Background: There is a pressing clinical requirement for an early simple test of severity in acute pancreatitis. We investigated the use of an assay of trypsinogen activation peptide (TAP).

Methods: We undertook a multicentre study in 246 patients (172 with acute pancreatitis [35 with severe disease], 74 controls). We assessed the predictive value of urinary TAP concentrations measured by a validated competitive immunoassay. We compared the results with those for plasma C-reactive protein and three clinicobiochemical scoring systems. TAP and C-reactive protein concentrations were analysed at set times after symptom onset and compared with the clinicobiochemical systems scores at key times during hospital stay.

Findings: At 24 h after symptom onset, the median urinary TAP concentration was 37 nmol/L (IQR 17-110) for severe and 15 nmol/L (5-35) for mild disease (p<0.001). The respective values for plasma C-reactive protein were 24 mg/L (3-34) and 25 mg/L (6-75; p=0.208). The sensitivity, specificity, positive predictive, and negative predictive values of the test to show severe acute pancreatitis compared with mild acute pancreatitis at 24 h were: for TAP (>35 nmol/L), 58%, 73%, 39%, and 86%, respectively, and for C-reactive protein (>150 mg/L), 0%, 90%, 0%, and 75%. 48 h after admission the values for the clinicobiochemical scoring systems were: APACHE II (> or =8), 56%, 64%, 30%, and 85%; Ranson score (> or =3), 89%, 64%, 38%, and 96%; and Glasgow score (> or =3), 77%, 75%, 44%, and 93%. At 48 h, the values for C-reactive protein were 86%, 61%, 37%, and 94% and for TAP were 83%, 72%, 44%, and 94%. Combined testing of C-reactive protein and TAP was not superior to TAP alone for accuracy.

Interpretation: Urinary TAP provided accurate severity prediction 24 h after onset of symptoms. This single marker of severity in acute pancreatitis deserves routine clinical application.