SANTA ANA PUEBLO, N.M. — A polygenic risk score that incorporates millions of variants may be able to identify young patients at high risk for MI, according to a keynote presentation at the American Society for Preventive Cardiology Congress on CVD.
Risk calculators like the American College of Cardiology/American Heart Association Pooled Cohort Equation are not useful in young patients because they are largely driven by age, Sekar Kathiresan, MD, director of the Center for Genomic Medicine at Massachusetts General Hospital, director of the Cardiovascular Disease Initiative at the Broad Institute and professor of medicine at Harvard Medical School, said in the presentation.
“In a population, that makes sense, but not in these young individuals in terms of appropriate risk assessment,” Kathiresan said.
An option to resolve this issue is to stratify patients based on inborn DNA variation to detect risk for early-onset disease.
“Most diseases have an inherent component, so we may be able to stratify at an early age —maybe even at birth — based on DNA,” Kathiresan said. “This is of course a major goal of genomic medicine.”
There are two major models that explain the genetic basis for MI at an early age: monogenic and polygenic. The monogenic model pertains to a single mutation, which is often sufficient to lead to early disease, whereas the polygenic model involves many variants, he said.
The traditional approach focuses on rare, monogenic mutations, according to the presentation. The main monogenic condition that leads to early-onset MI is familial hypercholesterolemia, which can increase the risk for MI by three- to fourfold, and is seen in 0.4% of the population, he said.
DNA variation or LDL measurements can be used to identify this monogenic risk group early in life, and then these patients can be treated by an intervention to lower LDL, he said.
In a recent study that’s currently under review, a team led by Amit V. Khera, MD, instructor of medicine at Massachusetts General Hospital and a Cardiology Today Next Gen Innovator, Harlan M. Krumholz, MD, SM, professor of medicine, investigative medicine and health policy, co-director of the Robert Wood Johnson Foundation Clinical Scholars Program at Yale University and director of the Center for Outcomes Research and Evaluation at Yale-New Haven Hospital and Kathiresan analyzed data from 2,081 patients with early-onset MI and 3,761 controls from the VIRGO trial and the MESA cohort. Whole genome sequencing was performed on all patients. Researchers found that in 100 patients with MI, 1.7% of patients had a familial hypercholesterolemia monogenic mutation, which contributed to a 3.8-fold increased risk for MI. The mean LDL for patients who carried the mutation was 206 mg/dL vs. 124 mg/dL in those who were not carriers.
“This is typically the way we find these patients because they show up with high LDL,” Kathiresan said. “We don’t often perform sequencing to find this population.”
Patients with early MI who were not identified by the monogenic model may be able to be identified through the polygenic model, where researchers take information from about 6.6 million gene variants to create a predictor.
“Each of those variants actually has a pretty small effect on disease risk, but we can add them all up into a score, a so-called polygenic risk score,” Kathiresan said.
Khera developed several of the genome-wide polygenic scores in a validation dataset of an estimated 125,000 patients from the U.K. Biobank. The best score was then evaluated in a testing dataset of an estimated 300,000 patients also from the U.K. Biobank.
“At the end of the day, we have a polygenic risk number for each person based on their genome, and the distribution of this number … is a beautiful bell curve,” Kathiresan said. “This is what you would see if you plotted LDL cholesterol in the population or HDL cholesterol, height, body mass index. What we have then is a quantitative metric of somebody’s liability to myocardial infarction that could be measured basically at birth.”
In the U.K. Biobank, the risk for CAD varied by more than 20-fold across polygenic score percentiles. The genome-wide polygenic score also had little correlation with MI risk factors that are currently measured in clinical practice.
For participants with highest 5% of polygenic scores, the risk for CAD was equivalent to those with familial hypercholesterolemia, Kathiresan said.
Researchers also tested this polygenic risk score in patients with MI from the VIRGO trial and controls from the MESA study. In this group, 17% of patients with MI had a high polygenic score.
“You have a group of people based on polygenic risk that are at very high risk and are not being captured right now in clinical practice and they’re often showing up essentially with an MI,” Kathiresan said.
Patients with high polygenic risk scores cannot always be distinguished by high LDL levels, according to the presentation.
“We’re not actually able to distinguish these individuals right now using traditional risk factors,” Kathiresan said. “You really need genotype, you really need genetic analysis.”
The polygenic risk score identified 10 times more patients compared with monogenic mutations, according to the presentation.
Polygenic risk for MI is modifiable through lifestyle and medications such as statins.
“Now it’s possible to score the polygenic component to any complex trait,” Kathiresan said. “You don’t need a whole genome sequence for this, by the way. You just need the standard single nucleotide polymorphisms you might get from 23 and Me, for example, to derive these scores.” – by Darlene Dobkowski
Kathiresan S. Keynote Address: Genetic Basis for Premature Heart Attack. Presented at: American Society for Preventive Cardiology Congress on CVD; July 27-29, 2018; Santa Ana Pueblo, New Mexico.
Disclosure: Kathiresan reports no relevant financial disclosures.