Effect of Mannitol on Cerebral Blood Volume inPatients With Head Injury

BACKGROUND: Mannitol has traditionally been the mainstay of medical therapy forintracranial hypertension in patients with head injury. We previously demonstrated thatmannitol reduces brain volume in patients with cerebral edema, although whether thisoccurs because of a reduction in brain water, blood volume, or both remains poorlyunderstood.OBJECTIVE: To test the hypothesis that mannitol acts by lowering blood viscosityleading to reflex vasoconstriction and a fall in cerebral blood volume (CBV).METHODS: We used 15O positron emission tomography to study 6 patients withtraumatic brain injuries requiring treatment for intracranial hypertension. Cerebral bloodflow (CBF), CBV, and cerebral metabolic rate for oxygen (CMRO2) were measured beforeand 1 hour after administration of 1.0 g/kg 20% mannitol.RESULTS: CBV rose from 4.1 6 0.4 to 4.2 6 0.2 mL/100 g (P = .3), while intracranialpressure fell from 21.56 4.9 to 13.7 6 5.1 mm Hg (P , .003) after mannitol. Bloodpressure, PaCO2, oxygen content, CBF, and CMRO2 did not change.CONCLUSION: A single bolus of 1 g/kg of 20% mannitol does not acutely lower CBV.Another mechanism, such as a reduction in brain water, may better explain mannitolsability to lower intracranial pressure and reduce mass effect.

KEY WORDS: Cerebral blood flow, Cerebral blood volume, Mannitol, Osmotic

Neurosurgery 70:12151219, 2012 DOI: 10.1227/NEU.0b013e3182417bc2 www.neurosurgery-online.com

Acute head injury is frequently complicatedby edema andmass effect. This increase inintracranial volume can elevate intracra-

nial pressure (ICP), critically reduce cerebralperfusion pressure (CPP), and lead to globalischemia. Additionally, focal edema and massesinduce pressure gradients across intracranial com-partments that can lead to tissue shifts andultimately herniation. Nonsurgical treatmentoptions are limited andmannitol has traditionallybeen the mainstay of medical therapy forintracranial hypertension in patients with headinjury.Despite its widespread use and proven effec-

tiveness in lowering ICP, the mechanism bywhich mannitol produces its effect on brainvolume and ICP remain poorly understood.

Two potential explanations have been proposed.One theory argues that mannitol lowers ICP byreducing cerebral blood volume (CBV), either byraising blood pressure1 or by reducing bloodviscosity, which induce reflex vasoconstriction.2

The second theory contends that mannitol actsby directly reducing brain water. The effect ofosmotic agents on CBV has never been assessedin humans.We previously demonstrated that hemispheric

volume falls 1 hour after administration of 1 g/kgof 20% mannitol to patients with ischemic,cerebral edema, and midline shift.3,4 The tech-nique used in those studies could not differen-tiate whether the observed response was due toa fall in brain water or CBV. To address thisunresolved question, we chose to measure CBVdirectly by using 15O positron emission tomog-raphy (PET) imaging. In a convenience sampleof head injury already receiving osmotic therapyto treat intracranial hypertension, we measuredcerebral blood flow (CBF), CBV, and cerebralmetabolic rate for oxygen (CMRO2) before andafter a bolus of 1 g/kg of mannitol. Our primary

Michael N. Diringer,MD, FCCM*

Michael T. Scalfani, BS, MCI*

Allyson R. Zazulia, MD*

Tom O. Videen, PhD*

Rajat Dhar, MD*

William J. Powers, MD

*Departments of Neurology and Neuro-

logical Surgery, Neurology/Neurosurgery

Intensive Care Unit, Washington University

School of Medicine, St. Louis, Missouri;

Department of Radiology, Washington

University School of Medicine, St. Louis,

Missouri; Department of Neurology,

School of Medicine, University of North

Carolina at Chapel Hill, Chapel Hill, North

Carolina

Correspondence:

Michael N. Diringer, MD, FCCM,

Washington University School of

Medicine,

Department of Neurology,

Campus Box 8111,

660 S Euclid Ave,

St Louis, MO 63110.

E-mail: diringerm@wustl.edu

Received, June 15, 2011.

