Women with obstetric fistula in Ethiopia: a 6-month follow up after surgical treatment


A Browning, Medical Director, Barhirdar Hamlin Fistula Centre, PO Box 1739, Barhirdar, Ethiopia. E-mail andrew_browning@hotmail.com


Objective  To quantify surgical and quality of life outcomes 6 months after obstetric fistula repair.

Design  A prospective study. Women were examined and questioned at discharge and at follow-up appointment.

Setting  The Barhirdar Hamlin Fistula Centre, a dedicated unit treating women with obstetric fistula in northern Ethiopia.

Population  All women admitted to the Barhirdar Hamlin Fistula Centre suffering from vesicovaginal (WF) and rectovaginal fistulae (RVF).

Methods  All women were asked to return for a follow-up appointment 6 months after surgical treatment. A standardised questionnaire and examination were used and information entered into a database.

Main outcomes measures  Urinary and anal continence status; extent of return to previous family situation, employment, social activities and reproductive capabilities.

Results  Continence status at discharge was largely maintained at 6 months and quality of life was improved. Many of those incontinent on discharge improved; a small number apparently cured at discharge had later recurrence of incontinence.

Conclusions  Surgical treatment for obstetric fistula is successful in maintaining continence and improving quality of life of women at 6-months follow up.


The obstetric fistula is a significant cause of maternal morbidity with up to 2 million women currently estimated to be untreated1,2 and 100 000–500 000 new cases occurring each year, predominantly in the developing world.2,3 The impact of fistula is much wider than physical injury, with up to 52% of sufferers being divorced because of their injuries, 92% having a stillbirth during the incident delivery and up to 97% screening positively for potential mental health disorders.4,5

After successful surgical closure of an obstetric fistula, 33–50% of cases may continue to suffer from urinary incontinence per urethram.6,7 It has been assumed that women who are discharged with residual urinary incontinence following fistula repair do improve with time, although there are no published follow-up studies to support this contention.

It has also been assumed that when a woman is cured of her incontinence she will remarry, reintegrate into her community and resume her previous activities; again this is unproven.

Most reports of fistula treatment have not included follow up after discharge from hospital; the centres treating these women are often at great distances from their homes, transport is often difficult and the associated costs have mitigated against return to hospital.8,9 The Barhirdar Hamlin Fistula Centre is a new unit situated in the north of Ethiopia, a rural area much closer to where the people live; hence, postoperative follow up seemed more likely to be achieved.


All women admitted to the Barhirdar Hamlin Fistula centre with vesicovaginal fistula (VVF) and/or rectovaginal fistula (RVF) were operated upon by the senior author or by a surgical trainee under supervision. Fistulae were described using the Goh classification, which denotes the site of urinary or faecal fistulae in relation to the urethral meatus or hymen, respectively, together detail of size and amount of scarring.10 All women were operated upon under spinal anaesthesia in the Trendelenburg position by the vaginal route. All operations employed a flap-splitting technique, mobilising the bladder and/or rectum from the surrounding structures until a tension-free closure was attainable. If the urethra was significantly involved in the defect (Goh type 3 and 4),10 a suburethral sling was created from pubococcygeal muscle.6 No Martius or other interposition grafts were used in the repair of either VVFs or RVFs.11 For all VVFs, a Foley catheter remained on free drainage for 10–14 days. The day after catheter removal, the woman was interviewed and examined for evidence of urinary incontinence by one or other authors. If the woman was still incontinent, a dye test was performed to check the integrity of fistula closure, a residual urine volume was measured and the woman was examined for stress incontinence leaks with coughing. The diagnosis of residual urethral incontinence, fistula breakdown or urinary retention with overflow incontinence was given.

A grading system for residual incontinence was been developed for use in this setting and used over the past 3 years; this reflects the severity of incontinence and the ability of the bladder to retain urine (Table 1). The grade is determined on history and then confirmed by observing the woman as the woman goes about her activities on the ward. While crude, a small local study on 36 women has shown 94% inter-observer and 97% intra-observer reliability.

