Abstract
Earlier intervention may delay or prevent heart failure (HF), a widespread health problem. However, it is not currently possible to identify those who are most at risk, especially before the appearance of any clinical signs. This study presents the development and subsequent validation of a novel genetic test for predicting the risk of HF, utilizing data from three independent cohorts of Australian and US subjects. We developed a first-phase test using the Baker Biobank case-control cohort, identifying 41 genetic variants indicative of HF risk through genome-wide interaction and association analyses. Subsequently, a second-phase test was designed. This identified 29 additional single-nucleotide polymorphisms. The combination of these two tests resulted in an aggregate test with a high predictive accuracy, achieving an Area Under the Curve of 0.93 and a balanced accuracy of 0.89. High genetic risk subjects in the Baker Biobank cohort had an odds ratio of 533.2. The test's robustness was validated by applying it to data from the Busselton Health Study and the Atherosclerosis Risk in Communities cohorts, yielding, respectively, Areas Under the Curve of 0.83 and 0.72, a balanced accuracy of 0.76 and 0.67, and Odds Ratios of 12.3 and 4.6. These results highlight the critical role of genetic factors in the development of heart failure and demonstrate this test's potential as a significant tool for clinical HF risk prediction.