Accepted, October 27, 2011.

Published Online, November 14, 2011.

Copyright 2011 by theCongress of Neurological Surgeons

ABBREVIATIONS: CBF, cerebral blood flow; CBV,

cerebral blood volume; CMRO2, cerebral metabolic

rate for oxygen; CPP, cerebral perfusion pressure;

GCS, Glasgow Coma Score; ICP, intracranial pres-

sure; TBI, traumatic brain injury

RESEARCHHUMANCLINICAL STUDIESTOPIC RESEARCHHUMANCLINICAL STUDIES

NEUROSURGERY VOLUME 70 | NUMBER 5 | MAY 2012 | 1215

Copyright Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

objective was to test whether mannitol acutely lowers CBV overa similar time frame as the fall in hemispheric volume.

MATERIALS AND METHODS

Eligible Patients

Traumatic brain injury (TBI) patients were eligible if they hada Glasgow Coma Score (GCS) score of# 11 and were receiving osmotictherapy for intracranial hypertension. Exclusion criteria included renalfailure (serum creatinine .1.5 mg/dL), congestive heart failure, cardiacischemia, and pregnancy. All patients were $ 18 years of age. TheHuman Research Protection Office and Radioactive Drugs ResearchCommittee of Washington University approved the study; writteninformed consent was obtained from a legally authorized surrogate.Based on our prior measurements of CBV,5,6 we calculated that, in

a sample size of 6 patients, a 2-tailed Wilcoxon rank sum test will give us80% power to detect a 10% change in CBV with an a= .05.

Clinical Management

In all patients, ICP was monitored continuously by using an intra-parenchymal device (Integra Camino Intracranial Pressure MonitoringKit, Integra, Plainsboro, New Jersey) and intracranial hypertension (ICP. 20 mm Hg) was treated by using a stepwise approach. The initial stepconsisted of sedation with benzodiazepines and opiates followed byintermittent boluses of 20% mannitol. Dose and timing were adjustedbased on the ICP response.The patients were treated in a consistent manner by a single neuro-

intensive care team. Intubation was performed in patients with markedimpairment of consciousness (typically GCS ,9), inability to maintainan adequate airway or manage secretions. Mannitol (20%) wasadministered intravenously by intermittent boluses (starting with a doseof 1 g/kg). Patients were weighed daily, fluid balances were assessedfrequently, and intravenous fluids were adjusted to keep the overall fluidbalance even and to maintain normal intravascular volume. Measure-ments of serum electrolytes, osmolality, and the osmotic gap wereperformed 2 to 4 times a day during osmotic therapy.

Study Protocol

The study was timed so that administration of mannitol between PETmeasurements coincided with treatment of elevated ICP. Baseline mea-surements of GCS, osmolality, and electrolytes were obtained and PETimaging performed to measure CBF, CBV, and CMRO2. Then, 1 g/kgof 20%mannitol was infused over 15 minutes. One hour after beginningthe infusion, PET imaging was repeated and GCS measured again. Heartrate, blood pressure, and CPP were continuously monitored andrecorded every 5 minutes. Serum osmolality and electrolytes weremeasured again 4 hours after completion of mannitol infusion. Duringthe study, the only intervention directed toward ICP management wasmannitol administration. None of the patients received vasopressors, andno interventions directed at changing blood pressure were used. Patientswith TBI had their admission computed tomography (CT) scans scoredaccording to the Marshall criteria.7

PET Methods

All patients were studied on the same Siemens CTI ECATEXACTHR47 PET Scanner located within the Neurology-Neurosurgery IntensiveCare Unit. A neurointensivist was present throughout the study, and all

ongoing therapies were continued throughout the duration of the study.Throughout the PET studies, every effort wasmade tomaintain a constantphysiological state. At the time of each image acquisition, physiologicaldata were recorded.Each scan was acquired in the two-dimensional (2D) mode. An

individual transmission scan was obtained and used for subsequentattenuation correction of emission scan data. All scans were calibrated forconversion of PET counts to quantitative radiotracer concentrations, aspreviously described.8

