| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Division of Surgical Oncology (LEM), University of Vermont, Burlington, Vermont; and the Departments of General Oncologic Surgery (CC, DZJC, LDW), Biostatistics (DDS), and Nursing Research (TB, GJ, BRF), City of Hope National Medical Center, Duarte, California.
Correspondence: Address correspondence and reprint requests to: Laurence E. McCahill, MD, University of Vermont College of Medicine, Division of Surgical Oncology, UHC Campus, 1 South Prospect Street, Burlington, VT 05401; Fax: 802-847-0574; E-mail: larry.mccahill{at}vtmednet.org
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: Fifty-nine patients were evaluated for major symptoms of intent to treat and were followed up until death or last clinical evaluation. Surgeons identified planned operations before surgery as either curative or palliative and estimated patient survival time. An independent observer assessed symptom relief. A palliative surgery outcome score was determined for each symptomatic patient.
Results: Surgeons identified 22 operations (37%) as palliative intent and 37 (63%) as curative intent. The median overall survival time was 14.9 months and did not differ between curative and palliative operations. Surgical morbidity was high but did not differ between palliative (41%) and curative (44%) operations. Thirty-three patients (56%) were symptomatic before surgery, and major symptom resolution was achieved after surgery in 26 (79%) of 33. Good to excellent palliation, defined as a palliative surgery outcome score >70, was achieved in 64% of symptomatic patients.
Conclusions: Most symptomatic patients with advanced malignancies undergoing major operations attained good to excellent symptom relief. Outcome measurements other than survival are feasible and can better define the role of surgery in multimodality palliative care. A new outcome measure to evaluate major palliative operations is proposed.
Key Words: Palliative surgery • Quality of life • Outcomes • Advanced malignancy
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In this study, we evaluated the ability of surgeons to define major operations before surgery as curative or palliative by using a set definition of palliative surgery developed from results of a prior survey.5 Additionally, we sought to establish the effectiveness of major surgery in alleviating significant and sometimes disabling symptoms for patients with advanced malignancies. To better determine the success of major surgery in patients likely to die as a result of their disease, we sought to specifically evaluate patients over an extended duration for both symptom recurrence and delayed surgical-related morbidity. Through analysis of patient outcomes, including initial symptom relief, symptom recurrence, surgical morbidity, and hospitalization time relative to the patient’s remaining survival time, we sought to establish a palliative surgery outcome score (PSOS). We hoped that this outcome score could more objectively define the role of surgery in the treatment of symptoms of advanced malignancy.
|
PATIENTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
BORDER="0">18 years old for study entry. The study enrollment began in November 2000 and concluded in June 2001. The City of Hope’s Research Subjects Protection Institutional Review Board approved the study. Informed consent was obtained from both the patients and surgeons participating in this study. Staff surgeons were given a questionnaire to complete before surgery. Surgeons were asked to classify the proposed operations as either palliative or curative, given the following operational definition of a palliative operation: "The proposed operation is largely intended for symptom relief OR avoidance of symptoms or conditions anticipated secondary to progressive local disease, AND is unlikely to alter the ultimate progression of disease in this patient or significantly impact patient survival."
Surgeons also categorized patients before surgery as either symptomatic or asymptomatic. Symptomatic patients were identified as those judged likely to obtain symptomatic relief from the proposed operation. Surgeons’ completed questionnaires were collected before surgery. Two to 4 weeks after the operation, surgeons completed a second questionnaire. They were asked to reclassify the completed operation as palliative, curative, or noncurative on the basis of additional information learned at surgery, the actual procedure completed, and review of final surgical pathology. A modified postoperative definition of palliative surgery was given that defined a palliative operation as follows: "The completed operation was intended solely for symptom relief or avoidance of symptoms or complications related to progressive local disease OR was curative in intent but did not result in complete resection of all gross disease in a patient with a poor prognosis."
The postoperative questionnaire also asked surgeons to rate the extent of disease encountered at operation as equal to, greater than, or less than that which was anticipated, on the basis of all available studies, before surgery.
