- Email: [email protected]
Mortality Increases for Octogenarians Undergoing Esophagogastrectomy for Esophageal Cancer
Alexander H. Moskovitz, MD, Nabil P. Rizk, MD, Ennapadam Venkatraman, PhD, Manjit S. Bains, MD, Raja M. Flores, MD, Bernard J. H. Park, MD, and Valerie W. Rusch, MD Thoracic Service, Department of Surgery, and Biostatistics Service, Department of Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, New York
Background. As the general population ages, it becomes increasingly important to understand the potential contribution of chronologic age to mortality after esophagectomy. Because this risk is poorly defined, we sought to determine whether extreme age (>80 years) is an independent risk factor after esophagectomy. Methods. We analyzed a prospectively maintained, single-institution database of 858 consecutive patients who underwent esophagectomy between January 1996 and May 2005. Data evaluated included patient demographics, medical comorbidity, types of resections performed, length of stay, postoperative adverse events, and overall survival. We used univariate, multivariate, and Kaplan–Meier analysis to determine the influence of age on postoperative morbidity, in-hospital survival, and overall survival. Results. Of 858 patients, 31 (10 female, 21 male) were older than 80 years of age. Preliminary analysis indicated that patients younger than 50 years (n ⴝ 107) had significantly fewer comorbidities; these were excluded from
the analysis. In the remaining 751 patients, the age older than 80 cohort was compared with patients aged 50 to 79. Patients aged 50 to 79 were grouped because of similar characteristics (length of stay, hospital death). There were no significant differences in comorbidities, types of resections, or postoperative complication type or severity between the two groups. Postoperative death, length of stay, and survival, however, were significantly worse in patients older than 80. In a logistic regression model controlling for comorbidity, age older than 80 was significantly associated with increased perioperative mortality (hazard-ratio, 3.9; p < 0.01). Conclusions. Patients older than 80 years have increased mortality risk after esophagectomy, independent of comorbidity. Octogenarian status should be a consideration in the management of these patients.
T
increased medical comorbidity and diminished physiologic reserve, leading to increased perioperative risk. Despite these physiologic changes, an age-associated increased operative morbidity and mortality has not been universally observed in the care of older surgical patients. Studies of older surgical patients undergoing pulmonary resection for cancer [6, 7], cardiac surgery [8], and colorectal resection [9, 10] have shown that thoracic and abdominal surgery can be performed safely in very old patients. The most commonly accepted explanation for these findings is that coexisting disease has more impact on perioperative morbidity and mortality than age alone [11]. Thus, if patient comorbidity is of paramount importance, then good surgical outcomes in the care of the older patient are based on careful patient selection. The impact of old age on outcomes after esophagectomy is controversial. Esophagectomy is a major operation, with significant associated mortality and morbidity. In a review of the largest cohort of esophagectomy patients analyzed to date, a mortality of 10% and morbidity of 50% was noted after esophagectomy in 1,777 patients who underwent surgery at 109 centers [12]. In this review, increasing age was associated with a very small increase (odds ratio, 1.01) in 30-day mortality.
he US population is aging. By the year 2030, the US population of individuals aged 65 and older is expected to double to more than 71 million, and the group of individuals aged 85 and older is expected to more than double to 9.6 million in 2030 from 4.2 million in 2000 [1]. Once an individual reaches the age of 80 in the United States, that person can reasonably expect to live another 8 years [2]. With this increasing age there is increasing incidence of cancer diagnosis [3] and an increasing number of older patients considered for major operations. In the older cancer patient the life-limiting event may be their cancer and not their remaining lifespan, and this has prompted the consideration of surgery in progressively older cancer patient populations. The widespread recognition of this demographic and surgical trend is already documented in the surgical literature [4, 5]. As candidates for surgery age, the operative risks may increase. Older age is often considered a surrogate for Accepted for publication June 20, 2006. Presented at the Forty-second Annual Meeting of The Society of Thoracic Surgeons, Chicago, IL, Jan 30 –Feb 1, 2006. Address correspondence to Dr Rizk, Thoracic Service, Department of Surgery, Memorial Sloan-Kettering Cancer Center, 1275 York Ave, Suite C883, New York, NY 10021; e-mail: [email protected].
© 2006 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2006;82:2031– 6) © 2006 by The Society of Thoracic Surgeons
0003-4975/06/$32.00 doi:10.1016/j.athoracsur.2006.06.053
GENERAL THORACIC
Mortality Increases for Octogenarians Undergoing Esophagogastrectomy for Esophageal Cancer
GENERAL THORACIC
2032
MOSKOVITZ ET AL ESOPHAGECTOMY IN OCTOGENARIANS
Others have noted significantly increased morbidity and mortality in esophagectomy patients older than the age of 70 [13, 14], but more recent investigation has found operative risk in septuagenarians to be similar to a younger cohort [15]. From these mixed data the risks of old age remain elusive. Additionally, as the population ages the definition of very old age may change. This has led some to conclude that esophageal cancer resection in the octogenarian offers good quality of life and acceptable longevity in selected patients [16, 17]. We recognize that patient selection is important, especially as age and comorbidity increase. We sought to quantify the age-related risk of esophagectomy for esophageal cancer in our population of patients and to determine whether extreme age (older than 80 years of age) is an independent risk factor after esophagectomy. A better understanding of the age-related risks of esophagectomy in octogenarians can help surgeons to educate their esophageal cancer patients and make informed therapeutic decisions. This will likely increase in importance as surgeons are increasingly involved in the care of older patients with esophageal cancer.
Patients and Methods We analyzed a prospectively maintained, single-institution database of 858 consecutive patients who underwent esophagectomy between January 1996 and May 2005. January 1996 was the initiation of an institutional electronic medical record, which facilitated uniform retrieval of treatment data. The composition and use of this database has been described previously [18]. All patients underwent esophagogastrectomy with curative intent for a diagnosis of malignancy in the esophagus or gastroesophageal junction. Preoperatively, cardiopulmonary fitness was consistently evaluated by noninvasive stress testing and pulmonary function testing. Cardiology consultation was obtained for patients who had a history of ischemic or valvular heart disease or whose stress test was abnormal. Information collected included demographics (age, sex), comorbid illnesses, types of procedures performed, pathologic tumor stage, operative details, postoperative adverse events, and disease-specific survival. Complications were defined and diagnosed as previously described [18]. Complications were recorded in the medical record (Disease Management System Software; MSKCC & CTP Inc, New York, NY) with a numerical grade (1 to 5) indicating clinical consequence and therapy required. This grading system is analogous to Common Terminology Criteria for Adverse Events guidelines [19]; grade 1 (mildest) complications have minimal clinical consequence and require minimal therapy whereas grade 5 complications are associated with patient death. We defined perioperative mortality as any death in the hospital or within 60 days of operation. Patients were excluded from the survival analysis if no long-term survival data could be obtained. This study required a waiver of consent and was performed with approval from the Memorial SloanKettering Institutional Review Board. The procedure performed was assigned according to the surgeon’s operative summary. Ivor Lewis, McKeown, tho-
Ann Thorac Surg 2006;82:2031– 6
racoabdominal, transhiatal, and transabdominal procedures were all performed at the surgeon’s discretion [18]. Procedures were grouped according to the presence of a thoracotomy. Gastric conduit was used as the neoesophagus unless the patient had undergone a prior gastric resection. Data analysis was performed using R statistical package (free software available through www.R-project.org). Age effects were examined with age as a continuous variable and by age group: younger than 50, 50 to 59, 60 to 69, 70 to 79, and older than 80. Local regression methods were used to study the comorbidity, complication, and mortality rates as a function of age. Wilcoxon rank-sum test was used to test for the association between age and in-hospital mortality. Survival was calculated from the date of the operation. Survival distributions were estimated using Kaplan–Meier curves and compared across groups using the log rank test. Odds ratio and hazards ratio were estimated using logistic regression and proportional hazards models, respectively.
