Diagnose and manage genitourinary and obstetric fistula

In one line

A genitourinary fistula is an abnormal epithelialised communication between the urinary tract and the genital tract that produces continuous, position-independent urinary leakage; in South Africa the dominant cause is still obstetric — ischaemic pressure necrosis from prolonged obstructed labour — and the single most important principle is that the first repair is the best chance of cure, so a fistula belongs with an experienced fistula surgeon who has staged it properly, not with whoever is on call.

The mechanism, the pathology and the simple bedside diagnosis are the groundwork; revise them at gynaecological fistulas. The consultant task here is to classify the defect accurately, decide when and by what route to operate, anticipate residual incontinence after an anatomically successful closure, and own the prevention chain that should make the disease disappear.

Mechanism & pathophysiology

Three aetiological mechanisms produce a genitourinary fistula, and they behave so differently that lumping them together is the commonest conceptual error.

Obstetric (ischaemic) fistula is a pressure injury. During prolonged obstructed labour the fetal head is impacted against the maternal pelvis, and the soft tissues of the bladder base, urethra, anterior vaginal wall and — posteriorly — the rectovaginal septum are crushed between the head and the symphysis or sacrum. Sustained compression for hours to days produces ischaemia, then pressure necrosis; the devitalised tissue sloughs several days postpartum, leaving a defect. This is why the classic obstetric fistula is not a clean surgical hole but a zone of avascular, fibrotic, scarred tissue, often with substantial loss of bladder and urethral substance, frequently bordering or involving the urethral closure mechanism and sometimes circumferential — the urethra effectively amputated from the bladder. The same ischaemic field explains the cluster of associated injuries that define the obstructed labour injury complex: vesicovaginal and/or rectovaginal fistula, urethral loss, vaginal stenosis, secondary infertility, foot-drop from lumbosacral plexus compression, and amenorrhoea. The fetus is almost always stillborn. The woman is characteristically young, short-statured, primiparous and poor, delivered late and far from emergency care — the fistula is the anatomical fossil of the three delays (deciding to seek care, reaching care, receiving care).

A crucial corollary, often missed: caesarean section does not abolish the risk if the labour was already long. By the time an obstructed labour reaches theatre the ischaemic insult to the bladder base may already be complete, so a fistula can still appear after a caesarean delivery — district data show fistulae occurring despite caesarean access precisely because the compression preceded the operation (Loposso 2015). The prevention target is therefore timely caesarean, not caesarean per se.

Iatrogenic (surgical) fistula is a different lesion. Here the bladder or ureter is injured directly — most often at hysterectomy (the classic post-hysterectomy vesicovaginal fistula presenting with leakage 1–2 weeks postoperatively as a devascularised suture line breaks down or an unrecognised cystotomy declares itself), at caesarean section, or at caesarean hysterectomy for placenta accreta. The tissue is otherwise healthy, the defect is usually small, high in the vaginal vault, and clean-edged, and the prognosis for closure is correspondingly excellent. Ureteric injury (ureterovaginal fistula) is the iatrogenic lesion that masquerades as a bladder fistula and must be actively excluded, because it leaks urine vaginally while the bladder fills and voids normally — a fundamentally different operation.

Malignant and radiation fistula is the third mechanism and the one that changes the entire plan. A fistula arising in a field of advanced or recurrent cervical, vaginal or bladder cancer, or in a pelvis that has had radiotherapy, is not a primary repair problem: the tissue is tumour-bearing or radiation-damaged and ischaemic, simple closure fails, and the priorities become tissue diagnosis (biopsy the edges — never assume), oncological staging, and usually urinary diversion or interposition rather than layered repair. Treating a malignant fistula as if it were an obstetric one wastes the patient's time and the surgeon's credibility.

Parity shapes the obstetric lesion in a way worth holding for the operating plan: primiparous women tend to develop distal fistulae with more scarring (the head sits low and long), while multiparous women tend toward proximal defects (Lorencz 2016). Distal and scarred is the harder, more incontinence-prone repair.

Hold one more pathophysiological distinction because it dictates whether continence is even achievable. The continence mechanism is the urethra plus the bladder neck and their support; an obstetric fistula that destroys this — a circumferential defect where the urethra is detached from the bladder, or one that takes out most of the urethral length — can be anatomically closed (the urinary tract made watertight) while leaving no functioning sphincter, so the woman is dry of the fistula but wet from sphincteric incontinence. By contrast, a high vesicovaginal or vesicocervical fistula that spares the urethra and bladder neck, once closed, leaves a continent woman. This is the physiological reason the site of the defect, not its size, is the strongest outcome lever, and why every classification that matters anchors on distance-from-the-meatus or urethral involvement.

Assessment

The history is usually diagnostic. Continuous, uncontrollable urinary leakage that wets the patient day and night and is independent of activity or position points to a fistula; leakage only on effort points to stress incontinence and a leak that comes and goes with urgency points to an overactive bladder — neither produces the constant wetness of a fistula. Date the leak against the index event: onset immediately or within days of an obstructed labour is obstetric; onset 1–2 weeks after a hysterectomy is iatrogenic. Take a full obstetric history (duration of labour, place and mode of delivery, fetal outcome), and screen deliberately for the rest of the injury complex — faecal leakage (rectovaginal fistula), difficulty walking (foot-drop), amenorrhoea, and the social devastation that is part of the diagnosis.

Examination is done in good light with the patient in the lithotomy or knee–chest position, and often it should be examination under anaesthesia when scarring, pain or a high defect prevents an adequate office assessment. Define:

• The site (vesicovaginal, urethrovaginal, vesicocervical, ureterovaginal, rectovaginal — and note that multiple defects coexist), the size, and the distance of the distal edge from the external urinary meatus (the basis of the Goh type).
• The state of the surrounding tissue — fibrosis, vaginal length and capacity, and whether the defect is circumferential (the urethra/bladder neck completely detached), because these, more than size, drive both the difficulty of repair and the risk of post-closure incontinence.
• Urethral length and competence, the single most prognostically loaded feature: urethral involvement independently predicts both failed closure and persistent incontinence after closure (Maljaars 2023).
• Associated injuries — a digital and, where available, endoanal assessment of the anal sphincter, and a neurological check for foot-drop.

The dye test (three-swab test) is the bedside investigation that localises and confirms a urinary fistula and distinguishes a vesical from a ureteric source. Three swabs are placed in the vagina (high, mid, low), the bladder is instilled with dilute methylene blue (or indigo carmine) through a catheter, and the swabs are inspected: blue staining of the upper/mid swabs confirms a vesicovaginal fistula; clear wetting of the top swab with no dye signals urine entering above the bladder — a ureterovaginal fistula — and mandates upper-tract imaging. A "double-dye" variant (oral phenazopyridine colouring renal urine orange while methylene blue fills the bladder) separates the two sources in one test.

Cystoscopy maps the defect's relationship to the trigone and ureteric orifices and is essential before any abdominal or complex repair. Imaging of the upper tracts is non-negotiable when a ureteric fistula is possible — CT urogram (or IVU where CT is unavailable) to exclude or define ureteric involvement, hydronephrosis and the level of any ureteric fistula before operating. Examination under anaesthesia is also when you confirm bladder capacity, plan the route, and decide whether an interposition graft will be needed.

Classification — and its honest limits

Two systems dominate, and the consultant must be able to use and contrast them.

Goh (the more widely used and the better validated) classifies by the distance of the distal fistula edge from the external urinary meatus — Type 1 furthest (most favourable), Type 4 closest to/involving the meatus (least favourable) — and adds a size descriptor (a/b/c) and a special-considerations descriptor (i/ii/iii) capturing fibrosis, vaginal length/capacity, circumferential defect and previous repair (Goh 2004). The further toward the meatus, the larger, the more scarred and the more circumferential, the worse the prognosis — that is the whole clinical message.

