Acta Neurol Scand.

, 1986:74:439-451

Key words: stroke; CBF; xenon-133; inhalation tomography

Cerebral blood flow in acute and chronic ischemic

stroke using xenon- 133 inhalation tomography
S. Vorstrup, 0. B. Paulson, N. A. Lassen
The Department of Neurology, Rigshospitalet,
Copenhagen, Denmark

ABSTRACT - Serial measurements of cerebral blood flow (CBF) were performed in 12 pa-
tients with acute symptoms of ischemic cerebrovascular disease. CBF was measured by xen-
on-133 inhalation and single photon emission computer tomography. Six patients had severe
strokes and large infarcts on the CT scan. They showed in the acute phase (Days 1-3) very
large low-flow areas, larger than the hypodense areas seen on the CT scan. The cerebral va-
soconstrictor and vasodilator capacity was tested in the acute phase following aminophylline
and acetazolamide, respectively. A preserved but reduced reactivity was seen at both tests in
all 6 cases in the infarct and the peri-infarct areas. On Days 5-25,4 of the patients had trans-
itory increases (59- 108%) of CBF, probably corresponding to lysis of an intracerebral em-
bolic occlusion. The other 2 patients showed on Days 7-15 only a moderate CBF increase
(appr. 20%),both had occlusion of the relevant internal carotid artery. In all 6 patients, CBF
studies at 2 and 6 months resembled the acute phase, showing large areas with reduced flow.
At the 6 months follow-up, the vasodilatory stress test was repeated, and all but one showed
a preserved but reduced vasoreactivity in the infarct and peri-infarct tissue. Of the remaining
6 patients, one had a pontine infarct and one had no lesions on the CT scan, both having nor-
mal angiograms and CBF maps. Four patients had small deep or subcortical Cr lesions, and
showed a slight, but persistent CBF reduction of about 6 4 % in the parietal region on the af-
fected side. No changes in the flow pattern were seen at the vasoreactive studies. A likely ex-
planation for the finding of superjacent low-flow areas is an intrahemispheric uncrossed dias-
chisis. This interpretation is discussed in relation to the peri-infarct low-flow area seen in the
6 cases with large infarcts.

Accepted for publication July 11, 1986

Following a moderate to severe ischemic stroke, hy-
podense lesions are seen on the CT scan within 24-
48 h in the majority of cases in regions correspond-
ing to the neurological symptoms (1).In the chronic
phase, these lesions are taken to represent complete
cerebral infarction (2,3) and a correlation between
the size of the lesion and the clinical symptoms have
been established. But, despite rather severe clinical
symptoms, some patients have only small CT-le-
sions, and some have only lacunar infarcts or no le-
sion at all (4). Lacunes were originally described as
post mortem findings in patients with pure motor or
sensory symptoms (5,6). As lacunes now can be
identified in vivo using CT scans, it has been recog-
nized that often more complex neurological symp-
toms are associated with these lesions (7,8). In ad-
dition to these diverse structural changes, angiogra-
phical studies in the acute and chronic phase of an
ischemic stroke have shown a wide spectrum of vas-
cular lesions, which range from occlusion of the
major extracranial or intracerebral arteries to nor-
mal findings (9-12).
Studies of the patho-physiological changes, i.e.
cerebral blood flow (CBF) and the cerebral meta-
bolic rate for oxygen (CMRO,) in ischemic stroke
have recently become possible by use of positron
emission tomography (PET) (13-16). In patients
with larger cerebral infarcts, the studies have shown
440 VORSTRUP
that potentially viable brain tissue may be present
within the first days after a stroke, such regions be-
ing defined as low-flow areas having an increased
oxygen extraction ratio with a preserved CMRO,.
Patients studied later than Day 3 invariably showed
a low oxygen extraction associated with either a re-
duced, a normal or even an increased CBF (16). In
the chronic phase, CBF and CMRO, are coupled
again, i.e. CBF reflects as in normal tissue the meta-
bolic demand (15). Rougemont et al. (17) mea-
sured CBF and CMRO, by PET in 7 patients with
symptoms of a moderate to severe stroke, having a
normal CT scan or a lacunar infarct. They were
studied in the acute or in the chronic phase. In none
of these cases were significant changes of CBF or
CMRO, observed. Thus, the patho-anatomical and
-physiological findings differ considerably in pa-
tients with cerebral ischemic stroke.
To study the time course of CBF changes in a ty-
pical stroke population to obtain a guidance for la-
ter series on therapeutic interventions, CBF was
measured in 12 consecutive patients with clinical
symptoms of an acute ischemic stroke. CBF was
studied by xenon- 133 inhalation and single-photon
emission computer tomography (SPECT) allowing
sequential measurements of CBF at much lower
cost than PET. In patients with larger infarcts our
studies show the same phases as the PET obser-
vations. In patients with smaller infarcts or lacunes
areas of slightly reduced flow located at a distance
of the lesion could be ascertained by the sequence
of CBF tomograms obtained in each patient.
Material and methods
Serial measurements were performed in a consecu-
tive series of 12 patients, suffering their first ische-
mic stroke with symptoms referable to the cerebral
hemispheres. The age limits were set from 20-75
years. All patients had a hemiparesis, and most had
other focal symptoms as well. The series comprised
4 women and 8 men, with an age ranging from 32-
66 years, mean 61 years. The first CBF studies were
performed as soon as possible after onset, in 11 pa-
tients by Day 2, and in one patient by Day 3 after
onset (acute phase). Subsequent CBF studies were
then performed 1-3 times within the first 2 weeks
(subacute phase) and repeated 2 and 6 months later
(chronic phase). A total of 112 CBF measurements
were performed, ranging from 8-12 studies in the
individual patient.
Cerebral vasoreactivity was tested in conjunction
with the CBF measurement in the acute phase fol-
lowing intravenous injections of aminophylline and
acetazolamide (DiamoxcR)). The effect of amino-
phylline is analoguous to a reduction of Pa CO,,
normally causing constriction of the cerebral vessels
and inducing slight hyperventilation (18,19). Dia-
moxR increases CBF by an effect corresponding to
inhalation of hypercapnic air. Both drugs leave
CMRO, unchanged (20,21). After the first baseline
measurement, 220 mg of aminophylline was inject-
ed over 5 min, and CBF was measured 10 min later.
Following this study, 1 gram of DiamoxR dissolved
in sterile water was injected over 30 s, and CBF
measured 20 min later. In the chronic phase at the
6-months follow-up vasoreactivity was tested 20
min after an intravenous injection of 1 g of Dia-
moxR.
X-ray transmission (CT) scanning was per-
formed within the first week after admission to ex-
clude a hematoma or a hemorrhagic infarct, and it
was repeated 8 weeks later for determination of the
definitive CT lesion. Angiography was performed
in 11 of the 12 patients by direct puncture of the
relevant internal carotid artery in 9 patients, 5-10
days (in average 7 days) after onset of the symp-
toms. In 2 patients the angiographical studies were
performed 4 and 6 weeks after onset; in one of
these patients (Case 2) 4-vessel angiography with a
cranial series was performed.
A clinical neurological examination was per-
formed in connection with every CBF measure-
ment. The clinical symptoms of paresis were graded
as: none, slight (25% reduction), moderate (50%
reduction), severe (movement on couch only) or
paralysis. One-step changes in these motor func-
tions in either arm, hand, leg or foot was considered
an improvement. The acute neurological findings
and the clinical course, the findings on the %week
CT scan, and the angiographical findings are pre-
sented in Table 1.
Only one patient had a history of previous neu-
rological symptoms. This patient (Case 12) had suf-
fered approximately 10 transient ischemic attacks
with paresthesia of the left hand (the same brain
areas subsequently involved in stroke). These
symptoms occurred unrelated to postural changes,
 CBF IN ISCHEMIC STROKE 441

