Prognostic Value of the Neutrophil Percentage-to-Albumin Ratio for All-Cause Mortality in Critically Ill Patients with Acute Myocardial Infarction: A Retrospective Cohort Study Based on the MIMIC-IV Database

BackgroundCritically ill patients with acute myocardial infarction (AMI) face high mortality, and existing risk stratification tools are limited. The neutrophil percentage-to-albumin ratio (NPAR), reflecting both systemic inflammation and nutritional status, may serve as a simple prognostic marker. MethodsData were extracted from the MIMIC-IV (version 2.0) database. Adult AMI patients with a first ICU admission between 2008 and 2019 were included. NPAR was calculated within 24 hours of ICU admission and categorized into quartiles. Cox regression models assessed associations with all-cause mortality. Kaplan-Meier analysis evaluated survival differences. Predictive performance of logistic regression, random forest, and XGBoost models was compared by ROC curves and AUC, and model interpretability was assessed using SHapley Additive exPlanations (SHAP). ResultsA total of 928 patients were included. Higher NPAR was associated with older age, greater illness severity, and higher mortality at all time points (p < 0.001). In univariate and partially adjusted Cox models, NPAR was an independent predictor of mortality, though significance diminished in the fully adjusted model (HR = 1.01, 95% CI 1.00-1.03, p = 0.055). Kaplan-Meier curves showed significantly poorer survival in higher quartiles (log-rank p < 0.0001). Logistic regression yielded the best predictive performance (AUC = 0.740), outperforming random forest (0.722) and XGBoost (0.709). SHAP analysis revealed a nonlinear effect of NPAR, with the strongest impact at intermediate levels, and age modified its contribution. ConclusionElevated NPAR is associated with increased mortality in critically ill AMI patients. As a readily available and low-cost biomarker, NPAR may aid early risk stratification, though further prospective validation is required.