Obstetric

The underlying factors responsible for the development of obstetric fistulae may be considered in physical, biosocial, cultural, and geographical or political terms. The basic physical factors responsible for obstetric fistula development include obstructed labour, accidental injury at the time of caesarean section, forceps delivery, craniotomy, symphysiotomy, traditional surgical practices including circumcision and ‘gishiri’, and complications of criminal abortion.

The overwhelming majority are complications of neglected obstructed labour. During normal labour, the bladder is displaced upwards and the anterior vaginal wall, bladder base and urethra are compressed between the fetal head and the posterior surface of the pubis. No harm results if this occurs for a short time, but in prolonged obstructed labour, the intervening tissues are devitalized by ischaemia. Usually, the anterior vaginal wall and underlying bladder neck are affected, although the area of necrosis is sometimes higher, in which case the anterior lip of the cervix and underlying trigone are involved. Compression of the soft tissues between the sacral promontory and the presenting part may occur at the same time, with necrosis at the posterior vaginal wall and underlying rectum. The devitalized area separates as a slough usually between the third and 10th day of the puerperium, with resulting incontinence (Figure 57.2).

Figure 57.2 Puerperal patient following prolonged obstructed labour. An area of devitalized tissue is seen on the anterior vaginal wall, about to slough with resultant fistula formation.

Accidental injury to the vaginal wall during a difficult operative delivery may involve damage to the underlying bladder wall, particularly if the tissues are devitalized by prolonged pressure. Forcible rotation of the head with Kielland’s forceps is particularly liable to produce such an injury by shearing stresses. The bladder is particularly exposed to injury following symphysiotomy if the pubic bones are too widely separated by forced abduction of the thighs. In these circumstances, the unsupported bladder neck is very likely to be damaged, especially if the head is rotated and extracted with forceps. The posterior wall of the bladder may be accidentally incised during lower segment caesarean section or repair of ruptured uterus, particularly if the bladder is not reflected sufficiently far downwards before the lower segment is opened. During the reflection itself, the bladder may be torn, especially if a previous operation has made it densely adherent to the lower segment. If the bladder injury goes undetected at the time, urinary leakage through the abdominal wound soon develops. This is usually followed by incontinence vaginally when the urine finds its way through the uterine incision. The abdominal leakage then dries up as the bladder drains through the resulting vesicocervical fistula. Alternatively, sutures may be passed through the posterior bladder wall during repair of the uterine incision. The appearance of urinary incontinence in these cases is delayed until the intervening bladder tissue caught up in the suture sloughs.

The perineum and posterior vaginal wall are, of course, at risk from even the most straightforward delivery, although primiparity, forceps delivery, birth weight over 4 kg and occipitoposterior position have been found to be significant risk factors associated with third-degree tears (Sultan et al 1994). Even when identified and repaired, this increases the risk of rectovaginal fistula.

Traditional surgical practices are undertaken in up to 98% of women in some parts of Africa (United Nations Population Fund 2007, World Health Organization 2008), and may play a role in the aetiology of both direct ‘surgical’ and obstetric fistulae. The practices of ‘angurya’ and ‘gishiri’ cutting [female genital mutilation (FGM) type IV] (World Health Organization 2008) are commonly employed to treat a wide variety of conditions including obstructed labour, infertility, dyspareunia, dysuria, amenorrhoea, goitre, backache and presumed tumour. The traditional cut made with a razor blade or knife through the vaginal introitus is sometimes superficial, but may result in fistula formation (Figure 57.3). Since cuts are usually made by linear incision into healthy tissues, repair is often much easier than for those resulting from pressure necrosis. Tahzib (1983, 1985) reported on the epidemiological determinants of vesicovaginal fistulae in northern Nigeria. In 84% of cases, obstructed labour was the major aetiological factor, 33% had undergone ‘gishiri’, and this was felt to be the main aetiological factor in 15%.

Figure 57.3 Vesicovaginal fistula resulting from ‘gishiri’ cutting (female genital mutilation type IV).