Table 1.  Incontinence grading system
Incontinence gradeDescription of incontinence
1Cured, no incontinence
2Incontinent with cough, strain or exertion
3Incontinent on walking
4Incontinent on walking, sitting and/or lying but voiding some urine
5Incontinent on walking, sitting and/or lying but not voiding any urine

A patient card was given to all women on discharge, and they were then asked to return to the hospital for follow up at 6 months following surgery. Women were offered reimbursement of transport costs plus an additional 50Birr (approximately GB£3) to cover any extra expenses.

At follow up, a structured interview was undertaken by one or other authors based on a standard data set. The interviews were performed in the local regional language (Amharic) and translated into English, for entry into a Microsoft Excel spreadsheet. This included the initial results of the surgery at the time of discharge, and current urinary and anal continence status; further questions related to social interaction, working ability, marital status, sexual intercourse, menstruation and subsequent pregnancies. A ten point Likert scale was used to assess the overall impact of fistula repair on quality of life; women were asked to grade impact as much worse, worse, slightly worse, a little worse, no different, a little improved, slightly improved, moderately improved, greatly improved and completely cured.


During the time span of the study, 404 women with urinary or RVF were admitted; 6 were considered inoperable do to almost complete bladder loss, thus 398 underwent surgical repair; of these 26 had an RVF of which 18 were combined with a VVF and 8 had an isolated lesion. All the isolated RVFs were in women who were raped at a young age (as young as 5 years old) or who were married at an early age (as young as 8 years old) and sustained trauma from the intercourse; they were excluded from the analysis as the aetiology for their injuries was not ischaemic necrosis from an obstructed labour and thus were not obstetric fistulae. Hence, we included 390 index cases of whom 240 (61.5%) returned for follow up. Eight women had a failed fistulae repair (2.1% of VVF cases) and 95 women in whom the fistula appeared to be closed had remaining urethral incontinence (24.3% of VVF cases). Seven of the 8 failed repairs and 92 of the 95 women discharged with urethral incontinence returned for follow up. Hence, the follow-up group contained a preponderance of those who were not completely cured at the time of discharge from hospital (96.1% of whom returned) than of those apparently cured at discharge (49.1% of whom returned).

However, of the 141 women attending for follow up who were cured at discharge, 22 came back now leaking with an incontinence grade of 2–4 (mode 2). In one of the seven women with apparently unsuccessful fistula repair on discharge, the fistula had healed by the time of the follow-up appointment; the other six persisted. Six women originally discharged with a closed fistula and continent came back with a recurrence or ‘late breakdown’ of the fistula. Most of the breakdowns occurred during physical activity including straining, sexual intercourse, heavy work or being on a very bumpy road in a bus.

Of the 92 women attending for follow up who were discharged with urethral incontinence, 49 (53.3%) returned with the same incontinence grade; 41 (44.6%) returned with their incontinence score improved or with their fistula now healed or retention resolved (Table 2).

Table 2.  Incontinence grade, failed repairs and urinary retention at discharge and follow up
Incontinence grade on dischargeIncontinence grade improved at follow upIncontinence grade unchanged at follow upIncontinence grade worse at follow upTotal
10 12685.1%2214.9%148
2975.0%325.0%0 12
31354.2%1145.8%0 24
5323.1%1076.9%0 13
Failed repair114.3%685.7%0 7
Urinary retention730.4%1565.2%14.3%23
Total41 175 24 240

Twenty-four women returned with their incontinence grade worse or with urinary retention or fistula recurrence; 22 of these were initially discharged completely cured with an incontinence grade 1.

These results were analysed for linear trend, comparing those women whose incontinence grade improved over time to those whose grade remained unchanged or worsened; those with a grade 1 were excluded as they could not improve. There was a significant trend for women with a lower initial incontinence grade to improve over time (= 0.021).

The relationship between the Goh classification of fistulae and the incontinence grade on discharge and follow up is summarised in Table 3. Again there was a significant linear trend for women with lower Goh classification to improve over time, that is those women with less urethral involvement in the fistula (Goh types 1 and 2) who were discharged with some urinary incontinence, were more likely to improve with time (= 0.005).