Waaldijk classifies by whether the urethral closure mechanism (continence mechanism) is involved — Type I not involving it, Type II involving it (subdivided IIA/IIB by degree of urethral involvement), Type III the miscellaneous/ureteric group — which keeps the surgeon's eye on the feature that most determines post-repair continence.

The honest, examinable point is that no classification predicts outcome well. When the major systems (WHO, Goh, Waaldijk, Tafesse) were tested head-to-head against actual closure in over 1,200 women, the best achieved an AUC of only ~0.62 — "poor to fair" (Frajzyngier 2013). Goh outperforms Waaldijk for predicting closure (Capes 2012), and the Panzi score distilled the three genuinely predictive descriptors — circumferential defect, proximity to the meatus, and size (Mukwege 2018) — but the take-home is that classification standardises description and audit, it does not replace the surgeon's assessment of tissue and urethra.

Management

Organise the plan immediate → ongoing → long-term, and let the first decision be the one that most often decides cure: who operates, when, and by what route.

Immediate. A fresh, small, recently formed fistula (typically iatrogenic, recognised within days, or a small post-obstructed-labour defect with minimal slough) may be given a trial of continuous bladder drainage for several weeks — a small proportion close spontaneously on catheter alone if urine is diverted from the defect while the tissue settles. For everything else, the discipline is delay and prepare, not operate hot. A frankly obstetric fistula is left to declare its final size, allowed to demarcate and for the surrounding inflammation and slough to resolve, and repaired electively at roughly 3 months — earlier surgery into oedematous, infected, sloughing tissue fails. The waiting interval is used to optimise the patient: treat infection, correct anaemia and malnutrition, manage the excoriating dermatitis of constant leakage (barrier protection), and address the psychological injury and social isolation.

Definitive repair — principles. The non-negotiables of fistula surgery are: adequate exposure, wide mobilisation of the bladder off the vagina so the repair is tension-free, excision of only frankly devitalised tissue (not aggressive debridement, which enlarges the defect), layered closure with the suture lines offset and non-overlapping, watertight closure confirmed by an on-table dye test, and interposition of a vascularised flap wherever the tissue is poor, the defect large, the repair under tension, or the fistula recurrent.

• Route — vaginal first. Most fistulae, including most obstetric ones, are repaired vaginally, which is less morbid and the default for the fistula surgeon. The Latzko technique (a partial colpocleisis — denuding and approximating the vaginal walls around a high, small vesicovaginal fistula, classically post-hysterectomy) is the elegant option for the small vault fistula. A standard layered dissection-and-repair handles most others.
• Abdominal/transvesical route is reserved for the fistula the vagina cannot reach or safely repair: a high, fixed, post-hysterectomy fistula; one intimately related to the ureteric orifices; a vesicouterine fistula; or where concomitant ureteric reimplantation is needed. Transvesical (transperitoneal or extraperitoneal) repair, with omental interposition, is the abdominal workhorse.
• Interposition flaps. The Martius graft — a labial fat pad (bulbocavernosus/labial fibrofatty flap) tunnelled from the labium majus to lie between the bladder and vaginal repair — is the classic vaginal interposition, used for large, scarred, irradiated or recurrent fistulae and where the urethra/bladder neck is reconstructed; it brings a fresh blood supply and a barrier between the offset suture lines. Omentum is the abdominal-route interposition of choice.
• Circumferential and urethral defects are the hard repairs: the bladder neck/urethra is reconstructed over a catheter, often with a Martius graft, and these are precisely the patients who close anatomically yet leak — see below.
• Ureteric fistula is not a vaginal repair: it is ureteric reimplantation (ureteroneocystostomy), with a psoas hitch or Boari flap for a higher injury, usually after a period of stenting/nephrostomy. A fresh, partial ureteric injury recognised early may close over a stent alone; an established ureterovaginal fistula needs reimplantation, and the consultant decision is the timing — drain the obstructed kidney and let inflammation settle, then reimplant onto healthy bladder.
• Urinary diversion is the salvage pathway, not a first option: for the irreparable fistula (multiple failed repairs, total urethral and bladder-neck loss, a tiny contracted bladder, or a malignant/radiation fistula), continent or incontinent diversion, or a Mainz-II (ureterosigmoidostomy) where ileal conduits are impractical, may be the only route to a dry, dignified patient. In SA this is a tertiary, multidisciplinary decision, taken only after the reconstructive options are genuinely exhausted.
• Recto-vaginal fistula, the posterior member of the injury complex, is repaired on its own principles — layered repair with the rectal and vaginal suture lines offset, interposition (Martius) for large or recurrent defects, and a concurrent anal-sphincter repair where the obstructed-labour injury has also disrupted the sphincter; a defunctioning stoma is reserved for large, complex or radiation rectovaginal fistulae.