Female circumcision has been practised in various forms in much of North Africa, being most prevalent in Ethiopia, Eritrea, Sudan, Egypt, Mali and Guinea, where currently over 75% of women are affected (World Health Organization 2008). The more extreme forms (previously referred to as ‘pharaonic circumcision’) involve removal of the labial minora (FGM type IIa), and may include removal of the clitoris (FGM type IIb) and labia majora (FGM type IIc) (see Figure 57.4). In FGM type III, the introitus may be reduced to pin-hole size by excision and apposition of the cut edges of the labia minora or majora (infibulation) (World Health Organization 2008). Whilst an association between FGM and both vesicovaginal and rectovaginal fistula is recognized (Lovel et al 2000), the strength of the association and the exact mechanisms are less certain and are currently under investigation (World Health Organization 2008).

Figure 57.4 Excisional type of female genital mutilation, with excision of labia minora and most of clitoris and labia majora (female genital mutilation type IIc).

In considering obstetric fistulae, it is important to consider the social, cultural and geographical influences just as much as the direct physical injury to the lower urinary tract. Obstructed labour is most often due to a contracted pelvis. This usually results from stunting of growth by malnutrition and untreated infections in childhood and adolescence. Where women retain a subservient role in society and standards of education are limited, early marriage and the absence of family planning services result in an early start to childbearing; where first pregnancies occur soon after the menarche, before growth of the pelvis is complete, this also contributes to obstruction in labour.

The influence of these factors is illustrated in the epidemiological studies alluded to earlier. Tahzib (1983) reported that over 50% of the cases of vesicovaginal fistulae seen in northern Nigeria were under 20 years of age, over 50% were in their first pregnancy, and only one in 500 had received any formal education. From the same area, Murphy (1981) reported that 88% of patients had married at 15 years of age or less, and 33% had delivered their first child before the age of 15 years. However, in different developing world societies, these factors do seem to have variable influence. For example, in south-east Nigeria (Hilton and Ward 1998) and the north-west frontier of Pakistan (World Health Organization 1989), fistula patients seem to be somewhat older and of higher parity; they also appear to have a higher literacy rate, and to be more likely to remain in a married relationship after the development of their fistula (Hilton and Ward 1998). It is likely that the development of fistulae here reflects other biosocial variations (Table 57.2). It is clear that in these populations, even where skilled maternity care is available, uptake may be poor. Mistrust of hospitals is commonplace, antenatal care is poorly attended, and delivery is commonly conducted at home by elderly relatives or unskilled traditional birth attendants. Where labour is prolonged, transfer to hospital may only be used as a last resort.

Surgical

Genital fistula may occur following a wide range of surgical procedures within the pelvis (Table 57.1, Figure 57.5A–F). It is often supposed that this complication results from direct injury to the lower urinary tract at the time of operation. Certainly, on occasions, this may be the case; careless, hurried or rough surgical technique makes injury to the lower urinary tract much more likely. However, of 300 fistulae referred to the author in the UK over the last 20 years, 207 have been associated with pelvic surgery and 147 followed hysterectomy; of these, only seven presented with leakage of urine on the first postoperative day. In other cases, compromise to the blood supply may result in tissue necrosis and subsequent leakage; alternatively, a small pelvic haematoma may develop in association with the vaginal vault, which subsequently becomes infected and discharges with the characteristic puff of haematuria 5–10 days later, with incontinence following shortly thereafter. Recent animal studies suggest that the inadvertent placement of vault sutures into the bladder wall at hysterectomy may not carry as great a risk of fistula formation as previously thought (Meeks et al 1997).

Figure 57.5 Urogenital fistulae of varying aetiologies: (A) following anterior colporrhaphy, (B) urethral diverticulectomy, (C) radical hysterectomy, (D) midurethral tape procedure for stress incontinence, (E) colposuspension (the fistula in the latter case was fixed retropubically in the midline at the bladder neck, and the photograph shows the patient prone, in the reverse lithotomy position), (F) subtrigonal phenol injection, (G) radiotherapy for carcinoma of cervix.

Although it is important to remember that the majority of surgical fistulae follow apparently straightforward hysterectomy in skilled hands, several risk factors may be identified that make direct injury more likely (Table 57.3). Obviously, anatomical distortion within the pelvis by ovarian tumour or fibroid will increase surgical difficulty, and abnormal adhesions between bladder and uterus or cervix following previous surgery or associated with previous sepsis, endometriosis or malignancy may make fistula formation more likely. Preoperative or early postoperative radiotherapy may decrease vascularity, and make the tissues generally less forgiving of poor technique.

Table 57.3 Risk factors for postoperative fistulae

Sources: Hilton P, Ward A 1998 Epidemiological and surgical aspects of urogenital fistulae: a review of 25 years experience in south-east Nigeria. International Urogynecology Journal and Pelvic Floor Dysfunction 9: 189–194.