Results Eight hundred fifty-eight esophagectomy patients were in the database. For preliminary analysis, these patients were divided into groups by age in years: younger than 50, 50 to 59, 60 to 69, 70 to 79, and older than 80. The characteristics of each cohort are shown in Table 1. Overall, older patients had more comorbidity, less induction therapy, longer length of hospital stay, and worse postoperative mortality. The proportion of female patients, the proportion of tumors located at the gastroesophageal junction, distribution of American Joint Committee on Cancer pathologic stage, and frequency of resection completeness (R0) were similar across all age groupings. Preliminary analysis indicated that patients younger than 50 (n ⫽ 107) had fewer comorbidities (p ⬍ 0.001), less mortality (p ⫽ 0.02), and shorter length of stay (p ⬍ 0.001) than patients age 50 and older. The younger than 50 group was then excluded from further analysis and comparison to avoid introducing bias related to the reduced rate of comorbidity and improved outcomes in this cohort of young patients. We then excluded the older than 80 cohort to determine differences among the three middle cohorts: 50 to 59, 60 to 69, and 70 to 79. We found no difference in mortality (p ⫽ 0.54) or length of stay (p ⫽ 0.45) between these three age cohorts and analyzed them as a group for the remainder the study. We then analyzed postoperative mortality rate by age as a continuous variable. From this analysis we determined that all octogenarian deaths occurred in the hospital (range, 0 to 50 days). Furthermore, we identified a continuum of increasing mortality with advancing age, with the most profoundly elevated mortality rate in our octogenarian patients (Fig 1). In the 751 patients age 50 or older, we evaluated patient characteristics and comorbidity. We compared the older than 80 cohort (n ⫽ 31, range, 80.0 to 88.6 years) to the patients aged 50 to 79. The type of operation and incidence of postoperative complication in the older than 80 and the 50 to 79 cohorts are compared in Table 2. Thoracotomy was used less often in the older than 80
MOSKOVITZ ET AL ESOPHAGECTOMY IN OCTOGENARIANS
2033
Table 1. Characteristics by Age Group Age Groups Characteristic No. of patients No. female Pulmonary comorbidity Cardiac comorbidity Diabetes Induction therapy Location (GE junction)a Median LOS (days, 95% CI) Perioperative mortality a
⬍50
50–59
60–70
70–79
⬎80
107 (12.5%) 79 (26.2%) 3 (2.8%)
228 (26.6%) 51 (22.4%) 21 (9.2%)
285 (33.2%) 76 (26.7%) 37 (13.0%)
207 (24.1%) 56 (27.1%) 28 (13.5%)
31 (3.6%) 10 (33.3%) 6 (19.4%)
2 (1.9%)
20 (8.8%)
48 (16.8%)
56 (27.1%)
8 (25.8%)
4 (3.7%) 73 (68.2%)
24 (10.5%) 144 (63.2%)
47 (16.5%) 141 (49.5%)
38 (18.4%) 80 (38.7%)
7 (22.6%) 6 (19.4%)
97 (90.7%)
199 (87.3%)
241 (84.6%)
177 (85.5%)
28 (90.3%)
17.6 (15.8–19.4)
16.4 (14.8–18.0)
17.9 (16.0–19.8)
12.6 (10.0–15.3) 2 (1.9%)
11 (4.8%)
16 (5.6%)
15 (7.3%)
26.0 (21.1–30.8) 6 (19.4%)
Includes Seiwert 1, 2, & 3 (see Stein and associates [20]).
Perioperative mortality was defined as death in the hospital or within 60 days of operation. CI ⫽ confidence interval;
GE ⫽ gastroesophageal;
LOS ⫽ hospital length of stay.
cohort, but this did not reach statistical significance (p ⫽ 0.07). Review of the medical record indicates that avoidance of thoracotomy was planned in some octogenarian patients to minimize anesthesia time and the morbidity of thoracotomy. Operative blood loss was available and recorded for 87% of patients in the database and was not significantly different between age groups. Pulmonary, infectious, and anastomotic leak complications occurred with similar frequency when analyzing age by cohort (older than 80 versus 50 to 79) and as a continuous variable. Although not apparent by our age cohort comparison, there was an overall trend of increasing atrial fibrillation with advancing age (p ⫽ 0.001). This atrial fibrillation accounted for all of an observed increase in cardiac complications with advancing age; no patient older than 80 had a cardiac complication other than medically managed atrial fibrillation. To better understand the nature of the postoperative morbidity in our older than 80 population we reexamined the medical records of our older than 80 cohort. All octogenarian patients had noninvasive cardiac stress testing. Pulmonary function test records were available for 71% (22 of 31) of the older than 80 group. In our octogenarians, 6 patients had induction therapy, but this did not influence in-hospital mortality in a separate multivariate analysis (p ⫽ 0.84). In our octogenarians, 5 of the 6 deaths occurred in women, 3 were related to respiratory complications, 2 were related to anastomotic leak, and 1 was related to intraoperative hemorrhage with hypotension and cardiac arrest. (Review of the medical record revealed that the hemorrhage was not quantified for this patient, but was considered massive and the cause of death.) Three recurrent nerve injuries were treated with medialization. The remaining complications were minor: infection (urinary tract, superficial
wound, unexplained fever; n ⫽ 10), atrial fibrillation (n ⫽ 8), pneumonia or pneumonitis (n ⫽ 6), small contained anastomotic leak seen on esophagram (n ⫽ 5), and chylothorax requiring chest tube reinsertion for resolution (n ⫽ 1). In patients aged 80 and older, we found that postoperative death and length of stay were significantly worse. To confirm the age-related mortality finding, we performed a multivariate analysis using a logistic regression model to control for comorbidity. In this analysis, age older than 80 was significantly associated with increased postoperative mortality (hazard ratio, 3.9; p ⬍ 0.01; 95% confidence interval, 1.5 to 10.6). To determine longer-term outcomes we then evaluated
Fig 1. Local linear logistic regression analysis of mortality rate by age as a continuous variable for all in-hospital deaths and any death within 2 months of operation. All deaths within 2 months of operation for patients age 75 and older occurred in the hospital. Mortality increases as a continuum, with the highest rate in patients older than 80 years of age. In this analysis, the estimated mortality rate for an 80year-old is 12% (95% confidence interval, 6.7 to 21.2).
