THE JOURNAL OF UROLOGY
Vol. 111, January
Copyright© 1974 by The Williams & Wilkins Co.
Printed in U.S.A.
DIAGNOSIS AND TREATMENT OF COLOVESICAL FISTULA FRANK P. MORSE, III
AND
STEPHEN P. DRETLER
From the Departments of Urology, Massachusetts General Hospital and Peter Bent Brigham Hospital, Harvard Medical School, Boston, Massachusetts
Adult colovesical fistulas occur as a consequence of inflammatory processes, neoplastic disease or trauma.'· 2 The non-specificity of clinical symptoms and physical signs and the unreliability of radiographic and endoscopic studies make the diagnosis of colovesical fistulas difficult. The study reported herein was undertaken to determine the clinical presentation of colovesical fistulas, the relative efficacy of radiographic and endoscopic studies and the results of surgical management. MATERIALS AND METHODS
A retrospective study of case records from our 2 hospitals from 1961 to the present revealed 39 adult patients with surgically, radiologically or autopsy proved colovesical fistulas. When evaluating each patient record, the following factors were considered: age, sex, cause of fistula, presenting clinical symptoms and physical signs, bacteriologic studies, accuracy of radiographic studies, reliability of endoscopic examinations and modes of therapy. RESULTS
Twenty-four men and 15 women ranging from 41 to 87 years old had colonic-vesical fistulization. The most common cause for fistula formation was inflammatory bowel disease, occurring in 27 patients (69 per cent): 18 men and 9 women. Twentyfour patients had diverticulitis, 2 had ulcerative colitis and in 1 the fistula was a result of an appendiceal abscess. Other causes of fistula formation are listed in table 1. The most common symptom of colovesical fistula was pneumaturia, occurring in 21 of 39 patients (17 men and 4 women). Fecaluria was noted in 20 patients. Seventeen patients had frequency, urgency and/or dysuria. Sixteen patients had fever and/or chills. Two patients presented in gramnegative sepsis and shock. Two patients had urine passing per rectum (table 2). Positive physical signs were few. Admission physical examination revealed a palpable abdominal or pelvic mass in 12 of 39 patients. In 2 Accepted for publication June 29, 1973. Read at annual meeting of Society of University Urology Residents, Williamsburg, Virginia, May 9-11, 1973.
1 Mayo, C. W. and Blunt, C. P.: Vesicosigmoidal fistulas complicating diverticulitis. Surg., Gynec. & Obst., 91: 612, 1950. • Pugh, J. I.: On the pathology and behaviour of acquired non-traumatic vesico-intestinal fistula. Brit. J. Surg., 51: 644, 1964.
22
additional patients, masses were discovered on examination with anesthesia. The results of urine culture were reported in 32 patients. Escherichia coli was the predominant organism in 25 patients. Four patients had mixed growth of other gram-negative bacteria. Four patients had no growth reported on urine culture; 1 was known to be on antibiotic treatment. Excretory urography (IVP) was performed on 36 patients and 29 showed some abnormality. An unequivocal diagnosis of colovesical fistula was made in 5 patients in whom dye entered the colon (see figure). In 14 patients the cystographic phase of the IVP showed a bladder wall irregularity and 1 patient had air within the bladder. Five IVPs showed evidence of hydronephrosis: 3 of these patients had carcinoma of the cervix and 2 had inflammatory bowel disease. Barium enemas were performed on 32 patients. Only 11 patients showed unequivocal fistulization between colon and bladder. However, all patients showed some colonic abnormality, that is diverticulum or mass. Cystograms were performed on 17 patients. Six demonstrated a bladder-colon communication. In 4 of the 6 IVP had failed to demonstrate fistulization. In addition, barium enema failed to demonstrate a fistula in 4 of these patients. Two patients had cystograms which showed bladder wall irregularities but were not diagnostic of colonic-vesical fistulization. Sigmoidosc(?pic examination confirmed the presence of only 1 fistula which was palpable by digital examination. Cystoscopic examination was performed on 28 patients; in 12 the fistula site could be identified. In 2 additional patients, 1 with barium in the bladder and 1 with feces in the bladder, a presumptive diagnosis of fistula could be made. Ten patients had localized areas of bladder wall inflammation but no identifiable fistula. Two patients had an entirely normal cystoscopic examination. Visible dyes (Congo red, indigo carmine, methylene blue and charcoal) were administered to 5 patients, orally or through a colostomy, and in 4 patients dye was found in the urine (table 3). Therapy for colovesical fistulas has included non-operative treatment, surgical palliation and 1, 2 or 3-stage surgical reconstruction. Of 22 patients who underwent complete surgical reconstruction, 16 underwent a 3-stage procedure (transverse colostomy, fistula resection and colostomy closure). Two major complications occurred: a pelvic abscess resulting in death and a recurrence of the
DIAGNOSIS AND TREATMENT OF COLOVESICAL FISTULA
fistula, Four patients had 2-stage reconstruction (primary resection with colostomy and colostomy without Two patients underwent 1-stage reconstruction; in 1 of the 2 a leak in the colon anastomosis which
TABLE
Type
No. Performed
!VP
:36
Barium enema
32 17 28 5
because of the 1'ABLE
3. Diagnostic studies Some
Unequivocal
Abnormality
Diagnosis
Ko.{':{.)