Regional CBFwasmeasured by bolus injection of 15O-labeled water byusing an adaptation of the Kety autoradiographic method.8,9 RegionalCBV was measured by using a brief inhalation of 15O-labeled carbonmonoxide,10,11 whereas CMRO2 and oxygen extraction fraction werederived from the CBF and CBVmeasurements and an inhalation of 15O-labeled oxygen.12 Quantitative measurements of arterial oxygen content(CaO2) were measured by oximetry.All PET scans for each patient were co-registered and aligned to the

initial baseline CBF study by use of Automated Image Registrationsoftware (AIR, Roger Woods, University of California, Los Angeles,California). An image mask was created for global measurements thatincluded the brain below the superior sagittal sinus down to the level of thepineal gland.

Data Analysis

The primary outcome was the change in global CBV from before toafter mannitol administration. Because of the low number of subjects,CBVwas compared before and after by use of the theWilcoxon rank-sumtest; a P , .05 was considered statistically significant. Similarly, othercontinuous clinical and physiological variables and PET measurementswere compared by use of the Wilcoxon rank-sum test; these analyseswere considered exploratory and uncorrected P values are reported.

RESULTS

Six patients with TBI were studied. Their clinical characteristicsare summarized in Table 1. The median admission GCS was 6.

TABLE 1. Admission Clinical Characteristicsa

Head Injury

No. of subjects 6

Age, y, mean 6 SD 30.2 6 11.9Female, n 1

Race, white/black 5/1

Admission GCS (median, range) 6 (3-8)

Midline shift, mm, mean 6 SD 3.5 6 3.9Marshall grade

Diffuse type 1 0

Diffuse type 2 2

Diffuse type 3 0

Diffuse type 4 0

5Evacuated mass lesion 3

6Nonevacuated mass lesion 1

Days to PET, mean 6 SD 3.4 6 1.8

aSD, standard deviation; GCS, Glasgow Coma Scale.

DIRINGER ET AL

1216 | VOLUME 70 | NUMBER 5 | MAY 2012 www.neurosurgery-online.com

Copyright Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

Admission CT scans were classified as Marshall grade 2 (diffuseinjury, cisterns present, midline shift, 5 mm, and/or no high- ormixed-density lesion.25 mL) in 2 cases, grade 5 evacuated masslesion in 3, and grade 6 non-evacuated mass lesion . 25 mL inone. All were intubated and receiving osmotic therapy withmannitol to treat elevated intracranial pressure.CBV was 4.16 0.4 and 4.2 6 0.2 mL/100 g before and after

mannitol (Table 2). At the same time, ICP fell from 21.5 6 4.9to 13.7 6 5.1 mm Hg (P , .003, Figure 1). A fall in ICP wasobserved in every patient (Figure 2). CPP and CBF tended to riseafter mannitol, although blood pressure was unchanged. oxygenextraction fraction and CMRO2 did not significantly change.

DISCUSSION

The ability of osmotic agents to reduce ICP is undisputed. Themagnitude of the response is determined by the dose,13 baselinelevel of ICP,14 and infusion rate.15 According to the Monro-Kellie doctrine, in order for ICP to fall, there must be a reductionin intracranial volume.16 A reduction in brain volume may occurthrough a reduction in CSF, brain tissue volume (water), or brainblood volume (CBV).Two theories have been proposed to explain how mannitol

could produce a fall in CBV. One argues that a bolus of mannitolproduces acute intravascular volume expansion and elevation inblood pressure, which lead to autoregulatory vasoconstriction,reducing CBV.1,17,18 However, our data indicate that boluses ofmannitol do not consistently raise blood pressure.The second theory proposes that mannitol acts through altering

blood rheology. This is based on the observation that mannitolproduced a rapid constriction of both arterioles and venules on thesurface of the brain.19 This theory states that mannitol reducesblood viscosity, in part, by increasing red blood cell deform-ability, which then raises CBF and improves oxygen delivery,leading to a reflex vasoconstriction and a fall in CBV. This effect

appears to be independent of any changes in hematocrit due tohemodilution.20

Our results do not support the hypotheses that mannitol acts byreducing CBV. Other studies in animals have noted similarfindings. In normal rats, CBV (measured withmagnetic resonanceimaging) increased in a time- and dose-dependent fashionfollowing mannitol administration.21 In a cat model of brainedema, CBV (measured by laser Doppler flowmetry) increasedfollowing mannitol.22 Studies in dogs with the use of labeled redblood cells found an early 20% to 25% increase in CBV 1 to3 minutes after mannitol bolus.23

