RESUMO
BACKGROUND/AIM: To examine the validity of Ranson's criteria in the prediction of the severity of acute pancreatitis (as judged by the occurrence of complications) in a high-altitude area of Saudi Arabia with a predominant biliary pancreatitis. MATERIALS AND METHODS: All consecutive cases of acute pancreatitis (AP) admitted to a tertiary care hospital over a two-and-half-year period were included in this prospective study. Ranson's criteria (RC) were used to determine the severity of the attack of AP, which was then correlated with the occurrence of complications. The validity of Ranson's score and that of each of its individual components was estimated. Using receiver operating characteristic (ROC) curve, new optimum values for these components were calculated and a new modified score was constructed. RESULTS: Seventy-three attacks of AP in 69 patients formed the material of this study. Ranson's prediction criteria classified 43.8% of the attacks as "severe", but only 22% of those attacks were associated with complications. Calcium level (<8 mg/dl) was the only criterion that was significantly associated with complications (Kappa = 0.32, P = 0.02). Using ROC curve to determine the optimum cut-off levels for prediction identified only four criteria, which were significantly associated with complications as compared with the original Ranson's cut-off levels. Those were: a serum glucose value of >or=160 mg/dl (P < 0.05), blood urea nitrogen rise of >or=35 mg/dl (P < 0.02) and an arterial Po(2) value of
