Review Articles

Hypertonic saline versus mannitol for the treatment of elevated intracranial pressure: A meta-analysis of randomized clinical trials*
Hooman Kamel, MD; Babak B. Navi, MD; Kazuma Nakagawa, MD; J. Claude Hemphill III, MD, MAS; Nerissa U. Ko, MD
Objectives: Randomized trials have suggested that hypertonic saline solutions may be superior to mannitol for the treatment of elevated intracranial pressure, but their impact on clinical practice has been limited, partly by their small size. We therefore combined their ndings in a meta-analysis. Data Sources: We searched for relevant studies in MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials (CENTRAL), Scopus, and ISI Web of Knowledge. Study Selection: Randomized trials were included if they directly compared equiosmolar doses of hypertonic sodium solutions to mannitol for the treatment of elevated intracranial pressure in human subjects undergoing quantitative intracranial pressure measurement. Data Extraction: Two investigators independently reviewed potentially eligible trials and extracted data using a preformed data collection sheet. Disagreements were resolved by consensus or by a third investigator if needed. We collected data on patient demographics, type of intracranial pathology, baseline intracranial pressure, osms per treatment dose, quantitative change in intracranial pressure, and prespecied adverse events. Our primary outcome was the proportion of successfully treated episodes of elevated intracranial pressure. Data Synthesis: Five trials comprising 112 patients with 184 episodes of elevated intracranial pressure met our inclusion criteria. In random-effects models, the relative risk of intracranial pressure control was 1.16 (95% condence interval, 1.00 1.33), and the difference in mean intracranial pressure reduction was 2.0 mm Hg (95% condence interval, 1.6 to 5.7), with both favoring hypertonic saline over mannitol. A mild degree of heterogeneity was present among the included trials. There were no signicant adverse events reported. Conclusions: We found that hypertonic saline is more effective than mannitol for the treatment of elevated intracranial pressure. Our meta-analysis is limited by the small number and size of eligible trials, but our ndings suggest that hypertonic saline may be superior to the current standard of care and argue for a large, multicenter, randomized trial to denitively establish the rst-line medical therapy for intracranial hypertension. (Crit Care Med 2011; 39:554 559) KEY WORDS: intracranial hypertension; intracranial pressure; mannitol; hypertonic saline; meta-analysis; randomized controlled trial

annitol is recommended as the rst-line osmotic agent for the treatment of intracranial hypertension attributable to traumatic brain injury (TBI) (1), intracerebral hemorrhage (2), malig-

*See also p. 601. From the Department of Neurology, University of California, San Francisco, CA. Supported, in part, by the Department of Neurology, University of California, San Francisco, CA. The authors have not disclosed any potential conicts of interest. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journals web site (www.ccmjournal.com). For information regarding this article, E-mail: hooman.kamel@ucsf.edu Copyright 2011 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: 10.1097/CCM.0b013e318206b9be

nant cerebral infarction (3), subarachnoid hemorrhage (4), and acute liver failure (5). Numerous studies have shown that mannitol is effective in decreasing intracranial pressure (ICP) (6), and at least one randomized clinical trial has shown that mannitol reduces mortality compared to barbiturates in patients with elevated ICP from TBI (7). However, mannitol has many clinically important adverse effects, such as renal failure and hypovolemia (6, 8). These adverse effects of mannitol have led to increasing enthusiasm about the use of hypertonic saline formulations, which can reduce ICP without causing volume contraction and with less risk of nephrotoxicity (9 12). Several randomized clinical trials have suggested that sodium-based hypertonic solutions may be superior to mannitol in reducing ICP (1317), but the impact of these studies on clinical practice has

been limited, partly because of the different specic formulations used and partly because of the small size of these studies (18). Therefore, to better understand the relative efcacy of these two forms of osmotic therapy, we performed a meta-analysis of randomized clinical trials comparing hypertonic sodium solutions and mannitol for the treatment of elevated ICP. This report of our study conforms to the recommendations of the QUORUM statement on the quality of reporting for meta-analyses of randomized clinical trials (19).