No.(%)
29 (80) 82()(10)
5 I 14) 11 (34) ti (:l:i) 14 (50) 4 (llll)
8 (47)
26 (92) 4
(80)
1. Cause of colouesical fistula
-·----------·-----
TABLE 4.
- - - -
Female Male --·-··-------17 7
Diverticulitis
Treatment of colouesical fistulas
Totals(~; I l\io. !'ls. 24 (60)
Carcinoma of colon
l
:1
4 (10)
Carcinoma of cer\'ix Ulcerative colitis
0
:J
:i (7 5) 2 (51 I (2,51
Diverticulitis Carcinoma of cervix
()
(2,51 (51
Ruptured appendix Prostatic abscess Two-stage procedure: Diverticulitis Carcinoma of colon
abscess
0
2
2 2 (51
0
1
Foreign body Unkno\vn
(2.5)
0
Three-stage procedure: l2
Carcinoma of colon
One-stage procedure:
* One patient had diagnosis of diverticulitis and ulcerative colitis.
TABLE
2. Symptoms manifested by patients u·ith
colouesical fistula (89 patients) No. Cases(%) -----
Pneumaturia Fecaluria
Hematuria Urine per rectum
--------------------21 (54) 20 (fil I 17 (44)
16 2 2
(41) (5)
(5)
Diverticulitis Colostomy with resection: Diverticulitis and ulcerative colitis Ulcerative colitis Carcinoma of cervix
Foreign body Colostomy with ilea! loop: Carcinoma of colon Traumatic Colostomy only: Diverticulitis Carcinoma of colon
Conservative: Traumatic Diverticulitis
Died: Diverticulitis Carcinoma of colon Foreign body
Unknown
patient's age, undergo additional treated solely by colostomy: 2 elderly patients diverticulitis and 1 with an colon cancer. One had conduit diversion and another had of the rectal fistula but unsuccessful of bladder and persistent fistulizatio:n necessitated ilea! conduit diversion. Four patients had no died of overwhelming zation. In 1 patient the fistula closed with drainage and the other patient was not be a suitable candidate for an (table DISCUSSION
IVP illustrates dye in colon in patient with colovesical fistula.
