In an attempt to decrease the infectious complications of acute pancreatitis and its high mortality, many investigators have conducted randomized prospective trials on the efficacy of prophylactic antibiotics. The results of these studies are conflicting, and many have called for a large multicenter study. Because multicenter trials are costly and difficult to organize, we believe that meta-analysis is a reasonable alternative. A meta- analysis of all eight previously published trials of prophylactic antibiotics in acute pancreatitis was performed. The end point was death. The Mantel-Haenszel statistic was used to summarize odds ratios across studies in a fixed effects model, after homogeneity was assessed. Sensitivity analysis was performed as appropriate. The meta-analysis of all eight trials showed a positive benefit for antibiotics in reducing mortality. Sensitivity analysis showed that the advantage was limited to patients with severe pancreatitis who received broad- spectrum antibiotics that achieve therapeutic pancreatic tissue levels. It is recommended that all patients with severe pancreatitis be treated with broad- spectrum antibiotics that achieve therapeutic levels in pancreatic tissue.