Ongoing (the post-operative regimen is part of the operation). A patent, never-blocked catheter is the linchpin — a bladder that distends against a fresh repair breaks it down — so free continuous drainage, generous fluids to keep the urine dilute and the catheter flushed, and vigilant nursing are core. Duration of catheterisation after a simple fistula repair can safely be 7 days: the multinational WHO/Fistula Care RCT showed 7-day drainage non-inferior to 14-day, with breakdown of 4% vs 3% (Barone 2015) — shorter catheterisation means earlier discharge and fewer catheter complications without sacrificing closure. Complex repairs are still drained longer (commonly ~14 days) by convention. Antibiotics: there is no good evidence that extended prophylaxis improves closure — an RCT found intra-operative prophylaxis did not reduce failed repair (Tomlinson 1998), and single-dose peri-operative cover is as effective as a prolonged course (Muleta 2010). Avoid early intercourse and confirm the repair held with a dye test before catheter removal.

Long-term. Outcome is reported as two separate endpoints, and conflating them is a classic error: anatomical closure (the hole is shut) and continence (she is dry). A fistula can be closed and the woman still wet. Counsel, follow up, and where the woman remains incontinent after a closed repair, investigate and manage that as stress or urgency incontinence in its own right. Beyond the bladder, long-term care owns the secondary infertility and amenorrhoea (often from intrauterine adhesions or hypothalamic suppression), vaginal stenosis (dilators), and — at least as important as the surgery — social and psychological reintegration: a woman cured of her leak after years of isolation needs active rehabilitation, not discharge.

Prevention — the part of the answer that should make the disease disappear

Obstetric fistula is, almost by definition, a preventable disease, and a complete consultant answer treats prevention as core management rather than a postscript. The chain is the same one that prevents the obstructed-labour deaths sitting next to fistula in the maternal-mortality data: delay first pregnancy beyond adolescence (immature pelvis plus social vulnerability is the classic at-risk profile), skilled birth attendance, partograph-monitored labour so obstruction is recognised before the bladder base is destroyed, and timely caesarean — timely being the operative word, since a caesarean after a day of obstruction cannot undo a completed ischaemic injury (Loposso 2015). At the system level this maps to the three delays, and the practical SA levers are functioning district maternity services, working referral and transport, and a culture that escalates a stalled labour rather than waiting. For an individual woman who has had a fistula repaired, prevention of recurrence in a future pregnancy means elective caesarean delivery of all subsequent pregnancies — a vaginal birth threatens the repair — counselled explicitly at discharge. Closing the loop on the index case (deliver her future pregnancies abdominally, and feed the failure back into the unit's perinatal audit) is part of managing the fistula in front of you.

