1566

Early Time Course of JV-Acetylaspartate,Creatine and Phosphocreatine, and Compounds

Containing Choline in the BrainAfter Acute Stroke

A Proton Magnetic Resonance Spectroscopy Study

Peter Gideon, MD; die Henriksen, MD, PhD; Bj0rn Sperling, MD; Pernille Christiansen, MD;Tom Skyh0j Olsen, MD, PhD; Henrik S. J0rgensen, MD; and Peter Arlien-S0borg, MD, PhD

Background and Purpose: The early time course after acute stroke of cerebral JV-acetylaspartate, creatineand phosphocreatine, and compounds containing choline was studied in vivo by means of localizedwater-suppressed proton magnetic resonance spectroscopy.

Methods: Eight patients with acute stroke were studied serially in the acute phase, 1 week after, and 2-4weeks after the onset of clinical symptoms. Ten healthy volunteers served as controls. A stimulated echo(STEAM) sequence was used for measurement of the brain metabolites in a volume of interest locatedwithin the infarcted area as visualized by magnetic resonance imaging. For quantification, the unsatur-ated water signal was used as the internal standard. Regional cerebral blood flow in the infarcted area wasmeasured relative to a symmetrically located unaffected area by means of single-photon emissioncomputed tomographic scanning, using 99mTc-labeled </,/-hexamethylenepropyleneamine oxime as the flowtracer.

Results: Relative regional cerebral blood flow was considerably reduced in the infarcted area in the acutephase. After 1 week, hyperemia was seen in all but one patient. The iV-acetylaspartate content wassignificantly reduced, with the loss appearing to occur between 6 and 24 hours after the stroke incident.The reduction in JV-acetylaspartate content was greater in the central part than in the peripheral part ofthe infarcted area. Creatine and phosphocreatine were also reduced in the infarcted area, whereas nosignificant change was seen in the choline content.

Conclusions: Assuming that A'-acetylaspartate content reflects neuronal survival or loss, our results maysuggest that treatment procedures with restoration of blood flow to severely ischemic areas should beinitiated within the first 6 hours after stroke onset. (Stroke 1992;23:1566-1572)

KEY WORDS • cerebral blood flow • cerebral ischemia • metabolism • nuclear magneticresonance

By means of proton magnetic resonance spectros-copy ('H-MRS), it has become possible to detecta number of metabolites in the brain, including

JV-acetylaspartate (NAA), creatine and phosphocreatine(Cr+PCr), and compounds containing choline (Cho).Using the unsaturated water signal as the internal stan-dard, average concentrations of these metabolites in avolume of interest (VOI) can be approximated by assum-ing an average fractional water concentration in braintissue.1-2 These studies suggest concentrations in normal

From the Danish Research Center of Magnetic Resonance(P.G., O.H., P.C.) and the Department of Neurology (T.S.O.,H.S.J., P.A.-S.), Hvidovre Hospital, and the Department of Clin-ical Physiology/Nuclear Medicine (B.S.), Bispebjerg Hospital,University of Copenhagen, Denmark.

Address for correspondence: Peter Gideon, MD, Danish Re-search Center of Magnetic Resonance, Hvidovre Hospital, DK-2650 Copenhagen, Denmark.

Received December 6, 1991; final revision received June 29,1992; accepted July 28, 1992.

brain tissue of 10-12 mM for NAA, 7-10 mM forCr+PCr, and 1-2 mM for Cho.1.2

The role of NAA in the brain is largely unknown.Koller et al3 showed that NAA existed in higher con-centrations in gray matter than in white matter anddisappeared after injection of the neurotoxin kainic acidinto the neurons. A number of studies indicate thatNAA is largely present within neurons.4 Thus, NAAmay be used as a neuronal marker to study the numberof viable neurons in the brain tissue.s

Phosphocreatine serves as a reserve for high-energyphosphates most prominently in neurons and musclecells6 and buffers cellular ATP reservoirs. The protonresonance peak in vivo represents the composite signalof Cr+PCr.67 The Cho peak observed probably repre-sents the major water-soluble Cho-containing com-pounds in the brain, including choline, glycerophospho-choline, and phosphocholine.6-7

