Construction of a Mortality Prediction Model for Cardiac Arrest Patients Using Lactate Clearance Rate and Albumin‐Corrected Anion Gap

Background: Cardiac arrest (CA) causes systemic circulatory failure, leading to postcardiac arrest syndrome (PCAS) and high mortality. Lactic clearance rate (LCR) and albumin-corrected anion gap (ACAG) reflect metabolic status but lack combined validation for predicting short-term mortality in CA patients. Objective: This study explores the predictive value of LCR at different time points and ACAG for 7-day mortality in post-CA patients and constructs a predictive model. Methods: Using the MIMIC-IV 3.0 database, we included eligible CA patients and calculated LCR at 6, 12, and 24 h, along with initial ACAG. ROC curves and logistic regression were used to assess predictive value, while restricted cubic splines (RCS) explored mortality risk. The best predictors were selected for model construction, with multivariate logistic regression identifying independent risk factors. Internal and external validations were performed. Results: < 0.05). The 24-h LCR combined with ACAG had the highest predictive efficacy (area under the ROC curve [AUC] = 0.716). Independent risk factors included 24-h LCR, ACAG, CK-MB, shockable rhythm, and CA etiology. The final model achieved an AUC of 0.775 (internal) and 0.743 (external). Conclusion: The 24-h LCR combined with ACAG predicts short-term mortality in post-CA patients who survive the first 24 h, supporting risk stratification after early postresuscitation stabilization.