Objective: To investigate the value of urinary trypsi-nogen activation peptide (TAP) in the early predic-tion of severe acute pancreatitis and to compare itwith acute physiology and chronic health evaluationⅡ (APACHE Ⅱ).Methods: We assessed the predictive value of urinaryTAP concentrations measured by a competitive en-zyme-linked immunosorbent assay. Urine sampleswere collected for detecting TAP concentrations atadmission, and 24, 48, and 72 h from 41 patientswith acute pancreatitis (12 with severe disease, 29with mild disease) who presented within 48 h the on-set of symptoms and from 11 control patients, whileAPACHE Ⅱ scores were recorded at 48 h after ad-mission.Results: The peak median urinary TAP concentrationwas seen at admission. The median urinary TAPconcentration at admission for severe pancreatitis (95nmol/L) was significantly higher than the medianfor patients with mild pancreatitis (20 nmol/L, P<0. 005) and controls (15 nmol/L, P<0. 005). TAPconcentrations were significantly higher in patientswith severe acute pancreatitis than the median in pa-tients with mild pancreatitis (P<0. 05) and controls(P<0. 05) on days 2 to 3. The median APACHE Ⅱscores of severe patients were significantly differentfrom those of mild patients (10.5 vs 6.0, P<0.01).The sensitivity, specificity, positive predictive, andnegative predictive values of an admission urinaryTAP≥35 nmol/L for severe pancreatitis were91.7%, 89.7%, 78.6% and 96.3%, whereas 48 hafter admission the values for APACHE Ⅱ scores(≥9) were 75.0%, 72.7%, 52.9% and 87.5%. Inprediction of disease severity, the urine TAP concen-tration was much better than APACHE Ⅱ at 48 h.Conclusions: Urinary TAP obtained at the first 48 hof the onset of symptoms can predict severe acutepancreatitis. In prediction of disease severity, theurinary TAP is much better than APACHE Ⅱ score.