The treatment of gastroduodenal outflow obstruction (GOO) caused by malignant diseases represents a significant challenge. Open surgical gastrojejunostomy (GJ) has been the treatment of choice, but it has high morbidity and mortality rates. More recently, endoscopic placement of self-expanding metallic stents (SEMS) has been proposed and the results of small, preliminary studies are encouraging. This study compared technical and clinical success, morbidity, mortality, and hospital stay in patients undergoing endoscopic and surgical treatment of GOO. Medical records of 60 consecutive patients with GOO seen between April 1997 and November 2002 were retrospectively reviewed. Because of extremely short life expectancy, 13 patients were treated by insertion of a double-lumen nasogastric-jejunal tube. The remaining 47 patients (28 men, 19 women; mean age 73.5 years, range 48-92 years) with unresectable pancreatic (33), gastric (7), metastatic lymph nodal (4), papillary (2), and biliary (1) tumors were treated by placement of a SEMS (24) or open surgical GJ (23). The technical success rates were similar, but clinical success was lower in the GJ group (92% vs. 56%, p = 0.0067). The SEMS group had a shorter length of hospital stay (3.0 [1.4] days vs. 24.1 [10.3], p < 0.001). Thirty-day mortality was 30% in the GJ group, and 0% in the SEMS group ( p = 0.004). Morbidity was higher in the GJ compared with the SEMS group (61% vs. 17%, p = 0.0021). Mean survival was longer in the SEMS group (96.1 [9.6] days vs. 70.2 [36.2] days, p = 0.0165 for a single test of hypothesis; Bonferroni correction for a multiple testing removes this significance), consequently, out-of-hospital survival was longer for the SEMS group (93.2 [9.3] days vs. 46.0 [31.5] days, p < 0.001). None of the endoscopic procedures required the assistance of an anesthesiologist or the use of an operating room. The results of this retrospective study suggest that SEMS insertion is better than surgical GJ for palliation of patients with GOO in terms of clinical success, morbidity, and mortality. Technical success rates were similar. SEMS placement should be proposed as the first-line treatment for relief of GOO. However, a randomized, comparative, prospective study of SEMS vs. laparoscopic GJ is needed.