Two factors must be considered when interpreting our results.First, the vasoconstrictive response to mannitol may occur withinthe first few minutes after mannitol administration and betransient. If this were the case, our measurements at 1 hour mightmiss the response. Such a brief response, that is, no longer

TABLE 2. Physiological Values Before and After Mannitola

Traumatic Brian Injury, n = 6 Baseline Post-Mannitol

CBFb 40.2 6 8.8 42.7 6 8.2CBVc 4.1 6 0.4 4.2 6 0.2CMRO2

b 1.8 6 0.4 1.8 6 0.2MAPd 98.2 6 5.6 96.3 6 8.7ICPd 21.5 6 4.9 13.7 6 5.1e

CPPd 76.7 6 5.9 82.6 6 10.8PaCO2 34.5 6 4.9 35.7 6 4.8PaO2 113 6 31.3 125 6 33.0

aCBF, cerebral blood flow; CBV, cerebral blood volume; MAP, mean arterial

pressure; ICP, intracranial pressure; CPP, cerebral perfusion pressure.bmL/100g per min.cmL/100g.dmm Hg.eP , .003.

FIGURE 1. ICP and CBV before and after 1 g/kg of 20% mannitol in patientswith head injury, poststroke edema, and intracerebral hemorrhage. ICP,intracranial pressure; CBV, cerebral blood volume.

FIGURE 2. Changes in ICP and CBV in each individual patient 1 hour afteradministration of 1 g/kg of 20% mannitol. ICP, intracranial pressure; CBV,cerebral blood volume.

EFFECT OF MANNITOL ON CBV

NEUROSURGERY VOLUME 70 | NUMBER 5 | MAY 2012 | 1217

Copyright Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

detectable at a time when both brain volume and ICP are stillfalling, is unlikely to account for a clinically important response.Second, we did not assess other modulators of cerebrovasculartone, autoregulation, or response to hyperventilation and hypoxiain our patients. If overall vascular reactivity were impaired, then itwould be reasonable to expect that the response to changes inviscosity would be compromised as well.The alternative hypothesis for mannitols ability to lower

ICP argues that it reduces brain water. Mannitol doses of 0.75 to1.25 g/kg reduced ICP and lowered brain water from 79.61% to77.96% (wet/dry weight method) in normal rabbits.24 In normalmonkeys, mannitol produced a parallel change in brain water(measured by CT density) and ICP. Animal models of diseasestates have yielded conflicting results, in part, because of differingmechanisms of injury, mannitol dosing, and correction ofmannitol-induced diuresis. Studies in humans have been limitedto pathological states. In intraoperative biopsies of white mattertaken from TBI patients who were already receiving mannitol,but none in the past 6 hours, brain water content fell from 80.1%to 75.3% following administration of a single dose of 0.28 g/kg ofmannitol.25 In patients with brain tumors, CT brain density(used to measure brain water content) rose progressivelyfollowing doses of 1 to 2 g/kg of mannitol.26

Although these data address the mechanistic question regardinghow mannitol reduces brain volume, important questions remainunanswered. The data do not address what occurs in mannitol-naive patients, because the subjects were studied after they hadalready received mannitol, nor do the data address the effects ofmultiple doses.Our data indicate that a bolus ofmannitol does not reduce CBV

and suggest that the impact of mannitol on ICP is mediated byanother mechanism such as a fall in brain water. Further study ofosmotic therapy is needed to better compare agents, doses, andkinetics and determine their relationship to changes in brain waterand volume and outcome.

CONCLUSION

We sought to test the hypothesis that mannitol reduces ICP inpatients with TBIs and intracranial hypertension by loweringCBV. Using 15O PET, we found that 1 g/kg of 20% mannitoldoes not lower CBV, and suggest that another mechanism, suchas a reduction in brain water, may better explain mannitolsability to lower ICP.