Patient outcomes were evaluated after surgery. A single independent evaluator (L.E.M.), who was not involved with any operation, evaluated both initial symptom relief and primary symptom recurrence. Evaluation was determined through direct patient encounters, patient telephone follow-up, evaluation of the medical record, and consultation with family members, study nurses, and primary care physicians and oncologists. Patients were followed up continuously through August 2002 or until the time of patient death. Events monitored included initial symptom relief for symptomatic patients, recurrence of symptoms that the operation was intended to treat, surgical complications, hospitalization time related to the initial surgery, management of surgical complications, and management of recurrent symptoms. To estimate the relative therapeutic benefit of surgery for symptomatic patients, a PSOS was developed. This score incorporated both elements of treatment success (symptom relief) and treatment failure (symptom recurrence and surgical complications) and monitored the hospitalization time associated with both. The PSOS indicates the percentage of postoperative days for which a patient was nonhospitalized, free of the symptom that the operation was intended to treat, and free of any major surgical complication in the 6 months after surgery. If a patient died within 6 months of surgery, the denominator used in calculating PSOS was the patient’s remaining days of life after surgery. Major complications were defined as any permanent complication that affected the patient’s mobility, comfort, or ability to return to his or her preoperative living situation or any temporary complication that prolonged inpatient treatment, required hospital re-admission, or required additional outpatient visits beyond routine postoperative follow-up.
Additional clinical factors were recorded for evaluation as predictors of patient survival time. Specific preoperative laboratory values recorded included serum albumin, blood urea nitrogen, creatinine, sodium, and bilirubin. Preoperative total body weight loss was also recorded and was defined by the patients’ measured weight at the time of surgery and their self-reported weight before the onset of their cancer diagnosis. All values were examined as continuous variables without predefined set points in this pilot evaluation.
Descriptive statistics from the surgeon questionnaire and the patients’ medical records were reported as proportions for discrete data and as medians and means with standard deviations and ranges for continuous data. The product-limit method of Kaplan and Meier was used to estimate overall patient survival and other time-to-event end points. Overall survival was calculated from the date of operation until death; patients were censored at last follow-up if they were still alive. Comparisons between curves were made by using the log-rank test. We compared categorical data by using 
2 tests for independence unless any patient group was prohibitively small for asymptotic assumptions. We fit Cox proportional hazards regression models for multivariate and continuous data on time-to-event end points. P values of <.05 were considered significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
1 year for 37 (65%) of 57 patients. In comparison, surgeons were more optimistic in their estimates of patient survival time when projecting the anticipated effect of the proposed operation. On average, surgeons who anticipated surgery extended patient survival time by an additional 5 months beyond the time estimated without surgery.
After a median follow-up of almost 15 months, overall patient survival from the date of surgery was compared with surgeons’ estimates. The 1-year survival rate for the patients whom surgeons expected to live >1 year was 85% with a median survival not yet reached, compared with a 1-year survival of 43% with a median survival of 9.4 months for patients anticipated to live 1 year without surgery, as shown in Fig. 1A (P < .0002). In contrast, the 1-year survival rate for patients whom surgeons had estimated to live >1 year, taking the proposed surgery into account, was just 61% (Fig. 1B). When taking the proposed surgery into account, surgeons had estimated that only 2 patient deaths would occur within 6 months of surgery, yet 17 patient deaths occurred within this time frame. Among the 37 patients for whom surgeons had estimated a life expectancy of 1 year if no surgery was performed, surgeons anticipated that the proposed operation would extend survival time for 30 (81%) of 37, and for 24 of these patients, they anticipated a survival time >1 year. Among the latter group, 14 (58%) of 24 died in <1 year after their completed operation. Even among patients identified as undergoing palliative-intent procedures, surgeons estimated that surgery would extend survival time in most of these patients. For the 10 palliative-intent patients for whom postoperative survival time was projected at >1 year, just 3 (30%) lived beyond 1 year.
