Calculating risks for heart disease

Researchers are studying how to update and personalize the tools doctors use to predict a patient’s chance of developing cardiovascular disease

The annual physical exam is often a routine event for many people. A primary care doctor looks at, among many things, their patient’s blood pressure, cholesterol, and body weight, to help glean the health of their patient's organs and systems, including their heart and blood vessels. But if the patient has, say, elevated cholesterol or is between ages 40-75, the doctor can use other tools to further determine their risks for heart disease.

One is called a cardiovascular disease risk calculator. A common version in the U.S. is the pooled cohort equation (PCE). Despite its fancy name, it basically and rather quickly calculates a patient’s 10-year likelihood of developing atherosclerosis, plaque buildup in the arteries that can rupture or block flood flow. In severe cases, this can lead to a fatal heart attack or stroke.

The PCE helps guide physician-patient discussions and decisions about ways to offset these risks. Using nine factors – including age, gender, race, cholesterol, blood pressure, smoking history, and diabetes – it generates one of four risk scores: low, borderline, intermediate, or high. If a patient has borderline or intermediate risks for developing heart disease, they may benefit from additional screenings. Patients with higher risks are often prescribed statins, a cholesterol-lowering treatment.

Yet, as useful as the PCE has been, it has limitations, some researchers say. Now, they are studying how to improve future models to allow for more personalized risk assessments.

Here are a few ideas:

Calculating social determinants of health

Multiple factors can influence cardiovascular risk – and they can start to have an effect as early as in childhood or adolescence, said Vasan S. Ramachandran, M.D., a principal investigator and director of the Framingham Heart Study and chief of preventive medicine and epidemiology at Boston University’s Schools of Medicine and Public Health. “Cardiovascular disease is a life course disease,” he explained.

But how do you input a lifetime of data into one calculation?

Population-based measures and existing research provide general insight. However, race, a social construct, can be integrated into these metrics. Current PCE equations, for example, exist for Black and white adults.

The challenge with integrating race into cardiovascular disease risk predictions is that it can skew scores for people with different skin color who may otherwise have identical risk factors and experiences, said Ramachandran, who has studied this phenomenon.

Finding a way to capture more nuanced variables that influence cardiovascular health could be a better starting point, he explained. These factors – like how easy or difficult it is for someone to get to a doctor’s appointment, pick up a prescription, or get enough sleep and exercise – could replace race and improve risk prediction, he said.

Certain PCE risk profiles tend to be race neutral – for example, an adult who has diabetes, high blood pressure, and a history of smoking will have a higher PCE score, regardless of race. But when race is part of the equation and drives treatment decisions, Ramachandran said, “you medicalize a community and don’t treat the root cause” of disease. As a result, physicians may overprescribe statins. Or they may miss patients who could benefit from secondary prevention.

This is one reason why Ramachandran encourages physicians to consider a patient’s unique medical history and experiences, as well as the factors that influence where they live, work and play – often called social determinants of health.

The American College of Cardiology and the American Heart Association reiterated this advice, along with other recommendations, through clinical guidelines for heart disease prevention released in 2019. They also encouraged physicians to consider risk-enhancing factors that aren’t captured in the PCE, such as having inflammatory conditions like rheumatoid arthritis or psoriasis; chronic kidney disease; and severe pregnancy complications, including preeclampsia.

Integrating an abundance of data into future risk calculator models may not happen at once, but it could be approached in stages, Ramachandran said. The information needs to be simple, but comprehensive enough to provide meaningful predictions for millions of Americans.

He added that when it comes to social determinants of health, researchers could consider using input from community-based socioeconomic reports, which provides information about factors like education, income, and having access to health care. They could also use local health data based on counties, zip codes, or taken from national geographical assessments.

Jerome L. Fleg, M.D., a medical officer within the Division of Cardiovascular Sciences at NHLBI, agrees with these concepts. Although, he cautions that finding a way to quantify multiple metrics into one risk score and demonstrating its clinical effectiveness is complex.

