998 Outcome of Palliative Urinary Diversion in the Treatment of Advanced Malignancies Bijan Shekarriz, M.D.1 Hodjat Shekarriz, M.D.2 Jyoti Upadhyay, M.D.1 Mousumi Banerjee, Ph.D.1 Herman Becker, M.D.3 J. Edson Pontes, M.D.1 David P. Wood, Jr., M.D.1 1 Department of Urology, Wayne State University, Karmanos Cancer Institute, Detroit, Michigan. 2 Klinik für Chirurgie, Medizinische Universität zu Lübeck, Lübeck, Germany. 3 Department of Urology, Marienkrankenhaus, Hamburg, Germany. BACKGROUND. It is unclear whether palliative endourologic or percutaneous urinary diversion in the treatment of advanced cancer provides significant improvement in quality or duration of life. The purpose of this study was to evaluate survival and performance status after endourologic palliative urinary diversion in patients with advanced malignancy and to compare the results for different malignancies. METHODS. One hundred three patients with advanced malignancies underwent palliative urinary diversion (stent or nephrostomy) between 1986 and 1997. Ninetytwo patients and 11 patients had bilateral and unilateral obstruction, respectively. Indications, complications, performance status after diversion, and survival for patients with different malignancies were identified and compared. A modified Karnofsky performance scale (KPS) was used for assessment of physical performance. A scale of 0 – 4 was used: 0) hospitalized until death; 1) bedridden at home, severe pain despite analgesia; 2) moderate disability, moderate pain despite analgesia; 3) mild disability, pain free with medication; and 4) normal. RESULTS. The mean age of patients was 68 years. The mean pre- and postoperative creatinine levels were 6 mg/dL and 3.3 mg/dL, respectively (P , 0.0001). The median survival and days of hospitalization were 112 and 45, respectively. The median postdiversion KPS score was 2 (range, 0 – 4), and 15% of patients never left the hospital. Overall, 51% required secondary percutaneous procedures, and 68.4% had complications (minor, 63%; major, 5.4%). CONCLUSIONS. Most patients with advanced cancers had poor performance status and survival after diversion. Eighty six percent had cancer-related symptoms despite the diversion. The average survival was 5 months, 50% of which was spent in the hospital. Primary endourologic procedures had a high failure rate, and additional procedures were required. Cancer 1999;85:998 –1003. © 1999 American Cancer Society. KEYWORDS: palliative urinary diversion, malignancy, ureteral stent, percutaneous nephrostomy, malignant ureteral obstruction. E Presented at the 93rd annual meeting of the American Urological Association, San Diego, California, May, 1998. Address for reprints: David P. Wood, Jr., M.D., Department of Urology, Wayne State University, 4160 John R. 1017, Detroit, MI 48201. Received June 1, 1998; revision received August 24, 1998; accepted August 24, 1998. © 1999 American Cancer Society xtrinsic ureteral obstruction secondary to malignancy is commonly a late manifestation of metastatic disease. Pelvic malignancies are the most common cause; however, metastatic processes secondary to nonpelvic malignancies, such as breast carcinoma and upper gastrointestinal tract malignancies, can also cause ureteral obstruction.1– 6 Most patients present with an advanced stage of the disease, with other sites of metastasis being commonly documented at the time of presentation. The optimal management of the malignant ureteral obstruction remains unclear. It has been reported that self-retaining ureteral stents have a high failure rate in extrinsic obstruction.7 Furthermore, the endoscopic insertion of ureteral stents may be technically difficult or even impossible in the presence of advanced pelvic involvement by a malignant process.4 Percutaneous Palliative Urinary Diversion in Advanced Malignancies/Shekarriz et al. nephrostomy catheters are a commonly used alternative as either a primary procedure or in the case of failure of transurethral procedures.8 Clinicians occasionally are confronted with the dilemma of performing palliative urinary diversion in terminally ill patients. Previous studies have addressed the technical aspects of urinary diversion and the associated complications. However, the palliative nature of these procedures makes the assessment of the performance status after diversion by means such as the Karnofsky performance scale (KPS) an important objective factor to determine the success of urinary diversion in this scenario. The aim of this study was to investigate the survival and performance status by KPS after palliative urinary diversion in patients with