2010;90:900 7 fluenced the survival outcome over time. In our series, operative mortality decreased from 9.3% in the first 5 years to 2.6% in the last 5 years. This decrease was, however, equally present in patients younger than 70 years of age and the elderly, with persisting doubled mortality in the older than 70 years group (Table 1b). In this study, the simple Steyerberg risk score that is based on 3,592 patients from different cohorts had good agreement between expected and observed 30-day mortality in both younger and older patients. As expected, the number of patients with comorbidity was significantly higher in the elderly patient group. Postoperative overall complication rate (39%) was comparable between the groups and was not different from the results of Alexiou and colleagues (35% versus 25%) [3], Ellis and coworkers (38% versus 32%) [12], Griffin and colleagues (45%) [17], and Baily and associates (50%) [19]. The surgical morbidity among patients aged 70 years or older (20%) was not significantly different from that among patients younger than 70 years of age (17%). These morbidity rates were comparable to rates previously reported between 10% and 52% by Griffin and associates [17], Poon and coworkers [11], Fang and associates [26], and Rizk and colleagues [27]. When the elderly group was analyzed separately, no risk factor could be identified to predict surgical complications. Nonsurgical morbidity, however, was significantly increased among patients aged 70 years or older, independent of preexisting comorbidity. This is in line with the studies of Fang and colleagues [26] and Poon and colleagues [11] who also found significantly more respiratory and cardiovascular complications for patients more than 70 years of age (29% and 38%) compared with 17% and 24% for younger patients. In our study, respiratory comorbidity was an independent predictor of nonsurgical complications and in-hospital mortality for patients aged 70 years or older. Although this study was not to develop a model that estimates absolute risks of outcomes and therefore bootstrap validation was not performed, our findings are well in line with the results of the Abunasra study [28], which showed that respiratory comorbidity and advanced age are risk factors for death after esophagectomy. A significantly decreased disease-specific 5-year survival rate was seen for older patients compared with the younger group. The 5-year disease-specific survival in both groups was in accordance with the results of other authors [3, 4, 10 12, 15], except for the series by Rahamim and colleagues [9] who found a 5-year survival rate of 14% (older patients) and 18% (younger patients) and that of Fang and associates [26], who described a high 5-year survival rate of 55% (older patients) and 59% (younger patients). Recently, a large literature review addressed the problems of reporting postoperative mortality and advocated the additional use of the 1-year survival [21], which was 10% lower in the elderly group in this study. Also in our study, the difference in survival was present especially in the first year after resection owing to increased nonsurgical morbidity. The decision to perform thoracoabdominal surgery on patients aged 70 years or older with preexisting respiratory disease should be made carefully, as only a limited group of patients eventually will benefit from the resection, whereas all patients are subjected to increased risk for early mortality and morbidity. Finding the balance is becoming more difficult in the light of published survival updates from the Transhiatal Versus Extended Esophagectomy (HIVEX) trial, which suggest that a more extended transthoracic approach for tumors of the distal esophagus with limited lymph node involvement may increase 5-year survival for selected patients [29]; and the same group showed that postoperative complications after esophagectomy are independently associated with early timing of death due to cancer recurrence [30].In our study, the disease-specific 5-year survival of patients more than 70 years of age with respiratory comorbidity who were operated on with a combined thoracolaparotomy was only 14% (data not shown), and especially in this group of patients, clinical studies to compare the outcome of alternative treatment options such as definitive radio(chemo)therapy should be considered. In conclusion, esophageal resection for elderly patients is accompanied by an increased nonsurgical complication rate and operative and in-hospital mortality, and decreased 1-year survival. In this study, that resulted in a small, but statistically significant, reduction in 5-year cancer-specific survival. Chronologic age itself should not be used as a determinant to offer surgical treatment, but for patients aged 70 years or older, special attention should be given to preexisting pulmonary diseases, and thoracoabdominal resection should be avoided. Both are risk factors for increased nonsurgical morbidity, which is closely linked to in-hospital mortality for this group of patients. We would like to thank Frank J. M. F Dor, MD, PhD for participation in the correction of the manuscript. References 1. Ginsberg RJ. Cancer treatment in the elderly. J Am Coll Surg 1998 Oct;187:427 8. 2. Lerut T, Coosemans W, De Leyn P, Decker G, Deneffe G, Van Raemdonck D. Is there a role for radical esophagectomy. Eur J Cardiothorac Surg 1999;16(Suppl 1): 44 7. 3. Alexiou C, Beggs D, Salama FD, Brackenbury ET, Morgan WE. Surgery for esophageal cancer in elderly patients: the view from Nottingham. J Thorac Cardiovasc Surg 1998;116: 545 53. 4. Thomas P, Doddoli C, Neville P, et al. Esophageal cancer resection in the elderly. Eur J Cardiothorac Surg 1996;10: 941 6. 5. Tsutsui S, Moriguchi S, Morita M, et al. Multivariate analysis of postoperative complications after esophageal resection. Ann Thorac Surg 1992;53:1052 6. 6. Adam DJ, Craig SR, Sang CT, Cameron EW, Walker WS. Esophagectomy for carcinoma in the octogenarian. Ann Thorac Surg 1996;61:190 4. 7. Ferguson MK, Martin TR, Reeder LB, Olak J. Mortality after esophagectomy: risk factor analysis. World J Surg 1997;21: 599 604.

