Abstract 16341: Application of Genomic Markers to Predict Outcome in Patients Receiving Left Ventricular Assist Device (LVAD) Therapy

Abstract

Introduction: LVADs are now used as a bridge to heart transplantation, or destination therapy in patients with advanced heart failure. LVAD therapy reverses left ventricular (LV) remodeling and improves survival in 50% of patients. Hypothesis: We hypothesize that patients with poor prognosis after LVAD therapy will have a different genomic profile as compared to patients with higher survival rate. We propose to use genome-wide expression analysis to determine the mRNA and microRNA profiling of LV tissue and peripheral blood and identify probes that predicts outcome in patients receiving LVAD therapy. Methods: Genome-wide RNA expression profiling of the LV apex core tissue, and blood samples of patients receiving LVAD were performed using Illumina HumanHT-12 v4 Expression BeadChip, to study the expression of 47000 mRNA and micro-RNA transcripts. Follow up echocardiographic parameters were used to determine the patients’ clinical response to LVAD therapy. We proposed that ≥10% reduction in LV end-diastolic diameter (LVEDD) at 3 months is considered a responder to LVAD therapy. The results of the gene expression analysis were integrated with clinical outcomes to identify RNA expression patterns that differentiate LVAD responders from non-responders. Results: The percent change in LVEDD at 3 months ranged from +12 to -44 (mean -12±0.21), and 45% of the patients were responders to LVAD therapy (n=35). We used GeneSpring GX12 to identify the differentially expressed transcripts, and filtered the probes by their correlation to % change in LVEDD as a continuous numerical measurement with correlation stringency of 0.85 <= r < =1.0. A total of 277 probes from LV samples were highly correlated with % change in LVEDD. The candidate mRNA, ncRNA, and microRNA transcripts (e.g. SOD1, 2, GPX1-4,TGF-β1, EGFR, AMPK, SRC) were categorized in 3 functional groups of oxidative stress, tissue fibrosis, and cellular energy depletion pathways using pathway studio. Profiling of blood samples did not reveal highly correlated probes. Conclusions: The myocardial transcriptome can predict the outcome in patients receiving LVAD therapy. The genetic markers, predictive of LV recovery after mechanical unloading, can help build new hypotheses for candidate pathways of LV remodeling.