advanced malignancies and to compare these results between different malignancies. MATERIALS AND METHODS Data were obtained from all patients who underwent urinary diversion for ureteral obstruction secondary to advanced malignant disease at two institutions between 1986 and 1997. A total of 103 patients with advanced malignancies resulting in ureteral obstruction underwent palliative urinary diversion. The diagnosis of urinary obstruction was made by renal ultrasound or computerized tomography (CT). Ninety-two patients had bilateral obstruction, and 11 patients had unilateral involvement. Endoscopic ureteral stent placement or percutaneous nephrostomies were performed according to standard techniques. Indications, complications, performance status after diversion, and survival were identified retrospectively and compared between different types of malignancies. The performance status prior to diversion was not available. For this comparison, patients with bilateral obstruction were divided into four groups based on the primary diagnosis: 1, bladder carcinoma (n 5 25); 2, gastrointestinal carcinoma (n 5 19); 3,Gynecological carcinoma (n 5 20); and 4, prostate carcinoma (n 5 28).The group of gynecological tumors included cervical, ovarian, and breast carcinomas. The primary tumor types are listed in Table 1. A modified KPS was used as an objective assessment of physical performance after diversion using a validated questionnaire. Informed consent was obtained, and the questionnaires were coded to maintain confidentiality. Since its introduction in 1948,9 KPS, which was designed initially for evaluating performance status in patients after chemotherapy treatment, has undergone modifications by other authors for application in different patient populations.10 –12 For the evaluation of patients after palliative urinary diversion, the degree of cancer-related pain is an im- 999 TABLE 1 Primary Malignancy in 92 Patients Type of malignancy No. of patients (%) Prostate Bladder Colorectal Gastric Breast Cervix Ovary Endometrium Fallopian tubes Total 28 (30.4) 25 (27.1) 17 (18.4) 2 (2.1) 5 (5.4) 8 (8.7) 5 (5.4) 1 (1.1) 1 (1.1) 92 (100) portant factor. Because pain is not specifically included in the original scoring system of Karnofsky, this scale was modified to address both performance and level of pain in our patients. We used the modified KPS by Köster et al.10 This scale includes nine categories. Because of the retrospective nature of our study in a small cohort of patients, we combined the nine categories into five categories for statistical analysis. A scale of 0 – 4 was used as follows: 0) hospitalized until death; 1) bedridden at home, severe pain despite analgesia; 2) moderate disability, moderate pain despite analgesia; 3) mild disability, pain free with medication; and 4) normal. Primary physicians and family members were contacted for this evaluation. The degree of social integration, level of pain by need for pain medication, patient’s functional independence, and overall performance status were addressed specifically with a detailed questionnaire. Generally, a performance status of 3 or 4 was considered good because of minimal limitation of daily activity and achievement of a pain free status. Statistical Analysis Analyses of variance and Kruskal–Wallis tests were used for all comparisons between the groups. For changes in pre- and postoperative creatinine level, analyses of variance on the difference were used. The t test was performed for comparison of survival and KPS between patients with and without disseminated metastasis. Pearson correlation coefficients were used to determine the correlation of preoperative creatinine levels and time from diagnosis to survival and KPS. RESULTS Bilateral Ureteral Obstruction Ninety-two patients had bilateral hydronephrosis (n 5 91) or unilateral hydronephrosis in a solitary kidney (n 5 1). In these patients, the main presentation was 1000 CANCER February 15, 1999 / Volume 85 / Number 4 TABLE 2 Comparison of Age and Pre- and Postoperative Creatinine Group All Bladder Gastrointestinal Gynecologic Prostate Age (yrs) 68 6 12 71 6 10 60 6 11 60 6 11 77 6 6 Preoperative creatinine 6.8 6 5.4 5.6 6 4.6 4.6 6 4.2 7.1 6 6.0 8.8 6 5.8 Postoperative creatinine 3.34 6 2.7 3.47 6 3.9 2.8 6 2.5 3.2 6 2.3 3.5 6 2.1 TABLE 3 Modified Karnofsky Performance Scale and Survival after Urinary Diversion P value 0.0001 0.0017 0.0045 0.0084 0.0002 Group Median KPS score (range)a Median survival in days (range) Bladder Gastrointestinal Gynecologic Prostate Total 1 (0–4) 2 (0–3) 1 (0–4) 2 (0–4) 2 (0–4) 122 (3–391) 90 (13–432) 65 (1–502) 173 (5–600) 112 (1–600) a low urine output with uremia. In addition, other symptoms, such as flank pain, hematuria, and urinary tract infection, were encountered. The mean age in all groups was 68 years 6 12.5 years (Table 2). Patients with prostate and bladder carcinomas were significantly older (P 5 0.0001). Preoperative and postoperative creatinine levels The pre- and postoperative creatinine levels in different groups are shown in Table 2. Overall, the mean preoperative creatinine level was 6. 