Annals of Surgical Oncology, 6(8):732738

Published by Lippincott Williams & Wilkins 1999 The Society of Surgical Oncology, Inc.

Urinary Diversion After Total Pelvic Exenteration for

Rectal Cancer
Paul Russo, MD, Bipin Ravindran, BA, Jared Katz, BA, Philip Paty, MD,
Jose Guillem, MD, and Alfred M. Cohen, MD

Background: Total cystectomy is indicated for the treatment of bulky primary rectal cancers as
well as previously treated, locally recurrent tumors that invade the bladder, prostate, seminal vesicle,
or urethra. We review a 10-year Memorial Sloan-Kettering Cancer Center experience with urinary
diversion in this setting.
Methods: Between April 1988 and June 1998, 47 patients underwent urinary diversion during a
total pelvic exenteration for rectal cancer. Charts and operative records were reviewed to determine
pathological findings, short-term and long-term urological complications, and survival.
Results: Forty-seven patients (25 males and 22 females; median age, 62 years; age range, 2779
years) were included. Sixteen (34%) patients underwent cystectomy for a primary rectal tumor
(including 1 for rectal sarcoma and 1 for synchronous invasive bladder cancer), and 31 (66%)
patients underwent surgery for a locally recurrent rectal cancer. Thirty (64%) patients underwent
preoperative, 18 (38%) underwent intraoperative, and 11 (23%) underwent postoperative radiotherapy. Twenty-six (55%) patients received preoperative and 16 (34%) underwent postoperative
chemotherapy. Two patients had continent ileal cecal reservoirs, 1 a colonic conduit, and the
remaining 45 had ileal conduits. The tumor invaded the bladder in 24 (51%) patients, the prostate
in 5 (11%) patients, and the seminal vesicle in 5 (11%) patients. Complete resection was achieved
in 42 (89%) patients. There were a total of eight complications in eight (17%) patients. There were
three early complications, two of which were ileoureteral anastomotic leaks, one managed by
reoperation, the second by percutaneous drainage, and one moderate hydronephrosis managed
expectantly. There were five late complications; three patients had ureteral stricture/stenosis, leading
to nephrectomy in one patient and percutaneous stenting in two patients. Two patients developed
late hydronephrosis, so far managed expectantly. There was one perioperative death. After a median
follow-up of 16.83 months, 20 patients were dead of the disease, 6 were alive with disease
recurrence, 2 were dead of other causes, and 19 had no evidence of disease. Three-year actuarial
disease-specific survival was 34%.
Conclusions: Complete resection of bulky primary or locally recurrent rectal cancer can be
performed with acceptable urological morbidity. Complete resection was obtained in 89% of
patients, with 72% having urological organ invasion. Overall urological complications of 17% are
acceptably low despite intensive perioperative radiation and chemotherapy. Disease-specific survival in these patients remains limited.
Key Words: Urinary diversionRectal cancerPelvic exenteration.

In 1999, there will be an estimated 130,000 new cases

of colorectal carcinoma in the United States, leading to

57,000 deaths.1 Colorectal carcinoma comprises 10% of

all cancer deaths, making it the second leading cause of
death because of cancer in the United States. Despite
improvements in preoperative imaging with abdominal
and pelvic computed tomography (CT) and magnetic
resonance imaging (MRI), as well as the preoperative
treatment strategies including combined radiotherapy
and chemotherapy,2 local recurrence in rectal cancer
remains a common form of treatment failure occurring in
approximately 20% of cases. Local pelvic recurrence
usually occurs within 2 years of the primary resection;

Received April 23, 1999; accepted August 18, 1999.

