Surgical Complications of Gynecologic Surgery Leroy R. Weekes, MD, Shobhana Anil Gandhi, MD and Anil Krishnakumar Gandhi, MD Los Angeles, California

Complications of gynecological surgery are considerable and when reviewed in detail are almost frightening. There is no substitute for experience and intimate knowledge of the intricate pelvic structures in health and disease. Anyone who is active in the field is sooner or later going to experience some difficulty whether it be due to his miscalculation or to innate conditions in the patient which are beyond his/her control. It is the responsibility of the pelvic surgeon to recognize the complication and apply proper corrective measures. The patient should not be given false hopes of sure success nor should she be deprived of whatever hope for success does exist. The surgeon who has never committed an error has never done an appreciable amount of surgery. Every honest operator will admit to his errors and hopefully profit by them. To avoid errors in judgement or technique is the aim of all, but since surgeons are made of the same material as the rest of mankind they naturally make their share of mistakes. When errors are committed at the operating table, they can often be corrected. I have witnessed surgeons making technical errors become tremendously upset and lose their ordinary good judgement. Often some improvised procedure can be done on the spot to correct or alleviate the error, but much can be accompllshed by anticipating the possibility preoperatively and carefully considering the method of correction.1

Any surgical procedure on the female pelvic organs subjects the patient to the risks of general anesthesia and surgery. The surgeon is wise who recalls that the preoperative evaluation must be as complete as possible, even in emergency situations. Shortcuts taken for convenience or for socioeconomic reasons should be condemned as inadequate and incompetent medical care. Dr. Weekes is clinical professor of obstetrics-gynecology, U niversity of Southern California School of Medicine and professor of obstetrics-gynecology, Charles Drew Post Graduate Medical School, Queen of Angeles Hospital, and Temple Hospital, Los Angeles, California. Requests for reprints should be addressed to Dr. Leroy R. Weekes, Julian W. Ross Medical Center, 1828 S Western Avenue, Los Angeles, CA 90006

The art and science of gynecologic surgery have kept pace with the great advancements and responsibilities that have benefited the practice of medicine and surgery in recent years. Gynecology has responded to the increasing demands of surgery of malignant and benign disease as well as congenital deformities in the female pelvis. We are caring for these problems in a bolder and more comprehensive manner than ever before. Although this task has been made much easier by the modern, well-trained anesthesiologist, the urologist, and the internist, nevertheless, the increased skill and confidence of the gynecologic surgeon is the main ingredient which gives the newer dimensions of effectiveness to his efforts.

Material and Methods This is a five-year study, based upon the caseload of the Queen of Angels Hospital. This hospital is unique in that it is operated by a non-religious organization which functions within the restrictions of the Catholic Hospital Code. It prohibits all procedures associated with interruption of pregnancy. In order to make this presentation as comprehensive as possible, illustrative cases were taken from the files of Temple Hospital which serves a comparable population but has no religious restric-

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION, VOL. 69, NO. 12, 1977

tions. This allowed us the opportunity to augment our five-year study with additional illustrative cases, some involving interruption of pregnancy during the first trimester. The number of abortions performed remains small, since the privilege is restricted to board-certified gynecologists, personal patients, and first trimester. The five-year review from the Queen of Angels Hospital revealed 228 bona fide complicated cases. Of these cases, ten complications were most often seen (Table 1). The frequency of occurrence parallels the order of presentation.

Case Reports Case 1 MC, a 65-year-old female, was admitted to Queen of Angels Hospital on September 29, 1976. She was admitted from a mental institution because the smell from a vaginal discharge was so offensive that other people did not want her in their presence. At the time of admission, examination showed a possible carcinoma of the endometrium and a D & C confirmed this impression. An IVP showed no abnormalities, but a barium enema and sigmoidoscopy confirmed a large recto-vaginal fistula due to tumor. Because of this, and the possible fixation of the tumor to the posterior wall of the rectum, it was decided in September that the patient should receive radiation therapy prior to attempting removal of the large tumor. She underwent radiation therapy to the whole pelvis (a total of 4,000 rads). She had a rest period of approximately four to five weeks and was readmitted when examination showed a hemoglobin level of 6.9 gm and a question of a fixed area in the pelvis. It was decided that she should be explored, and if indicated she would have an abdominoperineal resection. Blood transfusions were administered to the patient and her potassium and hemoglobin levels were corrected. 881

Case 2 Table 1. Frequent Complications in Gynecologic Surgery

Febrile postoperative course

98

Cystitis and upper urinary tract infection

51

Anemia secondary to acute blood loss

44

Vaginal cuff cellulitis

24

Vaginal cuff bleeding

21

Reactive ileus

11

Incisional abscess

6

Anterior abdominal wall hematoma

5

Laceration of bladder

4

Reaction to blood transfusion

3

The hematocrit volume rose to 44 percent. Adjustments were made for her diabetes, hypertension, and other medical complications. On January 17, 1977, she underwent a laparotomy, at which time lysis of adhesions showed not only a uterocolonic fistula, but also a vesicocolonic, and small bowelcolon fistula. Because of the large extent of the tumor and the presence of fecal material (due to the inability to completely prepare her bowel), it was decided that a total exenteration was warranted. Because of the patient's mental attitude, however, it was decided that this would not be compatible with her life. Our election was to divert the small bowel as a mucous fistula as well as the colon, perform a colostomy and hope that there was not significant disruption of tumor into the bladder. A small bowel anastomosis and colostomy were performed and the pelvis drained. Postoperatively, it was immediately noted that the patient had fecal material draining from her Foley catheter, but this seemed to clear after 24 to 48 hours. Her postoperative course was fairly unremarkable and she recovered fairly well from surgery. The patient then remained afebrile until the ninth postoperative day when she showed a pattern indicating an abscess. Her temperature would rise to as high as IOIF by 6 AM and fluctuate to 99F at 6 PM. This fluctuating pattern continued for the next few days. Later, she developed fecal sinuses draining through multiple areas on the 882

abdominal wall. These appeared to be draining well and the patient seemed to be comfortable with the drainage sites. There was not a rework-up on the number of fistulous tracts but on February 9, 1977, a hypaque study through her mucous fistulae showed a communication with the small bowel and a large sinus developing in the pelvis. This was thought to be in the vagina and uterus. It was decided that as long as the patient remained comfortable and seemed to be controlled, her tumor (which was unresectable) would likely kill her before the sinus and fistulous tracts would. These did not seem to bother her. The patient seemed to remain quite comfortable from February 14, 1977 on. It was attempted to transfer her to a convalescent home, but because of the multiple sinus tracts she could not be treated there and was referred back to the hospital almost immediately on the same day she was discharged. Her condition remained fairly stable for the next three to four weeks. She continued to have a temperature as high as 1 QOF. It was noted on March 2 that she was beginning to require more pain medication. On March 5, there were signs of increasing potassium, BUN, and creatinine levels. On March 6, she began bleeding profusely from the vagina and died immediately afterwards. Her diagnosis at death was adenocarcinoma of the rectum with involvement of the uterus, vagina, and small bowel, postradiation therapy, with onset diabetes mellitus and hypertension.

