(Stroke. 2000;31:2719.)
© 2000 American Heart Association, Inc.
Comments, Opinions, and Reviews |
From the Department of Neurology, University of Debrecen, Medical School, Debrecen, Hungary (D.B., I.F.); Department of Neurology, First University Hospital, West China University of Medical Sciences, Chengdu, Sichuan, China (M.L.); and Department of Internal Medicine, Federal University of Sao Paulo, Sao Paulo, Brazil (G.F.P.).
| Abstract |
|---|
|
|
|---|
Summary of ReviewWe tested whether there is any evidence from unconfounded randomized clinical trials that treatment with mannitol reduces short- and long-term case fatality and dependency if administered after ischemic stroke or cerebral parenchymal hemorrhage. Trials were identified by the standard search strategy of the Cochrane Collaboration Stroke Review Group. A supplementary MEDLINE search was performed, and the Chinese Stroke Trials Register and the Latin-American databank LILACS were checked. A search was performed of masters and PhD degree theses in the databank of Sao Paulo University and in abstracts of medical congresses on neurology and neurosurgery during 19651997 in Brazil. Investigators were contacted for unpublished information. Only truly randomized unconfounded clinical trials were eligible for inclusion. Two of the reviewers independently extracted data from the trials. Data synthesis and analysis was performed with the use of the Cochrane Review Manager software (RevMan version 4.0.4).
ConclusionsOnly 1 trial fulfilled the inclusion criteria. The number of included patients was small, and the follow-up was short. Case fatality, the proportion of dependent patients, and side effects were not reported and were not available from the investigators. As a result of lack of appropriate randomized trials, currently no conclusion can be drawn on the effects of mannitol in acute stroke. The routine use of mannitol in all patients with acute stroke is not supported by evidence from randomized controlled clinical trials.
Key Words: mannitol randomized controlled trials stroke, acute treatment outcome
| Introduction |
|---|
|
|
|---|
Cytotoxic edema develops in the first hours after the onset of ischemic stroke and can be detected as a decrease in the apparent diffusion coefficient (ADC) of water. This reduction in ADC lasts for 3 to 4 days and then increases as vasogenic edema develops.8 9 A decrease in ADC was also shown in acute hemorrhagic stroke,10 but in contrast to ischemic strokes, ADC in hemorrhagic stroke remains decreased even 100 days after stroke onset. Continuous monitoring of intracranial pressure (ICP) shows that the pronounced brain edema that develops during days 4 to 14 of an intracerebral hemorrhage could lead to an increase in ICP, requiring treatment.11
Corticosteroids, although frequently used to treat cerebral edema, did not increase survival after stroke.12 Osmotic diuretics, especially mannitol, are among the agents that are widely used in the treatment of cerebral edema.13 14 15 16 Mannitol is thought to decrease ICP by decreasing overall water content and cerebrospinal fluid volume and by reducing blood volume due to vasoconstriction.17 18 19 Mannitol may also improve cerebral perfusion by decreasing viscosity or by altering red blood cell rheology.20 As a free radical scavenger, mannitol may exert a protective effect against biochemical injury.21
Mannitol is reported to decrease cerebral edema, infarct size, and neurological deficit in several experimental models of ischemic stroke,22 23 24 25 26 although primarily when administered within 6 hours after stroke onset. The results of some animal studies do not prove the beneficial effects of mannitol in rodent27 or monkey28 models of ischemic stroke. The worsening of cerebral edema by multiple-dose mannitol was also reported in cats,29 and a harmful effect of a single very large dose was assumed although not found in complete middle cerebral artery infarcts in humans.30
Mannitol has been used in human ischemic brain damage for >30 years.31 Cerebral edema in humans is regularly treated with mannitol, which is known to decrease ICP in several diseases32 and was found to decrease case fatality in cerebral edema due to hepatic failure.33 In a study of middle cerebral artery territory stroke, treatment modalities for elevated ICP, including osmotherapy, were initially effective, but ICP control could only be sustained in a minority of patients.34
The most common complications of mannitol therapy are fluid and electrolyte imbalances, cardiopulmonary edema, and rebound cerebral edema.18 35 Mannitol might cause kidney failure36 37 38 in therapeutic doses, and hypersensitivity reactions may also occur.39
Although there are several reports that could not prove the beneficial effects of mannitol in ischemic or hemorrhagic strokes in humans,40 41 42 43 the guidelines of the American Heart Association currently recommend the use of mannitol in certain clinical conditions in selected cases of acute stroke.44 Mannitol is widely used in acute stroke throughout the world. Almost 70% of physicians in China use mannitol or glycerol routinely in acute stroke,45 and mannitol is used routinely in acute stroke in several European countries as well. For example, mannitol is listed among recommended therapeutic interventions by the consensus statement of the Hungarian Stroke Society for cases when increased ICP is proven in stroke.46 Although treatment with osmotic diuretics seems logical from a physiological point of view, and mannitol has some effect on the brain in ischemic stroke,47 it is presently not clear whether the routine use of mannitol results in increased survival and decreased dependency in stroke patients.48
In this systematic review, we sought to test whether treatment with mannitol reduces short- and long-term case fatality and dependency after ischemic stroke or cerebral parenchymal hemorrhage. A more detailed review will be published and updated in the Cochrane Database of Systematic Reviews.49
| Methods |
|---|
|
|
|---|
We planned to evaluate the effects of mannitol in patients within 2 weeks of onset of ischemic stroke or cerebral parenchymal hemorrhage. Trials with intravenous, oral, or rectal treatment schedules were planned to be included in the review. The analysis was to be stratified by treatment route.
