59ournal ofNeurology, Neurosurgery, and Psychiatry 1997;62:590-595

Early changes in peritumorous oedema andcontralateral white matter after dexamethasone:a study using proton magnetic resonance

spectroscopy

Paul Chumas, Barrie Condon, David Oluoch-Olunya, Stuart Griffiths, Donald Hadley,Graham Teasdale

AbstractAims-To study the mechanism of actionof steroids in patients with peritumorousoedema.Methods-To investigate early cerebralmetabolic changes proton magnetic reso-nance spectroscopy (1H-MRS) was usedbefore and 11 to 14 hours after treatmentwith dexamethasone (12 mg oral loadingand 4 mg four times daily maintenance).Nine patients (two men, seven women,mean age 54) with pronounced oedemaassociated with various intracranialtumours (two astrocytomas, three menin-giomas, two glioblastoma, and two metas-tases) were examined using MRI andMRS. SE1500/135 volume selected MRS(mean volume 21 ml) were performed onan oedematous region and a contralateralregion. All spectra were acquired withand without water suppression. Metabo-lite peak area ratios were determined.Results-Regions of oedema had signifi-cantly (P < 0-01) higher unsuppressedwater than the contralateral regions, asexpected. There was no change at thisearly time point after dexamethasone.The ratio ofthe area of choline containingcompounds to that creatine and phospho-creatine compounds was determined afterwhich the serial ratios of these before andafter were calculated (a serial ratio of 1-0would indicate no change in the choline tocreatine ratios after steroid administra-tion). The mean serial ratios for the areaof oedema were 1'02 (SEM 0.08) and 1-10(0.08) for the contralateral volume ofinterest, indicating no significantchanges. However, significant changes (P< 0.02) were found in the N-acetyl-aspar-tate (NAA)/choline serial ratios (0-86(0.06) in the area ofoedema, 1-20 (0 10) incontralateral brain) and the NAAlcreatineserial ratios (0.86 (0.08) for the oedema,1-25 (0-11) in contralateral brain).Conclusions-Such rapid changes may beexplained either by relatively large alter-ations in the relaxation characteristics ofNAA or, more controversially, by actualchanges in the amounts ofNAA. It is pro-posed that steroids act primarily by caus-ing early metabolic changes that are laterexpressed in improvements in intracra-nial volume relations.

(7 Neurol Neurosurg Psychiatry 1997;62:590-595)

Keywords: brain; dexamethasone; peritumorousoedema; spectroscopy; metabolism

Steroids are routinely employed in the treat-ment of patients with tumours of the CNS yettheir mode of action remains obscure. In par-ticular, the rapid and dramatic clinicalimprovement seen in some patients is difficultto explain on the basis of changes in blood-brain barrier permeability,' 2 blood flow or vol-ume,3-5 decrease in CSF production,' 6 orresolution of peritumorous oedema.7 We haveutilised proton magnetic resonance spec-troscopy ('H-MRS) to study peritumorousoedema before and after treatment with dex-amethasone to examine possible early meta-bolic changes.

Materials and methodsNine patients (two men, seven women, meanage 54) with primary space occupying braintumours were examined using a 1 5T SiemensMagnetom with a standard quadrature trans-mit/receive head coil. Collimated light localiserswere used to serially reposition the patient'shead with respect to the coil and magnet.SE1500/135 (256 averages) volume selectedMRS was performed on the oedematousregion and a contralateral region. All spectrawere obtained with water suppression (256averages) and without it (one average). Thechemical shift selective saturation CHESSsequence was used for water suppression.Global and then local shimming were used tobring the unsuppressed water linewidth (fullwidth at half maximum) in the selected vol-ume of interest (VOI) to within 10 Hz.A standard T2 weighted MRI sequence

(SE2200/80, 192 x 256, 20 x 5 mm slices inthe axial orientation) was acquired both beforeand after treatment to ensure that no macro-scopic changes in brain water had occurredover this short time frame. In addition "scout"sets of sagittal and coronal images wereacquired using an SE400/15 (128 x 256, 10x 10 mm slices) to guide positioning of theVOIs in conjunction with the T2 weightedaxial image (fig 1). These were placed withinthe oedema but excluding the tumour itself.To optimise signal to noise ratios VOIs werechosen to encompass as much oedematouswhite matter as possible, but not to be near theinner table of the skull to avoid line broaden-ing and lipid contamination of the spectra bymarrow and scalp. A second VOI was placed

