Table 1.

Publications examining predictors of fistula repair outcomesin developing country settings

Author, Year	Study Design	Population	Sample size	Outcome definition	Exposures of interest	Analytic approacha
Kirschner et al., 201019	Retrospective record review	Patients with vesicovaginal fistula; where unit of analysis was individual patient, analyses were restricted to women undergoing first repair	1084 records from 926 patients	Continence (dry vs wet), assessed at time of discharge	Patient characteristics (age at surgery, education, parity, number of living children, literacy, language group and marital status), clinical data (cause of fistula and number of previous surgeries) and surgical data (type/location of fistula, degree of fibrosis, surgical approach, and procedures performed)	Independent sample t-tests and Chi-square tests GEE bivariate and multivariate regression. Multivariate models adjusted for confounding by days in labor, number of living children, marital status, months with fistula and place of delivery
Muleta et al., 201024	RCT	Patients with obstetric fistula undergoing first repair	722 patients	Fistula closure, assessed after catheter removal and prior to discharge	Single-dose Gentamycin vs. extended antibiotic use. Extended antibiotic use included any one or combination of Amoxicillin (500mg IV and oral 6 hourly), chloramphenicol (500mg IV and oral 6 hourly), or co-trimexazole (800mg orally every 12 hours) for 7 days	Chi-square, risk difference
Nardos et al., 200917	Retrospective record review	Patients with obstetric vesicovaginal fistula undergoing first repairs via vaginal route	1045 patients	Fistula closureb, assessed after catheter removal and prior to discharge	Extent of urethral destruction, circumferential damage, extent of scarring, residual bladder size, repair technique (single vs double layer closure)	Logistic bivariate and multivariate regression
Lewis et al. 20099	Retrospective record review	Patients with genitourinary fistula	505 records from 435 patients	Continence (dry vs. wet), assessed via subjective appraisal after catheter removal and prior to discharge	Patient demographics (age), obstetric history (index pregnancy), and fistula parameters (number of prior repairs, fistula type, site and size, degree of fibrosis, and urethral status	Chi-square and Wilcoxon rank sum test; bivariate analyses stratified by primary vs. subsequent repair GEE multivariate regression
Olusegun et al. 200915	Retrospective record review	Patients with vesicovaginal fistula	37 patients	Continence (dry vs. wet) at discharge (personal communication H. Onah, July 2011)	Duration of fistula before repair	Chi-square
Safan et al. 200925	RCT	Patients with complicated fistula (defined as recurrence, local moderate to severe fibrosis, fistula location involving the bladder neck, and or size of the fistula being more than 1.5 cm in largest diameter)	38 patients	Continence (dry vs. wet), assessed at three months follow-up	Primary exposures were fibrin glue vs martius flap as interpositioning layer. Also examined parity, patient age, attempts of previous repairs, fistula size, and fistula location	Chi-square or Fisher’s Exact tests
Goh et al. 20086	Prospective	Patients with genitourinary fistula (women with rectovaginal fistula only or no bladder tissue excluded)	987 patients	Fistula closure and residual urinary incontinence following successful closure, assessed after catheter removal and prior to discharge	Components of Goh’s classification system: Fistula type (characterized by distance of fistula from external urinary meatus), size, “special considerations” (extent of fibrosis and vaginal length, and special circumstances such as previous repair, ureteric involvement, etc)	Chi-square test and logistic multivariate regression (residual incontinence only)
Morhason- Bello et al. 200820	Retrospective record review	Patients with midvaginal fistulas with no fibrosis, evidence of infection, urethral or bladder neck involvement and more than one previous repair attempt	71 patients	Continence three months following surgery	Abdominal versus vaginal route of repair	Fisher’s exact test