Animal studies of ischemic brain using phosphorousmagnetic resonance spectroscopy (31P-MRS) haveshown a rapid reduction in the PCr/inorganic phosphate

by guest on April 19, 2017

http://stroke.ahajournals.org/D

ownloaded from

Gideon et al Changes in Brain Metabolites After Stroke 1567

(Pi) ratio and acidic pH after the onset of ischemia,after which 31P-MRS spectra quickly returned to nor-mal.8-9 Proton spectroscopy has demonstrated rapidformation of lactate during ischemia after clipping ofthe carotid artery, and a gradual reduction of lactatecontent to normal values after restoration of bloodsupply.9 Studies of rat brain indicate that NAA,Cr+PCr, and Cho levels are markedly decreased afterthe induction of ischemia.10 It should be emphasized,however, that such animal models may not be directlycomparable to human pathology; differences in collat-eral perfusion may constitute a major problem.

Studies of brain infarction in humans using 31P-MRShave shown reduction in the PCr/Pi ratio only within thefirst week and acidic pH only within the first 32 hoursafter stroke, after which pH became alkalotic in thesubacute phase and subsequently returned to nor-mal.11-12 This is in agreement with earlier positronemission tomographic studies.13 No changes in meta-bolic ratios or pH were found in human chronic infarc-tion14; however, a decrease of up to 40% of the totalphosphorous signal of infarcted tissue was observed.

Preliminary studies using 'H-MRS in humans indi-cate that acute stroke is followed by loss of NAA in theinfarcted brain tissue.15-17 Recent studies indicate adecrease of NAA, Cr+PCr, and Cho after acutestroke.18-19 These studies, however, do not permit quan-tification because the Tl and T2 relaxation rates of themetabolites in question were not taken into account.

Therapeutic approaches aimed at reducing the neu-rological deficits after acute stroke depend on a betterunderstanding of the pathophysiological mechanismsoperating during severe cerebral ischemia. Thus, esti-mation of the early time course of neuronal death orsurvival after acute stroke in humans, as suggested bychanges in the NAA content, may be of particularclinical importance.

To our knowledge, no studies have been reporteddescribing serial measurements of early changes ofNAA, Cr+PCr, and Cho in infarcted brain tissue,including measurement of regional cerebral blood flow(rCBF) after acute stroke in humans. This study there-fore focuses on the following issues: early changes inNAA, PCr+Cr, and Cho contents after acute stroke;the regional variation of metabolite contents within theinfarcted area; and the time course of relative rCBF inthe infarcted area.

Subjects and MethodsOur study comprises eight patients with acute stroke

(age range, 27-80 [mean±SD, 58.4±19.7] years). 'H-MRS and rCBF measurements were carried out be-tween 6 and 52 hours after the onset of symptoms, withthe time between the 'H-MRS measurement and therCBF measurement ranging between 6 and 24 hours.The measurements were repeated 1 week and 2-4weeks after the stroke incident. For comparison, 10age-matched (50.1 ±22.9 [range, 23-80] years) healthyvolunteers were examined with 'H-MRS. The study wasapproved by the local ethics committee, and informedconsent was obtained in all cases.

rCBF MeasurementsThe rCBF was measured in eight slices using a

brain-dedicated single-photon emission computed to-

mography (SPECT) camera (Tomomatic 232), with 500MBq ""Tc-labeled d,/-hexamethylenepropyleneamineoxime injected intravenously used as a flow tracer. (Adetailed discussion of this method was given in a recentreview.20) The three-dimensional spatial resolution ofSPECT was 12-13 mm full-width at half-maximum.Linearization correction for back diffusion was carriedout according to Lassen et al.21 A contralateral, sym-metrically located region was used as reference for thecalculation of relative rCBF.

'H-MRS MeasurementsThe 'H-MRS experiments were carried out on a

Siemens Helicon SP 63/84 whole-body magnetic reso-nance imaging (MRI) scanner operating at 1.5 T using acircularly polarized headcoil.