MATERIALS AND METHODS

Inclusion Criteria. We included randomized clinical trials involving human subjects (without exclusions based on age, sex, or race) undergoing some method of quantitative ICP measurement (e.g., intraparenchymal monitor, external ventricular drain, lumbar drain, and others) and displaying evidence of ele-

554

Crit Care Med 2011 Vol. 39, No. 3

vated ICP, without regard for the underlying cause. Eligible trials directly compared the effect on ICP of equiosmolar doses (within 20%) of hypertonic sodium and mannitol solutions. We included trials using any formulation containing at least 3% sodium chloride or equivalent (e.g., hypertonic sodium plus lactate, starch, etc.). We included trials involving an unblinded or crossover design, as long as the treatment allocation was randomized. We excluded trials using varying doses or varying infusion times (e.g., titrating drug to effect), because this would have precluded standardized comparison of the two drugs. Data Sources. We performed an electronic search of MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials (CENTRAL), Scopus, and ISI Web of Knowledge for relevant studies published in any language between January 1967 and April 2010. We complemented this by using the Related Articles function on PubMed and searching the reference lists of relevant articles. To assess for possible publication bias, we searched for unpublished abstracts in Scopus and unreported studies in the http://clinicaltrials.gov registry, but we did not plan to include such studies in our analysis because of concerns about methodological limitations of unpublished trials (20). Search Strategy. We used text words and medical subject heading terms to identify studies involving the study population and intervention of interest (i.e., a comparison of mannitol and hypertonic saline in patients with elevated ICP; for full details of the search strategy, see Supplementary Data File 1 [Supplemental Digital Content 1, http://links.lww.com/CCM/A199]). To ensure the inclusion of studies not specically indexed to terms related to ICP, we included search terms for common causes of elevated ICP, including TBI, hemorrhage, and stroke. To limit our search to randomized clinical trials, we used the Cochrane highly sensitive search strategy for MEDLINE and a similar strategy for EMBASE (21). One investigator (H.K.) reviewed the titles and abstracts of all studies identied in this way and excluded those that were obviously irrelevant. The full articles of the remaining studies were retrieved and independently reviewed by two investigators (H.K. and B.N.) using a structured form to determine eligibility and extract data. Disagreements were resolved by consensus or by a third investigator (K.N.) if needed. We contacted study authors for clarications and further information as necessary. Quality Assessment. We formally evaluated the quality of eligible studies using the Cochrane Collaborations tool for assessing the risk of bias in randomized trials (22). Specically, studies were judged on the adequacy of the random sequence generation, allocation concealment, and blinding; the completeness of outcome data; the possibility of selective outcome reporting; and the existence of other potential sources of bias.

Data Extraction. For each trial, we recorded the number, age, and sex of patients; number of episodes of elevated ICP treated with a randomized allocation of a study drug; total osms per treatment dose; mean baseline ICP in each group; and the lowest mean ICP or the maximum mean change in ICP in each group within 60 mins of treatment. Our primary outcome was the proportion of cases in which randomized drug administration resulted in control of ICP, as dened in each study. We recorded the reported frequency of the following prespecied adverse events, as dened in each study: acute renal failure, pulmonary edema, hypotension, coagulopathy, and compartment syndrome from extravasation. We expected to include trials with crossover designs and multiple treatments per patient, so all data were extracted from intention-to-treat analyses only, with the unit of analysis being each randomized dose of study drug administered to treat a new episode of elevated ICP. In other words, crossover treatments for episodes of elevated ICP not responsive to the initial therapy were not included, but crossover treatments for new recurrent episodes of elevated ICP in the same patient were allowed as long as randomization was maintained. Statistical Analysis. In our primary analysis, we calculated the pooled relative risk of ICP control with hypertonic saline compared to mannitol. We expected heterogeneity among the trials from differences in hypertonic sodium formulations and subject populations; therefore, we used a random-effects model based on the method of DerSimonian and Laird. Because we expected a small number of eligible studies, and because randomeffects models may not perform as well with few studies, we also performed an analysis using a xed-effects model for comparison. In our secondary analysis, we used the same strategy to calculate the weighted difference between the two treatment groups in mean ICP reduction. If the mean change in ICP was not reported, we calculated it by subtracting the mean posttreatment ICP from the mean pretreatment ICP, and we then calculated the SD of the mean difference using the reported p value. If the change in ICP was not reported in terms of mean and SD, we used the median to represent the mean, because this has been shown to be reasonably accurate when the sample size is approximately 25 (23), and then calculated the SD from the p value of the difference in means within the study (24). To examine the effects of this statistical assumption regarding the mean and median, we performed sensitivity analyses by excluding such studies altogether or alternatively imputing the SD based on those of the other studies (25). We assessed for heterogeneity among studies using Cochrans Q statistic and Higgins I2 statistic; signicant heterogeneity was prespecied as p .10 or I2 50% (26). Because these statistical methods may be underpowered to detect heterogeneity, we explored pos-

