Evaluating Risk Factors Associated with Ureteric Involvement in Women with Urogenital Fistulae in Uganda

Introduction

Urogenital fistulas are defined as abnormal connections between the genital tract and urinary tract or rectum. The most-common etiologies include obstetric or iatrogenic causes. In developing countries, obstetric fistulas are frequently caused by prolonged or obstructed labor. In situations where emergency intervention is impracticable or impossible, impaction of the neonate's presenting part can cause local ischemia and tissue necrosis to adjacent maternal pelvic tissues. In contrast, urogenital fistulas are more-often attributed to iatrogenic causes in Western countries, such as secondary to surgery or radiation. In both etiologies, the resulting tissue damage can lead to aberrant connections between the vaginal tract and urinary or gastrointestinal tracts causing urinary and fecal incontinence. ¹ Ureteric injury can occur in both etiologies of fistula and may lead to an ureterovaginal fistula. Injury often occurs at the distal ureter where it crosses under the uterine arteries in close proximity to the cervix. ²

The prevalence of urogenital fistulas in sub-Saharan Africa is difficult to estimate and is likely underreported, as patients are more likely to live in rural areas where women may be difficult to reach. ¹ In Uganda, ∼2% of women have experienced obstetric fistulae, representing the third-highest rate of fistulae in the world. Many patients live with untreated fistulae for years due to obstacles, such as lack of knowledge that repair is possible, expense, and embarrassment. ³ Untreated fistulae lead to many medical complications, including infertility, chronic infection, dermatitis, and gastrointestinal disturbances. Additionally, cultural factors lead to significant social consequences that impact community involvement and quality of life. ⁴

Ureteric injury has been estimated to occur in ∼1 per 1000 cesarean sections in well-resourced settings.^5,6 The incidence of ureteric injury in resource-limited settings is unknown; however, studies have shown obstetric operations to be the main cause of ureteric injury in populations in these settings.^2,7 One retrospective review of 365 ureteric injuries in low-resource settings in Africa and Asia supported this trend, with 67.4% of ureteric injuries following obstetric procedures, 17.8% following gynecologic procedures, and 14.8% following surgical repair of obstetric fistulae. Of these, all patients except 2 anuric patients, presented with ureterovaginal fistulae. ² Of patients with obstetric procedures as the etiology of their ureterovaginal fistulae, 156 underwent cesarean sections, 18 sustained uterine ruptures requiring repair, and 68 underwent cesarean hysterectomies. ²

The current study was performed to identify risk factors for ureteric involvement to prepare for surgical planning and counseling better. Data have shown fistula location to be a significant indicator of successful surgical repair.^8,9 However, diagnosis of fistula location and determining fistula characteristics prior to surgery can be challenging. The best first diagnostic test is the double-dye tampon test, which is cost-effective and minimally invasive to patients. It is performed by injecting dye into the bladder and intravenously (IV), and examining staining patterns on the tampon to localize the fistula. ¹⁰ In evaluation of ureteral injury, imaging with computed tomography (CT) urography and IV pyelography can be used to classify fistula location and characteristics better. ¹⁰ Unfortunately, these tests are often not available in resource-limited settings and the primary surgeon may not know that the fistula involves the ureter until surgery is performed. ²

There are multiple classification systems of urogenital fistulae, with the first described system being the Goh classification system. Goh classification accounts for distance from the external urinary meatus (1–4), largest diameter(A, B, C), and fibrosis and special circumstances (i, ii, iii). ⁹ To the authors' knowledge, the association between Goh classification and ureteric involvement has not yet been described. However, studies have shown high fistula location to be associated with multiparity. ¹¹

One study used the Goh classification system to find that primiparous women were more likely to have distal fistulae (type 4), and multiparous women were more likely to have proximal fistulae, (type 1). ¹² In analyses using Goh classification primiparous women were also more likely to have moderate-to-severe scarring (category iii). While ureteric involvement is a special consideration for type iii, there was not a significant association with parity and ureteric involvement in that study. ¹²

Additionally, a study in Malawian women showed proximal fistula location was associated with both multiparity and peripartum operations, including cesarean delivery and hysterectomy during delivery. ¹¹ One theory for the correlation of parity and fistula location is the different causes of failed labor in these populations. In primiparous women, early fetal head descent and cephalopelvic disproportion may lead to more distal obstruction. This may also lead to longer duration of labor in primiparous women. Conversely, in multiparous women who have had prior successful vaginal delivery, cephalopelvic disproportion is less likely the cause of unsuccessful vaginal delivery. Therefore, arrest of descent may occur higher in the pelvis due to a different cause and may lead to a more proximal fistula.