Water-suppressed 'H-MRS was obtained using astimulated echo sequence as described by Frahm et al.22

The region of infarction was visualized using a mul-tislice, double-spin echo sequence in the transverseplane with a repetition time (TR) of 2.2 seconds andecho times (TE) of 15 and 90 msec. A VOI was placedwithin the pathological region, with care taken to avoidinclusion of cerebrospinal fluid. The magnetic field wasshimmed globally and locally within the selected VOI toa spectral width at half-maximum of the water signal of3-6 Hz in both regions. At the examinations 1 week and2-4 weeks after stroke, great care was taken to placethe VOI in the same location as before, guided byanatomic landmarks in the images (Figures 1A and 2A).The VOI ranged between 18 ml (3x3x2 cm3) and 27 ml(3x3x3 cm3). To evaluate possible regional variations,additional measurements in a smaller VOI were carriedout, with a volume of 8 ml (2x2x2 cm3) placed at thecenter of the first VOI.

The proton spectra from the VOI and the smallercentral VOI were measured with TE=46, 135, and 270msec, and TR=1.5 seconds. The middle interval (TM)was 30 msec. The number of acquisitions ranged be-tween 128 and 512. To estimate Tl saturation, supple-mentary measurements were carried out in all patientswith TE=46 msec and TR=6.0 seconds. For each echotime, the unsaturated water signal in the VOI wasobtained in one acquisition by setting the amplitude ofthe Gaussian water suppression radio frequency pulsesto zero.

Calculations and StatisticsThe signals in the time domain were multiplied by a

half Gaussian function with a half-width of 256 msec.After Fourier transformation and zero-order phasecorrection, the spectra were baseline corrected using asecond-order spline function. The areas under thepeaks were obtained manually using the software of themanufacturer (numerical integration). To minimize in-terindividual as well as intraindividual variations inmeasurement sensitivity, the calculated metabolite sig-nals were scaled with the unsaturated water signal(internal standard).2'23 Correction for the T2 decay wasdone by extrapolation of the signal versus echo time totime zero using a single exponential function.

The signals arising from the "peripheral volume"were calculated by subtracting the signals of the centralcube from those obtained in the total volume.

by guest on April 19, 2017

http://stroke.ahajournals.org/D

ownloaded from

ct%

» l«

?9

ttu

fly

3906

Str

okt

• 1

4 ho

urs

FIG

UR

E

1.

T2

-wei

gh

ted

m

ag

-

neti

c re

sona

nce

imag

e (t

op p

anel

)an

d pe

rfus

ion

(sin

gle-

phot

onem

issi

on

com

pute

d to

mog

raph

y[S

PE

CT

])

imag

e (b

otto

m p

anel

)ob

tain

ed i

n a

cute

pha

se.

Sele

cted

volu

mes

of

inte

rest

are

ind

icat

edin

to

p p

anel

. P

erfu

sion

im

age

show

s la

rge

perf

usio

n de

fect

on

left

sid

e co

rres

pond

ing

to o

cclu

-si

on o

f le

ft m

iddl

e ce

rebr

al a

rter

y.

Nor

m«l

u»d

«nd

Un

tcr,

120

FIG

UR

E

2.

T2

-wei

gh

ted

m

ag

-

neti

c re

sona

nce

imag

e (t

op

pane

l) a

nd p

erfu

sion

(si

ngle

-pho

-to

n em

issi

on c

ompu

ted

tom

ogra

-ph

y)

imag

e (b

otto

m

pane

l) o

b-

tain

ed 1

wee

k af

ter

stro

ke in

sam

epa

tien

t as

in F

igur

e 1.

Per

fusi

onim

age

now

sho

ws

a pa

rtia

l rep

er-

fusi

on,

wit

h ar

eas

of

hype

rem

ia.

CE

RE

TEC

ST

UD

YSL

ICE

6. O

fi *

40m

m

§8 C/2 B"

by guest on April 19, 2017http://stroke.ahajournals.org/Downloaded from

Gideon et al Changes in Brain Metabolites After Stroke

TABLE 1. Mean Relative Regional Cerebral Blood Flow for the Two Most Severely Affected Contiguous Slices

1569

Volume of interest

3x3x2 cm3

2x2x2 cm3

%

5751

1 Day

Range

32-8028-73

Time

%

9492

after stroke onset

1 Week

Range

41-16236-163

%

123129

2-4 Weeks

Range

74-16770-183

The Friedmann two-way analysis of variance by rankswas used for test of significance of the relative metab-olite signals with time within the patient group. TheMann-Whitney rank sum test for unpaired samples wasused for test of significance between the metabolitesignals from the patient group in the acute phase andthe control group. The level of significance was set at/?<0.05.