Table 1
Patient Presenting clinical Clinical course of Size (in cc) and location Angiographical findmgs
age, sex signs motor disturbances of infarct on CT scan (time since infarction)
2 months 6 months
1 73M Global aphasia Unchanged Unchanged 26 L basal ganglia Not performed
R hemiparalysis, and insula
hemiparesthesia and DCA
homonymous hemianopia
2 32F global aphasia Unchanged Unchanged 45 L temporo-parietal Normal (4 weeks)
R paralysis of arm and (Aphasia region
moderate paresis of leg improved)
Facial palsy
3 72F Severe L hemiparesis Unchanged Unchanged 48 R temporo-parietal R ICA occluded
and hemiparesthesia region R MCA stenosed (7 days)
4 70 F L hemiparalysis Unchanged Improved 6 1 R temporo-parietal R intracranial ICA and
hemiparesthesia and region MCA severely stenosed
homonymous hemianopia Severe DCA (6 weeks)
5 60M Severe L hemiparesis Improved Unchanged 39 R temporo-parietal R ICA occluded
and hemiparesthesia region - DCA (10 days)
6 57M Global aphasia severe R Unchanged Unchanged 91 L temporo-parietal L ICA severely stenosed
hemiparesis, hemiparesthesia region Poor filling of MCA-
and homonymous hemianopia Severe DCA territory (9 days)
7 55 M Moderate L hemiparesis Improved Unchanged Normal Normal (6 days)
8 34 M L hemiparalysis Improved Unchanged 8 R middle part of pons Normal (5 days)
9 74M Slight L paresis of arm Improved Unchanged 25 R temporo-occipital Normal
and leg region - Severe DCA (6 days)
10 69 M Slight R hemiparesis Improved Unchanged 7 L parietal lobe Stenosis of L ICA
Moderate DCA Normal intracranial
filling (7 days)
I1 69F R paralysis of arm and Unchanged Improved 0.5 L internal capsule Diffuse arteriosclerotic
severe paresis of leg changes - Normal
intracranial filling
(7 days)
12 6 2 M Slight L paresis of arm Unchanged Improved 0.5 R internal capsule Normal
and leg Severe DCA (7 days)
. ,

Abbreviations: M = male, F = female. R = right, L = left. ICA = internal carotid artery, MCA = middle cerebral artery,
DCA = diffuse cortical atrophy

