The aim of the study was to determine the clinical epidemiology of genitourinary fistulae as seen at Komfo Anokye Teaching Hospital in Kumasi, Ghana. A retrospective study was carried out from the hospital records and operative reports of all patients with genitourinary fistulae seen at Komfo Anokye Teaching Hospital between January 1977 and December 1992. Patient age, parity, type of fistula and cause of fistula were abstracted from the medical records. There were 164 cases of genitourinary fistula managed during the study period. There were 150 fistulae due to obstetric causes (91.5%), the vast majority of which were due to prolonged obstructed labor (121 cases, 73.8% of all fistulae), with a minority related to complications of lower-segment cesarean section (14 cases, 8.5% of all fistulae). In 5 cases (3.1%) patients developed a rectovaginal fistula owing to perineal tears and prolonged obstructed labor. During this time period there were 157 449 deliveries, giving an obstetric fistula rate of 1 fistula per 1000 deliveries. Obstetric fistulae were most common at the extremes of reproductive age and parity. Fourteen additional fistulae (8.5% of all cases) were due to gynecologic causes, most commonly from surgical injury occurring at the time of abdominal hysterectomy for leiomyomata uteri (12 cases, 7.3% of all fistulae). It was concluded that in Kumasi, Ghana, obstetric trauma from prolonged obstructed labor is the most common cause of genitourinary fistula formation. Such fistulae occur in older multiparous women as well in young primigravidae. Obstructed labour can, and does, occur in women who have previously undergone uneventful vaginal delivery. Birth attendants should be aware of that fact. Prompt referral for obstetric intervention should be made in obstructed labor, irrespective of the age and parity of the patient.