Giving steroids to head-injury patients not only fails to improve their chances of survival but actually appears to increase their risk of dying, a large international study has found.

Based on the findings, the researchers who conducted the study and other experts urged doctors against routinely using steroids for victims of head trauma.

"There are guidelines that say people shouldn't use steroids, but in practice a lot of people use steroids," said Ian Roberts of the London School of Hygiene & Tropical Medicine, who led the study being published tomorrow in the Lancet. "This is a widely used treatment of uncertain effectiveness, and a large trail has shown it to be ineffective and possibly harmful."

About 2 million Americans suffer head injuries each year, and about 52,000 die. Worldwide, about 3 million people die each year from head trauma.

For years, doctors routinely administered steroids to patients with moderate to severe head trauma to reduce swelling. But a number of studies have found that steroids have no benefit, leading experts to recommend against their routine use. Despite that, many trauma centers continue to use steroids.

To try to settle the issue, Roberts and his colleagues launched a study involving more than 10,000 head-injury patients at 239 hospitals in 49 countries.

Of the patients who received steroids within 48 hours of being hospitalized, 21 percent died, compared with 18 percent of those who received a placebo, the researchers reported.

The results were so striking that an independent panel monitoring the study stopped it early in May so the findings could be made public quickly.

"We had hoped for a benefit," Roberts said. "But there was no benefit, and in fact we found that it apparently causes harm."

It remains unclear how steroids cause harm to head trauma patients, but it could be that inflammation is actually helpful, Roberts said.

While many U.S. trauma experts say they have abandoned the routine use of steroids for such patients, a survey of trauma centers conducted in 2000 found that at least one-third still use them, according to Jamshid Ghajar, president of the Brain Trauma Foundation, who conducted the survey.

"It is still very prevalent," Ghajar said. "In spite of having guidelines that recommend against them, many trauma centers continue to use them."

Ghajar said he hopes the new study, the largest ever to examine the question, will persuade more doctors to discontinue the practice.

"This one definitely puts the nail in the coffin. It says definitively that it does increase mortality, so you shouldn't be using it," Ghajar said.