Outcomes — what closure and continence actually run at

Closure of a first obstetric fistula repair in experienced hands is in the order of 80–95% (WHO), but real LMIC cohorts are sobering and honest about the gap: a recent six-centre Ethiopian series reported 77.5% successful repair — below the WHO benchmark — with fistula length >2 cm and urethral damage independently lowering success (Zeleke 2025). Residual incontinence after anatomical closure is the dominant unsolved problem: even when the hole is shut, a substantial minority leak, and the strongest predictor is urethral/closure-mechanism involvement — repeat-surgery data show closed-but-incontinent rates such that continence was achieved in only ~64% after a first repeat repair, with urethral involvement raising the risk of failure 4–5-fold (Maljaars 2023). The first operation matters disproportionately because each failed attempt scars the field and lowers the odds of the next — which is the entire argument for concentrating fistula surgery in dedicated centres with experienced surgeons rather than dispersing it.

Guidelines compared

There are few hard guideline numbers here and many areas of consensus-by-practice; an honest account states where bodies align and where the evidence is simply thin.

The genuine divergence is less between guideline bodies than between the obstetric-fistula world and the gynaecological/iatrogenic-fistula world: timing (delay ~3 months for obstetric vs early repair for a clean iatrogenic fistula), routine interposition (frequent in obstetric, often unnecessary in small iatrogenic), and the dominant outcome problem (residual incontinence in obstetric vs near-universal cure in iatrogenic) all differ. Reading an SA obstetric fistula by the rules of a European post-hysterectomy fistula service is the conceptual trap.

The evidence & the controversy

The evidence base for fistula surgery is cohort- and consensus-led, with very few randomised trials — and a candidate who pretends otherwise is easily exposed. The handful of RCTs that exist address peri-operative management rather than the operation itself: the catheter-duration trial (Barone 2015) and the antibiotic-prophylaxis trials (Tomlinson 1998) are essentially the lot. The technique questions — vaginal versus abdominal route, Martius graft versus no graft, layered versus single-layer closure — rest on observational series and expert practice, not randomisation, largely because the disease is geographically concentrated in low-resource settings, the presentations are heterogeneous, and randomising a once-only best-chance operation is ethically and practically fraught. The right consultant posture is to say so plainly, defend practice from the best available cohort data and physiological first principles, and flag the absence of RCTs as a real limitation rather than papering over it.

The live controversy is timing. The traditional ~3-month delay rests on the logic of letting ischaemic tissue demarcate, but a body of work from high-volume centres (notably the Hamlin/Addis Ababa tradition and Waaldijk's early-repair experience) argues that early or immediate repair of selected, clean, small fistulae — catheter management for tiny defects, early surgery for others — shortens the woman's ordeal without harming closure. There is no RCT to settle the delay-versus-early question for the general obstetric fistula, so the defensible answer is selective: catheter trial for fresh small defects, early repair only for clean uncomplicated ones in expert hands, and the conventional ~3-month delay for the large, sloughing, scarred majority.

A second, structural debate is where repair should happen. The data that classification predicts outcome only weakly (Frajzyngier 2013), combined with the dominance of surgeon experience and urethral involvement as outcome drivers, is the evidence-based argument for centralising fistula care in dedicated units rather than letting every gynaecologist attempt the occasional repair — each failed first attempt lowers the ceiling for every subsequent one. For South Africa specifically, the uncomfortable current-events thread is that obstetric fistula is a marker of a failing emergency-obstetric pathway: where it still occurs, it indicts delays in reaching and receiving timely caesarean, and the same systemic failures show up in the perinatal and maternal mortality audits. The fistula in front of you is, epidemiologically, a sentinel of a health-system gap — and saying that is part of the answer.

Landmark trials & key evidence

A worked figure to hold: in the catheter RCT the absolute difference in breakdown between 7- and 14-day drainage was 0.8% with a confidence interval comfortably inside the pre-specified 10% non-inferiority margin — which is why shortening catheterisation for simple fistulae costs nothing in closure while saving a week of catheter-days per woman. Contrast that with the continence gap after closure (around a third still wet in repeat-surgery series): the unsolved problem is not whether the hole shuts but whether the urethra works.