Outcome Measures
We planned to extract information on the following outcomes: (1)
deaths from all causes within the scheduled treatment period; (2)
deaths from all causes at the end of the scheduled follow-up period;
and (3) the number of patients dead or dependent in activities of daily
living. Patients dependent in daily living are those who are unable to
care for themselves and need daily help from others, which corresponds
to a score of 3 to 5 on the modified Rankin Scale.50
Since it is likely that a number of outcome scales will have been used, we planned to extract data using the reported results to estimate the number of patients dead or dependent in activities of daily living. If the same continuous outcome scales were used in the studies, these data would also be analyzed. Data for length of hospital stay and quality of life at final follow-up were to be analyzed if available.
In regard to safety parameters, we planned to determine the rate of kidney failure, pulmonary edema, and pulmonary embolism and the frequency of other fatal and nonfatal adverse events in each treatment group.
Search Strategy for Identification of Studies
Trials were identified by the standard search strategy of the
Cochrane Collaboration Stroke Review Group. In addition to this,
MEDLINE searches were performed with the search terms (stroke OR
cerebrovascular) AND (mannitol OR mannit) and (cerebral OR
brain) AND (mannitol OR mannit) to look for further possible
trials of mannitol therapy in acute stroke. One of the reviewers (M.L.)
searched for studies in the Chinese language in the Chinese Stroke
Trials Register. Another reviewer (G.F.P.) searched in Spanish and
Portuguese languages in the Latin-American databank
LILACS51 with the search term mannitol and
its variations in the Portuguese and Spanish languages using the
strategy suggested by Castro et al52 and performed a
search of masters and PhD degree theses in the databank of Sao Paulo
University (BIREME/PAHO-WHO [Biblioteca Regionale Medicina/Panamerican
Health Organization of the World Health Organization]) and in
abstracts of medical congresses on neurology and neurosurgery during
19651997 in Brazil. We attempted to obtain missing data from the
published trials by correspondence with the trialists.
Methods Used to Select Trials for Inclusion
At least 2 of the 4 reviewers independently judged each
identified trial to determine whether it should be included in the
review. For studies reported in non-English languages, one of the
reviewers had to know the language, and the other reviewer made his/her
decision on the basis of the "data extract form" used by the
Cochrane Stroke Review Group. For trials reported in languages not
familiar for the reviewers, consultations were arranged to obtain a
reliable translation. Disagreements were solved by discussion among
reviewers. Before data analysis, all reviewers agreed on each
trial to be included.
Assessment of Methodological Quality of Included Trials
We planned to evaluate the trials for methodological quality,
including method of randomization and blinding, concealment of
randomization, balance in baseline prognostic factors between treatment
and control groups, whether CT scanning was performed, patient
follow-up, and whether an intention-to-treat analysis was
performed.
Data Collection From Included Trials
After reaching an agreement on which trials to include, 2 of the
reviewers (D.B. and I.F.) independently extracted data from the trials
and performed the data analysis. Accuracy of data extraction
was checked by comparing the 2 results.
Methods Used to Synthesize Data
Data synthesis and analysis were performed with the use
of the Cochrane Review Manager software, RevMan version 4.0.4. The Peto
odds ratio was used for calculating relative treatment effects. We also
planned to report the absolute risk difference for each of the primary
outcome measures (death and dependency).