8 6 5.4. The preoperative creatinine level was significantly higher in patients with gynecological and prostate carcinomas compared with other groups (P 5 0.02). The mean postoperative creatinine level was 3.3 6 2. 8; the creatinine level after urinary diversion decreased significantly in all groups (P , 0.0001; Table 2). The correlation of preoperative creatinine levels to survival approached statistical significance (P 5 0.06). Clinical stage and survival Overall, 70% of the patients had known disseminated metastasis at the time of diversion. The median survival was 112 days (range, 1– 600 days) in all patients; comparison of survival data between different types of cancer did not show any significant differences (P 5 0.1; Table 3). Patients with prostate carcinoma had the longest median survival of 173 days (range, 5– 600 days). For comparison of survival based on disease stage, patients with pelvic malignancies were divided into those with known disseminated disease at the time of diversion and those with locally advanced disease or regional nodal metastasis. Of the patients with pelvic tumors, 33% had locally advanced or regional metastases only. The median survival in these patients was 126 days (range, 2–510 days) compared with 100 days (range, 1– 600 days) in patients with pelvic tumors and disseminated metastasis (P 5 0.94). Prior treatment for primary carcinoma Most patients had undergone prior surgical or medical treatment for their disease. These surgical procedures A modified Karnofsky symptom score (KPS) was used: 0) hospitalized until death; 1) bedridden at home, pain under analgesia; 2) moderate disability, moderate pain under analgesia; 3) mild disability, pain free with medication; and 4) normal. included transurethral resection, simple prostatectomy, orchiectomy, total abdominal hysterectomy, mastectomy, and various intestinal procedures. Medical treatment included hormonal therapy and chemoradiation therapy. Overall, 62 patients (67.4%) had multimodal therapy, and 16 patients (17.4%) had at least one mode of therapy for their primary disease prior to urinary diversion. Only, 14 patients (15.2%) had no prior therapy at the time of diversion. Of these 14 patients, 7 were deemed incurable, and no treatment was rendered. The primary diagnosis was unknown at the time of diversion in 4 patients, and 3 patients refused treatment. Therefore, only 3 of the previously untreated patients had treatment after diversion. Comparison of survival between those patients with prior multimodal therapy and those who did not have any treatment or had at least one treatment did not show any significant differences (P 5 0.91). The time from diagnosis of the primary tumor to urinary diversion did not correlate with survival after urinary diversion (p 5 0.9). Karnofsky performance scale score and hospitalization after diversion The median KPS score in all patients was 2 (range, 0 – 4; Table 3). Patients with prostate carcinoma had a better KPS score; however, the differences did not reach statistical significance (P 5 0.56). Only 13 patients (14%) were free of pain with normal functional status after the procedure. The median KPS score in patients with known disseminated metastasis was 1 (range, 0 – 4) compared with 2.5 (range, 0 – 4) in those with locally advanced or local metastatic disease (P 5 0.03). Fourteen patients (15%) never left the hospital after urinary diversion. The mean time of hospitalization after procedures until death was 45 days 6 32 days, which constituted 52% of total survival (Table 4). There were no significant differences in survival between malignancy types (P 5 0.73). Palliative Urinary Diversion in Advanced Malignancies/Shekarriz et al. TABLE 4 Postdiversion Days in Hospital until Death and the Percentage of Survival Spent in Hospital Type of cancer Days hospitalized Survival time (%) Bladder Gastrointestinal Gynecologic Prostate Total 46 6 31 44 6 29 51 6 43 39 6 25 45 6 32 52 52 59 46 52 1001 TABLE 5 Complications of Urinary Diversion Patient group Minor (%)a Major (%)b Total (%) Bladder Prostate Gastrointestinal Gynecologic Total 25 15.2 10.8 12 63 3.2 0 