Kenneth D. Hatch
Jonathan S. Berek
The first series of pelvic exenterations for gynecologic cancer was published in 1946 by Alexander Brunschwig (1). This initial report was of 22 patients, 5 of whom died of the operation itself. His original procedure included sewing both ureters into the colon, which was then brought out as a colostomy. Since these beginnings, there have been major improvements in the selection of patients, operative technique, blood product use, antibiotic availability, and intensive medical management.
The operation gained wider acceptance when Bricker (2) published his technique of isolating a loop of ileum, closing one end, anastomosing the two ureters to this end, andbringing the other out as a stoma. This eliminated the hyperchloremic acidosis and markedly diminished the recurrent pyelonephritis and renal failure that were experienced with the wet colostomy. The popularity of the Bricker ileal loop was aided by the development of watertight stomal appliances.
Failure of the small bowel anastomosis to heal because of radiation fibrosis in some patients led to the use of a segment of nonirradiated transverse colon for the conduit (3). Further reductions in bowel complications occurred with the use of surgical staplers, which also decreased the operative time, blood loss, and subsequent medical complications (4). Further refinements in the urinary diversion led to the continent urinary reservoir, which is described in Chapter 20.
As a higher percentage of patients became long-term survivors, the desire to improve quality of life led to reconstructive techniques for the vagina and the colon. Today, the patient undergoing pelvic exenteration may have a colonic J-pouch rectal anastomosis, vaginal reconstruction, and continent urinary diversion, allowing her to enjoy a near-normal quality of life without major alterations in her physical appearance.
The terminology of pelvic exenteration has changed as the operations have been tailored to remove the tumor and only those organs that are involved. The total exenterationperformed by Brunschwig included the bladder, uterus, vagina, anus, rectum, and sigmoid colon (1). It usually included a large perineal phase (Fig. 22.1). This would lead to a permanent colostomy and urinary stoma. Rutledge et al. (5) and Symmonds et al. (6) reported decreased morbidity and acceptable survival when performing anterior exenteration,which removed the uterus, bladder, and various amounts of the vagina (Fig. 22.2). Total pelvic exenteration with rectosigmoid anastomosis (supralevator) became possible with the development of circular staplers. The rectum is excised to within 2 to 3 cm of the anal canal, and the levator support of the anal canal and perineal body is preserved (Fig. 22.3). The posterior exenteration removes the uterus, vagina, and portions of the rectosigmoid and anus. It is rarely performed today. Vaginal reconstructive techniques are discussed inChapter 20.
Figure 22.1 Total pelvic exenteration with perineal phase. This operation includes removal of the bladder, uterus, vagina, anus, rectum, and sigmoid colon, as well as performance of a perineal phase.
The most common indication for pelvic exenteration is recurrent or persistent cancer of the cervix after radiation therapy. Some of the early series reported exenterations as primary therapy for stage IVA cervical cancer and cancer of the vulva with urethral, vaginal, or rectal invasion. With modern radiation therapy, the use of exenteration as primary therapy is uncommon.
Exenteration has also been used for endometrial cancer, vaginal carcinoma, rhabdomyosarcoma, and other, miscellaneous rare tumors whenever ultraradical central resection of the cancer is feasible and there is no evidence of systemic or lymphatic spread.
Patients with endometrial cancer have a high likelihood of spread beyond the pelvis and are in general poor candidates for exenterative surgery. The survival rate for highly selected patients with endometrial cancer undergoing exenteration is less than 20% at 5 years (7). To debulk ovarian cancer optimally, a modified posterior exenteration is often performed, which includes en bloc resection of the pelvic peritoneum, uterus, tubes, ovaries, and a segment of rectosigmoid. It usually preserves most of the rectum and allows for a low rectal anastomosis. Because there is ovarian cancer left behind, the procedure violates the principle that exenterative surgery is meant to be curative. In the treatment of ovarian cancer, modified exenteration is performed as part of a cytoreductive procedure and is followed by chemotherapy.
The medical evaluation begins with histologic confirmation that cancer is present. The patient should have no other potentially fatal disease, and her general medical condition must be adequate for a prolonged operative procedure (up to 8 hours) with considerable fluid shifts and blood loss.