ResultsExamples of T2-weighted MRI and perfusion

(SPECT) images in the same patient obtained 1 day and1 week after clinical symptoms are shown in Figures 1and 2.

rCBFThe mean relative rCBF was reduced to about 51%

(range, 28-73%) at the center of the infarcted areas atthe time of the first examination. All but one patientdeveloped reperfusion hyperemia during the following2-4 weeks. The average values are summarized inTable 1.

'H-MRSExamples of 'H-MRS spectra obtained at day 1 and

after 1 week are shown in Figure 3. Resonances wereassigned, according to Michaelis et al,7 as follows: 1)Cr+PCr-methylene singlet at 3.94 ppm; 2) inositols at3.56 ppm; 3) Cho at 3.22 ppm; 4) Cr+PCr-methylsinglet at 3.03 ppm; 5) NAA-aspartyl group at 2.48,2.60, and 2.66 ppm; 6) NAA-methyl singlet at 2.01 ppm;and 7) lactate at 1.33 ppm. Peak 7 is known to belactate; spectra were also obtained at TE=135 msec and270 msec, and the peak exhibited the phase changescharacteristic for lactate.

Acute PhaseIn the acute phase, the NAA signal relative to water

of 1.8X10"4 (range, 0.0-4.1) was lower than that of4.0X10"4 (range, 3.1-4.9) observed in the control group(p<0.01; Figure 4). The measured NAA content versustime in the acute phase is shown in Figure 5. It is seenthat the NAA content decreased from the normal rangeto very low levels between 6 and 52 hours after the onsetof symptoms. In Figure 5, the estimated NAA concen-trations are shown, assuming a fractional water contentof 0.75.24 Using this assumption, the mean NAA con-centration was 11.1 (range, 8.6-13.6) mM in the controlgroup and 5.0 (range, 0.0-11.4) mM in the patientgroup.

The Cr+PCr concentrations in the patient group inthe acute phase were significantly lower than those inthe control group (/?<0.02), whereas no significantdifference in Cho concentrations was seen between thetwo groups (p>0.1; Figure 4). Assuming a fractionalwater content of 0.75, the mean Cr+PCr concentrationwas 7.2 (range, 5.8-9.5) mM in the control group and5.6 (range, 2.8-7.8) mM in the patient group in theacute phase. For Cho, the mean concentration was 1.8(range, 1.1-2.4) mM in the control group and 1.3(range, 0.7-1.9) mM in the patient group.

Lactate content was found in all patients in the acutephase.

Time CourseNo significant change in the NAA content was ob-

served 1 week or 2-4 weeks after acute stroke, and nosignificant changes in relative signals with time wereseen for Cho or Cr+PCr. The lactate content decreasedto undetectable levels 2-4 weeks after stroke in allpatients.

4.0 3.S

Chemical shift /ppm

i!o o'.5 o 4.0 315 2.5 2:0 1:5

Chemical shift/ppm

1.0 0.5 0.0

FIGURE 3. Example of proton spectrum obtained during acute phase (left panel) and 1 week after stroke (right panel) fromsame patient. Repetition time, 1.5 sec; echo time, 46 msec; and TM, 30 msec. Resonance peaks are as follows: 1, creatine andphosphocreatine (Cr+PCr); 2, inositols; 3, choline; 4, Cr+PCr; 5, N-acetylaspartate (NAA); 6, NAA; and 7, lactate. Verticalscales are identical. There is a marked decrease of NAA, Cr+PCr, and lactate signals and a smaller decrease in the choline signal.

by guest on April 19, 2017

http://stroke.ahajournals.org/D

ownloaded from

1570 Stroke Vol 23, No 11 November 1992

5 .

4 .

3 .

1 .

c

•

1

»

p

t

c

•

p

•

c

•

•

p

!

NAA Cr • PCr CHO

FIGURE 4. Metabolite signals relative to unsaturated watersignal in patients (P) obtained at examination in acute phasecompared with control subjects (C). Horizontal bars representmean values. N-Acetylaspartate (NAA) content is consider-ably reduced within the infarcted area. Cr+PCr, creatine andphosphocreatine; CHO, choline.

Regional Variation

In six patients, regional variation of the NAA contentwas estimated during the acute phase and 2-4 weeksafter. As shown in Figure 6, the NAA signal was lowerin the central part than in the outer part of the affectedarea during the acute phase in five of the six patientsstudied (p<0.05). In the sixth patient, no NAA signalwas seen in the total volume measured. The same trendwas seen at 2-4 weeks after the clinical incident.

