The MRC CRASH trial at 20: Time to reappraise corticosteroids for traumatic brain injury?

The final results of MRC CRASH were published 20 years ago and proved pivotal in the management of traumatic brain injury (TBI). The trial recruited 10,008 adults with a Glasgow Coma Score (GCS) of 14 or less within 8 h of injury from 49 countries. Patients were randomised to methylprednisolone or placebo with primary outcomes of death at 14 days and death or disability at 6 months. The trial stopped early for harm after demonstrating higher mortality in the intervention group (21.1% vs 17.9% at 14 days; 25.7% vs 22.3% at 6 months) although no difference in the composite outcome of death or disability at 6 months was found. ¹

Explanations for the unexpected mortality increase included undertreated hyperglycaemia and/or hypothalamo-pituitary axis suppression in the intervention group, neither of which were confirmed or refuted by the data available. ² Accepting the uncertainty from stopping a trial early (recruitment of 20,000 patients was planned), the authors argued the main limitation of their study was absence of judgement on the cause of death from participating clinicians. The trial became the basis for avoidance of steroids being the only level 1 recommendation (from over 40 recommendations made) in the current Brain Trauma Foundation Guidelines for management of severe TBI.

We believe the applicability of MRC CRASH to contemporary neurocritical care in the United Kingdom is limited and after two decades, reappraisal is warranted. From a 2025 perspective we query the lack of data on patients screened but not randomised, on concomitant treatments and on withdrawal of life sustaining treatment decisions. We are left to assume that the numbers not completing the intervention (at least 17% of those randomised) and not having CT scans (22%) are equally distributed between the groups. With such broad eligibility criteria (‘if . . . the treating doctor was substantially uncertain whether or not to treat with corticosteroid’), an exceptionally heterogenous group (GCS 14 with no or normal CT head, to GCS 3 with at least one unreactive pupil) and only 13% of patients recruited in the UK (prior to the advent of Major Trauma Centres) we contend it is hard to determine if these results apply to our patients today.

However, the most notable aspect is the intervention: the bolus of 2 g methylprednisolone is equivalent to 10 g of hydrocortisone (or 375 mg dexamethasone), and 0.4 g per hour (for 48 h) equivalent to a total of 96 g of hydrocortisone, doses that were high by the standards of contemporaneous trials and are astonishing now. Both the choice and dose of corticosteroid may be critical after severe TBI, with animal data favouring low dose hydrocortisone. ³ We highlight calls made at the time for continued considered use of steroid in TBI ⁴ and suggest that extrapolating from MRC CRASH to dismiss all steroids in all TBI is a step too far. Some patients, perhaps identified by simple biomarkers of the inflammatory response, ⁵ may have potential to benefit from low dose corticosteroid. MRC CRASH, though deservedly hugely influential, should not be the final chapter in the story.