Low rectovaginal fistulas

SCIENTIFIC PAPERS

Low

Rectovaginal Fistulas

Approach and Treatment Thomas R. Russell, MD, San Francisco, California Donald M. Gallagher, MD, San Francisco, California

Fistulous communications between the rectum and vagina, although uncommon, present symptoms that are unacceptable to the patient. The symptoms of passing flatus and feces from the vagina cause the patient to be anxious and, at times, hostile and litigious. In addition, some patients have gross fecal incontinence when there is associated sphincter muscle damage. Aside from the noxious physical symptoms of these fistulas, the patients are emotionally sensitive because the associated physical impairments cause loss of social contact, and this sensitivity must be reflected in the patient’s care. This was aptly described by T. H. Tanner in 1855: “There are few affections, unattended with danger to life, which give rise to greater anxiety or produce more disagreeable results than cases of rectovaginal fistula” [I]. Most patients are initially seen by gynecologists, and various surgical procedures with emphasis on a vaginal approach are advocated in obstetric and gynecologic literature [Z]. We emphasize that the primary opening of these fistulas is within the anorectal canal, not the vagina, and that correction necessitates the obliteration of this anorectal source of continued infection. The surgical principles advanced in the present study include full mobilization of the anterior rectal wall, wide excision of the internal opening of the fistula, repair of muscle tissue when there is attendant muscle damage, and avoidance of proximal colostomy [3]. These principles were advanced by Noble [4] in 1902, Elting [5] in 1912, and more recently by others [6,7]. Years ago, rectovaginal fistulas were more common than today. However, even under ideal conditions of childbirth and vaginal surgery, rectovaginal and From the Department of Surgery, University of California, San Francisco, California. Reprint requests should be addressed to Donald M. Gallagher, MD. 3838 California Street, San Francisco, California 94118. Presented at tfx ForiyIighth Annual Meeting of the Pacific Coast Surgical Association, Palm Springs, California, February 20-23. 1977.

Volume 134, July 1977

rectoperineal injuries will continue to occur. This is due to the lack of levator ani muscle support anteriorly as contrasted with the posterior wall. Also, the anterior anal crypt is a site of frequent anorectal infection, that is, anal fissures, abscesses, and fistulas. A midline episiotomy may extend into an anterior anal crypt with resultant infection. Factors other than childbirth may cause rectovaginal fistulas, such as inflammatory bowel disease, vaginal surgery, drainage of perianal abscess that has masqueraded as a Bartholin abscess, and direct extension of an anterior perianal abscess [8,9,10]. Clinical Material Two repairs of a low rectovaginal fistula, one without and one with associated sphincter muscle damage, are presented. Low rectovaginal fistulas have a direct tract between the lowest portion of the rectum and vagina deep to the anterior sphincter muscles. The more complex rectoperineal vaginal defects vary in degree of penetration and resultant damage to the anterior anal canal, anal sphincter muscles, and muscles of perineum. In each case, a low communication between the rectum and vagina exists between several millimeters and 2 cm from the mucocutaneous junction. Neither higher fistulas which are subject to a different approach, nor common anterior perianal fistulas are discussed herein. Rectovaginal Fistula

Physical examination reveals a depressed or “pit like” defect in t,he anterior midline crypt. Probing reveals an internal opening, as illustrated in Figure 1A. All repairs are preceded by mechanical bowel preparation. For exposure and maximal sphincter relaxation, the patient is placed in the prone (jackknife) position under regional block anesthesia. The procedure is delayed for weeks or months after occurrence of the rectovaginal fistula to allow the inflammatory reaction to subside. A semicircular incision is made in the anal skin anterior to the rectum several millimeters below the mucocutaneous line. (Figure 1R.) The dissection plane includes mucosa, submucosa, and a portion of the internal sphincter muscle. The internal sphincter muscle (not a superficial portion

13

Russell and Gallagher

i

Figure 1 1. A, low rectovaginal fistula. 6, semicircular anterior anal incision. of excision of fisfulous opening and adjacent scar tissue.

of the external sphincter) is exposed in the dissection, since under regional block anesthesia, the internal sphincter moves down and out to a more superficial location. Dissection is facilitated by mobilization laterally where one finds normal tissue, as opposed to the midline which is scarred. Mobilization of the anterior rectal wall is continued cephalad for at least 4 cm. Once one is above the fistula and the associated scar, the dissection plane is readily developed. The lower portion of the anterior rectal wall containing the fistulous opening and adjacent scar tissue is now excised. (Figure 1C.) Normal rectal wall is advanced and approximated without tension to the exposed internal sphincter muscle and anal skin at the site of the anterior anal skin incision. Approximately four to five sutures of 2-0 chromic catgut are used for this purpose. (Figure 2.) The portion of the tract that extends through the rectal vaginal septum and posterior vaginal floor remains open and acts as a natural dependent drainage route beneath the advanced rectal wall. Postoperatively, all patients are maintained on a liquid diet until they have a bowel movement. Then a regular diet is started supplemented with hydrophilic bulk agents. Sitz baths are used for comfort and cleanliness. RectovaginalFistula with Associated Sphincter Muscle and Perineal Injury (rectoperineal

vaginal fistula)

