Intended for healthcare professionals


Corticosteroids increase short term mortality in patients with head injury

BMJ 2004; 329 doi: (Published 14 October 2004) Cite this as: BMJ 2004;329:878
  1. Susan Mayor
  1. London

Corticosteroids, which have been used to treat head injuries for more than 30 years, increase the risk of death within the first two weeks after treatment, according to a study published last week (Lancet 2004;364: 1321-8.

The CRASH trial, a multicentre international collaboration funded by the UK Medical Research Council, was set up to establish the effects of corticosteroids in head injury after previous studies had shown inconclusive results. The study was designed to recruit 20 000 patients but was stopped in May 2004, when the data monitoring committee saw an increased risk of early death.

By that point, the trial included a total of 10 008 adults with head injury and a Glasgow coma score of 14 or less within eight hours of injury, recruited from 239 hospitals in 49 countries. They were randomly allocated to a 48 hour infusion of corticosteroids (methylprednisolone) or placebo.

Results showed that the relative risk of death from all causes within two weeks was 18% higher among patients randomised to corticosteroids compared with those given placebo. In all, 1052/4985 people died at two weeks in the corticosteroid group (21%) compared with 893/4979 among those randomised to placebo (18%) (relative risk 1.18, 95% confidence interval 1.09 to 1.27; P=0.0001). The relative increase in deaths due to corticosteroids was not influenced by injury severity (P=0.22) or time since injury (P=0.05).

Ian Roberts of the London School of Hygiene and Tropical Medicine, who is clinical coordinator of the trial, said, “Our results show there is no reduction in mortality with methylprednisolone in the two weeks after head injury.

“Head injury is devastating. Worldwide, millions of people, particularly young people, are killed or seriously disabled from their injuries each year. Obviously we'd have preferred to find out that corticosteroids improve patients' chances of surviving head injury, but our results are important because they'll improve patient care and protect future patients from increased risk of death from corticosteroids.” He said that the cause of the rise in risk of death within two weeks was unclear and raised the question of whether these drugs might also be harmful for patients with spinal injuries.

In an editorial accompanying the paper (Lancet 2004;364: 1291), Stefan Sauerland from the University of Cologne, Germany, said “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 non-existent. The results of CRASH are therefore a complete and alarming surprise for all.”

Dr Sauerland noted that inclusion of the result into a previous meta-analysis changed the absolute risk of death in the corticosteroid group from a 1% decrease to a 2% increase (Cochrane Collaboration handbook. In: Cochrane Library. Issue 1. Oxford: Update Software, 2001).

“In future, we should avoid trusting in underpowered clinical trials with surrogate rather than clinical endpoints,” he concluded.

Log in

Log in through your institution


* For online subscription