Prognostic Models for Predicting the Severity and Mortality in People With Acute Pancreatitis – A Systematic Review and Meta-Analysis

Introduction Acute pancreatitis is a common condition, and clinical scoring systems are used to identify patients who may develop severe acute pancreatitis (SAP). The primary aim was to assess the accuracy of Ranson, Glasgow-Imrie, acute physiology and chronic health evaluation (APACHE II), and bedside index of severity in acute pancreatitis (BISAP) scoring systems for predicting SAP and mortality.Materials and methods A search of multiple electronic databases was conducted, and we included all studies that reported accuracy data for one or more of the following scoring systems for adults within 48 h of admission; Ranson, Glasgow-Imrie, APACHE II, and BISAP. We included studies that reported the outcome of SAP (organ failure ≥48 h) or mortality. The meta-analytical summary of the area under the curve (AUC) for each score at different time points was determined.Results Eighty-seven studies that included 92 patient cohorts (65,942 participants) were included. The highest pooled AUC for each scoring system (and time point) in predicting SAP were 0.88 (0.83-0.92) for Glasgow-Imrie (within 48 h), 0.87 (0.81-0.91) for ranson (within 48 h), 0.85 (0.82-0.88) for BISAP (within 24 h), and 0.82 (0.78-0.85) for APACHE II (within 24 h). The highest pooled AUC for each scoring system in predicting mortality were 0.89 (0.84-0.92) for BISAP (within 24 h), 0.88(0.81-0.93) for Glasgow-Imrie (admission), 0.86(0.84-0.89) for APACHE II (within 24 h), and 0.86 (0.81-0.90) for Ranson (within 48 h).Conclusions There is no evidence that any of the four scoring systems are better than another to predict SAP or mortality. The accuracy changes based on the time at which they are performed.