Genome-Wide Polygenic Scores Identify Risk for Common Diseases

Batya Swift Yasgur, MA, LSW

August 24, 2018

Genome-wide polygenic scores (GPSs) can identify individuals at high risk for coronary artery disease (CAD), atrial fibrillation (AF), type 2 diabetes, chronic inflammatory bowel disease (IBD), and breast cancer, new research shows.

Investigators used summary statistical data from several large genome-wide association studies (GWASs) to find polygenic risk scores that identified individuals at high risk for common diseases.

The GPS identified 8% of the population as being at greater than threefold increased risk for CAD, 6.1% for AF, 3.5% for type 2 diabetes, 3.2% for IBD, and 1.5% for breast cancer.

The risk for CAD had a 20-fold higher prevalence than the carrier frequency of rare monogenic mutations that confer similar risk. The GPS for CAD (GPSCAD) involved over 6 million polygenic variants.

"We envision polygenic risk scores as a way to identify people at high or low risk for disease, perhaps as early as birth, and then use that information to target interventions — either lifestyle modifications or treatments — to prevent disease," senior author Sekar Kathiresan, MD, director, Massachusetts General Hospital, Center for Economic Medicine, Boston, and the Broad Institute's Cardiovascular Disease Initiative, Cambridge, Massachusetts, said in a press release.

The study was published online August 13 in Nature Genetics.

Insufficient Risk Stratification

Genes with rare mutations have been identified in several diseases (eg, familial hypercholesterolemia [FH]) and confer much higher risk of those diseases in heterozygous carriers, the authors write.

"Although the ascertainment of monogenic mutations can be highly relevant for carriers and their families, the vast majority of disease occurs in those without such mutations," they note.

Polygenic inheritance plays a greater role in most common diseases than do rare monogenic mutations, but it has been "unclear whether it is possible to create a GPS to identify individuals at clinically significant risk," the authors note.

Previous attempts to create GPSs "had only limited success, providing insufficient risk stratification for clinical utility," the authors observe.

The researchers set out to use larger studies and improved algorithms to "revisit the question of whether a GPS can identify subgroups of the population with a risk approaching or exceeding that of a monogenic mutation."

They created several candidate GPSs, based on summary statistics and imputation from recent large GWASs, largely in individuals of European ancestry.

They derived 24 predictors based on a pruning and thresholding method and used 7 additional predictors in their algorithm.

Data about the genotypes and phenotypes of participants (n = 409,258) was drawn from the UK Biobank.

Intensive CAD Prevention

The researchers started with a validation dataset of the participants in the UK Biobank phase 1 genotype data release (n = 120,280) to select the GPSs with the best performance (defined as the maximum area under the receiver-operator curve [AUC]).

They then assessed the performance of an independent testing set comprising the participants in the UK Biobank phase 2 genotype data release (n = 288,978).

For each disease, "the discriminative capacity within the testing dataset was nearly identical to that observed in the validation dataset."

CAD was used an example of the method.

Polygenic predictors were derived from a GWAS (n = 184,305 participants) and evaluated based on their ability to detect the participants in the UK Biobank validation dataset diagnosed with CAD.

The predictors were found to have AUCs ranging from 0.79 to 0.81 in the validation set, with the best predictor (GPSCAD) involving 6,630,150 variants.

The predictor performed equally well in the testing dataset (AUC, 0.81).

The next step was to investigate whether GPSCAD could identify individuals with risk similar to the threefold increased risk conferred by an FH mutation.

They found that the median GPSCAD percentile score was 69 for individuals with CAD and 49 for individuals without CAD.

Eight percent of the population was found to have inherited a genetic predisposition that conferred at least a threefold increased risk for CAD.

"Strikingly," the polygenic score identified 20-fold more people at similar or greater risk than were found by familial FH mutations in previous studies, the authors report.

An advantage of GPSCAD is that it could be assessed from the time of birth, well before development of the risk factors typically used in clinical practice to predict CAD, the authors note.

"Making high GPSCAD individuals aware of their inherited susceptibility may facilitate intensive prevention efforts," they comment.

