An early predictive model for acute respiratory distress-syndrome related to pancreatitis in pregnancy: an 8-year multicenter analysis.

BACKGROUND: Acute respiratory distress syndrome (ARDS) related to acute pancreatitis in pregnancy (APIP) is associated with a higher risk of maternal and fetal death. This study aimed to explore early predictors and develop a predictive model for ARDS associated with APIP aligned with the updated global ARDS definition. METHODS: The APIP data of two hospitals over an 8-year period were retrospectively collected. The variables were analyzed using Least Absolute Shrinkage and Selection Operator regression, and binary logistic regression to build a predictive model, visualized with a nomogram. The performance of the predictive model was then evaluated. RESULTS: Of 6597 patients with acute pancreatitis, 103 pregnant patients were included, and 24 pregnant patients had ARDS. Lower oxygen saturation as measured by pulse oximetry to fraction of inspired oxygen (SpO2/FiO2) ratio, elevated heart rate (HR), and total cholesterol (TCH) were identified as independent risk factors for ARDS in APIP. Compared to previous scoring systems, the predictive model was more discriminating between APIP and ARDS, with an area under the receiver operating characteristic curve of 0.926 (95% CI 0.864-0.988). Notably, the new model performed best when the prediction cutoff was set at 0.205 (sensitivity = 0.823, specificity = 0.958). Calibration and decision curve analyses confirmed the strong clinical utility and accurate risk prediction. CONCLUSION: A new accurate utility predictive model for ARDS related to APIP, including three simple variables (HR, TCH, and SpO2/FiO2), was constructed at admission. Elevated total cholesterol level was first identified as an independent risk factor for ARDS in patients with APIP.