Abstract 4357: Soluble ST-2 Concentrations to Predict Sudden Cardiac Death in Stable Heart Failure Patients
Background & Aim: The soluble form of the ST2 receptor (sST2) is released, like B-natriopeptides, by cardiomyocytes in response to mechanical overload. sST2 has been shown useful for predicting short- and long-term risk in heart failure (HF). Sudden cardiac death (SCD) is a poorly predictable HF-complication; we previously showed that NT-proBNP and left atrial size (LA) were prognostic of SCD in a cohort of stable HF patients. Since NT-proBNP and sST2 share some common features, we evaluated the role of sST2 to predict SCD in HF.
Patients & Methods: A 1:2 case:control design was used. Ninety-nine stable HF patients were included and followed during 3 years (942±323 days); 36 suffered SCD (Cases) and 63 survived (Controls). Cases and Controls were matched by age (66±11 years), sex (83% male) and LVEF (mean: 29%). NT-proBNP and sST2 were measured by immunoassays in serum/ plasma samples collected at inclusion; clinical information was recorded at inclusion and during follow-up.
Results: Cases had more often prior history of ischemic disease (72 vs 44%) and AMI (64 vs 41%), lower eGFR (54 vs 64 mL/min) and higher NYHA class (50 vs 22% in class III), LA (27 vs 23 mm/m2) and NT-proBNP (2873 vs 966 ng/L) than Controls (all variables, p<0.05). sST2 was higher in Cases (0.232 vs 0.121 ng/L, p=0.002) and unrelated to eGFR. ROC analysis of sST2 and NT-proBNP showed optimal cut-points to predict SCD of 0.152 and 2005 ng/L, respectively. sST2 values > 0.152 ng/L were found in 78% of Cases and 44% of Controls (p<0.001). In multivariable analysis sST2, NT-proBNP, eGFR and LA were independent predictors of SCD. When sST2 and NT-proBNP concentrations (dichotomized by the set cut-points) were analyzed combined as a single variable the OR to predict SCD raised to 21.857 (95% CI: 2.655–179.909, p=0.004). eGFR and indexed LA size were weaker predictors of SCD (OR: 1.665 and 1.157, respectively). The area under ROC of the combined variable to predict SCD was 0.869 (95% CI: 0.798 – 0.940).
Conclusions: sST2 is a novel biomarker for SCD prediction in stable HF patients. A multimarker approach using sST2 and NT-proBNP improves the identification of HF patients at high risk of SCD.