Kelly J, Kwast B 1993 Epidemiologic study of vesico-vaginal fistula in Ethiopia. International Urogynecology Journal 4: 278–281.

Murphy M 1981 Social consequences of vesico-vaginal fistula in Northern Nigeria. Journal of Biosocial Sciences 13: 139–150.

Tahzib F 1983 Epidemiological determinants of vesicovaginal fistulas. British Journal of Obstetrics and Gynaecology 90: 387–391.

World Health Organization 1989 The Prevention and Treatment of Obstetric Fistulae: a Report of a Technical Working Group. WHO, Geneva.

Issues of training and surgical technique are, of course, also important. The ability to locate and, if necessary, dissect out the ureter must be part of routine gynaecological training, as should the first aid management of lower urinary tract injury when it arises. The use of gauze swabs to separate the bladder from the cervix at caesarean section or hysterectomy should be discouraged; sharp dissection with knife or scissors does less harm, especially where the tissues are abnormally adherent.

It has recently been shown that there is a high incidence of abnormalities of lower urinary tract function in fistula patients (Hilton 1998); whether these abnormalities antedate the surgery, or develop with or as a consequence of the fistula, cannot be answered from this data. However, it is likely that patients with a habit of infrequent voiding, or with inefficient detrusor contractility, may be at increased risk of postoperative urinary retention; if this is not recognized early and managed appropriately, the risk of fistula formation may be increased.

Radiation

As noted above, preoperative pelvic irradiation increases the risk of postoperative fistula development, but irradiation itself may be a cause of fistula (Figure 57.5G). The obliterative endarteritis associated with ionizing radiation in therapeutic dosage proceeds over many years, and may be aetiological in fistula formation long after the primary malignancy has been treated. Of the 28 radiation fistulae in the author’s series, the fistulae developed at intervals between 1 and 30 years following radiotherapy. Not only does this ischaemia produce the fistulae, it also causes significant damage in the adjacent tissues, so ordinary surgical repair has a high likelihood of failure and modified surgical techniques are required.

Malignancy

The treatment of pelvic malignancy by either surgery or radiotherapy is associated with a risk of fistula development, but tissue loss associated with malignant disease itself may result in genital tract fistula. Carcinoma of cervix, vagina and rectum are the most common malignancies to present in this way. It is relatively unusual for urothelial tumours to present with fistula formation, other than following surgery or radiotherapy. The development of a fistula may be a distressing part of the terminal phase of malignant disease; it is, nevertheless, one deserving not simply compassion but full consideration of the therapeutic or palliative possibilities.

Inflammatory bowel disease

Inflammatory bowel disease is the most significant cause of intestinogenital fistulae in the UK, although these rarely present directly to the gynaecologist. Crohn’s disease is by far the most important, and appears to be increasing in frequency in the Western world. A total fistula rate approaching 40% has been reported, and the involvement of the genital tract in females may be up to 7%.

Ulcerative colitis has a small incidence of low rectovaginal fistulae. Diverticular disease can produce colovaginal fistulae and, rarely, colouterine fistulae, with surprisingly few symptoms attributable to the intestinal pathology. The possibility should not be overlooked if an elderly woman complains of feculent discharge or becomes incontinent without concomitant urinary problems (Figure 57.6).

Figure 57.6 Barium enema from patient complaining of vaginal discharge; demonstrates extensive diverticular disease and obvious colovaginal fistula.

Miscellaneous

Among other miscellaneous causes of fistulae in the genital tract, the following should be considered.

