534
Local Cerebral Blood Flow Following
Transient Cerebral Ischemia
I. Onset of Impaired Reperfusion within the First Hour Following
Global Ischemia
CALVIN L. MILLER, P H . D . , D O U G L A S G.
KARL ALEXANDER, M.B.,
LAMPARD, P H . D . ,
A N D WILLIAM A.
BROWN, P H . D .
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
SUMMARY Using the hydrogen clearance technique, local cerebral blood flow (LCBF) in 22 dogs was estimated at 6 parietal sites prior to and following 5 min of total global ischemia. Ischemia was immediately
followed by an initial reactive hyperemia during which the electrocorticogram (ECoG) usually began to
recover, and within the first 30 min, most of the LCBPs decreased to subnormal values. This onset of bypoperfusion was accompanied by a concomitant decrease in ECoG activity. Two animals that maintained normal
local perfusion after the initial hyperemia recovered ECoG activity quickly. These results suggest that the subsequent poor reperfusion was caused by an increased microvascular resistance rather than by blood aggregates,
increased blood viscosity, or a variety of other mechanisms which have been proposed. Increased vascular tonus
was, at least, partly responsible for the increased vascular resistance. This report supports the hypothesis that
impaired reperfusion (which occurs some time after an initial hyperemia) may be responsible for ultimate
neuronal death, rather than the period of global ischemic hypoxia per se.
Stroke, Vol 11, No 5, 1980
SEVERAL REPORTS have indicated that the
vulnerability of the central nervous system to brief
periods of ischemic hypoxia may be due to impaired
recirculation, rather than to immediate cell death.
Hossmann, Lechtape-Gruter, and Hossmann1
demonstrated the return of neuronal functions (EEG,
pyramidal response, evoked potentials) in cats after 30
or 60 minutes of complete cerebral ischemia.
Hossmann and Kleihues2 concluded that normothermic nerve cells can withstand 60 min of complete
ischemia and that revival of nerve cells is hindered
(presumably) by deficits in recirculation. Additional
support to this hypothesis has been given by Yatsu,
Lee, and Liao3 and others4'6 who have shown that
cerebral energy metabolism is not always irreversibly
damaged and is not primarily vulnerable to ischemia.
The cause of impaired cerebral microcirculation
following ischemia has been attributed to many factors. Ames and co-workers8 concluded that the "noreflow phenomenon" was due to increased blood
viscosity and endothelial and perivascular glial cell
swelling. Fischer and Ames 7 '' noted that perfusion
deficits appeared in a scattered pattern and did not
appear to be caused by platelet microemboli or pial
arterial spasm. Chiang et al." implicated swollen endothelial "blebs" which they claimed obstructed the
capillary lumen. Little, Kerr and Sundt10 related impaired microcirculation to the compression of
capillaries by perivascular glial swelling, but also
noted that severe neuronal injury preceded the microvascular obstruction. Conversely, Cuypers and
Matakas11 concluded that no-reflow was produced by
blood aggregates which obstructed larger vessels
rather than capillaries. Wade et al.12 proposed that the
no-reflow state is due to an increase in potassium concentration in brain extracellular fluid which causes a
pronounced contraction of vascular smooth muscle.
Klatzo13 concluded that arterial spasm was responsible for the no-reflow phenomenon. Hekmatpanah14
demonstrated capillary microthrombi and red cell
aggregates during circulatory standstill. Microcirculatory blockage has also been attributed to consumption coagulopathy18 and to an "ischemic tissueblood interface" reaction that reduces blood flow by
an unspecified mechanism."
Some investigations have shown that cerebral
ischemia may be initially followed by hyperemia, but
with a subsequent hypoperfusion within one hour.1' 17
During the initial post-ischemic hyperemia, the
cerebral mean requirement of oxygen (CMROi) may
be temporarily reduced, but the CMRO2 may subsequently increase beyond control values.18 Perfusion
deficits, accompanied by an elevated CMROi, result
in a prolonged cerebral hypoxia that may contribute
to post-ischemic cerebral pathogenesis.19 The etiology
of perfusion deficits has still not been resolved which
emphasizes the need to investigate post-ischemic
cerebral circulation so that methods of restoring adequate blood flow may be devised.
