INTRODUCTION: Vesico-vaginal fistula (VVF) is one of the most important diseases related to quality of life in women's diseases and postoperative complications. We aimed to identify preoperative predictive factors of fistula repair failure and to develop a predictive model for this disorder. PATIENTS AND METHODS: The data from patients who underwent VVF repair from January 2010 and December 2020 were recorded and analyzed. The therapeutic results were assessed after a follow-up of at least six months. A successful VVF repair was defined as: closed VVF as noted on visual inspection of both the bladder and the vagina, no subjective complaints of vaginal leakage, no evidence of leakage during Valsalva and cough from the vaginal closure area using a half-speculum. Variables associated with fistula repair failure in the univariate analysis were included in a multivariate model using binary logistic regression to determine the independent factors of this disorder. All statistical analyses were performed using SPSS, and significance was set at 0.05%. RESULTS: Ninety-eight patients were identified. The median age of patients was 50 (interquartile range: 42 - 55) years. Failure of surgical treatment of VVF has been reported in 32 (32.7%) cases. Patients with failed surgical treatment had the higher size of fistula (p = 0.03), supratrigonal fistula (p = 0.02) and vaginal fibrosis (p = 0.001). On multivariate analysis, vaginal fibrosis (Adjusted OR = 4.2
95% CI = 1.398 - 12.739
p = 0.01) and supratrigonal VVF (Adjusted OR = 3.3
95% CI = 1.18 - 9.26
p = 0.02) were independent factors of fistula repair failure. These two predictors were used to calculate the probability of combined success given by the following formula:ⅇ CONCLUSION: Treatment failure was, in our series, correlated with the quality of the vaginal tissue and the seat of the fistula. This study demonstrated that vaginal fibrosis and supratrigonal fistula location are independent risk factors for failure of surgical treatment of VVF.