• Infection

Lymphogranuloma venereum

Schistosomiasis

Tuberculosis

Actinomycosis

Measles

Noma vaginae

• Other

Penetrating trauma

Coital injury

Neglected pessary

Other foreign body

Catheter-related injury

United Kingdom

The prevalence of genital fistulae obviously varies from country to country and continent to continent as the main causative factors vary. Accurate figures are impossible to obtain since those areas with the highest overall prevalence are also those with the poorest systems of health data collection. UK National Health Service data reported approximately 150 operations for vesicovaginal and urethrovaginal fistula per year in England and Wales in the early 1990s (Hilton 1997). The most recent data from Health Episode Statistics for England suggest an average of 95 operations for vesicovaginal fistula and 10 operations for urethrovaginal fistula per year in England (Department of Health 2009). It is tempting to suggest that this apparent reduction in incidence of fistulae may reflect changes in the practice of hysterectomy. However, assuming that 50% of urogenital fistulae in the UK follow hysterectomy (as in the author’s series), the rate of fistula formation following hysterectomy seems, if anything, to be increasing; from approximately one in 1200 hysterectomies in the period 1990–1991 to 1994–1995 to approximately one in 750 in 2003–2004 to 2007–2008 (Department of Health 2009). Whilst there are other possible explanations, it may be that as fewer hysterectomies are undertaken, those that remain are the more difficult procedures. Studies from Finland suggest a similar rate of posthysterectomy fistulae overall, with approximately one per 1000 abdominal hysterectomies and one per 450 laparoscopic hysterectomies (Harkki-Siren et al 1998). Similarly, ureteric injury may be up to six times as common following laparoscopic hysterectomy compared with open hysterectomy (Harkki-Siren et al 1998), and two to 10 times as common following radical hysterectomy and exenteration (Averette et al 1993, Bladou et al 1995, Emmert and Köhler 1996). Sultan et al (1994) reported that third-degree tears followed 0.6% of vaginal deliveries, and a rectovaginal fistula resulted in 6% of these (one in 3000 deliveries overall).

Developing world

In the developing world, many fistula cases are unknown to medical services, being separated from their husbands and ostracized from society. Although the true prevalence in the developing world is unknown, particularly high prevalence rates are reported in Nigeria, Ethiopia, Sudan and Chad. The estimated prevalence in the developing world is one to two per 1000 deliveries, with perhaps 50,000–100,000 new cases each year. Although several units exist in Nigeria, Ethiopia and Sudan, which deal with 100–700 cases per year, this does not come close to meeting the demand, and there are estimated to be perhaps 500,000 to 2 million untreated cases worldwide (Waaldijk and Armiya’u 1993).

Microbiology

Urinary infection is surprisingly uncommon in fistula patients, although urine culture should be undertaken and appropriate antibiotic therapy instituted, especially where there have been previous attempts at surgery. Urine culture is not easily obtained when a fistula produces severe incontinence, but a pipette may be used to obtain a small intravesical sample for investigation.

Dye studies

When the diagnosis is in doubt, it is important to confirm that the discharge is urinary and that the leakage is extraurethral rather than urethral, and to establish the site of leakage. Although other imaging techniques undoubtedly have a role (see below), carefully conducted dye studies remain the investigation of first choice.

Excessive vaginal discharge, or the drainage of serum from a pelvic haematoma following surgery, may simulate a urinary fistula. If the fluid is in sufficient quantity to be collected, biochemical analysis of its urea content in comparison to that of urine and serum will confirm its origin. Phenazopyridine is no longer available in the UK, although it has previously been used orally (200 mg tds) to stain the urine and hence confirm the presence of a fistula; alternatively, indigo carmine may be used intravenously for the same purpose. The identification of the site of a fistula is best carried out by the instillation of coloured dye (methylene blue or indigo carmine) into the bladder via catheter with the patient in the lithotomy position. The traditional ‘three swab test’ has its limitations and is not recommended; the examination is best carried out with direct inspection, and multiple fistulae may be located in this way. It is important to be alert for leakage around the catheter, which may spill back into the vagina creating the impression of a fistula. It is also important to ensure that adequate distension of the bladder occurs, as some fistulae do not leak at small volumes; conversely, some fistulae with an oblique track through the bladder wall may leak at small volumes but not at capacity. If leakage of clear fluid continues after dye instillation, a ureteric fistula is likely; previously, this could be confirmed by a ‘two dye test’, using phenazopyridine to stain the renal urine, and methylene blue to stain the bladder contents (Raghavaiah 1974). However, intravesical followed by intravenous injection of dye may be required nowadays.

Dye tests are less useful for intestinal fistulae, although a carmine marker taken orally may confirm their presence. Rectal distension with air via a sigmoidoscope may be of more value; if the patient is kept in a slight head-down position and the vagina filled with saline, the bubbling of any air leaked through a low fistula may be detected.