Table 3.  Urinary incontinence grade at discharge and follow up by Goh classification
Goh classificationIncontinence grade improved at follow upIncontinence grade unchanged at follow upIncontinence grade worse at follow upTotal
Cured (grade = 1)Not cured (grade = 2–5)
Total42 126 48 24 240

A number of further treatment interventions were carried out during the follow-up period. Of those women with persistent urethral incontinence, 23 were admitted for a secondary operation for stress urinary incontinence, 5 used urethral plugs and 9 were encouraged to continue pelvic floor exercises as their incontinence was mild. Three women had a completely closed vagina secondary to scarring and all underwent vaginoplasty. One woman had a bladder stone and six required treatment for urethral stricture; six women with recurrence late of their fistula were admitted for further repair.

Six women were pregnant at the time of their 6-month follow-up visit; all were admitted either immediately or subsequently with a view to elective caesarean delivery.

All 18 of the RVFs closed at the time of discharge and they remained so at follow up. Only two women reported volunteered symptoms of flatal incontinence and one complained of both flatal and loose stool incontinence.

Various biosocial outcomes are given in Table 4. Only 17 women were having sexual intercourse before the operation, that is while suffering incontinence, three of these suffered pain during intercourse. By 6 months after fistula repair, one-third of women had resumed sexual activity, and in 90%, this was without pain. Of those who were not having intercourse following surgery, 80% were divorced, separated or widowed.

Table 4.  Biosocial variables pre-operatively and at follow up
 Pre-operatively (= 240)At follow up (= 240)
Having sexual intercourse177.1%8435.0%
 Dysparuenia (% of those sexually active)317.6%910.7%
Reasons for apareunia (% of those not sexually active)
 Divorced/separated 12777.0%
 Awaiting permission 84.8%
 Afraid damaging the repair 148.5%
 Widowed 42.4%
 Scarred vagina 31.8%
Marital status
Menstrual cycle 
Reason for amenorrhoea (% of those amenorrhoeic)
 Early postdelivery8157.9% 
 Caesarean hysterectomy1410.0%1420.0%
 Pregnant 68.6%
 Depoprovera 45.7%
Living arrangements
 Living alone4016.7%4016.7%
 With husband10041.7%9238.3%
 With family member10041.7%10845.0%
 No 7631.7%
 Yes 16468.3%
Attending social gatherings
Overall satisfaction score
 Satisfied (score 8–10) 19982.9%
 Unsatisfied (score <8) 4117.1%

When asked if the operation had improved their life on Likert scale, 199 (82.9%) reported being at least moderately improved. Only one woman described being worse; the woman had urinary retention and despite being dry with continuing intermittent catheterisation claimed that the woman had been better off being continually incontinent with her VVF than self-catheterising but dry.


Although it has often been assumed that follow up of women with obstetric fistula would be impractical in view of the social, environmental and economic circumstances in which most arise, we have shown that in this particular setting at least, this is not necessarily the case. This is perhaps facilitated by our unit’s rural setting, being closer to patients’ homes and more easily accessible for them. Our practice of reimbursement of transport costs and additional expenses may also have acted as further incentive. Nevertheless, it should be recognised that our follow-up rate was only 62%, and even in this setting, those women with residual symptoms at the time of discharge were twice as likely to attend for follow up as those free from symptoms on discharge.

Urinary incontinence persisting following repair of obstetric fistula is complex and still poorly understood. In most cases, the urethra has been affected and the continence mechanisms damaged. The repair may be anatomically satisfactory but functionally inadequate. Detrusor overactivity and impairment of bladder compliance may play a role in this.12

Despite or perhaps because of this complexity, some symptoms of women do improve with time, although others apparently cured or with only mild incontinence at discharge returned with recurrence or worsening of their symptoms. It could be argued that these represent a lack of sensitivity of vaginal examination for the diagnosis of residual fistula in women operated upon within the previous 2 weeks. Alternatively, women may be so pleased to have obtained some level of benefit from surgery that they exaggerate the extent of their improvement.

61.5% of women with grade 4 incontinence at discharge reported their symptoms to be improved at follow up compared with 23.1% of those with grade 5 incontinence at discharge. This still left most of them with significant symptoms, as the improvement tended only to be by a single grade. In comparison, many of those discharged with grade 3 incontinence returned free of symptoms. This observation has lead to a management policy within the Barhirdar Hamlin Fistula Centre such that women with an incontinence grades 4 and 5 at discharge are taught how to use the urethral plug;13 those with grades 2 and 3 are taught how to perform pelvic floor exercises.