The purpose of this investigation was to closely
follow the changes in local cerebral blood flow after
total ischemia and to determine whether there was any
evidence to support (or reject) the hypothesis that
cerebral recovery may be hindered by poor reperfusion rather than by immediate death of hypoxic nerve
cells.
From the Department of Electrical Engineering, Monash University, Clayton, Victoria, Australia, 3168.
Dr. Alexander is from the Intensive Care Unit, Royal Children's
Hospital, Parkville, Victoria, Australia, 3053.
Methods
Anesthesia was induced in adult, healthy mongrel
dogs, 15 to 25 kg, with intravenous thiopentone
sodium (Intraval, May & Baker) and maintained with
CBF AFTER TIA. I. IMPAIRED REPERFUSION/A/iV/er et al.
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
intravenous pentobarbitone sodium (Sagatal, May &
Baker) on demand. Following endotracheal intubation
and insertion of femoral arterial and venous catheters
for pressure monitoring and blood sampling, artificial
ventilation was maintained with air and oxygen (40%
Oj), 20 breaths per minute, at a tidal volume which
kept the arterial Pco, at about the normal canine level
of 38 mm Hg;10 this Paco, produced an end-expired
CO, (FEco2) typically between 4.5% and 5% as
measured by a Godart Capnograph. A Radiometer
BMS Mk2 blood microsystem was used for serial
measurement of arterial pH and blood gases. Body
temperature was maintained at 38°C and was continuously monitored using a rectal thermistor. The
ventilation system was computer controlled as
described by Lampard, Coles, and Brown."
A Cardiovascular Instrument cardiac output computer, which employs the thermodilution technique,
was used for cardiac output estimates. A thermistor
was firmly fixed in the pulmonary artery with a purse
string suture; bleeding from the artery was slight and
transient.
Local Cerebral Blood Flow Measurement
The hydrogen clearance technique"' M was used for
estimating local cerebral blood flow. Hydrogen
clearance curves were essentially monoexponential
with a small second order effect. Only the first order
time constant was calculated. From the first order rate
function, A0"kt, k was calculated from the monoexponential segment by dividing the half-time (time in
seconds taken for the function to decay to half its initial value, Ao) by 0.693; multiplication of k by 60
seconds/minute and 100 gm tissue yielded the local
cerebral blood flow in units of ml blood/minute/100
gm tissue.
Platinum wire of 0.3 mm was silver soldered to 1
mm Belling-Lee connectors and insulated with Epoxylite 6001 thermosetting epoxy. Insulation (0.5 mm)
was removed from the electrode tip. This insulation
technique yields an electrode which is impermeable to
physiological fluids or saline for at least 24 hours. Absolute insulation of the electrode is a prerequisite in
obtaining inter-electrode independence and smooth,
noiseless graphs. If the electrode resistance dropped
below 20 megaohms, considerable "cross-talk" occurred among electrodes, which resulted in unstable
baselines, and it became difficult to properly bias the
electrodes to +300 millivolts. Adequate electrode insulation against ultimate leakage is a commonly encountered problem, especially in chronic preparations,
as more fully discussed by Donaldson."
Following reflection of the frontal and temporal
muscles along the external sagittal crest, the parietal
bone was perforated lateral to the sagittal sinus with a
No. 4 dental burr, care being exercised to avoid abrasion or laceration of the dura. In the event of bleeding
from the bone the hole was stoppered with bone wax;
otherwise, bleeding was slight or unnoticeable.
Because of extreme differences among mongrel skulls,
it was not possible to stereotaxically implant the elec-
535
trodes. However, electrodes were standardly placed in
parietal lateral, ectolateral, and middle suprasylvian
areas at depths ranging from 1 to 8 mm. A rigid
micromanipulator was used to prevent electrode
movement and minimize tissue trauma while the electrodes were fixed to the skull with Surgical Simplex
bone cement. Consistent blood flow measurements
from acutely implanted platinum electrodes depends
on the immobilization of the electrode after insertion
and on the absence of subdural bleeding. An electrode
depth of at least 1 mm was necessary to avoid inaccurate flow estimates resulting from the rapid diffusion of hydrogen from the surface of the cortex."