Exam traps & red flags

• Conflating closure with continence. "The fistula is closed" is not "she is dry" — report and counsel the two endpoints separately; residual incontinence (driven by urethral involvement) is the dominant unsolved problem.
• Missing the ureteric fistula. Continuous vaginal leakage with a normally voiding bladder and a clear (undyed) top swab is a ureterovaginal fistula — it needs upper-tract imaging and ureteric reimplantation, not a vaginal bladder repair. Operating on it as a VVF fails.
• Not biopsying a fistula in a cancer field. A fistula in advanced/recurrent cervical or vaginal cancer, or an irradiated pelvis, is a malignant/radiation fistula until proven otherwise — biopsy the edges, stage, and think diversion/interposition, not primary layered repair.
• Operating hot, or letting the wrong person operate. Repairing into sloughing, infected, oedematous tissue, or attempting an obstetric fistula occasionally rather than referring to an experienced centre, squanders the best (first) chance of cure.
• Assuming caesarean prevents fistula. If the labour was already long and the bladder base ischaemic, a fistula can follow even a caesarean — the prevention target is timely caesarean and partograph-driven recognition of obstruction.
• Blocked catheter after repair. A distended bladder ruptures a fresh repair — a blocked or kinked catheter in the early post-operative period is a surgical emergency, not a nursing inconvenience.
• Forgetting the rest of the injury complex. Screen every obstetric-fistula patient for rectovaginal fistula, anal-sphincter injury, foot-drop, secondary infertility/amenorrhoea, vaginal stenosis and the psychosocial injury — the bladder is only one component.
• Over-trusting the classification number. Goh/Waaldijk types standardise description and audit but predict outcome only "poorly to fairly"; weight the tissue, the urethra and surgeon experience above the type.
• Over-claiming an evidence base. The field is cohort- and consensus-led with almost no RCTs on technique — claiming randomised proof for Martius grafts or route is wrong; name the limitation honestly.

Evidence anchors

• Barone MA, et al. Breakdown of simple female genital fistula repair after 7 day versus 14 day postoperative bladder catheterisation: a randomised, controlled, open-label, non-inferiority trial. Lancet 2015
• Goh JTW. A new classification for female genital tract fistula. Aust N Z J Obstet Gynaecol 2004
• Capes T, et al. Comparison of two classification systems for vesicovaginal fistula. Int Urogynecol J 2012
• Frajzyngier V, et al. Development and comparison of prognostic scoring systems for surgical closure of genitourinary fistula. Am J Obstet Gynecol 2013
• Maljaars LP, et al. Predictors of outcomes in patients with repeat surgery for obstetric fistula. Int Urogynecol J 2023
• Zeleke LB, et al. Successful surgical repair outcomes and determinants among obstetric fistula patients: a multi-centre study from Ethiopia. Midwifery 2025
• Tomlinson AJ, Thornton JG. A randomised controlled trial of antibiotic prophylaxis for vesico-vaginal fistula repair. BJOG 1998
• Mukwege D, et al. Panzi score as a parsimonious indicator of urogenital fistula severity. Int J Gynaecol Obstet 2018
• de Bernis L. Obstetric fistula: guiding principles for clinical management and programme development, a new WHO guideline. Int J Gynaecol Obstet 2007
• Loposso MN, et al. Obstetric fistula in a district hospital in DR Congo: fistula still occur despite access to caesarean section. Neurourol Urodyn 2015
• Lorencz E, et al. Parity as a predictor of obstetric fistula classification. Int Urogynecol J 2016
• WHO — 10 facts on obstetric fistula
• FIGO Global Competency-Based Fistula Surgery Training Manual (FIGO and partners) — standardised competency-based training for fistula surgeons (standard/advanced/expert tiers).
• South Africa NDoH Maternity Care Guidelines — partograph-monitored labour, recognition of obstruction and timely caesarean as the obstetric-fistula prevention chain; established fistula referred for sub-specialist repair.