GENERAL THORACIC
Ann Thorac Surg 2006;82:2031– 6
GENERAL THORACIC
2034
MOSKOVITZ ET AL ESOPHAGECTOMY IN OCTOGENARIANS
Ann Thorac Surg 2006;82:2031– 6
Table 2. Operation Performed and Complications
Age Group (y) 50–79 Older than 80
Thoracotomy
Operative Blood Loss (mean, 95% CI)
550/720 (76.5%) 19/31 (61.3%) p ⫽ 0.07
482 mL (453–511) 475 mL (335–615) p ⫽ 0.92
Pulmonary Complication Grade 2–3a
Grade 4–5
161/720 34/720 (22.4%) (4.7%) 6/31 3/31 (19.4%) (9.7%) p ⫽ 0.49
Cardiac Complication Grade 1–2
Grade 3–5
146/720 6/720 (20.3%) (0.8%) 8/31 0/31 (25.8%) (0.0%) p ⫽ 0.91
Infection Grade 1–2
Grade 3–5
149/720 29/720 (20.7%) (4.0%) 10/31 0/31 (32.3%) (0.0%) p ⫽ 0.41
Anastomotic Leak Grade 1–2
Grade 3–5
68/720 63/720 (9.4%) (8.8%) 5/31 2/31 (16.1%) (6.5%) p ⫽ 0.43
a Complication grading is based on institutional database (Disease Management System Software; MSKCC & CTP Inc) that is analogous to Common Terminology Criteria for Adverse Events guidelines [19]. Complications are graded based on clinical consequence and therapy required: grade 1, oral medication or bedside care; grade 2, intravenous medical therapy or bedside procedure; grade 3, operative intervention required (eg, tracheostomy); grade 4, chronic disability or deficit; grade 5, patient death associated with this event.
our survival data, excluding patients with a diagnosis other than adenocarcinoma or squamous cancer of the esophagus. We examined age as a continuous variable and by age cohort. Although the age effect is relatively constant until patients reach an age in the mid-70s, older than 80 years of age the diminished survival is pronounced. Overall survival was worse for our older than 80 cohort (n ⫽ 31, median, 16.8 months) than for younger patients. We found that patients aged 75 to 80 (n ⫽ 76) also had slightly worse overall survival (median, 29.1 months) than younger patients. Our youngest patients aged 50 to 75 (n ⫽ 637) had the best overall survival (median, 46.7 months). The difference among these three groups was significant (p ⫽ 0.01). Similar to our findings of a continuum of increasing mortality in our oldest patients, the diminished survival was progressive with advancing age. There is a significant decrease in overall survival in patients older than 75 years of age, but the shortest survival was for our octogenarian patients (Fig 2). Disease-free survival was also shorter in the older than 80 cohort than in the younger patients (data not shown), but reliable disease recurrence information was not available in this retrospective analysis. Our octoge-
Fig 2. Overall survival (calculated from the date of operation) for patients who had either esophageal adenocarcinoma or squamous cell carcinoma. Patients aged 80 years or older had a significantly worse survival. The number of patients at risk in each age group is shown at the bottom margin.
narian patients did not derive the same survival benefit from esophagectomy as younger patients, despite having a similar spectrum of tumor stage and frequency of complete (R0) resection (data not shown).
Comment At our institution, esophagectomy in patients older than 80 years of age is associated with more than a threefold increased risk of in-hospital mortality when compared with younger patients. We found this age-related risk in a highly selected group of patients who were offered esophagectomy after careful assessment of comorbidities and cardiopulmonary fitness. The longer length of stay and worse in-hospital mortality suggest that octogenarian status alone should be a consideration in the therapeutic management of these patients. Although the total number of octogenarians in this study is small (n ⫽ 31), our cohort size is similar to the largest previously studied octogenarian cohorts undergoing esophagectomy [16, 21]. Adam and colleagues [16] similarly investigated outcomes in a retrospective review of octogenarian esophagectomy patients at a single institution. In their study, octogenarian in-hospital mortality was 16% (5 of 31 patients). Twelve of their patients experienced moderate to severe complications, and of the 5 deaths, 2 were related to pneumonia and 2 were related to anastomotic leak with sepsis. The strong influence of pulmonary and anastomotic complications on their observed postoperative mortality is similar to our findings. Alexiou and colleagues [21] analyzed a cohort of 36 octogenarian esophagectomy patients at their institution in whom the incidence of preoperative comorbidity was lower than the younger patient cohorts. In this highly selected group, their postoperative mortality of 5.6% (2 of 36 patients) and hospital length of stay (median, 13.5 days) were similar to their younger patients. Although respiratory complications were the most common postoperative adverse event, anastomotic leak complications resulted in the two octogenarian deaths. Taken together with our findings, these results confirm both the importance of avoiding technical complications [18], such as anastomotic leak, and the need for aggressive postoperative respiratory care to minimize perioperative risk. In
contrast to others, our results suggest that careful patient selection, meticulous operation, and aggressive perioperative care at a high-volume center may not be enough to compensate for the increased age-related risk of esophagectomy in octogenarians. We have previously noted that our morbidity and mortality rates are comparable to other high-volume institutions [18]. National benchmark mortality for esophagectomy in those age 65 years and older has been reported to range from 8% to 23% [22]. In this study, we found septuagenarian mortality (6.3%) to be similar to patients aged 50 to 69 (4.3%). The independent agerelated risk of esophagectomy is a continuum with the highest risk seen in the oldest patients. At our center, the highest mortality rate is only approached in the octogenarian patient. An 80-year-old patient at our center has a 12% postoperative mortality risk. In our octogenarian patients we observed a 19.4% mortality rate, compared with an overall mortality of 5.1%. Although atrial fibrillation occurred more frequently in our older patients, major cardiac events were not a cause of postoperative mortality. Notably, the observed postoperative rate of pneumonia, infection, and anastomotic leak in octogenarians was similar to that of younger patients. Thus, despite a similar rate of postoperative complication, the octogenarian patients spent longer in the hospital and were less likely to leave the hospital alive. This suggests octogenarian patients may have had an inability to survive complications that are survivable as a younger patient, with more physiologic capacity for recovery. Our increased octogenarian mortality may reflect a diminished physiologic reserve in the octogenarian patient. As a single-institution study, our analysis is potentially subject to complication-reporting bias. Memorial Sloan-Kettering Cancer Center is a high-volume, National Cancer Institute– designated, comprehensive cancer center. The relatively uniform perioperative care, anesthesia care, follow-up evaluation, and medical record documentation diminish the potential for confounding health system factors affecting our observed outcomes. Additionally, our ability to access a uniform medical record and follow-up documentation minimizes the potential for underreporting of complications. The nature and severity of complications are discussed at a weekly thoracic surgical meeting to ensure correct entry into a database of postoperative adverse events and are reconfirmed through chart review. Our institutional system of complication reporting is accessible and redundant, and minimizes the risk of underreporting. An interesting finding in our analysis was the disproportionate number of female deaths in the older than 80 cohort. Five of the 6 octogenarian deaths occurred in women. Overall, there was no difference in the proportion of female patients in the older than 80 and 50 to 79 cohorts. In the older than 80 cohort, our patient numbers were too small to make valid conclusions about the mortality risk of being female. However, in our 50 to 79 cohort, female patients did not have a statistically different in-hospital mortality than
MOSKOVITZ ET AL ESOPHAGECTOMY IN OCTOGENARIANS
2035
our male patients (6% female versus 4.7% male; p ⫽ 0.48). It is unlikely that a patient’s sex would have an influence in the older than 80 cohort that is not present in a younger population, but this remains a topic for further investigation. Another finding of interest is the diminished survival in octogenarians who survive to hospital discharge. We found a shorter overall survival in our octogenarian cohort, despite similar rates of comorbidity, distribution of tumor stage, choice of operation, resection completeness, and postoperative complication. In contrast, Alexiou and colleagues [21] found no significant survival difference (median, 21 months; 95% confidence interval, 11.0 to 31.0) in their octogenarian patients compared with younger cohorts. This led them to conclude that esophagectomy can be safely offered to carefully selected patients with an expectation of survival benefit similar to younger patients. In our series, the reasons for diminished overall survival in octogenarians remain unclear. The possibility that octogenarians undergoing esophagectomy and surviving to hospital discharge do not derive the overall survival benefit seen in younger patients remains an area for further investigation. In summary, our data show that patients older than 80 years have age-related increased mortality risk after esophagectomy, independent of comorbidity. Octogenarian status should be a consideration in the management of these patients. Additional methods to improve the perioperative care of our highest-risk patients are currently under investigation, including initiating a postoperative pulmonary rehabilitation program. The authors wish to thank Kathleen Meneses for her excellent support in database management, and Janet Creason for her assistance in manuscript preparation.