Local cerebral blood flow estimates were measured
from 5 or (usually) 6 electrode sites in each dog.
Tissue damage during electrode insertion may
preclude reliable washout curves until healing has occurred,"- " but a series of experiments in this
laboratory showed LCBF estimates to be constant under controlled conditions."
Stainless steel Sherman bone screws were located
parieto-occipitally as the reference electrodes for the
platinum electrodes. The epidural electrocorticogram
(ECoG) was measured from bilateral Sherman bone
screws inserted into frontal bone. Bone screw depth
was carefully matched to skull thickness to avoid compression of the cortex with the screw tip. Epidural intracranial pressure (ICP) was measured using a
stainless steel adapter inserted into parietal bone.
Care was taken to ensure that there was no bleeding
from the bone that would clot the lumen of the ICP
adapter.
Mean arterial pressure (MAP), systolic pressure,
diastolic pressure, pulse rate, FEco2, I.C.P., rectal
temperature, and the mean rectified voltage (MRV) of
the ECoG were recorded continuously on a Brush 8
channel recorder.
Technique Tor Producing Total Cerebral Ischemia
Following a left thoracotomy and removal of the
fourth rib, the brachiocephalic and left subclavian
arteries were exposed and cotton thread was passed
under each artery near its origin. A length of Silastic
tubing (3 mm o.d.) was passed twice around the
descending aorta to form a loop. This loop allowed
complete constriction of the aorta when the ends of
the Silastic tubing were drawn. Dissection around the
heart and ascending arteries was cautiously performed
to avoid any trauma to the vagosympathetic trunk,
cardiovagal branches and other nerves in the area.
Because of the collateral circulation to the dog
brain arising from the descending aorta, complete
cerebral ischemia cannot be achieved by clamping
only the brachiocephalic and left subclavian arteries.
However, if the descending aorta is also occluded, it is
necessary to provide a pressure vent (or "shunt") for
the left ventricle to prevent blood congestion in the
heart and lungs. To provide this shunt, a catheter was
passed from the femoral artery into the aortic arch.
During the ischemia, this catheter allowed diversion of
536
STROKE
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
arterial blood to a one liter Viaflex bag (Travenol
Labs).
To produce the total cerebral ischemia, the brachiocephalic and left subclavian arteries were clamped at
their origin. Assuming 86 ml blood per kg body
weight," heparin was injected intravenously at a dose
of one unit per ml blood. Immediately after the
heparin injection, the intra-aortic catheter was connected to a one-liter plastic blood collection bag
(Travenol Viaflex) and the descending aorta was constricted around the intra-aortic catheter by drawing
the ends of the Silastic loop. Blood returned from the
blood collection reservoir via the femoral vein. It is
important that the catheters have the widest possible
lumen so that blood can flow rapidly and freely into
and out of the blood bag.
The ECoG became flat (isoelectric) typically within
30 seconds and this flattening served as the starting
point of the 5 minute period of ischemia. At the end of
5 minutes of isoelectric ECoG, the constricture
around the descending aorta was removed, the clamps
were removed from the ascending arteries, and the
intra-aortic catheter was withdrawn. Arterial Pco 3
quickly returned to pre-ischemia levels (36-40 mm
Hg), but arterial pH was predictably low. Arterial pH
was corrected by intravenous infusion of 4.2% sodium
bicarbonate at 1 ml/minute until pHa was within a
normal range (7.36 to 7.44); usually, this required
about 60 to 80 ml of 4.2% bicarbonate. Bicarbonate
was not infused until the local blood flows were
demonstrated to be reduced below control values.
Simultaneous LCBF estimates were recorded from 6
electrode sites in each dog, except for 3 animals which
had 5 electrode sites. This resulted in a total of 117
electrode sites for the 20 dogs used in these experiments.
