Intrapancreatic activation of trypsinogens is believed to occur either as a cause or a consequence of acute pancreatitis and to be associated with the more severe forms of the disease. Trypsinogen-activation peptides (TAPs) were measured in plasma, urine, and ascites of rats (n = 54) assigned to different pancreatitis-inducing regimens reproducing the entire spectrum of severity. Compared with survivors, nonsurvivors at 9 hours had significantly higher TAP levels in plasma at 3 hours (P = 0.0001), urine (peak, 1-4 hours) (P = 0.004), and ascites (P = 0.0001) after death. Stepwise discriminant analysis showed that TAP in urine and plasma were the most accurate predictors of outcome (88.2% of animals) compared with other routine laboratory parameters. Morphometric analysis showed that the best histopathologic correlates of TAP elevation were acinar necrosis and intrapancreatic hemorrhage. In a second series of experiments using a homogeneous technique of induction producing pancreatitis with a mortality of 55% at 48 hours, plasma TAP level at 3 hours (cutoff, 0.5 nmol/L) and/or urinary TAP level (peak, 1-6 hours; cutoff, 25 nmol/L) accurately predicted outcome in 85% of animals. It is concluded that the TAP assay gives an accurate early prediction of outcome in different pancreatitis models and correlates best with acinar necrosis and hemorrhage.