Ann Thorac Surg CIJS ET AL 2010;90:900 7 OUTCOME OF ESOPHAGECTOMY IN THE ELDERLY 907 8. Ferguson MK, Durkin AE. Preoperative prediction of the risk of pulmonary complications after esophagectomy for cancer. J Thorac Cardiovasc Surg 2002;123:661 9. 9. Rahamim JS, Murphy GJ, Awan Y, Junemann-Ramirez M. The effect of age on the outcome of surgical treatment for carcinoma of the oesophagus and gastric cardia. Eur J Cardiothorac Surg 2003;23:805 10. 10. Jougon JB, Ballester M, Duffy J, et al. Esophagectomy for cancer in the patient aged 70 years and older. Ann Thorac Surg 1997;63:1423 7. 11. Poon RT, Law SY, Chu KM, Branicki FJ, Wong J. Esophagectomy for carcinoma of the esophagus in the elderly: results of current surgical management. Ann Surg 1998;227:357 64. 12. Ellis FH, Williamson WA, Heatley GJ. Cancer of the esophagus and cardia: does age influence treatment selection and surgical outcomes? J Am Coll Surg 1998;187:345 51. 13. Xijiang Z, Xizeng Z, Xishan H, Hongjing J. Surgical treatment for carcinoma of the esophagus in the elderly patient. Ann Thorac Cardiovasc Surg 1999;5:182 6. 14. Johansson J, Walther B. Clinical outcome and long-term survival rates after esophagectomy are not determined by age over 70 years. J Gastrointest Surg 2000;4:55 62. 15. Kinugasa S, Tachibana M, Yoshimura H, et al. Esophageal resection in elderly esophageal carcinoma patients: improvement in postoperative complications. Ann Thorac Surg 2001;71:414 8. 16. Avendano CE, Flume PA, Silvestri GA, King LB, Reed CE. Pulmonary complications after esophagectomy. Ann Thorac Surg 2002;73:922 6. 17. Griffin SM, Shaw IH, Dresner SM. Early complications after Ivor-Lewis subtotal esophagectomy with two-field lymphadenectomy: risk factors and management. J Am Coll Surg 2002;194:285 97. 18. Sabel MS, Smith JL, Nava HR, Mollen K, Douglass HO, Gibbs JF. Esophageal resection for carcinoma in patients older than 70 years. Ann Surg Oncol 2002;9:210 4. 19. Bailey SH, Bull DA, Harpole DH, et al. Outcomes after esophagectomy: a ten-year prospective cohort. Ann Thorac Surg 2003;75:217 22. 20. Internullo E, Moons J, Nafteux P, et al. Outcome after esophagectomy for cancer of the esophagus and GEJ in patients aged over 75 years. Eur J Cardiothorac Surg 2008; 33:1096 104. 21. Jamieson GG, Mathew G, Ludemann R, Wayman J, Myers JC, Devitt PG. Postoperative mortality following oesophagectomy and problems in reporting its rate. Br J Surg 2004;91: 943 7. 22. de Manzoni G, Pedrazzani C, Di Leo A, et al. Experience of endoscopic ultrasound in staging adenocarcinoma of the cardia. Eur J Surg Oncol 1999;25:595 8. 23. Orringer MB, Sloan H. Esophagectomy without thoracotomy. J Thorac Cardiovasc Surg 1978;76:643 54. 24. Hulscher JB, van Sandick JW, de Boer AG, et al. Extended transthoracic resection compared with limited transhiatal resection for adenocarcinoma of the esophagus. N Engl J Med 2002;347:1662 9. 25. Steyerberg EW, Neville BA, Koppert LB, et al. Surgical mortality in patients with esophageal cancer: development and validation of a simple risk score. J Clin Oncol 2006;24: 4277 84. 26. Fang W, Igaki H, Tachimori Y, Sato H, Daiko H, Kato H. Three-field lymph node dissection for esophageal cancer in elderly patients over 70 years of age. Ann Thorac Surg 2001;72:867 71. 27. Rizk NP, Bach PB, Schrag D, et al. The impact of complications on outcomes after resection for esophageal and gastroesophageal junction carcinoma. J Am Coll Surg 2004;198: 42 50. 28. Abunasra H, Lewis S, Beggs L, Duffy J, Beggs D, Morgan E. Predictors of operative death after oesophagectomy for carcinoma. Br J Surg 2005;92:1029 33. 29. Omloo JM, Lagarde SM, Hulscher JB, et al. Extended transthoracic resection compared with limited transhiatal resection for adenocarcinoma of the mid/distal esophagus: fiveyear survival of a randomized clinical trial. Ann Surg 2007; 246:992 1001. 30. Lagarde SM, Reitsma JB, Maris AK, et al. Preoperative prediction of the occurrence and severity of complications after esophagectomy for cancer with use of a nomogram. Ann Thorac Surg 2008;85:1938 45.