Current issues of ACP Journal Club are published in Annals of Internal Medicine


Review: Insufficient evidence exists for the effectiveness of 5 commonly used interventions for severe head injury

ACP J Club. 1999 May-June;130:66. doi:10.7326/ACPJC-1999-130-3-066

Source Citation

Roberts I, Schierhout G, Alderson P. Absence of evidence for the effectiveness of five interventions routinely used in the intensive care management of severe head injury: a systematic review. J Neurol Neurosurg Psychiatry. 1998 Nov;65:729-33.



What is the effectiveness of 5 interventions (hyperventilation, mannitol, cerebrospinal fluid [CSF] drainage, barbiturates, and corticosteroids) routinely used for intensive care of severe head injury?

Data sources

Published and unpublished randomized controlled trials available by August 1996 were identified by searching the Cochrane Controlled Trials Register, MEDLINE, and EMBASE/Excerpta Medica; hand searching 43 international journals, conference abstracts, and the proceedings of 16 neurosurgical meetings; reviewing the bibliographies of trials and review articles; and contacting the authors of all identified trials.

Study selection

Studies were selected if they compared 1 intervention of hyperventilation, mannitol, CSF drainage, barbiturates, or corticosteroids with no intervention; the method of allocation precluded previous knowledge of the next treatment (e.g., allocation by alternate or odd and even record numbers was not acceptable); and treatment groups were unconfounded (differed only for the treatment of interest).

Data extraction

Data were extracted on strategy for concealing allocation to groups, number of randomized participants, and outcomes. Main outcome was death; another outcome was combined death and disability.

Main results

No studies of CSF drainage were found. Treatment and control groups did not differ for death in 1 trial of 77 patients that compared hyperventilation with normoventilation (25% vs 34%, {relative risk reduction [RRR] 27%, 95% CI -49 to 64}*) or in 1 trial of 41 patients that compared mannitol with placebo (25% vs 14%, {relative risk increase [RRI] 75%, CI -52 to 538}*). Meta-analyses of 2 of 5 trials of barbiturates (n = 126, 58% vs 52%, {RRI 12%, CI -19 to 54}*) and 13 trials of corticosteroids (n = 2073, 35% vs 37%, {RRR 5%, CI -7 to 16}*) found no group differences for death. A combined end point of death or disability did not differ in 1 study of hyperventilation {RRI 14%, CI -18 to 58}*, 1 study of barbiturates {RRR 4%, CI -49 to 38}*, or a meta-analysis of 9 trials of corticosteroids {RRI 1%, CI -9 to 11}*.


Available evidence about the effectiveness of 5 intensive care interventions routinely used to treat severe head injury (hyperventilation, mannitol, cerebrospinal fluid drainage, barbiturates, and corticosteroids) is inconclusive. None of these interventions were shown to decrease mortality or combined mortality and disability.

Source of funding: NHS Research and Development Programme.

For correspondence: Dr. I. Roberts, Child Health Monitoring Unit, Department of Epidemiology and Public Health, Institute of Child Health, 30 Guilford Street, London WC1N 1EH, England, UK. FAX 44-171-242-2723.

*Numbers calculated from data in article.


The modern-day treatment of severe head injury has been used for approximately 40 years. Compared with other conditions with similar mortality rates (e.g., myocardial infarction), treatment methods and patient outcomes have not improved. With the advent of new imaging techniques, such as computed tomography and magnetic resonance imaging, the pathologic diagnosis of head injury has certainly improved. These techniques, however, have not contributed to improved outcomes in patients with head injuries.

The systematic review by Roberts and colleagues used methods consistent with the standard statistical process of meta-analysis. The authors used clear and consistent criteria for selecting studies. The finding that corticosteroids may confer a survival benefit of up to 6% (upper limit of 95% CI) is still in agreement with the 1996 U.S. Brain Trauma Foundation guidelines, which suggest that corticosteroids are not indicated in patients with severe head injuries (1). The wide CIs associated with the results of this meta-analysis confirm that all of the studies of individual treatments were underpowered—even combining them did not provide a definitive answer. A phase IIa trial, Corticosteroid Randomisation after Significant Head Injury (CRASH), will test the feasibility of conducting a worldwide trial (n = 20 000) to determine the effect of corticosteroids on death and disability in patients with severe head injuries (

Roberts and colleagues conclude that evidence on the treatment of head injury gives us no clear guidance on which treatments are beneficial and which are harmful. Until more randomized trials are done, the treatment of severe head injury must rely on fastidious critical care, with the goal of controlling such factors as hypoxia and shock that are harmful to patients with brain injuries.

Daniel L Herr, MD, MS
Washington Hospital CenterWashington, DC, USA


1. Guidelines for the management of severe head injury. Brain Trauma Foundation, American Association of Neurological Surgeons, Joint Section on Neurotrauma and Critical Care. J Neurotrauma. 1996;13:641-734.