Intervention Review

You have free access to this content

Mannitol for acute traumatic brain injury

  1. Abel Wakai1,*,
  2. Aileen McCabe1,
  3. Ian Roberts2,
  4. Gillian Schierhout3

Editorial Group: Cochrane Injuries Group

Published Online: 5 AUG 2013

Assessed as up-to-date: 20 APR 2009

DOI: 10.1002/14651858.CD001049.pub5


How to Cite

Wakai A, McCabe A, Roberts I, Schierhout G. Mannitol for acute traumatic brain injury. Cochrane Database of Systematic Reviews 2013, Issue 8. Art. No.: CD001049. DOI: 10.1002/14651858.CD001049.pub5.

Author Information

  1. 1

    Division of Population Health Sciences (PHS), Royal College of Surgeons in Ireland, Emergency Care Research Unit (ECRU), Dublin 2, Ireland

  2. 2

    London School of Hygiene & Tropical Medicine, Cochrane Injuries Group, London, UK

  3. 3

    London School of Hygiene & Tropical Medicine, c/o Cochrane Injuries Group, London, UK

*Abel Wakai, Emergency Care Research Unit (ECRU), Division of Population Health Sciences (PHS), Royal College of Surgeons in Ireland, 123 St. Stephen's Green, Dublin 2, Ireland. awakai@rcsi.ie. abelwakai@gmail.com.

Publication History

  1. Publication Status: Edited (no change to conclusions)
  2. Published Online: 5 AUG 2013

SEARCH

 
Characteristics of included studies [ordered by study ID]
Sayre 1996

MethodsRandomised and double blind placebo controlled trial.


ParticipantsModerate and severe head injury (CGS 11 and less), above 18 years of age.


InterventionsMannitol received 5ml/kg of 20% mannitol through large-bore intravenous catheter over five minutes; placebo group were given5ml/kg of 0.9% saline solution.


OutcomesDeath was reported. Length of follow-up was two hours after hospital admission.


NotesTesting pre-hospital administration of mannitol.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskInformation not provided in published study.

Allocation concealment (selection bias)Low riskQuote: "The study solutions were prepared by the pharmacy department and were identified by a code number. An envelope showing the code number identified the contents of that IV bag. If the treating physicians at the receiving trauma center believed it was necessary to know whether the patient had received mannitol or placebo, the envelope could be opened".

Blinding (performance bias and detection bias)
All outcomes
Low riskComment: Blinding is implied but not explicitly stated.

Quote: "This was a prospective, randomized, double-blind, placebo-controlled clinical trial of the effect of mannitol on systolic BP in multiply injured patients with severe head injuries".

Quote: "The study solutions were prepared by the pharmacy department and were identified by a code number. An envelope showing the code number identified the contents of that IV bag. If the treating physicians at the receiving trauma center believed it was necessary to know whether the patient had received mannitol or placebo, the envelope could be opened".

Quote: "At the conclusion of the 2-hour monitoring period, the EM resident and the investigator recording the patient data were asked to provide their guesses as to whether mannitol or placebo had been administered to the patient. For those patients actually given mannitol, the resident correctly picked mannitol in 11 of 13 cases for which a guess was given; and the investigators correctly picked mannitol in 13 of 14 cases. For those patients actually given placebo, the resident correctly picked placebo in 5 of 10 cases for which a guess was given; and the investigators correctly picked placebo for 13 of 16 cases".

Schwartz 1984

MethodsRandomised single blind trial.


ParticipantsSevere head injury (GCS<8). Participants had to have already been resuscitated and had raised ICP for more than 15 minutes.


InterventionsThe mannitol group received 20% mannitol solution at an initial dose of 1g per kg. Additional increments were given to maintain the intracranial pressure at less than 20 torr, within a serum osmolarity of mannitol of 320mOs/L.

The pentobarbital group received pentobarbital as an IV bolus of up to 10mg/kg, followed by a continuous infusion at 0.5-3 mg/kg/hr, provided that CPP remained above 50 torr, Additional increments of pentobarbital were given to maintain the intracranial pressure at less than 20 torr.


