Abstract 344: Circulating MicroRNA Signature as a Predictive Biomarker for Postoperative Heart Failure

Abstract

Background: Advancements in cardiac surgical techniques have led to decreasing operative risk. However, postoperative heart failure (PoHF) continues to be a major risk factor for adverse cardiac events in 20-35% of patients after cardiac surgery, with a 10-fold increase in 30-day mortality. Prediction of PoHF is challenging, particularly in patients with preserved ventricular function. Circulating microRNAs (miRNAs) recently were identified to predict HF or AF after surgery, but their role in predicting PoHF is not known. This study aimed to find novel noninvasive circulating biomarkers along with clinical factors that can identify patients at risk of developing PoHF immediately after surgery. Methods: Patients undergoing CABG surgery with no previous history of HF, ventricular or supraventricular tachycardia were recruited, and preoperative blood assessed for circulating levels of protein biomarkers using ELISA. Differences in relative plasma levels of 13 miRNAs between the PoHF and no-PoHF groups were assessed by qPCR. Preoperative echocardiography was obtained. SAS was used for statistical analysis and ROC curve. Results: Out of 68 patients, 13 developed PoHF (19.1%, mean age 64.1±11.6y, 53.8% males), whereas 55 (mean age 68.3±12.4y) remained free of HF. Patients who developed PoHF had lower LVEF (51.4±13.7 vs 58.2±9.9, P<0.05) with no differences in prevalence of hypertension, diabetes, hyperlipidemia, obesity, previous myocardial infarction, stroke, COPD, sleep apnea, or use of cardiac medications. The correlation matrix of all 13 miRNAs was transformed in a principal component (PC), resulting in 3 main clusters with eigenvalue >1. PC cluster2 consisted of miR-23a, -23b, -25 and -26a2, principally involved in oxidatives stress, fibrosis and contractility, and had the strongest association (AUC=0.797; P<0.01) with PoHF. A model combining PC cluster2 with age and LVEF improved sensitivity and specificity of the model to identify patients at risk of PoHF (AUC=0.880; 95% CL, 0.761-0.991; P<0.001) Conclusion: Our study demonstrates that miR-23a, -23b, -25 and -26a2 may be useful predictors of PoHF. Circulating miRNA as biomarkers may have diagnostic potential to preoperatively, noninvasively identify patients at risk of developing PoHF.