Figure 22.2 Anterior pelvic exenteration. This operation includes removal of the bladder, uterus, and varying amounts of the vagina, depending on the extent of disease.
The search for metastatic disease is imperative. The physical examination should include careful palpation of the peripheral lymph nodes and fine-needle aspiration (FNA) cytologic analysis if any suspicious nodes are found. Particular attention should be paid to the groin and supraclavicular nodes. A random biopsy of nonsuspicious supraclavicular lymph nodes has been advocated but is not routinely practiced (8). A computed tomographic (CT) scan of the lungs detects disease missed on routine chest radiography. An abdominal and pelvic CT scan is mandatory to detect liver metastasis and enlarged paraaortic nodes. CT-directed FNA cytologic analysis of any abnormalities should be undertaken. CT scanning should not be relied on for determining resectability on the basis of apparent absence of fatty planes lateral to the tumor.
Magnetic resonance imaging (MRI) has been evaluated for preoperative assessment of candidates for pelvic exenteration (9). Twenty-three patients were evaluated before pelvic exenteration for presence and location of the recurrent tumor; tumor extension to the bladder, rectum, or pelvic sidewall; and presence and location of lymphadenopathy. In four patients (17.4%), the MRI was falsely positive for pelvic sidewall infiltration, and in one patient (4.3%), it was falsely negative.
Lai and colleagues in Taipei evaluated the PET scan for the restaging of cervical carcinoma at the time of first recurrence (10). Forty patients had a PET scan, together with computed tomography and/or magnetic resonance imaging. Twenty-two patients (55%) had their treatment modified due to the PET findings. PET was significantly superior to CT/MRI (sensitivity = 92% vs. 60%; p <0.0001) in identifying metastatic lesions. In addition, when compared with an earlier cohort of patients who did not undergo restaging with PET, there was a significantly better 2-year overall survival (72% vs. 36%; p = 0.02).
Husain et al. (11) used FDG PET to determine metastatic disease prior to pelvic exenteration or radical resection in 27 patients with recurrent cervical or vaginal cancers. They found that FDG PET had a high sensitivity (100%), and a specificity of 73% in detecting sites of extra-pelvic metastasis. Chung (12) performed PET CT scans in 52 patients suspected of having recurrence of a cervical cancer. Twenty-eight of 32 patients (87.5%) with positive scans were proven to have recurrent disease. Seventeen of 20 patients (85%) with negative PET CT scans had no evidence of disease. The sensitivity was 90.2% and specificity 81.0%.
Figure 22.3 Supralevator total pelvic exenteration. This operation removes the uterus, vagina, and portions of the rectosigmoid colon with colonic reanastamosis.
The high sensitivity of the PET CT scan may allow the clinician to perform needle-guided biopsies or minimally invasive procedures to confirm the metastatic disease and avoid an aborted exenteration attempt. On the other hand, the 70% to 80% specificity requires that patients undergo exploration if the minimally invasive techniques do not document metastatic disease. Thus, the PET scan is an important addition to the preoperative investigation of a candidate for pelvic exenteration.
Extension of the tumor to the pelvic sidewall is a contraindication to exenteration; however, this may be difficult for even the most experienced examiner to determine because of radiation fibrosis. If any question of resectability arises, the patient should be given the benefit of exploratory laparotomy and parametrial biopsies.
Laparoscopy has been described as useful in the assessment of lymph nodes as well as the resectability of disease in the pelvis. In the hands of highly skilled laparoscopic surgeons, this may be an option (13).
The clinical triad of unilateral leg edema, sciatic pain, and ureteral obstruction is nearly always indicative of unresectable cancer on the posterolateral pelvic sidewall.
Despite careful preoperative evaluation, there is approximately a 30% risk that patients will undergo exploratory laparotomy and be judged unsuitable candidates for exenteration.Miller et al. (14) reported that 111 of 394 patients (28.2%) undergoing exploration at the University of Texas M. D. Anderson Cancer Center had findings that led to abortion of the exenterative procedure. Reasons for aborting the procedure were peritoneal disease in 49 patients (44%); nodal metastasis in 45 (40%); parametrial fixation in 15 (13%); and hepatic or bowel involvement in 5 (4.5%).