MD, a 61-year-old female, was admitted to Temple Hospital in August 1975 complaining of lower abdominal pain. A laparotomy was performed and disclosed extensive abdominal carcinomatosis primarily in the ovaries. The ovaries were removed and the omentum resected. Her recovery was slow and was complicated by fever of unknown origin. She left the hospital two weeks later on August 27, 1975. At home, she improved momentarily but on the morning of admission she developed a fistula with passage of stool through the lower transverse abdominal incision. A fter admission a gastrografin enema disclosed a fistula between the sigmoid colon and the bladder and in turn a communication with the skin through the wound. There was obviously an intraperitoneal but walled off abscess associated with the fistula. On September 8, 1975, a defunctionalizing right transverse colostomy was performed. She recovered rapidly from her sigmoid procedures. The fistula continued to drain but in less amounts. She was referred to an oncologist for chemotherapy.

Case 3 This 33-year-old female, para 1, grav 3 (ectopic pregnancy) entered Temple Hospital on March 5, 1977 requesting a therapeutic abortion because of her depressed state. She had a ten-week period of amenorrhea. During the course of the abortion, the uterus was perforated by a number 12 suction curet. Influenced by the likely possibility of intraperitoneal hemorrhage, an exploratory laparotomy was immediately performed. About 350cc of bright red blood were detected in the pelvic cavity and a point of active bleeding was detected in the anterior wall of the uterus. The perforation was repaired. Although the hemoglobin level dropped to 8 gm, a blood transfusion was not given. The patient left the hospital in stable condition with hematinics and outpatient instruction.

Case 4 LM, a 29-year-old female, para 6, grav 6, was admitted to Temple Hospital in September 1972. She had an abnormal Pap smear in February 1972. Shortly after a D & C and cervical conization revealed cervical carcinoma

JOU RNAL OF THE NATIONAL MEDICAL ASSOCIATION, VOL. 69, NO. 12, 1977

in situ. The patient subsequently underwent an abdominal hysterectomy in April 1972. After surgery the wound healed properly, but she had a rather profuse vaginal discharge. Two months after surgery the discharge increased and eventually urine began draining through the vagina. As an outpatient, she had an IV pyelogram which was normal. The bladder shadow was normal. On the post voiding film, extravasation into the vagina was noted. Cystoscopic examination revealed a vesico-vaginal fistula. The patient was subsequently readmitted for a fistula repair and the operation and postoperative course were uneventful.

Case 5 The patient, a 23-year-old para 1, grav 3, ab 1, was admitted to Temple Hospital for a therapeutic abortion because of her depressed state. The pregnancy was considered to be eight to nine weeks. During the course of the suction curettage, the uterus was ruptured. Exploratory laparotomy was recommended because of the size of the curet. Upon opening the peritoneal cavity about 1 00cc of bright red blood was noted. A perforation of the uterus was indeed found at the cervical corporeal junction just above the bladder reflection. The uterus was about tenweek gestation size. Since the abortion had not been completed from below, a ,small transverse incision was made and the remaining products of gestation were removed. The incision was closed along with the perforation site.

Case 6 This patient, a 40-year-old para. 2, grav 2, was admitted to Temple Hospital on April 10, 1975, complaining of urinary incontinence. She delivered, when 18 years of age, a child after an extremely difficult labor. The child died three months later of polycystic disease of the lungs. She lost a subsequent child delivered by cesarian section of the same disease. Because of menorrhagia and this genetic taint the patient had a supravaginal hysterectomy about five years later. She subsequently divorced her husband and approximately 12 years later, remarried. One and one half years after marriage, she suddenly be-

came totally incontinent of urine. In April 1973 a vesicovaginal fistula was repaired. In October 1974, following a severe coughing bout, she suddenly became incontinent again. At this time it was noted that she had a urethrovaginal fistula. In December 1974 this was repaired. A number 8 Foley catheter was left in the urethra with back-up from a suprapubic catheter. About one month later while straining at stool, she expelled the catheter and broke open her repair. She continued to drain until April 10, 1975 when she was admitted for another attempt at repair. At this time a plastic recontruction of the urethra was performed along with closure of the vesicovaginal fistula. She left the hospital dry with the aid of suprapubic and uretheral catheters.

Complications Anatomic Complications The pelvic surgeon is constantly aware of the potential anatomical complications which may occur in the course of the operation and in the

postoperative period.2 They can involve (1) blood vessels (2) nerves (3) the small intestine, large intestine, and appendix (4) the urinary tract (bladder, urethra, and ureters); and (5) retroperitoneal spaces.

Blood Vessels The single vessel most important to the pelvic surgeon is the hypogastric artery with its numerous and important visceral branches and 'associated venous plexus. At the origin of the hypogastric artery, where it branches from the common iliac, lies a most important landmark near the pelvic brim. The ureter courses over the pelvic brim so near the hypogastric artery that the inexperienced surgeon may be confused unless he watches for either pulsati'ons or peristalsis. At this point, the ureter may be identified and then manipulated as may be necessary in the course of the surgery. The hypogastric artery may be identified and ligated where there is otherwise uncontrollable pelvic arterial hemorrhage. The proximity of the ureter to the uterine artery in the base of the broad ligament can never be overemphasized. More serious compli-

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION, VOL. 69, NO. 12 1977

cations are encountered at this point than in any other area of pelvic surgery. Here, the ureter may be injured either by crushing, ligation, or transection. Also, at this point, control of the uterine artery may be lost. This leads to troublesome hemorrhage, hematoma formation, and injury to the base of the bladder, the ureter, or the venous plexus. Firm, well-controlled pressure with an empty sponge forceps may save the patient's life. When hemorrhage becomes so uncontrollable that one cannot ligate individual points, one must then rely on properly applied tamponade pressure and warm saline packs. If these fail, the experienced surgeon must fall back on maneuvers that he has learned through previous encounters and diligent observation. The ovarian vessels, arising from either the aorta or the renal arteries and coursing downward through the infundibulopelvic ligaments, must be respected because of the proximity of the ureter at the pelvic brim. When control of these vessels is lost, there is danger of formation of a troublesome retroperitoneal hematoma in the lateral gutters of the abdomen. The surgeon must remember that the infundibulopelvic ligament with the accompanying ureter may be greatly displaced or distorted in the case where extensive adnexal pathology is present.