Imaging

Examination under anaesthesia

Careful examination, if necessary under anaesthetic, may be required to determine the presence of a fistula, and is deemed by several authorities to be an essential preliminary to definitive surgical treatment (Chassar Moir 1967, Lawson 1978, Jonas and Petri 1984, Lawson and Hudson 1987). A malleable silver probe is invaluable for exploration of the vaginal walls, and tissue forceps or plastic surgical skin hooks are helpful to put tension on the tissues for the identification of small fistulae. If the probe passes directly into the rectum, it may be felt digitally or seen via the proctoscope. In the bladder or urethra, it may be identified by a metallic click against a silver catheter, or seen by a cystoscope; in either case, the diagnosis is then obvious.

It is also important at the time of examination to assess the available access for repair vaginally, and the mobility of the tissues (Figure 57.7A,B). The decision between the vaginal and abdominal approaches to surgery is thus made; when the vaginal route is chosen, it may be appropriate to select between the more conventional supine lithotomy with a head-down tilt, and the prone (reverse lithotomy) with head-up tilt (Figure 57.9). This may be particularly useful in allowing the operator to look down on to bladder neck and subsymphysial fistulae, and is also of advantage in some massive fistulae in encouraging the reduction of the prolapsed bladder mucosa (Lawson 1967).

Figure 57.9 Patient in prone (reverse lithotomy) position with head-up tilt in preparation for repair of a fixed subsymphysial fistula.

Endoscopy

Cystoscopy

Although some authorities suggest that endoscopy has little role in the evaluation of fistulae, it is the author’s practice to perform cystourethroscopy in all but the largest defects (Figure 57.10). Although in some obstetric and radiation fistulae, the size of the defect and the extent of tissue loss and scarring may make it difficult to distend the bladder, much useful information is still obtained. The exact level and position of the fistula should be determined, and its relationship to the ureteric orifices and bladder neck are particularly important. With urethral and bladder neck fistulae, the failure to pass a cystoscope or sound may indicate that there has been circumferential loss of the proximal urethra; a circumstance which is of considerable importance in determining the appropriate surgical technique and the likelihood of subsequent urethral incompetence. Similar considerations may apply to investigation of the lower bowel following major obstetric injuries in which segmental circumferential loss of the upper rectum may have occurred.

Figure 57.10 Endoscopic appearance of (A) vesicovaginal fistula following total abdominal hysterectomy, (B) urethrovaginal fistula following midurethral tape procedure.

The condition of the tissues must be assessed carefully. Persistence of slough means that surgery should be deferred, and this is particularly important in obstetric and postradiation cases. Biopsy from the edge of a fistula should be taken in radiation fistulae, if persistent or recurrent malignancy is suspected. Malignant change has been reported in a longstanding benign fistula, so where there is any doubt at all about the nature of the tissues, biopsy should be undertaken (Hudson 1968). In areas of endemicity, evidence of schistosomiasis, tuberculosis and lymphogranuloma may become apparent in biopsy material, and again it is important that specific antimicrobial treatment is instituted prior to definitive surgery. If calculi are identified clinically, radiologically or endoscopically in the bladder, vaginal vault or diverticulum, their removal prior to any attempt at surgical correction is essential.

Immediate management

Before epithelialization is complete, an abnormal communication between viscera will tend to close spontaneously, provided that the natural outflow is unobstructed. Normal continence mechanisms, however, involve the intermittent physiological contraction of urethral and anal sphincters (see Chapter 50). As a result, although spontaneous closure of genital tract fistulae does occur, it is the exception rather than the rule. Bypassing the sphincter mechanisms, or diverting flow around the fistula (e.g. by urinary catheterization or defunctioning colostomy), may encourage closure.

The early management is of critical importance, and depends on the aetiology and site of the lesion. If surgical trauma is recognized within the first 24 h post operatively, immediate repair may be appropriate, provided that extravasation of urine into the tissues has not been great. The majority of surgical fistulae are, however, recognized between 5 and 21 days post operatively, and these should be treated with continuous bladder drainage. It is worth persisting with this line of management in vesicovaginal or urethrovaginal fistulae for 6–8 weeks since spontaneous closure may occur within this period (Davits and Miranda 1991). Although only 8% of the fistulae in the author’s series are known to have healed with catheter drainage, it is possible that the actual rate of spontaneous healing in surgical fistulae is as high as 20%.

Obstetric fistulae developing after obstructed labour should also be treated by continuous bladder drainage (Waaldijk 1994a), combined with antibiotics to limit tissue damage from infection. Indeed, if a patient is known to have been in obstructed labour for any significant length of time, or is recognized to have areas of slough on the vaginal walls in the puerperium, or has any indication of urine leakage, prophylactic catheterization should be undertaken; with this approach, spontaneous healing has been reported in up to 28% of cases (Waaldijk 1997). If sloughing of the rectal wall has also occurred, faecal discharge will adversely affect spontaneous healing, and temporary defunctioning colostomy should be performed.