Preoperative Patient Preparation
The patient must be counseled extensively concerning the seriousness of the operation. She should be prepared to spend several days in the intensive care unit and have a prolonged hospitalization of up to several weeks. She must understand that her sexual functioning will be permanently altered and that she may have one or two stomas. In addition, there is no guarantee of cure. The most difficult subject to broach is the possibility that she may have unresectable disease and that the procedure will need to be aborted.
At least a mechanical bowel preparation is given (see Chapter 20, Table 20.1). Intravenous fluids are started at the time of the bowel preparation to avoid dehydration. The patient should have the stoma sites marked by the ostomy team, and management of the ostomies should be discussed. If the patient is severely malnourished, total parenteral nutrition (TPN) should be started in advance of surgery. Because these patients may not have significant oral caloric intake for a week or longer, postoperative TPN is commonly given.
The patient is placed in the low lithotomy position using stirrups that support the hips, knees, and thighs that can be repositioned during the surgery. This position allows the operators to perform the abdominal and perineal phases of the operation simultaneously. Intermittent pneumatic compression devices are applied to the calves as prophylaxis against deep venous thrombosis. Combined epidural and general anesthesia allow the epidural to be maintained after surgery for better pain control while keeping the patient alert and able to maintain better respiratory function.
The abdominal incision is made in the midline and should be adequate for exploration of the upper abdomen as well as for performing the pelvic surgery. The liver and omentum should be palpated carefully. The rest of the abdomen is explored, and the paraaortic nodes are palpated. Both the right and left paraaortic nodes are sent for frozen-section analysis. If these are negative, pelvic spaces are opened by dividing the round ligament at the pelvic sidewall. The prevesical, paravesical, pararectal, and presacral spaces are all developed and the ligaments are evaluated for resectability. Enlarged or suspicious pelvic lymph nodes should be removed and sent for frozen-section evaluation. More than one positive pelvic node, positive paraaortic nodes, peritoneal breakthrough of tumor, or tumor implants in the abdomen or pelvis should lead to abandonment of the operation.
The procedure begins by ligating the internal iliac artery just after it crosses the internal iliac vein. This sacrifices the uterine artery, vesical artery, and obliterated umbilical artery. The remainder of the hypogastric artery is left intact. It carries the internal pudendal and inferior hemorrhoidal arteries that are important in maintaining the blood supply to the anal canal and lower rectum, where a potential low rectal anastomosis will be performed. The obturator artery should also be preserved because it is the major blood supply to the gracilis muscle, and a gracilis neovagina may be planned. The cardinal ligaments are divided at the sidewall and the broad attachments of the rectum to the sacrum are divided. The vaginal attachments to the tendinous arch are divided. The vaginal arteries and vein are located at the lateral margin of this pedicle. The specimen is completely mobilized, and the penetration of the rectum and vagina through the pubococcygeal muscle can be identified. Various sites for ligation of pubococcygeal muscle for total exenteration versus anterior exenteration are identified (Fig. 22.4).
Anterior exenteration may be planned for lesions confined to the cervix and the anterior upper vagina. The uterus, cervix, bladder, urethra, and anterior vagina are removed, and the posterior vagina and rectum are preserved. Intraoperative bimanual palpation helps select the appropriate patient. The peritoneal reflection of the cul-de-sac can be incised and the rectum dropped away with a finger in the rectum and a finger in the vagina to ensure that the tumor is adequately resected. One surgeon conducts the perineal phase and the other surgeon conducts the abdominal phase.
Figure 22.4 Cross-sectional diagram of pelvis showing lines of excision through the pubococcygeus muscle for anterior and total exenterations.
The perineal incision includes the urethral meatus and the anterior vagina. A long curved clamp is placed beneath the pubis and directed caudad and anterior to the urethra. Another clamp is placed lateral to the pubourethral ligaments and directed out under the symphysis pubis, first at 2 o'clock and then at 10 o'clock. This isolates the right and left pubourethral ligaments, which can be clamped, divided, and ligated. The posterior vaginal incision is made under direct vision from below, insuring a surgical margin of at least 4 cm. The specimen is then ready to be removed. Hemostasis is provided by suture ligatures, and a pelvic pack is placed while the urinary diversion is performed. The omentum is mobilized and brought down the left paracolic gutter into the pelvis. It is used to cover the denuded area of the rectum and may provide a receptacle for neovaginal construction by a split-thickness skin graft. The omentum is sewn to the posterior vaginal mucosa over the rectum and to the pelvic sidewalls. The skin is harvested and placed around a sterile mold. It is then placed into the cylinder formed by the omentum. If there is not enough omentum, the bulbocavernosus flaps may be used (15).
