S224 Evaluating Lipase Elevation Thresholds for Diagnosing Acute Pancreatitis Using Cross-Sectional Imaging as a Gold Standard

Introduction: Serum lipase elevation greater than 3 times the upper limit of normal (ULN) is a key criterion for the diagnosis of acute pancreatitis (AP). Since the sensitivity (∼80%) and specificity (∼90%) of lipase remains suboptimal, this study investigates whether different lipase thresholds may improve diagnostic accuracy for AP. Methods: This is a single-center retrospective analysis of 1,603 patients evaluated for AP with an elevated serum lipase and an abdominal computed tomography or magnetic resonance imaging obtained within 48 hours of lipase testing. Patients were identified using a text alert system with any abnormal lipase value. A receiver operator curve, sensitivity, and specificity were calculated using cross-sectional imaging as the gold standard for AP. Youden’s J statistic was calculated to evaluate the optimal lipase threshold in predicting positive imaging. Results: Youden’s J statistic found an optimal diagnostic cut-off of 2.73xULN with a sensitivity of 0.71 and specificity of 0.82. A cut-off of 3xULN yielded a sensitivity of (64%) and a specificity of (84%). A threshold of 1.34xULN lead to optimal sensitivity (sensitivity 90%, specificity 40%) and 4.3xULN led to optimal specificity (sensitivity 56%, specificity 90%). There was a trend for men to have a lower sensitivity (58%) than women (sensitivity 74%) for a 3xULN threshold with similar specificity (80%). Conclusion: The current lipase threshold of 3xULN indeed represents the best combination of sensitivity and specificity, but gender may impact sensitivity. Yet almost 30% of lipase levels below 3xULN had pancreatitis by imaging. This suggests that cross-sectional imaging should be pursued to rule out acute pancreatitis when the pre-test probability remains high.