Gut Check is a periodic look at health claims made by studies, newsmakers, or conventional wisdom. We ask: Should you believe this?
The Claim: The most influential calculator of the risk of having a heart attack or stroke overestimates the likelihood of those events by 500 percent or more. More than 7,000 people access the online tool every day, according to the American College of Cardiology. And physicians use it routinely in their offices to identify people who should be prescribed cholesterol-lowering statins. That suggests many people have been needlessly prescribed the drugs.
Tell Me More:
The risk tool, released by the American Heart Association and the cardiology group in 2013, combines age, sex, smoking status, cholesterol level, blood pressure, and other factors to calculate a person’s risk of having a heart attack, stroke, or fatal coronary disease in the next 10 years. The idea is to guide therapy more precisely than cholesterol levels alone do; most people with a risk of 7.5 percent or more are advised to take a statin, while at 5 percent they’re told they and their doctors should consider doing so.
To see how well the predictions matched reality, researchers at Kaiser Permanente, the large health care provider in California, followed 307,591 nondiabetics aged 40 to 75, none diagnosed with atherosclerosis or prescribed statins. That matched the population the calculator is aimed at. Starting in 2008, the researchers compared the calculated risks to the actual incidence of heart attack, stroke, or death from blocked arteries.
article continues after advertisement
By 2013 there had been 2,061 such events; according to the risk calculator, there should have been 10,151. Among the nondiabetic patients whose calculated risk was 2.5 percent to 5 percent, the actual chance of suffering a cardiovascular event was roughly one-fourth that amount, the researchers reported in the current Journal of the American College of Cardiology. In the highest-risk group, with a predicted risk above 5 percent, actual risk was only 1.85 percent. Actual risks were closer to predictions for people with diabetes, whom the researchers analyzed separately.
The overestimates occurred for both sexes and all races and ethnicities.
The risk calculator was criticized as soon as it was released in 2013, and the skeptics said the new study confirms their suspicions. If physicians use the risk predictions to identify people who should be on a statin, as the cardiology groups recommend, then at least some of those prescribed the powerful drugs “won’t benefit from them,” said epidemiologist Nancy Cook of Brigham and Women’s Hospital in Boston, who was not involved in the new study.
The new research strikes at the heart of the risk calculator by saying its core algorithm errs on the high side. It is the largest study to find a severe overestimate of risk, which gives it credibility. And unlike some predecessors, the Kaiser study cannot be criticized for “missing” some strokes and heart attacks; it has complete records for everyone. “We had full follow-up on all of our patients,” said senior author Dr. Alan Go.
Another point in its favor is that it jibes with recent research. A 2015 study called MESA, of about 7,000 people, discovered that the risk calculator overestimated the odds of stroke or heart attack by 86 percent in men and 67 percent in women. Of people predicted to have a risk of 7.5 percent to 10 percent, only 3 percent of men or 5 percent of women actually had heart attacks or strokes.
The heart groups criticized MESA for including patients taking statins, arguing that the drugs were the reason that the incidence of heart attacks and strokes was below what the calculator spat out. The Kaiser study excluded people on statins, so that criticism doesn’t apply.
But a related one might. Excluding people who started taking statins during the study “substantially skewed their [final] population to one with the lowest risk of having a heart attack or stroke,” said Dr. Donald Lloyd-Jones of Northwestern University Feinberg School of Medicine, a member of the task force that developed the risk calculator. “Kaiser is renowned for identifying high-risk people and getting them on therapy.” He acknowledged, however, that the calculator is meant for people not already being treated with statins.
The new study reported data for only five years, while the calculator’s risks are for 10. But the chance of a heart attack or stroke rose by about the same amount annually, a linear trend that after 10 years would still have left the number of cardiovascular events well below what the calculator predicts.
It is not clear why the algorithm overestimates risks. One possibility is that the patients it is based on participated in 1990s-era research and might not be representative of today’s population. For one thing, they were older and whiter.
A more interesting explanation is that the inputs for the algorithm — blood pressure, cholesterol, age, and the others — have different consequences in 2016 than they did in the 1990s.
For instance, said Kaiser’s Go, blood pressure drugs (which more people are taking than in the 1990s) “might have beneficial effects apart from their effects on hypertension, so a blood pressure of 150/80 might give you a lower risk today than it did in the 1990s.” Yet the calculator assumes that 150/80, and hypertension generally, imparts the same risk of heart attack and stroke as it did decades ago.
Another likely flaw is that the calculator doesn’t take into account diet and exercise. “If people have a healthier lifestyle, the same ‘inputs’ might add up to less risk” than the calculator shows, said Kaiser cardiologist Dr. Jamal Rana, the study’s lead author. That is, being this age with this cholesterol level and this blood pressure adds up to a lower risk because you’re exercising, avoiding trans fats, and loading up on kale.
The new study is not definitive, but more and more evidence says Americans are being told they have a much greater chance of having a stroke or heart attack than they actually do, with millions unnecessarily being prescribed statins.