References 1. U.S. Department of Health and Human Services, Administration on Aging. A profile of older Americans: 2004. Available at: www.aoa.gov. Accessed on January 11, 2006. 2. Arias E. United States Life Tables, 2002. National Vital Statistics Reports, 2004;53:1–39. 3. United States Cancer Statistics 2002. Available at: www. cdc.gov/cancer/npcr/uscs. Accessed on December 10, 2005. 4. Loran DB, Zweischenberger JB. Thoracic surgery in the elderly. J Am Coll Surg 2004:199 :773– 84. 5. Reiss R, Deutsch A, Nudelman I. Surgical problems in octogenarians: epidemiological analysis of 1,083 consecutive admissions. World J Surg 1992;16:1017–21. 6. Brock MV, Kim MP, Hooker CM, et al. Pulmonary resection in octogenarian with stage I nonsmall cell lung cancer: a 22-year experience. Ann Thorac Surg 2004;77:271–7. 7. Pagni S, McKelvey A, Riordan C, Federico JA, Ponn RB. Pulmonary resection for malignancy in the elderly: is age still a risk factor? Eur J Cardiothorac Surg 1998;14:40 –5. 8. Bacchetta MD, Ko W, Girardi LN, et al. Outcomes of cardiac surgery in nonagenarians: a10-year experience. Ann Thorac Surg 2003;75:1215–20. 9. Puig-La Calle J, Quayle J, Thaler HT, et al. Favorable short-term and long-term outcome after elective radical rectal cancer resection in patients 75 years of age or older. Dis Colon Rectum 2000;43:1704 –9.
GENERAL THORACIC
Ann Thorac Surg 2006;82:2031– 6
GENERAL THORACIC
2036
MOSKOVITZ ET AL ESOPHAGECTOMY IN OCTOGENARIANS
10. Spivak H, Vande Maele D, Friedman I, Nussbaum M. Colorectal surgery in octogenarians. J Am Coll Surg 1996;183:46–50. 11. Dunlop WE, Rosenblood L, Lawrason L, et al. Effects of age and severity of illness on outcome and length of stay in geriatric patients. Am J Surg 1993;165:577– 80. 12. Bailey SH, Bull DA, Harpole DH, et al. Outcomes after esophagectomy: a ten-year prospective cohort. Ann Thorac Surg 2003;75:217–22. 13. Griffin S, Desai J, Charlton M, Townsend E, Fountain SW. Factors influencing mortality and morbidity following oesophageal resection. Eur J Cardiothorac Surg 1989;3:419 –24. 14. Naunheim KS, Hanosh J, Zwischenberger J, et al. Esophagectomy in the septuagenarian. Ann Thorac Surg 1993;56:880 – 4. 15. Ellis FH Jr, 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. 16. Adam DJ, Craig SR, Sang CTM, Cameron EWJ, Walker WS. Esophagectomy for carcinoma in the octogenarian. Ann Thorac Surg 1996;61:190 – 4.
Ann Thorac Surg 2006;82:2031– 6
17. Kuwano D, Morita M, Baba K, et al. Surgical treatment of esophageal carcinoma in patients eighty years of age and older. J Surg Oncol 1993;52:36 –9. 18. 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. 19. Cancer Therapy Evaluation Program. Common terminology criteria for adverse events, version 3.0. National Cancer Institute, National Institutes of Health, Department of Health and Human Services, March 31, 2003. Available at: http://ctep.cancer.govAccessed on December 10, 2005. 20. Stein HJ, Feith M, Siewert JR. Cancer of the esophagogastric junction. Surg Oncol 2000;9:35– 41. 21. 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. 22. Birkmeyer JD, Siewers AE, Finlaysonn E, et al. Hospital volume and surgical mortality in the United States. N Engl J Med 2002;346:1128 –37.
DISCUSSION DR MARK J. KRASNA (Baltimore, MD): I really enjoyed your presentation. I have two quick comments and then a question. Someone is going to have to go to the dictionary later this afternoon and see if the octogenarian is defined as the eighth decade or in the 80s. I think some of these patients might be nonagenarians. The other comment is, at least for The Annals publications, you are going to need to report 90-day mortality as well as in-hospital mortality, so you should definitely include that in your analysis. Your results are very different from both Tom Rice’s at the Cleveland Clinic as well as the group at MD Anderson from last year, and I wonder if you could comment what the differences between your sets are. Last year the MD Anderson group showed excellent results in elderly patients with esophageal cancer, although they did not have as many patients in their 80s, but did not actually show a significant difference, including those patients who got neoadjuvant therapy. Tom Rice’s group at Cleveland Clinic about a year ago also showed data on elderly patients and found that age was an independent risk factor for death only when associated with thoracotomy. So my other specific question to you is, would you now consider still doing esophagectomies on these 80-year-olds and older but doing them as a nonthoracotomy transhiatal esophagectomy? Thank you. DR MOSKOVITZ: We are continuing to perform the standard operation at Memorial that is done in the majority of the patients, which is Ivor Lewis esophagectomy: thoracotomyassociated esophagectomy. We continue to apply careful selection criteria. We are using this data primarily to educate our patients in what we believe is an increased risk in our octogenarian population. DR KRASNA: Well, if that’s the case, then if there’s no difference in stage between the two groups and yet you are offering these patients a different treatment, which is surgery without induction therapy, as you showed us, why not offer them a different operation? I think it’s something to consider since there is a significantly increased mortality. DR MOSKOVITZ: Thank you.
DR STEPHEN G. SWISHER (Houston, TX): Could you tell me what your mortality was for your 80-year-olds who had gotten induction chemoradiation versus those who didn’t? DR MOSKOVITZ: There were 6 octogenarian patients who had induction chemoradiation and 2 of them are included as the 6 octogenarian deaths. DR KEMP H. KERNSTINE (Los Angeles, CA): You were talking about induction therapy. Did you have any of the patients that you included in the induction group that had “definitive chemoradiotherapy” and then you went on and tried to resect them, as a “salvage esophagectomy”? It’s my impression that that’s more common in the older patient, because the medical oncologists will say, “You are older and can’t have any surgery.” The patients are selected out by the medical oncologists to get chemoradiotherapy, then they recur a year later and then they are sent to us. Did you have any patients like that? The reason that’s important is because those patients tend to have a higher morbidity and mortality after their procedure, and that might influence your results. It may not be age alone. DR MOSKOVITZ: I can tell you that in our octogenarians, from the records that I reviewed, 3 of the 6 had our standard approach at Memorial, which includes 5,040 Gy of radiation combined with the chemotherapy, cisplatin-based chemo, and I’m not sure precisely what the other 3 had, so I can’t comment on that. DR JOE B. PUTNAM (Nashville, TN): Before we completely abandon surgery for esophageal cancer in this population, do you have any idea as to the results of the medical oncology/ radiation oncology therapy for these octogenarians and anticipated complications or mortality from that treatment and the subsequent survival at your institution? DR MOSKOVITZ: That’s an excellent question. It’s very difficult to ascertain in our institution, but I think that is the next step. The question when there is an increased mortality in the octogenarians undergoing operation is what is the risk to the best alternative therapy, and that may be as high or higher. So we’re taking our data and looking forward to think about how we can study this in a prospective fashion, and we are using the data to educate our patients, still trying to provide the best operative care that we can.
