Background: Heart failure (HF) is a leading cause of avoidable, premature death and carries a 25% 1-year mortality. Any condition that damages the heart can cause HF: ~10% of people develop HF within 5 years after an acute coronary event. However, the rate at which HF develops varies considerably between people and is difficult to predict. Aim: We aimed to investigate whether large structural DNA variants (copy-number variants, CNVs), which can cause major disruption via deletion or duplication of genes, influence risk of HF in 3,934 people with acute coronary syndromes (e.g. unstable angina or a heart attack). Methods: Patients from the New Zealand Coronary Disease Cohort Study (CDCS, n=1,987) were genotyped with Affymetrix PMDA arrays (850K SNPs). Patients from the Multi-Ethnic New Zealand Study of Acute Coronary Syndromes (MENZACS, n=1,947) were genotyped with Illumina GSA arrays (650K SNPs). We used probe intensity data from these arrays to determine whether regions of DNA had been duplicated or deleted using PennCNV. Results: We will present results from a CNV genome-wide association study for risk of fatal or non-fatal HF over a median 5-years follow-up. To date, we have generated preliminary CNV calls for 1,848 CDCS patients (n=408 with HF) and detected >22,000 duplications and >15,000 deletions after quality filtering. Among the deletions, >1,200 were unique to patients who subsequently developed HF. These overlapped 151 genes, of which 5 are known HF-related genes, including ITGA10, which is highly expressed in the heart and associated with dilated cardiomyopathy. Conclusions: Our study will be the first CNV genome-wide association study for HF. This work may discover CNVs that help identify people with acute coronary syndromes who would benefit from more intensive treatment and management to reduce risk of progression to HF.