Results
The mean LCBF was 43.0 ml/min/100 gm tissue,
with a standard error of ±4.8 ml/min/100 gm and a
range of 19 to 198 ml/min/100 gm. The wide range
and relatively large standard error emphasize the
heterogeneity of flow around cortical and subcortical
electrode sites. For ease in comparative analysis each
post-ischemic LCBF is expressed as a percentage of its
pre-ischemic control value of 100%. Figure 1 shows
the local cerebral blood flow from 6 electrode sites in
one representative animal following the 5 minute
period of ischemia; the changes indicated are typically
seen after the ischemia. Most of the local flows were
initially greatly increased, although occasionally some
electrode sites demonstrated an initial decrease below
control flow. It is apparent that the initial flow increase is followed by a decrease which results in most
of the LCBF's falling substantially below the control.
All LCBF estimates for the first hour following
ischemia are shown in figure 2. Except for 9 sites, all
of the electrode sites recorded an increase in LCBF
after ischemia. For these 9 sites, it is not possible to
determine whether the flows were initially increased
and then began to diminish, or whether theflowswere
VOL 11, No
5, SEPTEMBER-OCTOBER
1980
LOCAL CBF FOLLOWING
5 MINUTE S
ISCHEMIA
6001
21 JULY 78
NORMAL PaC02
FIGURE I. Changes in local CBF at 6 electrode sites for
one representative dog. Each symbol represents an electrode
site. PacOt was normal within 5 minutes after the ischemia
and the pH, was corrected to 7.4 after one hour. Note that
most flows remain subnormal during subsequent hours.
initially decreased below control when cranial circulation was restored. After the ischemic episode, it took
at least 5 minutes to reposition the animal, rebias the
electrodes and resume estimates of LCBF's.
After the initial hyperemia, bloodflowsappeared to
decrease relatively gradually (fig. 2), but occasionally
a precipitous drop in blood flow was noted. This was
exemplified by a sudden decrease in the wash-out of
hydrogen and the introduction of a significant second
order component in the decay curve (fig. 3). This acute
decrease in flow is a difficult event to capture using a
gas clearance method. In one animal, there was a
simultaneous acute decrease inflowat 3 of 6 electrode
sites. This was accompanied by a marked decrease in
the ECoG mean rectified voltage.
Intracranial pressure (ICP) was never elevated during the period of flow decrease. During the ischemia,
ICP always decreased, indicating the partial emptying
of blood from the cerebral venules and a reduction in
cerebral blood volume. Following restoration of the
circulation, ICP usually increased sharply (not more
than 40 mm Hg) and then recovered to pre-ischemic
levels of about 3 mm Hg within 10 minutes (fig. 4). In
subsequent hours, the ICP never showed a tendency to
increase.
537
CBF AFTER TIA. I. IMPAIRED REPERFUSI0N/A/;7/er et al.
TIME AFTER
ISCHEMIA
10 mln I—I—»0
50
100
200
300
Igniter'
20 mln
. -V. t .•!* .
50
200
30 mln
(go
FIGURE 2. Line graph of all LCBF estimates
from all animals showing the progression from
|
high initial flows to severely depressed flows
WO within the first hour.
50
iO min
S0m,n
50
I
i
50
no
50
100
LOCAL CBF'
I
i
200
300
I
200
300
I ' ' ' ' I
SO mln
200
300
X OF CONTROL
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
action consisted of adequate ventilation and restoration of blood pressure and cardiac output. When pHa
was corrected (typically 1 to 1.5 hours after ischemia),
sodium bicarbonate (4.2%) was infused at a rate not
exceeding 0.05 ml/min/kg. The correction of pHa did
not cause any significant increase or decrease in
LCBF.
Cardiac output (CO) was not measured in all
animals but measurements in 3 animals showed an immediate post-ischemic increase in CO that returned to
nearly normal within the first hour. In most animals,
the increase in cardiac output was accompanied by a
transient hypertension (150 to 190 mm Hg systolic
BP) that decreased to no less than 120 mm Hg systolic
After restoration of circulation, end-expired CO2
was elevated, but returned to normal within 5 to 10
minutes. The post-ischemic arterial pH was between
7.25 and 7.3, and was not corrected with sodium bicarbonate during the first hour post-ischemia. It was
believed best to avoid the complication of introducing
another variable to the many physiological alterations
that were happening during and immediately after the
ischemia. Bicarbonate may have varying effects
depending on the rate of intravenous infusion and the
rate of change of the cerebrospinal fluid pH. Also, it
was important to initially observe the behavior of the
post-ischemic brain blood flow without interference
from other therapies or drugs. The only resuscitative
ACUTE DECREASE OF
POST-ISCHEMIC LOCAL CBF
FIRST ORDER DECAY: 92-5 ml/min/lOOgm
SECOND ORDER DECAY: I6'5
10
MINUTES
/5
ml/min/lOOgm
20
FIGURE 3. Sudden introduction of second-order component in the hydrogen wash-out curve,
implying an acute decrease in blood flow at that electrode site.