DiscussionThe main results of the present study are that the

NAA content in infarcted brain tissue is considerablyreduced compared with normal brain tissue; this is inagreement with earlier reports.15-19 The decrease ap-pears to occur mainly between 6 (in one patient) and 24hours after the clinical incident. The loss of NAA seemsto be greater in the central than in the outer part of theVOI, suggesting a heterogeneous distribution of NAAwithin the infarcted area as visualized by MRI. TheCr+PCr content also appears to be reduced in theinfarcted area, whereas no significant changes occurredin the Cho content.

Methodological ConsiderationsQuantification of the metabolite concentrations in

the brain using the unsaturated water signal as theinternal standard depends on conditions related toexperimental as well as biological factors.2 The experi-mental factors include signal linearity, with the amountof metabolites and good volume selection avoidingsignals from the brain tissue outside the VOI. More-over, the spectral resolution must be sufficient to avoidoverlapping of peaks in the spectrum.

The biological factors include the water concentra-tion remaining constant during the actual changes inpathophysiological conditions and known for absolutequantification. The metabolites studied should be fullyMRS visible, and the Tl and T2 relaxation behaviormust be known.

Even with all these conditions fulfilled, it is possibleto calculate only the average concentration of themetabolites in the selected VOI.

The instrumental factors have been tested in a seriesof phantom studies using water solutions of NAA atdifferent concentrations within the biological range.2

The results show an excellent linearity of the NAAsignal with selected volume size as well as with NAAconcentration. The absolute error in calculation ofNAA concentration was approximately 1 mM. In vitrostudies on a two-compartment phantom showed lessthan 1% contamination by signals from outside theselected volume.2

The same linearity of metabolite signals versus vol-ume size (8 ml, 15.6 ml, and 27 ml) was found in theoccipital lobe in five healthy volunteers.2 This findingindicates that the fractional water content of braintissue does not vary significantly under the presentmeasurement conditions.

When using an external standard,1 variations in signalamplitudes due to Bl field inhomogeneity and staticfield inhomogeneity may occur. Furthermore, the use ofan external standard is dependent on instrumentalsettings, including receiver gain, coil sensitivity, andradio-frequency inhomogeneity.12

Concerning the biological factors, the Tl and T2relaxation behavior was estimated at each investigation.The Tl and T2 relaxation rates of the three metabolitesdid not differ significantly from those measured in thecontrol group; furthermore, no significant changes wereseen during the time course of infarction.25 The T2relaxation rate of water in the acute phase was notsignificantly longer in the patient group compared withthe control group; however, at the later examinationsthe calculated T2 relaxation rates of water were signif-icantly longer in the patient group, indicating edema.25

30 30

Tm60

hour.

FIGURE 5. N-Acetylaspartate content (Estimated NAAcone mM) measured in acute phase at first examination ofeach patient plotted against time after onset of clinical symp-toms. Note that value obtained after approximately 6 hours iswithin normal range.

by guest on April 19, 2017

http://stroke.ahajournals.org/D

ownloaded from

Gideon et al Changes in Brain Metabolites After Stroke 1571

6-1

'•£ 2-

0-

outer central outer central•15

•10 i

-5

L 0

dayl week 3

FIGURE 6. Regional distribution of N-acetylaspartate(NAA) content in infarcted area obtained in acute phase and2-4 weeks after stroke, shown as relative signal and millimo-lar concentration (cone). NAA loss is more pronounced incentral than in outer part of lesion.

The changes in the T2 relaxation rates of water with timein the patient group were not significant.25 We used themetabolite ratios in the acute phase, in which the T2 ofwater was not significantly longer, in the comparison withthe control group, as the same assumed fractional watercontent could reasonably be used for the calculation.

It is not known at present whether the metabolitesmeasured in this study are all fully MRS visible. How-ever, the measured NAA concentrations tend to behigher than those measured by direct chemical analysis,4

which suggests that NAA at least may be fully MRSvisible.

