Abstract 17397: Discovery and Validation of Novel Genes Predicting Cardiac Remodeling on β-blockade
Background: Gene arrays allow for genome wide transcriptional analysis. Gene array studies are limited by false discovery (FD) due to multiple hypothesis testing. Our goal was to identify novel genes which predict cardiac remodeling on beta-blocker therapy while controlling for FD.
Methods: Endomyocardial biopsies were obtained from a randomized multicenter trial involving 49 patients with idiopathic dilated cardiomyopathy. Patients were randomized to metoprolol, metoprolol + doxazosin, or carvedilol 1:1:1. At baseline, patients had endomyocardial biopsies done for gene expression analysis of 54,675 probe sets. Left ventricular ejection fraction (LVEF) was measured by radionuclide ventriculography at baseline, 3 months and 12 months. Patients were randomly assigned to subgroup 1 (discovery) or subgroup 2 (validation), stratifying for sex, treatment group, age, race and baseline LVEF. Bivariate analysis was conducted in SAS for baseline genes associated with changes in LVEF over time. Only genes found to be concordant in the discovery and validation samples were retained. A p value < 0.001 was considered significant.
Results: Baseline patient characteristics included: LVEF = 27 ± 11 % mean age 45 ± 13 years, percent male 73%, percent Caucasian 69%.
Conclusions: We have identified 11 novel genes that are associated with reverse remodeling due to beta-blocker therapy in idiopathic dilated cardiomyopathy patients. These genes are involved in cell differentiation, mitochondrial electron transport, ubiquitination, oxidative metabolism, tumor suppression and apoptosis. Randomizing patients into discovery and validation sets is a practical strategy for controlling FD in gene array studies. Molecular pathology from endomyocardial biopsies may identify poor-responders to pharmacotherapy.
- © 2010 by American Heart Association, Inc.