Disclosures

This work was supported by National Institutes of Health grant NS035966.The authors have no personal financial or institutional interest in any of the drugs,materials, or devices described in this article.

REFERENCES

1. Rosner MJ, Coley I. Cerebral perfusion pressure: a hemodynamic mechanism ofmannitol and the postmannitol hemogram. Neurosurgery. 1987;21(2):147-156.

2. Muizelaar JP, Wei EP, Kontos HA, Becker DP. Mannitol causes compensatorycerebral vasoconstriction and vasodilation in response to blood viscosity changes.J Neurosurg. 1983;59(5):822-828.

3. Manno EM, Adams RE, Derdeyn CP, Powers WJ, Diringer MN. The effects ofmannitol on cerebral edema after large hemispheric cerebral infarct. Neurology.1999;52(3):583-587.

4. Videen TO, Zazulia AR, Manno EM, et al. Mannitol bolus preferentially shrinksnon-infarcted brain in patients with ischemic stroke. Neurology. 2001;57(11):2120-2122.

5. Diringer MN, Yundt K, Videen TO, et al. No reduction in cerebral metabolism asa result of early moderate hyperventilation following severe traumatic brain injury.J Neurosurg. 2000;92(1):7-13.

6. Diringer MN, Videen TO, Yundt K, et al. Regional cerebrovascular and metaboliceffects of hyperventilation after severe traumatic brain injury. J Neurosurg. 2002;96(1):103-108.

7. Marshall LF, Marshall SB, Klauber MR, et al. The diagnosis of head injury requiresa classification based on computed axial tomography. J Neurotrauma. 1992;9(suppl 1):S287-S292.

8. Herscovitch P, Markham J, Raichle ME. Brain blood flow measured withintravenous H2(15)O. I. Theory and error analysis. J Nucl Med. 1983;24(9):782-789.

9. Raichle ME, Martin WR, Herscovitch P, Mintun MA, Markham J. Brain bloodflow measured with intravenous H2(15)O. II. Implementation and validation.J Nucl Med. 1983;24(9):790-798.

10. Martin WR, Powers WJ, Raichle ME. Cerebral blood volume measured withinhaled C15O and positron emission tomography. J Cereb Blood Flow Metab.1987;7(4):421-426.

11. Videen TO, Perlmutter JS, Herscovitch P, Raichle ME. Brain blood volume, flow,and oxygen utilization measured with 15O radiotracers and positron emissiontomography: revised metabolic computations. J Cereb Blood Flow Metab. 1987;7(4):513-516.

12. Mintun MA, Raichle ME, Martin WR, Herscovitch P. Brain oxygen utilizationmeasured with O-15 radiotracers and positron emission tomography. J Nucl Med.1984;25(2):177-187.

13. James HE. Methodology for the control of intracranial pressure with hypertonicmannitol. Acta Neurochir (Wien). 1980;51(3-4):161-172.

14. McGraw CP, Alexander E Jr, Howard G. Effect of dose and dose schedule on theresponse of intracranial pressure to mannitol. Surg Neurol. 1978;10(2):127-130.

15. Roberts PA, Pollay M, Engles C, Pendleton B, Reynolds E, Stevens FA. Effect onintracranial pressure of furosemide combined with varying doses and administra-tion rates of mannitol. J Neurosurg. 1987;66(3):440-446.

16. Monro A. Observations on the Structure and Functions of the Nervous System.Edinburgh, United Kindom: Creech and Johnson; 1783.

17. Rosner MJ. Introduction to cerebral perfusion pressure management. NeurosurgClin N Am. 1995;6(4):761-773.

18. Rosner MJ, Rosner SD, Johnson AH. Cerebral perfusion pressure: managementprotocol and clinical results. J Neurosurg. 1995;83(6):949-962.

19. Muizelaar JP, van der Poel HG, Li ZC, Kontos HA, Levasseur JE. Pial arteriolarvessel diameter and CO2 reactivity during prolonged hyperventilation in therabbit. J Neurosurg. 1988;69(6):923-927.