Nardos et al. 200813	Retrospective record review	Patients with obstetric fistula (women with rectovaginal fistula only excluded)	212 patients	Fistula closure and residual incontinence, assessed after catheter removal and prior to discharge (differences at 6- month follow-up not tested)	3 duration of catheterization groups: 10 days (group 1), 12 days (group 2), and 14 days (group 3)	Unspecified (chi-square assumed); bivariate analyses stratified by components of Goh classification system
Raassen et al. 20087	Prospective	Patients with obstetric fistula undergoing first- time repair	581 patients	Fistula closure assessed via dye test prior to catheter removal (14-21 days following surgery) and residual urinary incontinence following successful closure assessed after catheter removal prior to discharge	Patient characteristics (age and duration of leakage) and components of Waaldijk classification system (type of fistula characterized by extent of involvement of closing mechanism and presence of circumferential defect, exceptional fistulas and size)	Chi-square and Fisher’s Exact tests and logistic multivariate regression (closure only)
Holme et al. 20078	Retrospective record review	Patients with obstetric fistula	259 patients	Closure, not closed, residual incontinence; time period unspecified	Scarring	Spearman correlation
Browning 200611	Retrospective record review	Patients with obstetric fistula (women with rectovaginal fistula only excluded)	413 repairs	Fistula closure assessed via dye test prior to catheter removal (14-21 days following surgery) and residual urinary incontinence following successful closure assessed after catheter removal prior to discharge	Martius graft	Fisher’s Exact test or Chi-Square with continuity correction; bivariate analyses stratified by components of Goh classification system and other fistula characteristics
Browning 200618	Retrospective record review	Patients with obstetric fistula (women with breakdown of repair, lack of bladder tissue and rectovaginal fistula only excluded)	481 women	Residual incontinence following fistula closure, assessed following catheter removal and prior to discharge	Urethral involvement, repeat surgery, size of fistula, size of bladder, location of ureter, scarring, flap required, presence of rvf, number of vvf, age, parity, duration labor, time since delivery, diameter of fistula, delivery method and outcome of delivery	T-test, Mann-Whitney U test, and logistic multivariate regression
Chigbu et al. 200612	Retrospective record review	Patients with juxtacervical vesicovaginal fistula	78 women	Fistula closure at either 6 weeks or 3 months (personal communication H. Onah, July 2011)	Route of repair (vaginal vs. abdominal)	T-tests and Chi-square tests
Melah et al. 200610	Retrospective record review	Patients with vesicovaginal fistula	80 women	Fistula closure and residual incontinence following closure; time period of assessment unspecified	Early (less than 3 months) vs. late (after 3 months) closure	Chi-square
Kriplani et al. 200516	Retrospective record review	Patients with genital fistula (radiation fistulas excluded)	34 women	Continence following catheter removal	Age, parity, duration of fistula, route of repair, etiology	Levene’s test of equality of variances and Chi square with Yates correction
Murray et al. 200221	Retrospective record review	Patients with obstetric fistula	55 women	Residual incontinence following fistula closure, assessed between four weeks and three months following repair	Mean fistula diameter	Wilcoxon signed rank sum test
Rangnekar et al. 200014	Retrospective record review	Patients with urinary-vaginal fistulas (excluded fistulas situated high on the posterior wall of the bladder and fistulas greater than 1.5cm in size)	46 women	Fistula closure assessed via dye test prior to catheter removal and residual incontinence following closure, assessed with urodynamic test 3 weeks postoperatively.	Martius flap repair	Fisher’s exact test
Tomlinson and Thornton 199823	RCT	Patients with obstetric vesico- vaginal fistula	79 women	Fistula closure and continued incontinence (positive pad test) at hospital discharge.	500 mg ampicillin	Mann-Whitney (non-parametric tests)
Bland and Gelfand 197022	Prospective	Patients with vesicovaginal fistula	60 women	Closed fistula 6 weeks after repair	Urinary bilharziasis defined by presence of ova on bladder biosopsy or urine examination or rectal snip	Chi square with Yates correction

^aOnly the analytic approach for the outcome of interest is reported

^bUnless otherwise specified, fistula closure was assessed using dye test if the patient reported urine leakage