To determine the effect of early aggressive parenteral support in pancreatitis, 54 patients with acute pancreatitis were randomized to receive either conventional therapy (control group) or conventional therapy plus the institution of total parenteral nutrition within 24 hours. The two groups were similar demographically. The total parenteral nutrition group had a significantly higher rate of catheterrelated sepsis than did an additional group of contemporaneous patients without pancreatitis who received total parental nutrition (10.5 percent and 1.47 percent, respectively; p < 0.01). There was no advantage to the use of early total parenteral nutrition; that is, there was no difference in the number of days to oral intake, total hospital stay, or number of complications of pancreatitis. Patients with zero or one Ranson's criterion on admission were more likely to be eating by the seventh hospital day than were those with two or more Ranson's criteria (80 percent and 54 percent, respectively; p < 0.05). The early institution of total parenteral nutrition in patients with acute pancreatitis did not appear to improve the outcome. Its use should be limited to prolonged periods of no oral intake or treatment of a specific complication, such as a pseudocyst.