Rectus Muscle Hematoma Subfascial bleeding resulting in a hematoma may be mistaken for an intraperitoneal disorder.3 Characteristic findings are exquisite tenderness and spasm, and a mass in the involved rectus muscle with no signs on the opposite side. Etiologic factors may include extensive undermining of the rectus during incision, placement of retractors beneath the muscles, suture of the muscle when the incision is being closed, and poor hemostasis. The hematoma should be evacuated, the muscle elevated, and bleeding vessels ligated. Anticipated management is almost never adequate.

Nerves Anatomic complications involving nerve fibers and trunks are rather rare in pelvic surgery, however, if they do 883

occur, they can be serious. Prolonged pressure by the blades of the deep abdominal retractor, the Deaver retractor, the Balfour selfretaining retractor, or the O'SullivanO'Connor self-retaining retractor can cause this injury. Great care must be used to insure that pressure from these instruments is not so prolonged in any area that the nerves can be injured. It is advisable for the surgeon to use as shallow a retraction as possible, consistent with adequate exposure, and to shift or relieve the pressure periodically during a long surgical procedure.

Small and Large Intestine and Appendix Anatomical complications which may be encountered in pelvic surgery involving the intestines or the appendix are usually incidental because the gynecologist rarely operates primarily upon the bowel. However, one must expect involvement of the intestinal tract at any time because of malignant and benign neoplasms, endometriosis, pelvic inflammatory disease, and congenital anomalies. Whenever this type of complication is anticipated, it is wise to have the bowel prepared pre-

operatively. Injury to a loop of adherent bowel, upon entering into the lower abdominal cavity, is not a rare occurrence. The chance for this is greatly increased by adhesions from previous operations or by the pressure of lower abdominal pathology. Recognizing this, the surgeon will be extremely careful in entering the peritoneal cavity, anticipating the possibility of injury and immediately recognizing it if it should occur. When it does occur, prompt closure must be done. This does not mean that the primary operation should be categorically discontinued. It is a test of the surgeon's technical ability and experience to avoid injury to the involved bowel or to recognize an injury and treat it. In the event that large segments of the bowel are unexpectedly involved to a depth where simple excision is not feasible, resection must be performed. Simple incidental appendectomy is frequently performed by the pelvic surgeon; however, the numerous vagaries of its position, size, and blood supply must be taken into account. It is disturbing and frustrating to the 884

surgeon who has completed a technically excellent pelvic operation to be confronted by a complication of the incidental appendectomy. Although there is little chance of injury to the rectum in the usual repair of a rectocele, complications may occur if perforation takes place or if ischemia and necrosis should result from too vigorous dissection or inaccurate reapproximation of the rectal musculature or the rectovaginal fascia.

The Urinary Tract Serious urinary tract complications for the pelvic surgeon are usually due to the intimate relationship of the urinary tract in the female true pelvis and the genital system. This is not surprising considering the close anatomical relationships and the physiologic, pathologic, and hormonal common denominators. The student of surgical anatomy and even the more experienced pelvic surgeon find it difficult to visualize accurately the relationship of the course of the ureter in its three dimensional progression from the kidney to the bladder. The following points may be useful to remind the operator of the presence of the ureter and its location in the course of the operation: 1. At the pelvic brim, it courses over the external iliac artery and vein approximately 1 cm lateral to the origin of the hypogastric artery. The surgeon may need to take advantage of the relationship at times to trace the ureter in its pelvic course if the pathology and anatomic relationships are confusing or if injury is suspected. 2. The ureter practically always remains attached to the parietal peritoneum in its course through the lateral aspect of the true pelvis and in the extraperitoneal space of the lateral pelvis. 3. The ureter tunnels through the base of the broad ligament approximately 2 cm lateral to the uterus and beneath the bridge of the uterine artery ("water under the bridge"). 4. The intramural portion of the ureter at the ureterovesical junction can sometimes be injured because of its intimate relationship with the vaginal canal and the bladder base. It must be remembered that very gentle dissection is needed in cases where mobilization of the lower ureters and

bladder must be done, not only to prevent direct trauma to the lumen, but also to preserve as much of the blood supply as possible. In the course of retropubic dissection toward the bladder neck, such as the Marshall-Marchetti-Krantz retropubic urethropexy, one needs to remember the closely involved relationships of the venous plexus about the bladder neck and urethra. Some complication can be encountered from toublesome bleeding. This bleeding is venous and usually can be controlled by pressure without ligation.

Urinary Tract: Prevention of Injury Prevention of injury to the bladder and ureter should be the aim of every pelvic surgeon. This requires first a thorough knowledge of the normal anatomy of the urinary tract and its possible anatomic variations, such as ureteral reduplication, pelvic kidney, or absence of a kidney (congenital or surgical). In other words, as much information as is practical should be obtained before pelvic laparotomy. This does not mean that every patient must have cystoscopy, intravenous pyelograms, ureteral catheters, and a bladder full of methylene blue before surgery. It does require, however, that sound judgment be used to determine which patient is likely to be technically difficult or have the urinary tract involved in the pathology to be treated, and is hence a candidate for trouble with the urinary tract at surgery. Certainly with large, probably adherent and fixed adnexal masses, or probable intraligamentory cysts or tumors, preoperative intravenous pyelograms are most useful in determining not only the normalcy of the urinary tract but also in revealing ureters in an abnormal position or any degree of ureteral obstruction. If the operator expects the case to be technically difficult with probable involvement of the ureters by inflammatory reaction, endometriosis, or tumor, or if preoperative pyelography indicates this to be the case, then cystoscopy with passage of a number 6 or 9 Teflon catheter up the ureter, immediately prior to laparotomy, is a wise precaution. Small catheters are difficult to feel and do not drain as well. These are kept in place by taping them to a Foley catheter placed in the bladder; each ureteral catheter should

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION, VOL. 69, NO. 12, 1977

be identified (right to left) so that the side can be identified should the need arise.