and had always subsided in less than 12 h. One pa-
tient had arterial hypertension, and one had cardiac
insufficiency, symptoms that in both cases were well
controlled by appropriate therapy. A third patient
(Case 2) had a congenital heart disease (see case
history).
The study was approved by the Copenhagen Dis-
trict Ethical Committee.
CBF measurement and CT scanning
For measurement of CBF, xenon-133 inhalation
was used combined with a single-photon emission
computer tomograph (Tomomatic 64, Medimatic
Inc., Hellerup, Denmark) previously described
(22). The detector device consists of 64 NaI crys-
tals, which are placed in 4 banks, and covered by
photomultiplier tubes to constitute a total of 192
detectors. These detectors are equipped with a sen-
sitive collimator. Measurements of the isotope con-
centration during the rotation of the detectors
yields the isotope uptake, distribution, and wash-
out from 3 slices of brain tissue simultaneously.
Each slice is 2 cm thick, and separated by an inter-
track distance of 2 cm. The resolution is 1.5-1.7 cm
in the horizontal plane and 2,O cm in the axial
plane, when measured as full width half maximum
using a line source. CBF is determined from a series
of isotope concentration pictures using a modified
filtered back-projection algorithm (23).
The patients were studied resting in a supine po-
sition in most cases with eyes closed. No speaking
was allowed during the studies. The patient’s head
442 VORSTRUP
was carefully positioned in the aperture of the in-
strument, so that the slices were oriented in parallel
to the orbitomeatal (OM) plane, and centered 1
cm, 5 cm, and 9 cm above it by use of an inflatable
cuff. Correct symmetrical positioning of the pa-
tient’s head could be ascertained by comparing the
localization of the low flow areas corresponding to
the petrous bones on the lowest slice (Slice 1).
End-tidal p C 0 , was measured before, during the
third minute, and immediately after the CBF study
using a spectroscopic infrared analyzer. Blood pres-
sure was measured by ausculation at the end of eve-
ry CBF measurement.
The CT scans were performed using either a
EM1 1010 Head Scanner or a Siemens Somatom.
The size of the hypodense lesion was calculated on
the CT scan performed 2 months after onset as the
product of the 2 largest diameters at a right angle to
each other, and the height of the lesion estimated
from the number of slices. For the 6 patients having
large infarcts, this value was then divided by 4 to
yield a more correct value for a spherical lesion.
Data analysis
Calculation of the average CBF in a given region
was performed by the computer after encircling the
region. This yielded the average flow value in the
region of interest (ROI), as well as the average flow
value in a symmetrically placed region in the con-
tralateral hemisphere. In this way, mean hemisphe-
ric flow values were obtained from Slices 2 and 3,
and in every ROI. Positioning of the ROI was per-
formed on the flow map from the acute study, and
CBF values from the similar ROIs were then calcu-
lated from all successive CBF studies. The size of
the ROI ranged from 10-36cm2.
When the 8-week CT scan showed large hypo-
dense lesions (6 patients), delineation of the region
of complete infarction on the flow map was made
by comparing the corresponding levels of the CT
scan and the CBF tomogram. This region was in all
+
cases largest on Slice 2 (OM 5 cm). Calculations
of CBF were also performed in the low-flow areas
surrounding the region corresponding to the com-
plete infarct, in the following referred to as the peri-
infarct areas. The size of these ROIs were in the
same order as descibed above, and the areas chosen
from either Slice 2 or 3. In the remaining patients,
the CT lesions were either small (less than 25cc) or
it was not possible to detect hypodense lesions in
the supratentorial brain tissue. In 4 of these cases,
rather extensive areas with slightly reduced flow
were observed in the early studies in the diseased
hemisphere. This region was found in either Slice
2 or 3, and the time courses of these flow
changes were analyzed.
Finally, CBF was calculated in the cerebellar
hemispheres in all 12 patients.
The degree of side-to-side asymmetry was evalu-
ated by an asymmetry index: the difference of CBF
in the symmetrical ROIs calculated as a percentage
of the higher CBF value.
Controls. For comparison, mean cerebral and re-
gional flow values were obtained from a normal
material of 9 elderly controls without symptoms of
cerebrovascular disease with a mean age of 63 years
(range from 57-69 years). Regional CBF values
were calculated from the territories of arterial sup-
ply, i.e. the territory of the anterior cerebral artery,
the cortical part of the middle cerebral artery and
the posterior cerebral artery (24). The mean values
of the left to right side difference as a percentage of
the highest side, ranged from 1.2% * 1.4% (SD)
(posterior cerebral territory) to 4.8% _+ 2.6% (cor-
tical part of the middle cerebral artery territory).
The size of these areas ranged from 15.7 cm2 to
27.2 cm2. Based on these results a significant
asymmetry was defined when the side-to-side diffe-
rence taken as a percentage of the higher value (the
asymmetry index) was equal to or exceeded 10%.
CBF was repeated 20 min after an intravenous dose
of 1 g DiamoxR. Mean CBF increased by 13% to
46%, in average by 31%. Calculation of the
changes in side-to-side asymmetry after DiamoxR
for the same vascular regions as descibed above was
performed, and enhancement of the side-to-side
asymmetry by 10% or more by DiamoxR indicated
a significant change in the flow pattern ( ~ ~ 0 . 0 5 ) .
To evaluate the time course changes of the de-
gree of side-to-side asymmetry for a ROI as used in
the 4 patients having small CT lesions and supeja-
cent low-flow areas, calculations were performed in
a normal material of 10 persons, (mean age 43
years) where CBF had been measured 3 times one
week apart. This material has previously been de-
scribed (24). Only 1 of the 10 controls showed a

COMPLETE INFARCTION
_/.-.
.-.
-
.. .*
8oi
1 3 5 7 9 11 13 16 17 19 2 6
days rnonlhr
Fig. 1. Six stroke cases with large infarcts on C
T scan.
The graphs show the time course of CBF changes in
the region corresponding to the infarcted area as evi-
denced by the 8-week CT scan. The values are ex-
pressed as percentages of the first CBF study. Hyper-
emia was not observed in the acute studies, but trans-
ition through a hyperemic phase was seen in 4 patients,
most likely due to lysis of an intracerebral embolus.
flow pattern where the flow in the ROI was lowest
in the same hemisphere during 3 consecutive mea-
surements one week apart. The values of the side-
to-side asymmetry for this patient were 2%, 2%,
and 6%.
ResuIts
CBF changes with large infarcts on CT (6 pa-
tients)
Time course changes. The 6 patients (cases 1-6)
with large infarcts (26-91 cc) on the %week CT
scan all showed even more extensive areas with de-
creased CBF in the acute phase, Days 1-3.
F m of these 6 patients (Cases 1-4) showed on
Days 5-9 a marked increase in CBF in regions
corresponding to the CT lesions. The increases
ranges from 59-108%, taken as a percentage of the
CBF value in the acute flow study (Fig. 1). In the
peri-infarct low-flow areas more modest increases
of about 30% were seen on Days 5-25 in 2 of the
patients, whereas the other 2 showed no such in-
creases (Fig. 2). Three patients were angiographed:
one study (on Day 9) showed an occlusion of the
middle cerebral artery (MCA). In 2 other studies
performed 4 and 6 weeks after onset, one patient
CBF IN ISCHEMIC STROKE 443
showed a moderate stenosis of the MCA, whereas
the other had a normal angiogram.
In the remaining 2 patients with large infarcts
(Cases 5,6), flow increases of about 20% were seen
in the regions of complete infarction in the studies
from Days 7-14. In the pen-infarct low-flow regions
in the same hemisphere, no increase was seen (Figs.
1,2). These patients were angiographed on Days 7
and 10, and both showed an occlusion of the ipsila-
teral internal carotid artery.
At the 2-month study, as well as in the 6-month
Time
from onset, follow-up, a large area with very low
flow values, much like the acute CBF studies, were
seen in all 6 patients. However, the low-flow areas
tended to be somewhat larger and more sharply
demarcated.
The time course changes in the non-affected
hemisphere were also studied. In the acute phase,
Days 7-10, 2 months and 6 months after onset the
mean hemispheric flow values were 56 f 5 ml/
100g/min (SD), 49 f 4 m1/100g/min (ASD), 53 f
3 m1/100g/min (ASD), and 56 f 5 mI/lOOg/min
(ASD), respectively. ASD indicates the standard de-
viation of the CBF change from the acute study. For
the ROI located symmetrically to the area of com-
plete infarction, the same values were 59 i 7 ml/
lOOg/min (SD), 52 f 5 m1/100g/min (ASD), 55 f
3 m1/100g/min (ASD), and 57 f 3 m1/100g/min
(ASD); a two-way analysis of variance followed by
a Dunnett test was applied to these data (26). Using
INCOMPLETE INFARCTlON
180
--
160
"1 1 3 5 7 9 11 13 15 17 19
*
2
I.
6 Time
aoys mO"lhl
Fig. 2. Six stroke cases with large infarcts on CT scan.
The graphs show the time course of CBF changes ex-
pressed as percentages of the value in the first CBF
study in the peri-infarct regions. Only moderate CBF
changes with increases of about 20% are seen.
444 VORSTRUP

