Establishment and Diagnostic Value of an Early Prediction Model for Acute Pancreatitis Complicated With Acute Kidney Injury

Abstract

To establish an early prediction model for acute pancreatitis (AP) complicated with acute kidney injury (AKI) and evaluate its diagnostic value. AP patients were recruited from the Emergency Department at Peking University People's Hospital in 2021 and stratified into AKI and control (no AKI) groups. Their clinical data were analyzed. The risk for AKI development was determined using logistic analyses to establish a risk prediction model, whose diagnostic value was analyzed using a receiver operating characteristic curve. There was no significant difference in the basic renal function between the AKI (n = 79) and control (n = 179) groups. The increased triglyceride glucose index (odds ratio [OR], 2.613; 95% confidence interval [CI], 1.324–5.158; P = 0.006), age (OR, 1.076; 95% CI, 1.016–1.140; P = 0.013), and procalcitonin (OR, 1.377; 95% CI, 1.096–1.730, P = 0.006) were associated with AKI development. A model was established for prediction of AKI (sensitivity 79.75%, specificity 96.65%). The area under the receiver operating characteristic curve was 0.856 which was superior to the Ranson, Bedside Index for Severity in AP, and Acute Physiology and Chronic Health Evaluation II scores (0.856 vs 0.691 vs 0.745 vs 0.705). The prediction model based on age, triglyceride glucose, and procalcitonin is valuable for the prediction of AP-related AKI.