In acute pancreatitis (AP) tumor necrosis factor-alpha mediates multi-organ failure; in animal models its blockade with pentoxifylline ameliorates AP. The efficacy of pentoxifylline in predicted severe AP (pSAP) was tested in a double-blinded, randomized, control trial. Twenty-eight patients with pSAP were randomized within 72 hours of diagnosis to pentoxifylline or placebo. Baseline characteristics were similar in both groups. The pentoxifylline group had fewer intensive care unit admissions and shorter intensive care unit and hospital stays of longer than 4 days (all P < .05). Patients receiving pentoxifylline had no adverse effects. Pentoxifylline within 72 hours of pSAP is safe; a larger study of pentoxifylline in AP is needed to confirm efficacy. ClinicalTrials.gov number: NCT01292005.