From the Department of Surgery, Urology (PR, BR, JK) and Colorectal (PP, JG, AMC) Services, Memorial Sloan-Kettering Cancer
Center, New York, New York.
Presented at the 52nd Annual Meeting of Society of Surgical Oncology, Orlando, Florida, March 4 7, 1999.
Address correspondence and reprint requests to: Paul Russo, MD,
Urology Service, Memorial Sloan-Kettering Cancer Center, 1275 York
Avenue, New York, NY 10021; Fax: 212-717-3175; E-mail:
[email protected]

732

URINARY DIVERSION AFTER EXENTERATION

and, in many cases, patients remain free of systemic
nodal, hepatic, or pulmonary metastases. Combinations
of systemic chemotherapy and radiotherapy are effective
treatments for advanced colorectal cancers but are curative in less than 5% of patients, making surgical salvage
the major therapeutic strategy. Factors associated with an
increased risk for pelvic recurrence include tumors between 0 and 11 cm from the anal verge and an increasing
Dukes stage.39
The involvement by a local tumor recurrence, or, less
commonly, a bulky, primary nonmetastatic tumor of
adjacent urological organs including bladder, prostate,
and seminal vesicles, suggests that an exenterative operation could salvage the patient. In the late 1940s and
early 1950s, surgical investigators such as Brunschwig10
and Bricker11 reported their initial experiences with salvage pelvic exenteration for locally recurrent pelvic cancers. In these heroic operations, significant perioperative
mortality (10%) was experienced, and, for those patients
who survived the procedure, 5-year survival rates of less
than 30% were observed. Relapse after salvage exenteration was associated with survival of less than 1 year.
During this initial experience, the urinary diversions
ranged from cutaneous ureterostomy and wet colostomy
to cecal reservoirs and, finally, the durable ileal conduit
popularized by Bricker.11 Increasing surgical experience
with total pelvic exenteration, coupled with improvements in perioperative care, blood transfusions, and anesthesia techniques, substantially decreased the associated surgical morbidity, allowing this form of salvage
surgery to assume its present place in the treatment of
recurrent rectal cancer.12,13
Despite four decades of experience with total pelvic
exenteration for recurrent rectal cancer, as well as improvements in patient selection because of the development of CT and MRI scanning, this salvage operation
remains a challenging procedure, with approximately
25% to 35% long-term survivors reported.9 Although
there are no absolute guidelines for patient selection for
total pelvic exenteration for rectal cancer, it has been our
practice, in general, to limit this operation to patients
without evidence of extrapelvic cancer. The potential
perioperative morbidity and anticipated decline in quality of life after such procedures are risks worth taking if
there is a 25% to 30% chance of long-term survival. The
opportunity to perform such procedures in an intraoperative brachytherapy suite increases the enthusiasm when
preoperative imaging studies indicate the possibility of a
close surgical margin; however, the long-term benefits of
this approach have not been confirmed. In unusual cases
of unremitting and crippling local symptoms of bladder
or prostatic invasion, even in the presence of metastatic

733

disease, a palliative pelvic exenteration can be offered to

the patient if the medical condition of the patient can
withstand such a procedure. For bulky and unresectable
local tumors involving the bladder, formal urinary diversion is generally no longer done. The main reason for this
is the evolution of the endourological techniques of percutaneous nephrostomy. Patients with this degree of malignant urinary tract involvement rarely survive more
than 1 year, and major surgical urinary diversion is not
warranted. To qualify for a total pelvic exenteration for
recurrent rectal cancer, patients must be medically fit,
with a performance status of more than 90%, which will
increase their likelihood of surviving the perioperative
period.9,14
The purpose of the current study was to assess the
urological complications related to urinary diversion at
the time of total pelvic exenteration for colorectal cancer
and to describe the urological surgeons role in these
procedures.
PATIENTS AND METHODS
All patient urinary diversions at the time of total pelvic
exenteration for rectal cancer, at Memorial Sloan-Kettering Cancer Center, were reviewed between April 1988
and June 1998. Charts, operative records, pathological
findings, short- and long-term urological complications,
and survival were assessed. Data were entered in a clinical database (SPSS, Chicago, IL), and the Kaplan-Meier
method was used to estimate disease-free survival.
RESULTS
A total of 47 patients (25 males and 22 females;
median age, 62 years; age range, 2779 years) underwent
urinary diversion at the time of total pelvic exenteration.
There were 16 (34%) patients who underwent cystectomy for the treatment of bulky primary rectal tumor
(including 1 for rectal sarcoma and 1 for synchronous
bladder carcinoma) and 31 (66%) patients who underwent surgery for a locally recurrent rectal cancer. Thirty
(64%) patients underwent preoperative radiotherapy, 18
(38%) patients underwent intraoperative radiotherapy,15
and 11 (23%) patients underwent postoperative radiotherapy. Twenty-six (55%) patients received preoperative systemic chemotherapy and 16 (34%) patients underwent postoperative systemic chemotherapy (Table 1).
The urinary diversions used were as follows: A total of
44 (94%) patients had ileal conduits,11 2 (4%) patients
had ileal cecal continent reservoirs (Indiana Pouch),16
and 1 (2%) patient had a colonic conduit.17,18 In 34
(72%) patients, a urological organ was invaded by the
Ann Surg Oncol, Vol. 6, No. 8, 1999

734

PAUL RUSSO ET AL.