Background. As the general population ages, it becomes increasingly important to understand the potential contribution of chronologic age to mortality after esophagectomy. Because this risk is poorly defined, we sought to determine whether extreme age (>80 years) is an independent risk factor after esophagectomy. Methods. We analyzed a prospectively maintained, single-institution database of 858 consecutive patients who underwent esophagectomy between January 1996 and May 2005. Data evaluated included patient demographics, medical comorbidity, types of resections performed, length of stay, postoperative adverse events, and overall survival. We used univariate, multivariate, and Kaplan–Meier analysis to determine the influence of age on postoperative morbidity, in-hospital survival, and overall survival. Results. Of 858 patients, 31 (10 female, 21 male) were older than 80 years of age. Preliminary analysis indicated that patients younger than 50 years (n ⴝ 107) had significantly fewer comorbidities; these were excluded from
the analysis. In the remaining 751 patients, the age older than 80 cohort was compared with patients aged 50 to 79. Patients aged 50 to 79 were grouped because of similar characteristics (length of stay, hospital death). There were no significant differences in comorbidities, types of resections, or postoperative complication type or severity between the two groups. Postoperative death, length of stay, and survival, however, were significantly worse in patients older than 80. In a logistic regression model controlling for comorbidity, age older than 80 was significantly associated with increased perioperative mortality (hazard-ratio, 3.9; p < 0.01). Conclusions. Patients older than 80 years have increased mortality risk after esophagectomy, independent of comorbidity. Octogenarian status should be a consideration in the management of these patients.
T
increased medical comorbidity and diminished physiologic reserve, leading to increased perioperative risk. Despite these physiologic changes, an age-associated increased operative morbidity and mortality has not been universally observed in the care of older surgical patients. Studies of older surgical patients undergoing pulmonary resection for cancer [6, 7], cardiac surgery [8], and colorectal resection [9, 10] have shown that thoracic and abdominal surgery can be performed safely in very old patients. The most commonly accepted explanation for these findings is that coexisting disease has more impact on perioperative morbidity and mortality than age alone [11]. Thus, if patient comorbidity is of paramount importance, then good surgical outcomes in the care of the older patient are based on careful patient selection. The impact of old age on outcomes after esophagectomy is controversial. Esophagectomy is a major operation, with significant associated mortality and morbidity. In a review of the largest cohort of esophagectomy patients analyzed to date, a mortality of 10% and morbidity of 50% was noted after esophagectomy in 1,777 patients who underwent surgery at 109 centers [12]. In this review, increasing age was associated with a very small increase (odds ratio, 1.01) in 30-day mortality.
he US population is aging. By the year 2030, the US population of individuals aged 65 and older is expected to double to more than 71 million, and the group of individuals aged 85 and older is expected to more than double to 9.6 million in 2030 from 4.2 million in 2000 [1]. Once an individual reaches the age of 80 in the United States, that person can reasonably expect to live another 8 years [2]. With this increasing age there is increasing incidence of cancer diagnosis [3] and an increasing number of older patients considered for major operations. In the older cancer patient the life-limiting event may be their cancer and not their remaining lifespan, and this has prompted the consideration of surgery in progressively older cancer patient populations. The widespread recognition of this demographic and surgical trend is already documented in the surgical literature [4, 5]. As candidates for surgery age, the operative risks may increase. Older age is often considered a surrogate for Accepted for publication June 20, 2006. Presented at the Forty-second Annual Meeting of The Society of Thoracic Surgeons, Chicago, IL, Jan 30 –Feb 1, 2006. Address correspondence to Dr Rizk, Thoracic Service, Department of Surgery, Memorial Sloan-Kettering Cancer Center, 1275 York Ave, Suite C883, New York, NY 10021; e-mail: [email protected].
© 2006 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2006;82:2031– 6) © 2006 by The Society of Thoracic Surgeons
0003-4975/06/$32.00 doi:10.1016/j.athoracsur.2006.06.053
GENERAL THORACIC
Mortality Increases for Octogenarians Undergoing Esophagogastrectomy for Esophageal Cancer
GENERAL THORACIC
2032
MOSKOVITZ ET AL ESOPHAGECTOMY IN OCTOGENARIANS
Others have noted significantly increased morbidity and mortality in esophagectomy patients older than the age of 70 [13, 14], but more recent investigation has found operative risk in septuagenarians to be similar to a younger cohort [15]. From these mixed data the risks of old age remain elusive. Additionally, as the population ages the definition of very old age may change. This has led some to conclude that esophageal cancer resection in the octogenarian offers good quality of life and acceptable longevity in selected patients [16, 17]. We recognize that patient selection is important, especially as age and comorbidity increase. We sought to quantify the age-related risk of esophagectomy for esophageal cancer in our population of patients and to determine whether extreme age (older than 80 years of age) is an independent risk factor after esophagectomy. A better understanding of the age-related risks of esophagectomy in octogenarians can help surgeons to educate their esophageal cancer patients and make informed therapeutic decisions. This will likely increase in importance as surgeons are increasingly involved in the care of older patients with esophageal cancer.
Patients and Methods We analyzed a prospectively maintained, single-institution database of 858 consecutive patients who underwent esophagectomy between January 1996 and May 2005. January 1996 was the initiation of an institutional electronic medical record, which facilitated uniform retrieval of treatment data. The composition and use of this database has been described previously [18]. All patients underwent esophagogastrectomy with curative intent for a diagnosis of malignancy in the esophagus or gastroesophageal junction. Preoperatively, cardiopulmonary fitness was consistently evaluated by noninvasive stress testing and pulmonary function testing. Cardiology consultation was obtained for patients who had a history of ischemic or valvular heart disease or whose stress test was abnormal. Information collected included demographics (age, sex), comorbid illnesses, types of procedures performed, pathologic tumor stage, operative details, postoperative adverse events, and disease-specific survival. Complications were defined and diagnosed as previously described [18]. Complications were recorded in the medical record (Disease Management System Software; MSKCC & CTP Inc, New York, NY) with a numerical grade (1 to 5) indicating clinical consequence and therapy required. This grading system is analogous to Common Terminology Criteria for Adverse Events guidelines [19]; grade 1 (mildest) complications have minimal clinical consequence and require minimal therapy whereas grade 5 complications are associated with patient death. We defined perioperative mortality as any death in the hospital or within 60 days of operation. Patients were excluded from the survival analysis if no long-term survival data could be obtained. This study required a waiver of consent and was performed with approval from the Memorial SloanKettering Institutional Review Board. The procedure performed was assigned according to the surgeon’s operative summary. Ivor Lewis, McKeown, tho-
Ann Thorac Surg 2006;82:2031– 6
racoabdominal, transhiatal, and transabdominal procedures were all performed at the surgeon’s discretion [18]. Procedures were grouped according to the presence of a thoracotomy. Gastric conduit was used as the neoesophagus unless the patient had undergone a prior gastric resection. Data analysis was performed using R statistical package (free software available through www.R-project.org). Age effects were examined with age as a continuous variable and by age group: younger than 50, 50 to 59, 60 to 69, 70 to 79, and older than 80. Local regression methods were used to study the comorbidity, complication, and mortality rates as a function of age. Wilcoxon rank-sum test was used to test for the association between age and in-hospital mortality. Survival was calculated from the date of the operation. Survival distributions were estimated using Kaplan–Meier curves and compared across groups using the log rank test. Odds ratio and hazards ratio were estimated using logistic regression and proportional hazards models, respectively.
