Controversy persists regarding the role of prophylactic gastrojejunostomy in patients with unresectable periampullary adenocarcinoma. In review of the retrospective series, presence of gastric outlet obstruction (GOO) has been claimed to be a bad prognostic sign. This prospective study aimed to clarify the necessity of routine prophylactic gastrojejunostomy in patients with unresectable periampullary adenocarcinoma. The incidence and prognostic significance of GOO were also evaluated. Sixty-six patients with unresectable periampullary adenocarcinoma were enrolled. They were divided into 2 groups to receive either a single biliary bypass or a double bypass (concomitant gastric and biliary bypasses) if they had GOO. The single bypass group was followed up to assess the incidence of GOO and subsequent need of a gastric bypass. Prognostic factors were evaluated by univariate and multivariate analyses. Forty-four (67%) of the overall 66 patients presented with GOO at the time of diagnosis. There was no statistical difference regarding the morbidity and mortality between the 2 groups, except longer (7 days) hospital stay in the double bypass group. Seven (31.8%) of the 22 patients in the single bypass group subsequently developed GOO an average of 6.2 +/- 1.0 months after their initial biliary bypass. By univariate analysis, significant prognostic factors for unresectable periampullary adenocarcinoma were: GOO (p = 0.0379), pancreatic head origin (p = 0. 0146 by univariate analysis), and distant metastasis (p < 0.0001). After multivariate analysis, only pancreatic head origin and distant metastasis remained significant independent factors of poor prognosis. In conclusion, gastrojejunostomy should be performed prophylactically in addition to a biliary bypass in patients with unresectable periampullary adenocarcinoma. The presence of GOO is not an independent factor of poor prognosis, but a reflection of the aggressive biologic behavior of pancreatic head adenocarcinoma.