Objective: The aim of this paper is to determine which is the best ear
ly prognostic score in acute pancreatitis among those we have studied.
Design: Prospective study of acute pancreatitis cases during a 5 year
period. Patients: 113 patients, biliary etiology in 93 cases, (averag
e age: 63 years) and 66% females. The diagnostic accuracy of Ranson, I
mrie, Osborne and apache II scores was analyzed. Results: There were c
omplications in 12%, and mortality was 8%. Among all the systems anali
zed the Ranson criteria achieved the highest sensitivity (79%), althou
gh sensitivity was improved with Apache II at admission to the hospita
l (86%). We have obtained the best specificity with the Apache II (89%
) and with the Osborne criteria (88%). The modification of Ranson's cr
iteria for biliary etiology didn't improve the sensitivity (56%) to de
tect severe cases of acute pancreatitis. Conclusions: The Apache II sy
stem was the best for the early detection of severe acute pancreatitis
. Its diagnostic sensitivity was higher than the multiple laboratory c
riteria analyzed.