Human pancreatic secretory trypsin inhibitor in the assessment of the severity of acute pancreatitis. A comparison with C-reactive protein.

We assessed the ability of serum human pancreatic secretory trypsin inhibitor (hPSTI) to establish the severity of acute pancreatitis and compared it in this respect to that of serum C-reactive protein (CRP). Of 26 patients studied with acute pancreatitis, 16 had mild pancreatitis, and 10, severe disease. Initial studies were performed at onset of the disease in 20 patients, on the second day of illness in two, and on the third day of illness in the remaining four. In all, serum hPSTI and CRP concentrations were determined on admission and daily for the following 5 days using commercial kits; Ranson's score was evaluated within the first 48 h of admission. Sixty-three healthy subjects and 31 patients with nonpancreatic acute abdomen were also studied. Values of 70 ng/ml for serum hPSTI and 10 mg/dl for serum CRP were taken as limits to distinguish severe from mild-to-moderate acute pancreatitis. When assessed within the first 24 h of pain, serum hPSTI correctly classified 71% of the patients with severe acute pancreatitis, whereas serum CRP did so for 29%. In subsequent days, the two markers showed a similar sensitivity in predicting severe acute pancreatitis. Serum hPSTI and CRP were alike in excluding a diagnosis of severe acute pancreatitis. Ranson's score correctly identified 50% of patients with severe illness and 63% of patients with mild pancreatitis. This study indicates that, when assessed within 24 h of pain onset, serum hPSTI is a better predictor of the severity of acute pancreatitis than serum CRP or Ranson's criteria.