With regard to the fractional water concentration, weused a value of 75% (representing a mean value of whitematter [71%] and gray matter [81%]24) in the calcula-tion of the absolute concentrations. The water contentmay increase due to formation of edema during thehyperemic period in the first week after stroke. Themaximal error is 33%, if the water concentration is100%, assuming that all water is MRS visible. The erroris probably less than 10% underestimation of the con-centrations, as we did not observe any significant changein the T2 relaxation behavior of the water signal withtime.25

Regarding spectral resolution, we cannot exclude thatother Af-acetyl moieties may contribute to the resonancepeak observed at 2 ppm. Frahm et al26 have found aresonance peak at 2.045 ppm in a human brain regionconsisting mainly of white matter and suggest that thisaccounts for up to 20% of the peak area assigned toNAA. They tentatively assigned this peak to Af-acetyl-neuraminic acid. Thus, the concentration of NAA maybe overestimated by 20% in normal white matter corre-sponding to about 2 mM.

Based on these considerations, we suggest that eventhough the absolute concentration values can be re-garded only as first-order approximations with an un-certainty of approximately 20%, methodological errorsdo not hamper the main conclusions indicated above.

The calculated metabolite concentrations in the age-matched control group in this study are in agreementwith the values reported earlier.1-2

With regard to the rCBF measurements, we calcu-lated the relative perfusion as the ratio of the count rate

in the infarcted and symmetrically located nonaffectedarea, corrected for early washout of the tracer accordingto Lassen et al.21 However, our results may be influ-enced by scatter of radiation, yielding an overestimationof the amount of tracer in the infarcted area in the acutehypoperfusion phase.20 On the other hand, in the laterphase the hyperemic perfusion may be underestimateddue to initial washout of the tracer. Thus, the calculatedrelative rCBF in the acute phase reflects maximumvalues, and those obtained in the hyperemic phaserepresent minimum values. All in all, this indicates thatthe observed increase in relative perfusion with timerepresents a minimum value.

Pathophysiological ImplicationsHeterogeneous distribution of NAA within the infarcted

area. Our results indicate a heterogeneous distributionof NAA within the affected volume shown by MRI.Partial volume effects are probably of minor importancebecause we found a clearly reduced NAA content in theouter volume as well as in the inner volume. Thetendency toward a greater loss of NAA in the centralpart of the lesion is compatible with the finding that thereduction in blood flow is greater in the central part aswell (Table 1). Thus, these findings support the conceptof a "penumbra zone" with viable neurons in theischemic periphery of the lesion.27

Time course of NAA content. Our results may suggestthat the loss of NAA occurs between 6 and 24 hoursafter the onset of clinical symptoms. Ideally, repetitivestudies on the same patients during the acute phasewithin 24 hours should be carried out. This, however, isalmost impossible to accomplish in practice. Becausethere is evidence suggesting that NAA for all purposes islocated exclusively within neurons,3'4 the observed de-crease in NAA concentration indicates neuronal loss.Assuming that the NAA concentration correlates withthe number of surviving neurons,3 our results suggestthat neuronal damage or loss occurs after 6 hours ofsevere ischemia. This is in agreement with experimentalfindings in animals.828 The question now is whether theneuronal damage is reversible when the blood flow isreestablished during the reperfusion hyperemia. It mustbe emphasized that we did not observe any increase ofthe NAA content during hyperemia in the patients whoshowed no NAA content during the ischemic phase. Wetherefore believe that the loss of NAA is mostlyirreversible.

Creatine+phosphocreatine and choline. We observed asignificant decrease in the Cr+PCr content at the firstexamination. One important methodological implicationis that this peak does not remain stable and is thereforeunsuitable as an internal standard for in vivo quantifica-tion of metabolite concentrations in the brain by MRS.29

In the acute phase, the decrease in Cr+PCr may reflectchanges in energy metabolism with reduced uptake orincreased loss of Cr. However, in the later phase, duringhyperemia the persistent reduction in the Cr+PCr con-tent supports earlier findings that gliotic tissue containsless Cr+PCr than does neuronal tissue.6 The Cho con-tent was within the normal range, and we did not observesignificant changes with time. Duijn et al,19 in a recentstudy, found that Cho was reduced approximately 54%compared to the contralateral hemisphere.

by guest on April 19, 2017

http://stroke.ahajournals.org/D

ownloaded from

1572 Stroke Vol 23, No 11 November 1992

Clinical ImplicationsAssuming that NAA concentration in fact reflects the

number of intact neurons, 'H-MRS may become animportant clinical tool for assessment of the indicationfor treatment and follow-up studies of the effect. Ourresults support earlier findings28 that treatment proce-dures, in order to prevent neuronal damage after severeischemia by restoration of blood flow, should be initi-ated within 6 hours after stroke. It should be empha-sized, however, that the prognostic information of therecorded NAA loss in terms of clinical outcome remainsto be clarified in controlled clinical studies.