Supralevator Total Exenteration
Supralevator total exenteration with low rectal anastomosis for patients whose disease extends off the cervix on to the posterior vagina should have the segment of rectum removeden bloc with the specimen. This usually entails resection of the rectum to within 6 cm of the anal verge (Fig. 22.3). To remove the specimen, it is best to divide the sigmoid with the stapler to allow for easier exposure to the presacral space. The space is developed in the median avascular plane down to where the rectum exits between the levator muscles. The superior rectal and middle rectal arteries are sacrificed. The incision in the vaginal mucosa is 1 to 2 cm inside the hymenal ring. The supralevator attachments of the bladder, urethra, and vagina are divided, leaving the specimen attached only by the rectum. The hand is placed to encircle the rectum, and traction is placed cephalad. The thoracoabdominal stapling device is then placed across the lower rectum with a 4-cm margin. Preservation of some of the lower rectum is desirable for the patient to have better continence and stool storage functions. The specimen is then removed from the field. Hemostasis is provided, and a pack is placed while the urinary diversion is performed. The left colon is mobilized, sacrificing the sigmoidal arteries and leaving the inferior mesenteric vessels. The sigmoid is then used for a colonic J-pouch, and a low anastomosis is performed using the stapling device. The omentum should be mobilized and brought down to reinforce the stapled anastomosis and to cover the denuded area in the pelvis.
Because there is more of the vagina removed in this operation than in the anterior exenteration, the omentum may not be satisfactory for a split-thickness skin-grafted neovagina.The patient is more likely to require a myocutaneous graft from the gracilis muscles in the medial thigh or the rectus abdominis muscle. Because of the smaller opening in the vaginal introitus, the rectus abdominis myocutaneous graft is preferred.
Total Exenteration with Perineal Phase
If the tumor has extended down the lower vagina and involves the levator muscles, it is necessary to remove them for a chance of cure. The specimen is mobilized from above in a way similar to that described in the preceding operations (Fig. 22.5). The perineal incision is made around the anus and as far lateral as necessary to gain clearance from the tumor. The anococcygeal and pubococcygeal muscles are divided as necessary for margins. This leaves a large pelvic and perineal defect, which is best filled with bilateral gracilis myocutaneous flaps. Alternatively, the rectus abdominis muscle can be used. The omentum is harvested and used as a pedicle flap to provide additional blood supply and a barrier to bowel adhesions. A permanent colostomy is placed, and urinary diversion is undertaken.
Figure 22.5 A surgically removed specimen from a total pelvic exenteration. Note the bladder above with a fistulous tract to the vagina, and the rectum below.
Posterior exenterations are now rarely performed except occasionally for cancer of the vulva involving the rectum after radiation therapy. When cervical cancer recurs after radiation therapy, even if it is confined to the posterior vagina and rectum, the distal ureters, bladder, and urethra should be removed to avoid the morbidity and mortality of urinary tract fistulae, stenosis, and denervation.
Low Rectal Anastomosis During Pelvic Exenteration
The introduction of the end-to-end circular stapling device has greatly facilitated and popularized the performance of low rectal resection and reanastomosis for a variety of general surgical and gynecologic malignancies. The automatic circular stapling device has many advantages over the traditional hand-sewn anastomosis. It allows use of a shorter anal or rectal stump, causes less tissue inflammation, creates a higher collagen content, and facilitates faster healing (16). These are most likely due to a better blood supply at the stapled anastomosis compared with a sutured anastomosis (17,18).
The anastomotic leak rate for low rectal anastomosis is reported to be less than 8% (19,20) in patients without previous radiation. The most important variables in the anastomotic leak rate are the distance from the anus to the anastomosis, the vascularity of the cut ends, the tension on the anastomotic line, and the elimination of the pelvic cavity (21,22). Graffner et al. (22) showed in a randomized series that the anastomotic leak rate in previously unirradiated patients is the same for those patients with diverting colostomies as for those without.