1.1 1.1 5.4 28.2 15.2 11.9 13.1 68.4 a b Minor complications: hematuria, catheter blockage or dislodgement, or urinary tract infection. Major complications: significant bleeding, bladder tamponade (requiring surgical intervention). Technical failure and complications The success rate of primary endoscopic stent placement was evaluated. Fifty-one percent of patients required percutaneous procedures for adequate diversion secondary to unsuccessful stent placement. Overall, 68.4% had procedural related complications in the postoperative course (Table 5). Sixty-three percent had minor (hematuria, catheter blockage requiring manipulation and irrigation, and nephrostomy dislodgment), and 5.4% had major complications (significant bleeding, bladder tamponade, or complications that required reoperation). In 30% of patients, dislodgment of the percutaneous nephrostomy occurred, which required replacement. Unilateral Urinary Obstruction Eleven patients had unilateral ureteral obstruction. Their primary diagnoses were carcinoma of the prostate (n 5 3), bladder (n 5 3), colon (n 5 3), and breast (n 5 2). Six patients (54.5%) had known metastatic disease at the time of presentation. Four patients were asymptomatic at the time of diagnosis, six presented with flank pain, and one had hematuria. Flank pain resolved in these patients after urinary diversion. The mean pre- and postoperative creatinine levels were 1.43 and 1.48, respectively. The mean survival was 293 days. Average hospitalization was 33.7 days, which constituted 11.2% of the overall survival. DISCUSSION Open surgical placement of nephrostomy tubes for the management of malignant ureteral obstruction was common practice until the 1970s. This procedure was associated with a high complication rate, a significant perioperative mortality rate, and an average length of survival of 6 months.5 The introduction of percutaneous techniques associated with the evolution of the field of endourology, the use of ultrasound and CT guidance, and the routine use of perioperative antibiotics have dramatically reduced the morbidity of ureteral stent placement. Furthermore, the availability of new stenting biomaterial for endoscopic insertion has contributed to safe urinary diversion. This is reflected in our data, in which the morbidity associated with the procedures was mostly minor; however, 63% of patients had complications. Significant major complications occurred in 5.4% of patients with no procedural mortality. A difficult ethical question arises in patients with incurable malignancies who are diagnosed with malignant ureteral obstructions. Decompression of the obstruction may merely prolong the patient’s suffering. Patients who present with bilateral obstruction and uremia have a short life expectancy if their disease remains untreated.1 These patients appear to present in a more advanced stage of the disease. The functional status of these patients is generally poor at the time of presentation secondary to preexisting conditions related to their malignancy. In a few studies on palliative urinary diversion in patients with advanced malignancies, the impact on quality of life after diversion has been addressed.5,13,14 Most of these studies were conducted prior to the routine use of endoscopic diversion. The assessment of quality of life is difficult, and different arbitrary definitions have been used.5,15 The KPS is the most widely used method of quantifying cancer patients’ ability to function.15 The KPS measures the patient’s ability to carry on the normal activity of everyday life or degree of dependence on others to perform these activities. Furthermore, the KPS has contributed to rehabilitating patients with cancer and can serve as a helpful tool both in the initial assessment as well as the ongoing assessment of prognosis. Although studies have suggested that, in a palliative setting, assessment of the quality of life using the KPS is more appropriate than survival per se,16 we used this scale as a way to quantify objectively the functional status of cancer patients after diversion rather than to assess their “quality of life,” which is more difficult. Our data suggest that the majority of patients with bilateral ureteral obstruction secondary to malignancy 1002 CANCER February 15, 1999 / Volume 85 / Number 4 had a poor performance status for the remainder of their life after diversion. Overall, 86% were symptomatic mainly from cancer-related pain after urinary diversion, and 15% never left the hospital. Survival of patients after palliative urinary diversion in earlier studies has been generally poor.1,4,5,17 The prognostic