The primary analysis was planned to include all trials irrespective of whether the trials primarily intended to include patients only with ischemic stroke or only hemorrhagic stroke. Separate subgroup analyses on outcome in patients with confirmed ischemic stroke and in patients with hemorrhagic stroke confirmed by CT scan or cerebrospinal fluid examination were planned to be completed. A sensitivity analysis was to be performed to compare the overall results when all trials were included with the results when the analysis was restricted to trials in which patients had a CT scan.
Since mannitol might be more effective in patients with more
severe brain edema, we planned to perform a subgroup analysis
for those patients who had a decreased level of consciousness. A
further subgroup analysis was planned for studies with
intravenous administration of mannitol. Trials would be
analyzed separately if a dose of
1 g/kg of mannitol was given
because it was found that administration of
1.0 g/kg per dose
consistently reduced ICP from control values, but dosages <1
g/kg per dose did not always reduce ICP.13 A separate
analysis was planned for studies in which the total daily dose
was divided into
3 doses because the edema-reducing effect of
mannitol lasts for a few hours. An additional analysis was
planned for studies in which mannitol was given for >7 days because
edema around the ischemic or hemorrhagic region might persist
for longer periods of time.
| Results |
|---|
|
|
|---|
Methodological Quality of Included Studies
The method of randomization was not reported, and no information
about the concealment of randomization was available. The initial
number of patients randomized to the control and mannitol groups was
not reported, and no intention-to-treat analysis was performed.
Follow-up was not performed after 10 days. Case fatality and the
proportion of those who were dependent were not reported, and data
could not be obtained from the investigators.
Results of Studies
The reported outcome of the only included study was the change in
clinical condition, categorized as improved, unchanged, or worsened.
Fourteen of 41 controls (34%) and 12 of 36 mannitol-treated patients
(33%) improved, whereas the number of those whose condition worsened
was 18 of 41 (44%) and 16 of 36 (44%). With the use of the Peto odds
ratio method, neither beneficial nor harmful effects of mannitol could
be found. Case fatality, the proportion of dependent patients at the
end of the follow-up, and side effects were not reported and were not
available from the investigators. The planned outcome analyses
and sensitivity analyses could not be performed because of lack
of appropriate trials.
| Discussion |
|---|
|
|
|---|
Although the beneficial effect of mannitol in acute stroke cannot be excluded, and the use of mannitol in certain clinical conditions in selected cases of acute stroke might be appropriate, the routine use of mannitol in all patients with acute stroke cannot be recommended at this time. The clinical efficacy of mannitol in acute stroke has not yet been properly evaluated. On the basis of some animal experiments and clinical observations, mannitol has effects that might be beneficial in acute stroke. To prove this hypothesis, placebo-controlled, unconfounded, properly randomized clinical studies should be designed and performed. In these studies, the registration of early and late case fatality and of valid and reliable measures of dependency and disability is recommended. Brain CT should be part of the study protocol, long-term follow-up (eg, 3 and 6 months) is mandatory, and intention-to-treat analysis should be performed.
| Acknowledgments |
|---|
| Footnotes |
|---|
Received June 28, 2000; revision received August 14, 2000; accepted August 21, 2000.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
M. A. Koenig, M. Bryan, J. L. Lewin III, M. A. Mirski, R. G. Geocadin, and R. D. Stevens Reversal of transtentorial herniation with hypertonic saline Neurology, March 25, 2008; 70(13): 1023 - 1029. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. Bereczki, L. Mihalka, S. Szatmari, K. Fekete, D. Di Cesar, B. Fulesdi, L. Csiba, and I. Fekete Mannitol Use in Acute Stroke: Case Fatality at 30 Days and 1 Year Stroke, July 1, 2003; 34(7): 1730 - 1735. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Schwarz, D. Georgiadis, A. Aschoff, and S. Schwab Effects of Hypertonic (10%) Saline in Patients With Raised Intracranial Pressure After Stroke Stroke, January 1, 2002; 33(1): 136 - 140. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. Restrepo, I. E. Silverman, D. Bereczki, I. Fekete, M. Liu, and G. F. do Prado Osmotherapy: A Call to Arms Response Stroke, March 1, 2001; 32(3): 811 - 812. [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Stroke Home | Subscriptions | Archives | Feedback | Authors | Help | AHA Journals Home | Search Copyright © 2000 American Heart Association, Inc. All rights reserved. Unauthorized use prohibited. |