Department ofNeurosurgeryP ChumasD Oluoch-OlunyaS GriffithsG TeasdaleDepartment of ClinicalPhysicsB CondonDepartment ofNeuroradiology,Institute ofNeurological Sciences,Southern GeneralHospital, Glasgow,ScotlandD HadleyCorrespondence to:Dr Paul Chumas, TheGeneral Infirmary at Leeds,Great George Street, Leeds,West Yorkshire LS 1 3EX.Received 26 October 1995and in revised form15 May 1996Accepted 22 November 1996

590

on 31 August 2018 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.62.6.590 on 1 June 1997. D

ownloaded from

Early changes in peritumorous oedema and contralateral white matter after dexamethasone: a study using proton magnetic resonance spectroscopy

Figure 1 Images showingtypical positioning of VOIover oedema: (A) axial,(B) sagittal, (C) coronal.

B

tion used was the interactive phase correctionof the frequency data to make the baseline aslinear as possible over the frequency range ofthe metabolites of interest. There is alwayssome degree of subjectivity inherent in choos-ing integration limits for peak area calculationsso to avoid introducing between observererrors all peak area calculations were per-formed by one observer (BC).

Ethical approval was obtained for this studyand each patient signed a consent form beforeparticipating. Recruitment of patients was diffi-

over the contralateral brain region and a spec-trum acquired. Figure 2 shows a T2 weightedpositioning image. Hard copy films of the"scout" sets and VOI placements were

obtained and these were used as references toguide positioning of the VOI in the post treat-ment acquisitions to obtain as close a spatialcorrespondence as possible.

After acquisition a fast Fourier transformwas applied to convert the time domain signalsinto frequency domain spectra. No filteringwas applied and the only baseline manipula-

Figure 2 T2 axial slice showing typical positioning ofcontralateral VOI.

A

C

591

on 31 August 2018 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.62.6.590 on 1 June 1997. D

ownloaded from

Chumas, Condon, Oluoch-Olunya, Griffiths, Hadley, Teasdale

Mean (SEM) serial metabolite ratio changes before andafter dexamethasone

ContralateralSerial ratios Oedema side P value

NAAICho 0-86 (0-16) 1-20 (0 10) < 0-02NAAICr 0-86 (0 08) 1-25 (0 11) < 0-02Cho/Cr 1-02 (0-08) 1 10 (0 08) -

Mean 1 00 would represent no change.

cult due to the fact that referring physicians Predexamethasonehad often commenced steroid treatmentbefore transfer. In addition, ethical considera- (a

tions in delaying treatment even for the few .2'hours necessary to secure an MR time slot l)meant that patients with florid symptoms andsigns of raised intracranial pressure were notstudied. None of the patients showed dramaticimprovement after dexamethasone over the Postdexamethasonetime scale of the study.

ResultsComparison of the T2 weighted images dis-closed no visible changes in oedema contrast _or extent after dexamethasone treatment. 3.5 1.8Similarly there were no significant serialchanges in the unsuppressed water peaks Beither in the oedematous or normal sides,which support previous findings.7 The VOIsover oedema had significantly higher (P <0 01) unsuppressed water peak areas thanthose over the contralateral side, as expected.The area ratios calculated were N-acetyl

aspartate (NAA)/choline (Cho), NAA/creatine+ phosphocreatine (Cr) and Cho/Cr, whereCho refers to the peak reflecting choline con-taining compounds and Cr refers to the peak _redeametasonreflecting creatine and phosphocreatine com- CPredexamethasonepounds. As expected the VOI over oedema- I Atous brain before dexamethasone treatmentshowed some reduction in the NAA/Choratios with a mean (SD) ratio of 1-52 (089)compared with 1-84 (024) in the contralateralVOI. For NAA/Cr ratios the values were 1-78(036) and 1-90 (028) respectively. Choline Postdexamethasoneto creatine ratios differed more in the oedema-tous VOI with a mean of 1-79 (1-45) com-