Ureteral Injuries Gynecologic surgery is responsible for most ureteral injuries.4 The "easy" operations, the simple abdominal hysterectomy, and not the technically difficult pelvic one, are responsible for most ureteral injuries. Total abdominal hysterectomy accounts for almost 50 percent of the genitourinary fistulas and perhaps 80-90 percent of all surgical ureteral injuries. This problem will persist until a most important surgical axiom is applied routinely during the accomplishment of all pelvic operations. During all dissections, the contiguous structures subject to injury must be exposed. This step will not only avoid injuries to the ureter but also will facilitate an equally important aspect, that is, urinary-tract injuries must be recognized at the time of operation. With recognitions and adequate repair, problems such as fistula formation and serious morbidity and subsequent litagation can be avoided almost entirely. Because the gynecologic surgeon frequently will find that urologic consultation is not available at the time of urinary-tract injury, he or she must be aware of and familiar with the various u r e teral reconstruction procedures that may be required. The gynecologic surgeon must devote time and study to the management of urinary tract injuries before their occurrence. All pelvic surgeons will eventually encounter ureteral problems. The methods of bladder molilization and ureteroneocystostomy should be within the ability of all who operate within the pelvis. When extensive damage has occurred and a urologist is not available, the gynecologist who is unfamiliar with the more demanding techniques (ureteroureterostomy, bladder flaps, ileal conduits) should avoid additional damage to the urinary tract and accomplish a simple catheter ureterostomy deferring definitive repair to a urologist.

Prevention of Urinary Fistulas "Prevention is better than cure" is a self-evident truth that might have been written on the subject of urinary

fistulas. In regard to ureteral injuries it seems that the gynecologist must, at all times, satisfy himself that the ureter is clear of his clamp, knife, or scissors. Throughout its whole length in the pelvis, the ureter may be injured. Uterine and tubo-ovarian swelling which grows out into the tops of the broad ligament will almost certainly have the ureter pressed against their lower or lateral margins. The isolation of the ureter in these cases is best accomplished by dividing the round ligament (gateway to the broad ligament) cutting the peritoneum backward, and then with the fingers dissecting downward and backward to separate the mass from the great vessels in the side wall of the pelvis. After this, the broad ligament is opened and the ureter can be readily identified and kept clear with a thumb inside and the fingers behind the posterior leaf. At the lower end of the broad ligament, the ureter can be avoided if in pushing the bladder clear of the cervix, one reminder is to clear the vaginal angles. In order to minimize the risk of injury, they should be palpated in every case.

Ureteral Injuries during Surgery In recent years, there has been an apparent increase in the incidence of operative ureteral injuries despite a longer training period for surgeons.5 Ureteral injuries may result from complete or partial ligation by a suture, inadvertent application of hemostats, complete or partial transection, excision of longitudinal segment of ureter, compromise of ureteral blood supply, or kinking by a ligature. The most frequent injuries appear to be complete ligation and complete transection. Patients who have flank pain after operation, whose temperature rises, and in whom adynamic ileus or abdominal distension develop, should be suspected of sustaining ureteral injury and radiographic studies should be carried out immediately. Usually cystoscopy and passage of ureteral catheters are essential to determine the nature of the lesion and to determine if intraperitoneal or extraperitoneal extravasation is occurring. The ideal time for treatment is during the initial operation. Most injuries not reported at this time may be repaired during convalescence. Nephrostomy is

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION, VOL. 69, NO. 12, 1977

the simplest surgical procedure when bilateral ureteral ligation is diagnosed early in the postoperative period. To avoid ureteral injury, the surgeon should be thoroughly familiar with the areas where the ureter most often is injured and should handle the ureter with extreme care. An urogram should be obtained preoperatively. When injury is detected at operation, end-to-end anastomosis is the procedure of choice if technically possible, otherwise reimplantation of the ureter into the bladder is recommended.

Preoperative Excretory Urography Preoperative urography can provide information regarding the relationship of ureters to abnormal pelvic structures. Although urography may detect a congenital anomaly such as a double ureter, a pelvic kidney, ureteral obstruction, or perhaps intrinsic urinary tract disease, the procedure is of little help in preventing ureteral injuries during surgery. Generally, the patient with ureteral injury has had a technically easy hysterectomy for minimal disease, with production of minimal distortion of pelvic anatomy. That is, the excretory urogram usually shows the ureter in normal position. Regardless of the radiographic demonstration of the position of the ureter, the surgeon must identify the ureters above the level of the disease and throughout their pelvic course, and must be aware of their position throughout the operation.

Preoperative Placement of Ureteral Catheters Preoperative insertion of ureteral catheters may enable the surgeon to avoid ureteral injuries or at least facilitate their recognition. Although this has been advocated repeatedly during the past 30 years, largely by urologists who so frequently are responsible for repairing ureteral injuries, it has not become a particularly accepted or effective method of avoiding injury, as noted by Higgins.6 This remains true for several reasons: 1. Insertion of catheters into the ureters of all patients subjected to pelvic surgery is inconvenient and even undesirable; 2. Accurate preoperative selection 885

of patients in whom ureteral catheters might provide assistance is impossible; 3. The simple gynecologic operation, not the anticipated difficult dissection, accounts for most ureteral injuries; 4. Catheters give the surgeon a false sense of security by allowing him or her to think that there is no need to search for and avoid the ureters; 5. With fibrotic pelvic conditions (chronic pelvic inflammatory disease, endometriosis, prior irradiation) that make ureteral identification difficult, the lower portions of the ureters cannot be palpated even with a catheter in place, although their dissection and exposure remain essential; 6. Dissection of the ureter in the presence of a hard unyielding intraureteral catheter increases the trauma to the ureter; and 7. During a difficult pelvic dissection, when the identification or integrity of the ureter is in doubt, the most logical expeditious measure of assistance is an interim (extraperitoneal) cystotomy. Cystotomy can be performed quickly, without risk, and allows prompt identification of ureteral orifices for the insertion of ureteral catheters. This is considerably less cumbersome than the routine preoperative (or the occasional intraoperative) insertion of ureteral catheters by the urologist via cystoscope.

