To determine whether treatment of patients with large (>15mm) impacted upper ureteral stones depended on stone location, we prospectively evaluated the therapeutic outcomes, complications, safety, and effectiveness of percutaneous nephrolithotomy (PCNL) and ureteroscopic lithotomy (URSL) in patients with stones higher and lower than the upper border of the fourth lumbar vertebra. Of the 174 patients analyzed, 83 (47.7%) underwent PCNL and 91 (52.3%) underwent URSL; all patients were followed up 1 month later and every 6 months for 18 months. Mean operation time (108.76 +/- 19.36 vs. 63.56 +/- 16.38 minutes, p<0.05) and postoperative hospital stay (2.49 +/- 1.23 vs. 5.36 +/- 1.98 days, p<0.05) were significantly longer in the PCNL than in the URSL group. The overall stone-free rates after 1 month were 96.4% and 75.8%, respectively, differing significantly for stones higher (97.8% vs. 57.5%, p<0.05) but not lower (94.7% vs. 90.2%) than the upper border of the fourth lumbar vertebra. The stone-retropulsion rate of URSL differed significantly for stones higher and lower than the upper border of the 4th lumbar vertebra (47.5% vs. 9.8%, p<0.05). Postprocedural complication rates were comparable in the URSL and PCNL groups, although the rate of auxiliary or salvage procedures was higher in the URSL group. The efficiency quotients (EQ) for PCNL and URSL were 0.93 and 0.59, respectively, with EQs in the URSL group differing significantly for stones higher and lower than the upper border of the fourth lumbar vertebra (0.40 vs. 0.82, p<0.05). Our findings indicate that treatment of impacted upper ureteral stones is dependent on stone location relative to the upper border of the fourth lumbar vertebra. URSL is unsuitable for stones at a higher location, whereas URSL and PCNL were equally effective for stones at a lower location.