Sudden cardiac death causes about half of the roughly 500,000 cardiovascular deaths that happen every year in the US. The vast majority of these deaths are due to coronary artery disease (CAD). Although there are treatments that can prevent mortality, many people never use them because they don't know there is a problem. But scientists have now created a better method for predicting CAD compared to current methods. When it was developed, the method integrated a large amount of information collected over a decade about study participants, including genetics, medical history, and lifestyle factors. The findings have been reported in Nature Medicine.
“I think more precise and personalized risk prediction could motivate patients to engage in early prevention,” said senior study author Ali Torkamani, Ph.D., a professor and director of Genomics and Genome Informatics at the Scripps Research Translational Institute. “Our model first predicts the risk that a person will develop CAD, and then it provides information to allow personalized intervention.”
Cholesterol and other stuff can build up on arterial walls to create plaque, and this plaque is the cause of coronary artery disease. The plaque blocks blood vessels, and most people don't know until it causes a heart attack. But this problem can be prevented.
CAD usually harms people who are in their mid-50s or older, but it can start to develop much sooner, sometimes even in teenagers. "There's a lot of room for us to take action,” noted first study author Shang-Fu ‘Shaun’ Chen, a former doctoral student in the Torkamani lab at Scripps.
In this work, the researchers aimed to create a machine learning model that could factor in a variety of characteristics, and at first, included 2,000 features. They supplied the model with data from the UK Biobank, and then tested its predictive value to determine if it identified the individuals who were at risk of coronary artery disease within ten years.
The effort whittled the risk factor list down to 53 data points, including some physical measurements, mental illness, sleep duration, certain genetic variants, and family medical history.
Compared to other risk factors, genetic predisposition was the strongest predictor of CAD risk by far; genetic predisposition of related conditions like high blood pressure, high cholesterol, and diabetes was significant.
“The higher your genetic risk for one of those traits: high cholesterol levels or high blood pressure levels or high diabetes risk, the greater benefit you get from intervening on that particular aspect through medication or lifestyle changes,” Torkamani noted.
The machine learning model performed better than the model now used in the clinic, and could predict twice as many CAD events. CAD arose in 62.9% of people in the highest risk group, and only 0.3% of people in the lowest risk group. It was also good at identifying those who were at risk, while avoiding inaccurate diagnoses for people who would not go on to develop CAD, noted the researchers.
When the validity of the model was tested with data from the National Institutes of Health's All of Us project, an effort that includes more diversity than the UK Biobank, it could also predict CAD just as well in those with European, African and Hispanic ancestries.
Now, a long-term clinical trial is planned to determine whether CAD can be prevented by warning people of their risk.
“We think the most important thing is for patients to be aware of their individual risks so that they can receive the appropriate treatments and make lifestyle changes,” said Chen.
Sources: Scripps Research Institute, Nature Medicine