Validation of the Harmless Acute Pancreatitis Score in Predicting Nonsevere Course of Acute Pancreatitis

Background: The Harmless Acute Pancreatitis Score (HAPS) is a scoring algorithm to identify patients with nonsevere acute pancreatitis. The aim of this study was to evaluate the reproducibility of HAPS outside its original study setting. Method: Baseline information of all hospitalized patients with acute pancreatitis at Karolinska University Hospital, Stockholm, Sweden, between 2004 and 2009 was collected. The parameters constituting HAPS were signs of peritonitis, hematocrit and serum creatinine levels. Since hematocrit was not available in all patients, complete sample analysis was performed by replacing hematocrit with hemoglobin (strongly correlated with hematocrit; r = 0.86). Results: In total, 531 patients with a first-time or a recurrent attack of acute pancreatitis were included. Among 353 patients with complete information on parameters constituting HAPS, 79 patients were predicted to have a nonsevere course, of whom 1 patient developed severe acute pancreatitis. The specificity of HAPS in predicting a nonsevere course of acute pancreatitis was 96.3% (95% CI: 81.0-99.9) with a corresponding positive predictive value of 98.7% (95% CI: 93.1-100). Complete sample analysis replacing hematocrit with hemoglobin level predicted a nonsevere course in 182 patients, of whom 2 patients had severe acute pancreatitis (94.3% specificity and 98.9% positive predictive value). Conclusion: HAPS is a highly specific scoring algorithm that predicts a nonsevere course of acute pancreatitis. Therefore, HAPS might be an additional tool in the clinical assessment of acute pancreatitis where early screening is important to treat the patients at an optimal level of care.