Objectives To examine the long-term results of hypospadias fistula repair, the factors involved in recurrence and the outcome in cases where this has occurred. Patients and methods The study comprised 113 children undergoing urethrocutaneous fistula repair between 1984 and 1996. Most of the fistulae were closed in two to three layers, with or without a transpositional skin flap. Tunica vaginalis or a scrotal dartos flap was used in patients with inadequate vascularized tissue adjacent to the fistula. Success rates were calculated for each attempt at fistula repair until the patient was cured. Results The median (range) age at primary fistula repair was 40 (18-169) months and the median follow-up after the most recent repair 7.5 (2.3-17) years. The overall success rate of primary fistula repair was 71%. Fistulae which were >2 mm (11 of 21, 52%) were more likely to recur than were those < or = 2 mm (22 of 92, 24%). Recurrence did not relate to the initial form of hypospadias repair, to the means of skin closure nor, with the exception of multiple lesions, to the location of the fistula. The success rates of subsequent repairs were 70% at the second and 50% at the third, fourth and fifth repairs. One child was cured at the sixth attempt. The use of tunica vaginalis or scrotal dartos as a 'waterproofing' layer was limited to the third or subsequent repairs and was successful in five of six cases. Conclusion A simple layered closure with or with no transpositional skin flap is effective in 71% of repairs. For recurrent fistulae, tissues from an unscarred area (tunica vaginalis or scrotal dartos layer) should be used to cover the fistula.