Many people living with the heart rhythm disorder known as atrial fibrillation may be taking unneeded blood thinners, a new study suggests.

These blood thinners, which include aspirin, Plavix and warfarin, are believed to reduce the risk of stroke that can come with atrial fibrillation. But for many atrial fibrillation patients with a low stroke risk, the medications might actually increase both bleeding and stroke risk, researchers reported.

The way most doctors decide whether a patient needs a blood thinner is by using a simple score called CHADS2, which assigns points to patients based on age and other medical risks. A score of 2 is usually needed to recommend a blood thinner, the researchers explained.

But, “people are realizing that the CHADS2 scores are putting too many people above the threshold — it’s pretty easy to get a 2,” explained study author Benjamin Horne, an adjunct assistant professor of biomedical informatics at Intermountain Medical Center Heart Institute in Utah.

For some patients with low CHADS2 scores, the risk of bleeding outweighs the risk of stroke, he added.

“It’s better than flipping a coin, but there are many other scores out there that are more predictive,” Horne said. “The problem with those scores is that it is difficult and time-consuming to use.”

The CHADS2 score breaks down this way: C stands for congestive heart failure, H for high blood pressure, A for age 75 or older, and D for diabetes. S stands for stroke, and the 2 gives an extra point for a previous stroke.

For the study, Horne and his colleagues collected data on nearly 57,000 patients with atrial fibrillation and a CHADS2 score of 0-2. Patients were divided into groups receiving aspirin, Plavix or warfarin or no blood thinner.

At three and five years, the rates of stroke, mini-stroke and major bleeding were higher with any blood thinner, compared with no treatment, the researchers found. The rates of these outcomes were lower among patients taking warfarin than among those taking aspirin or Plavix, the study authors added.

The findings were to be presented Friday at the American College of Cardiology’s annual meeting, in Washington, D.C. Research presented at meetings is considered preliminary until published in a peer-reviewed journal.

Horne said that Intermountain has developed a risk score using a blood test that can help doctors make a more precise decision about a patient’s risk for stroke. When used along with the CHADS2 score, it might prevent low-risk patients from being put on a blood thinner, he said.

But one heart rhythm expert was less certain.

“We have to take this study with caution,” said Dr. Apoor Patel, director of complex ablations in the department of electrophysiology at Northwell Health’s Sandra Atlas Bass Heart Hospital in Manhasset, N.Y.

It’s controversial whether patients with low CHADS2 scores should take blood thinners, Patel said. “It’s something we struggle with every day in clinical practice,” he said.

Stroke risk varies among patients, even those with a CHAD score of just 1, he said.

“I wouldn’t use this one study alone to change practice. When you have a patient with a low CHADS2 score, you have to make a decision about the pros and cons of anticoagulation [blood thinners], and you have to take into account risk factors that aren’t in the score,” Patel said.

Conditions not in the score that can make people more prone to stroke include kidney dysfunction, obesity, smoking and alcohol use, and many others, Patel said.

“When you are faced with a patient with a low CHADS2 score, you have to make a decision patient by patient,” he said. “You have to take into account not just a patient’s score, but a patient’s preferences, as well as risk factors not in the score.”