sible sources of heterogeneity with prespecied subgroup and meta-regression analyses based on the following categories: type of hypertonic sodium formulation (containing starch/dextran or not), osms per dose ( or 300 mosm, corresponding to a standard 50-g dose of mannitol), baseline mean ICP ( or 25 mm Hg, because some studies used this as the treatment threshold instead of 20 mm Hg), type of intracranial pathology (TBI, subarachnoid hemorrhage, intracerebral hemorrhage, and stroke), and age (adult vs pediatric subjects).

RESULTS
Five studies enrolling a total of 112 adult patients met our eligibility criteria (13, 14, 17, 18, 27) (see Supplemental Fig. 1, which describes the ow diagram of search results and study selection [Supplemental Digital Content 2, http://links.lww.com/CCM/A200]). We did not nd any potentially eligible unpublished abstracts or unreported trials, and a funnel plot did not reveal visual evidence of publication bias. All studies were limited by the absence of blinding, but otherwise the overall methodological quality of the included studies was fairly high, especially because the main outcomes were quantied objective ICP readings. All the trials were small (the largest enrolled 40 patients), but several amplied their sample size by including multiple episodes of elevated ICP per patient (13, 14, 17). A et al (27) and Francony et al (18) compared mannitol to hypertonic sodium chloride, Battison et al (14) and Schwarz et al (13) used hypertonic sodium chloride plus starch, and Ichai et al (17) used hypertonic sodium lactate. Most studies included a mix of patients with TBI, stroke, intracerebral hemorrhage, and subarachnoid hemorrhage (13, 14, 18); one study exclusively consisted of patients with TBI (17) and one exclusively consisted of patients with recently resected supratentorial tumors (27) (Table 1). The average age of patients was 38 yrs ( 13 yrs), and 64% were men. None of the studies reported the occurrence of any of our prespecied adverse events. Among the 112 patients in the eligible trials, a total of 184 episodes of intracranial hypertension were randomly treated with mannitol or a hypertonic sodium solution. Mannitol was effective in controlling elevated ICP in 69 of 89 episodes (78%; 95% condence interval [CI], 67% 86%), whereas hypertonic sodium solutions were effective in 88 of 95 episodes (93%; 95% CI, 85%97%). The pooled relative risk of ICP control using
555

Crit Care Med 2011 Vol. 39, No. 3

Table 1. Study characteristics Study Study Characteristics Study year Country Subjects (n) Age (yr) Cases Battison et al (14) 2005 UK 9 Not available TBI (n 6) Subarachnoid hemorrhage (n 3) 20% 7.5% sodium chloride 6% dextran-70 249 mosm 250 mosm 24.0 (18.8, 25.9)a 22.0 (20.126.3)a 18 mm Hg

A et al (27) 2003 Egypt 40 36 (6) Tumor (n 40)

Francony et al (18) 2008 France 20 40 (14) TBI (n 17) Intracranial hemorrhage (n 2) Stroke (n 1) 20% 7.45% sodium chloride 255 mosm 255 mosm 31 (6) 27 (3) 20% below baseline 10 of 10 (100%) 9 of 10 (90%) 14 (8) 10 (5)

Ichai et al (17) 2009 France 34 36 (15) 34)

Schwarz et al (13) 1998 Germany 9 57 (11) Stroke (n 8) Intracranial hemorrhage (n 1) 20% 7.5% sodium chloride hydroxyethyl starch (60 g/L) 220 mosm 257 mosm 26.1 (1.5) 28.6 (4.8) 10% below baseline

TBI (n

Mannitol formulation Sodium formulation

20% 3.0% sodium chloride 5.49 mosm/kg 5.49 mosm/kg 31 (4) 31 (4) 20 mm Hg

20% Sodium lactate

Mannitol dose Sodium dose Baseline ICP (mm Hg) Mannitol Sodium Denition of ICP control ICP control Mannitol Sodium ICP decrease Mannitol Sodium

1.74 mosm/kg 1.65 mosm/kg N/A N/A Decrease 5 mm Hg or absolute 20 mm Hg 19 of 27 (70%) 28 of 31 (90%) 5 (2) 8 (2)