Ureterovaginal Fistula The main objection to radical hysterectomy and pelvic node dissection as a method of primary treatment for cervical carcinoma is the inherent danger to the ureters that may result in ureterovaginal fistula formation.7 From July 1969 to the present, the following changes in operative technique have been effected. After completion of radical hysterectomy and lymph node dissection, the vaginal vault is closed. Large plastic or rubber catheters are placed retroperitoneally into each obturator fossa and brought out through stab wounds in each lower quadrant and connected to continuous suction to drain lymphatic fluid and blood. Surgeons have devised procedures as guides for lessening the risk of postoperative ureteral vaginal fistula from devascularization. Green et a18 recommend ureteral suspension. This is car886

ried out so that the lower 5 to 6 cm of the denuded ureter is sutured to the superior vesical artery by means of interrupted 5-0 chromic catgut sutures with an atraumatic needle. This replaces the ureter in a gently curving near normal course and anchors it adjacent to the sidewall and well above the pelvic hollow. The position of the adjoining artery may also improve the ureter's peristaltic action. Complete and meticulous reperitonealization of the pelvis is always carried out. The adnexa are only removed when diseased and in peri or postmenopausa-l patients. A suprapubic catheter is used because it is better tolerated; spontaneous voiding time is shorter; and there is lower incidence of urinarytract infection. Furthermore, repeated catheterization may be avoided as residual urine may be measured by means of the three-way stopcock, before the suprapubic catheter is removed completely. The ureterovaginal fistula is usually located in the lower third of the ureter and characteristically manifests itself between 12 and 18 days postoperatively. Once the leakage of urine occurs in the vagina, distinction between the ureterovesicovaginal and the vesicovaginal fistula can be made by the dye test. Localization of the affected side is more difficult and both ureters must be tested to avoid repair of the unaffected side. When the diagnosis of ureterovaginal fistula is established, it is generally agreed that the preferred treatment is ureteroneocystotomy as the bladder wall furnishes an excellent blood supply to the reimplantation site. End-to-end anastomosis of the ureter is not advocated if ureteroneocystotomy is possible, for it may result in breakdown necrosis or ureteral stricture. It is difficult to be precise as to the time of repair of the fistula, but most authors currently advocate waiting for two to six months. Waiting for at least six months after operation has resulted in, not only, spontaneous healing of the fistula but also better surgical repair. Prior to 1965, a postoperative fistula was repaired as soon as the patient's condition permitted. But the operative difficulties encountered at that time were extreme, with prompt breakdown of the repair in the majority of patients. After 1965, repair was done after a delay period of at least six

months provided acute problems (eg, rapid progression of azotemia or obstructive uropathy) did not develop. When necessary temporary nephrostomy is undertaken to delay the operative procedure, the perineum is kept dry by a Tassette cup or a Foley catheter.

Urethrovaginal Fistula The urethrovaginal fistula is usually not so frequently seen, comprising 15 percent of the urinary tract fistulas in the vagina.9 Necrosis and fistula formation may follow anterior colporrhaphy without proper attention to surgical techniques. It may be related to thin flaps, wide dissection, inadequate blood supply, hematoma, infection, or to suture suspension of the bladder neck. Repeated intermittent catheterization for long periods promotes fistula formation. Excessively long catheterization is preferable to too early removal of the catheter. Customarily, drainage is continued for seven to nine days. Most failures after excision of ureteral diverticulae are related to infection. The repair of the urethrovaginal fistula may be tedious but usually not so tedious as the high vesicovaginal fistula. The principle error that can be made in the urethral repair is improper closure of the urethral ostium, and excessively wide mobilization of the urethra prior to closure. It must be remembered that since the urethral circumference is rather limited, wide mobilization may lead to necrosis or actual mechanical obstruction. A safeguard against this is the use of a 20F indwelling urethral catheter during and following the repair. This is not too large to distend the urethra excessively but large enough to show the defect and also to allow adequate postoperative drainage. Antibiotics or sulfonomides of choice are given when an indwelling catheter is in place 72 hours or longer.

Vesicovaginal Fistula The cause of the vesicovaginal fistula is probably a consequence of local bruising, too much handling of the tissues, vigorous blunt dissection, infection, hematoma formation, and the faulty placing of sutures. If gynecologists and others in training can be

JOURNAL OF THE NATIONAL MEDICAL

ASSOCIATION, VOL. 69, NO.

12, 1977

helped to acquire a "feeling" for tissue and a constant awareness of the importance of the blood supply, then something may be done to prevent these injuries. In particular, over enthusiastic surgery must be avoided on tissues possibly already devitalized because of previous surgery and radiotherapy or even advancing years.

General Principles of Fistula Repair There are a few principles of proven value. The general physical condition of the patient and the tissue itself must be optimal - even if this means postponement of the repair. Although the surgeon may be anxious to get on with the task, particularly if he has had pressure put on him by the patient and her family, he must realize that the chance for success is greatly enhanced by choosing the proper time to operate. If the fistula had not been detected at the time of the error and repaired immediately, postponement of the repair for the proper length of time is of paramount importance. The length of this time varies in the opinion of many experienced surgeons. TeLinde advocates waiting as long as six months before the attempt at repair is made. Moir' 1 suggests that at least three months be allowed to elapse, and Counseller12 had also suggested this same length of time. The patient must be properly prepared psychologically for the vagaries associated with successful repair of a fistula. The gynecologist should clearly explain the odds for and against chances of success with the first attempt at repair. The patient must not be given false hope of sure success nor should she be deprived of what hope for success does exist.

Hemorrhage and Infection Vaginal hysterectomy is notoriously more likely to give rise to p o s t o p erative infection than abdominal hysterectomy13 and the risk of postoperative hemorrhage at an early stage is also more common. For these reasons much attention has been devoted to means for preventing these problems or at least reducing them to an absolute minimum. During surgery the nurse and anesthesiologist are in a much better posi-

tion to estimate blood loss than the surgeon or his assistant. If a surgeon repeatedly incurs a blood loss greater than 300 to 400 cc, he should ascribe it to faulty technique or poor selection of cases and take the necessary steps to re-educate himnself. Blood loss greater than 500 cc should be replaced even if the patient starts with a high hematocrit volume, or the bleeding tendency may increase. An attitude of allowing continuous oozing of blood from denuded surfaces on the basis that it is uncontrollable is to be condemned. Heavy packing at the conclusion of the operation is no substitute for good technique and may cram blood back into the retroperitoneal space or the peritoneal cavity instead of allowing it to emerge from the vagina where it can be seen and treated. At least two postoperative evaluations of the hematocrit volume, one within 24 hours postoperatively and the other two days later, gives the best assessment of blood loss and should be routine.

