A Simple 3-Lab-Value Score Predicts Death Risk in Heart Disease Patients

Coronary artery disease is widely understood as a localized blockage problem, but mounting evidence shows it is also a systemic endothelial disease. Damaged and dysfunctional blood vessel linings throughout the body — not just at the site of a coronary stenosis — contribute meaningfully to long-term mortality. Yet no widely validated, easily calculated biomarker of systemic endothelial dysfunction exists for routine cardiology practice. This study asked whether EASIX, a formula originally created in hematology to flag endothelial crises after stem cell transplantation, could fill that gap in CAD patients.

The study enrolled 3,217 patients who underwent coronary catheterization at University Hospital Heidelberg between January 2005 and December 2017. All had confirmed CAD (stenosis ≥ 25% in any segment). The training cohort comprised 1,283 patients whose EASIX was measured at a median of 52 days before catheterization, deliberately chosen to reflect a baseline endothelial health state rather than an acute ischemic event. The independent validation cohort comprised 1,934 patients whose EASIX (labeled EASIXval) was measured at a median of 174 days after catheterization. EASIX is calculated as: LDH (U/L) × Creatinine (mg/dL) ÷ Platelet count (per nL). The median age across both cohorts was 71 years; 70.7% were male.

In the training cohort, log2-transformed EASIX was a robust predictor of overall survival in multivariable Cox regression: hazard ratio 1.29 per log2 unit (95% CI 1.18–1.41, p < 0.001), adjusting for age, sex, LVEF, presence of ≥ 90% stenosis, arterial hypertension, and diabetes. Kaplan–Meier curves stratified by EASIX quartile showed progressively worse survival with each ascending quartile. Two pre-established external cut-offs were tested: EASIX ≥ 0.88 (the 75th percentile of healthy volunteers) carried a hazard ratio of 1.63 in training; EASIX ≥ 2.32 (a previously validated sepsis-risk threshold in transplant patients) carried a hazard ratio of 3.57 in training. Pre-catheterization EASIX correlated strongly with post-catheterization EASIXval (Spearman rho = 0.70), confirming measurement stability across time.

In the independent validation cohort, EASIXval confirmed the predictive signal with an even stronger hazard ratio of 1.53 per log2 (95% CI 1.42–1.64, p < 0.001). The pre-established cut-offs replicated closely: HR 1.67 for the ≥ 0.88 threshold and HR 4.65 for the ≥ 2.32 threshold (both p < 0.0001). Model performance was assessed via integrated Brier score and two concordance indices (Harrell C-Index and Gönen & Heller K), and the Cox model built on the training cohort predicted validation cohort survival with minimal misspecification — the regression coefficient of the offset prognostic index was near zero, confirming transportability of the model.

Importantly, EASIX predicted mortality independently of hs-cTnT and NT-proBNP — established cardiac biomarkers — suggesting it captures a distinct biological dimension: systemic endothelial stress rather than direct myocardial injury. This is clinically meaningful because EASIX components (creatinine, LDH, platelet count) are measured in virtually every routine blood panel, making the index immediately computable without any additional cost. The authors propose that EASIX could serve as a stratification tool to identify CAD patients who might benefit most from more intensive cardiovascular risk factor management or emerging therapies targeting endothelial dysfunction. However, causality cannot be inferred from this retrospective design, and the single-center academic hospital setting may limit generalizability to community cardiology populations.