References1. Narayana PA, Johnston D, Flamig DP: In vivo proton magnetic

resonance spectroscopy studies of human brain. Magn Reson Imag-ing 1991;9:303-308

2. Christiansen P, Henriksen O, Stubgaard M, Gideon P, LarssonHBW: In vivo quantification of brain metabolites by 'H-MRS usingwater as an internal standard. Magn Reson Imaging (in press)

3. Roller KJ, Zaczek R, Coyle J: Af-acetyl-aspartylglutamate:Regional levels in rat brain and the effects of brain lesions asdetermined by a new HPLC method. J Neurochem 1984;43:1136-1142

4. Birken D, Oldendorf WH: Af-Acetyl-Aspartic acid: A literaturereview of a compound prominent in 'H-NMR spectroscopic studiesof brain. Neurosci Biobehav Rev 1989;13:23-31

5. Arnold DL, Matthews PM, Francis G, Antel J: Proton magneticresonance spectroscopy of human brain in vivo in the evaluation ofmultiple sclerosis: Assessment of the load of disease. Magn ResonMed 1990;14:154-159

6. Miller B: A review of chemical issues in 'H NMR spectroscopy:Af-acetyl-L-aspartate, crcatinc and choline. NMR Biomed 1991;4:47-52

7. Michaelis T, Merboldt K-D, Hanicke W, Gyngell M, Bruhn H,Frahm J: On the identification of cerebral metabolites in localized'H-NMR spectra of human brain in vivo. NMR Biomed 1991;4:90-98

8. Williams SR, Crockard A, Gadian DG: Cerebral ischemia studiedby nuclear magnetic resonance spectroscopy. Cerebrovasc BrainMetabRev 1989;1:91-114

9. Gadian DG, Frackowiak RSJ, Crockard HA, Proctor E, Allen K,Williams SR, Ross Russell RW: Acute cerebral ischemia: Concur-rent changes in cerebral blood flow, energy metabolites, pH, andlactate measured with hydrogen clearance and 3'P and 'H nuclearmagnetic resonance spectroscopy: I. Methodology. J Cereb BloodFlow Melab 1987;7:199-206

10. Peres M, Bourgeois D, Roussel S, Lefur Y, Devoulon P, Remy C,Barrere B, Decorps M, Pinard E, Riche D, Benabid A-L, Seylaz J:Two-dimensional 'H spectroscopic imaging for evaluating the localmetabolic response to focal ischemia in the conscious rat. NMRBiomed 1992;5:11-19

11. Levine S, Welch K, Helpern J, Bruce R, Smith M: Clinical inves-tigation of ischemic stroke by serial 31-phosphorous NMR spec-troscopy. (abstract) Proc Soc Magn Reson Med 1987;2:536

12. Bottomley PA: Human in vivo NMR spectroscopy in diagnosticmedicine: Clinical tool or research probe? Radiology 1989;170:l-15

13. Hakim AM, Pokrupa RP, Villanueva J, Diksic M, Evans AC,Thompson CJ, Meyer E, Yamamoto YL, Feindel WH: The effectof spontaneous reperfusion on metabolic function in early humancerebral infarcts. Ann Neurol 1987;21:279-289

14. Bottomley PA, Drayer BP, Scott Smith L: Chronic adult cerebralinfarction studied by phosphorous NMR spectroscopy. Radiology1986;160:763-766

15. Bruhn H, Frahm J, Gyngell ML, Merboldt KD, Hanicke W, SauterR: Cerebral metabolism in man after acute stroke: New observa-tion using localized proton NMR spectroscopy. Magn Reson Med1989;9:126-131

16. Berkelbach van der Sprenkel JW, Luyten PR, van Rijen PC,Tulleken CAP, den Hollander JA: Cerebral lactate detected byregional proton magnetic resonance spectroscopy in a patient withcerebral infarction. Stroke 1988;19:1556-1560

