A Clinical Model for the Early Diagnosis of Acute Pancreatitis in the Emergency Department

Abstract

Objective: This study aimed to develop a diagnostic model that predicts acute pancreatitis (AP) risk before imaging.

Methods: Emergency department patients with serum lipase elevated to 3 times the upper limit of normal or greater were identified retrospectively (September 1, 2013-August 31, 2015). An AP diagnosis was established by expert review of full hospitalization records. Candidate predictors included demographic and clinical characteristics at presentation. Using a derivation set, a multivariable logistic regression model and corresponding point-based scoring system was developed to predict AP. Discrimination accuracy and calibration were assessed in a separate validation set.

Results: In 319 eligible patients, 182 (57%) had AP. The final model (area under curve, 0.92) included 8 predictors: number of prior AP episodes; history of cholelithiasis; no abdominal surgery (prior 2 months); time elapsed from symptom onset; pain localized to epigastrium, of progressively worsening severity, and severity level at presentation; and extent of lipase elevation. At a diagnostic risk threshold of 8 points or higher (≥99%), the model identified AP with a sensitivity of 45%, and a specificity and a positive predictive value of 100%.

Conclusions: In emergency department patients with lipase elevated to 3 times the upper limit of normal or greater, this model helps identify AP risk before imaging. Prospective validation studies are needed to confirm diagnostic accuracy.