Figure 3 Mean (SEM) 1.5 _serialpeak area ratios for 3.5 1.8oedema and contralateralside. NAACr Chemical shift (ppm)

NAA/Cho Figure 4 Examples of the smaUl but observable changes inT I spectra after dexamethasone for (A) oedematous and (B)IT

C contralateral VOI, with relative reductions in NAA in the|Cho/Cr oedematous region, and relative increases in thecontralateral region. The predexamethasone spectra haveux Cho/Cr been displaced vertically for ease of comparison

01

-1.0 F NAAtoNAA/Cr pared with 1-04 (0-19) on the contralateral

~~~~~~~~side.We are, however, interested specifically inI the changes to these relative peak areas after

drug administration. Thus for both the oede-matous side and the contralateral side wecalculated the following serial ratios: post-NAA/Cho/pre-NAA/Cho; post-NAA/Cr/pre-NAA/Cr; post-Cho/Cr/pre-Cho/Cr. If

> there had been no change then all these serial0.5 Oedema Contralateral region ratios would yield a value of 1 00. Table 1

592

on 31 August 2018 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.62.6.590 on 1 June 1997. D

ownloaded from

Early changes in peritumorous oedema and contralateral white matter after dexamethasone: a study using proton magnetic resonance spectroscopy 593

gives the actual mean values obtained from thenine patients. Figure 3 displays these graphi-cally. A t test was used to determine the statis-tical differences between serial ratio changes inthe oedematous side and serial ratio changesin the contralateral side. Combined SDs wereused in these calculations as the sample sizewas small (< 30).

There were pronounced differences in theway the metabolite ratios change. Whereascholine/creatine serial ratios did not changesignificantly, serial ratios involving NAAshowed a significant divergence between theoedema and contralateral VOIs, with NAApeak areas apparently decreasing by an averageof 14% on the oedematous side and increasingby an average of 22% on the contralateral side(see discussion).

Figure 4 (A and B) shows serial spectra overoedematous and contralateral regions. Smallrelative decreases in the NAA in the oedema-tous region and relative increases in the con-tralateral region are observable.

Negative peaks at the chemical shift posi-tion corresponding to lactate were found in allpatients in the oedema spectra, and in oneinstance in the contralateral spectra. Thesepeak areas were generally small compared withthe metabolite peak areas and showed no evi-dence of change after dexamethasone.

DiscussionSteroids have been used in the treatment ofbrain tumours since the early 1950s whenIngraham et al noted benefit from their use inpatients with craniopharyngioma. However, itwas not until the report of Galicich et al in1961 that the wider application of steroids forall CNS tumours was recognised.9 Over thepast 30 years, numerous studies have beenperformed to attempt to elucidate the mecha-nisms of action but none have been able toexplain the rapid clinical response. In the pre-sent study, significant metabolic changes wereseen within 14 hours, both in the zone ofoedema around the tumour and in the con-tralateral hemisphere. A finding all the moresurprising considering the largely asympto-matic nature of many of the patients and thelack of obvious clinical change. We proposethat steroids have a primary metabolicaction-of which the changes in NAA seen inthis study represent only a small part. Anychanges in blood-brain barrier permeabilityand resolution of the oedema being secondaryto the altered metabolic milieu. The time scaleof these changes is in keeping with possibleenzyme induction or inhibition.

PREVIOUS CONCEPTS OF THE ACTIONS OFSTEROIDS IN PATIENTS WITH A BRAIN TUMOURAfter treatment with steroids clinical improve-ment is usually seen within the first 24hours.'01' Previous investigators have mea-sured the intracranial pressure in patients withcerebral tumours before and after treatmentwith steroids and shown that although there is arapid decrease in the frequency and amplitudeof plateau waves, the baseline intracranial

