LONDON -- Doctors have been giving steroids to head-trauma patients for more than 30 years, but the first major study of the practice has shown they are useless and may even have killed thousands of people.
Experts said the findings, published this week in The Lancet medical journal, are ''a complete and alarming surprise for all."
Head trauma, usually from car crashes, violence or falls, is the leading cause of death and disability in children and young adults in the developed world, and globally second only to HIV in causing the death of people under 40. About 3 million people die each year from head trauma.
Steroids had been used for decades in head trauma patients because it was believed their anti-inflammatory effects could help bring down swelling in the brain.
There was never reliable evidence, however, that steroids helped in cases of severe head injury, and their use has been controversial.
In the mid-1990s, existing evidence indicated they were probably not effective. Many doctors stopped using them, but not all.
The latest study, involving hundreds of doctors in 49 countries treating more than 10,000 patients, was stopped halfway through recruitment as it became clear steroids did not help and could be harmful.
''Most clinicians expected the trial to confirm the benefits of steroids, while others suspected that the effectiveness of steroids would turn out to be small or nonexistent," said Dr. Stefan Sauerland and Dr. Marc Maegele, experts from the University of Witten-Herdecke and the University of Cologne in Germany.
''Instead of a treatment benefit or at least equality, there were actually 159 excess deaths in the steroid group," the experts wrote in an independent critique of the study. ''When extrapolating the results to the annual incidence of severe head injuries worldwide, it is frightening to calculate how many patients might have been harmed by steroids."
Sauerland, who was not connected with the research, estimated doctors killed about 10,000 head-trauma patients with steroids in the 1980s and earlier.
The study, coordinated at the London School of Hygiene and Tropical Medicine, involved 10,008 adults with severe head injuries who were randomly allocated either a steroid drip or a fake drip for 48 hours after being admitted to the emergency room.
Within two weeks, 21 percent of the patients given steroids had died, compared with 18 percent of those given the fake drug. The results were the same regardless of how quickly the treatment was administered and regardless of the type or severity of the head injury.
Dr. Philip Stieg, chair of neurological surgery at Weill Cornell Medical Center in New York, said he was not sure the study settles the question.
''They picked one dose of steroids. Would a different dose be more effective?" asked Stieg, who was not connected with the research.
It also remains to be seen whether patients getting steroids fare better in the long term from the disabling effects of the trauma, he noted.
''If the long-term, one-year follow up data suggest that steroids improve morbidity, that the patient has a better neurological outcome, then physicians would reevaluate their use of steroids," he said.
Sauerland said, however, that it was unlikely any long-term benefit in disability reduction could outweigh the concern about the higher death rate early in treatment.
''This is a tombstone trial," Sauerland said by telephone. ''It will close the issue for ever."
Dr. Ian Roberts, coordinator of the trial, said many other treatments commonly used in trauma patients also were open to question. Approaches such as hyperventilation, barbiturates and fluid resuscitation have not been proven to help, he said.
''Trauma care has really been the poor relation of medical research," he said. ''Maybe there are other surprising and alarming results out there. We don't know."