Background and goals: Acute pancreatitis runs an unpredictable course. We prospectively analyzed the prognostic usefulness of four different scoring systems in separately assessing three variables; acute pancreatitis severity, development of organ failure and pancreatic necrosis.
Study: 78 patients with acute pancreatitis were studied prospectively. Data pertinent to scoring systems were recorded 24 hours (APACHE II and III scores), 48 hours (Ranson score) and 72 hours (Balthazar computed tomography severity index) after admission. Statistical analysis was performed by using receiver operating characteristic curves and by comparing likelihood ratios of positive test (LRPT) for all three outcome variables.
Results: 44 patients were classified as mild and 34 as severe pancreatitis. When we compared LRPT, only that for the Balthazar score (11.2157) was able to generate large and conclusive changes from pretest to post-test probability in acute pancreatitis severity prediction. LRPT were 2.4157 for Ranson, 4.0980 for APACHE II and 3.6670 for APACHE III score. The APACHE II and III scores and Ranson criteria performed slightly better than the Balthazar score in predicting organ failure (LRPT: 4.0667, 3.2892, 3.0362 and 1.7941 respectively), while when predicting pancreatic necrosis the APACHE II and III performed slightly better than the Ranson score (LRPT: 2.0769, 2.7500 and 1.7813 respectively).
Conclusions: In all outcome measures the APACHE scores generate small and of similar extent changes in probability. The Balthazar score is superior to other scoring systems in predicting acute pancreatitis severity and pancreatic necrosis. However the Ranson and APACHE scores perform slightly better with respect to organ failure prediction.