Results Eight hundred fifty-eight esophagectomy patients were in the database. For preliminary analysis, these patients were divided into groups by age in years: younger than 50, 50 to 59, 60 to 69, 70 to 79, and older than 80. The characteristics of each cohort are shown in Table 1. Overall, older patients had more comorbidity, less induction therapy, longer length of hospital stay, and worse postoperative mortality. The proportion of female patients, the proportion of tumors located at the gastroesophageal junction, distribution of American Joint Committee on Cancer pathologic stage, and frequency of resection completeness (R0) were similar across all age groupings. Preliminary analysis indicated that patients younger than 50 (n ⫽ 107) had fewer comorbidities (p ⬍ 0.001), less mortality (p ⫽ 0.02), and shorter length of stay (p ⬍ 0.001) than patients age 50 and older. The younger than 50 group was then excluded from further analysis and comparison to avoid introducing bias related to the reduced rate of comorbidity and improved outcomes in this cohort of young patients. We then excluded the older than 80 cohort to determine differences among the three middle cohorts: 50 to 59, 60 to 69, and 70 to 79. We found no difference in mortality (p ⫽ 0.54) or length of stay (p ⫽ 0.45) between these three age cohorts and analyzed them as a group for the remainder the study. We then analyzed postoperative mortality rate by age as a continuous variable. From this analysis we determined that all octogenarian deaths occurred in the hospital (range, 0 to 50 days). Furthermore, we identified a continuum of increasing mortality with advancing age, with the most profoundly elevated mortality rate in our octogenarian patients (Fig 1). In the 751 patients age 50 or older, we evaluated patient characteristics and comorbidity. We compared the older than 80 cohort (n ⫽ 31, range, 80.0 to 88.6 years) to the patients aged 50 to 79. The type of operation and incidence of postoperative complication in the older than 80 and the 50 to 79 cohorts are compared in Table 2. Thoracotomy was used less often in the older than 80
MOSKOVITZ ET AL ESOPHAGECTOMY IN OCTOGENARIANS
2033
Table 1. Characteristics by Age Group Age Groups Characteristic No. of patients No. female Pulmonary comorbidity Cardiac comorbidity Diabetes Induction therapy Location (GE junction)a Median LOS (days, 95% CI) Perioperative mortality a
⬍50
50–59
60–70
70–79
⬎80
107 (12.5%) 79 (26.2%) 3 (2.8%)
228 (26.6%) 51 (22.4%) 21 (9.2%)
285 (33.2%) 76 (26.7%) 37 (13.0%)
207 (24.1%) 56 (27.1%) 28 (13.5%)
31 (3.6%) 10 (33.3%) 6 (19.4%)
2 (1.9%)
20 (8.8%)
48 (16.8%)
56 (27.1%)
8 (25.8%)
4 (3.7%) 73 (68.2%)
24 (10.5%) 144 (63.2%)
47 (16.5%) 141 (49.5%)
38 (18.4%) 80 (38.7%)
7 (22.6%) 6 (19.4%)
97 (90.7%)
199 (87.3%)
241 (84.6%)
177 (85.5%)
28 (90.3%)
17.6 (15.8–19.4)
16.4 (14.8–18.0)
17.9 (16.0–19.8)
12.6 (10.0–15.3) 2 (1.9%)
11 (4.8%)
16 (5.6%)
15 (7.3%)
26.0 (21.1–30.8) 6 (19.4%)
Includes Seiwert 1, 2, & 3 (see Stein and associates [20]).
Perioperative mortality was defined as death in the hospital or within 60 days of operation. CI ⫽ confidence interval;
GE ⫽ gastroesophageal;
LOS ⫽ hospital length of stay.
cohort, but this did not reach statistical significance (p ⫽ 0.07). Review of the medical record indicates that avoidance of thoracotomy was planned in some octogenarian patients to minimize anesthesia time and the morbidity of thoracotomy. Operative blood loss was available and recorded for 87% of patients in the database and was not significantly different between age groups. Pulmonary, infectious, and anastomotic leak complications occurred with similar frequency when analyzing age by cohort (older than 80 versus 50 to 79) and as a continuous variable. Although not apparent by our age cohort comparison, there was an overall trend of increasing atrial fibrillation with advancing age (p ⫽ 0.001). This atrial fibrillation accounted for all of an observed increase in cardiac complications with advancing age; no patient older than 80 had a cardiac complication other than medically managed atrial fibrillation. To better understand the nature of the postoperative morbidity in our older than 80 population we reexamined the medical records of our older than 80 cohort. All octogenarian patients had noninvasive cardiac stress testing. Pulmonary function test records were available for 71% (22 of 31) of the older than 80 group. In our octogenarians, 6 patients had induction therapy, but this did not influence in-hospital mortality in a separate multivariate analysis (p ⫽ 0.84). In our octogenarians, 5 of the 6 deaths occurred in women, 3 were related to respiratory complications, 2 were related to anastomotic leak, and 1 was related to intraoperative hemorrhage with hypotension and cardiac arrest. (Review of the medical record revealed that the hemorrhage was not quantified for this patient, but was considered massive and the cause of death.) Three recurrent nerve injuries were treated with medialization. The remaining complications were minor: infection (urinary tract, superficial
wound, unexplained fever; n ⫽ 10), atrial fibrillation (n ⫽ 8), pneumonia or pneumonitis (n ⫽ 6), small contained anastomotic leak seen on esophagram (n ⫽ 5), and chylothorax requiring chest tube reinsertion for resolution (n ⫽ 1). In patients aged 80 and older, we found that postoperative death and length of stay were significantly worse. To confirm the age-related mortality finding, we performed a multivariate analysis using a logistic regression model to control for comorbidity. In this analysis, age older than 80 was significantly associated with increased postoperative mortality (hazard ratio, 3.9; p ⬍ 0.01; 95% confidence interval, 1.5 to 10.6). To determine longer-term outcomes we then evaluated
Fig 1. Local linear logistic regression analysis of mortality rate by age as a continuous variable for all in-hospital deaths and any death within 2 months of operation. All deaths within 2 months of operation for patients age 75 and older occurred in the hospital. Mortality increases as a continuum, with the highest rate in patients older than 80 years of age. In this analysis, the estimated mortality rate for an 80year-old is 12% (95% confidence interval, 6.7 to 21.2).