16 of 20 (80%) 19 of 20 (95%) 13 (5) 12 (5)

14 of 18 (78%) 16 of 18 (89%) 7.5 (5.811.8)a 13.0 (11.517.3)a

10 of 14 (71%) 16 of 16 (100%) 4.6 (4.7) 11.0 (7.1)

ICP, intracranial pressure; TBI, traumatic brain injury. Continuous measures are expressed as mean (SD), except ICP measurements from Battison et al,a which are expressed as median (interquartile range). ICP control denotes the proportion of episodes of elevated ICP successfully treated with a randomized dose of study medication. ICP decrease denotes the maximum decrease in mean ICP (mm Hg) within 60 mins of study drug administration.

(p .26, I2 24% in our primary analysis). We did not nd any eligible pediatric trials, and we did not have sufcient data to perform subgroup analysis based on the type of intracranial pathology. In our other prespecied subgroup analyses, there were no clear differences in effect size among subgroups based on the type of hypertonic saline formulation, dose per treatment, or baseline mean ICP. There were too few eligible studies to perform meta-regression.

DISCUSSION
Figure 1. Forest plot showing relative risk (RR) of successful control of elevated intracranial pressure. CI, condence interval.

hypertonic saline compared to mannitol was 1.16 (95% CI, 1.00 1.33, p .046; Fig. 1). The weighted mean difference in ICP reduction using hypertonic saline compared to mannitol was 2.0 mm Hg (95% CI, 1.6 to 5.7 mm Hg, p .276; Fig. 2). In comparison, when a xedeffects model was applied, our point estimates did not substantially change but
556

the condence intervals narrowed; the relative risk of ICP control favoring hypertonic saline was 1.20 (95% CI, 1.05 1.36, p .007), and the weighted mean difference in ICP reduction using hypertonic saline compared to mannitol was 2.0 mm (95% CI, 0.13.8 mm Hg, p .036). Our prespecied tests indicated a mild degree of heterogeneity among trials

In this meta-analysis of ve randomized clinical trials, we found that hypertonic sodium solutions are more effective than mannitol in controlling episodes of elevated ICP. We also found a trend toward greater quantitative ICP reduction in the hypertonic sodium group. Individually, two of the ve trials reported statistically greater quantitative ICP reduction using hypertonic sodium compared to mannitol, one showed a trend favoring hypertonic sodium, one showed equivalence, and one showed a statistically sigCrit Care Med 2011 Vol. 39, No. 3

ptosis (42). Overall, the conicting nature of these basic science considerations suggests that the preferred agent for the treatment of intracranial hypertension ultimately must be established through large, well-performed, randomized human trials.

CONCLUSIONS
The need for more rigorous clinical data prompted us to perform this metaanalysis of existing randomized trials. Our study is limited by the small number and size of eligible trials, as well as their methodological differences, lack of rigorous adverse event reporting, and failure to control for clustering among patients receiving multiple doses of study drug. These limitations preclude a rm recommendation to preferentially use hypertonic saline as rst-line therapy, but our ndings indicate that hypertonic sodium solutions are, on average, superior to mannitol for the treatment of elevated ICP and argue for a large multicenter study comparing these two therapies. Our results suggest an absolute effect size of approximately 15% in favor of hypertonic saline, and a trial would require approximately 100 patients in each arm to have 80% power to detect this difference in the rate of ICP control between the two groups. Ideally, such a trial would incorporate a uniform hypertonic sodium solution and, more importantly, also include assessment of long-term neurologic outcomes. While we await such a trial, the balance of evidence indicates that hypertonic saline is more effective than mannitol for the treatment of intracranial hypertension.

Figure 2. Forest plot showing difference in mean quantitative reduction of intracranial pressure (mm Hg). WMD, weighted mean difference; CI, condence interval.

nicant advantage to mannitol. None individually were sufciently powered to show a difference in the proportion of episodes of elevated ICP that were successfully treated, but the pooled results indicate that hypertonic sodium solutions are, on average, more effective. In selecting our primary outcome, we chose each studys clinical denition of ICP control, because this is a more clinically relevant end point than the mean quantitative decrease in ICP, which can mask signicant intragroup heterogeneity of treatment effect. Regardless, in our secondary analysis, there was a strong trend toward greater quantitative ICP reduction in patients treated with hypertonic saline compared to mannitol; this became statistically signicant when a xed-effects model was used. The lack of a uniform treatment effect among all the trials is not surprising; we expected heterogeneity from methodological differences among the trials, specically in terms of study population, treatment thresholds, and types of hypertonic sodium formulation. Our meta-analysis revealed only a mild degree of heterogeneity, and our prespecied subgroup analyses did not reveal an explanation for differences in effect size. Only one of the ve trials (by Francony et al [18]) failed to show a trend favoring hypertonic sodium formulations, and we noted no obvious methodological differences between this trial and the others that would explain this disparity. However, this study was the smallest of the ve included trials, and its discordant results may be attributable to the effect of chance. This explanation is supported by the results of seven observational studies
Crit Care Med 2011 Vol. 39, No. 3

demonstrating the effectiveness of hypertonic saline in lowering ICP after failure of standard mannitol therapy (28 34). Another observational study found no signicant difference between the efcacy of mannitol and hypertonic saline, but the latter had a longerlasting effect (35). Furthermore, two randomized trials that did not meet our inclusion criteria support the superiority of hypertonic saline therapy. A trial by Vialet et al (15) was not eligible because different osms per dose were used in the two treatment groups (361 osms for hypertonic saline vs. 175 osms for mannitol); nevertheless, they found that 7.5% sodium chloride was more effective in treating refractory intracranial hypertension. A trial by Harutjunyan et al (16) did not qualify because study drugs were titrated to effect, making a standardized comparison difcult; nevertheless, they found that 7.2% sodium chloride plus hydroxyethyl starch was superior to mannitol in reducing ICP. In addition to the clinical evidence reviewed, there are also numerous biochemical and physiologic considerations when choosing between hypertonic saline and mannitol for the treatment of elevated ICP. For example, hypertonic saline solutions result in signicant volume expansion, improved cardiac output, improved regional blood ow, improved cerebrospinal uid absorption (36), and benecial immunomodulation (11). In addition, hypertonic saline may be superior to mannitol in regard to brain oxygenation and cerebral hemodynamics (33). However, mannitol may improve microvascular cerebral blood ow (37), reduce blood viscosity (38), improve blood rheology (6, 39), reduce cerebrospinal uid production (40), promote freeradical scavenging (41), and inhibit apo-

REFERENCES
1. Brain Trauma Foundation, American Association of Neurological Surgeons, Congress of Neurological Surgeons, et al: Guidelines for the management of severe traumatic brain injury. II. Hyperosmolar therapy. J Neurotrauma 2007; 24:S14 S20 2. Broderick J, Connolly S, Feldmann E, et al: Guidelines for the management of spontaneous intracerebral hemorrhage in adults: 2007 update: A guideline from the American Heart Association/American Stroke Association Stroke Council, High Blood Pressure Research Council, and the Quality of Care and Outcomes in Research Interdisciplinary Working Group. Circulation 2007; 116: e391 e413 3. Adams HP Jr, del Zoppo G, Alberts MJ, et al: Guidelines for the early management of

557

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

adults with ischemic stroke: A guideline from the American Heart Association/ American Stroke Association Stroke Council, Clinical Cardiology Council, Cardiovascular Radiology and Intervention Council, and the Atherosclerotic Peripheral Vascular Disease and Quality of Care Outcomes in Research Interdisciplinary Working Groups: The American Academy of Neurology afrms the value of this guideline as an educational tool for neurologists. Stroke 2007; 38:16551711 Bederson JB, Connolly ES Jr, Batjer HH, et al: Guidelines for the management of aneurysmal subarachnoid hemorrhage: A statement for healthcare professionals from a special writing group of the Stroke Council, American Heart Association. Stroke 2009; 40:994 1025 Stravitz RT, Kramer AH, Davern T, et al: Intensive care of patients with acute liver failure: Recommendations of the U.S. Acute Liver Failure Study Group. Crit Care Med 2007; 35:2498 2508 Diringer MN, Zazulia AR: Osmotic therapy: Fact and ction. Neurocrit Care 2004; 1:219 233 Schwartz ML, Tator CH, Rowed DW, et al: The University of Toronto head injury treatment study: A prospective, randomized comparison of pentobarbital and mannitol. Can J Neurol Sci 1984; 11:434 440 Oken DE: Renal and extrarenal considerations in high-dose mannitol therapy. Ren Fail 1994; 16:147159 Bentsen G, Breivik H, Lundar T, et al: Predictable reduction of intracranial hypertension with hypertonic saline hydroxyethyl starch: A prospective clinical trial in critically ill patients with subarachnoid haemorrhage. Acta Anaesthesiol Scand 2004; 48: 1089 1095 Bentsen G, Breivik H, Lundar T, et al: Hypertonic saline (7.2%) in 6% hydroxyethyl starch reduces intracranial pressure and improves hemodynamics in a placebo-controlled study involving stable patients with subarachnoid hemorrhage. Crit Care Med 2006; 34:29122917 White H, Cook D, Venkatesh B: The use of hypertonic saline for treating intracranial hypertension after traumatic brain injury. Anesth Analg 2006; 102:1836 1846 Forsyth LL, Liu-DeRyke X, Parker D Jr, et al: Role of hypertonic saline for the management of intracranial hypertension after stroke and traumatic brain injury. Pharmacotherapy 2008; 28:469 484 Schwarz S, Schwab S, Bertram M, et al: Effects of hypertonic saline hydroxyethyl starch solution and mannitol in patients with increased intracranial pressure after stroke. Stroke 1998; 29:1550 1555 Battison C, Andrews PJ, Graham C, et al: Randomized, controlled trial on the effect of a 20% mannitol solution and a 7.5% saline/6% dextran solution on increased intra-

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

cranial pressure after brain injury. Crit Care Med 2005; 33:196 202; discussion 257298 Vialet R, Albanese J, Thomachot L, et al: ` Isovolume hypertonic solutes (sodium chloride or mannitol) in the treatment of refractory posttraumatic intracranial hypertension: 2 mL/kg 7.5% saline is more effective than 2 mL/kg 20% mannitol. Crit Care Med 2003; 31:16831687 Harutjunyan L, Holz C, Rieger A, et al: Efciency of 7.2% hypertonic saline hydroxyethyl starch 200/0.5 versus mannitol 15% in the treatment of increased intracranial pressure in neurosurgical patientsA randomized clinical trial [ISRCTN62699180]. Crit Care 2005; 9:R530 R540 Ichai C, Armando G, Orban JC, et al: Sodium lactate versus mannitol in the treatment of intracranial hypertensive episodes in severe traumatic brain-injured patients. Intensive Care Med 2009; 35:471 479 Francony G, Fauvage B, Falcon D, et al: Equimolar doses of mannitol and hypertonic saline in the treatment of increased intracranial pressure. Crit Care Med 2008; 36: 795 800 Moher D, Cook DJ, Eastwood S, et al: Improving the quality of reports of metaanalyses of randomised controlled trials: the QUOROM statement. Quality of Reporting of Meta-analyses. Lancet 1999; 354: 1896 1900 Cook DJ, Guyatt GH, Ryan G, et al: Should unpublished data be included in metaanalyses? Current convictions and controversies. JAMA 1993; 269:2749 2753 Lefebvre C, Manheimer E, Glanville J: Searching for studies. In: Cochrane Handbook for Systematic Reviews of Interventions, version 5.0.2. Higgins JPT, Greene S (Eds). Available at: http://www.cochranehandbook.org. Accessed April 26, 2010 Higgins J, Altman D: Assessing risk of bias in included studies. In: Cochrane Handbook for Systematic Reviews of Interventions, version 5.0.2. Higgins JPT, Greene S (Eds). Available at: http://www.cochrane-handbook.org. Accessed April 26, 2010 Hozo SP, Djulbegovic B, Hozo I: Estimating the mean and variance from the median, range, and the size of a sample. BMC Med Res Methodol 2005; 5:13 Higgins J, Deeks J: Selecting studies and collecting data. In: Cochrane Handbook for Systematic Reviews of Interventions, version 5.0.2. Higgins JPT, Greene S (Eds). Available at: http://www.cochrane-handbook.org. Accessed April 26, 2010 Furukawa TA, Barbui C, Cipriani A, et al: Imputing missing standard deviations in meta-analyses can provide accurate results. J Clin Epidemiol 2006; 59:710 Deeks J, Higgins J, Altman D: Analysing data and undertaking meta-analyses. In: Cochrane Handbook for Systematic Reviews of Interventions, version 5.0.2. Higgins JPT, Greene S (Eds). Available at: http://www.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