COMPLETE INFARCTION INCOMPLETE INFARCTION

oj , 0’ I

B A D B A D

/:
!/. 504
Asymmetry

I*/

OJ I

B A D
Fig. 3. Vasoreactive studies in 6 stroke cases with large infarcts on C T scan, studied in the acute (Days 1-3) and
chronic (6 months) phase. B = baseline study, A = aminophylline, D = DiamoxR,all performed in the acute phase.
B, = baseline study and D, = DiamoxKtest performed in the chronic phase. Following aminophylline, a decrease
in side-to-side asymmetry was seen in all cases, both in the infarct and the peri-infarct tissue, whereas DiamoxRen-
hanced the asymmetry anew. In the chronic phase, a significant change in flow distribution was noted, in the peri-
infarct areas in one patient, but the finding of a CBF increase indicated a preserved vasodilational capacity.

the acute-flow study as reference value, a significant
CBF decrease in the ROI was observed at Days
7-10.
All 6 patients with large hemispheric CT lesions
showed a significant reduction of CBF in the con-
tralateral cerebellar hemisphere, present through-
out the whole study period.
Vasoreactive studies in the acute and chronicphases.
Aminophylline caused a reduction of CBF in both
hemispheres in all cases, but variable changes were
seen in the infarct and peri-infarct areas (Fig. 3).
The degree of side-to side asymmetry, however,
decreased in all cases. After DiamoxK, the flow
asymmetry worsened again. The cases who showed
a CBF increase in the ROI after aminophylline all
showed a CBF decrease after DiamoxR, and vice
versa. One patient showed improvement in motor
function after aminophylline. No changes were seen
in the clinical symptoms after DiamoxR.
The DiamoxR test was repeated at the 6-month
follow up. At this point, enhanced side-to-side as-
symmetry was seen in all cases in the area of com-
plete infarction. In the pen-infarct area, a signifi-
cant change in flow-distribution was seen in one
case (the asymmetry was enhanced from 28% at

Fig. 4. Stroke case with large CT lesion (Case 2). This patient suffered sudden onset of a severe right-sided hemi-
paresis and global aphasia. Corresponding levels of the 8-week CT scan and the flow maps are presented, placed 5
and 9 cm above the orbito-meatal (OM) plane (Slices 2 and 3, respectively). The scale ranges from 0 - 96 ml/
100g/min, and white indicates the highest flow value. The right hemisphere is oriented to the right, nose upward.
The CT scan shows a well-demarcated hypodense lesion the temporal lobe of the left hemisphere. No lesion was
seen above OM+7 cm. In the acute phase, the flow pictures showed a severely hypoperfused area present on both
Slices 2 and 3. Hyperemia in the area of infarction was seen on Day 7. In the chronic phase, the large low-flow
areas reappeared. A 4-vessel angiogram including cerebral series performed 6 weeks after onset was normal.
CBF IN ISCHEMIC STROKE 445
446 VORSTRUP

the baseline study to 40% after DiamoxR),indicat- lesions of the extracranial neck vessels, and normal fill-
ing a limited vasodilational capacity in this one case ing of the intracranial vessels which indicated that no
only. vascular obstructions were present any longer.