GENERAL THORACIC
Ann Thorac Surg 2006;82:2031– 6
GENERAL THORACIC
2034
MOSKOVITZ ET AL ESOPHAGECTOMY IN OCTOGENARIANS
Ann Thorac Surg 2006;82:2031– 6
Table 2. Operation Performed and Complications
Age Group (y) 50–79 Older than 80
Thoracotomy
Operative Blood Loss (mean, 95% CI)
550/720 (76.5%) 19/31 (61.3%) p ⫽ 0.07
482 mL (453–511) 475 mL (335–615) p ⫽ 0.92
Pulmonary Complication Grade 2–3a
Grade 4–5
161/720 34/720 (22.4%) (4.7%) 6/31 3/31 (19.4%) (9.7%) p ⫽ 0.49
Cardiac Complication Grade 1–2
Grade 3–5
146/720 6/720 (20.3%) (0.8%) 8/31 0/31 (25.8%) (0.0%) p ⫽ 0.91
Infection Grade 1–2
Grade 3–5
149/720 29/720 (20.7%) (4.0%) 10/31 0/31 (32.3%) (0.0%) p ⫽ 0.41
Anastomotic Leak Grade 1–2
Grade 3–5
68/720 63/720 (9.4%) (8.8%) 5/31 2/31 (16.1%) (6.5%) p ⫽ 0.43
a Complication grading is based on institutional database (Disease Management System Software; MSKCC & CTP Inc) that is analogous to Common Terminology Criteria for Adverse Events guidelines [19]. Complications are graded based on clinical consequence and therapy required: grade 1, oral medication or bedside care; grade 2, intravenous medical therapy or bedside procedure; grade 3, operative intervention required (eg, tracheostomy); grade 4, chronic disability or deficit; grade 5, patient death associated with this event.
our survival data, excluding patients with a diagnosis other than adenocarcinoma or squamous cancer of the esophagus. We examined age as a continuous variable and by age cohort. Although the age effect is relatively constant until patients reach an age in the mid-70s, older than 80 years of age the diminished survival is pronounced. Overall survival was worse for our older than 80 cohort (n ⫽ 31, median, 16.8 months) than for younger patients. We found that patients aged 75 to 80 (n ⫽ 76) also had slightly worse overall survival (median, 29.1 months) than younger patients. Our youngest patients aged 50 to 75 (n ⫽ 637) had the best overall survival (median, 46.7 months). The difference among these three groups was significant (p ⫽ 0.01). Similar to our findings of a continuum of increasing mortality in our oldest patients, the diminished survival was progressive with advancing age. There is a significant decrease in overall survival in patients older than 75 years of age, but the shortest survival was for our octogenarian patients (Fig 2). Disease-free survival was also shorter in the older than 80 cohort than in the younger patients (data not shown), but reliable disease recurrence information was not available in this retrospective analysis. Our octoge-
Fig 2. Overall survival (calculated from the date of operation) for patients who had either esophageal adenocarcinoma or squamous cell carcinoma. Patients aged 80 years or older had a significantly worse survival. The number of patients at risk in each age group is shown at the bottom margin.
narian patients did not derive the same survival benefit from esophagectomy as younger patients, despite having a similar spectrum of tumor stage and frequency of complete (R0) resection (data not shown).
Comment At our institution, esophagectomy in patients older than 80 years of age is associated with more than a threefold increased risk of in-hospital mortality when compared with younger patients. We found this age-related risk in a highly selected group of patients who were offered esophagectomy after careful assessment of comorbidities and cardiopulmonary fitness. The longer length of stay and worse in-hospital mortality suggest that octogenarian status alone should be a consideration in the therapeutic management of these patients. Although the total number of octogenarians in this study is small (n ⫽ 31), our cohort size is similar to the largest previously studied octogenarian cohorts undergoing esophagectomy [16, 21]. Adam and colleagues [16] similarly investigated outcomes in a retrospective review of octogenarian esophagectomy patients at a single institution. In their study, octogenarian in-hospital mortality was 16% (5 of 31 patients). Twelve of their patients experienced moderate to severe complications, and of the 5 deaths, 2 were related to pneumonia and 2 were related to anastomotic leak with sepsis. The strong influence of pulmonary and anastomotic complications on their observed postoperative mortality is similar to our findings. Alexiou and colleagues [21] analyzed a cohort of 36 octogenarian esophagectomy patients at their institution in whom the incidence of preoperative comorbidity was lower than the younger patient cohorts. In this highly selected group, their postoperative mortality of 5.6% (2 of 36 patients) and hospital length of stay (median, 13.5 days) were similar to their younger patients. Although respiratory complications were the most common postoperative adverse event, anastomotic leak complications resulted in the two octogenarian deaths. Taken together with our findings, these results confirm both the importance of avoiding technical complications [18], such as anastomotic leak, and the need for aggressive postoperative respiratory care to minimize perioperative risk. In
contrast to others, our results suggest that careful patient selection, meticulous operation, and aggressive perioperative care at a high-volume center may not be enough to compensate for the increased age-related risk of esophagectomy in octogenarians. We have previously noted that our morbidity and mortality rates are comparable to other high-volume institutions [18]. National benchmark mortality for esophagectomy in those age 65 years and older has been reported to range from 8% to 23% [22]. In this study, we found septuagenarian mortality (6.3%) to be similar to patients aged 50 to 69 (4.3%). The independent agerelated risk of esophagectomy is a continuum with the highest risk seen in the oldest patients. At our center, the highest mortality rate is only approached in the octogenarian patient. An 80-year-old patient at our center has a 12% postoperative mortality risk. In our octogenarian patients we observed a 19.4% mortality rate, compared with an overall mortality of 5.1%. Although atrial fibrillation occurred more frequently in our older patients, major cardiac events were not a cause of postoperative mortality. Notably, the observed postoperative rate of pneumonia, infection, and anastomotic leak in octogenarians was similar to that of younger patients. Thus, despite a similar rate of postoperative complication, the octogenarian patients spent longer in the hospital and were less likely to leave the hospital alive. This suggests octogenarian patients may have had an inability to survive complications that are survivable as a younger patient, with more physiologic capacity for recovery. Our increased octogenarian mortality may reflect a diminished physiologic reserve in the octogenarian patient. As a single-institution study, our analysis is potentially subject to complication-reporting bias. Memorial Sloan-Kettering Cancer Center is a high-volume, National Cancer Institute– designated, comprehensive cancer center. The relatively uniform perioperative care, anesthesia care, follow-up evaluation, and medical record documentation diminish the potential for confounding health system factors affecting our observed outcomes. Additionally, our ability to access a uniform medical record and follow-up documentation minimizes the potential for underreporting of complications. The nature and severity of complications are discussed at a weekly thoracic surgical meeting to ensure correct entry into a database of postoperative adverse events and are reconfirmed through chart review. Our institutional system of complication reporting is accessible and redundant, and minimizes the risk of underreporting. An interesting finding in our analysis was the disproportionate number of female deaths in the older than 80 cohort. Five of the 6 octogenarian deaths occurred in women. Overall, there was no difference in the proportion of female patients in the older than 80 and 50 to 79 cohorts. In the older than 80 cohort, our patient numbers were too small to make valid conclusions about the mortality risk of being female. However, in our 50 to 79 cohort, female patients did not have a statistically different in-hospital mortality than
MOSKOVITZ ET AL ESOPHAGECTOMY IN OCTOGENARIANS
2035
our male patients (6% female versus 4.7% male; p ⫽ 0.48). It is unlikely that a patient’s sex would have an influence in the older than 80 cohort that is not present in a younger population, but this remains a topic for further investigation. Another finding of interest is the diminished survival in octogenarians who survive to hospital discharge. We found a shorter overall survival in our octogenarian cohort, despite similar rates of comorbidity, distribution of tumor stage, choice of operation, resection completeness, and postoperative complication. In contrast, Alexiou and colleagues [21] found no significant survival difference (median, 21 months; 95% confidence interval, 11.0 to 31.0) in their octogenarian patients compared with younger cohorts. This led them to conclude that esophagectomy can be safely offered to carefully selected patients with an expectation of survival benefit similar to younger patients. In our series, the reasons for diminished overall survival in octogenarians remain unclear. The possibility that octogenarians undergoing esophagectomy and surviving to hospital discharge do not derive the overall survival benefit seen in younger patients remains an area for further investigation. In summary, our data show that patients older than 80 years have age-related increased mortality risk after esophagectomy, independent of comorbidity. Octogenarian status should be a consideration in the management of these patients. Additional methods to improve the perioperative care of our highest-risk patients are currently under investigation, including initiating a postoperative pulmonary rehabilitation program. The authors wish to thank Kathleen Meneses for her excellent support in database management, and Janet Creason for her assistance in manuscript preparation.
