OBJECTIVE To compare complication and success rates of antegrade and retrograde endopyelotomy performed over 10 years and to define possible risk factors associated with treatment failure. METHODS From 1994 to 2004, 61 patients underwent a total of 68 endoscopic treatments: 19 antegrade and 49 retrograde endopyelotomy procedures. Antegrade endopyelotomy was always performed using diathermy. In the first 18 procedures retrograde endopyelotomy was performed using diathermy. In the most recent 30 procedures the incision was made using holmium laser. Endoluminal ultrasound was used in 78% of retrograde endopyelotomy and in 5% of antegrade endopyelotomy. RESULTS The retrograde endopyelotomy patients demonstrated significantly lower complication rates (12.5% vs. 42%) and shorter hospital stay (1.5 vs. 7 days) than the antegrade endopyelotomy patients. The mean follow up of the patients who remained free from disease recurrence during the study period was 46 and 24 months for the antegrade and retrograde endopyelotomy group, respectively. The overall success rate (mean time to failure) of antegrade and retrograde endopyelotomy was 56% (31 months) and 70% (17 months), respectively. There was no statistically significant increase in the overall success rate of retrograde endopyelotomy using endoluminal ultrasound per se. Stratifying retrograde endopyelotomy by the type of energy used for the incision, the overall success rate (mean time to failure) was 80% (10 months) and 53% (21 months) for Holmium laser and diathermy, respectively (p = 0.0626). CONCLUSIONS The overall success of antegrade and retrograde endopyelotomy in this series appears to be largely a factor of lead-time bias and is similar enough to recommend retrograde endopyelotomy with holmium laser on the basis of its relative safety and shorter hospital stay.