Case 2 history CBF changes with small or absent CT-lesions

This patient, a 32-year-old woman, suffered an acute
stroke with a right-sided facial palsy and a severe hemi-
paresis, slight hemiparesthesia, and global aphasia. She
had a congenital heart disease, a patent ductus arterio-
sus and coarctation of the aorta with pulmonary hyper-
tension. At admission, atrial fibrillation was present.
The patient received digoxin and diuretics, but she was
not being treated with anticoagulants. Embolism from
the heart was considered the pathogenetic factor in this
patient.
The acute CBF study performed on Day 1 showed a
large hypoperfused area confined to the territory of the
anterior part of the MCA (Fig. 4). This was taken as
evidence of embolism to the anterior branches of the
MCA. On Day 7, the CBF study focal hyperemia in the
previous low-flow area in Slice 2, suggesting lysis of the
embolus. CT showed a large infarct in Slice 2, but no
hypodensities were seen above the level of OM 7cm.
The late CBF studies performed approximately 2
and 6 months later showed large low-flow areas corre-
sponding to the CI-lesion. However, in rather wide pe-
ri-infarct regions which almost comprised the entire left
hemisphere, very low flow values were now present. In
this patient, the angiogram was performed 6 weeks af-
ter the acute episode. This showed no atherosclerotic
Table 2
Time course of the degree of side-to-side asymmetry ("6)in patients
with small infarcts.
~ ~ ~ ~~~ ~~
(6 patients).
Time course changes. Two patients in this group
(Cases 7,8) had no C T lesions in the cerebral hem-
ispheres; in one case, however, the repeated CT
scan revealed a hemi pontine infarct explaining the
severe hemiparesis. Significant focal changes or
persistence of an asymmetrical flow pattern were
not seen in these 2 patients. Both had normal an-
giograms.
Four patients showed minor CT lesions. Cases 9
and 10 showed small subcortical infarcts of about 6
and 2 cc on the 8-week CT scan. In one case, the
infarct was seen only on the C T level 3 cm above
the orbito-meatal plane, and the infarct was there-
fore between the planes normally used for CBF
studies. However, large regions with a marginally
+ decreased flow were seen in the ipsilateral hemis-
phere in both 2 and 3 in the early as well as in the
chronic phase (Table 2). The angiogram showed
normal intracranical filling. Case 10 had a subcorti-
cal infarct in the parietal region. In this region, and
the overlying areas variable CBF changes were seen
in the early phase, but in the chronic phase a low
flow was seen. Also this case had a normal intra-
cerebral angiogram.
Two patients (Cases 11,12) showed lacunar in-

farcts of less than 0.5 cc on the CT scan, in both

Flow no. Case 9 Case 10 Case I1 Case 12
cases located to the internal capsule (5-6 cm above
1 Acute phase 8/12 2 12 8
the OM plane). The CBF map showed no abnor-
2 Aminophylline test 15 2 I I
3 Diamox" test 7 6 6 13 mality at the site of the CT lesion or in the immedi-
4 Subacute phase - 6 -13 12 11 ate surroundings. Surprisingly, a slight CBF reduc-
5 Subacute phase -10 0 8 tion was seen in both patients in the posterior part
6 Subacute phase 10 6 5 of the parietal lobe in Slices 2 and 3 ipsilateral to
7 2 months 9 8 6 9
8 6 months I 9 6 18
the lacunar infarction, but a significant side-to-side
9 Diamox" test 5 10 4 17 asymmetry exceeding the 10% level was obtained
Positive values = decreased, negative values = increased flow in dis-
only in some of the individual studies. However, the
eased side compared to the symmetrical region in the contralateral repeated CBF studies allowed for the recognition of
hemisphere. this slight but persisting flow reduction (Table 2).
Fig. 5. Stroke case with lacunar infarct (Case 11). This patient suffered a severe right-sided hemiparesis. Corre-
sponding levels of the 8-week Cr scan and the flow maps are presented, both placed 5 and 9 cm above the orbi-
tomeatal plane, representing Slices, 2 and 3, respectively.The scale ranges from 0 - 96 ml/lOOg/min. The CT scan
showed a small lacunar infarct in the left hemisphere. The CBF studies showed a rather large area with slightly re-
duced flow in the posterior part of Slice 3 as demarcated by the dotted line. This moderate change was seen
throughout all successive CBF measurements. The carotid angiogram was normal.
CBF IN ISCHEMIC STROKE 447
448 VORSTRUP
The angiographical studies were normal.
A transient, crossed cerebellar diaschisis was seen
on Days 4 and 9 in the 2 patients with lacunar in-
farcts, i.e. the cerebellar side-to-side asymmetry ex-
ceeded 10%. In the remaining, no significant chan-
ges were seen.
Vasoreactive studies in the acute and chronic phase.
In this group of patients CBF changed in all regions
such that there was no change in the flow pattern.
Case 11 history
This patient, a 69-year-old woman suffered an acute
stroke with a moderate, right-sided facial palsy and
hemiparesis combined with aphasia. In the following
days, the aphasic symptoms gradually declined, but at
the same time her arm became paralysed. For several
years, she had received treatment for arterial hyperten-
sion, to keep her blood pressure at about 200/ 100
mmHg.
The first CBF measurement was performed on Day
2 after onset of the symptoms and showed a quite large
area with reduced flow (12% less than the symmetrical
region) in the parietal lobe of Slice 3. This moderate
change was seen throughout all successive CBF mea-
surements (Fig. 5). The CT scan showed a lacunar in-
+ +
farct on Slices OM 5cm and OM 6cm. The angio-
gram showed no lesions.
Discussion
Xenon-133 inhalation and SPECT is well suited to
the performance of serial CBF studies. Measure-
ments may be repeated with an interval of 20 rnin,
which is needed to allow the remaining activity
from the previous study to decline to acceptable le-
vels. Analyses of simulated data have shown, that
despite the varying and unknown lambda values
within the brain tissues, the use of a fixed lambda
only causes minor errors in the final CBF values
(27). On the other hand, it is recognized that the
CBF values on the low-flow areas are severly influ-
enced by scattered radiation, an effect depending
on the size and the localization of the low-flow area
as well as on the difference between the flow values
in the hypoperfused area and its surroundings.
Studies have been performed on water filled cylin-
drical perspex phantoms containing cold spots of
varying diameters within hot surroundings. These
analyses showed, that in the centre of a cold spot
having a diameter of 5 cm, an activity of about 55%
of the surrounding could be measured (28).
CBF changes with large CBF lesions
Despite the methodological limitations, focal low-
flow areas were easily identified in all stroke pa-
tients who had large CT lesions, but the lowest CBF
value recorded even in the most severe case with the
largest low-flow area was about 20-26 m1/100g/
min, i.e. in regions where true flow values could be
expected to be much lower.
The vasomotor function in the acute stage. Fol-
lowing arninophylline, all 6 patients showed a re-
duced reactivity in the infarct and peri-infarct areas
as indicated by the decrease in side-to-side
asymmetry. Following DiamoxK, redistribution of
flow in favour of the healthy hemisphere was seen
with aggravation of the side-to-side asymmetry. The
changes in the absolute CBF values during the va-
soactive tests did not permit conclusions. It should
be noted, however, that DiamoxR apparently only
abolished the vasoconstriction produced by amino-
phylline. Nevertheless, these findings demonstrate,
that although the vasoreactivity is reduced in the is-
chemic tissues, CBF may be modified by vasoactive
substances. Except for one patient, no changes in
the clinical symptoms were noted. Larger series in
this subgroup, including long term clinical evalua-
tion, will be needed to identify the drugs that in-
crease CBF in the ischemic areas and thereby im-
prove the clinical outcome.
The repeated CBF studies in the early phase
showed a transition through a hyperemic phase with
an increase of CBF in the area of complete infarc-
tion in 4 of the 6 cases on Days 5-9. The increase in
CBF ranges from 59-108%, taken as a percentage
of the value measured at the first CBF study (Fig.1).
Although neither acute nor serial angiographical
studies were possible in the present series, it is most
likely that the CBF increase was due to fragmenta-
tion and lysis of an embolus with reopening of the
previously occluded vessel, although metabolic der-
angements may also occur. This has been described
by several authors (14-16, 29-31), in some relating
the angiographical findings directly to the changes
in CBF (29,32), Dissolution of an intracerebral em-
bolus with luxury perfusion may occur within the
first days after a stroke, but delayed recanalisation
is not an uncommon finding (10,31,33). In the