Considering these limited data specific to fistulas with ureteric involvement, the current research was conducted to examine associated patient and fistula characteristics further to determine the risk factors for ureteric involvement. This may be particularly useful for stratifying a patient's risk preoperatively when advanced imaging is not available. This would enable surgeons in low-resource settings to prepare better for surgery and counsel patients accordingly. The hypothesis for this research was that ureteric fistulae would be associated with more-proximal locations, shorter labor, multiparity, and increased scarring.

Materials and Methods

This was a retrospective cohort study performed at Kitovu Mission Hospital in Masaka, Uganda, and was reviewed by the University of South Florida's (Tampa, Florida, USA) institutional review board (USF IRB; #Pro00039861). The primary objective was to identify the risk factors associated with ureteric involvement to prepare better for surgical planning and counseling. Kitovu Hospital is a private hospital that provides specialized medical and surgical care, including urogenital fistula repair, and receives referrals from throughout Uganda and neighboring African countries.

In the current study, inclusion criteria were women of any age who underwent urogenital fistula repair at Kitovu Mission Hospital between 2013 and 2019. Those who had concomitant rectovaginal fistulae were included if the urogenital fistula was addressed during surgery. Surgeries were performed by experienced fistula surgeons from Uganda and other countries.

All procedures were performed vaginally with spinal anesthesia when possible. If indicated by surgeon discretion, an abdominal approach was used with general anesthesia. Standard surgical technique included trimming and subsequent closure of fistulae in multiple tension-free layers. Intravesical instillation of dye was then performed intraoperatively to confirm watertight integrity of the fistula repair. Perioperative antibiotics were administered to all patients. Urinary drainage via catheter was maintained for 7–14 days postoperatively, with shorter length reserved for less-complex repairs. During this time, patients were admitted as inpatients and data were recorded into a surgical database. This database was used to identify patients who underwent urogenital fistula repair during the study timeframe and who met the inclusion criteria.

The cases were then divided into 2 groups: (1) urogenital fistula without ureteric involvement and (2) urogenital fistula with ureteric involvement. Ureteric involvement was determined intraoperatively during fistula repair and typically resulted in operative reimplantation of ureters. Characteristics between the 2 groups were compared to determine differences and potential risk factors for ureteric involvement. Variables were collected via the Kitovu Fistula Lynch Structured Patient Record. These variables included sociodemographic characteristics, relevant past medical history, operative details, and postoperative complications.

Results

A total of 546 patients with urogenital fistulae met inclusion criteria. Of these, 50 patients had ureteric involvement (9.2%). All analyses compared patients with ureteric involvement to patients without ureteric involvement. The patients' demographics and clinical characteristics are in Table 1. There were no significant differences between the groups in age, body mass index, length of symptoms, urinary incontinence, fecal incontinence, prior fistula repair, concomitant bowel involvement, or International Consultation on Incontinence Questionnaire score. Patients with ureteric involvement had significantly higher parity of 5.38 ± 2.88, compared with 3.55 ± 2.80 in patients without ureteric involvement (p < 0.0001). Patients with ureteric involvement also had shorter duration of labor at 48.8 ± 27.5 hours, compared with 63.17 ± 49.79 hours in patients without ureteric involvement (p = 0.003).

Whether the baby was born at home, with a birth attendant, at a health center, or at a hospital was not significantly different between the groups. All patients with ureteric involvement gave birth at a hospital (95.6%) or a health center (4.4%). Patients without ureteric involvement gave birth at a hospital (85.3%), health center (8.4%), home with a birth attendant (0.5%), or home without a birth attendant (5.9%).

Delivery type distribution was significantly different between the 2 groups (p = 0.016). There were higher rates of cesarean sections and cesareans with hysterectomies in women with ureteric involvement, compared to those without ureteric involvement (73.9% and 13% versus 61.4% and 5.8%, respectively). Delivery outcomes measured by stillbirth, live birth, and early neonatal death were not significantly different between the 2 groups. Further information on delivery characteristics is in Table 2.

In patients with ureteric involvement and without ureteric involvement, the etiology of the majority of fistulae was obstetric (92% versus 96%, respectively), followed by surgical complications (4% versus 4%, respectively) and congenital origins (4% versus 0%, respectively). Additional information about fistula characteristics is in Table 3.