This type of fistula is less common than the previously described rectovaginal fistula. Figure 3A illustrates the more extensive injury in which there is involvement of the anterior anal canal and perineal structures as well as a rectovaginal or rectoperineal opening. The internal opening is in a depressed scar of the anterior midline anal crypt. There is attenuation, separation, scarring, and loss of anterior sphincter muscle and perineal body. The midline

14

G mobilizafion of anterior rectal wall with line

defect is covered by a depressed V-shaped scar and sulcus. Gross fecal incontinence is the result, and repair includes restoration of sphincter muscle and perineal structures as well as excision of the fistulous tract. Proper skin incisions, which are important to gain exposure for complete mobilization of muscle tissue and repair without tension, cannot be obtained with incisions that are primarily in the midline or in a radial direction. Exposure is begun with a long transverse incision across the entire perineum as well as an anterior anal skin incision several millimeters below the mucocutaneous junction. (Figure 3B.) These two incisions are joined, or meet, in the midline. Skin flaps are developed and turned back in order to visualize scar and muscle which has retracted from the midline and lies laterally and posteriorly. (Figure 3C.) With scissors dissection beneath the skin flaps, perianal skin is completely mobilized free of underlying sphincter muscle and carried to the posterior portion of the anal canal, as illustrated in Figure 4A. Similarly, skin is dissected free of the transverse muscle of the perineum and scar across the perineum. The anterior anal skin incision is mobilized as previously described. The involved infected scarred tissue of the anterior wall is excised. Muscle repair to restore the perineal body is now begun. Scar tissue that is attached and partially replaces muscle is not excised, as it affords better tensile strength than muscle itself. Allis clamps are placed on bundles on each side, attempting to identify corresponding bundles. Sphincter musclds can be differentiated from the muscles of the transverse perineal group by the direction of their fibers. The various portions of sphincter muscles cannot be individually identified. The deeper portion of the scarred transverse muscles of the perineum are brought to the midline and sutured together using 2-O chromic catgut. (Figure 4B.) In a similar

The American Journal of Surgery

Low Rectovaginal Fistulas

Figure 2. Completion of repair by advancing rectal wafl and sufuring t0 the infernal sphincter muscle and anal skin. The tract deep to the rectum is nof closed.

Figure 3. A, re&ovagina/ fistula with dfsruption of sphincter and perineal muscles anteriorly with attenuated perineal bodyB, transverse and semicircular incisions used. C, ear/y mobifization of skin flaps and anterior rectal waK fashion, the deeper sphincter muscles of the anus are brought together across the anterior midline. At times, the catgut sutures will unite both sphincter muscles of the anus and transverse muscles of the perineum. After this deep repair, the more superficial muscle and scar bundles are united and sutured in the midline. (Figure JR.)

Volume134,July1977

The previously mobilized anterior rectal wall, now devoid of infected tissue, is sutured to the innermost portion of the sphincter muscle and anterior anal skin. Next, the skin flaps are brought across the midline. The perineum, as it presently lies in the midline, is now of much greater thickness and is no longer t.hinned out and widened in a

15

Russell and Gallagher

a

b

Figure 4. A, full mofS/ization of anal skin and perineal skin from attached muscle and scar tissue. B, approximation of scarred external sphincter and perineal muscles. advantage of excellent secondary healing and tissues need only to he approximated.

by contraction,

Results

Figure 5. Redundant scarred skin is excised and the remaining skin is approximated in the mtdline. transverse direction. As a result, there is an excess of scarred skin. (Figure 5.) One to 4 cm of redundant scarred skin is excised and normal skin is approximated with fine catgut sutures. Perianal and perineal wounds have the