pressure remains increased for some days.3 12 14Miller and Leech showed a favourable alter-ation of the volume-pressure curve within 24hours after starting steroids, indicating achange in brain elasticity with a greater abilityto compensate.'4 It may be that this changerelates to the decreased CSF production seenafter dexamethasone.6 '5 However, similardegrees of reduction in CSF production canbe seen after treatment with acetazolamide'° 16which does not benefit patients with braintumours. Furtherm-ore, the response tosteroids is variable with no effect on CSF pro-duction after dexamethasone in rhesus mon-keys. 7

Other investigators have concentrated onchanges in regional cerebral blood flow(rCBF) and regional cerebral blood volume(rCBV) but again very variable results havebeen obtained. Reulen et al showed anincrease in rCBF after five to seven days oftreatment with dexamethasone5 whereasLeenders et al reported a decrease in rCBFand rCBV within one to five days of treat-ment.4 Permeability of capillaries in thetumour has been shown to decrease both inexperimental models,2 and in humans' butalthough these changes may occur within thefirst 24 hours, MRI has failed to show signifi-cant changes in relaxation times that wouldindicate altered water content in either thetumour or the surrounding brain for up to sixdays after treatment with dexamethasone.7Thus although capillary permeability mayimprove, there is not a significant change inthe volume of oedema. More recently,Andersen et al analysed changes in the TI(spin lattice or longitudinal relaxation time) inthe zone of oedema around the tumour andfound considerable heterogeneity.'8 Althoughoverall there was a change of only 2% in meanTi after 24 hours of treatment with dexam-ethasone, the zone of oedema around thetumour that had the highest Ti values(termed regions of "superoedema") showedchanges of up to 13%.18 These authors haveconfirmed the slow time scale of oedemaabsorption with only a 12% change being seenin seven days.'9 Again, they noted considerableheterogeneity, with the areas of "super-oedema" showing changes of 50% over sevendays. 9

PROTON MAGNETIC RESONANCE SPECTROSCOPYProton magnetic resonance spectroscopy ('H-MRS) is being used increasingly often as anon-invasive method for analysing regionalcerebral metabolism in vivo. Changes in 'H-MRS have been shown in multiple sclerosis,epilepsy, cerebral infarction, AIDS, and braintumours.20 The hope has been to develop atumour specific metabolic profile to act as afingerprint and aid in preoperative diagnosis.After, initially, a lack of reliable correlation,Preul et al have reported for certain tumours a99% correlation between spectroscopic dataand final pathological diagnosis.2' As a generalrule in cerebral tumours, there is a decrease inthe relative concentration of NAA and crea-tine/phosphocreatine together with an increase

on 31 August 2018 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.62.6.590 on 1 June 1997. D

ownloaded from

Chumas, Condon, Oluoch-Olunya, Griffiths, Hadley, Teasdale

in choline and lactate.22 Despite the fact thatdexamethasone is used routinely in thesepatients no previous study has analysed thechanges in the spectra as a consequence of thisdrug.

Forty years after its identification by Tallan etal, the role of NAA in cerebral metabolismremains uncertain.23 It is found only in thenervous system and is second only to gluta-mate in total concentration of free aminoacids.24 In2 H-MRS, NAA produces a sharpspectral peak suggesting that it may be mobileand in free solution.24 Although regardedwidely as a neuronal marker, there is stilldebate about its site of production. In particu-lar, apparent increases in NAA seen during therecovery phase of various pathological statescasts some doubt on this assumption.25-27Some have suggested that NAA is involved inlipid synthesis in the production of myelin;24 28or as a regulator of protein synthesis; or as astorage form of aspartate; or it is a breakdownproduct of another compound.24

