Skip to content

Polygenic Score Predicts Risk of Dilated Cardiomyopathy

Key findings

  • This study of cardiac MRI measurements from 36,041 participants was performed to evaluate the relationship between common genetic variants, left ventricular measurements on cardiac MRI and risk of dilated cardiomyopathy (DCM)
  • The researchers identified 45 previously unreported loci associated with left ventricular structure and function
  • There was substantial overlap between the novel genetic loci and known cardiomyopathy genes
  • A polygenic score reflecting left ventricular end-systolic volume robustly predicted incident DCM

Rare variants in dozens of genes have been associated with dilated cardiomyopathy (DCM). However, these variants yield a genetic diagnosis of DCM in only about 40% of cases. Also, many individuals with rare DCM-associated variants do not manifest clinical disease.

Researchers at Massachusetts General Hospital recently took a different approach to studying DCM: probing the link between common genetic variants and left ventricular measurements on cardiac MRI. Through this work, they were able to produce a polygenic score that robustly predicted DCM in the general population. James P. Pirruccello, MD, cardiologist, and Krishna G. Aragam, MD, preventive cardiologist, of the Corrigan Minehan Heart Center at Massachusetts General Hospital, and colleagues published the findings in Nature Communications.

Study Design

From the population-based UK Biobank, the researchers obtained cardiac MRI readings on 36,041 participants who did not have a diagnosis of congestive heart failure, coronary artery disease or DCM at the time of imaging. For these individuals, seven MRI-derived phenotypes were available:

  • Left ventricular end-diastolic volume (LVEDV)
  • Left ventricular end-systolic volume (LVESV)
  • Stroke volume (SV)
  • The same three traits indexed for body surface area (LVEDVi, LVESVi and SVi)
  • Left ventricular ejection fraction (LVEF)

Variants Associated with Cardiac Structure and Function

Based on a series of genome-wide association studies, the researchers identified common genetic variants associated with the seven cardiac MRI phenotypes:

  • 57 distinct loci were associated with at least one phenotype at a genome-wide significance threshold of P < 5 × 10−8
  • 45 of those loci had not been described in prior common variant analyses of cardiac imaging phenotypes

The novel genetic loci substantially overlapped with known cardiomyopathy genes.

Polygenic Scores

The researchers developed a polygenic score for each of the seven phenotypes, then used the scores to assess the 358,556 individuals in the UK Biobank who had no cardiac MRI data available. All but two of the scores significantly predicted incident DCM.

For the best-performing score, which comprised 28 single-nucleotide polymorphisms associated with LVESVi, the hazard ratio was 1.58 per standard deviation increase in the score (P = 6.4 × 10−18).

Carriers of Rare Variants

The researchers studied the LVESVi polygenic score in 59 carriers of TTN truncating mutations, a rare variant that's nevertheless identified in 15%–20% of DCM cases. Increases in the score were associated with changes in LVEDV, LVESV and LVEF.

These latter findings suggest the penetrance of high-impact rare variants may be influenced by carriers' polygenic backgrounds. For example, individuals with a DCM-linked rare variant but a favorable background of common genetic variants may be less likely to develop a reduced LVEF and therefore be protected from DCM.

Learn more about the Corrigan Minehan Heart Center

Refer a patient to the Corrigan Minehan Heart Center

Related topics


Researchers at Massachusetts General Hospital show that increased activin type II pathway receptor signaling contributes to heart failure and aging, and the findings may lead to new therapeutic targets in heart disease.


In this video, Matthew Nayor, MD, cardiologist in the Heart Failure and Transplantation Program in the Corrigan Minehan Heart Center, discusses his research on identifying and predicting heart failure using data from the Framingham Heart Study to understand protein measurements in blood.