16

Thirty-two patients, aged twenty-three to fiftyseven years, who have undergone either of the repairs just described and have been followed for an average of more than ten years are included in this review. Although additional patients have undergone operation, they are not included because of inadequate follow-up. Ten of the thirty-two patients had gross fecal incontinence. The most common cause of the fistulas-perianal infection involving the anterior midline anal cryptoccurred in eighteen patients (15 rectovaginal, 3 rectoperineal). Six of the patients with rectovaginal and two with rectoperineal fistulas had had one to six previous operative procedures that had failed. Five patients (3 rectovaginal, 2 rectoperineal) operated on by us had recurrences requiring additional surgery to close the fistula. Fistulas of all the patients eventually closed with one exception (rectovaginal) which remains a current failure after four operative attempts. In appraising the results, thirteen (10 rectovaginal, 3 rectoperineal) were evaluated as good and four (rectovaginal) as fair. Patients in the latter category have occasional discomfort or soiling. The second most common cause-rectal injury occurring at childbirth, with a tear or episiotomy involving the rectum-occurred in twelve patients (6 rectovaginal, 6 rectoperineal), of whom seven (2 rectovaginal, 5 rectoperineal) had had one to three previous attempts at repair through the vagina. Good

The American Journal of Surgery

Low Rectovaginal Fistulas

surgical results were obtained in eleven patients; one result (rectovaginal) was fair. There were no recurrences. Although additional patients were seen with inflammatory proctitis and/or colitis associated with rectovaginal fistula, only two patients underwent operation, after conservative treatment failed. Both repairs had favorable results. However, the disease redeveloped years later in one of the patients who consequently required colectomy. We do not advocate this type of surgery in the presence of inflammatory bowel disease. Of all the repairs performed, (22 rectovaginal, 10 rectoperineal), twenty-five (15 rectovaginal, 10 rectoperineal) were considered good, six (rectovaginal) were considered fair, and one (rectovaginal) is currently a failure. Half of the patients had undergone one to six previous attempts at repair. Colostomy was not employed in this series, even in the setting of multiple recurrences. Results in patients with childbirth injury were excellent and clearly, from this review, were superior to the results in those patients whose primary problem was infection.

acceptable recurrence rate was achieved. Colostomy was not used in this series. References 1. Hudson CN: Acquired fistulae between the intestine and the vagina. Ann Co/l Surg fngl46: 21, 1970. 2. Lescher TC, Pratt JH: Vaginal repair of the simple rectovaginal fistula. Surg Gynecol Dbstet 124: 1317, 1967. 3. Gallagher DM, Scarborough RA: Repair of low rectovaginal fistula. Dis Co/ Rectum 5(3): 193, 1962. 4. Noble GH: A new operation for complete laceration of the perineum designed for the purpose of eliminating danger of infection from the rectum. Trans Am Gynecol Sot 27: 357, 1902. 5. Elting AW: The treatment of fistula in ano. Am Surg 56: 774, 1912. 6. Laird DR: Procedures used in treatment of complicated fistulas. AmJSurg 76: 701. 1948. 7. Mengert WF, Fish SA: Anterior rectal wall advancement: technic for repair of complete perineal laceration and rectovaginal fistula. Obstet Gynecol!? 262. 1955. 8. Faulconer HT, Muldoon JP: Rectovaginal fistula in patients with colitis. Dis Co/on Rectum 5: 413. 1975. 9. Belt RL Jr:, Repair of anorectal vaginal fistula utilizing segmental advancement of the internal sphincter muscle. Dis Co/on Rectum 12(2): 99, 1969. 10. Given FT Jr: Rectovaginal fistula. Am ./ Dbstet Gynecol108( 1): 41, 1970.

Discussion Comments

All series of rectal fistula involving the vagina have a significant recurrence rate requiring multiple procedures. Due to socially unacceptable symptoms, these patients are unhappy and will go to great measures to obtain closure of a rectovaginal fistula. The common denominator causing breakdown of repair is infection. Prior to repair of childbirth injury or acute anorectal infection, a lengthy period of time should ensue for acute inflammation to resolve. Then, by following the principles previously mentioned of advancing full thickness rectal wall to cover the defect and excising all infected tissue, one can expect a high success rate. Our success with fistulas induced by infection was less favorable than with fistulas induced by injuries at childbirth. Although there were recurrences, closure can be accomplished with additional repairs. Diverting colostomy was not necessary in our experience.

Experience

rectovaginal sphincter

with

thirty-two

fistula

with

damage

patients

or without

is reported.

with a low

attendant

The technic

used is

advancement of the anterior rectal wall with excision of the infected anal glandular tissue and repair of muscle tissue when indicated. Anorectal infection and childbirth