TABLE 1. Urinary diversion after exenteration: patient characteristics
Chemotherapy
Patient groups

Radiotherapy

Number (%)

Preop

Postop

Preop

IORT

Postop

Node negative

28 (60)

15

15

Node positive

5 (10)

No nodes seen

14 (30)

11

Status
5 AWD
5 DOD
18 NED
1 AWD
4 DOD
3 AWD
1 DOC
4 DOD
6 NED

Preop, preoperatively; Postop, postoperatively; IORT, intraoperative radiation therapy; AWD, alive with
disease; DOD, dead of the disease; NED, no evidence of disease; DOC, dead of other causes.

tumor. This included the bladder in 24 (51%) patients,

the prostate in 5 (11%) patients, and the seminal vesicle
in 5 (11%) patients. A complete resection of the tumor
with negative surgical margins was achieved in 42 (89%)
patients.
Urological complications were seen in 8 (17%) of 47
patients. There were three early complications, two of
which were ileoureteral anastomotic leaks. One required
immediate open exploration and primary repair of a
partial dehiscence of the ileoureteral anastomosis. The
second patient underwent percutaneous nephrostomy
proximal urinary diversion and percutaneous drainage of
a sequestered intra-abdominal urinoma. Later, an antegrade percutaneous nephroureteral stent was placed

through the nephrostomy into the ileal conduit. The final

patient, with moderate hydronephrosis, was managed
expectantly.
There were five late complications; three patients had
ureteral stricture/stenosis, leading to nephrectomy in one
patient and percutaneous stenting in two patients. Two
patients developed late hydronephrosis, so far managed
expectantly. In our experience, most significant complications were clinically detected within 12 months of the
operation (Fig. 1).
After a median follow-up of 16.83 months, 20 patients
were dead of the disease, 6 were alive with disease
recurrence, 2 were dead of other causes, and 19 were
alive without clinical evidence of disease. Three-year

Fig. 1. Urological complications from

time of urinary diversion after exenteration for rectal cancer.

Ann Surg Oncol, Vol. 6, No. 8, 1999

URINARY DIVERSION AFTER EXENTERATION

actuarial disease-specific survival was 34%. The median
survival was 33.43 months. There was one perioperative
death (Fig. 2).
DISCUSSION
The post-World War II development of surgical technique in total pelvic exenteration occurred in parallel
with marked improvements in anesthesiology, blood
transfusion technology, antibiotic therapy, and perioperative nutritional support.12 This major salvage cancer
operation in selected, heavily pretreated patients, remains a surgical option today. Patients with pretreated,
locally recurrent rectal cancer, with the disease restricted
to the pelvis with direct invasion of bladder, prostate,
seminal vesicals, and urethra, are considered candidates
for exenterative surgery. Often transaxial imaging by CT
or MRI scanning demonstrates a close proximity between the recurrent tumor and the urological organ, with
direct invasion not clearly demonstrated. Cystoendoscopy may confirm direct invasion or secondary edema in
the bladder trigone, usually from tumor infiltration of the
bladder perivesical soft tissues and bladder wall. The
search for clinically occult metastatic disease by chest
CT scanning and, recently, by positron emission tomographic scanning, makes the final recommendation for
exenteration with curative intent easier to propose to the
patient if those tests are negative.

