BACKGROUND/AIMS: No single classification system has so far effectively predicted the severity for Acute Pancreatitis (AP). This study compares the effectiveness of classification systems: Original Atlanta (OAC), Revised Atlanta (RAC), Determinant based classification (DBC), PANC 3, Harmless AP Score (HAPS), Japanese Severity Score (JSS), Symptoms Nutrition Necrosis Antibiotics and Pain (SNNAP), and Beside Index of Severity for AP (BISAP) in predicting outcomes in AP. METHODS: Scores for BISAP, Panc 3, HAPS, SNNAP, OAC, RAC, and DBC were calculated for 221 adult patients hospitalized for AP. Receiver Operating Characteristic curve analysis and Akaike Information Criteria were used to compare the effectiveness of predicting need for surgery, intensive care unit (ICU) admission, readmission within 30 days, and length of hospital stay. RESULTS: Both the RAC and the DBC strongly predict the length of hospital stay (p < 0.0001 for both) and ICU admission (p < 0.0001 for both). Additionally, both BISAP and PANC 3 showed weak predictive capacity at identifying length of stay and ICU admission. CONCLUSIONS: We suggest that BISAP and PANC3 be obtained within the initial 24 h of hospitalization to offer an early prediction of length of stay and ICU admission. Subsequently, RAC and DBC can offer further information later in the course of the disease.
BACKGROUND/AIMS: No single classification system has so far effectively predicted the severity for Acute Pancreatitis (AP). This study compares the effectiveness of classification systems: Original Atlanta (OAC), Revised Atlanta (RAC), Determinant based classification (DBC), PANC 3, Harmless AP Score (HAPS), Japanese Severity Score (JSS), Symptoms Nutrition Necrosis Antibiotics and Pain (SNNAP), and Beside Index of Severity for AP (BISAP) in predicting outcomes in AP. METHODS: Scores for BISAP, Panc 3, HAPS, SNNAP, OAC, RAC, and DBC were calculated for 221 adult patients hospitalized for AP. Receiver Operating Characteristic curve analysis and Akaike Information Criteria were used to compare the effectiveness of predicting need for surgery, intensive care unit (ICU) admission, readmission within 30 days, and length of hospital stay. RESULTS: Both the RAC and the DBC strongly predict the length of hospital stay (p < 0.0001 for both) and ICU admission (p < 0.0001 for both). Additionally, both BISAP and PANC 3 showed weak predictive capacity at identifying length of stay and ICU admission. CONCLUSIONS: We suggest that BISAP and PANC3 be obtained within the initial 24 h of hospitalization to offer an early prediction of length of stay and ICU admission. Subsequently, RAC and DBC can offer further information later in the course of the disease.