present study, none of our patients showed hyper-
emia in the acute stage, i.e. Days 1-3 (Fig.1). In the
2 patients without hyperemia it could be assumed
that the occlusion of the ICA combined with poor
collateral flow caused the cerebral symptoms. A
permanent occlusion of a major intracerebral artery
could also explain these findings as cerebral emboli-
sation is a common finding in patients with ICA oc-
clusions (34).
Analyses of the time course changes of CBF in
the contralateral hemisphere revealed a significant
decrease in CBF in the mirror locus to the infarct on
Days 7-10. Hcledt-Rasmussen et a1 originally de-
scribed a decreased CBF in the hemisphere contral-
ateral to the ischemic lesion, and denoted this find-
ing a transneuronal depression (35). The pheno-
menon was later described by other authors (15).
Other factors such as increased intracranial pressure
owing to brain edema after stroke should, however,
also be considered as pathogenetic factors (36).
The finding in the present study of a delayed onset,
of its transient nature, and the circumscript location
of the lesion may support both theories.
In the chronic phase 2 and 6 months after onset,
the flow maps were most like the early flow studies
with no tendency toward diminuition of the low-
flow area. In some cases, even enlargement and
sharper demarcation were seen. Thus, an increase
in CBF does not accompany the gradual improve-
ment noted in most stroke patients. The same con-
clusions were reached by Demeurisse et a1 (37) who
measured CBF 15, 30 and 60 days after the acute
symptoms using xenon- 133 inhalation with station-
ary detectors. Neither mean hemispheric nor re-
gional CBF increased contrasting the clinical im-
provement seen in most cases.
The DiamoxR test was repeated at the 6-month
follow-up. At this point, the collateral flow capacity
may be assessed from the regional CBF increase. In
regions with a compromised collateral capacity, the
perfusion pressure will decrease but initially the
cerebral autoregulation maintains CBF by a com-
pensatory dilatation of the arterioles. However,
when additional vasodilatory stress (DiamoxR) is
applied in such regions, the flow response is res-
tricted compared with unaffected regions. In the
present study, an increase in CBF after DiamoxK
was observed in all 6 patients in the peri-infarct
areas, although one patient did have a significant
29
CBF IN ISCHEMIC STROKE 449
enhancement of the side-to-side asymmetry indi-
cating a restricted collateral capacity. Thus, in these
6 cases we may consider the low-flow CT negative
areas to represent a final “functional” lesion with
CBF adjusted to the lowered metabolic demands.
The pathogenetic lesion causing these low-flow
areas still remains mostly unsettled, but some au-
thors have suggested, that deafferentiation (discon-
nection) due to undercutting of the nerve fibers
could be an explanation (35,38). Also, it could be
assumed, that in the regions surrounding the in-
farcted area direct injury to the most vulnerable tis-
sue component - the nerve cells - could occur
(39,40). In either case the reduced flow level repre-
sents an adaptation to reduced metabolic demands.
It could have been speculated, that the low-flow
areas in the CT intact regions represented a state of
“chronic penumbra” (41). Astrup and co-workers
(42) coined the term the “acute ischemic penumb-
ra” to the blood flow range within which the neu-
ronal tissue is structurally intact maintaining the ion
homeostasis, but functionally inactive. The tolerat-
ed time limits before irreversible ischemic damage
occurs have yet to be assessed, although experi-
mental studies have been performed to extend
these results to the chronic phase (43,44). It is diffi-
cult, however, to conceive that the very narrow
range of perfusion pressure required to keep CBF
just below the lower limit for neuronal function and
above the threshold for irreversible ischemic da-
mage can be maintained over longer periods in
man. A preserved vasodilatational capacity as indi-
cated by the DiamoxR test in the present study
seems to exclude that the peri-infarct low-flow re-
gions observed in our cases reflects a chronic pen-
umbra.
CBF changes with small or no CT lesions
Two patients in the present series showed lacunar
lesions on the CT scan, and another 2 showed only
smaller CT lesions. These patients showed quite
large regions with slight reduced CBF, but the de-
gree of side-to-side asymmetry only occasionally
reached significant levels. However, as this CBF
pattern was seen in nearly all the CBF studies, both
in the early and in the chronic phase, this was taken
as suggestive evidence of an intracerebral diaschisis.
Recently, Rougemont et al (17) reported single
measurements of CBF and CMRO, in 7 patients
450 VORSTRUP
who had clinical syndromes of lacunar infarction. In
agreement with our study, the calculation of side-
to-side asymmetry in the single cases was insignifi-
cant. Similar PET results were reported by Wise et