A small proportion of women returned with late breakdown of their repairs, that is they were discharged continent and return with some degree of extraurethral incontinence. This has also been observed in Uganda (B. Hancock, pers. comm.). Although five of these six women reported an exacerbating physical event for the recurrence of their incontinence, research into the causes and associations of late fistula breakdown is required.

The follow-up survey also revealed that by 6 months after fistula repair, there was a significant effect on the women’s lives with many being reintegrated into society, sexual relationships could be restored for many and fertility resumed for some women even in the short term.

It could be argued that in this cultural setting with largely illiterate and disempowered women, their answers to the questions that were put to them reflected a wish to please the health provider who cared for them during their illnesses and did not give their true feelings or situation. While this is possible, it is thought to be unlikely as the women who still remained with incontinence answered mainly in the negative to the questions relating to their social situation and quality of life. It is acknowledged that neither the incontinence grading scheme used nor the follow-up questionnaire have been effectively validated. We felt, however, that a simple pragmatic approach to the assessment of outcome that could be translated into the local language and easily administered to our women was necessary. We accept that follow-up interviews carried out by an independent third party, perhaps in the community setting away from the hospital an point of care, might have given more valid responses.

It is noteworthy that at 6 months, the women living arrangements had not necessarily changed, that is most of the women who were living alone at the time of their injuries were still living alone. Follow up at 6 months is clearly inadequate to fully evaluate changes in social situation and longer term follow up is necessary to see changes in this regard, for example in getting remarried. Likewise, 6 months too short a time to make any conclusions about fertility rates following fistula repair.

It is noteworthy that 41.3% of women who were cured of their incontinence had resumed a sexual relationship at the time of follow up, whereas only 21% of those with residual symptoms had resumed intercourse. It is also interesting to note that at follow up, 44.8% of those women who were cured of their incontinence were still divorced or separated while of those who still suffered from some urinary incontinence, 62.7% were divorced or separated. There was very little change in the number of new divorces or remarriages between discharge and follow up, so why an injury with poorer outcome from repair might be more likely to result in divorce at presentation is unknown. The link may be made in that primiparous women are more likely to be divorced (and more likely to have no living children which in itself can lead to divorce in this culture) and are also more likely to have a fistula involving the urethra. This has been proved to give a worse outcome postrepair.14 Of the primiparous women, 66.7% were divorced, only 3% had living children and 83.8% had a fistula involving the urethra; whereas among multiparous women, 45.4% were divorced, 72.3% had living children and 45.4% had injuries involving the urethra.


In an appropriate healthcare setting, follow up of women following repair of obstetric VVF is feasible. There is a trend for those women who are discharged still incontinent to improve over time, especially if the fistula was of a less severe type; others, however, may demonstrate persistence or recurrence of symptoms. Further research is needed into ways to prevent and treat residual incontinence and into what factors might lead to late breakdown of following fistula repair.

Quality of life of women does seem to improve following fistula repair with reintegration into society, although it is possible that both the prognosis for successful restoration of continence and for reintegration may be poorer for primiparous women. However, 6 months follow up is inadequate to fully evaluate the social impact of treatment and longer term follow up with validated quality of life scales is needed.

Disclosure of interests

The authors have no financial or other interests to declare.

Contribution of authorship

A.B. conceived and designed the study, contributed to the data collection and performed or supervised all the surgery. He correlated and analysed the data, wrote and redrafted the paper and approved of the final version for publication. B.M. contributed to the design of the study, contributed to the data collection and helped revise and redraft the paper. He also approved the final version of the paper for publication.

Ethical approval

No ethics board exists in the hospital where the study was undertaken.


There was no specific funding for the study. The hospital paid the women transport fees and extra expenses as part of the normal hospital expenses.


The authors wish to thank Julie Morris, Head of Medical Statistics, University Hospital of South Manchester, Manchester, UK, for her help with analysing the data and also the staff of the Barhirdar Hamlin Fistula Centre, Ethiopia, for their cooperation.