STROKE
538
LOCAL CBF AND THE ECoG
FOLLOWING ISCHEMIA
300
i
250
o
a:
h
21 JULY 78
1ICO
38mmHg
200
o
o
o
150
o
o
start pHa
correct/or
•
100
o
o
•
1
1 I
50
I
j
I
i
i
ISCHEMIA
J
-I
>-//
•
—i
i
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
5
i
6
7
HOURS
A0
20
0
1
2
3
5
i
6
HOURS
7
40
O. | 20
\
0
-in
1
0
1
2
3
4
5
1
6
HOURS
—H
7
FIGURE 4. Slight recovery of the ECoG during hyperemia,
and the subsequent decrease in ECoG following the decrease
in LCBFs.
BP. The difference between systolic and diastolic
pressures was at least 30 mm Hg. Heart rate was normal or slightly increased. These cardiovascular
measurements demonstrated that reduced LCBFs
were not due to cardiovascular impairment. This also
emphasizes that the technique for producing the
ischemia preserves heart function. It is important to
have a healthy heart after the ischemia so that changes
in LCBF can be confidently ascribed to the cerebral
vasculature rather than to a failing heart and a milieu
of supportive drugs.
The electrocorticogram (ECoG) is a sensitive indicator of the decrease of LCBFs following the initial
hyperemia. During the initial post-ischemic hyperemia, the ECoG mean rectified voltage (MRV) began
to gradually recover, but then there was a subsequent
decrease concomitant with the reduction in LCBFs
(fig. 4). This effect was noted in 14 of the 20 animals.
In the remaining 6 dogs, the ECoG MRV remained
nearly isoelectric during the hyperemia. Even though
the local flows remained depressed, most of the
VOL 11, No 5, SEPTEMBER-OCTOBER
1980
animals gradually recovered control levels of ECoG
MRV within 8 hours following ischemia. Recovery of
electrocorticographic activity was defined as the
return of ECoG MRV to approximately 100% of the
control (pre-ischemic) levels. In 2 animals, the ECoG
MRV recovered quickly to control levels and the
LCBFs did not decrease below control after the initial
hyperemia. These 2 dogs were apparently not as
vulnerable to ischemia as the other 20 dogs and,
therefore, have not been included in figure 2.
During ischemia, the pupils were fully dilated.
When circulation was re-established, the pupils
recovered normal diameter (about 2 mm) within 10
minutes. Mild and transient convulsions and tremors
were observed during and after the ischemia; these
events typically subsided within the subsequent hour.
Because of the ischemic insult, the animals did not
require any further anesthetic. The animals had
typically been anesthetized for 6 to 8 hours prior to
ischemia.
Discussion
Five minutes of ischemia was chosen as this period
is reported to be the threshold for irreversible brain
damage in animals and man.10"" Other studies have
shown that animals may survive 12 to 15 minutes of
ischemia"" and have an apparent clinical recovery
and normal EEG, but these periods are long in comparison to the shorter periods compatible with permanent damage. Although several studies have shown
that energy metabolism returns quickly after
ischemia,4'8 resumption of cerebral metabolism does
not imply recovery of cerebral function." In animals it
is difficult to assess clinical recovery in terms of subtle
changes in fine motor skills, personality, and intellect.
Definite histological evidence of brain damage may
be accompanied by a normal scalp EEG and apparent
clinical recovery."'" In the present experiments, we
were interested in the initial changes in cerebral blood
flow that may contribute to neuronal death. These
results show that 5 minutes of total cerebral ischemia
in dogs caused an initial reactive hyperemia that was
succeeded by a decrease in local cerebral blood flows
to subnormal values. Thirty minutes after ischemia,
most of the flows had fallen to subnormal levels (figs.
