This study investigates the prognostic value of early CT in acute pancreatitis, the role of pancreatic necrosis as a indicator of prognosis, and the need for the routine use of IV iodinated contrast material in early CT to assess prognosis in these patients. We conducted a retrospective review of 148 patients who underwent unenhanced and contrast-enhanced helical CT within 72 hr after onset of symptoms of a first episode of acute pancreatitis. Patients were classified by CT grade and grouped into two categories (mild: grades A, B, C; and severe: grades D and E) that were correlated with complications and death. In the grades including patients with pancreatic necrosis, it was also correlated with complications and death. All complications (n = 15) and deaths (n = 4) occurred in patients with a CT grade of severe disease; differences as compared with mild grade were significant (p < 0.001 and p < 0.03, respectively). CT grade had a sensitivity and specificity of 100% and 61.6%, respectively, for predicting morbidity and 100% and 56.9% for predicting mortality. The 13 patients with necrosis were all in the severe group (p < 0.001). Necrosis detection on early CT had a sensitivity and specificity of 53.3% and 90.2%, respectively, for predicting morbidity and 75% and 83.8% for mortality. Early unenhanced CT alone was a good indicator of severity of acute pancreatitis in our selected population. CT grade was sensitive for predicting outcome in acute pancreatitis. Pancreatic necrosis, estimated on early, contrast-enhanced CT and seen only in patients having severe disease, was a specific predictor of morbidity and mortality. These findings lead us to suggest that the use of iodinated contrast material to assess necrosis can be reserved for only those patients classified as having severe disease on unenhanced CT.