al. in 6 cases having smaller CT lesions (16). Taka-
no et a1 (45) using stationary equipment and intra-
carotid injection of xenon-133 reported on 10 pa-
tients with smaller deep ischemic lesions, and in
that study, all the patients showed a decrease of
CBF in a small circumscribed area in the ipsilateral
hemisphere, mostly confined to the central sulcus.
They suggested, that the reduction in flow in the re-
mote cortical regions ipsilateral to the deep seated
infarction was due to damage of the neuronal fibers
passing through the lesion. As thrombosis of the
small perforating striatal arteries is assumed to be
the pathogenetic factor in patients with lacunes
(46), widespread ischemic damage is not likely to
have caused the remote, large, low-flow areas ob-
served in our patients. Likewise, our results of the
DiamoxR test gave no evidence of a restricted
collateral capacity at the 6-month follow-up, sup-
porting that undercutting of neuronal pathways
causing intracerebral diaschisis is a likely explana-
tion. This phenomenon corresponds to the crossed
cerebellar diaschisis originally described by Baron
et al (47), later reconfirmed by other series of
stroke patients including our own study of the time
course of the cerebellar flow changes (48,49).
Conclusion
Although CBF measurement only covers part of
the pathophysiological changes after a stroke, the
present study emphasizes, that quite differing flow
patterns are seen in ischemic stroke. With the
knowledge derived from studies on the cerebral
metabolism, this technique may be used to monitor
changes in CBF, in particular in the acute stage
where treatment with Ca-blockers, hemodilution,
or thrombolytic agents etc. could be evaluated.
Acknowledgements
This study was supported by grants from the Danish Medical Re-
search Council and Danish Hospital Foundation for Medical Re-
search, Region of Copenhagen, The Faroe Islands and Greenland.
Pia Poulsen is acknowledged for her excellent secretarial assist-
ance.
References
1. Cronqvist S, Brismar J. Kellin K, S~derstremEC. Computer
assisted axial tomography in cerebrovascular lesions. Acta
Radiol Diag 1975:16135-145.
2. Messina AV. Cranial computerized tomography. A radiolo-
gic-pathologic correlation. Arch Neurol 1977: 34602-608.
3. Kistler JP, Hochberg FH, Brooks BR, Richardson EP,
Schnur J. Computerized axial tomography: Clinicopathologic
correlation. Neurology (Minneap) 1975: 25201-209.
4. Pullicino P, Nelson RF, Kendall BE, Marshall J. Small deep
infarcts diagnosed on computed tomography. Neurology
1980:30: 1090-1096.
5. Fisher CM, Curry HB. Pure motor hemiplegia of vascular
origin. Arch Neurol 1965: 1 3 130-140.
6. Fisher CM. Pure sensory stroke involving face, arm and leg.
Neurology (Minneap) 1965: 1576-80.
7. Nelson RF, Pullicino P, Kendall BE, Marshall J. Computed
tomography in patients presenting with lacunar syndromes.
Stroke 1980: 11:256-261.
8. Weisberg LA. Lacunar infarcts. Clinical and computed tomo-
graphic correlations. Arch Neurol 1982:3937-40.
9. Alter M, Kieffer S, Resch J, Ansari K. Cerebral infarction:
Clinical and angiographic correlations. Neurology (Minneap)
1972: 22590-602.
10. Cronqvist S. Regional cerebral blood flow and angiography
in apoplexy. Acta Radiol Diag 1968: 7521-534.
11. Marshall J. Angiography in the investigation of ischemic ep-
isodes in the territory of the internal carotid artery. Lancet
1971: ~ 7 1 9 - 7 2 1 .
12. Paulson OB. Cerebral Apoplexy: Pathogenesis, pathophy-
siology and therapy as illustrated by regional blood flow mea-
surements in the brain. Stroke 1971:2:327-360.
13. Ackerman RH, Alpert NM, Correia JA, Finklestein S, Davis
SM, Kelley RE, Donnan G A , Dalton JG, Taveras JM. Posi-
tron imaging in ischemic stroke disease. Ann Neurol
1 9 8 4 : ( ~ ~ p pII5) ~ 1 2 6 - ~ 1 3 0 .
14. Kuhl DE, Phelps ME, Kowell AP, Metter EJ, Selin C , Win-
ter J. Effects of stroke on local cerebral metabolism and per-
fusion: Mapping by emission computed tomography of FDG-
18 and NH, -13. Ann Neurol 1980:8:47-60.
15. Lenzi GL, Frackowiak RSJ, Jones T. Cerebral oxygen meta-
bolism and blood flow in human cerebral ischemic infarction.
J Cereb Blood Flow Metabol 1982:2321-335.
16. Wise RJS, Bernardi S, Frackowiak RSJ, Legg NJ, Jones T.
Serial observations on the pathophysiology of acute stroke.
The transition from ischaemia to infarction as reflected in re-
gional oxygen extraction. Brain 1983:106: 197-222.
17. Rougemont D, Baron JC, Lebrun-Grandie P, Bousser MG,
Soisson I, Comar D. Debit sanguin cerebral et extraction
d'oxygene dans les hemiplegies lacunaires. etude semi-quan-
titative par I'oxygkne- 15 et la tomographie d'emission. Path
Biol 1982:30:295-302.
18. Gottstein U, Bernsmeier A, Sebening H, Steiner K. Der Ein-
fluss von Euphyllin auf die Hirndurchblutung und den zere-
bralen Stoffwechsel des Menschen. Med Klinik
1961:561589-1592.
19. Skinhej E, Paulson OB. The mechanism of action of amino-
phylline upon cerebral vascular disorders. Acta Neurol Scand
1Y70:46 129-140.