735

In this current series, 16 patients (34%) underwent

total pelvic exenteration for primary rectal cancer, including 1 patient with a perirectal sarcoma and another
with a simultaneous muscle invasive bladder cancer. In
these patients, the indications for exenteration were
based on direct urological organ invasion, the absence of
systemic disease, and the absence of significant comorbid medical conditions that would make the operation
unsafe.
The urinary diversion component of this operation has
always been a particularly critical one because of the
potential for urinary leaks and obstruction, with morbid,
and potentially fatal, effects. Many centers use two surgical teams, allowing a fresh urological reconstructive
team to enter the operating room, after the completion of
the exenteration and possibly intraoperative radiotherapy, to complete the urinary diversion in an expeditious
manner. The urological surgeon plays an important role
in preoperative assessment, as well as intraoperative
care, of patients undergoing resection of recurrent rectal
cancer. In general, prospective candidates for pelvic exenteration are referred for outpatient cystoendoscopy to
determine if there is direct bladder or prostatic urethral
invasion. At the time of that visit, urinary diversion,
either continent or ileal conduit, is presented to the
patient and the family by the urologist and the nursing
staff. At the time of the operation, the colorectal surgeon
often asks the urological surgeon to place ureteral stents

Fig. 2. Disease-specific survival from

time of pelvic exenteration and urinary
diversion for rectal cancer.

Ann Surg Oncol, Vol. 6, No. 8, 1999

736

PAUL RUSSO ET AL.

TABLE 2. Pelvic exenteration series for rectal cancer
Series

No. of Patients Perioperative mortality (%)

Ledesma (1981)
Boey (1982)
Jakowatz (1985)
Wanebo (1987)
Yeung (1993)

30
49
104
28
50

3 (10)
13 (27)
12 (12)
3 (12)
7 (14)

Brophy (1994)
MSKCC (1999)
Total

35
47
343

16 (47)
1 (2)
57 (17)

Median survival

Survival at 5 years (%)

N/A
N/A
N/A
36 mo
19 mo (curative)
10 mo (palliative)
20 mo
33 mo
24 mo (avg)

15 (50)
19 (39)
28 (27)
6 (21)
3 (6)
N/A
9 (19)
80 (23)

N/A, not available; MSKCC, Memorial Sloan-Kettering Cancer Center.

to aid in identification of previously irradiated ureters

and to assist in the deep pelvic dissection. In this series,
72% of patients had a urological organ invaded by rectal
cancer and 28% did not but still underwent exenteration
to clear the tumor from adherent tissues and to provide a
surgical margin. In such uncertain cases, in the operating
room, every effort is made to avoid exenteration, by
excision of seminal vesicals or resection of the posterior
capsule of the prostate, or by performing a partial cystectomy with ureteral reimplant. The ultimate decision to
remove the bladder is made by the colorectal surgeon to
secure a complete surgical resection.
After radiotherapy, the pelvic ureters may be surrounded by a dense fibrotic reaction. Care must be used
during the mobilization of the ureters not to dissect too
closely to them and, hence, cause a devascularization
that may lead later to stenosis and ureteral obstruction. If
ureteral length is short, a Wallace-type anastomosis
(spatulated ureters are sewn to each other and then to the
butt end of the conduit) can be performed. If there is
adequate ureteral length, then a standard Bricker anastomosis is done.19 After the ureteral ileal anastomosis is
complete, every attempt should be made to retroperitonealize the ureter so that if a small urinary leak occurs, it
will not allow for free intraperitoneal extravasation of
urine. A sequestered urinoma can be drained percutaneously in conjunction with proximal urinary diversion
with a percutaneous nephrostomy and, thus, avoid reoperation. One patient in the current series with an early
urinary leak was managed in this way.
The final decision concerning the type of urinary diversion is made by the urologist and colorectal surgeon
and takes into consideration the operative findings (i.e.,
complete resection versus clinical positive margin, positive nodes) that may worsen the immediate prognosis as
well as the condition of the patient. If the operation is
long and the patients condition is not stable, an expeditious ileal conduit should be performed. If a continent
urinary reservoir is contemplated, sufficient remnant coAnn Surg Oncol, Vol. 6, No. 8, 1999

