COMPARATIVE PREDICTIVE ACCURACY OF HAPS AND BISAP FOR EARLY SEVERITY ASSESSMENT IN ACUTE PANCREATITIS

Abstract

Background and Objectives: Acute pancreatitis (AP) ranges from mild self-limiting disease to life-threatening severe acute pancreatitis (SAP) with multi-organ failure. Early and accurate risk stratification is essential for guiding triage and management decisions. The Bedside Index of Severity in Acute Pancreatitis (BISAP) and the Harmless Acute Pancreatitis Score (HAPS) represent two widely used early scoring tools with contrasting designs. This retrospective study compared their diagnostic performance for predicting SAP and clinical outcomes.

Methods: A retrospective study was conducted on 41 patients admitted with acute pancreatitis over a 24-month period. BISAP and HAPS scores were calculated using clinical and laboratory data obtained at admission. Disease severity was classified according to the Revised Atlanta Classification (2012). The primary outcomes were severe acute pancreatitis in-hospital mortality and ICU admission. Diagnostic performance was assessed using receiver operating characteristic analysis together with sensitivity, specificity, positive predictive value and negative predictive value. Wilson 95% confidence intervals were calculated for all estimates.

Results: Among the 41 patients included in the study, 26 (63.4%) developed severe acute pancreatitis and 5 (12.2%) died during hospital admission. The mean ICU stay was 5.7 ± 7.8 days. BISAP performed better than HAPS in predicting severe disease with an AUROC of 0.712 compared with 0.494. Using a cutoff value of 2 or higher, BISAP achieved 100% specificity and a positive predictive value of 100% but sensitivity remained low at 42.9% with a negative predictive value of 44.4%. HAPS showed lower overall performance with a sensitivity of 37.5% and specificity of 66.7% when scores of 1 or more were considered indicative of non-harmless disease. Neither score showed a statistically significant association with severe acute pancreatitis at the predefined thresholds on Chi-square testing (BISAP p=0.120 and HAPS p=0.814) which was likely related to the small sample size.

Conclusions: In this cohort BISAP performed better than HAPS for identifying patients at risk of severe acute pancreatitis mainly because of its high specificity. However neither score was sensitive enough to exclude severe disease when used alone. HAPS was originally designed to identify low-risk patients suitable for early discharge and therefore performed poorly in this population where severe disease was common. Neither BISAP nor HAPS predicted in-hospital mortality with acceptable accuracy in our cohort. The inclusion of biomarkers such as CRP or procalcitonin may improve prediction and deserves further evaluation. Because only 41 patients were available for analysis these findings should be interpreted cautiously and require confirmation in larger studies.