Utility of the "harmless acute pancreatitis score" in predicting a non-severe course of acute pancreatitis: a pilot study in an Indian cohort.

backgroundSeveral severity predictors have been tested for assessing acute pancreatitis (AP). The recently described harmless acute pancreatitis score (HAPS) could be an ideal predictor for Indian patients at the community level. We validate this system in the current study.

methodsThis was a prospective pilot study conducted at a tertiary center from July 2010 to December 2011. Consecutive directly admitted patients over 18 years with a documented first episode of AP were enrolled and followed for at least 12 months after discharge/till death. HAPS was defined as absence of rebound abdominal tenderness, serum creatinine of <2 mg/dL, and hematocrit of <43 for male and <39.6 for female patients at the time of admission; and it was considered positive if the patient fulfilled all three criteria. Study outcomes included total hospital stay, need for intensive care unit (ICU), ICU stay, development of local complications, organ failure, hospital-acquired infections (including infected necrosis), and in-hospital mortality.

resultsThere were a total of 103 patients directly admitted with AP during the study period, out of which, 23 were excluded. Forty-seven (58.8 %) patients were positive for HAPS. Of these, 44 (93.6 %) had an eventual non-severe (mild) clinical course with odds ratio 17.6 (95 % CI 4.5-68.3). Sensitivity, specificity, positive and negative predictive value, and receiver operating characteristics area under the curve of HAPS as a predictor of non-severe disease were 76.3 (66.9-86.4), 85.7 (78.0-96.8), 93.8 (88.5-98.6), 56.6 (45.4-73.6), and 84.8 (76.9-92.7) respectively.

conclusionThis study validated the utility of HAPS for directly admitted patients with AP in India. Large-scale multicenter community-based studies need to be performed.