Objective: To evaluate the long-term outcome of secondary endopyelotomy after failed primary intervention for uretero-pelvic junction (UPJ) obstruction and to assess the effect of preoperative parameters on treatment outcome. Methods: Twenty patients (13 men, seven women; mean age 30.7 years) who underwent secondary endopyelotomy after the failure of a primary intervention for the treatment of congenital UPJ obstruction were included in this retrospective analysis. Mean interval from primary treatment to secondary endopyelotomy was 27.2 months (range 3-123 months). The diagnosis of failure of the primary treatment was based on symptoms and the results of imaging studies. Treatment success was defined as symptomatic relief with either stable or improved renal function and improved wash-out shown on diuretic renogram or excretory urography. Results: Mean follow-up was 47.2 months (range 6.2-138.8 months). Success rates were as follows: overall, 70%; after primary dismembered pyeloplasty, 66.7%; after primary endopyelotomy, 57.1%; after primary balloon dilatation, 100%. Kaplan-Meier estimates of success were 64.4% at 5 years. Six patients in whom the procedure failed at a mean of 13.8 months (range 4-33 months) were treated with open pyeloplasty (four patients), simple nephrectomy (one), and a repeat endopyelotomy (one). Grade 4 hydronephrosis and significant obstruction occurred more often in the failure group. Conclusions: Endopyelotomy is an acceptable minimally invasive secondary treatment option for UPJ obstruction. Preoperative severe hydronephrosis and the presence of a significant obstruction seem to be risk factors for the failure of a secondary endopyelotomy.