Let’s talk about controls – The example of heart failure

Poster Abstract: Maria Tsakiroglou, Clinical Lecturer in Pharmacology and Therapeutics, University of Liverpool 

Abstract

Aims: Heart failure (HF) represents the final stage of most cardiovascular diseases and is linked with metabolic, endocrine, inflammatory and kidney disorders. In genome-wide association studies (GWAS), the presence or absence of HF is frequently used to define cases and controls, respectively. While case identification often relies on clinically validated diagnostic criteria, control definition usually lacks equivalent scrutiny to confirm the absence of HF resulting in a potentially invalid comparator group.

Methods: Using the UK Biobank (UKB), we explored the presence of subclinical HF (“cryptic” cases) among putative controls and hypothesised that phenotypic refinement of controls, to more accurately exclude cryptic cases, could enhance genetic discovery. Based on the Universal Definition of HF and epidemiological evidence, we established eligibility criteria to minimize cryptic cases among controls; individuals meeting these criteria were designated super-controls. 

Conclusions: We found that up to one in four UKB participants classified as controls may have undiagnosed HF. Our framework identified 21,992 HF cases and 83,945 super-controls. Of super-controls with available prescription data, 17% had been prescribed at least one cardioprotective medication, compared with 43% of traditional controls. The Super-control GWAS identified 14 genome-wide significant loci (p < 5 x 10-8). Of these ten had been reported in prior meta-GWAS with sample sizes 20-23 times larger, while four were novel for HF but previously associated with risk factors. This proof-of-concept demonstrates that refining control definitions to create super-control groups significantly enhances genetic discovery in HF and can support genome-informed therapeutic strategies.