1, 2). Not all the electrode sites recorded an initial
hyperemia. Within the first 10 minutes, 9 of 92 electrode sites recorded subnormal blood flows (fig. 2). It
is not possible to determine whether these electrode
sites were initially hyperemic and then decreased
before resumption of recording, or whetherflowswere
depressed upon restoration of arterial flow.
The ECoG was a sensitive monitor of the hyperemia
and the subsequent decrease in local CBFs. The postischemic rise in the ECoG activity (during hyperemia)
was followed by a diminution of ECoG activity that
was concomitant with the LCBF reduction to subnormal levels (fig. 4). This decrease in LCBFs was seen in
20 of 22 dogs. The remaining 2 animals displayed an
initial hyperemia that was followed by LCBFs returning to approximately normal values; the ECoG mean
CBF AFTER TIA. I. IMPAIRED REPERFUSION/A////er et al.
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
rectified voltage (MRV) recovered quickly to control
levels within 2 hours in these animals.
The decrease in LCBF could be a normal response
to reduced oxygen demand by the brain; a reduced
CMROj would be expected to result in lower LCBFs.
Hossmann et al.40 correlated CMRO, with postischemic recovery in cats. Post-ischemic CMRO2 was
always reduced within the first hour after ischemia,
but then increased to normal levels. If the O2 uptake
remained low, the animals had no electrophysiological recovery but if Oa uptake returned to
normal, animals showed good recovery. Yatsu et al.8
concluded that the energy metabolism was not
primarily vulnerable to ischemia and that the
machinery for energy synthesis remained intact
following 5 minutes of ischemia. In fact, rather than
metabolism being depressed after 5 minutes of
ischemia, mitochondrial ATP synthesis was greater
than control. Schutz et al.5 also reported brain mitochondrial function to be intact after ischemia.
Nemoto18 reported that after 16 minutes of ischemia
in cats, the CMRO, was not different from control
CMRO, within the first 30 minutes after ischemia;
over the subsequent 2 hours, CMRO2 increased by 2fold.
Catecholamine release during ischemia has been
shown to increase CMRO2, a condition that may lead
to a secondary hypoxia if oxygen availability is not
proportionately increased.18
Post-ischemic brain hypermetabolism may incite
neuron damage." After global ischemia in dogs,
Snyder et al.17 noted a reduced CMRO2 and CBF
coincident with an abnormally low cerebral venous
oxygen content, which indicated that a hypoxic state
existed. They negate "the idea that post-ischemic
reduction in CBF is simply a matching of CBF to
reduced cerebral metabolic demands (reduced
CMRO,)". Lang et al.41 also concluded that postischemic reduced CBF is not a normal response to
reduced CMRO2; decreased oxygen consumption was
the result of metabolic derangement and reduced oxygen availability.
The present experiments did not measure CMRO2,
but we believe that it is unlikely that diminished CBF
after ischemia is a normal consequence of a noninjurious reduced CMROj. Because of the possibility
of normal or elevated O2 demand, reduced CBF may
represent a post-ischemic hypoxia that may initiate
nerve cell death. The reduced local bloodflows(within
30 minutes post-ischemia) seen in these experiments
are most likely causing an oxygen debt that is being
reflected by the concomitant decrease in ECoG activity (fig. 4).
These results indicate that a general "no-reflow"
phenomenon was not present following the ischemic
insult, but that gross impairment of local reperfusion
occurred some time after the resumption of arterial
flow. This is in contrast to other investigations which
have reported that vessels develop blockage during (or
immediately subsequent to) the ischemia."-7> '• "• "
Hence it appears that the impairment in reperfusion
occurs well after the end of ischemia. The initial rise in
539
ECoG activity was followed by a decrease that coincided with the decrease in local CBF. It is tempting to
speculate that the ECoG would have recovered
quickly if the post-ischemic LCBFs could have been
maintained at normal or supranormal levels.