lon is required to isolate the ileocecal segment and ascending colon plus 10 to 12 cm of distal ileum that is
required to form the catheterizable efferent limb. If, after
such mobilization, sufficient colon does not exist,
thereby leaving the patient with watery stool through the
colostomy, then a standard ileal conduit should be performed. In young patients with adequate remnant colon,
a continent reservoir should be performed because there
is a substantial improvement in the patients mental
quality of life.20 Despite our willingness to perform such
a continent diversion, only two patients in this series
qualified for the procedure.
After exenteration, the pelvis is completely empty and
there is a tendency for intestine, including the newly
created urinary diversion, to drop into the pelvis and
possibly obstruct the ureters. Care must be used to tack
the conduit to the cut peritoneal reflections from the butt
end at the pelvic brim all the way to the stoma site, which
is usually marked by an enteral stomal therapist before
the operation, in the right lower quadrant. If an Indiana
pouch is performed, the efferent limb (that used for
catheterization) is usually brought out through the umbilicus for easy and direct patient access. If there is
omentum available, it should be brought into the pelvis
to prevent adhesions from the bowel to the resection
sites, to fill the dead space, and to decrease the possibility
of abscess formation. If postoperative radiotherapy is
planned, a synthetic mesh can be placed to exclude the
bowel from the pelvis. All rents in the intestinal mesenteries should be closed to prevent internal herniation and
later bowel obstruction. Drains are placed in the pelvis to
drain lymphatic fluid and serum that pools there in the
immediate postoperative period.
If there is excessive pelvic drainage, the fluid should
be sent for blood urea nitrogen/creatinine to rule out the
possibility of an early urinary leak. In general, we recommend stenting all such urinary diversions with 8.5-F
urinary diversion stents, which are carried through the
ileal ureteral anastomosis to the kidneys bilaterally and

URINARY DIVERSION AFTER EXENTERATION

exit the ileostomy stoma or efferent limb of the Indiana
pouch. In the recovery room, an abdominal x-ray is
performed to confirm that the stents are in the kidney
bilaterally. In patients who were heavily irradiated, we
leave the stents in for 4 to 6 weeks. For patients not
previously irradiated and with an uncomplicated postoperative course, the stents are left for 10 to 14 days and
removed after a single dose of intravenous gentamicin, to
protect from any Gram-negative organisms that may
have colonized them.
The creation of the ileal conduit stoma is a critically
important portion of this operation and, if not performed
properly, can hamper the postoperative recovery by the
constant leak of urine around an ill-fitting appliance.
Care must be used, to allow for an adequate fascial
incision for easy placement of the conduit through the
fascia, to prevent venous congestion. After such long
operations, the bowel can have a great deal of edema, and
every attempt should be made to achieve sufficient
length and satisfactory eversion of the ileal stoma. Careful inspection, after the suturing of the ileal conduit to the
skin, should guarantee that there is no tension on small
intestinal mesentery, which causes ischemia. Stents emanating from the stoma should be sutured to the skin
separately.
From the urological perspective, postoperative follow-up, in conjunction with the primary colorectal
surgeon, should include a check for stomal stenosis, a
chemistry work-up for serum blood urea nitrogen/
creatinine, and upper tract imaging studies with CT,
intravenous pyelography, or ultrasonography to detect
ureteral obstruction.
In this contemporary series of urinary diversion after
total pelvic exenteration in patients with rectal cancer, a
urological organ was invaded in 72% of patients, which
confirms the accuracy of preoperative planning. A complete resection was possible in 89% of patients. Urological complications were seen in 8 (17%) of 47 patients.
There were three early complications (two ileoureteral
anastomotic leaks and one moderate hydronephrosis).
One leak required a repeat operation, whereas the other
was managed endourologically, and the patient with
moderate hydronephrosis was managed expectantly.
There were five late complications; three patients had
ureteral stricture/stenosis, leading to nephrectomy in one
patient and percutaneous stenting in two patients. Two
patients developed late hydronephrosis, so far managed
expectantly. This overall complication rate of 17% compares favorably with complications observed in patients
undergoing urinary diversion after cystectomy for invasive bladder cancer, in which ureteral leaks occur in 3%
to 14% of cases and ureteral obstruction occurs in 3% to

737

8% of cases.19 Our results in this contemporary series of

pelvic exenterations with urinary diversion suggest that
the urological portion of this procedure can be performed
with acceptable morbidity despite previous and subsequent chemotherapy and radiotherapy.
Unfortunately, despite improvements in operative
techniques and perioperative care, long-term survival of
less than 30% for patients undergoing this salvage operation does not represent a significant difference in the
outcome during the last 30 years9 (Table 2). These results
with exenteration correlate closely with our overall experience with surgical salvage of recurrent rectal carcinoma after previous curative resection in which 5-year
survival is 24%.21 A search for newer adjunctive systemic chemotherapies is required to improve the overall
survival of patients undergoing complete pelvic exenteration for colorectal cancer.