Goh classifications for type and size were not significantly different between the 2 groups. There was a trend toward more proximal fistulae in patients with patients with ureteric involvement, with 64% having type 1, 27% having type 2, and 9% having type 3. No patients with ureteric involvement had type 4 fistulae. There was a statistically significant difference between the 2 groups for Goh classification of scarring (p = 0.013). Results for ureteric involvement versus no ureteric involvement for scarring showed type i (27% versus 56%), type ii (9% versus 20%), and type iii (64% versus 24%), all respectively. Additional details on Goh classifications are in Table 4.

Discussion

Historically, the majority of patients with urogenital fistulae differ in developed countries, compared with developing countries, with the leading etiologies being iatrogenic and obstetric, respectively. Obstetric fistulae in Africa have been a leading cause of morbidity for decades and are largely attributed to lack of access to emergency obstetric care. Urogenital fistulae with ureteric involvement accounts for a small proportion of all fistulae, and there are limited data on etiology and associated risk factors. The current study showed that a large proportion of ureteric fistulae in Uganda were secondary to surgical injuries by cesareans and cesarean hysterectomies. This may indicate that access to emergency obstetric care is improving; however the rates of ureteric injury during these operations raises concern. Injury to ureters can be difficult to identify during surgery and may develop over time. One method for identifying an ureteric injury promptly is intraoperative cystoscopy. ¹⁰ Unfortunately, in under-resourced settings, cystoscopy may not be available and injury may not be identified until a ureterovaginal fistula has developed. ²

One theory for the high rates of ureteric injuries during cesarean delivery and cesarean hysterectomy is decreased surgeon familiarity, as these procedures may be performed less frequently. Women may present for obstetric intervention later and may have histories of prior operations, increasing the difficulty of surgical intervention. In unfortunate cases when a fetus is determined to have demised, alternative delivery methods that avoid intra-abdominal surgery may be considered. While less-commonly performed in contemporary obstetrics, these options include destructive delivery, symphysiotomy, and vacuum extraction, and these approaches may be more-appropriate for delivery of a nonviable fetus, as they may confer less risk of ureteric injury and subsequent fistulae. ¹³

Additionally, in resource-limited settings without access to advanced imaging modalities such as CT urography and IV pyelography, the surgeon may be unaware that a fistula involves a ureter until the time of surgery. For this reason, this research was conducted to identify risk factors in women with ureteric fistulae to aid in raising clinical suspicion. The study results corroborated existing data that the most-common cause of ureterovaginal fistulae is obstetric procedures in low-resource countries.^2,7 The results also showed that there is a statistically significant association between parity and ureteric involvement, with higher parity being associated with ureteric involvement.

A previous study examined the relationship between high fistulae—which included fistulae involving the vaginal apex, cervix, uterus, or ureters—and parity as well as cesarean section. That study found high fistulae was significantly associated with higher parity and history of cesarean section. ¹¹ The current study's results corroborated that study, finding significant associations between ureteric involvement and parity, delivery type, and prior nonfistula operations. This is significant in that it strengthens existing evidence and confirms that these relationships hold true for ureteric involvement.

To the authors' knowledge, there are limited data on the relationship between Goh classification and determination of ureteric involvement in urogenital fistulae. While the current study's results were limited by missing data for these variables, the extant data showed a novel trend toward more-proximal fistula location, smaller fistula size, and type iii (indicating increased scarring or special consideration, represented as Goh classification 1Aiii) in patients with ureteral involvement. Multivariate regression was not feasible with this data, but further studies examining this relationship may contribute more to improve diagnosis and management of advanced urogenital fistulae.

The current study's data contribute to the resources that surgeons can use to identify patients who may be at increased risk of having complex urogenital fistulae, such as patients with ureteral involvement. This may be particularly beneficial in resource-limited settings where advanced imaging or other evaluation is not available or practicable. In a woman presenting with urinary leakage from her vagina with an unknown fistula location, clinical suspicion should increase in if she has higher parity and/or previous cesarean delivery or cesarean hysterectomy.

Conclusions

There are limited data on risk factors for urogenital fistulae with ureteric involvement. The data from the current study helps fill this gap in knowledge by identifying significant relationships between ureteric involvement and higher parity, shorter duration of labor, and cesarean or cesarean hysterectomy. This is also first study, to the authors' knowledge, that examined the relationship between Goh classification and determination of ureteric involvement in urogenital fistulae, showing a trend toward Goh classification 1Aiii. These data can be used to stratify patients' risk preoperatively and assist in surgical planning, particularly in resource-limited settings.