In the present study we focused on changesin the zone around the tumour and found thatthe relative concentration of NAA apparentlyeither decreased on the side of the oedema orincreased on the contralateral side within 14hours after treatment with dexamethasone.This implies that either the amount ofNAA ischanging, or there is a change in the relaxationbehaviour of NAA. The degree of changes inrelaxation time necessary to produce suchlarge changes in apparent amounts of NAApresent can be estimated by applying the stan-dard spin echo signal equations with thesepulse timing variables, and using previous esti-mates of NAA relaxation times262729 of 1650ms (T1) and 330 ms (T2). Such calculationsshow that to account for the increased NAAsignal in the contralateral region, either a 40%reduction in Ti or a 170% increase in T2would be required. The apparent 14% reduc-tion in oedematous NAA signal would requirearound a 30% increase in TI or a 30%decrease in T2.

Such large T2 changes in oedematous NAAare not without precedent (although over themuch longer time course of oedema resolu-tion). Using two different TEs and two differ-ent TRs Kamada et al were able to estimateTI and T2 for the oedematous NAA associ-ated with brain tumours.2627 The Ti showedsimilar values to the NAA of normal brain butthe T2 was found to have decreased by about50%.26 27 The use of steroid treatment is notmentioned, but it is highly likely that all thepatients had received treatment so their resultswould relate to our findings after treatment. Intwo patients who had long term follow up agradual "normalisation of T2" occurred as theoedema resolved.2627 No changes in contralat-eral NAA were found.

Large changes in metabolite T2 values over ashort time after initiation of treatment wouldtherefore be required to explain these results.Changes in binding (and hence alterations inrelaxation time) may indeed provide the expla-nation but it is also possible that we havedetected increases in the transport out of a free

and mobile molecule because of metabolicimprovement. Other possibilities includeincreased utilisation (perhaps for lipid or pro-tein synthesis for repair of damaged myelin) orincreased breakdown (white matter havingthree times as much hydrolase activity as greymatter-despite only having half the concen-tration ofNAA) 24

If we are seeing real changes in NAA ratherthan relaxation effects, the apparent rise inNAA on the contralateral side is more difficultto explain but could represent either increasedproduction or decreased utilisation or break-down of this compound. McIntosh andCooper have previously reported that drugsthat normally raise cerebral concentrations of5-hydroxytryptamine also increase concentra-tions of NAA.30

CEREBRAL METABOLIC CHANGES AFrERDEXAMETHASONECorticosteroids cause systemic metabolicchanges which include hyperglycaemia andcatabolic effects that are usually mediated viaintracellular receptors acting on gene expres-sion.31 Likewise within the brain, steroids havebeen shown to act both genomically and non-genomically,3' in the second case via directaction on cell surfaces, alteration in ion per-meability, and by the release of neurohor-mones and neurotransmitters.31 In the brain,two types of receptors bind glucocorticoidswith high affinity: the type I mineralocorticoidreceptor and the type II glucocorticoid recep-tors32 and although the anatomical distributionof these receptors varies, studies have shownuptake in neurons and all classes of astrocytes(including oligodendrocytes).32 Studies using[3H] dexamethasone have shown uptake in thecerebral hemispheres of about 50% (with theremainder evenly divided between brainstemand cerebellum) in control animals and thisincreases to 75-80% after cerebral damage.33Before trauma, 75% of the dexamethasone islocated in astrocytes and 25% in neuronswhereas after cerebral insult this changes to48% bound to astrocytes and 42% to neu-rons.33 At the subcellular level, accumulationoccurs in the microsomal, lysosomal, andcytoplasmic fractions of the damaged cells.33

