Abstract 13316: Multi-Ancestry Genome-Wide Association Analysis in 1.9 Million Individuals Identifies 49 New Genetic Loci for Heart Failure and Related Subtypes

Abstract

Introduction: Characterization of the genetic etiology of heart failure (HF) provides an opportunitiy to advance understanding of causal human biology and to develop genomic tools for prediction and patient stratification. We conducted a multi-ancestry genome-wide association meta-analysis of HF and related subtypes in 1.9 million individuals. Methods: We investigated 10,199,961 common genetic variants for association with all-cause HF (145,795 cases); non-ischemic HF, defined as the absence of antecedent myocardial infarction and abnormal loading conditions (44,030 cases); and non-ischemic HF stratified by left ventricular ejection fraction (LVEF) (5,407 cases of LVEF <50% and 3,846 cases of LVEF 50%). By synthesizing evidence from functionally-informed fine-mapping, in silico gene prioritization, and multi-tissue transcriptome-wide association analysis, we prioritized likely causal variants, effector genes, and tissues for HF. We tested for enrichment of genes linked to Mendelian disorders within genomic loci for HF and assessed the genetic correlation of HF-related traits and their causal inter-relationships using Mendelian randomization. Results: We identified 56 independent genomic loci, including 39 novel, associated with all-cause HF; and additional 10 novel independent genomic loci among non-ischemic HF subtypes at P < 5 x 10-8. Genomic loci for HF were enriched for genes associated with Mendelian disorders of cardiomyopathy (Padjusted = 0.01) and arrhythmia (Padjusted = 0.0006). In 30 / 56 loci for all-cause HF, functionally-informed fine-mapping identified 5 or fewer likely causal variants. Genetically predicted expression levels of 266 genes across 49 human tissues were associated with HF at a false discovery rate <1%, 95 (36%) of which were also identified in heart tissue-specific analysis. Genetic correlation and causal associations were detected between upstream risk factors and traits of cardiac structure and function, yielding new insights into HF endophenotype relationships. Conclusion: These findings extend our understanding of the etiology of the HF syndrome and provide new opportunities to inform the development of new therapeutic strategies.