Assessment of the endothelial activation and stress index in predicting all-cause mortality among patients resuscitated after cardiac arrest: a retrospective cohort study using the MIMIC-IV database

Abstract Background Endothelial Activation and Stress Index (EASIX) is a reliable surrogate biomarker of endothelial dysfunction, a critical pathophysiological process in ischemia-reperfusion injury following cardiac arrest. This study aimed to investigate the predictive ability of EASIX for in-hospital mortality in patients resuscitated after cardiac arrest and to assess its potential to enhance existing predictive models for critically ill patients. Methods We extracted data on patients diagnosed with cardiac arrest from the Medical Information Mart for Intensive Care (MIMIC-IV) database. The EASIX score was calculated using the formula: lactate dehydrogenase (U/L) × creatinine (mg/dL) / platelet count (10^9/L). Patients were stratified into tertiles based on log2-transformed EASIX scores. The primary and secondary outcomes were ICU and in-hospital mortality, respectively. Cox regression was used to evaluate the association between log2-EASIX and mortality. To investigate the nature of this association, restricted cubic spline analysis was applied to determine if it was linear. The predictive ability of EASIX for death risk was assessed using the Receiver Operating Characteristic (ROC) curve, C-index, Net Reclassification Improvement (NRI), and Integrated Discrimination Improvement (IDI). Results A total of 905 patients were included, with a mean age of 64.82 ± 16.81 years, of whom 337 (37.24%) were female. Cox regression analysis showed that compared with the first tertile, the second and third tertiles had higher ICU mortality risks [HR (95% CI) 1.41 (1.02-1.93); HR (95% CI) 2.39 (1.67-3.42)]; P for trend < 0.001], with similar findings for in-hospital mortality. Restricted cubic spline analysis revealed an approximately linear relationship between log2-EASIX and ICU and in-hospital mortality. Subgroup analysis found no interaction among patients with different comorbidities (diabetes, heart failure, chronic kidney disease, stroke, chronic lung disease) or different support treatments (Continuous Renal Replacement Therapy and mechanical ventilation). Additionally, incorporating log2(EASIX) levels into the Acute Physiology Score 3 and Sequential Organ Failure Assessment scores significantly improved the prediction performance of all-cause mortality, with increases in the C-index (from 0.673 to 0.688; 0.638 to 0.666), positive IDI values (0.027, 95% CI: 0.009-0.054, P < 0.001 and 0.022, 95% CI: 0.007-0.044, P < 0.001), and NRI values (0.143, 95% CI: 0.055-0.227, P < 0.001; 0.113, 95% CI: 0.022-0.187, P < 0.001). Conclusions EASIX is an independent risk factor for ICU and in-hospital mortality in patients resuscitated after cardiac arrest. Adding EASIX to existing models can enhance their predictive performance, providing clinical value in identifying high-risk patients.