Lactate-to-albumin ratio improves prediction of perioperative mortality in lung transplant patients

Abstract

Backgrounds Perioperative mortality remains elevated among lung transplant recipients. The lactate-to-albumin ratio (LAR) is a prognostic indicator, but its role in predicting mortality in lung transplant patients remains unclear. The study aims to assess the incremental prognostic value of LAR and develop a more accurate prediction model for perioperative mortality in lung transplant patients. Methods This study conducted a single-center, retrospective analysis involving 379 lung transplant recipients. A base model was developed to predict 30-day postoperative mortality and a modified model was then constructed by incorporating LAR. Model performance was evaluated using the area under the curve (AUC), Brier score, and calibration curves. DeLong and likelihood ratio tests compared models, while continuous Net Reclassification Improvement (NRI) and Integrated Discrimination Improvement (IDI) quantified accuracy improvements. Results Of 379 patients, 51 (13.46%) died within 30 days post-transplant. LAR was found to be an independent predictor of perioperative mortality in lung transplant patients. Other predictors in the modified model included blood transfusion, cold ischemia time, and fluid balance. The inclusion of LAR significantly improved the modified model’s discriminatory ability (AUC: 0.807 vs. 0.764; p = 0.024) and calibration (Brier score: 0.0927 vs. 0.1025). The likelihood ratio test confirmed a significantly improved model fit (p < 0.0001). The continuous NRI and IDI were 0.643 and 0.090, respectively (p < 0.001). A nomogram and an interactive web-based tool were developed to facilitate individualized risk assessment. Conclusion The modified model incorporating LAR provides a more accurate and comprehensive assessment of perioperative mortality risk.