20. Gottstein U, Paulson OB. The effect of intracarotid amino-
phylline infusion on the cerebral circulation. Stroke
1972:3560-565.
21. Vorstrup S, Henriksen L, Paulson OB. Effect of acetarola-
mide on cerebral blood flow and cerebral metabolic rate for
oxygen. J Clin Invest 1984:7 4 1634-1639.
22. Stokely EM, Sveinsdottir E, Lassen NA, Rommer P. A single
photon dynamic computer assisted tomograph (DCAT) for
imaging brain function in multiple cross sections. J Comput
Assist Tomogr 1980:4230-240.
23. Celsis P, Goldman T , Henriksen L, Lassen NA. A method
for calculating regional cerebral blood flow from emission
computed tomography of inert gas concentrations. J Comput
Assist Tomogr 198 1:641-645.
24. Berman SA, Hayman LA, Hinck VC. Correlation of C T
cerebral vascular territories with function: 3. Middle cerebral
Artery. AJR 1984:142 1035-1040.
25. Vorstrup S, Lassen NA, Henriksen L, Haase J , Lindewald H,
Boysen G , Paulson OB. C B F before and after extracranial-
intracranial bypass surgery in patients with ischemic cerebro-
vascular disease studied with xenon-133 inahalation tomo-
graphy. Stroke 1985:16:616-626.
26. Zar JH. Biostatistical analyses. Prentice Hall Inc., Englewood
Cliffs, New Jersey, 1974.
27. Holm S, Vorstrup S, Lassen NA. Tomographic cerebral
blood flow calculation. J Cereb Blood Flow Metahol
1985:5615-618.
28. Holm S, Vorstrup S, Lassen NA, Paulson OB. Physical fac-
tors affecting calculated cerebral blood flow values in hypo-
perfused areas in single photon emission computerized tomo-
graphy. In: Hartman A, Hoyer S eds. Cerebral Blood Flow
and Metabolism Measurement. Heidelherg: Springer-Verlag,
1985:234-237.
2Y. Paulson OB. Regional cerehral blood flow in apoplexy due lo
occlusion of the middle cerebral artery. Neurology (Min-
neap) 1970:20:63-77.
30. Skyhej Olsen T , Lassen NA. A dynamic concept of middle
cerebral artery occlusion and cerebral infarction in the acute
state based on interpreting severe hyperemia as a sign of em-
holic migration. Stroke 1984:15458-468.
31. Uemura K,Goto K, Ishii K, Ito Z Hen R, Kawakami H. Se-
quential changes of regional cerebral circulation in cerebral
infarction. Neuroradiol 1978:16228-232.
32. Hsedt-Rasmussen K, Skinhej E, Paulson OB, Ewald J, Bjer-
rum JK, Fahrenkrug A, Lassen NA. Regional cerebral blood
flow in acute apoplexy. The "luxury perfusion syndrome" of
brain tissue. Arch. Neurology (Chic) 1967:17271-281.
33. Dalal PM, Shah PM, Aiyar RR. Angiographic study of cere-
bral embolism. Lancet 1965:2358-361.
34. Castaigne P, Lhermitte F, Gautier JC, Escourelle R, De-
rousse C. lnternal carotid artery occlusion. Brain
1970:93231-258.
29'
CBF IN ISCHEMIC STROKE 451
35. Heedt-Rasmussen K, Skinhej E. Transneural depression of
the cerebral hemispheric metabolism in man. Acta Neurol
Scand 1964:4@.41-46.
36. Ropper AH, Shafran B. Brain edema after stroke. Clinical
syndrome and intracranial pressure. Arch Neurol
1984:41:26-29.
37. Demeurisse G, Verhas M, Capon A, Paternot J. Lack of evo-
lution of the cerebral blood flow during clinical recovery of
stroke. Stroke 1983.1477-81.
38. Slater R, Reivich M, Goldberg H, Banka R, Greenberg J. Di-
aschisis with cerebral infarction. Stroke 1977:8686-690.
40. Metter EJ, Maziotta JC, Itabashi HH, Mankovich NJ, Phelps
ME, Kuhl DE. Comparison of glucose metabolism, x-ray CT,
and postmortem data in a patient with multipe infarcts. Neu-
rology 1985:35 1695-1701.
41. Lassen NA, Vorstrup S. Ischemic penumbra results in incom-
plete infarction: Is the sleeping beauty dead? Stroke
1984:15755.
42. Astrup J, Symon L, Branston NM, Lassen NA. Cortical
evoked potential and extracellular K+ and Hf at critical levels
of brain ischemia. Stroke 1977:851-57.
43. Mies G,Auer LM, Ebhardt G, Traupe H, Heiss WD. Flow
and neuronal density in tissue surrounding chronic infarction.
Stroke 1983:1422-27.
44. Strong AJ, Tomlinson BE, Venables GS, Gibson G, Hardy
JA. The cortical ischaemic penumbra associated with occlu-
sion of the middle cerebral artery in the cat. 2. Studies of his-
topathology, water content, and in vitro neurotransmitter up-
take. J Cereb Blood Flow Metabol 1983:397-108.
45. Takano T,Kimura K, Nakamura M, Fukunago R, Kusunoki
M, Etani H, Matsumoto M, Yoneda S, Abe. Effect of small
deep hemispheric infarction on the ipsilateral cortical blood
flow in man. Stroke 1985.1664-69.
46. Fisher CM. Lacunes: Small, deep cerebral infarcts. Neurol-
ogy 1965:15:774-784.
47. Baron JC, Bousser MG, Comar D, Castaigne DP. "Crossed
cerebellar diaschisis" in human supratentorial infarction.
Trans Am Neurol Assoc 1980:105459-461.
48. Meneghetti G , Vorstrup S, Mickey B, Lindewald H, Lassen
NA. Crossed cerebellar diaschisis in ischemic stroke: A study
of regional cerebral blood flow by 133-Xe inhalation and
single photon emission computerized tomography. J Cereb
Blood Flow Metabol 1984:4235-240.
Address
Olaf 6. Paulson, M.D.
Dept. of Neurology
Rigshospitalet
9,Blegdamsvej
DK-2100Copenhagen
Denmark