There are a few reports in the gynecologic literature concerning the low rectal anastomosis in women with previous pelvic radiation therapy. Berek et al. (23) reported 11 patients with no anastomotic leaks, and 7 of these patients had their bowel continuity reestablished with the end-to-end stapling device. Harris and Wheeless (24) reported 17 patients with a 12.4% anastomotic leak rate and a 12.4% stricture rate. Both groups advised using a diverting colostomy in the previously radiated patient. Hatch et al. (21) reported using a diverting colostomy in 12 of 31 previously irradiated patients. Six patients (50%) later had non-cancer-related rectovaginal fistulae requiring permanent colostomy. Of the 19 patients without protective colostomies, 6 (31.6%) had non-cancer-related rectovaginal fistulae. In the series of Hatch et al. (21), the most important factor in fistula prevention was the use of an omental wrap to bring a new blood supply to the irradiated pelvis. For patients who did not have a diverting colostomy, total parenteral nutrition was used for 14 to 21 days.
Mirhashemi et al. (20) conducted a risk factor analysis of 77 patients at the University of Miami who had low rectal anastomosis after exenterative surgery. The indications for the surgery were recurrent cervical cancer (33); ovarian cancer (27); recurrent vaginal cancer (7); recurrent endometrial cancer (4); colon cancer (3); and endometriosis (3). Previous radiation was the major factor in anastomotic leak rate, with 35% of the irradiated patients and 7.5% of the nonirradiated patients having a leak or a fistula. Protective colostomy did not make a difference. Of the 40 patients who had total pelvic exenteration with low rectal anastomosis, 36 had received pelvic radiation therapy. A protective colostomy was used in 12 of these patients, and 6 developed fistulae. Of the 24 who did not have protective colostomies, 6 developed an anastomotic leak or fistulae. Only 1 of the 37 patients who had posterior exenteration and low rectal anastomosis had previous radiation therapy, and this patient had an anastomotic leak. Of the remaining 36 patients, 3 had an anastomotic leak. Protective colostomies were not used on any of these patients (20).
Removal of the rectum alters the physiology of stool storage and defecation. The rectum is the reservoir for the collection of feces and transmits impulses to the sensory nerves to initiate the urge to defecate. Inhibitory reflexes from the rectum to the anus are necessary while the rectum is filling to ensure continence. After resection of most of the rectum, reservoir capacity, sensation, and recto-anal reflex are significantly altered (25). The most important factor in restoring normal bowel function is restoration of the reservoir capacity. Capacity can be increased by preserving as much rectum as possible or by a colonic J-pouch. The length of rectum necessary for return to acceptable function is 6 cm or more (26,27). When the anastomosis is above 12 cm, there is little alteration of function (28).
Table 22.1 Randomized Comparison of Colonic J-Pouch versus Coloanal Anastamosis in 100 Patients
The colonic J-pouch has been popularized by colorectal surgeons to treat rectal cancer with low rectal resection. It has replaced coloanal anastomosis because of its superior results. Studies comparing colonic J-pouch with coloanal anastomosis have shown (i) a decreased anastomotic leak rate; (ii) a better continence rate; (iii) fewer stools per day; (iv) better control of urgency; and (v) better control of flatus (29,30,31,32). Prospective, randomized trials have confirmed the observational studies (33,34) (Table 22.1).
The most significant drawback to the colonic J-pouch is the inability of some patients to empty the pouch. This is most likely because of the length of the staple line used to construct the pouch. Hida et al. (35) prospectively randomized patients to a 5-cm versus a 10-cm pouch and found the 5-cm pouch to be superior for evacuation without compromising the other parameters (Table 22.2). Most authors report using a diverting colostomy when creating the colonic J-pouch, which has led to a decrease in the anastomotic leak rate.
Harris et al. have reported the long-term function of J-pouch anal anastomosis in 119 consecutive randomized patients with colorectal cancer from the Cleveland Clinic. Patients who had J-pouch versus coloanal anastomosis had significantly better continence scores at 5 to 9 years after surgery and fewer nocturnal bowel movements (36).