|
The median overall survival for all patients was 14.9 months, with a median follow-up of 14.7 months. Thirty patients died during the follow-up period, including the three postoperative deaths and one death attributable to complications of chemotherapy. The remaining 26 patient deaths were secondary to progressive malignant disease. On the basis of the surgeons’ preoperative classifications, the median survival time for palliative-intent patients was 10.0 months, and the median survival time for curative-intent patients was not yet reached, a difference that did not reach statistical significance (P = .0815). By using the postoperative procedure designations, the median survival for postoperatively defined curative patients (median not yet reached) was longer than the survival time for the 10 patients identified as noncurative (7.7 months); all of these patients’ operations had a curative intent, but the patients were unresectable at operation (P = .02; Fig. 2). The median survival time for postoperatively defined palliative patients was 11.0 months, which was less than that for postoperatively defined curative patients (P = .0458).
|
|
Symptom Outcomes
Among the 33 patients designated as symptomatic before surgery, symptom resolution was documented in 26 (79%) of 33, including 16 (80%) of 20 palliative-intent patients and 10 (77%) of 13 symptomatic curative-intent patients. For seven patients, symptoms either never resolved (jaundice, n = 2; pain, n = 2; and bowel obstruction, n = 3) or were considered to have negligible improvement after surgery. For most symptomatic patients, symptom improvement was readily apparent and allowed patients to return to their prior home living situation. For preoperative palliative-intent patients, 15 (68%) of 22 returned to independent living after a major operation. With continued long-term follow-up, symptoms initially alleviated by surgery recurred in 8 (24%) of 33 patients at a median of 4.9 months after surgery. A PSOS was calculated for each of the 33 symptomatic patients and is shown in Table 4. We considered a PSOS of >70 as good to excellent surgical palliation because it indicated that a patient lived at least 70% of the study period (remaining days of life or the 6 months after surgery) both nonhospitalized and free of both the symptom treated surgically and any major surgical morbidity. A PSOS of >70 was achieved in 21 (64%) of 33 symptomatic patients. Among patients with a relatively short postoperative survival time (<6 months), a PSOS >70 was achieved in just 4 (36%) of 11.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This study asked surgeons to consider carefully the term palliative by using a definition developed on the basis of a prior survey of the Society of Surgical Oncology.5 We found relatively few instances (4 of 14) of curative-intent procedures reclassified as palliative after surgery. It seems that when surgeons are asked prospectively to identify a palliative benefit for an asymptomatic patient for whom a curative resection cannot be completed, most do not. Instead, they identified these procedures as noncurative. Although this may seem a relatively straightforward result, it calls into question the wider usage of "palliative surgery" in the surgical literature that segregates cancer operations as curative or palliative on the basis of residual disease status at procedure completion. Although the nontherapeutic benefit of making an incision and diagnosing carcinomatosis without further resection is well understood by most practicing surgeons, this study helps demonstrate the negative implications of identifying such procedures as palliative. Patients undergoing a noncurative procedure in this study had a 3.3-month shorter median survival time than patients undergoing successfully completed palliative procedures. Grouping unresectable patients with patients undergoing successful palliative procedures has resulted in shorter reported survival times for palliative patients, leading some to question the utility of performing major palliative operations.13,14 Additionally, potential improvement in quality of life for symptomatic palliative-intent patients is likely to be underestimated by the inclusion of asymptomatic curative-intent patients whose disease is found to be unresectable.
From the findings of this study and our prior survey, we would propose that future prospective studies of palliative surgery ideally be limited to symptomatic patients. Although palliation of anticipated symptoms remains a reasonable pursuit, future studies evaluating palliative benefits should clearly distinguish palliation of symptomatic patients from anticipated palliation of currently asymptomatic patients. The strength of this study is somewhat diminished by the inclusion of two patients designated as palliative who were asymptomatic but were considered palliative on the basis of a surgeon’s perception of a high likelihood of subsequent symptoms. Measurement of effective palliation in patients for whom symptoms are anticipated is problematic and requires a control arm of patients not undergoing surgery to determine the true incidence of symptom development in a well-matched patient cohort.