the fact that diseases of the colon most common cause for colonic-vesical fistulization, the clinical symptoms are ily urinary in origin. Pneumaturia was more corn mon in men, the shorter urethra makes the sensation of air
24
MORS_E AND DRETLER
less marked. Although pneumaturia may occur as a result of gas-producing organisms in the urinary tract, especially in diabetics, 3 its presence can lead to the possible diagnosis of enteric-vesical fistula. Although pneumaturia is the most commonly occurring symptom, fecaluria and cystitis are more distressing and may be the symptoms which cause a patient to seek medical attention. Patients with persistent urinary tract infection, especially men with coincident inflammatory bowel disease, should be suspected of having colovesical fistulization and appropriate studies should be made. Symptoms of inflammatory bowel disease may be minimal and when fistula formation is suspected a careful history must be obtained. Since abdominal or pelvic masses were found on physical examination in only 31 per cent of patients, their absence should not deter further diagnostic evaluation. If the search for a colovesical fistula is to include further diagnostic studies, it seems reasonable to start with the most productive, easiest and least expensive, that is the administration of a visible oral dye, such as charcoal, and the observation of the urinary sediment. If dye appears in the urine, a diagnosis of fistula is confirmed. Radiographic studies are necessary for localization of the fistula. Cystography (35 per cent) and barium enema (34 per cent) had the highest incidence of unequivocal fistula demonstration. Although IVP showed some abnormality in 29 of 36 patients, a definite diagnosis of fistula (dye in the colon) could be made in only 5 patients (14 per cent). It is likely that the higher concentration of dye in cystography and barium enemas and the higher intracavitary pressures generated, resulted in the higher yield of positive diagnoses. If delayed films are taken after barium enema, it is possible that the diagnosis of fistula could be made more frequently. One patient in this series did not have a fistula demonstrated by barium enema, yet barium was found in the bladder when endoscopic examination was performed 24 hours later. Cystoscopic examination showed some abnormality suggestive or diagnostic of fistulization in 26 of 28 patients. However, we think that the converse is more significant, that is that 2 patients had a bladder which appeared entirely normal at cystoscopy, yet had a colovesical fistula. When the index of suspicion is high, non-diagnostic cystoscopy must be accompanied by normal IVP, cystography and barium enema examination before the • Campbell, M. F. and Harrison, J. H.: Urology, 3rd ed. Philadelphia: W. B. Saunders Co., vol. 1, p. 208, 1970.
diagnosis of colovesical fistula may be excluded. In surgical management of colovesical fistulas, the patient's age, general condition and the cause and severity of the primary disease must be considered. Palliative colostomy may be the only procedure indicated in certain cases. In patients who are thought to be candidates for reconstruction, it must be decided who are suitable for a 1-stage resection and who must be treated more conservatively by a staged procedure. There are reports of excellent results following staged procedures and many authors consider this the safest method of management.'-• Recently several authors have advocated the use of 1-stage resection. 7 - 9 There is general agreement that certain criteria regarding patient selection must be met for 1-stage resection to be successful. The patient's general condition must be good, the fistula must not involve a third organ and must be inflammatory in origin, no abscess should be present, the disease should be localized and the fistula should be small. 10 Patients in whom these criteria are not met will benefit from staged operations. SUMMARY
The diagnosis of colovesical fistula may be difficult and the treatment involves significant morbidity and mortality rates. Thirty-nine patients with proved fistula were retrospectively studied. The results suggest that no single diagnostic test is clearly superior and that a thorough examination should be made. The use of visible dyes may be helpful in confirming the diagnosis of fistula. Treatment with staged reconstruction is safer in the poor-risk patient but I-stage operation is a reasonable alternative if certain criteria are fulfilled. • Ward, J. N., Lavengood, R. W., Jr., Nay, H. R. and Draper, J. W.: Diagnosis and treatment of colovesical fistulas. Surg., Gynec. & Obst., 130: 1082, 1970. • Smithwick, R.H.: Surgical treatment of diverticulitis of the sigmoid. Amer. J. Surg., 99: 192, 1960. • Moore, F. D.: The gastrointestinal tract and the acute abdomen. In: Surgery. Edited by R. Warren. Philadelphia: W. B. Saunders Co., chapt. 25, 1963. 7 Ewell, G. H.: Intestinovesical fistula. Amer. J. Surg., 82: 597, 1951. • Scardino, P. L. and Lippitt, W. H.: Vesical intestinal fistula. J. Urol., 99: 752, 1968. • Carpenter, W. S., Allaben, R. D. and Kambouris, A. A.: One-stage resections for colovesical fistulas. J. Urol., 108: 265, 1972. 10 Giffin, J. M., Butcher, H. R., Jr. and Ackerman, L. V.: Surgical management of colonic diverticulitis. Arch. Surg., 94: 619, 1967.