The management of ureterovaginal fistulae is beyond the skill of most gynaecologists, and even those with specialist skills in urogynaecology are likely to liaise with urological and radiological colleagues in this situation. Whilst in the past, ureteric reimplantation has been looked on as the preferred approach in most cases, the use of stents inserted either in retrograde (endoscopically) or antegrade (percutaneously) fashion is successful in the majority of cases unless there is complete obstruction (Chang et al 1987, Schwartz and Stoller 2000). Where stenting cannot be achieved, or does not result in closure of the fistula, definitive surgery should be undertaken at 4–6 weeks, unless progressive calyceal dilatation or impairment of function demands earlier intervention.

The immediate management of obstetric third-degree tears has traditionally been thought to lead to good outcome. Recent data suggest that approximately half of patients sustaining such injuries will have subsequent defaecatory symptoms as a result of sphincter disruption rather than denervation, and 6% may end up with rectovaginal fistulae (Sultan et al 1994). It has not been investigated whether primary repair by a more experienced surgeon or delayed repair would improve outcome.

It is important to appreciate that some fistulae may be associated with minimal symptoms, and even if persistent, these may not require surgical treatment. Small distal urethrovaginal fistulae, uterovesical fistulae with menouria and some low rectovaginal fistulae may fall into this category.

Palliation and skin care

During the waiting period between diagnosis and repair, incontinence pads should be provided in generous quantities so that patients can continue to function socially to some extent. Fistula patients usually leak much greater quantities of urine than those with urethral incontinence from whatever cause, and this needs to be recognized in terms of provision of supplies.

The vulval skin may be at considerable risk from ammoniacal dermatitis, and liberal use of silicone barrier cream should be encouraged. Steroid therapy has been advocated in the past as a means of reducing tissue oedema and fibrosis, although these benefits are refuted and there may be a risk of compromise to subsequent healing (Jonas and Petri 1984). Local oestrogen has been recommended by some authors (Kelly 1983, Jonas and Petri 1984), and whilst empirically, one might expect benefit in postmenopausal women or those obstetric fistula patients with prolonged amenorrhoea, the evidence for this is lacking.

Nutrition

Due to social ostracism and the effects of prolonged sepsis, patients with obstetric fistulae may also suffer from malnutrition and anaemia. To maximize the prospects for postoperative healing, it is essential that the general health of the patient should be optimized. Where there is severe inflammatory bowel disease, the question of an elemental diet or even total parenteral nutrition may need to be considered in consultation with the gastroenterologist involved.

Physiotherapy

Obstetric fistulae are commonly associated with lower limb weakness, foot drop and limb contracture. In a group of 479 patients studied prospectively, 27% had signs of peroneal nerve weakness at presentation, and a further 38%, whilst having no current signs, gave a history of relevant symptoms (Waaldijk and Elkins 1994). Early involvement of the physiotherapist in preoperative management and rehabilitation of such patients is essential.

Antimicrobial therapy

In tropical countries, the treatment of malaria, typhoid, tuberculosis and parasitic infections should be rigorously undertaken before elective surgery. There is no evidence of benefit from prophylactic antibiotics in the management of obstetric fistulae (Tomlinson and Thornton 1998). Opinions differ on the desirability of prophylactic antibiotic cover for surgery in the developed world, with some avoiding their use other than in the treatment of specific infection, and some advocating broad-spectrum treatment in all cases. The author’s current practice is for single-dose prophylaxis with co-amoxiclav in urinary fistulae, and 5 days cover for intestinal fistulae using metronidazole and cefuroxime. Only symptomatic urinary tract infections need to be treated in the catheterized patient.

Bowel preparation

Although surgeons vary in the extent to which they prepare the bowel prior to rectovaginal fistula repair, it is the author’s preference to carry out formal preparation in all cases of intestinogenital fistula, whatever the level of the lesion. A low-residue diet should be advised for 1 week prior to admission, followed by a fluid-only diet for 48 h before surgery. Polyethylene glycol 3350 (Kleneprep), four sachets in 4 l of water over a 4-h period, or sodium picosulphate (Picolax) 10 mg repeated after 6 h is given orally on the day before operation. Bowel wash-out should be carried out on the evening before surgery, and if bowel content is not completely clear, this should be repeated on the morning of surgery.

