NEW YORK – A pair of new studies suggests that polygenic risk scores (PRS) for coronary artery disease (CAD) such as myocardial infarction might offer some predictive improvement when added to conventional clinical evaluation approaches, though genetic scores alone do not appear to have the predictive wherewithal to replace existing strategies.
For the first of the studies, published in the Journal of the American Medical Association on Tuesday, a team from the UK, Greece, and the Netherlands first established, tested, and tweaked CAD PRS based on risk SNPs identified in prior genome-wide association studies and on genetic data for nearly 16,000 CAD cases and about as many unaffected, age- and sex-matched controls.
When they tested that PRS — alone or in combination with an established clinical risk score called the "pooled cohort equation" — in 352,660 UK Biobank participants between the ages of 40 and 69 years, they found that it did improve their ability to predict CAD when it was used in addition to the existing clinical approach.
On the other hand, the PRS alone did not perform as well as the combined PRS and pooled cohort equation approach or the pooled cohort equation on its own. The team also estimated that it would get a "statistically significant, yet modest, improvement in the predictive accuracy for incident CAD and improved risk stratification for only a small proportion of individuals," by combining PRS and the existing clinical prediction approach.
From these and other findings, Imperial College London epidemiology and biostatistics researcher Ioanna Tzoulaki and her co-authors suggested that the "use of genetic information over the pooled cohort equations model warrants further investigation before clinical implementation."
Based on the existing pooled cohort equation, for example, the American College of Cardiology and American Heart Association recommends lipid-lowering treatments for those with a 7.5 percent or higher estimated risk of developing atherosclerotic cardiovascular disease over the course of a decade. At that risk level, the researchers estimated that the PRS plus pooled cohort equation could potentially reclassify some 4.4 percent of CAD cases and less than one percent of controls.
"The number of people meaningfully changing risk category and, therefore, receiving different treatment strategies based on genetic information is relatively small, with improvements mainly seen among cases reclassified to higher risk by addition of polygenic risk score to pooled cohort equations," the authors wrote, noting that the "relative benefit of [correct versus incorrect] reclassifications in cases and non-cases needs to take into account the risk-benefit profile of statins in a decision analysis and subsequent economic evaluation."
For their own analysis in JAMA, investigators from Vanderbilt University and elsewhere considered coronary heart disease events in more than 4,800 individuals between the ages of 45 and 79 years from the "Atherosclerosis Risk in Communities" (ARIC) study and almost 2,400 adult participants in the "Multi-Ethnic Study of Atherosclerosis" (MESA) study.
That retrospective analysis compared the coronary heart disease (CHD) prediction accuracy of the pooled cohort equation with that of a polygenic risk score based on weighted allele data spanning more than 6.6 million SNPs.
"The clinical utility of new risk markers such as the [PRS] depends on the ability to predict future CHD events, not on the strength of the associations with prevalent CHD," first and corresponding author Jonathan Mosley, a researcher at Vanderbilt, and his colleagues wrote.
His team found that although the CHD PRS did correspond with individuals' risk of experiencing heart disease over a decade, it did not significantly increase the ability to predict such disease when added to the existing pooled cohort equations for individuals in the ARIC or the MESA study.
Likewise, the researchers did not see significant reclassification when they incorporated the PRS into the type of 10-year risk prediction model used for prescribing statins, again assuming a threshold of 7.5 percent.
"Neither the proportions of individuals categorized as high risk or low risk, nor the observed event rates in each group, were substantially altered by the polygenic risk score," they reported, "suggesting that implementation may have limited effect at the population level."
In a related JAMA editorial, cardiology, preventive medicine, and public health researchers from Northwestern University and Loyola University Chicago noted that "available data do not support the clinical utility of CAD polygenic risk scores (in their current form) in middle-aged adults of European descent."
"In the meantime," corresponding author Sadiya Khan, a cardiology researcher at Northwestern, and her editorial co-authors suggested, "the best approach for prevention of CAD continues to be a combination of population-wide risk factor approaches for the entire population and addition of drug therapies and lifestyle interventions according to guidelines developed by the American Heart Association and American College of Cardiology."