values of prior treatment, the stage of disease, and the period from diagnosis to diversion all have been debated. The site of the initial tumor has been thought to be the most important prognostic factor for survival, with prostate and cervical carcinomas having the longest reported survivals.10,17 It has been reported that nonurologic malignancies, especially gastric and pancreatic carcinomas, are worse offenders than urologic malignancies.6 In our series, the average survival of patients with bilateral obstruction after urinary diversion was between 4 months and 7 months. Although we found a trend toward better survival and quality of life in patients with prostate carcinoma, the differences were not significant. This may reflect the natural history of the disease and the fact that urinary obstruction may occur earlier during the course of this disease.4 Intuitively, patients with locally advanced tumors causing ureteral obstruction should do better than those with metastatic disease. One study reported a 7.6-month mean survival time in 20 patients who underwent percutaneous nephrostomy placement. However, their disease stages were not stated.18 Conversely, patients with cervical carcinoma who had disseminated metastasis did poorly.14 The majority of our patients had disseminated disease at the time of diversion. In patients with pelvic tumors, the performance status was significantly better in those with locally advanced (including lymph node metastasis) disease compared with those with disseminated disease (P 5 0.03); however, no statistically significant survival benefit was noted. The prognostic value of prior or subsequent treatment has been controversial. Watkinson et al.8 found that only patients who had treatment options after urinary diversion benefited from percutaneous nephrostomy. In their patients with relapsed disease and no conventional treatment options available, only 5.5% benefited from the procedure. The average survival was 38 days in this group, and 61% never left the hospital. In our series, 67% of patients had maximized multimodal treatment prior to diversion with no noted survival difference compared with patients who received no treatment or only a single modality of treatment prior to diversion. The majority of patients with prostate carcinoma (78%) had prior hormonal therapy, making a comparison with those who received no treatment difficult. Overall, only three pa- tients with no prior therapy underwent treatment after diversion. This may explain the fact that no difference was found in survival between those with multimodal and those with some or no prior treatment. It also highlights the observation that few patients (3.2%) receive systemic treatment after urinary diversion. In accordance with previous published data, no correlation existed between the period of time from diagnosis of malignancy to diversion with survival. Overall, no factor could be identified as a predictor of improved survival or performance status after diversion in our patient cohort. Zadra et al.4 reported symptomatic relief of pain in patients with unilateral ureteral obstruction. They did not recommend urinary diversion in asymptomatic patients with unilateral obstruction and stable contralateral renal function who failed endoscopic stenting. In another study, unilateral or bilateral obstruction did not appear to be an important factor in quality of survival.13 In our series, although the number of patients with unilateral obstruction was small, decompression resulted in resolution of symptoms in those patients with isolated flank pain. In asymptomatic patients, urinary diversion may be of little benefit. The type of urinary diversion in many instances is related to clinician biases, availability, and urgency of the diversion. A success rate of 37– 47% using primary endoscopic ureteral stents has been reported.4,13,19 Prostatic and bladder tumors were more difficult to divert endoscopically, because the ureteral orifices were difficult to see or were invaded by the tumor. Therefore, primary percutaneous nephrostomy in patients with prostate or bladder carcinomas has been advocated.4 In our patients, primary endoscopic stent placement had a high failure rate of 51%, necessitating percutaneous procedures. The failure rate in patients with prostate and bladder carcinomas was 55%. This is in agreement with the previously reported studies. Furthermore, Docimo and Dewolf7 reported a high postoperative failure rate (defined as occlusion within 30 days) of ureteral stents in 46% of patients with extrinsic ureteral obstruction. In patients with advanced malignancies, ureteral obstruction is commonly the result of external compression by metastatic process, which may explain the lower success rate of stenting. Although the percutaneous approach under ultrasound guidance has a high technical success rate, long term management of nephrostomy catheters is cumbersome and is associated with an inferior quality of life compared with internal stents.13 Perinephric abscess rates as high as 15% have been reported.18 Complications, such as bleeding, may be significant, and revisions may become necessary secondary to malpo- Palliative Urinary Diversion in Advanced Malignancies/Shekarriz et al. sition or blockage.17,18 Dislocation of the tube may occur in up to 40% of patients. With the advent of new percutaneous techniques, conversion of a percutaneous nephrostomy to an internal double-J stent using flexible guide wires is possible.20 This will eliminate the need for external collection devices and will improve patient independence. In conclusion, the indications for palliative urinary diversion should be individualized. Patients with prostate carcinoma showed a trend toward better survival and performance status compared with patients with other tumor sites. No other factors (tumor stage, metastasis, time from diagnosis to diversion) seemed to play a significant prognostic role in deciding which patient benefited from diversion. In an era of advances in the field of endourology with unquestionable improvement in mortality and morbidity, the quality of life and survival of patients with malignant ureteral obstruction remains poor. Patients and family members should be informed of the poor outcomes after palliative procedures to guide them toward appropriate decision making. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. REFERENCES 1. 2. 3. 4. 5. Brin EN, Schiff M, Weiss RM. Palliative urinary diversion for pelvic malignancy. J Urol 1975;113:619 –22. Lee SK, Jones HW III. Prognostic significance of ureteral obstruction in primary cervical cancer. Int J Gynecol Obstet 1994;44:59 – 65. Paul AB, Love C, Chisholm GD. The management of bilateral ureteric obstruction and renal failure in advanced prostate cancer. Br J Urol 1994;74:642–5. Zadra JA, Jewett MAS, Keresteci AG, Rankin JT, Louis E, Grey RR, et al. Nonoperative urinary diversion for malignant ureteral obstruction. Cancer 1987;60:1353–7. Holden S, Mcphee M, Grabstald H. The rationale of urinary diversion in cancer patients. J Urol 1979;121:19 –21. 16. 17. 18. 19. 20. 1003 Donat SM, Russo P. Ureteral decompression in advanced nonurologic malignancies. Ann Surg Oncol 1996;3:393–9. Docimo SG, Dewolf WC. High failure rate of indwelling ureteral stents in patients with extrinsic obstruction. Experience at two institutions. J Urol 1989;142:277–9. Watkinson AF, A’Hern RP, Jones A, King DM, Moskovic EC. The role of percutaneous nephrostomy in malignant urinary tract obstruction. Clin Radiol 1992;47:32–5. Karnofsky DA, Abelmann WH, Craver LF, Burchenal, JH. The use of nitrogen mustards in the palliative treatment of cancer. Cancer 1948;1(4):634 –56. Köster R, Gebbensleben B, Stützer H, Salzberger B, Ahrens P, Rohde H. Quality of life in gastric cancer. Karnofsky’s scale and Spitzer’s index in comparison at the time of surgery in a cohort of 1081 patients. Scand J Gastroenterol 1987; 22(Suppl 133):102– 6. Melvin TA, Ozbek IN, Eberle DE. Recognition of depression. Hosp J 1995;10:39 – 46. Ifudo O, Paul HR, Homel P, Friedman EA. Predictive value of functional status for mortality in patients on maintenance hemodialysis. Am J Nephrol 1998;18:109 –16. Fallon B, Onley L, Culp DA. Nephrostomy in cancer patients: to do or not to do? Br J Urol 1980;52:237– 42. Emmert C, Rassler J, Kohler U. Survival and quality of life after percutaneous nephrostomy for malignant ureteric obstruction in patients with terminal cervical cancer. Arch Gynecol Obstet 1997;259:147–51. O’Toole DM, Golden AM. Evaluating cancer patients for rehabilitation. West J Med 1991;155:384 –7. Enck RE. Prognostication of survival in hospice care. Am J Hosp Palliat Care 1990;7:11–3. Culkin D, Wheeler JS, Marsans R, Nam SI, Canning JR. Percutaneous nephrostomy for palliation of metastatic ureteral obstruction. Urology 1987;30:229 –31. Bodner D, Kursh E, Resnick M. Palliative nephrostomy for relief of ureteral obstruction secondary to malignancy. Urology 1984;24:8 –10. Huebner WA, Plas EG, Porpaczy P. Hydronephrosis in malignant tumors: rationale and efficiency of endo-urological diversion. Eur J Surg Oncol 1993;19:27–32. Evans PA, Nisbet AP, Saxton HM. Anterograde ureteral stents in malignant disease. J Intervent Radiol 1988;3:9 –12.