Previous research has demonstrated that a high polygenic risk for CAD may be offset by adherence to a healthy lifestyle or cholesterol-lowering therapy.

Other Diseases, Populations

Similar to the findings for CAD, the shape of the observed risk gradient for all the other diseases was consistent with predicted risk based only on the GPS.

In AF, the polygenic predictor identified 6.1% of the population at threefold or greater risk, with the top 1% having a 4.63-fold risk.

Screening for AF may have "maximal utility in those with high GPSAF," the authors suggest.

The polygenic predictor for type 2 diabetes identified 3.5% of the population at least threefold risk, with the top 1% having 3.30-fold risk, who might be candidates for more targeted preventive interventions.

The GPS identified 3.2% of the population at threefold or greater risk for IBD and the top 1% had 3.87-fold risk, potentially enabling "enrichment of a clinical trial population to assess a novel preventive therapy."

For breast cancer, the GPS identified 1.5% of the population at threefold or greater risk, with 0.1% of women having  fivefold or greater risk for breast cancer, corresponding to a breast cancer prevalence of 19.0% in this group vs 4.2% in the remaining 99.9% of the distribution.

"The role of screening mammograms for asymptomatic middle-aged women has remained controversial," but knowledge of GPSBC "may inform clinical decision-making about the appropriate age to recommend screening," the authors observe.

"These results show that, for a number of common diseases, polygenic risk scores can now identify a substantially larger fraction of the population than is found by rare monogenic mutations at comparable or greater disease risk," they conclude.

While the research was conducted primarily in individuals of European ancestry, they note that they "expect the relative impact of genetic risk strata to be generalizable across study populations."

Not Yet Ready for Clinical Use

Commenting on the study for | Medscape Cardiology, Thomas Quertermous, MD, William G. Irwin Professor of Medicine and director, Division of Cardiovascular Medicine (Research), Stanford University, California, who was not involved, with the study called it a "great, important paper by a great group at a good time."

Quertermous believes the study "will do a lot to get people to think seriously about genetic testing, not just for people with rare syndromes and Mendelian traits, but thinking about genotyping in larger populations to assess complete disease, which has not been on the table before."

He added that although the findings are "not yet ready for use as of today," there is likely to be "a debate that will probably last a decade or two as to how the information should be used and brought into the clinic."

He noted that it is still unclear whether this approach actually will affect outcomes, since those studies have not yet been completed.

Also commenting on the study for | Medscape Cardiology, John Mandrola, MD, clinical electrophysiologist, Baptist Medical Associates, Louisville, Kentucky, who was not involved with the study, said he is "optimistic" about the approach.

Although it is "still early and there aren't clinical outcomes yet — we need to do these studies — these findings are still important," he said.

"You can find out this risk when you're younger without waiting until your forties to get a coronary calcium score, so interventions might be implemented sooner," he said. 

Kathiresan and his team are considering potential partnerships to commercialize GPS tests in the coming year, according to the press release.

A web-based service might provide risk scores and interpretation to individuals who already have direct-to-consumer genotyping profiles.

"If you don't have the genotypes already, it would have to be offered," and "we're in discussions with several companies as to how to get this available to the general population," Kathiresan said.

The analysis was supported by the Harvard Catalyst (funded by the National Institutes of Health), the National Lipid Association, the National Heart, Lung, and Blood Institute of the US National Institutes of Health, the National Human Genome Research Institute of the US National Institutes of Health, the Doris Duke Charitable Foundation, the Foundation Leducq, and the Ofer and Shelly Nemirovsky Research Scholar Award from Massachusetts General Hospital. Kathiresan is listed as a co-inventor on a patent application for the use of genetic risk scores to determine risk and guide therapy and is also supported by a grant from Bayer AG to the Broad Institute focused on the genetics and therapeutics of myocardial infarction and atrial fibrillation. The other authors' disclosures are listed on the original paper. Quertermous and Mandrola have disclosed no relevant financial relationships.

Nat Genet. Published online August 13, 2018.  Abstract

For more Medscape Cardiology news, join us on Facebook and Twitter


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: