Journal of the Neurological Sciences 261 (2007) 157 166 www.elsevier.

com/locate/jns

Hypertonic saline: First-line therapy for cerebral edema?

Wendy C. Ziai, Thomas J.K. Toung, Anish Bhardwaj
Neurosciences Critical Care Division, Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA Neurosciences Critical Care Division, Department of Neurological Surgery, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA Neurosciences Critical Care Division, Department of Anesthesiology and Critical Care Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA Available online 21 June 2007

Abstract This article highlights the experimental and clinical data, controversies and postulated mechanisms surrounding osmotherapy with hypertonic saline (HS) solutions in the neurocritical care arena and builds on previous reviews on the subject. Special attention is focused on HS therapy on commonly encountered clinical paradigms of acute brain injury including traumatic brain injury (TBI), post-operative retraction edema, intracranial hemorrhage (ICH), tumor-associated cerebral edema, and ischemia associated with ischemic stroke. 2007 Elsevier B.V. All rights reserved.
Keywords: Cerebral edema; Hypertonic saline; Intracranial hemorrhage

1. Introduction Elevated intracranial pressure (ICP) from diverse etiologies results in secondary brain injury by reducing cerebral perfusion pressure (CPP) to critical levels from ischemia and by causing distortion and displacement of brain tissue resulting in compression of vital structures (herniation syndromes) [15,7]. Conventional strategies for resuscitation of patients with intracranial hypertension require both pharmacologic therapies and more definitive surgical interventions. The overriding goal of these measures is to maintain adequate cerebral blood flow (CBF) at a level sufficient to meet neuronal metabolic requirements to prevent cerebral ischemia. Regardless of etiology of elevated ICP, osmotherapy remains the cornerstone of medical therapy for brain resuscitation. The premise for the use of osmotic agents is to reduce the volume of intracranial contents coupled with other routine measures (head elevation to 3060 to augment cerebral venous return, avoidance of
Corresponding author. Present address: Neurosciences Critical Care Program, Oregon Health & Science University, 3181 SW Sam Jackson Park Road, L-226, Portland, OR 97239-3098, USA. Tel.: +1 503 418 1472; fax: +1 503 418 1495. E-mail address: bhardwaj@ohsu.edu (A. Bhardwaj). 0022-510X/$ - see front matter 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jns.2007.04.048

dehydration, hyperventilation, blood pressure augmentation with vasopressors to maintain CPP) and allow time for more definitive treatments (metabolic suppression with pharmacological coma, cerebrospinal fluid drainage, surgical decompression) [13]. Cerebral edema is a defined as an increase in brain water content. Most cases of brain injury resulting in elevated ICP begin as focal cerebral edema. Classically, though simplistic, cerebral edema has traditionally been classified into 3 major types: cytotoxic, vasogenic, and interstitial (hydrocephalic) [13,6,7]. Most brain insults involve a combination of these fundamental mechanisms although one can predominate depending on the type and duration of injury. Cytotoxic edema results from cellular swelling involving neurons, glia and endothelial cells from energy failure (as in ischemia) and affects both gray and white matter. This type of edema is resilient to any known medical treatment. Vasogenic edema results from increased capillary permeability causing breakdown of the blood brain barrier (BBB) (trauma, tumor, cerebral abscess and other inflammatory conditions) and affects mostly white matter. This type of edema responds to both steroids (notably tumor edema) and osmotherapy. Other causes include issue hypoxia and water intoxication that may be responsive to osmotherapy but resilient to steroids [1 3,6,7]. Interstitial edema is a consequence of impaired

158

W.C. Ziai et al. / Journal of the Neurological Sciences 261 (2007) 157166

absorption of cerebrospinal fluid (CSF) and increases in transependymal flow of CSF resulting in acute hydrocephalus. This edema subtype is also not responsive to steroids and the response to osmotic agents is questionable [7]. Although HS solutions have existed for several decades, renewed interest in their use for the treatment of cerebral edema of diverse etiologies has resurfaced just over a decade [15]. A considerable body of evidence from both experimental and clinical literature has emerged suggesting that this therapy may be an effective alternative to other conventional osmotic agents, especially in patients with head trauma or post-operative cerebral edema. The controversies for the use of HS solutions as first-line agents exist in large part due to paucity of clinical studies delineating direct comparisons of HS therapy with conventional therapies, and inadequate knowledge regarding safety, optimum duration and dose, selectivity of benefit for specific cerebral lesions and cost effectiveness [13]. 2. Historical background The concept of osmotherapy for central nervous system disorders dates back to 1919 when two research fellows from Water Reed Army Medical Center were working at the Johns Hopkins Medical School. Weed and McKibben discovered that intravenous injection of concentrated 30% sodium chloride solution into anesthetized cats caused the normal convexity of the brain to shrink 3 to 4 mm below the inner table of the skull (as directly visualized via a craniotomy) with a maximal response at 15 to 30 min after each injection [8]. Conversely, injection of hypotonic solutions resulted in brain herniation via the craniotomy site. These sets of observations have constituted the first description of osmotherapy in the medical literature. In 1927, intravenous administration of concentrated urea was brought into clinical use by Fremont-Smith and Forbes [9] and investigated further by Javid and Settlage [10,11] in the 1950s. About the same time the use of concentrated solutions of human plasma proteins by Hughes et al. [12] was attempted, although this was abandoned quickly due to high cost and concern for allergic reactions. The disadvantages of urea as an osmotic agent were also quickly realized based its distribution throughout total body water (thereby compromising its ability to maintain an effective osmotic gradient), reports of a rebound overshoot effect on ICP, clinical evidence of toxicity, the instability of urea solutions and time required for preparation. It was not long before Wise and Charter in 1962 [13], reported their experience and recommendations for mannitol (alcohol derivative of simple sugar mannose) solution (20%25%), which had the benefits of a longer duration of ICP control, less rebound overshoot, stability in solution, lack of toxicity and relatively low cost. Interestingly, higher doses were used in this study (2.5 to 3 g/kg), although favorable results were also reported with 1 g/kg [14]. To date, mannitol has remained the osmotic agent of choice since its inception. It was not until the 1980s that the

trauma literature reported beneficial effects of small volumes of HS solutions for resuscitation for both human and animal hemorrhagic shock. In these studies, rapid improvement in cardiac index, systemic blood pressure, tissue perfusion and survival were demonstrated with various combinations of HS therapy [15]. The cerebral effects of HS solutions were first demonstrated by Todd et al. [16], whereby neurologically intact rabbits hemodiluted with hypertonic lactated Ringer's solution (480 mosM/L) showed decreased ICP, decreased total brain water content and enhance cerebral blood flow (CBF). These results prompted further laboratory and clinical investigation with HS in the setting of brain injury. 2.1. Pathophysiologic basis for osmotherapy It is suggested that like most osmotic agents, mannitol and HS generally exert similar mechanisms of action in the brain; an early effect (15 to 20 min) on ICP due to optimization of rheological properties of the blood resulting in decreased blood viscosity and hematocrit (volume, rigidity and cohesiveness of red blood cell membranes), increasing CBF and oxygen delivery, resulting in reflex autoregulatory vasoconstriction of cerebral arterioles that reduces CBV and ICP [1719]; this is followed by osmotic shrinkage of brain cells which reaches peak effect at 15 to 30 min after administration and may last from 90 min to 6 h depending on the specific etiology [20] resulting in reduced brain water content and ICP. The rheologic effects are most effective with rapid bolus administration rather than continuous infusion [20,21]. Other properties of mannitol include reduction in systemic vascular resistance (and hence afterload), combined with transiently increased preload and a mild positive ionotropic effect resulting in improved cardiac output and oxygen delivery [22,23], and scavenging of toxic oxygen free radicals with potential cytoprotection [24]. However, intravascular volume is often reduced following its diuretic effect and fluid replacement is an important component of mannitol therapy to avoid both hypovolemia resulting in secondary ischemic injury or elevation of ICP due to reflex vasodilation of cerebral arterioles [20]. HS solutions which are available and used in concentrations ranging from 2% to 23.4% produce increasing osmotic gradients with higher concentrations although there is little clinical evidence for choosing one concentration over another in terms of attenuating brain water content [4]. HS solutions have a different mechanism of action of diuresis compared to mannitol which is freely filtered at the glomerulus and decreases the reabsorption of water (and to a lesser extent sodium) [25] accounting for its diuretic effect and hyponatremia. It is postulated that HS produces its diuretic effect from stimulation of atrial natriuretic peptide (ANP) release rather than direct osmotic diuresis which accounts for its ability to augment intravascular volume and cardiac performance, avoiding hypotension and hypovolemia [26]. Improved CBF and oxygen delivery are believed to occur via dehydration of cerebrovascular endothelial cells,

W.C. Ziai et al. / Journal of the Neurological Sciences 261 (2007) 157166

159

increasing vessel diameter and improving deformability of red blood cells [26]. HS may also produce more complex therapeutic actions including reducing the inflammatory response and immunomodulatory effects via decreasing endothelial cell edema, reducing leukocyte adherence and migration which may further attenuate secondary brain injury [4,26]. In TBI, HS may improve cellular function by re-establishing electrochemical gradients, restoring normal resting membrane potential and may interrupt cell hyperstimulation and subsequent cell death [26]. Assuming the osmolality of plasma, interstitial fluid and cytosol are roughly equal in the normal state, and given that cell membranes are freely permeable to water, osmotic agents in concentrations used clinically draw water from the extracellular (interstitial) and possibly the intracellular spaces of the brain into the intravascular compartment, and dilute the plasma. This largely depends on an intact BBB which is relatively impermeable to small solutes and water [27]. Osmotic solutions can be graded by their tonicity, the effective osmotic forces exerted by solutions in adjoining compartments [7]. The tonicity is expressed mathematically by the Osmotic reflection coefficient () with values ranging from 0 (a freely permeable particle with no osmotic force) to 1, (a completely impermeable particle with ideal osmotic activity). Mannitol has a reflection coefficient of 0.9. Glycerol ( = 0.48) and urea ( = 0.59) are even less osmotically active and have potential for rebound cerebral edema from reversal of osmotic gradient. Sodium chloride has a reflection coefficient of 1 and is theoretically an ideal osmotic agent, and more effective than mannitol. Other properties which are desirable for osmotic agents include being pharmacologically inert, nontoxic, and rapidly excreted. Both mannitol and HS meet these goals under usual clinical use although potential the toxicity profile for HS is still not fully ascertained. The bulk flow of water from brain tissue in response to an osmotic challenge may therefore be significantly limited by anatomical constituents of the BBB, mainly small capillary pore size and reduced number of intercellular clefts. This feature, referred to as the low hydraulic conductivity (Lp) of the cerebral microvasculature implies the need for huge osmotic forces to achieve significant movement of water from the brain and may explain the importance of nonosmotic actions of HS in reducing ICP [7]. Another experimental observation is that a substantially smaller change in brain volume is observed following administration of HS than predicted by osmotic behavior alone [28]. Furthermore, studies using ion-sensitive microelectrodes have demonstrated that while extracellular volume decreases, intracellular water content is maintained by uptake of potassium and sodium chloride [29]. In fact, a rapid regulatory increase in intracellular volume occurs counterintuitive to osmotic dehydration theory. This regulatory mechanism may be blocked by osmotic and loop diuretics such as furosemide which inhibit membrane transport of ions, explaining their synergistic prolongation of ICP

reduction in addition to the mechanism of enhanced diuresis [28]. During prolonged elevation of plasma osmolality, which occurs with ongoing osmotherapy for brain edema over days, excess brain electrolytes are replaced by organic solutes (previously termed idiogenic osmoles) which are now known to be the same osmolytes used by all organisms for volume regulation and include: myo-inositol, taurine, glycerylphosphorylcholine, and betaine, cotransported along with sodium from the extracellular to the intracellular compartment [30]. The slow loss of brain organic osmolytes, especially myo-inositol may explain prolonged edema states and delayed rebound edema when osmotherapy is withdrawn. 2.2. Conventional osmotic agents versus HS solutions in treatment of elevated ICP Although a large scale randomized controlled trial directly comparing equiosmolar doses of mannitol and HS has yet to be undertaken, several clinical studies suggest that HS solutions may be more effective in lowering elevated ICP than mannitol. A prospective randomized study comparing the effects of equal volumes of 7.5% HS to 20% mannitol during elective neurosurgical procedures showed a higher increase in serum osmolality at 15 min after administration of HS although mean arterial pressure, central venous pressure and CSF pressure were not different between the 2 treatment groups [31]. Schwarz et al. compared 100 mL HS hydroxyethyl starch (HSHES) with 40 g mannitol in 9 patients with stroke and 30 episodes of ICP crisis [32]. Treatment was effective in reducing ICP N 10% below baseline in all 16 HSHES-treated and in 10 of 14 mannitoltreated episodes. The rise in serum osmolarity in the HSHES group (10.5 mmol/L) was more than in the mannitol group (6.2 mmol/L). In both pediatric and adult TBI, HS has been used effectively to reduce elevated ICP which was refractory to mannitol administration [3335]. A randomized controlled crossover trial administered equiosmolar bolus infusions of either 200 mL of 20% mannitol or 100 mL of 7.5% HS and 6% dextran-70 solution (HSD) in a randomized fashion to 9 patients with ICP N 20 mm Hg [36]. The result was a significantly greater decrease in ICP and longer duration of effect with HSD compared to mannitol. This study had limitations, however, based on suboptimal administration of mannitol (overly rapid bolus administration and urinary output not specifically replaced) and for choosing a dextran-based HS solution [37]. Addition of dextran to HS significantly increases cost and use of a similar colloid, hetastarch has been associated with bleeding complications in patients with subarachnoid hemorrhage [38]. An earlier randomized study of 20% mannitol and 7.5% HS in 20 patients with TBI, refractory intracranial hypertension (N 25 mm Hg) and persistent coma also demonstrated more effective ICP reduction in the HS-treated group with fewer ICP refractory episodes and fewer interventions [39]. While there was no difference in clinical

160

W.C. Ziai et al. / Journal of the Neurological Sciences 261 (2007) 157166

outcome, this study did not compare equiosmolar doses of the 2 osmotic agents. 2.3. Clinical evidence in brain injury paradigms 2.3.1. Traumatic brain injury and post-neurosurgical procedures Previous experimental studies in animal models have demonstrated a beneficial effect of HS in reducing ICP and brain water content under conditions of hemorrhagic shock without brain injury [4042] as well as in brain injury models of TBI [43]. Although lactate Ringers (LR) remains the resuscitation fluid of choice for trauma patients, a number of clinical studies report positive or improved results with various combinations of HS solutions in patients with TBI with or without hypotension. Brain injury resulting from external trauma or associated with neurosurgical procedures (e.g. prolonged brain retraction) may result in intracranial hypertension as a consequence of cerebral edema, reduced CBF, and consequent secondary ischemic injury. Accumulation of extracellular glutamate leading to cell death may be ameliorated by HS solutions and reestablish the normal direction of sodium glutamate cotransporters [44]. Worthley et al. [45] reported 2 patients with traumatic cerebral edema and elevated ICP refractory to mannitol and furosemide who were each given a single bolus of 30% HS. In both cases, ICP decreased into the normal range for 12 to 24 h with improvement in intravascular volume and renal function. In a prospective trial in 6 patients with severe TBI refractory to standard therapy, intravenous bolus administration of 7.5% HS with 6% hydroxyethyl starch significantly lowered ICP and improved CPP at 30 min without affecting arterial blood pressure. Plasma sodium normalized within 30 min [46]. In a double-blind crossover study of 18 pediatric patients who sustained TBI [47], 3% HS was compared to 0.9% saline following initial resuscitation. Single boluses of 10 mL/kg of each solution were administered and ICP monitored for a 2 h period before cross over. HS treatment was found to significantly reduce elevated ICP by 4 mm Hg for 2 h post-infusion versus no change with 0.9% saline treatment. The maximal serum sodium concentration occurred at 30 min after bolus administration of 3% HS and reached a modest increase to 152 mEq/L. ICP was reduced by 44% at 30 min after administration without any effect on blood pressure following administration of repeated boluses of 7.5% HS in 6% hydroxyethyl starch to 6 patients with severe TBI [48]. Another pediatric cohort of 32 children with severe TBI (GCS b 8) with ICP monitoring was studied in a prospective, randomized controlled fashion comparing Ringer's lactate (LR) with 2% HS (268 mmol/L) as resuscitative fluids over 3 days after injury [49]. A radiographic bias existed in this study with more diffuse head injury in the LR group (6/17) compared with the HS group (3/15). Although ICP and CPP did not differ significantly between groups, a significant correlation between serum sodium concentration and ICP

occurred after 8 h of treatment in the HS group with fewer interventions required to maintain ICP b 15 mm Hg. Other differences between the two groups included a higher frequency of acute respiratory distress syndrome, pneumonia, cardiac arrhythmia, and sepsis in the LR group. The HS group had significantly shorter ICU stays and shorter mechanical ventilation times although survival rates and total duration of hospital stay were similar in the two groups. Neurologic outcomes were not assessed in this study. Serum sodium concentrations were in the range 145 to 155 mEq/L in the HS group compared to 130145 mEq/L in the LR group. No adverse effects of hypertonic resuscitation were observed in this study. Shackford et al. [50] performed a prospective randomized clinical trial using 1.6% HS or LR as resuscitative fluids during the first 5 days of ICU care in 34 patients with moderate to severe TBI. Any hemodynamic instability (systolic blood pressure b 90 mm Hg or urine output b 0.5 mL/kg) was treated with the study solution to restore hemodynamic stability in addition to continuous infusions of 0.9% and 0.45% saline in the HS and LR groups respectively. Although the treatments effectively lowered ICP in both groups without significant difference in ICP between groups at any time after entry, the HS group had a lower admission GCS score and a higher initial mean ICP and required significantly more interventions. No conclusions regarding the efficacy of HS or LR on lowering ICP could be drawn from this study. However, there were no adverse effects from the use of HS on renal, cerebral or pulmonary function. Qureshi et al. [51] retrospectively reviewed 36 patients admitted with severe TBI (GCS 8) treated with HS (2% or 3%) within 48 h of admission for a mean of 72 85 h (mean SD). Compared with 46 patients who did not receive HS, the HS group was more likely to have a penetrating TBI, a mass lesion on CT scan and a higher requirement for pentobarbital coma. There was no difference in the frequency of other interventions for cerebral resuscitation (hyperventilation, mannitol, CSF drainage and vasopressor use). After adjusting for differences between the two groups, in-hospital mortality was higher in the HS group (OR 3.1, 95% CI 1.1 10.2), suggesting that prolonged infusions of HS may not favorably impact requirement for other interventions or survival. The maximum serum sodium concentration in the HS group was 156 mEq/L compared with 145 mEq/L in the standard treatment group. Given the small number of patients and disparity in severity of TBI, it is difficult to draw definitive conclusions from this study. However, it was suggested that the use of bolus administrations or short infusions of HS (up to 24 h) as opposed to prolonged infusions may have impacted outcome differently due to potential for rebound edema from continuous infusions (over 7287 h in this study). Qureshi et al. also investigated the effect of 3% HS on ICP and lateral brain displacement on CT scans in a heterogeneous group of 27 patients with cerebral edema of various etiologies including 8 patients with TBI and 5 patients with post-operative edema (4 with tumor resections and 1 aneurysmal clipping) [52]. The protocol which

W.C. Ziai et al. / Journal of the Neurological Sciences 261 (2007) 157166

161

involved continuous intravenous infusion of 3% saline/ acetate to increase the serum sodium concentration to 145 to 155 mmol/L demonstrated reduction of the mean ICP within the first 12 h correlating with an increase in the serum sodium concentration in patients with TBI (r2 = 0.91, p = 0.03) and post-operative edema (r2 = 0.82, p = 0.06), but not for patients with non-traumatic ICH or cerebral infarction. In the same 2 subgroups of patients, repeat CT scans within 72 h of infusion showed reduction in lateral displacement of the brain. Beyond 3 days, however, the beneficial effect of HS therapy in the TBI group was diminished and 4 patients required pentobarbital coma due to refractory intracranial hypertension. The selective benefit of HS in patients with TBI may be secondary to the underlying mechanism of edema being predominantly vasogenic and due to extensive edema on the pre-treatment CT scans without associated hematomas [52]. Presently, the effect of HS solutions on cerebral water content in clinical TBI is not well studied. A case report of a patient with TBI (multiple hemorrhagic contusions) and refractory intracranial hypertension to mannitol treated with a small volume infusion (1.5 mL/kg) of 18% HS demonstrated that cerebral water content was dramatically reduced on MRI in both healthy and edematous brain regions following treatment, and associated with ICP reduction of 109% lasting approximately 6 h [53]. HS has been evaluated clinically in acute resuscitation protocols for hemorrhagic shock with and without cerebral injury. Pre-hospital resuscitation with HS in patients with TBI and hypotension has been shown to be efficacious, with beneficial effects on blood pressure and survival rates [54,55]. Compared with Ringer's lactate, hypertonic fluids exert a positive ionotropic effect restoring cardiac output, expanding circulating blood volume by extracting volume from the intracellular compartment, especially from endothelial and red blood cells, and inducing pre-capillary vasodilation through a direct action on vascular smooth muscle facilitating oxygen delivery [56]. Intravascular volume is restored with significantly less volume and at lower capillary hydrostatic pressure than with LR [5759]. Multiple clinical trials have evaluated the use of small volume resuscitation with HS solutions. In most cases, 6% dextran-70 is added to the hypertonic solution to enhance the duration and intensity of volume expansion without loss of the hemodynamic effects [56]. Holcroft et al. [60] administered a 250 mL pre-hospital bolus injection of 7.5% HS/ dextran (HSD) to 10 patients with trauma with and without associated TBI. Despite the use of strictly LR for in-hospital resuscitation, small number of patients and greater severity of injury in the LR group, the overall survival was better in the HSD group. An expanded study of 166 patients with TBI showed improved in-hospital survival with HSD treatment (32%) compared to those treated with LR (16%, p b 0.05). The overall survival in the expanded study was, however, not significantly different between HSD and LR groups [61]. A subsequent ambulance transport study [54] observed improved survival over predicted for hypotensive trauma

patients treated with 250 mL of 7.5% HS compared with normal saline. A cohort analysis was performed on individual patient data from 6 previous prospective randomized double-blind trials to evaluate the efficacy of HSD in patients with TBI and hypotension [55]. After adjusting for potential confounding factors, a survival benefit emerged with an in-hospital survival of 38% in the HSD group compared with 27% in the standard of care group (OR, 2.1; p = 0.048). A meta-analysis of controlled clinical studies evaluating efficacy of 7.5% HS and 6% dextran-70 in treating patients with trauma indicated that HS may double the survival rate of the subgroup of patients with both hemorrhagic shock and TBI [62]. Although favorable effects in terms of survival have been demonstrated from the use of HS solutions in the trauma resuscitation setting, results should be interpreted with caution as all clinical trials have used relatively small volumes of hypertonic fluid together with conventional crystalloid therapy and have been unable to define a specific patient population that benefits most from this therapy. A separate issue in the trauma literature is whether hypertonic solutions should be given alone especially in the setting of uncontrolled hemorrhagic shock in which HS may increase blood loss from injured blood vessels and accelerate death [63]. The use of HS should not necessarily be restricted to supratentorial pathologic processes as supported by one case report of a 14 year old with TBI afflicted with flaccid tetraparesis secondary to pontine contusion and brain stem compression, who 2 weeks after admission showed evidence of angiographic vasospasm of the intracranial vertebral arteries and MRI evidence of ischemic brain stem damage. Although the patient's neurologic status had been stable, treatment with 2 subsequent HS infusions (2.7% and 5.4%) for a total of 60 h accompanied by relative hypervolemia resulted in improvement of somatosensory evoked potentials and transient improvement in motor performance. Brain stem auditory evoked potentials were unchanged. It is arguable that the transient clinical improvement was the result of improved perfusion alone in this case because the peak serum sodium level was only 143 mEq/L (serum osmolarity of 292 mosM/L) although treatment with 5.4% HS concentration produced the best improvement in motor function [64]. 2.3.2. Brain tumor and neurosurgical procedures In a well-characterized model of experimental brain tumor, Toung et al. [65] demonstrated that continuous intravenous infusion of 7.5% HS and maintenance of serum sodium between 145155 mEq/L for 48 h attenuates water content more effectively than high dose bolus mannitol (2 g/ kg) or furosemide. Thus far, one prospective randomized trial has investigated the role of HS in the setting of elective supratentorial neurosurgical procedures [31]. Treatment during the 4 h of surgery with 2.5 mL/kg of 20% mannitol (1400 mosM/kg) was compared to 7.5% HS (2560 mosM/ kg) producing similar effects on CSF pressure and clinical

162

W.C. Ziai et al. / Journal of the Neurological Sciences 261 (2007) 157166

assessment of brain bulk although the different osmolalities of the two solutions precluded direct comparison. 2.3.3. Intracerebral hemorrhage (ICH) To date, there have been no clinical trials of HS therapy in the ICH paradigm. However, several experimental studies in animal models suggest a benefit with this therapy. Qureshi et al. [66] demonstrated the reversal of transtentorial herniation and restoration of regional CBF and CMRO2 with a bolus administration of 23.4% HS in a canine model of ICH. In the same model, the investigators in a subsequent study compared the effects of equiosmolar doses of 20% mannitol, 3% and 23.4% HS [67]. While there was an immediate decrease in ICP in all treatment groups, after 2 h only animals receiving continuous 3% HS continued to demonstrate decreased ICP compared to pre-treatment values. Animals in this treatment group also demonstrated higher CPP and significantly lower hemispheric water content compared to those that receive 23.4% HS or mannitol. There were no significant differences in regional CBF, oxygen extraction or consumption between different treatment groups. It should be noted, however, that malignant and delayed rebound edema has been reported following the use of HS in patients with ICH [68]. In an earlier retrospective study by Qureshi et al. [52], there was no correlation between ICP reduction with HS therapy in patients with ICH. 2.3.4. Subarachnoid hemorrhage (SAH) An experimental study in the rat model of SAH demonstrated attenuation of elevated ICP with 7.5% HS treatment with sustained improvement in neurologic recovery particularly in treatment arm comprising of 7.5% HS plus 6% dextran 70 [69]. Based on the premise that HS may have significant rheologic effects, it has been used in patients with cerebral vasospasm following SAH to enhance regional CBF. Suarez et al. [70] in a retrospective series reported the effects of a continuous infusion of HS in 29 patients with mild hyponatremia and symptomatic vasospasm following SAH. No adverse effects were reported and there were no changes in cerebral blood flow velocities that were detected by transcranial Doppler ultrasonography. Tseng et al. [71] reported the effects of repeated bolus infusions of 23.4% HS (17 episodes) in patients with high grade SAH. Xenon-CT determined regional CBF was improved with HS treatment as compared to pre-treatment with accentuation of CPP over a period of observation for 3 h. 2.3.5. Ischemic stroke Experimental studies with the use of HS in the ischemic stroke paradigm have yielded mixed results. Bhardwaj et al. demonstrated that injury volume is worsened with HS treatment as a continuous infusion when begun at the time of reperfusion in a well-characterized transient focal ischemia rodent model [72]. The mechanism of this detrimental effect was not due to impaired CBF. Little is known about the

differential response of neurons and glia to HS solutions during evolution of cerebral infarction. For example, in vitro studies have demonstrated that hypertonichyperoncotic saline differentially affects healthy and glutamate-injured primary hippocampal neurons and astrocytes [73]. However, brain water content was decreased significantly when treatment was begun 24 h following onset of focal ischemia and serum Na+ was maintained 145155 mEq/L [74]. Likewise, attenuation in brain edema produced with HS in this model was comparable to that achieved with large doses (2 g/kg every 6 h) of bolus mannitol [74]. In a subsequent study, brain water was markedly attenuated in the ipsilateral as well as the contralateral hemispheres when onset of treatment was delayed for 6 h following permanent focal ischemia [75]. There may be competing (anti-neuronal and anti-edema) effects of HS solutions in ischemic stroke, and the beneficial osmotic effects on stroke-associated cerebral edema may be dependent on timing of the onset and duration of therapy in relation to maturation of the lesion following ischemic stroke. Very few clinical studies have investigated the beneficial role of HS in ischemic stroke. Schwarz et al. [32] demonstrated 7.5% HS6.5% hydroxyethylstarch attenuated elevated ICP more reliably than 20% mannitol in 9 patients with large hemispheric strokes. The same group subsequently reported effectiveness of treatment with 10% HS in controlling ICP elevations that were refractory to treatment with mannitol in 9 patients with acute ischemic stroke [76]. 2.4. Toxicity profile and safety considerations Therapeutic concerns with mannitol include significant and well known systemic side effects including hypotension, hemolysis, hyperkalemia, renal insufficiency and pulmonary edema [15,7,15]. The literature suggests that side effect profile of HS therapy is much better in comparison to mannitol, but some theoretical complications are possible. To date, there have been no Phase 1 safety trials with HS solutions. Myelin injury is a well-known complication of rapid over-correction of preexisting hyponatremia. Studies on patients treated with HS have not shown evidence of myelin injury on MRI or post-mortem examination [34,35] despite their serum sodium levels in excess of 180 mEq/L [34]. However, the threshold for myelin injury due to a change in serum sodium from a normonatremic to a sustained hypernatremic state is ill defined. In nave-uninjured rats, induced hypernatremia (145155 mEq/L) with HS does not cause myelin injury ( 17 mEq/L) [33] and a change of 35 40 mEq/L is required to induce myelinolysis [77]. Other potential neurologic complications of rapid changes in sodium and plasma osmolality include symptoms and signs of encephalopathy (confusion, lethargy, seizures, and occasionally coma) [13,15]. Subdural hematomas or effusions may occur due to shearing of bridging veins as a result of hyperosmolar contracture of the brain. Risk of rapid volume expansion is a concern in patients with poor cardiovascular

W.C. Ziai et al. / Journal of the Neurological Sciences 261 (2007) 157166

163

reserve and history of heart failure, neurogenic cardiac stun or pulmonary edema. Anecdotal observations suggest that, like mannitol, bolus administration of HS can paradoxically induce transient acute hypotension. Rapid expansion of the plasma volume could lead to hypokalemia and cardiac arrhythmias [13,15]. Use of HS solutions as a mixture of chloride:acetate (50:50) is recommended and frequently utilized in clinical practice as an increased plasma chloride concentration could result in metabolic (hyperchloremic) acidemia [13,15]. HS-induced coagulopathy may enhance bleeding from dilution of plasma coagulation factors, prolonged activated prothrombin and partial thromboplastin times and decreased platelet aggregation [15]. Whenever possible, slow infusion of HS solutions is recommended because rapid changes in osmotic gradients in the serum may lead to hemolysis. Phlebitis may also occur if concentrated solutions are given through the peripheral route. Thus, a central venous route of administration is recommended. Thus far, there are no reports of toxicity or organ system failure from HS other than ventilatory failure secondary to pulmonary edema in patients with poor cardiovascular reserve [52]. Thus caution is advised in the use of HS in patients in this subset of patients. Although not rigorously studied, rapid withdrawal of therapy with HS may result in rebound cerebral edema, leading to elevated ICP or herniation syndromes (Table 1). 3. Future perspectives While the complex mechanisms of both osmotic and nonosmotic action of mannitol have been extensively studied in the cerebral resuscitation paradigm, little is known of action of HS beyond its osmotic effects. Although some literature suggests that serum osmolality of 300320 mosM/L is optimal in patients with brain injury [78,79], this recommendation is not supported by systematic laboratory-based or clinical studies. Much of the evidence for this goal is indirectly supported by studies that utilized mannitol where its beneficial effects were offset by significant systemic side effects [7], when serum osmolality exceeded 320 mosM/L. However, these observations have not been substantiated [80]. We recently demonstrated improved mortality and potent anti-edema action with HS treatment to target levels of N 350 mosM/L when treatment onset was delayed for 24 h following onset of ischemia in a well-characterized animal model of ischemic stroke [75]. The efficacy of HS solutions following global cerebral ischemia has not been investigated, although HS solutions have recently been shown to enhance CBF following cardiopulmonary resuscitation in a swine model of cardiac arrest [81]. Few studies have demonstrated non-osmotic effects of HS on CBF and CSF dynamics, but other potential mechanisms of its action require further study. For example, HS has been shown to modulate inflammatory molecules [8284], regulate neutrophilendothelial cell interactions [85] and attenuate polymorphonuclear neutrophil cytotoxicity [86].

While many inflammatory mediators have been implicated in modulating BBB permeability following acute brain injury, the anti-inflammatory in vivo effects of HS in brain injury remain unexplored. Newer molecular mechanisms of cerebral edema are now being considered by several laboratories. In this regard, the glial membrane water channel, aquaporin4 (AQP4), has received recent attention in the pathogenesis of cerebral edema [87]. Induced hyponatremia causes a pronounced and rapid increase in AQP4 [88] and mice deficient in AQP4 have significantly less cerebral edema following water intoxication as well as following experimental stroke [89]. Conversely, it is not known whether alteration in AQP4 water channels plays an important role in determining the therapeutic efficacy of HS in these brain injury paradigms. Furthermore, timing, duration, most efficacious method of instituting therapy (intravenous bolus versus continuous infusion) and specific lesions responsive to HS remains unclear at the present and require carefully controlled experiments in appropriate animal models of brain injury.
Table 1 Comparison of mannitol and hypertonic saline (HS) Characteristic Dosing (Neurotrauma guidelines) Relative effectiveness Mannitol HS 3060 cm3 of 23.4% saline infused over 20 min [4] Reflection coefficient = 1.0 Potentially greater and more prolonged effect Yes Diuretic action via ANP Augments intravascular volume, maintains MAP, CVP, CO

2 mL/kg 20% mannitol infused over 20 min [79] Reflection coefficient = 0.9 Effectiveness may decrease with repeated administration Rheologic effect Yes Diuretic effect Osmotic diuretic Hemodynamic effect Diuresis may compromise intravascular volume causing hypotension, hypovolemia, rebound intracranial hypertension if not replaced Proposed cellular Antioxidant via free effects radical scavenging Maximum serum osmolarity at osmotic effects on brain are observed Potential for rebound edema Half-life Adverse effects 320 mosM/L

Restores resting membrane potential and cell volume Modulates inflammation Up to 360 mosM/L

Yes 24 h Hypotension, rebound elevation in ICP, hyperkalemia, hemolysis, renal failure

Yes Not known Rebound elevation in ICP, hypokalemia, congestive heart failure, hemolysis, coagulopathy, central pontine myelinolysis

MAP: Mean arterial pressure; CVP: central venous pressure; CO: cardiac output; ICP: intracranial pressure. Guidelines refer to trauma patients with elevated ICP, clinical herniation or progressive neurologic deterioration (before ICP monitoring). Adapted from [90].

164

W.C. Ziai et al. / Journal of the Neurological Sciences 261 (2007) 157166 [10] Javid M, Settlage P. Effect of urea on cerebrospinal fluid pressure in human subjects; preliminary report. J Am Med Assoc Mar 17 1956;160: 9439. [11] Javid M. Urea in intracranial surgery. A new method. J Neurosurg 1961;18:517. [12] Hughes J, Mudd S, Strecker EA. Reduction of elevated intracranial pressure by concentrated solutions of human lyophile serum. Arch Neurol Psychiatr 1938;12:127787. [13] Wise BL, Chater N. The value of hypertonic mannitol solution in decreasing brain mass and lowering cerebro-spinal-fluid pressure. J Neurosurg 1962;19:103843. [14] Shenkin HA, Goluboff B, Haft H. The use of mannitol for the reduction of intracranial pressure in intracranial surgery. J Neurosurg 1962;19: 897901. [15] Zornow MH. Hypertonic saline as a safe and efficacious treatment of intracranial hypertension. J Neurosurg Anesthesiol 1996;8:1757. [16] Todd MM, Tommasino C, Moore S. Cerebral effects of isovolemic hemodilution with hypertonic saline solution. J Neurosurg 1985;63: 9448. [17] Muizelaar JP, Wei EP, Kontos HA, Becker DB. Mannitol causes compensatory cerebral vasoconstriction and vasodilation in response to blood viscosity changes. J Neurosurg 1983;59:8228. [18] Hudak ML, Jones Jr MD, Popel AS, Koehler RC, Traystman RJ, Zeger SL. Hemodilution causes size-dependent constriction of pial arterioles in the cat. Am J Physiol 1989;257:H9127. [19] Burke AM, Quest DO, Chien S, Cerri C. The effects of mannitol on blood viscosity. J Neurosurg 1981;55:5503. [20] Brain Trauma Foundation, American Association of Neurological Surgeons, Joint Section on Neurotrauma and Critical Care. Guidelines for the management of severe head injury. J Neurotrauma 2000;6-7: 449627. [21] Cruz J, Minoja G, Okuchi K. Improving clinical outcomes from acute subdural hematomas with the emergency preoperative administration of high doses of mannitol: a randomized trial. Neurosurgery 2001;4: 86471. [22] Cote CJ, Greenhow DE, Marshall BE. The hypotensive response to rapid intravenous administration of hypertonic solutions in man and in the rabbit. Anesthesiology 1979;50:305. [23] Willerson JT, Curry GC, Atkins JM. Influence of hypertonic mannitol on ventricular performance and coronary blood flow in patients. Circulation 1975;51:1095100. [24] Suzuki J, Imaizumi I, Kayama T, Yoshimoto T. Chemiluminescence in hypoxic brain: Second report: cerebral protective effect of mannitol, vitamin E, and glucocorticoid. Stroke 1985;16:695700. [25] Nau R. Osmotherapy for elevated intracranial pressure: a critical reappraisal. Clin Pharmacokinet 2000;1:2340. [26] Doyle J, Davis D, Hoyt D. The use of hypertonic saline in the treatment of traumatic brain injury. J Trauma Inj Infect Crit Care 2001;2:36783. [27] Raichle MD. Neurogenic control of bloodbrain permeability. Acta Neuropathol (Berl) 1983;SVIII:759. [28] McManus ML, Churchwell KB, Strange K. Regulation of cell volume in health and disease. NEJM 1995;333:12606. [29] Cserr HF, DePasquale M, Nicholson C, Patlak CS, Pettigrew KD, Rice ME. Extracellular volume decreases while cell volume is maintained by ion uptake in rat brain during acute hypernatremia. J Physiol 1991;442:27795. [30] Strange K. Regulation of solute and water balance and cell volume in the central nervous system. J Am Soc Nephrol 1992;3:1227. [31] Gemma M, Cozzi S, Tommasino C, Mungo M, Calvi MR, Cipriani A, et al. 7.5% hypertonic saline versus 20% mannitol during elective neurosurgical supratentorial procedures. J Neurosurg Anesthesiol 1997;9:32934. [32] Schwarz S, Schwab S, Bertram M, Aschoff A, Hacke W. Effects of hypertonic saline hydroxyethyl starch solution and mannitol in patients with increased intracranial pressure after stroke. Stroke 1998;29: 105563.

In the clinical paradigm, there are no large randomized clinical trials to date comparing the conventional osmotic agent mannitol versus HS (equiosmolar concentrations) for cerebral resuscitation (reversal of cerebral herniation, ICP lowering effects, CPP augmentation, rebound cerebral edema) and long-term functional outcomes in critically ill brain injured patients with ICH, SAH, TBI, brain tumor, global cerebral ischemia following cardiac arrest and ischemic stroke. 4. Summary and conclusions Osmotherapy remains the cornerstone of medical therapy for cerebral edema in patients with acute brain injury with or without elevated ICP. While mannitol has remained the osmotic agent of choice for several decades, experimental studies, small case series and a few randomized trials coupled with a more desirable toxicity profile suggest that HS solutions may be more desirable therapeutic agents. Osmotherapy with HS is particularly promising in patients with TBI, brain tumor and post-operative cerebral edema. Further studies, in carefully controlled experimental animal models and randomized clinical trials, are required to determine the safety, timing of onset of therapy, optimum duration of benefit and the particular brain injury paradigms that are most likely to benefit from this therapy. Until these definitive trials are performed, caution is advised in clinical use of these solutions as first line therapy in acute brain injury. Acknowledgments This work is supported in part by the Clinical Research Training Fellowship Award from the American Academy of Neurology (WCZ) and US Public Health Service National Institutes of Health grant NS046379 (AB, TJKT). References
[1] Bhardwaj A, Ulatowski JA. Cerebral edema: hypertonic saline solutions. Curr Treat Options Neurol 1999;1:17987. [2] Harukuni I, Kirsch JR, Bhardwaj A. Cerebral resuscitation: role of osmotherapy. J Anesth 2002;16:22937. [3] Bhardwaj A, Ultaowski JA. Hypertonic saline solutions in brain injury. Curr Opin Crit Care 2004;10:12631. [4] Qureshi AI, Suarez JI. Use of hypertonic saline solutions in treatment of cerebral edema and intracranial hypertension. Crit Care Med 2000;28:330113. [5] Ogden AT, Mayer SA, Connolly SE. Hyperosmolar agents in neurosurgical practice: the evolving role of hypertonic saline. Neurosurgery 2005;57:20715. [6] Klatzo I. Neuropathological aspects of cerebral edema. J Neuropathol Exp Neurol 1967;26:114. [7] Paczynski RP. Osmotherapy. Basic concepts and controversies. Crit Care Med 1997;13:10529. [8] Weed LH, McKibben PS. Experimental alteration of brain bulk. Am J Physiol 1919;48:53155. [9] Fremont-Smith F, Forbes HS. Intraocular and intracranial pressure: an experimental study. Arch Neurol Psychiatr 1927;18:55064.

W.C. Ziai et al. / Journal of the Neurological Sciences 261 (2007) 157166 [33] Horn P, Mnch E, Vajkoczy P, Herrmann P, Quintel M, Schilling L, et al. Hypertonic saline solution for control of elevated intracranial pressure in patients with exhausted response to mannitol and barbiturates. Neurol Res 1999;21:75864. [34] Peterson B, Khanna S, Fisher B, Marshall L. Prolonged hypernatremia controls elevated intracranial pressure in head-injured pediatric patients. Crit Care Med 2000;4:113643. [35] Khanna S, Davis D, Peterson B, Fisher B, Tung H, O'Quigley J, et al. Use of hypertonic saline in the treatment of severe refractory posttraumatic intracranial hypertension in pediatric traumatic brain injury. Crit Care Med 2000;4:114451. [36] Battison C, Andrews PJD, Graham C, Petty T. Randomized controlled trial on the effect of a 20% mannitol solution and a 7.5% saline/6% dextran solution on increased intracranial pressure after brain injury. Crit Care Med 2005;33:196202. [37] Schrot RJ, Muizelaar JP. Sugar or salt? Crit Care Med 2005;33:2578. [38] Trumble ER, Muizelaar JP, Myseros JS, Choi SC, Warren BB. Coagulopathy with the use of hetastarch in the treatment of vasospasm. J Neurosurg 1995;82:447. [39] Vialet R, Albanese J, Thomachot L, Antonini F, Bourgouin A, Alliez B, et al. Isovolume hypertonic solutes (sodium chloride or mannitol) in the treatment of refractory posttraumatic intracranial hypertension: 2 ml/kg 7.5% is more effective than 2 mL/kg 20% mannitol. Crit Care Med 2003;31:16837. [40] Ducey JP, Mozingo DW, Lamiell JM, Okerburg C, Gueller GE. A comparison of the cerebral and cardiovascular effects of complete resuscitation with isotonic and hypertonic saline, hetastarch, and whole blood following hemorrhage. J Trauma 1989;29:15108. [41] Schmoker J. Hypertonic fluid resuscitation improves cerebraloxygen delivery and reduces intracranial pressure after hemorrhagic shock. J Trauma 1991;31:160713. [42] Prough DS, Whitley JM, Taylor CL, Deal DD, DeWitt DS. Regional cerebral blood flow following resuscitation from hemorrhagic shock with hypertonic saline. Influence of a subdural mass. Anesthesiology 1991;75:31927. [43] Mirski MA, Denchev DI, Schnitzer MS, Hanley DF. Comparison between hypertonic saline and mannitol in the reduction of elevated intracranial pressure in a rodent model of acute cerebral injury. J Neurosurg Anesthesiol 2000;12:33444. [44] Phillis JW, Song D, O'Reagan MH. Effect of hyperosmolarity and ion substitutions on amino acid efflux from the ischemic rat cerebral cortex. Brain Res 1999;828:111. [45] Worthley LI, Cooper DJ, Jones N. Treatment of resistant intracranial hypertension with hypertonic saline: report of two cases. J Neurosurg 1988;68:47881. [46] Hartl R, Ghajar J, Hochleuthner H, Mauritz W. Treatment of refractory intracranial hypertension in severe traumatic brain injury with repetitive hypertonic/hyperoncotic infusions. Zentralbl Chir 1997;122:1815. [47] Fisher B, Thomas D, Peterson B. Hypertonic saline lowers raised intracranial pressure in children after head trauma. J Neurosurg Anesthesiol 1992;4:410. [48] Hartl R, Ghajar J, Hochleuthner H, Mauritz W. Hypertonic/ hyperoncotic saline reliably reduces ICP in severely head-injured patients with intracranial hypertension. Acta Neurochir Suppl (Wien) 1997;70:1269. [49] Simma B, Burger Falk M, Sacher P, Fanconi S. A prospective, randomized, and controlled study of fluid management in children with severe head injury: lactated Ringer's solution versus hypertonic saline. Crit Care Med 1998;26:126570. [50] Shackford SR, Bourguignon PR, Wald SL, Rogers FB, Osler TM, Clark DE. Hypertonic saline resuscitation of patients with head injury: a prospective, randomized clinical trial. J Trauma 1998;44:508. [51] Qureshi AI, Suarez JI, Castro A, Bhardwaj A. Use of hypertonic saline/ acetate infusion in treatment of cerebral edema in patients with head trauma. Experience at a single center. J Trauma 1999;47:65965. [52] Qureshi AI, Suarez JI, Bhardwaj A, Mirski M, Schnitzer MS, Hanley DF, et al. Use of hypertonic (3%) saline/acetate infusion in the

165

[53]

[54]

[55]

[56] [57]

[58] [59]

[60]

[61]

[62]

[63] [64]

[65]

[66]

[67]

[68]

[69]

[70]

[71]

[72]