Whatever the mechanism of action at thecellular level, steroids have been shown to pre-serve cerebral energy reserves.3436 In particu-lar, in the neonatal period hydrocortisone hasbeen shown to increase brain glucose, glyco-gen, B-hydroxybutarate, and ATP concentra-tions after decapitation in mice.3536 Recentstudies using pretreatment with small doses ofdexamethasone in a neonatal rat model ofhypoxia-ischaemia have shown virtually com-plete protection.37 Even in adult models ofischaemia-in which no beneficial effect fromtreatment with steroids has been shown andthere is even some evidence of a detrimentaleffect-prior treatment with dexamethasonehas maintained ATP concentrations and elec-trical activity.34 Studies of tumour metabolismindicate a preference for non-oxidative metab-olism with structural and kinetic changes inrate limiting enzymes and a reversion to fetal

594

on 31 August 2018 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.62.6.590 on 1 June 1997. D

ownloaded from

Early changes in peritumorous oedema and contralateral white matter after dexamethasone: a study using proton magnetic resonance spectroscopy 595

type enzymes.38 It may be, therefore, that dex-amethasone acts by inducing changes in theseenzymes, which in turn increase local energyreserves and improvement in cellular home-ostasis.

ConclusionThe study confirms that there are significantearly changes in cerebral metabolites aftertreatment with dexamethasone and that thesechanges can be monitored in vivo over thetime that clinical benefit occurs. Our work alsoshows that the consequences of steroid admin-istration must now be taken into account inthe previous reports in which brain tumourswere studied with MRS. Further studies areunderway to examine changes in tissue energystatus by using phosphorus MRS and to exam-ine metabolite changes within the tumours.

The research reported in this paper was supported by theScottish Hospitals Endowment Research Trust.

1 Jarden JO, Dhawan V, Poltorak A, et al. Positron emissiontomographic measurement of blood-to-brain and blood-to-tumor transport of '2Rb: the effect of dexamethasoneand whole brain radiation therapy. Ann Neurol 1985;18:636-46.

2 Shapiro WR, Hiesiger EM, Cooney GA, et al. Temporaleffects of dexamethasone on blood-to-brain and blood-to-tumor transport of 14C-alpha-aminoisobutyric acid inrat C6 glioma. J Neurooncol 1990;8:197-204.

3 Brooks D, Beaney R, Leenders K. The effect of dexam-ethasone therapy on cerebral haemodynamics, oxygenutilization and blood-brain barrier permeability inpatients with brain tumours. In: Capildeo R, ed. Steroids indiseases of the central nervous system. Chichester: JohnWiley, 1989:59-68.

4 Leenders K, Beaney R, Brooks D, et al. Dexamethasonetreatment of brain tumour patients: effects on regionalcerebral blood flow, blood volume, and oxygen utiliza-tion. Neurology 1985;35:1610-6.

5 Reulen H, Hadjidimos H, Schurmann K. The effect of dex-amethasone on water and electrolyte content and onrCBF in perifocal brain edema in man. In: Reulen H,Schurmann K, eds. Steroids and brain edema. Berlin:Springer-Verlag, 1972:239-52.

6 Sato 0, Hara M, Asai T, et al. The effect of dexamethasonephosphate on the production rate of cerebrospinal fluidin the spinal subarachnoid space of dogs. J Neurosurg1973;39:480-4.

7 Bell B, Kean D, Smith M, et al. Brain water measured bymagnetic resonance imaging. Correlation with direct esti-mation and changes after mannitol and dexamethasone.Lancet 1987;i:66-8.

8 Ingraham F, Matson D, Mclaurin R. Cortisone and ACTHas an adjunct to the surgery of craniopharyngiomas. NEnglJ Med 1952;246:568-71.

9 Galicich J, French L, Melby J. Use of dexamethasone in thetreatment of cerebral edema associated with braintumors. Lancet 1961;81:46-53.

10 Kirkham S. The palliation of cerebral tumours with high-dose dexamethasone: a review. Palliat Med 1988;2:27-33.

11 Maxwell R, Long D, French L. The clinical effects of a syn-thetic gluco-corticoid used for brain edema in the prac-tice of neurosurgery. In: Reulen H, Schurmann K, eds.Steroids and brain edema. Berlin: Springer-Verlag, 1972:219-32.

12 Alberti E, Hartmann A, Schutz H, et al. The effects of largedoses of dexamethasone on the cerebrospinal fluid pres-sure in patients with supratentorial tumours. J Neurol1978;217:173-181.

13 Brock M, Zillig C, Wiegand H, et al. The influence of dex-amethasone therapy in ICP in patients with tumours of

the posterior fossa. In: Beks J, Bosch D, Brock M, eds.Intracranial pressure III. Berlin: Springer-Verlag, 1976:235-42.

