Methods:
An IRB approved retrospective cohort study was performed to identify the incidence of and risk factors for perioperative blood transfusions after surgical repair of urogenital fistula in Uganda. The study was conducted among women who underwent fistula repair between 2013-2019 at the Kitovu Hospital in Masaka, Uganda. Kitovu Hospital is one of three private, faith-based hospitals under the partnership of the USAID Fistula Care Plus. Complex fistula cases from many geographical areas are often referred to the hospital which has a specialized fistula repair unit. All types of urogenital fistula were included (vesicovaginal, urethrovaginal, vesicouterine, and those with ureteral involvement). Those who had concomitant rectovaginal fistulas were included so long as they underwent repair of the urogenital fistula. A surgical database, established in 2013 from patient charts, was used to identify those who underwent urogenital fistula repair during that time period and determine eligibility for inclusion. After identification of included patients within our cohort, a retrospective chart review was performed to gather all remaining variables of interest including patient demographics, medical and surgical history, as well as perioperative details. Primary outcome was the need for perioperative blood transfusion. Patient characteristics were then compared between those who did and did not require perioperative blood transfusion after urogenital fistula repair.
Descriptive statistics were calculated for a number of demographic and clinical characteristics to describe the full sample and women with and without a transfusion. Results were reported as mean with standard deviation (or median with interquartile range) for continuous variables and frequency with percentages for categorical variables. Logistic regression was used to estimate unadjusted odds ratios (ORs) and 95% confidence intervals (CIs) representing the association between each risk factor of interest and blood transfusion. Risk factors were chosen a priori based on previous literature and clinical knowledge and included patient age, time with fistula, type of delivery, delivery outcome, cause of fistula, history of fistula repair, HIV status, presence of foot drop, surgical approach, anesthesia type, concomitant surgery, use of a graft or flap as part of their surgery, a concomitant sling procedure, complications, use of ureteric catheter, and surgeon training. A multivariable logistic regression model was used to estimate adjusted ORs and 95% CIs; only covariates that were significantly associated with risk of blood transfusion in univariate analyses were included in the adjusted model. Categorical subgroups were created for covariates with missing data and included in the models. All statistical tests were two-sided and declared significant at p < 0.05. Statistical analyses were performed with SAS software, version 9.4 (SAS Institute, Inc., Cary, NC).