Adequate Pelvic Exposure In order to avoid injury to the urinary tract during pelvic surgery, adequate pelvic exposure throughout the operation is necessary. In the presence of any significant pelvic disease, a small incision (whether transverse or vertical) does not provide adequate exposure for the careful dissection of the pelvic structures at risk. An inadequate incision leads to timid and inadequate local dissection, inadequate identification, and inadequate mobilization of ureters, bladder base, and rectum. This promotes blind clamping, blind suturing, and poor surgical technique in general.

Sepsis One of the most feared complications of vaginal or abdominal surgery is sepsis. This occurs most commonly when blood loss has been heavy during the operation or when blood loss has accumulated behind the closure. The age and condition of the patient, the duration of the procedure, and the quality of technique in avoiding massive necrosis of tissue are all of the utmost importance as predisposing factors. Preexisting inflammation in the fallopian tubes may lead to acute

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION, VOL. 69, NO. 12, 1977

salpingitis when the tubes are severed from the uterus and left in situ. Much controversy has arisen in recent years as to the value of prophylactic antibiotics in reducing the incidence of pelvic abscess and salpingitis p ostoperatively. 1 4 Unfortunately,1 5 the available studies have defects in experimental design or the patients have been inadequately followed. Good criteria as to what constitutes a diagnosis of infection have not been agreed upon. As a result they do not provide definitive answers as to the desirability of using antibacterial medication on a routine basis beginning just prior to or during surgery, nor have they established rational guidelines as to the necessary dosage and duration of administration of drugs to be used. In addition, the advantages of prophylaxis must be weighed against drug reactions, the danger of super-infection by secondary invaders, and the development of resistant strains among the hospital and community populations. From a practical viewpoint, it seems logical to begin treatment, using therapeutic dosage levels at the time of surgery, and to continue it for not more than 72 hours. This will provide effective suppression of the organisms which enter at the time of surgery and will not last long enough to foster resistance or obscure the diagnosis of smoldering occult infections. In most cases, when the risk of infection is not great, patients should be individualized and infections treated therapeutically after documentation by culture and physical examination. There are three groups of organisms we must consider when treating patients with severe pelvic infections: the aerobic enteric bacilli, especially E. coli, the enterococci, and the anaerobes. Prior to our concern about anaerobes, seriously ill patients were often treated with penicillin plus an aminoglycoside (Streptomycin, Gentamycin, or Kanamycin). Such a regimen provides excellent coverage for enterococci, gram-negative enteric bacilli, and practically all anaerobes except for Bacteroides fragilis, which are sensitive only to clindamycin and chloramphenecol. We are beginning to recognize that the bacteriology of pelvic infections involves a complete ecology of multiple micro-organisms. Because of the potential toxicity of clindamycin and chloramphenecol, as 887

well as the protracted course of Bacteroides infections, the addition of one of these drugs can generally be deferred until its use is dictated by microbiologic data or a failure of clinical response. Persistent fever generally indicates a lack of the appropriate antibiotic, an undrained collection of pus, or concomitant pelvic thrombophlebitis.

Conization of Cervix According to the literature there is a relation between the incidence of complications and the interval between cone biopsy16 and hysterectomy. Data reviewed on 230 cases with cone biopsy followed by hysterectomy indicate this. The interval from cone biopsy to hysterectomy ranged from two days to two months, but nearly all operations were done within a month after cone biopsy. Pelvic complications occurred in 22 patients. In all instances, they followed surgery performed four to 14 days after cone biopsy. Hematoma or abscess in the abdominal wound developed in 26 patients. Only a few wound complications occurred in patients having hysterectomy more than two weeks after cone biopsy. A febrile postoperative course was found in 58 percent of cases. Fever most frequently followed hysterectomy done four to 14 days after biopsy. Hysterectomy should be postponed for three to four weeks or preferably five to six weeks after cone biopsy. Otherwise, the operation should be performed within 48 hours after cone biopsy. The best procedure appears to be cone biopsy and hysterectomy in the same stage. This shortens the hospital stay and avoids the common complicatiops of cone biopsy, particularly secondary hemorrhage. This approach requires frozen section examination of cone biopsy specimens.

monary embolization may be fatal. Because of the frequency with which thromboembolic phenomena are encountered in patients who have undergone pelvic surgery via the abdominal route, the use of anticoagiilants after surgery is worthwhile. Studies have shown considerable decrease in thrombophlebitis and pulmonary emboli under such a routine. Although risk is involved with the use of any medication, care and common sense will minimize complications. When thrombophlebitis occurs, prompt treatment is indicated. Any elevation of temperature after operation should be evaluated. Appropriate work-up should include checking the legs for evidence of thrombophlebitis and the calves for tenderness and for Homan's sign. Varicosities or redness or tenderness of specific vessels may be noted and when a clinical diagnosis is made, treatment should be started

immediately. If the patient has not been treated with anticoagulants, administration of heparin and warfarin (Coumadin) should be started simultaneously. The heparin should be continued until the Coumadin reaches therapeutic ranges. In all instances, the blood must be monitored daily by the prothrombin time for the Coumadin and clotting time for the heparin. If the diagnosis of pulmonary embolization is suspected, roentgenograms of the chest, and electrocardiogram and preparation for embolectomy should be made immediately. Arteriograms may be of value, but should not delay pulmonary embolectomy. Immediate thoracic consultation is required and effective teamwork must exist if the patient is to survive the result. Consideration must be given clipping or ligation of the vena cava to prevent subsequent embolization.

Uterine Perforation During D & C

Thromboembolic Disease Thromboembolic disease is an inclusive term for vascular clotting problems and their sequellae. Virchow's triad - venous stasis, hypercoagulability, and endothelial injury - is established as the etiologic basis for this disease. Postoperative complications of thrombophlebitis and pul888

McElen reviewed a computerized study of 299 instances of diagnostic dilatation and curettage performed during a 20-year period has revealed that recognized uterine perforation occurred in 19 instances. 7 This represents a perforation incidence of 0.63 percent. Although infection or uterine bleeding following a curettage is dis-

turbing to the gynecologist, accidental uterine perforation is usually more disconcerting and alarming because of the obvious iatrogenic implications. If the reported grotesque instances of intra-abdominal mutilation are excluded, we feel the gynecologist can be comforted by the knowledge that the seriousness of this complication is not great, provided a simple perforation is immediately recognized and properly managed.

Elective Abortion Experience is reported with elective first trimester abortion of 16,410 pregnancies during a 31-month period by Reproductive Health Services of St Louis,1 8 a free standing clinic. Incidence of complications was 1.54 percent. The most common of these were incomplete evacuation, excessive postabortal bleeding, and uterine perforation. In problems with unquestioned perforations the use of laparoscopy has been very valuable in determining the exact nature of a perforation, in avoiding the unnecessary laparotomy, and in giving intra-abdominal visual guidance to concomitant suction evacuation in cases of an incompleted procedure. This implementation occurs because any procedure that attempts to empty a pregnant uterus is a "blind" one and it is impossible to be certain that all tissue is removed. In spite of attempts to make the evacuation method as thorough as possible, incidence of this complication remains high. The symptoms are those of retained tissue, severe bleeding, and cramps that are not relieved by oral medications. Because of this close follow-up, these patients usually return to the clinic for a second evacuation. Thirty-one percent have evacuations performed by their own physicians elsewhere. Perforation of the pregnant uterus is an ever present hazard in the performance of abortion, and occur in 34 patients in the present study. Difficult cervical dilation and acute displacement of the uterus alert the physician to the possibility of the danger. The passage of the uterine sound cervical dilator and the suction curette is therefore done more cautiously in such patients. Although a number of these patients have repeat evacuations in the clinic after a period of observation, the

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION, VOL. 69, NO. 12, 1977

use of laparoscopy provides a more accurate assessment of the injury and allows a safer repeat evacuation of the uterus in the presence of this complication. When the abortion is completed by vaginal suction, together with laparoscopic observation of the uterus, the laparoscopist can guide the suction curette away from the perforated area watching the tenting of the uterine wall by the curette and the contraction of the uterus as it is emptied. It is, thus, well established that the morbidity and mortality associated with first trimester abortion is much lower than those performed in the second trimester. The results of this study suggest the difference seen within the first trimester revealing that the earlier in pregnancy the abortion is performed the less chance there will be of major complications. The incidence of ectopic pregnancy in the study is low (4 per 10,000 pregnancies) this is a condition which must always be considered at the time of the abortion. The presence of chorionic villi on microscopic study of the evacuated tissue and a negative cul-desac tap are helpful.

Obesity Obesity, especially when the abdomen is large, will pose some frustrating problems to even the most skilled technician. Tumor excision will be less complete and hemorrhage will be poorly controlled. Guidelines in cases of excessive obesity are elusive. The actual difficulty encountered may only become evident after laparotomy since the proportion of exagenous fat varies. In addition the obese patient is subject to cardiovascular strain which would be aggravated by the operation. During recovery the presence of excess fat conceals the development of intraabdominal complications. Little data exists to define the nature and extent of the risks imposed by obesity in gynecologic surgery. Pitkin19 reviewed 25 years of experience with obesity in relation to the outcome of a single operation, total abdominal hysterectomy, at University Hospital gynecological service. Data on all patients having this operation in 1948 (73 who weighed 200 pounds or more on admission) were reviewed. Non-obese controls were matched with

300 obese patients. Surgical techniques were fairly constant throughout the review period. The obese patients were about four years older on the average than the controls. The respective mean weights were 229 and 139 pounds. Endometrial carcinoma was nearly three times more often the indication for operation in the obese group. The only other diagnosis which was significantly more frequent in obese patients was endometrial hyperplasia. About half the obese patients were hypertensive compared with 28 percent of the controls. Diabetes was found in 17 percent of the obese and six percent of the nonobese. The mean operating time was 3.1 hours on obese patients and 2.4 hours on the controls - a significant difference. Blood loss was excessive in 47 percent of the obese patients and in 24 percent of the nonobese. Three obese patients died postoperatively, compared with none of the nonobese. The deaths were due to pulmonary embolism, myocardial infarction, and presumed malignant hyperpyrexia associated with myotonia congenita. Over 35 percent of the obese patients were hospitalized for over 12 days after operation, compared with 11 percent of the nonobese - a significant difference. Abdominal hysterectomy is associated with increased risks in obese women. Obesity itself should rarely, if ever, contraindicate necessary surgery. However, in situations in which surgery is more elective, its complications should be kept in mind.

Gynecologic Surgery on Elderly Patients In order to evaluate the risks associated with gynecologic operations on elderly patients (60 years of age or over), the authors performed this retrospective study of 573 operations.20 More than half the patients (53 percent) were operated on because of uterine prolapse. An abdominal operation was done in 33 percent, a vaginal approach in 59 percent, and a combined abdominal and vaginal procedure on nine percent. The total amount of postoperative complications was 26 percent and higher after abdominal (36 percent) than

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION, VOL. 69, NO. 12, 1977

vaginal operation (19 percent). Most complications were mild, eg, fever (16 percent) and wound complications (four percent). Five patients (0.87 percent) died within one month postoperatively. But only in two cases was death caused specifically by postoperative comnplications, namely cardiac infarction and pulmonary embolism. It appears that best results can be obtained with cooperation between the gynecologist, internist, and anesthetist. The chronologic age in itself is only seldom a contraindication to operative treatment.

Death As is the case of any surgical procedure, the ultimate complication is death. Fortunately in gynecologic surgery, where most of our procedures are elective and on fairly good risk patients, this is an infrequent complication. The exact mortality figures across the country are impossible to ascertain, since late casualties may escape the notice of the surgeons and reported figures are usually derived from large services where quality of care is excellent. On a first class service an estimate of 6.1 percent would probably be reasonable.2 1 Death may occur on the operating table as a result of factors beyond the control of the surgeon, but every surgeon must be ready to play his role in the prompt resuscitation of patients suffering from cardiac arrest, coronary occlusion, respiratory paralysis, etc. Postgraduate education in all hospitals must include the regular and constant repetition of the measures necessary to save the patient in an emergency and the individual responsibilities of the surgeon and the other members of the operating team. Deaths in the hospital during the postoperative course are usually the result of hemorrhage, infection, embolism, or intercurrent disease; later deaths can sometimes be correlated with urinary tract damage and/or pyelonephritis which had its conception at the time of surgery.

Comment Even the most skilled and mature 889

surgeon will inevitably make errors. There are such variations in human anatomy that no amount of previous experience and knowledge will prevent this from being a factor in contributing to errors. Also, there are individual responses of tissue healing and strength that are completely unpredictable. The presence of one or all of these factors in a given surgical problem may contribute to the failure or to the less than satisfactory end result of the operation. These self-induced complications are possibly the most disturbing because of direct reflection upon the skill and judgment of the surgeon. This in turn may produce a sense of frustration, guilt, and disappointment. Even more seriously, such casualties promote invitations for litigation in malpractice suits. In the past two decades this has been increasingly true, to the degree of causing the surgeon to frequently compromise his better judgment of technique and patient care. The most common error to be anticipated with abdominal hysterectomy is failure to recognize the various tissues of the pelvis, bladder, bowel, ureter, and blood vessels. Each has different characteristics, but disease may have distorted the anatomic position of the various organs as well as their identifying features. The most serious error is not to identify and repair damage to these organs. The safeguard is knowledge of the surgical area and avoidance of blind manipulation in unrecognized areas. The anatomic position of the bladder subjects it to increased risk during pelvic operations. Inasmuch as pelvic operations require low abdominal incising, the relationship of the bladder to the lower portion of the abdominal wall necessitates extreme awareness. This is true both in making and repairing the incision. The intimate relationship between the cervix and the bladder also requires gentle dissection to avoid trauma during the required separation of these organs. The most common errors in abdominal hysterectomy which may lead to major complications are as follows: 1. Damage to the bowel or bladder when making incision due to failure to recognize the differences in these tissues; 2. Damage to the bladder when freeing the uterus from the bladder, particularly when there is scarring 890

from serious injury or disease; 3. Damage to the ureters while dissecting and clamping the infundibulopelvic ligament caused by distortion with tumor or previous surgery; 4. Damage to the ureter when, in clamping cardinal or uterosacrals, clamps are placed extrafascially, or too far from the cervix; 5. Damage to the rectosigmoid when clamping the uterosacral ligaments; 6. Damage to the ureters or bladder while peritonizing by including these organs in the sutures; 7. Adhesions and obstructions from failure to peritonize and cover raw surfaces; 8. Hernia or fistula from inclusion of bladder or bowel in the peritoneal closure; 9. Hematoma or abscess with failure to obliterate dead space and obtain food hemostasis with incisional closure; and 10. Failure to recognize and repair any damage which may have occurred. Every surgeon has his favorite operation, technique, maxims, and superstitions and as long as these are being used in the best interest of the patient, all is well. The risk is that he may lose flexibility, assume a posture of omnipotence, and become careless about the application of his expertise. Rumors of accidents, complications, and late results must be regularly received by the surgeon and his peers to determine whether recognized standards of competence are being maintained.' Our medicolegal consultant, John L. Moore, Jr, advises us that the competent physician should not be overly concerned about malpractice exposure. He claims that if a physician is well trained, hard working, and not the touting or promoter prototype of an unethical professional, court and jury alike will differentiate very well between unprofessional negligence on the one hand and accident or error due to anatomic vagaries or poor tissue healing or failure of the patient to adhere to instructions, on the other. In summary, the fact that an error has been made in surgical management is not usually res ipsa loquitur of negligence, inexperience, or indifference. It most frequently is related to those many variable factors in conscientious and skilled surgical care that are beyond the control of the operator.

Acknowledgement The author wishes to express appreciation to Mrs. Adrianne Bass and Mrs. Bobbie Leonard, Medical Record Department, Queen of Angels Hospital; to Mrs. Mary Foster and Mrs. Susanne Deitch, Queen of Angels Staff for research and typing; Mrs. Marilyn Steifel, Medical Record Department, Temple Hospital; and to Mrs. Rosa Moore, Charles Drew Post Graduate Medical School for research, and my thanks for their assitance in the preparation of this paper.

Literature Cited 1. TeLinde RW: Errors, safeguards, salvage. In Ridley JH (ed): Gynecologic Surgery. Baltimore, Williams and Wilkins, 1974, p 297 2. Ridley JH: Gynecologic surgery: Anatomical complications of pelvic surgery. Amer Surg 42:706-712, 1971 3. Schiffer MA, Hellman LM: Rectus muscle hematoma or complication of gynecologic and obstetrieal procedures. Obstet Gynecol 35:231-234, 1970 4. Symmonds RE: Ureteral injuries associated with gynecologic surgery: Prevention and management. Clin Obstet Gynecol 19(3):623-644, 1976 5. Higgins CC: Ureteral injuries during surgery. JAMA 190:82-88, 1967 6. Higgins CC: Ureteral injuries during surgery: A review of 87 cases. JAMA 199:118, 1967 7. Macaset MA, Nelson JH: Ureterovaginal fistula as complication of radical pelvic surgery. Am J Obstet Gynecol 124:757-760, 1976 8. Green TH, Meigs JV, Ulfelder A, et al: Urologic complications of radical Wertheim hysterectomy: Incidence, etiology, management, and prevention. Obstet Gynecol 20:293-312, 1962 9. Gray LA: Ureterovaginal fistula. Am J Obstet Gynecol 101:28-36, 1968 10. TeLinde RW: Textbook of Operative Gynecology, ed 4. Philadelphia, JB Lippincott, 1970, p 588 11. Moir JL: The Vesico Vaginal Fistula, ed 2. London, Bailliere, Tindall, Cassell, 1967, p 55 12. Counseller VS: Surgical and postoperative treatment of large vesicovaginal and rectovaginal fistula. Surg Gynecol Obstet 74:738-747, 1943 13. Ledger WJ: Postoperative pelvic infections. Clin Obstet Gynecol 12:265-282,

1969 14. Allen JL, Rampone JF, Wheeless CR: Use of prophylactic antibiotics in elective major gyriecologic operations. Obstet Gynecol 39:218-224, 1972 15. Ledger WJ, Sweet RL, Headington JT: Prophylactic cephaloridine in the prevention of postoperative pelvic infections in premenopausal women undergoing vaginal hysterectomy. J Obstet Gynecol 115:766-774, 1973 16. Skarrup P, Berget U, Syzepenski MK: Incidence of complications following hysterectomy in relation to time interval between cone biopsy of cervix and hysterectomy. Acta Obstet Gynecol Scand 50:3!1-324, 1971 17. McElen TW: Study of uterine perforation occutring during 299 instances of diagnostic curettage. Int J Obstet Gynecol

7:243-251, 1969 18. Georee J, Wulff L, Jr, Michael S: Complications seen in a free standing clinic. Freeman Obstet Gynecol 49:351-357, 1977 19. Pitkin RM: Abdominal hysterectomy in obese women. Surg Gynecol Obstet 142:532-536, 1976 20. Swondja L, Ylekorhala 0, Jarvinen PA: Gynecologic surgery in elderly patients. Ann thir Gynaecol Fenn 64(6):388-393, 1975 21. Gray LA: Indications, techniques, and complications in vaginal hysterectomy. Obstet Gynecol 28:718-722, 1966

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION, VOL. 69, NO. 12, 1977