REFERENCES
1. Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics,
1999. CA Cancer J Clin 1999;49:8 31.
2. Cohen AM, Minsky BD, Schilsky RL. Cancer of the rectum. In:
DeVita VT, Hellman S, Rosenberg SA, eds. Cancer: Principles
and Practices of Oncology. 5th ed. Philadelphia: LippincottRaven, 1997:11971233.
3. Pilipshen SJ, Heilweil M, Quan SH, Sternberg SS, Enker WE.
Patterns of pelvic recurrence following definitive resections of
rectal cancer. Cancer 1984;53:1354 62.
4. Sagar PM, Pemberton JH. Surgical management of locally recurrent rectal cancer. Br J Surg 1996;83:293304.
5. Welch JP, Donaldson GA. Detection and treatment of recurrent
cancer of the colon and rectum. Am J Surg 1978;135:50511.
6. Gunderson LI, Sosin H. Areas of failure found at reoperation
(second or symptomatic look) following curative surgery for
adenocarcinoma of the rectum. Cancer 1974;34:1278 92.
7. Rao AR, Kagan AR, Chan PM, Gilbert HA, Nussbaum H, Huntz
BL. Patterns of recurrence following curative resection alone for
adenocarcinoma of the rectum and sigmoid colon. Cancer 1981;
48:14925.
8. Rich T, Gunderson LI, Lew R, Galdibini JJ, Cohen AM, Donaldson G. Patterns of recurrence of rectal cancer after potentially
curative surgery. Cancer 1983;52:131729.
9. Petros JG, Lopez MJ. Pelvic exenteration for carcinoma of the
colon and rectum. Surg Clin North Am 1994;3:257 66.
10. Brunschwig A. Complete excision of pelvic viscera for advanced
carcinoma: a one-stage abdominoperineal operation with endocolostomy and bilateral ureteral reimplantation into the colon above
the colostomy. Cancer 1948;1:177 83.
11. Bricker EM. Bladder substitution after pelvic evisceration. Surg
Clin North Am 1950;30:151121.
12. Bricker EM. Evolution of radical pelvic surgery. Surg Clin North
Am 1994;3:197203.
13. Cohen AM, Minsky BD. Aggressive surgical management of locally advanced primary and recurrent rectal cancer. Dis Colon
Rectum 1990;33:432 8.
14. Wanebo HJ, Gaker DL, Whitehill R, et al. Pelvic recurrence of
rectal cancer: options for curative resection. Ann Surg 1987;205:
48295.
15. Harrison LB, Enker WE, Anderson LL. High dose intraoperative
radiation therapy for colorectal cancer. Oncology 1995;9:737 41.

Ann Surg Oncol, Vol. 6, No. 8, 1999

738

PAUL RUSSO ET AL.

16. Rowland RG, Mitchell ME, Bihrle R, et al. Indiana continent

urinary reservoir. J Urol 1987;137:1136 9.
17. Hill JT, Ransley PG. The colonic conduit: a better method of
urinary diversion? Br J Urol 1983;55:629 31.
18. Ritchie JP. Sigmoid conduit urinary diversion. Urol Clin North Am
1986;13:22531.
19. Seigne J, McDougal WS. Urinary diversion. Surg Clin North Am
1994;3:30722.

Ann Surg Oncol, Vol. 6, No. 8, 1999

20. McGuire M, Grimaldi G, Groatas J, Russo P. The type of urinary

diversion after radical cystectomy significantly impacts on the patients quality of life. Presented at the 51st Annual Meeting of the
Society of Surgical Oncology, San Diego, CA, March 2125, 1998.
21. Salo JC, Paty PB, Guillem J, Minsky BD, Harrison LB, Cohen
AM. Surgical salvage of recurrent rectal carcinoma after curative resection: a 10 year experience. Ann Surg Oncol 1999;6:
17177.