This study is in agreement with prior reports of high complication rates for surgery for patients with advanced malignancies and stresses the significance of judgment and experience in patient selection.15–17 Contrary to some previous studies, however, we found that complication rates were not significantly different between curative-intent and palliative-intent procedures (P = .8125). This is likely because our patient eligibility criteria were based on disease stage alone. It was our perception from previous work that surgeons remain very individualistic in their interpretation of a procedure’s curative potential, and, in limiting enrollment to procedures considered palliative, we would be subject to individual surgeons’ biases.7 We have demonstrated that for similarly staged patients undergoing either curative-intent or palliative-intent procedures, complication rates are the same. This finding challenges the concept that palliative surgery has higher risk than curative surgery.
This study demonstrated that surgeons were generally accurate in predicting patient survival time. Their accuracy was better when they were asked to predict patient survival in the absence of further treatment (i.e., if the proposed surgical procedure were not performed). Other studies have suggested that the prognostic skills of clinicians have severe limitations.18,19 Surgeons in this study, however, did overestimate the effect of the proposed surgery on patient survival time in this advanced cancer population. Even among palliative-intent patients, an extension of survival time was anticipated for more than half, yet these patients rarely lived beyond the survival time originally estimated for them in the absence of surgery. The importance of accurately estimating patient survival time in this advanced cancer population cannot be underemphasized. Patients with survivals <6 months were less likely to achieve good palliation in this study, as defined by a PSOS >70, with just 36% achieving good palliation after surgery. Other recent trials evaluating the efficacy of surgical palliation have also reported a correlation between survival times of <3 months and the failure to achieve successful surgical palliation.16 It is possible that such patients might be more effectively managed by a multidisciplinary healthcare team focusing on palliative care and avoiding all but low-risk surgical interventions.
This study demonstrated clinical criteria that may assist surgeons in identifying patients with shorter survival times. Both preoperative serum albumin and preoperative total body weight loss were predictors of survival time in this small patient cohort. Other studies have reported low serum albumin as a predictor of survival in cancer patients undergoing surgical resection.20,21 Total body weight loss was also predictive of survival time in this study and is an inexpensive parameter to obtain. These factors may prove valuable in decision making for patients who are considered at high risk for operative complications or who have symptoms that are questionably suitable for surgical palliation. Validation of these clinical parameters as meaningful in terms of preoperative patient selection has yet to be confirmed.
The major focus of this study was the prospective and systematic evaluation of the effectiveness of surgery in alleviating symptoms of advanced cancers. Most previous studies have attempted to quantify this retrospectively on the basis of the duration of symptom relief and time spent out of the hospital, two concepts we considered important in defining effective palliation.15 We chose to include both symptomatic curative-intent patients and palliative-intent patients in evaluating the success of surgery in symptom relief. Available data suggest that most patients with stage III and IV solid tumors undergoing surgery with curative intent (including R0 resections) are more likely than not to die as a result of their disease. Nevertheless, patients undergoing curative-intent procedures may be highly symptomatic, and the symptom relief provided by surgery in patients undergoing curative-intent surgery has rarely been evaluated. In this study, 30% of patients undergoing curative-intent surgery were symptomatic. By calculating an outcome score that incorporates initial symptom relief, long-term symptom relief, and treatment-related morbidity, we found that good to excellent palliation was achieved in 64% of symptomatic patients. This percentage represents patients who had sustained symptom relief and remained out of the hospital for >70% of their remaining days of life. Our report of 64% with successful surgical palliation compares favorably to a recent similar prospective study of palliation in patients with advanced cancer, which reported successful palliation in 46% and a median duration of just 3.4 months of successful palliation.16 Our higher reported success of surgical palliation is likely attributable to patient selection, because the report by Miner et al.16 included more patients with malignant biliary, gastric, and bowel obstruction, all clinical scenarios that are notoriously challenging to palliate. In our report, the most common symptom with intent to palliate was pain. Additionally, the study of Miner et al. required surgeons to identify proposed major operations as palliative procedures for study entry, rather than using the criteria of disease stage, as in this study. This likely resulted in the selection of patients with more advanced disease. The median survival for all palliative-intent patients in the report of Miner et al. was 3.5 months, with an in-hospital mortality of 31%, versus a median survival of 10.0 months and a 3% in-hospital mortality for palliative-intent patients in this report; this difference is likely attributable to selection of more advanced or poorer-prognosis disease in their report. Both studies confirm that operative morbidity is high, symptom relief is not guaranteed, and the duration of initially successful relief may be limited.