injuries

Volume 134, July 1977

were the common

causes, An

Russell R. Klein f San Rafael, CA) : In discussing this problem with my colleagues in obstetrics I was informed that each one sees one or two cases in a lifetime; therefore, as the authors have pointed out, the condition is very rare. Most obstetricians, of course, are much more familiar with a vaginal approach, so they usually repair the rectovaginal fistula either by excising the tract and reapproximating all the layers or converting it to a third degree laceration and then doing a primary repair. They tell me their results are excellent. I do not think that they are ever going to change their method, and few of them are even aware that it is possible to eradicate this problem through a rectal approach. The authors have alluded to their technic as a perineal approach. Since there is virtually no incision on the outside, I really think a better term would be transanal repair or anterior advancement of the rectal wall. The authors have indicated that 2-O chromic catgut is the suture of choice. This is rather interesting, because with all our advancements we still have to rely on O-Ochromic catgut. Maybe the authors will give this matter further study, as to whether there is some synthetic suture material available that may be more reliable. (Slide) As Doctor Russell pointed out, the mucocutaneous line where most of the-fistulas start is at this level and is approximately in the midportion of the internal sphincter. (Slide) This is a specimen of the area operated on. The skin of the anal canal; the mucocutaneous line; a hypertrophied anal papilla; and a crypt. The authors have advised us to make a horizontal incision at this level just distal to the primary opening, which is usually here.

17

Russell and Gallagher

(Slide) This is the internal sphincter. The authors have made a transverse incision here at the dentate line, utilizing a portion of the internal sphincter to give a viable flap that can be elevated and then brought down after cutting off the involved scar and infected area. This report presents an excellent exposure. The results are not perfect but they are very good, and anyone who feels at all familiar in this area will find it is a very satisfactory approach.

Carl E. Lischer (St. Louis, MO): I thought it would be of some interest if I were to describe a new operative approach to this problem, devised by Doctor Eugene Bricker and not yet published. (Slide) Most rectovaginal fistulas are traumatic or iatrogenic and have the following important characteristics: (1) There is no associated loss of tissue. (2) The fistulous tract is formed by tissue that has normal vasculature. (3) The fistulas are amenable to excision and anatomic repair in layers, using local tissues with a good chance of healing. (Slide) On the other hand, post radiation fistulas have the following characteristics: (1) There is always some loss of tissue from sloughing, and the loss may be extensive. (2) The fistulous defect is bordered by tissue that has been severely damaged by radiation. (3) The defect is usually associated with scar contracture of the rectum which decreases its volume and circumference. (4) Repair by the use of local tissue to close the defect is impossible. (5) Permanent colostomy is usually the only solution. (Slide) However, permanent colostomy has been avoided in a few very small postradiation fistulas which have been closed with local tissue (How many?). In addition, the best chance for successful closure is resecting the irradiated portion of the rectum and low anastomosis. (Slide) At the Ochsner Clinic, 136 patients with rectovaginal fistulas were seen from 1949 to 1970, nine (6.6 percent) of which were caused by radiation for pelvic cancer. (Slide) Gerald Marks reported on eight cases of combined abdominotranssacral reconstruction of radiation-injured rectum (Am J Surg 131: 54,1976). (Slide) An alternative approach to the problem, suggested by Doctor Bricker, (1) leaves the remaining rectal

wall and its innervation undisturbed, (2) transplants vascularized bowel wall to close the defect and adds volume and circumference to the rectum, and (3) during healing of the repair, leaves the patient with a protecting colostomy which can be permanent with no further operations if the reconstruction is a failure. (Slides) With this approach, the fistula is exposed through the abdomen. A rim of scar tissue is excised and the proximal rectosigmoid colon is turned down on itself, the open end being anastomosed to the fistulous opening in the rectum. No sutures are placed on the vaginal side of the fistula. At a later date the proximal colostomy opening is sutured end-to-side to the intrapelvic loop of rectosigmoid colon. (Slides) Doctor Bricker has performed this operation on two patients with postradiation rectovaginal fistula. Both patients are completely continent of gas as well as liquid and solid feces, and both report normal sexual function. Donald M. Gallagher (closing): As to why we use 2-O chromic catgut for the repair, I can only reply that this has evolved as a matter of convenience. Doctor Robert Scarborough, who taught us the procedures described, used “pull out” stainless steel wire sutures and had excellent results. Doctor Lischer has presented the problem of rectovaginal fistulas resulting from radiation injury. This a subject in itself. Such fistulas occur at a higher level. We have attempted advancement of the anterior rectal wall for this problem and have failed. We appreciate Doctor Lischer presenting us with a unique approach to the problem of radiation injury. In closing, I emphasize the marked sensitivity of the patient, for I know of no surgical problem more associated with professional liability than that of the patient with a rectovaginal fistula. Rectovaginal fistulas are less frequent today because of the advances made in childbirth; however, they continue to occur. The anterior portion of the anorectal canal is vulnerable to injury since there is a relative absence of levator muscle support anteriorly, and the anterior anal crypt is a source of infection. Successful repair cannot be promised, but every attempt should be made, for the resulting noxious symptoms of these fistulas affect every social aspect of a woman’s life.

The American Journal 01 Surgery