Association between monocyte-to-lymphocyte ratio and mortality in patients with acute pancreatitis requiring intensive care unit admission: a retrospective cohort study and predictive model establishment based on machine learning.

The purpose of this study was to evaluate the predictive value of monocyte-to-lymphocyte ratio (MLR) on the short-term (28 days) and long-term (365 days) mortality risk in patients with acute pancreatitis (AP) using multiple statistical and machine learning (ML) models. Studies selected 1,044 eligible AP patients from the MIMIC-IV database and divided them into four groups based on their MLR values (MLR<0.32; 0.32 ≤ MLR<0.57; 0.57 ≤ MLR<1; MLR ≥ 1). Findings revealed that MLR demonstrated a U-shaped relationship with patient mortality risk, with the minimal mortality risk occurring at an MLR of approximately 0.57. Cox regression model analysis showed that after adjusting for multiple parameters, MLR was still significantly associated with the risk of death. Moreover, ML model analysis identified that MLR has potential value in predicting AP patient outcomes. This study suggests that MLR can be used as a potential indicator to assess prognostic risk in critically ill patients with AP to support clinical decision-making.