OBJECTIVE To compare the results of repairing pelvi-ureteric junction (PUJ) obstruction by percutaneous endopyelotomy in children with a similar series carried out in adults. METHODS Nine children with primary (six) or secondary (three) PUJ obstruction were treated using a one-stage cold-knife percutaneous endopyelotomy. The success and morbidity rates were compared with a series of 61 adults with primary (46) or secondary (15) PUJ obstruction treated similarly. RESULTS In children, endopyelotomy was successful in five of six with primary and two of three with secondary PUJ obstruction, with a mean follow-up of 30 months (range 18-56). In the adults, endopyelotomy was successful in 38 of 46 (83%) with primary and 12 of 15 with secondary PUJ obstruction, an overall success rate of 82%, with a mean follow-up of 42 months (range 9-86). There was no statistical difference in the success rates with primary and secondary endopyelotomy between adults and children (P = 0.58). Failures were associated with high-grade hydronephrosis, a stenotic segment > 1.5 cm long and technical problems. Morbidity occurred in one of nine children and 3.2% of the adults. CONCLUSIONS This early experience suggests that percutaneous endopyelotomy can be performed safely and successfully in children with primary PUJ obstruction. However, in secondary stenosis, the results were less than optimal. Larger series should be analysed to form definitive conclusions on the role of endopyelotomy for the treatment of PUJ obstruction in children, given the high rate of success of open pyeloplasty and its minimal morbidity.