OutcomesDeath reported. Follow-up to three months.


NotesFor subsequent episodes of raised ICP, cross-over between the groups was allowed.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskQuote: "The starting drug was determined by a random number generator."

Allocation concealment (selection bias)Low riskQuote: "The mechanism of randomization was the opening of a serially-numbered, sealed envelope taken from one of two packages of envelopes, one for patients who had had intracranial hematomas removed and one for those who developed raised intracranial pressure from brain injury alone."

Blinding (performance bias and detection bias)
All outcomes
Unclear risk.

Smith 1986

MethodsRandomised controlled trial; not reported as double-blind.


ParticipantsSevere head injury (GCS 8 or less).


InterventionsCompared two regimens of guiding the administration of mannitol, one according to neurological signs and physiological measurements, and the other using ICP monitoring. The protocol for the first group used neurological signs, GCS, ventilatory values and arterial blood gas to guide administration. In the second group, a 250ml bolus of mannitol (20%) was given if ICP >25mmHg, and subsequent boluses were given incrementally.


OutcomesDeaths and neurological disability were reported. Follow-up was one year.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo information provided in published study.

Allocation concealment (selection bias)Low riskQuote: "The patients were randomly assigned to one of two groups by means of sealed envelopes containing cards designating the assigned protocol: Group I (mannitol therapy based on ICP monitoring) or Group II (empirical mannitol therapy)".

Blinding (performance bias and detection bias)
All outcomes
Unclear riskNo information provided in published study.

Vialet 2003

MethodsRandomised single blind trial.


ParticipantsSevere head injury (GCS<8) who required intravenous infusions of an osmotic agent to treat episodes of intracranial hypertension resistant to standard therapy.


InterventionsCompared two osmotic agents, mannitol and hypertonic saline solution. The mannitol group received 20% mannitol solution. The hypertonic saline group received 7.5% hypertonic saline. The infused volume was the same for both solutions: 2 ml/kg body weight in 20 minutes.


OutcomesDeath and neurological disability reported. Follow-up to three months.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo information provided in published study.

Allocation concealment (selection bias)Unclear riskQuote: "After randomization, the patients received either 20% mannitol (1160 mOsm/kg/H2O) or 7.5% HSS (2400 mOsm/kg/H2O)".

Blinding (performance bias and detection bias)
All outcomes
High riskQuote: "Because 20% mannitol can crystallize at ambient temperature, injections could not be performed in a blinded manner."

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Battison 2005Cross-over design.

Cruz 2001Concerns about the integrity of the trial data. Manuscripts does not indicate where the patients were randomised and despite an investigation by the Cochrane Injuries Group this remains open to question.

Cruz 2002Concerns about the integrity of the trial data. Manuscripts does not indicate where the patients were randomised and despite an investigation by the Cochrane Injuries Group this remains open to question.

Cruz 2004Concerns about the integrity of the trial data. Manuscripts does not indicate where the patients were randomised and despite an investigation by the Cochrane Injuries Group this remains open to question.

Fortune 1995Controlled follow up study; no means to reduce bias in the selection of the control group was used.

Francony 2008Participants were not all head-injured patients.

Gaab 1989Randomised to a diuretic, not necessarily mannitol.

Harutjunyan 2005Participants were not all head-injured patients.

Ichai 2009Cross-over design.

Levin 1979Participants were not all head injured; had a cross-over design.

Midgely 1993Cross-over design.

Smedema 1993Participants were not all head-injured patients.

 
Comparison 1. ICP-directed mannitol ('treatment') compared to mannitol according to neurological signs ('control')

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Death1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

 2 Death or severe disability1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

 
Comparison 2. Mannitol versus pentobarbital

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Death1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

 
Comparison 3. Mannitol versus hypertonic saline

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Death1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

 
Comparison 4. Mannitol versus placebo

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Death1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only