References 1. U.S. Department of Health and Human Services, Administration on Aging. A profile of older Americans: 2004. Available at: www.aoa.gov. Accessed on January 11, 2006. 2. Arias E. United States Life Tables, 2002. National Vital Statistics Reports, 2004;53:1–39. 3. United States Cancer Statistics 2002. Available at: www. cdc.gov/cancer/npcr/uscs. Accessed on December 10, 2005. 4. Loran DB, Zweischenberger JB. Thoracic surgery in the elderly. J Am Coll Surg 2004:199 :773– 84. 5. Reiss R, Deutsch A, Nudelman I. Surgical problems in octogenarians: epidemiological analysis of 1,083 consecutive admissions. World J Surg 1992;16:1017–21. 6. Brock MV, Kim MP, Hooker CM, et al. Pulmonary resection in octogenarian with stage I nonsmall cell lung cancer: a 22-year experience. Ann Thorac Surg 2004;77:271–7. 7. Pagni S, McKelvey A, Riordan C, Federico JA, Ponn RB. Pulmonary resection for malignancy in the elderly: is age still a risk factor? Eur J Cardiothorac Surg 1998;14:40 –5. 8. Bacchetta MD, Ko W, Girardi LN, et al. Outcomes of cardiac surgery in nonagenarians: a10-year experience. Ann Thorac Surg 2003;75:1215–20. 9. Puig-La Calle J, Quayle J, Thaler HT, et al. Favorable short-term and long-term outcome after elective radical rectal cancer resection in patients 75 years of age or older. Dis Colon Rectum 2000;43:1704 –9.
GENERAL THORACIC
Ann Thorac Surg 2006;82:2031– 6
GENERAL THORACIC
2036
MOSKOVITZ ET AL ESOPHAGECTOMY IN OCTOGENARIANS
10. Spivak H, Vande Maele D, Friedman I, Nussbaum M. Colorectal surgery in octogenarians. J Am Coll Surg 1996;183:46–50. 11. Dunlop WE, Rosenblood L, Lawrason L, et al. Effects of age and severity of illness on outcome and length of stay in geriatric patients. Am J Surg 1993;165:577– 80. 12. Bailey SH, Bull DA, Harpole DH, et al. Outcomes after esophagectomy: a ten-year prospective cohort. Ann Thorac Surg 2003;75:217–22. 13. Griffin S, Desai J, Charlton M, Townsend E, Fountain SW. Factors influencing mortality and morbidity following oesophageal resection. Eur J Cardiothorac Surg 1989;3:419 –24. 14. Naunheim KS, Hanosh J, Zwischenberger J, et al. Esophagectomy in the septuagenarian. Ann Thorac Surg 1993;56:880 – 4. 15. Ellis FH Jr, 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. 16. Adam DJ, Craig SR, Sang CTM, Cameron EWJ, Walker WS. Esophagectomy for carcinoma in the octogenarian. Ann Thorac Surg 1996;61:190 – 4.
Ann Thorac Surg 2006;82:2031– 6
17. Kuwano D, Morita M, Baba K, et al. Surgical treatment of esophageal carcinoma in patients eighty years of age and older. J Surg Oncol 1993;52:36 –9. 18. 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. 19. Cancer Therapy Evaluation Program. Common terminology criteria for adverse events, version 3.0. National Cancer Institute, National Institutes of Health, Department of Health and Human Services, March 31, 2003. Available at: http://ctep.cancer.govAccessed on December 10, 2005. 20. Stein HJ, Feith M, Siewert JR. Cancer of the esophagogastric junction. Surg Oncol 2000;9:35– 41. 21. 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. 22. Birkmeyer JD, Siewers AE, Finlaysonn E, et al. Hospital volume and surgical mortality in the United States. N Engl J Med 2002;346:1128 –37.
DISCUSSION DR MARK J. KRASNA (Baltimore, MD): I really enjoyed your presentation. I have two quick comments and then a question. Someone is going to have to go to the dictionary later this afternoon and see if the octogenarian is defined as the eighth decade or in the 80s. I think some of these patients might be nonagenarians. The other comment is, at least for The Annals publications, you are going to need to report 90-day mortality as well as in-hospital mortality, so you should definitely include that in your analysis. Your results are very different from both Tom Rice’s at the Cleveland Clinic as well as the group at MD Anderson from last year, and I wonder if you could comment what the differences between your sets are. Last year the MD Anderson group showed excellent results in elderly patients with esophageal cancer, although they did not have as many patients in their 80s, but did not actually show a significant difference, including those patients who got neoadjuvant therapy. Tom Rice’s group at Cleveland Clinic about a year ago also showed data on elderly patients and found that age was an independent risk factor for death only when associated with thoracotomy. So my other specific question to you is, would you now consider still doing esophagectomies on these 80-year-olds and older but doing them as a nonthoracotomy transhiatal esophagectomy? Thank you. DR MOSKOVITZ: We are continuing to perform the standard operation at Memorial that is done in the majority of the patients, which is Ivor Lewis esophagectomy: thoracotomyassociated esophagectomy. We continue to apply careful selection criteria. We are using this data primarily to educate our patients in what we believe is an increased risk in our octogenarian population. DR KRASNA: Well, if that’s the case, then if there’s no difference in stage between the two groups and yet you are offering these patients a different treatment, which is surgery without induction therapy, as you showed us, why not offer them a different operation? I think it’s something to consider since there is a significantly increased mortality. DR MOSKOVITZ: Thank you.
DR STEPHEN G. SWISHER (Houston, TX): Could you tell me what your mortality was for your 80-year-olds who had gotten induction chemoradiation versus those who didn’t? DR MOSKOVITZ: There were 6 octogenarian patients who had induction chemoradiation and 2 of them are included as the 6 octogenarian deaths. DR KEMP H. KERNSTINE (Los Angeles, CA): You were talking about induction therapy. Did you have any of the patients that you included in the induction group that had “definitive chemoradiotherapy” and then you went on and tried to resect them, as a “salvage esophagectomy”? It’s my impression that that’s more common in the older patient, because the medical oncologists will say, “You are older and can’t have any surgery.” The patients are selected out by the medical oncologists to get chemoradiotherapy, then they recur a year later and then they are sent to us. Did you have any patients like that? The reason that’s important is because those patients tend to have a higher morbidity and mortality after their procedure, and that might influence your results. It may not be age alone. DR MOSKOVITZ: I can tell you that in our octogenarians, from the records that I reviewed, 3 of the 6 had our standard approach at Memorial, which includes 5,040 Gy of radiation combined with the chemotherapy, cisplatin-based chemo, and I’m not sure precisely what the other 3 had, so I can’t comment on that. DR JOE B. PUTNAM (Nashville, TN): Before we completely abandon surgery for esophageal cancer in this population, do you have any idea as to the results of the medical oncology/ radiation oncology therapy for these octogenarians and anticipated complications or mortality from that treatment and the subsequent survival at your institution? DR MOSKOVITZ: That’s an excellent question. It’s very difficult to ascertain in our institution, but I think that is the next step. The question when there is an increased mortality in the octogenarians undergoing operation is what is the risk to the best alternative therapy, and that may be as high or higher. So we’re taking our data and looking forward to think about how we can study this in a prospective fashion, and we are using the data to educate our patients, still trying to provide the best operative care that we can.