20. Burke AM, Quest DO, Chien S, Cerri C. The effects of mannitol on bloodviscosity. J Neurosurg. 1981;55(4):550-553.

21. Lin W, Paczynski RP, Kuppusamy K, Hsu CY, Haacke EM. Quantitativemeasurements of regional cerebral blood volume usingMRI in rats: effects of arterialcarbon dioxide tension and mannitol. Magn Reson Med. 1997;38(3):420-428.

22. Inao S, Kuchiwaki H, Wachi A, et al. Effect of mannitol on intracranial pressure-volume status and cerebral haemodynamics in brain oedema. Acta Neurochir Suppl(Wien). 1990;51:401-403.

23. Ravussin P, Archer DP, Meyer E, Abou-Madi M, Yamamoto L, Trop D. Theeffects of rapid infusions of saline and mannitol on cerebral blood volume andintracranial pressure in dogs. Can Anaesth Soc J. 1985;32(5):506-515.

24. Donato T, Shapira Y, Artru A, Powers K. Effect of mannitol on cerebrospinal fluiddynamics and brain tissue edema. Anesth Analg. 1994;78(1):58-66.

25. Nath F, Galbraith S. The effect of mannitol on cerebral white matter watercontent. J Neurosurg. 1986;65(1):41-43.

26. Cascino T, Baglivo J, Conti J, Szewczykowski J, Posner JB, Rottenberg DA.Quantitative CT assessment of furosemide- and mannitol-induced changes inbrain water content. Neurology. 1983;33(7):898-903.

DIRINGER ET AL

1218 | VOLUME 70 | NUMBER 5 | MAY 2012 www.neurosurgery-online.com

Copyright Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

COMMENT

T his is an important and valuable study that explores the mechanism ofaction of mannitol. Despite being the workhorse medication fortraumatic brain injury (TBI)management, themechanism throughwhichmannitol acts is still unknown. The authors address the question withpositron emission tomography (PET) imaging in humans, and argue thatif vasoconstriction is the source of decreased volume in the brain, cerebralblood volume (CBV) should decrease. They studied patients with TBIwho were receiving standing doses of mannitol and found no change inCBV by PET imaging in TBI patients.The study cohort is small and included 2 patients with computed

tomography (CT) scans graded as Marshall grade 2. One would not havepredicted these patients with no MLS or cistern effacement to havehad intracranial pressure (ICP) issues requiring mannitol treatment.Interestingly, there also was a small decrease in CBV in 2 patients, and itwould be interesting to know if these 2 patients, possibly with a lowerlesion burden and a larger area of intact BBB, had a different CBVresponse to mannitol.Of minor note, the authors do not provide any information as to the

status of autoregulation in their TBI patients and how a pressure-passive vspressure-active state of autoregulation might affect the interpretationof their results. With the decrease in ICP following mannitol treatment,they observed a compensatory, but not significant, increase in cerebral

perfusion pressure (CPP). How changes in CPP might influence CBVwith different states of autoregulation would have added to the discussion.In this study, patients received multiple doses of mannitol and were

not nave to this medication when the PET studies were performed.Repeated mannitol dosing raises the possibility of rebound edema,possibly due to mannitol crossing the BBB and acting as an osmol.Theoretically, this could potentially contribute to the increase in CBVsuch as was observed following mannitol administration, and alter theinterpretation of their conclusions. PET studies on patients nave tomannitol would be very helpful to refute this possibility.This study has markedly advanced our understanding of the mecha-

nisms underlying the therapeutic efficacy of mannitol. The authors haveshown that mannitol treatment of raised ICP in the setting of acute TBImanagement does not decrease CBV, and have thereby cogently arguedthat mannitol does not act by autoregulatory vasoconstriction induced byvolume expansion or by changes in blood viscosity. These data support,but do not prove the alternate hypothesis, that mannitol decreases ICP bydecreasing brain water. Further studies, perhaps employing PET togetherwith magnetic resonance imaging of brain water, may ultimately enablea comprehensive understanding of one of the most common medicationsused in the neurointensive care unit.

Shirley I. StiverSan Francisco, California

EFFECT OF MANNITOL ON CBV

NEUROSURGERY VOLUME 70 | NUMBER 5 | MAY 2012 | 1219

Copyright Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.