14 Miller J, Leech P. The effect of mannitol and steroid therapyon intracranial volume-pressure relationships in patients. JNeurosurg 1975;42:274-81.

15 Johnston I, Gilday DC, Hendricks EB. Experimentaleffects of steroid withdrawal on cerebrospinal fluidabsorption. J7 Neurosurg 1975;42:690-5.

16 Oppelt W, Patlak C, Rall D. Effects of certain drugs oncerebrospinal fluid production in the dog. Am J Physiol1964;206:247-50.

17 Martins A, Ramirez A, Solomon L, et al. The effect of dex-amethasone on the rate of formation of cerebrospinalfluid in the monkey. JNeurosurg 1974;41:550-4.

18 Andersen C, Haselgrove JC, Doenstrup S, et al. Resorptionof peritumoural oedema in cerebral gliomas during dex-amethasone treatment evaluated by NMR relaxationtime imaging. Acta Neurochir (Wein) 1993;122:218-24.

19 Andersen C, Astrup J, Gyldensted C. Quantitation of peri-tumoural oedema and the effect of steroids using NMR-relaxation time imaging and blood-brain barrier analysis.Acta Neurochir Suppl (Wien) 1994;60:413-5.

20 Duyn J, Gillen J, Sobering G, et al. Multisection protonMR spectroscopic imaging of the brain. Radiology 1993;188:277-82.

21 Preul M, Caramanos Z, Villemure J, et al. Preoperativediagnosis of the five most common types of supratentorialbrain tumors using in vivo biochemistry [abstract]. 7Neurosurg 1995;82:341A.

22 Henriksen 0, Wieslander S, Gjerris F, et al. In vivo 1H-spectroscopy of human intracranial tumors at 1 5 tesla.Preliminary experience at a clinical installation. ActaRadiol 1991;32:95-9.

23 Tallan H, More S, Stein W. N-acetyl-L-aspartic acid inbrain. Y Biol Chem 1956;224:41-5.

24 Birken D, Oldendorf W. N-acetyl-L-aspartic acid: a litera-ture review of a compound prominent in 1H-NMR spec-troscopic studies of brain. Neurosci Biobehav Rev 1989;13:23-31.

25 Arnold DL. Reversible reduction of NAA after acute cen-tral nervous system damage. Proceedings of the Society ofMagnetic Resonance in Medicine. 1992:643.

26 Kamada K, Houkin K, Iwasaki Y, et al. In vivo proton mag-netic resonance spectroscopy for metabolic changes ofhuman brain edema. Neurol Med Chir (Tokyo) 1994;34:676-81.

27 Kamada K, Houkin K, Hida K, et al. Localized protonspectroscopy of focal brain pathology in humans: signifi-cant effects of edema on spin-spin relaxation time. MagnReson Med 1994;31:537-40.

28 D'Adamo A, Gidez L, Yatsu F. Acetyl transport mecha-nisms. Involvement of N-acetyl aspartic acid in de novofatty acid biosynthesis in the developing rat brain. ExpBrain Res 1968;5:267-73.

29 Frahm J, Bruhn H, Gyngell M. Localised proton NMRspectroscopy in different regions of the human brain invivo. Relaxation time of cerebral metabolites. MagnReson Med 1989;11:47-63.

30 McIntosh J, Cooper J. Function of N-acetyl aspartic acid inthe brain: effects of certain drugs. Nature 1964;203:658.

31 McEwen BS. Non-genomic and genomic effects of steroidson neural activity. TIPS Reviews 1991;12:141-7.

32 Vielkind U, Walencewicz A, Levine JM. Type II glucocor-ticoid receptors are expressed in oligodendrocytes andastrocytes. J Neurosci Res 1990;27:360-73.

33 Kostron H, Fischer J. Regional, cellular, and subcellulardistribution of [3H]dexamethasone in rat brain edema.Surg Neurol 1983;20:48-54.

34 Koide T, Wieloch TW, Siesjo BK. Chronic dexamethasonepretreatment aggravates ischemic neuronal necrosis. JCereb Blood Flow Metab 1986;6:395-404.

35 Thurston JH, Pierce RW. Increase of glucose and highenergy phosphate reserve in the brain after hydrocorti-sone.,7 Neurochem 1969;16:107-1 1.

36 Thurston JH, Hauhart RE, Dirgo JA, et al. Mechanisms ofincreased brain glucose and glycogen after hydrocorti-sone: possible clinical significance. Ann Neurol 1980;7:515-23.

37 Chumas PD, Del Bigio MR, Drake JM, et al. A comparisonof the protective effect of dexamethasone to other poten-tial prophylactic agents in a neonatal rat model of cere-bral hypoxia-ischemia. J Neurosurg 1993;79:414-20.

38 Whittle I. The biology of glioma. In: Miller J, Teasdale G,eds. Current Neurosurgery. New York: ChurchillLivingstone, 1992:255-84.

on 31 August 2018 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.62.6.590 on 1 June 1997. D

ownloaded from