Cyclooxygenase-2 (COX-2) inhibitors are being developed as chemopreventive and anticancer agents. This study aimed to determine the biological effect of the COX-2 inhibitor celecoxib in pancreatic cancer as an early step to the further development of the agent in this disease. Eight patients scheduled for resection of an infiltrating adenocarcinoma of the pancreas were randomized to receive celecoxib at a dose of 400 mg twice daily or placebo for 5 to 15 days before the surgery. In addition, carcinomas from nine additional patients were xenografted in nude mice, expanded, and treated with vehicle or celecoxib for 28 days. Celecoxib markedly decreased the intra-tumor levels of prostaglandin E2 in patient carcinomas and in the heterotransplanted xenografts. However, this effect did not result in inhibition of cell proliferation or microvessel density (as assessed by Ki67 and CD31 staining). In addition, a panel of markers, including bcl-2, COX-1, COX-2, and VEGF, did not change with treatment in a significant manner. Furthermore, there was no evidence of antitumor effects in the xenografted carcinomas. In summary, celecoxib efficiently inhibited the synthesis of prostaglandin E2 both in pancreatic cancer surgical specimens and in xenografted carcinomas but did not exert evident antitumor, antiproliferative, or antiangiogenic effect as a single agent. The direct pancreatic cancer xenograft model proved to be a valuable tool for drug evaluation and biological studies and showed similar results to those observed in resected pancreatic cancer specimens.