17. Fenstermacher MJ, Narayana PA: Serial proton magnetic reso-nance spectroscopy of ischemic brain injury in humans. InvestRadiol 1990;9:1034-1039

18. Graham GD, Blamire AM, Howseman AM, Rothman DL, FayadPB, Brass LM, Petroff OAC, Shulman RG, Prichard JW: Protonmagnetic resonance spectroscopy of cerebral lactate and othermetabolites in stroke patients. Stroke 1992;23:333-340

19. Duijn JH, Matson GB, Maudsley AA, Hugg JW, Weiner MW:Human brain infarction: Proton MR spectroscopy. Radiology 1992;183:711-718

20. Andersen AR: Tc-<t/-hexamethylene-propyleneamine oxime(Tc-HMPAO): Basic kinetic studies of a tracer of cerebral bloodflow, in Murray Harper A (ed): Cerebrovascular and Brain Metab-olism Reviews. New York, Raven Press, Publishers, 1989, vol I,pp 288-318

21. Lassen NA, Andersen AR, Friberg L, Paulson OB: The retentionof [Tc]-<t'-HM-PAO in the human brain after intracarotid bolusinjection: A kinetic analysis. J Cereb Blood Flow Metab 1988;6:13-22

22. Frahm J, Bruhn H, Gyngell ML, Merboldt KD, Hanicke W, SauterR: Localized high-resolution proton NMR spectroscopy usingstimulated echoes: Initial applications to human brain in vivo.Magn Reson Med 1989;9:79-93

23. Tofts PS, Wray S: Measuring metabolite concentrations by NMRspectroscopy. NMR Biomed 1988;l:l-10

24. Norton W, Poduslo S, Suzuki K: Subacute sclerosing leukoenceph-alitis: II. Chemical studies including abnormal myelin and anabnormal ganglioside pattern. J Neuropathol Exp Neurol 1966;25:582-597

25. Gideon P, Henriksen O: In vivo relaxation of /V-acetyl-aspartate,creatine plus phosphocreatine, and choline containing compoundsduring the course of brain infarction. Magn Reson Imaging (in press)

26. Frahm J, Michaelis T, Bruhn H, Gyngell ML, Hanicke W, Mer-boldt KD: On the /V-acetyl resonance in proton MR spectra ofhuman brain in vivo, (abstract) Proc Soc Magn Reson Med 1991 ;1:390

27. Olsen TS, Larsen B, Herning M, Skriver ES, Lassen NA: Bloodflow and vascular reactivity in collaterally perfused brain tissue,evidence of an ischemic penumbra in patients with acute stroke.Stroke 1983;14:332-341

28. Crowell RM, Marcoux FW, DeGirolami U: Variability and revers-ibility of focal cerebral ischemia in unanesthetized monkeys. Neu-rology 1981;31:1295-1302

29. Frahm J, Bruhn H, Gyngell ML, Merboldt KD, Hanicke W, SauterR: Localized proton NMR spectroscopy in different regions of thehuman brain in vivo: Relaxation times and concentrations of cere-bral metabolites. Magn Reson Med 1989;ll:47-63

by guest on April 19, 2017

http://stroke.ahajournals.org/D

ownloaded from

Arlien-SøborgP Gideon, O Henriksen, B Sperling, P Christiansen, T S Olsen, H S Jørgensen and P

spectroscopy study.containing choline in the brain after acute stroke. A proton magnetic resonance

Early time course of N-acetylaspartate, creatine and phosphocreatine, and compounds

Print ISSN: 0039-2499. Online ISSN: 1524-4628 Copyright © 1992 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Stroke doi: 10.1161/01.STR.23.11.1566

1992;23:1566-1572Stroke. 

http://stroke.ahajournals.org/content/23/11/1566the World Wide Web at:

The online version of this article, along with updated information and services, is located on

  http://stroke.ahajournals.org//subscriptions/

is online at: Stroke Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer information about this process is available in the

located, click Request Permissions in the middle column of the Web page under Services. FurtherEditorial Office. Once the online version of the published article for which permission is being requested is

can be obtained via RightsLink, a service of the Copyright Clearance Center, not theStrokepublished in Requests for permissions to reproduce figures, tables, or portions of articles originallyPermissions:

by guest on April 19, 2017

http://stroke.ahajournals.org/D

ownloaded from