The study compared the use of stents, small tubes that prop open blood vessels with a common surgical procedure for cleaning out blockages in the carotid artery. Deaths and strokes were more than twice as common in patients treated with stents, the researchers found.

The findings conflict with influential earlier research, which found stents to be safer.

"The jury is still out," said Dr. Anthony Furlan, a vascular neurologist at the Cleveland Clinic.

Furlan wrote an editorial that accompanies the new study, published in Thursday's New England Journal of Medicine.

At issue is the best way to treat patients who develop blockages in the carotid, a key artery that supplies blood to the brain. The blockages are caused by plaque, a cheesy buildup of fat and cholesterol.

Many people with partial blockages do not experience symptoms. But the condition can cause a mini-stroke with temporary symptoms such as loss or blurring of vision in one eye, or weakness or numbness in an arm or leg. If left untreated, a blockage can lead to a major stroke or death.

In the standard treatment, doctors clamp off the artery and clean it out surgically. But that can be risky, especially for patients with heart damage or problems in the other carotid artery.

Doctors developed another treatment, in which they use a catheter to string a wire mesh stent into the artery that expands and props the artery open.

U.S. doctors have been doing carotid stenting since the mid-1990s. But the Food and Drug Administration has approved stenting only in patients who have symptoms from an artery that is blocked 70 percent or more and for whom surgery would be highly risky.

In the new study, French researchers recruited adult patients who had symptoms, a 60 percent or more blockage of the carotid artery, and an average surgery risk. Of the 520 participants, half got surgery and half got stents.

A month after undergoing the procedures, nearly 10 percent of the stented patients suffered death or stroke, versus only 4 percent of the surgical patients.

Here are the numbers: In the stent group, two people died and 23 others suffered a stroke. In the surgical group, three died and seven more suffered stroke.

In contrast, a 2004 U.S. study of patients with high surgery risk found the incidence of death, stroke or heart attack was about 5.5 percent in patients treated with a stent and 10 percent in patients treated with surgery. The difference was attributed to fewer heart attacks in the stent patients.

Some doctors criticized the new study. They noted that initially, the physicians who placed stents did not use an umbrella-like device that is placed beyond the blockage to catch any plaque that breaks off. The 2004 study used the devices.

U.S. standards require use of the device, said Dr. Christopher Cates, an Emory University cardiologist who has performed more than 600 carotid stent procedures.

The researchers strengthened safety guidelines after the study was under way to include the devices, but strokes and deaths had already occurred, Cates said. "That may have been the difference," he said.

The French study also allowed physicians to use a range of different stents, and doctors were not required to have worked with individual products more than twice before, Cates and others said.

A U.S. government-funded study now under way has enrolled 1,472 people. It requires a cerebral protection system and more physician training with the stent used.

"That's our best hope for a more definitive answer," Furlan said.