treatment of cerebral edema: effect on intracranial pressure and lateral displacement of the brain. Crit Care Med 1998;26:4406. Saltarini M, Massarutti D, Baldassarre M, Nardi G, De Colle C, Fabris G. Determination of cerebral water content by magnetic resonance imaging after small volume infusion of 18% hypertonic saline solution in a patient with refractory intracranial hypertension. Eur J Emerg Med 2002;9:2625. Vassar MJ, Fischer RP, O'Brien PE, Bachulis BL, Chambers JA, Hoyt DB, et al. A multicenter trial for resuscitation of injured patients with 7.5% sodium chloride. The effect of added dextran 70. The Multicenter Group for the Study of Hypertonic Saline in Trauma Patients. Arch Surg 1993;128:100311. Wade CE, Grady JJ, Kramer GC, Younes RN, Gehlsen K, Holcroft JW. Individual patient cohort analysis of the efficacy of hypertonic saline/ dextran in patients with traumatic brain injury and hypotension. J Trauma 1997;42:S615. Rocha e Silva M. Hypertonic saline resuscitation. Medicina (B Aires) 1998;58:393402. Shackford SR, Norton CH, Todd MM. Renal, cerebral, and pulmonary effects of hypertonic resuscitation in a porcine model of hemorrhagic shock. Surgery 1988;104:55360. Velasco IT, Pontieri M, Rocha e Silva JR. Hyperosmotic NaCl and severe hemorrhagic shock. Am J Physiol 1980;239:H66473. Nakayama S, Sibley L, Gunther RA. Small volume resuscitation with hypertonic saline (2300 mOsm/liter) during hemorrhagic shock. Circ Shock 1984;12:14959. Holcroft JW, Vassar MJ, Turner JE, Derlet RW, Kramer GC. 3% NaCl and 7.5% NaCl/dextran 70 in the resuscitation of severely injured patients. Ann Surg 1987;206:27988. Vassar MJ, Perry CA, Gannaway WL, Holcroft JW. 7.5% sodium chloride/dextran for resuscitation of trauma patients undergoing helicopter transport. Arch Surg 1991;128:100311. Wade CE, Kramer GC, Grady JJ, Fabian TC, Younes RN. Efficacy of hypertonic 7.5% saline and 6% dextran-70 in treating trauma: a metaanalysis of controlled clinical studies. Surgery 1997;122:60916. Krausz MM, Amstislavsky T. Hypertonic sodium acetate treatment of hemorrhagic shock in awake rats. Shock 1995;4:5660. Gemma M, Cozzi S, Piccoli S, Magrin S, De Vitis A, Cenzato M. Hypertonic saline fluid therapy following brain stem trauma. J Neurosurg Anesthesiol 1996;8:13741. Toung TJK, Tyler B, Brem H, Traystman RJ, Kirsch JR, Hurn PD, et al. Hypertonic saline ameliorates cerebral edema associated with experimental brain tumor. J Neurosurg Anesthesiol 2002;14: 18793. Qureshi AI, Wilson DA, Traystman RJ. Treatment of transtentorial herniation unresponsive to hyperventilation using hypertonic saline in dogs: effect on cerebral blood flow and metabolism. J Neurosurg Anesthesiol 2002;14:2230. Qureshi AI, Wilson DA, Traystman RJ. Treatment of elevated intracranial pressure in experimental intracerebral hemorrhage: comparison between mannitol and hypertonic saline. Neurosurgery 1999;44:105564. Qureshi AI, Suarez AI, Bhardwaj A. Malignant cerebral edema in patients with hypertensive intracerebral hemorrhage associated with hypertonic saline infusion: a rebound phenomenon? J Neurosurg Anesthesiol 1998;10:18892. Zausinger S, Thal SC, Kreimeier U, Messmer K, Schmid-Elsaesser R. Hypertonic fluid resuscitation from subarachnoid hemorrhage in rats. Neurosurgery 2004;55:67986. Suarez JI, Qureshi AI, Parekh PD, Razumovsky A, Tamargo RJ, Bhardwaj A, et al. Administration of hypertonic (3%) sodium chloride/ acetate in hyponatremic patients with symptomatic vasospasm following subarachnoid hemorrhage. J Neurosurg Anesthesiol 1999;11:17884. Tseng MY, Al-Rawi PG, Pickard JD, Rasulo FA, Kirkpatrick PJ. Effect of hypertonic saline on cerebral blood flow in poor-grade patients with subarachnoid hemorrhage. Stroke 2003;34:138996. Bhardwaj A, Harukuni I, Murphy SJ, Alkayed NJ, Crain BJ, Koehler RC, et al. Hypertonic saline worsens infarct volume after transient focal ischemia in rats. Stroke 2000;31:1694701.

166

W.C. Ziai et al. / Journal of the Neurological Sciences 261 (2007) 157166 [82] Arbabi S, Garcia I, Bauer G, Maier RV. Hypertonic saline induces prostacyclin production via extracellular signal-regulated kinase (ERK) activation. J Surg Res 1999;83:1416. [83] Oreopoulos GD, Hamilton J, Rizoli SB, Fan J, Lu Z, Li YH, et al. In vivo and in vitro modulation of intracellular adhesion molecule (ICAM)-1 expression by hypertonicity. Shock 2000;14:40914. [84] Oreopoulos GD, Bradwell S, Lu Z, Fan J, Khadaroo R, Marshall JC, et al. Synergistic induction of IL-10 by hypertonic saline solution and lipopolysaccharides in murine peritoneal macrophages. Surgery 2001: 5765. [85] Angle N, Cabello-Passini R, Hoyt DB, Loomis WH, Shreve A, Namiki S, et al. Hypertonic saline infusion: can it regulate human neutrophil function? Shock 2000;14:5038. [86] Ciesla DJ, Moore EE, Zallen G, Biffl WL, Silliman CC. Hypertonic saline attenuation of polymorphonuclear neutrophil cytotoxicity: timing is everything. J Trauma 2000;48:38895. [87] Wen H, Nagelhus EA, Amiry-Moghaddam M, Arge P, Ottersen OP, Nielsen S. Ontogeny of water transport in rat brain: postnatal expression of the aquaporin-4 water channel. Eur J Neurosci 1999;11:93545. [88] Vajda Z, Promeneur D, Doczi T, Sulyok E, Frokiaer J, Ottersen OP, et al. Increased aquaporin-4 immunoreactivity in rat brain in response to systemic hyponatremia. Biochem Biophys Res Commun 2000;270:495503. [89] Manley GT, Fujimura M, Ma T, Noshita N, Filiz F, Bollen AW, et al. Aquaporin-4 deletion in mice reduces brain edema after acute water intoxication and ischemic stroke. Nat Med 2000;6:15963. [90] Knapp JM. Hyperosmolar therapy in the treatment of severe head injury in children. Mannitol and hypertonic saline. AACN Clin Issues 2005;16:199211.

[73] Himmelseher S, Pfenninger E, Morin P, Kochs E. Hypertonic hyperoncotic saline differentially affects healthy and glutamate-injured primary rat hippocampal neurons and cerebral astrocytes. J Neurosurg Anesthesiol 2001;13:12030. [74] Toung TJK, Hurn PD, Traystman RJ, Bhardwaj A. Global brain water increases after experimental focal cerebral ischemia: effect of hypertonic saline. Crit Care Med 2002;30:6449. [75] Toung TJ, Chang Y, Lin J, Bhardwaj A. Increases in lung and brain water following experimental stroke: effect of mannitol and hypertonic saline. Crit Care Med 2005;33:2038. [76] Schwarz S, Georgiadis D, Aschoff A, Schwab S. Effects of hypertonic (10%) saline in patients with raised intracranial pressure after stroke. Stroke 2002;33:13640. [77] Soupart A, Penninckx R, Namias B, Stenuit A, Perier O, Decaux G. Brain myelinolysis following hypernatremia in rats. J Neuropathol Exp Neurol 1996;55:10613. [78] Ropper AH, Gress DR, Diringer MN, Green DM, Mayer SA, Bleck TP. Treatment of intracranial hypertension. Neurological and neurosurgical intensive care. New York: Raven Press Ltd; 2003. p. 2651. [79] Brain Trauma Foundation, American Association of Neurological Surgeons, Joint Section on Neurotrauma and Critical Care: use of mannitol: Brain Trauma Foundation, The American Association of Neurological Surgeons, The Joint Section on Neurotrauma and Critical Care. J Neurotrauma 2000;17:215. [80] Diringer MN, Zazulia AR. Osmotic therapy: fact and fiction. Neurocrit Care 2004;1:21934. [81] Krep H, Briel M, Sinn D, Hagendorff A, Hoeft A, Fischer M. Effects of hypertonic versus isotonic infusion therapy on regional cerebral blood flow after experimental cardiac arrest cardiopulmonary resuscitation in pigs. Resuscitation 2004;63:7383.