Counselling

Surgical fistula patients are usually previously healthy individuals who entered hospital for what was expected to be a routine procedure, and they end up with symptoms infinitely worse than their initial complaint. Obstetric fistula patients in the developing world are social outcasts. In both situations, therefore, these women are invariably devastated by their situation; significant impact on their mental health has been objectively confirmed (Browning et al 2007). It is vital that they understand the nature of the problem, why it has arisen, and the plan for management at all stages. Confident but realistic counselling by the surgeon is essential, and the involvement of nursing staff or counsellors with experience of fistula patients is also highly desirable. The support given by previously treated sufferers can also be of immense value in maintaining patient morale, especially where a delay prior to definitive treatment is required.

General principles of surgical treatment

Details of individual operations are outside the scope of this chapter, and readers wishing for further information about this aspect of fistula management are referred to operative surgical texts or more specific texts on the subject (Chassar Moir 1967, Lawson and Hudson 1987, Zacharin 1988, Mundy 1993, Waaldijk 1994b).

Instruments

All operators have their own favoured instruments, although those described in the treatise by Chassar Moir (1967) and Lawson (1967) are eminently suitable for repair by any route (Figure 57.11).

Figure 57.11 Fistula repair instruments.

Dissection

Great care must be taken over the initial dissection of the fistula, and one should probably take as long over this as over the repair itself. Preliminary infiltration with 1 : 200,000 solution of adrenaline may help to separate planes and reduce oozing. The fistula should be circumcised in the most convenient orientation, depending on size and access. All things being equal, a longitudinal incision should be made around urethral or midvaginal fistulae, so that during repair, sutures tend to close the bladder neck; conversely, vault fistulae are better handled by a transverse elliptical incision, so that sutures do not tend to approximate the ureters during repair (Figure 57.12A).

Figure 57.12 Series of operative photographs demonstrating steps of repair of typical posthysterectomy vault fistula. (A) Fistula circumscribed using no. 12 scalpel, (B) sharp dissection around fistula edge, (C) fistula fully mobilized, (D) vaginal scar edge trimmed, (E) first two sutures of first layer of repair in place lateral to the angles of the repair, (F) second layer completed, sutures catching back of vaginal flaps to close off dead space, (G) testing the repair with methylene blue dye instillation, (H) final layer of mattress sutures in the vaginal wall.

The tissue planes are often obliterated by scarring, and dissection close to a fistula should therefore be undertaken with a scalpel or scissors (Figure 57.12B). Sharp dissection is easier with counter traction applied by skin hooks, tissue forceps or retraction sutures. Blunt dissection with small pledgets may be helpful once the planes are established, and provided one is away from the fistula edge. Wide mobilization should be performed, so that tension on the repair is minimized (Figure 57.12C).

There are arguments over the benefits of excision of the fistula track itself. Excision of the bladder walls is probably unwise, as it enlarges the defect and may increase the amount of bleeding into the bladder. However, limited excision of the scarred vaginal wall is usually appropriate (Figure 57.12D).

Bleeding is rarely troublesome with vaginal procedures, except occasionally with proximal urethrovaginal fistulae. Diathermy is best avoided to prevent further devascularization of tissues, and pressure or under-running sutures are preferred.

Specific repair techniques

Postoperative management

Results

It is difficult to compare the results of treatment in different series since the lesions involved and the techniques of repair vary so greatly. Cure rates should be considered in terms of closure at first operation, and vary from 60% to 98% (Turner-Warwick et al 1967, Hamlin and Nicholson 1969, Chassar Moir 1973, Hudson et al 1975, Goodwin and Scardino 1979, Patil et al 1980, O’Conor 1980, Wein et al 1980, Elkins et al 1988, Lee et al 1988, World Health Organization 1989, Hilton 1995, Hilton and Ward 1998). In ideal circumstances, one might anticipate an 80% continence rate, 10% failures and, in the case of obstetric fistulae at least, 10% suffering from postfistula stress incontinence.

