Background: Gemcitabine is still standard chemotherapy in pancreatic cancer. Pancreatic cancer patients often have cachexia with body weight loss. It had been reported that eicosapentaenoic acid (EPA) reduced proinflammatory cytokines, leading to improve inflammatory status of cachexia and attenuation of body weight loss in patients with advanced pancreatic cancer (Barber, et al, 1999).We hypothesized that EPA-enriched oral supplement may prevent cancer cachexia and prolong survival in advanced pancreatic cancer with chemotherapy. The aim of this randomized study is to evaluate the efficacy and safety of gemcitabine with an EPA enriched oral supplement compared with gemcitabine monotherapy in patients with unresectable advanced pancreatic cancer. Methods: This is a multicenter randomized phase II study. Patients will be randomized in a 2:1 ratio to Arm A or Arm B. In Arm A, 1,000mg/m² of gemcitabine will be administered in a 30 minute intravenous infusion on days 1, 8, and 15. This cycle will be repeated every 4 weeks until disease progression. In addition, a maximum of 2 packs of an EPA-enriched oral supplement (Prosure, Abbott) will be administered every day. Arm B patients will receive gemcitabine monotherapy. Major inclusion criteria include patients over 20 years old with recurrent or unresectable pancreatic cancer that is histologically verified, an ECOG PS of 0,1 or 2, sufficient oral intake. The primary endpoint is the evaluation of 1-year survival. The secondary endpoints are progression-free survival, response rate, adverse events, QOL evaluation, body weight and performance status. A sample size of 66 patients was chosen on the basis of the randomized phase II selection design by Simon et al (1985). The design has a 78% correct selection probability if 1-year survival probabilities are 35% and 25% for two treatment arms. To date, 34 patients have been enrolled.