Scientists have developed a blood test that can predict whether someone is at high risk of a heart attack, stroke, heart failure, or dying from one of these conditions within the next four years, with roughly twice the accuracy of existing risk scores.
The test, which is already available in the US, could enable doctors to determine whether patients’ existing medications are working, or whether they need additional drugs to reduce their risk.
It could also hasten the development of new cardiovascular drugs by providing a faster means of assessing whether drug candidates are working during clinical trials.
Protein analysis can provide a more accurate snapshot of what someone’s organs, tissues and cells are doing at any given moment in time.
Stephen Williams at SomaLogic in Boulder, Colorado, and colleagues used machine learning to analyze 5,000 proteins in plasma samples from 22,849 people, and identify a protein signature that could predict the four-year likelihood of heart attack, stroke, heart failure or death.
When validated in 11,609 of the participants, they found that their model outperformed existing risk prediction tools, which use someone’s age, sex, race, medical history, cholesterol and blood pressure to characterize risk.
“Our ability to stratify risk across populations is more than twice as good as existing risks scores,” said Williams, whose results were published in Science Translational Medicine.
Importantly, the test could also accurately evaluate risk in people who have previously suffered a cardiovascular event and are taking drugs to reduce their risk, which existing risk assessments struggle to do.
The tool is already being used in four health systems within the US, and SomaLogic is in talks about the possibility of introducing it to the UK.