A new procedure called a coloplasty has been developed at the Cleveland Clinic to improve on the poor bowel function after either a coloanal anastomosis or a colonic J-pouch anastomosis (37) (see Chapter 20). In a randomized study of the three techniques, the coloplasty and colonic J-pouch patients had significantly more favorable compliance, reservoir volume, and fewer bowel movements per day than the straight anastomotic group. The advantage of the coloplasty was that it could be used in a narrow pelvis. This may apply to the female patient who is having vaginal reconstruction with myocutaneous graphs, where space for anastomosis is diminished (37).
Table 22.2 Randomized Study of 5-cm J-Pouch versus 10-cm J-Pouch
There are some important anatomic considerations for patients undergoing pelvic exenteration with a continent urinary diversion. The continent urinary diversion uses the right colic artery up to its anastomosis with the middle colic artery. A colonic J-pouch uses the sigmoidal and left colic vessels. Adequate mobilization of the descending and left colon requires mobilization of the splenic flexure and rotation of the left colon into the pelvis. If a diverting loop colostomy is performed, it may interrupt the vascular supply from the marginal artery of Drummond. Care must be taken to preserve this vascular supply so that the colonic J-pouch and the resultant colorectal anastomosis have an adequate blood supply.
Husain et al. reported the experience at Memorial Sloan-Kettering in 13 patients who had total pelvic exenteration and low rectal anastomosis with a continent urinary diversion. Of these, seven leaked early and two had fistulae later, for a 30% success rate. They recommend against low rectal anastomosis when a continent diversion is used (38). Because of the vascular problems associated with a loop colostomy, the surgeon should consider a loop ileostomy for diversion of the fecal stream while the bowel anastomoses heal.
The overall survival rate for patients with pelvic exenteration and low rectal anastomosis at the University of Alabama at Birmingham was 68%. This was superior to that of patients with anterior exenteration (53%) (39), although the difference was not statistically significant. For both groups of patients, survival significantly improved if there was no spread of disease beyond the cervix and/or vagina. Patients with disease confined to the cervix and/or vagina who underwent a total pelvic exenteration and a low rectal resection had a corrected survival rate of 94%, versus 70% for patients who underwent an anterior exenteration. Although this difference is not statistically significant, it suggests that the more extensive procedure may improve survival by virtue of its larger tissue margin around apparently confined tumor. The survival rate for patients with disease in the bladder, rectum, or parametria was 38%.
Techniques for urinary diversion are demonstrated in Chapter 20. The selection of the proper urinary diversion technique depends on a number of factors. The majority of women undergoing pelvic exenteration have had high doses of pelvic radiation therapy, which leads to fibrosis and lack of vascularity in the distal ileum. This increases the risk of anastomotic breakdown and both bowel and urinary fistulae.
Most centers currently prefer the transverse colon when a urinary conduit is chosen as the urinary diversion method in a patient who has had full dose radiotherapy. This leads to fewer bowel and ureterocolonic anastomotic leaks (40). The colon absorbs water, sodium, and chlorides. This may lead to hypochloremic acidosis with hyponatremia and hyperkalemia if there is urinary retention due to stomal stricture, or when a long segment of colon is used. When a 10 to 15 cm length of colon is used and the stoma remains open, the complications of electrolyte imbalance are rarely encountered.
Continent urinary diversion is preferred for those patients with motor skills and motivation to maintain the emptying and irrigation that it requires. It gives the advantage of avoiding the external appliance and helps restore the patient's self-image. A greater degree of renal function is necessary for the continent reservoirs versus the conduits. The glomerular filtration rate (GFT) should be 40 mL/min. or greater. The serum creatinine should be <2 mg/dL and there should be no urinary proteinuria. Because the cecum and ascending colon will absorb electrolytes, it is important that the patient empty the reservoir three times daily and irrigate once daily. The patient may experience diarrhea because the bowel has been shortened and the distal ileum has been taken out of the gastrointestinal stream. This leads to decreased bile acid absorption and steatorrhea. This can be treated with cholestyramine and with motility agents such as Lomotil and Imodium.
In the patients who have significant fibrosis in the pelvis after radiation therapy, the ureters should be cut above the pelvic brim so the ureterointestinal anastomosis has a lower fistula and stricture rate. The left ureter will need to be brought across the midline above the inferior mesenteric artery (IMA) to provide the appropriate length and to decrease the risk of stricture caused by the kinking at the level of the IMA. The continent reservoirs and the transverse colon conduit are ideal for patients with short ureters because the anastomosis is in the mid- to upper abdomen.
Patients are best managed in an intensive care unit with an arterial line and central venous catheter. Central catheter monitoring facilitates administration of blood products, colloids, and crystalloids, particularly in those patients whose urine output is not a reliable predictor of fluid status. Patients have a large abdominal and pelvic peritoneal defect that exudes serum, and they may have significant third-space fluid shifts. Inadequate fluid replacement may lead to intravascular compromise and decreased perfusion of the kidneys. The hematocrit should be kept stable above 30%, and the prothrombin and partial thromboplastin times should be kept normal with fresh frozen plasma. The central catheter can also be used for TPN.
A first-generation cephalosporin is given immediately before surgery for infectious prophylaxis. It is continued after surgery until the patient has remained afebrile for 48 hours. If febrile episodes persist or become severe, antibiotics are changed based on culture results. If no cultures are available, antibiotic therapy is extended to cover anaerobic and gram-negative organisms. If there is fecal spill during surgery, antibiotic coverage is usually extended to anaerobic and gram-negative organisms.
Although the mortality rate is less than 5%, as many as 50% of patients may have a major complication (41,42,43). The most significant intraoperative complication is hemorrhage, with blood loss of 1,500 to 4,000 mL being typical (44,45). Postoperative hemorrhage is often handled by percutaneous embolization because reexploration carries a high morbidity. The length of surgery (4 to 8 hours), large volume of blood loss, and inability to monitor urinary output because of the urinary diversion make the accurate replacement of fluids very difficult. The central catheter is invaluable in monitoring the replacement of blood, colloids, and crystalloids, which may reach 1,500 mL/hour during intraoperative management.
Nonsurgical complications, such as myocardial infarction, pulmonary embolism, heart failure, stroke, and multiorgan failure, account for a 2% to 3% mortality rate and are slightly more common in the elderly patient.
A small bowel anastomotic leak or fistula is a serious complication, with a mortality rate of 20% to 50%. The incidence of small bowel fistulae ranged from 10% to 32%(41,42,43,44,45,46) in patients who had an ileo-ileal anastomosis in previously irradiated bowel. Small bowel fistulae have been virtually eliminated by the use of transverse colon conduits and attention to pelvic floor reconstruction. Today, the continent urinary diversion commonly practiced uses an ileocolonic anastomosis with a low small bowel fistula rate.
The incidence of small bowel obstruction is 4% to 9%. Initially conservative management with nasogastric decompression and TPN should be attempted because reoperation has been associated with an 8% to 10% risk of mortality. The obstructions are most common in the distal ileum at the site of the ileal anastomosis. Avoiding the ileal anastomosis and using pelvic floor reconstruction has decreased the morbidity of small bowel obstruction.
Urinary Tract Complications
The standard urinary diversion for several decades was the urinary conduit using a segment of terminal ileum. The high complication rate of the ileo-ileal anastomosis led to development of the transverse colon conduit (47). There have been no bowel anastomotic leaks reported with this technique, and ureterocolonic anastomotic leaks also are rare.
The continent urinary diversion using the Miami pouch (see Chapter 20) also has a low rate of intestinal fistula formation and urinary leaks. If urinary leaks or fistulae do occur, conservative management with percutaneous drainage is recommended. The mortality rate from surgical reexploration for urinary complications may reach 50%.
The most common long-term complication is pyelonephritis, requiring rehospitalization in 14% of patients. The incidence of ureteral stricture has been decreased by the use of ureteral stents and is approximately 8% (48).
The 5-year survival rate has improved significantly over time (Table 22.3) (5, 6, 41, 42, 46, 49,50,51,52). Patients who have had anterior exenterations have a better survival rate (30% to 60%) than those with total exenteration (20% to 46%), no doubt reflecting the smaller dimensions of the recurrent disease. The clinical factors that have been reported to affect survival most significantly are length of time from initial radiation therapy to exenteration (50), size of the central mass (51,53), and preoperative pelvic sidewall fixation determined by clinical examination (52).
The important pathologic factors are positive nodes, positive margins, and spread of tumor to adjacent organs. The occurrence of metastatic cancer in the pelvic lymph nodes after radiation therapy is a poor prognostic finding at the time of exenteration (Table 22.4). Stanhope and Symmonds (55) achieved the highest 5-year survival rate at 23%. In their analysis, they eliminated confounding high-risk factors, such as positive margins and metastasis to other peritoneal surfaces.
Table 22.3 Operative Mortality and 5-Year Survival Rates for Pelvic Exenteration
Table 22.4 Survival of Patients With Positive Nodes at Time of Exenteration for Postirradiation Recurrence
Rutledge et al. (5) in 1977 reported a 6.6% 5-year survival rate in 30 patients with positive nodes. This publication included patients who had positive pelvic and inguinal nodes and those who died of operative complications. Ten years later, Rutledge and McGuffee (56) reported a 26.3% survival rate in 41 patients with positive nodes. They noted an increase in the incidence of positive nodes in the later cases and suggested that the patients were more highly selected to eliminate other risk factors, and fewer died of operative complications. There was also a decrease in the number of posterior exenterations performed. These patients had vulvar, urethral, and rectal cancers and had been managed more aggressively despite significant risk factors for higher recurrence rates. The 5-year survival rate was 21.9% for recurrent cervical cancer after radiation therapy, after eliminating death from other causes. Given this rate of survival, patients with positive pelvic nodes and no other poor prognostic factors can be considered candidates for exenteration.
Morley et al. (42) reported a 73% 5-year survival rate for 57 patients with squamous cell cancer of the cervix versus 22% for 9 patients with cervical adenocarcinoma. Crozier et al. (57) reported a median survival of 38 months for 35 patients with adenocarcinoma and 25 months for 70 control patients with squamous cell carcinoma. They concluded that patients with cervical adenocarcinomas who meet the criteria for pelvic exenteration have results similar to those of patients with squamous carcinomas.
Chronologic age is not a contraindication to exenteration. Matthews et al. (45) compared 63 patients aged 65 years or older with 363 patients younger than 65 years who underwent pelvic exenteration. The operative mortality rates were 11% and 8.5%, and the 5-year survival rates were 46% and 45%, respectively.
Quality of Life
The quality of life after pelvic exenteration is significantly improved by organ reconstruction. Hawigorst-Knapstein et al. (58) reported 28 patients who were periodically assessed in a prospective study by examination, interview, and questionnaires in the postoperative period. The women were divided into groups of two, one, or no ostomies. A separate comparison was made of women with or without vaginal reconstruction. At all points of evaluation, the patients' quality of life was most affected by worries about progression of the tumor. One year after surgery, the patients with two ostomies reported a significantly lower quality of life and poorer body image than patients with no ostomy.
Those women with vaginal reconstruction reported fewer problems in all categories related to quality of life and significantly fewer sexual problems.
Ratliff et al. (59) prospectively evaluated 95 patients who underwent pelvic exenteration and gracilis myocutaneous vaginal reconstruction. Forty patients completed the study, and 21 (52.5%) reported that they had not resumed sexual activity after surgery. Of the 19 patients who resumed sexual activity, 84% did so within 1 year of surgery. The most common problems were in adjusting to the self-consciousness of the urostomy or colostomy. Vaginal dryness and vaginal discharge were also significant problems. These findings indicate the need for adequate counseling after the exenterative surgery.
While some European centers have advocated pelvic exenteration for patients with advanced primary cervical cancer (60), most centers restrict the operation to patients with a central recurrence following chemoradiation. The only exception would be patients with stage IVA disease and a recto- or vesico-vaginal fistula.
Pelvic exenteration provides the only hope for cure in women with recurrent pelvic malignancies after radiation therapy. Most procedures are done for recurrent cervical cancer. Operative morbidity and mortality can be decreased by careful patient selection, attention to intraoperative technique, excellent postoperative care, and early management of complications. The 5-year survival rate is acceptable given the lack of satisfactory alternative treatments. With modern reconstructive and rehabilitative techniques, the patient can maintain a near-normal lifestyle, but sexual functioning will always be significantly impaired.