cochrane-handbook.org. Accessed April 26, 2010 A W, Abdallah I, Yosry M, et al: The effect of a single dose of 3% hypertonic saline versus mannitol 20% for treatment of brain oedema and intracranial hypertension in postoperative patients with supratentorial gliomata. Egyptian J Anaesthesia 2003; 19: 371379 Schatzmann C, Heissler HE, Konig K, et al: Treatment of elevated intracranial pressure by infusions of 10% saline in severely head injured patients. Acta Neurochir Suppl 1998; 71:3133 Suarez JI, Qureshi AI, Bhardwaj A, et al: Treatment of refractory intracranial hypertension with 23.4% saline. Crit Care Med 1998; 26:1118 1122 Horn P, Munch E, Vajkoczy P, et al: Hyper tonic saline solution for control of elevated intracranial pressure in patients with exhausted response to mannitol and barbiturates. Neurol Res 1999; 21:758 764 Schwarz S, Georgiadis D, Aschoff A, et al: Effects of hypertonic (10%) saline in patients with raised intracranial pressure after stroke. Stroke 2002; 33:136 140 Koenig MA, Bryan M, Lewin JL III, et al: Reversal of transtentorial herniation with hypertonic saline. Neurol 2008; 70:10231029 Oddo M, Levine JM, Frangos S, et al: Effect of mannitol and hypertonic saline on cerebral oxygenation in patients with severe traumatic brain injury and refractory intracranial hypertension. J Neurol Neurosurg Psychiatry 2009; 80:916 920 Kerwin AJ, Schinco MA, Tepas JJ III, et al: The use of 23.4% hypertonic saline for the management of elevated intracranial pressure in patients with severe traumatic brain injury: A pilot study. J Trauma 2009; 67: 277282 Ware ML, Nemani VM, Meeker M, et al: Effects of 23.4% sodium chloride solution in reducing intracranial pressure in patients with traumatic brain injury: A preliminary study. Neurosurg 2005; 57:727736 Paczynski RP: Osmotherapy. Basic concepts and controversies. Crit Care Clin 1997; 13: 105129 Muizelaar JP, Lutz HA III, Becker DP: Effect of mannitol on ICP and CBF and correlation with pressure autoregulation in severely head-injured patients. J Neurosurg 1984; 61: 700 706 Burke AM, Quest DO, Chien S, et al: The effects of mannitol on blood viscosity. J Neurosurg 1981; 55:550 553 Mendelow AD, Teasdale GM, Russell T, et al: Effect of mannitol on cerebral blood ow and cerebral perfusion pressure in human head injury. J Neurosurg 1985; 63:43 48 Donato T, Shapira Y, Artru A, et al: Effect of mannitol on cerebrospinal uid dynamics and brain tissue edema. Anesth Analg 1994; 78:58 66

558

Crit Care Med 2011 Vol. 39, No. 3

41. Alvarez B, Ferrer-Sueta G, Radi R: Slowing of peroxynitrite decomposition in the presence of mannitol and ethanol. Free Radic Biol Med 1998; 24:13311337 42. Korenkov AI, Pahnke J, Frei K, et al: Treatment with nimodipine or mannitol reduces programmed cell death and infarct size following focal cerebral ischemia. Neurosurg Rev 2000; 23:145150 43. Cooper DJ, Myles PS, McDermott FT, et al: Prehospital hypertonic saline resuscitation of

patients with hypotension and severe traumatic brain injury: A randomized controlled trial. JAMA 2004; 291:1350 1357 44. Murphy N, Auzinger G, Bernel W, et al: The effect of hypertonic sodium chloride on intracranial pressure in patients with acute liver failure. Hepatology 2004; 39: 464 470 45. Rozet I, Tontisirin N, Muangman S, et al: Effect of equiosmolar solutions of mannitol versus hypertonic saline on intraoperative

brain relaxation and electrolyte balance. Anesthesiology 2007; 107:697704 46. Wu CT, Chen LC, Kuo CP, et al: A comparison of 3% hypertonic saline and mannitol for brain relaxation during elective supratentorial brain tumor surgery. Anesth Analg 110:903907 47. Gemma M, Cozzi S, Tommasino C, et al: 7.5% hypertonic saline versus 20% mannitol during elective neurosurgical supratentorial procedures. J Neurosurg Anesthesiol 1997; 9:329 334

Crit Care Med 2011 Vol. 39, No. 3

559