A PSOS, as proposed in this study, seems to be an advancement in quantitatively measuring the success of surgery in a high-risk patient population in which surgical morbidity can quickly mask any gains in symptom relief. A major limitation of this report is that our proposed scoring system has not yet been validated against more established quality-of-life measures that are well validated and reliable.11 Studies evaluating the validity of PSOS are under way. This study was also limited by small patient numbers and a heterogeneous patient population, which did not allow conclusions to be drawn regarding the likelihood of specific symptom relief for different disease entities. Further work in this area with larger patient cohorts is planned and will likely require multicenter cooperation.
The benefits of palliative surgery have been called into question by those who have specifically noted the significantly diminished survival time in patients undergoing palliative surgery compared with curative surgery.13 Such conclusions speak little to the quality of symptom relief obtained by these procedures and highlight the limitations of surgery literature that focuses outcome evaluations only on treatment end points of patient survival and local disease recurrence, when the most valued end point in the palliative setting is symptom resolution.5 The prospective and extended evaluation of symptom alleviation in this study, however, increases the accuracy of the conclusion that successful palliative outcomes can be achieved with surgery in carefully selected patients.
Symptom relief for patients with advanced malignancies can be achieved and documented through a careful assessment of preoperative symptoms and use of an objective measurement (PSOS). Validation of this measure against more established quality-of-life measures is needed. Future studies evaluating the effectiveness of palliative surgery will be enhanced by adhering to a preoperative definition of palliative surgery that allows for careful assessment of both symptom severity and the effectiveness of surgery or other interventions in alleviating the symptoms of disease. The potential for surgery to play an important role in the treatment and palliation of advanced malignancies is apparent and will likely improve with further refinement in patient symptom and disease selection criteria that are most likely to benefit from surgical intervention.
|
ACKNOWLEDGMENTS |
|---|
The acknowledgments are available online at www.annalssurgicaloncology.org.
|
FOOTNOTES |
|---|
Received for publication June 14, 2002. Accepted for publication March 25, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. J. Brasel Quality-of-life Assessment in Palliative Care American Journal of Hospice and Palliative Medicine, June 1, 2007; 24(3): 231 - 235. [Abstract] [PDF]
|
|
|
|
 |
|
 S. L. Wong When a Cure Isn't Possible... Do All Roads Lead to Palliation? Ann. Surg. Oncol., April 1, 2007; 14(4): 1245 - 1246. [Full Text] [PDF]
|
|
|
|
|
|
Y. D. Podnos and L. D. Wagman The Surgeon and Palliative Care Ann. Surg. Oncol., April 1, 2007; 14(4): 1257 - 1263. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. J. Yeh, S. Singer, M. F. Brennan, and D. P. Jaques Effectiveness of Palliative Procedures for Intra-Abdominal Sarcomas Ann. Surg. Oncol., December 1, 2005; 12(12): 1084 - 1089. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Galante, T. L. Bowles, V. P. Khatri, P. D. Schneider, J. E. Goodnight Jr, and R. J. Bold Experience and Attitudes of Surgeons Toward Palliation in Cancer Arch Surg, September 1, 2005; 140(9): 873 - 880. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