Ramachandran concurs. A goal to strive for, he adds, is to identify the best metrics to use so that race no longer serves as a proxy for variables that better predict cardiovascular health while creating adjustable framework. The models should be able to integrate research findings from the past, but be designed for the future.

Integrating a composite lipid score into screenings

Researchers in the Division of Intramural Research at NHLBI are also studying ways to enhance cardiovascular disease risk predictions.

They created a tool that’s called an estimated atherosclerotic cardiovascular disease risk score, eASCVD for short. This tool, said Anna Wolska, Ph.D., a staff scientist in the Lipoprotein Metabolism Laboratory, “can be paired with fasting lipid tests in the lab to quickly and automatically identify patients who could benefit from advanced cardiovascular disease risk screenings.”

After patients have their bloodwork taken, an eASCVD could be calculated at no additional cost by the laboratory by combining total cholesterol, HDL cholesterol, triglycerides, and age into a composite lipid score. The eASCVD was created to work with tools, like the PCE, so physicians can better assess a patient’s risk for developing atherosclerotic heart disease. Similar to the PCE, the eASCVD works best at identifying patients with higher or lower risks.

Adding a composite lipid score to lab work would provide the biggest benefits for physicians since it's designed to automate screening and treatment decisions, they explained in a study published in Clinical Chemistry. However, the researchers envision the eASCVD, which still needs to be tested in clinical settings, would also strengthen patient education by fostering discussions about lipids and cardiovascular health.

Considering coronary artery calcium scans and polygenic risks

Medical researchers are studying how additional assessments could further personalize cardiovascular risk assessments. One example is looking at how polygenic risk scores, which provide general insight into a person’s inherited risk for disease, may help with coronary artery disease predictions.

For example, genomic predictions for coronary heart disease could guide decisions about if a patient may benefit from having coronary artery calcium imaging, which is used to detect traces of or accumulating plaque in the arteries. Physicians often use coronary artery calcium scans to assist with treatment decisions, such as if a patient has borderline or medium-level atherosclerotic risks.

“Coronary calcium scores have great appeal because they directly assess disease in the major coronary arteries, which are responsible for acute coronary events, the largest subclass of atherosclerotic events,” said Fleg.

A coronary artery calcium score of zero means the patient has no detectable calcified plaque and often has a lower risk for heart disease. This could move them from having an intermediate to lower 10-year risk. In that case, primary prevention through heart-healthy living might be recommended. If that same patient had a coronary artery calcium score of 1-99, indicating visible plaque accumulation, they may benefit from starting statin therapy.

Through papers published in JACC: Cardiovascular Imaging, researchers describe how adults with a higher genetic predisposition for coronary artery disease and who have intermediate risks may also benefit from having coronary artery calcium scans earlier in life, such as at age 35 or 40. Conversely, an adult in that same age group with low genetic risks and no cardiovascular risk factors may never need coronary artery imaging.

Yet, while these concepts appear promising, physicians are finding that pairing genetic insight with atherosclerotic risk predictions hasn’t significantly changed their current recommendations for screening and assessments.

One exception has been using single genes to identify people with rare or uncommon conditions, like familial hypercholesterolemia. This inherited condition stems from genetic mutations that can prevent the body from removing extra cholesterol from the bloodstream. If undetected or untreated, the person develops extremely high cholesterol levels that could significantly increase their risk for developing atherosclerosis and for having a heart attack early in life.

However, for most people, general genomic insight – the output of polygenic risk scores – hasn’t outperformed or significantly enhanced traditional cardiovascular risk predictions, including the PCE.

“Improving risk prediction will be an ongoing effort,” Fleg explained. This effort requires input from clinicians and data scientists and will continue to evolve.

“The reality is that atherosclerotic disease risk prediction will never be perfect,” he added, “but will hopefully be improved to allow for better targeting of preventive therapies.”

To learn more about heart and vascular health, visit https://www.nhlbi.nih.gov/health-topics/education-and-awareness/heart-truth.