Of the 300 patients in the author’s series managed in the UK, 25 (8.3%) healed without operation, 10 (3.3%) declined surgery, 12 (4.0%) were telephone referrals only and either healed spontaneously or were treated at their local hospital, five (1.7%) underwent primary urinary diversion, and four died from malignancy or the effects of radiation prior to treatment. Of the remaining 244 who have undergone repair surgery, 232 (95.1%) were closed at their first operation, although of these, 12 (5.2%) have some residual incontinence. More specifically, the rate of closure at first operation in the author’s series is 97% in surgical cases, 95% in miscellaneous cases, 89% in obstetric cases and 88% in radiotherapy cases. Of the 130 fistulae following simple hysterectomy undergoing surgical treatment in Newcastle, 128 (98.5%) were cured at their first operation (Table 57.5).

Long-term outcome has rarely been considered. Dolan et al (2008) reported a median 4 year follow-up of women undergoing fistula repair in the UK. Despite their fistulae being anatomically closed, most women reported one or more urinary symptoms in the long term, although only one in eight found these symptoms significantly problematic. Frequency, nocturia, and stress or urge incontinence were the most common troublesome symptoms, and there was no obvious trend to improvement over time.

Of the largely obstetric fistulae reviewed from Nigeria by Hilton and Ward (1998), 81.2% were cured by the first operation and 97.7% were eventually successfully anatomically repaired, with only 0.6% undergoing urinary diversion. It is commonly assumed that follow-up of obstetric fistula patients after discharge from hospital is not feasible; however, two recent studies have found this not to be the case (Browning and Menber 2008, Nielsen et al 2009). Both studies found fistula closure to be associated with improvement in quality of life and reintegration into society. In the longer-term follow-up study of Browning and Menber (2008), although incontinence persisted in some patients, there was a tendency to further improvement over time.

A law of diminishing returns has been reported previously (Lawson 1978). Although repeat operations are certainly justified, the success rate decreases progressively with increasing numbers of previous unsuccessful procedures. In the series reported from Nigeria, although the cure rate at first operation was 81.7%, the success rate for those patients requiring two or more procedures fell to 65% (Hilton and Ward 1998). It cannot be overemphasized that the best prospect for cure is at the first operation, and there is no place for the well-intentioned occasional fistula surgeon, be they gynaecologist or urologist.

Postfistula stress incontinence

Stress incontinence has long been recognized as a complication of vesicovaginal fistulae (Lawson 1967). It is most likely to occur in obstetric fistula patients when the injury involves the sphincter mechanism, particularly if there is tissue loss (Waaldijk 1989, Browning 2006b), although it has also been reported in a large proportion of surgical fistulae involving the urethra or bladder neck (Hilton 1998). It affects at least 10% of all fistula patients, and its impact in terms of the patient’s quality of life and mental health may be as great as that of the fistula itself (Browning et al 2007).

The extent of scarring in the area means that conventional approaches to bladder neck elevation may be technically difficult and of limited success. The use of a labial musculofat graft in the initial repair may reduce the likelihood of the complication (Waaldijk 1994b), and a number of other techniques have been attempted (Hudson et al 1975, Waaldijk 1994b). The implantation of an artificial urinary sphincter or neourethral reconstruction may be appropriate from a theoretical point of view (Hilton 1990), but the former would be prohibitively expensive, and both would be excessively morbid, for use in the developing world. Periurethral injections hold promise as a minimally invasive technique particularly appropriate in the situation of urethral insufficiency with a relatively fixed immobile urethra. Although initial success has been reported in postfistula stress incontinent patients (Hilton et al 1998), the long-term outcomes are disappointing.

Subsequent pregnancy

Although many patients with obstetric vesicovaginal fistula will experience amenorrhoea for 2 years or more after the causative pregnancy (Hilton and Ward 1998), where tissue loss is not great, hypothalamic function often returns to normal immediately after successful repair, and fertility may be relatively normal. Kelly (1979) reported on 33 patients who became pregnant within 1 year of fistula repair, 12 of whom were delivered vaginally without damage to the repair. His criteria for attempting vaginal delivery were: that the fistula arose from a non-recurring cause (i.e. malpresentation as opposed to pelvic contraction), that an interposition graft had been utilized in the repair, and that the labour was conducted under skilled supervision in hospital (Kelly 1979). However, most other authorities have emphasized the need for caesarean section in any subsequent pregnancy (Lawson 1978). Most recently, in a study from Ethiopia, Browning (2009) described no recurrent fistulae, and 47 livebirths from 49 pregnancies delivered by a policy of elective caesarean section (in the rare cases where dates were certain) or emergency caesarean section at premature rupture of membranes or onset of labour.