The hyperemia may be explained by vasodilation
due to tissue acidosis from lactic acid and carbon
dioxide, or by a slight increase in extracellular
potassium." Some of the increases in local blood flow
were remarkable. For example, one electrode site
went from a control flow of 35 ml/min/100 gm to a
postischemic flow of 230 ml/min/100 gm. Figure 2
shows that 27flowswere initially greater than 400% of
control.
It is unlikely that transient brain swelling associated
with hyperemia was responsible for a subsequent artifactual decrease of LCBF around each electrode site.
In previous experiments from this laboratory," hypercapnea was never followed by low LCBF when the
animals were returned to normocapnea; similarly,
when animals were hyperventilated,28 a return to normocapnea was never associated with LCBF changes
from previous control values.
Other investigators have reported a reactive
hyperemia following either total or focal ischemia.
Cuypers and Matakas11 found that sustained
hyperemia (either spontaneous or sympathomimetically induced) was beneficial in preventing
post-ischemic no-reflow and intracranial hypertension. Osburne and Halsey44 reported that postischemic hyperemia in gerbils was associated with ultimate clinical and EEG recovery; regional blood
flows demonstrated a reactive hyperemia that later
declined to subnormal levels. Hossmann et al.1
reported that post-ischemic hyperemia was a prerequisite for the recovery of brain electrical activity; their
results also demonstrated a transient phase of
hyperemia that was followed by generally subnormal
cerebral blood flow." Safar et al." showed that
therapeutic measures aimed at improving the cerebral
circulation resulted in eventual clinical recovery
following 12 minutes of cardiac arrest in dogs. Our
results suggest, that initial hyperemia may be
beneficial in restoring the ECoG, and only when the
hyperemic phase ended, did the ECoG demonstrate an
abrupt decrease in activity (fig. 4).
Conversely, Mchedlishvili et al.48 concluded that
post-ischemic hyperemia is a contributory factor in
the development of brain edema. Heiss et al.47 also
concluded that reactive hyperemia in focal (middle
cerebral artery occlusion) ischemia appeared to potentiate cerebral edema. Other investigations have also
implicated hyperemia as a harmful consequence of
ischemia4860 but some of these involved prolonged
focal ischemia and/or induced arterial hypertension;
therefore, comparisons with transient ischemia and
reactive hyperemia (without arterial hypertension)
may not be valid and should be interpreted cautiously.
When comparing these results against those from
other laboratories, it is important to recognize the
differences in experimental design and duration of
ischemia. While the conclusions drawn by other in-
540
STROKE
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
vestigators are usually convincing, the methodologies
are often so different as to make comparisons difficult.
This is an important consideration and is further
emphasized by Molinari and Laurent.61
The results of these experiments do not support
some hypotheses presented by other investigators concerning the etiology of "no-reflow." Obstruction of
capillaries by swollen astroglia,*' " • " blood
aggregates and thrombi,11' M microthrombi," swollen
erythrocytes" and increased blood viscosity8 have
been implicated as major antecedents of "no-reflow."
Wade et al.ia reported that high concentration of extracellular potassium and consequent vascular constriction may account for "no-reflow." However, if
any of these effects were significant contributors to the
derangement of local CBF, then it would be expected
that these would be manifest by initial low flows. It
may be possible to mount an argument that would explain how one of the above hypotheses may have a
delayed effect after a period of hyperemia. In all our
experiments, ischemia was immediately followed by
hyperemia. Even at those sites which displayed
depressed initial flows, LCBFs were between 52% and
89% of control (fig. 2). This is an "impaired reflow"
rather than a "no-reflow" and implies vascular
narrowing rather than occlusion. Only at later times
(30 to 60 minutes) did the LCBFs become so seriously
decreased as to constitute in some instances a nearly
"zero" flow.
It is uncertain whether the hyperemia per se was
responsible for the subsequent decrease in LCBFs.
While some reports have concluded that excess reperfusion may contribute to brain edema,47 none of the
animals in the present experiments demonstrated
raised ICP after the hyperemic phase. Local edema
(e.g. perivascular glial swelling) may have developed,
but if this were a significant effect, then the ICP would
be expected to become elevated because most of the
LCBFs were subnormal, and widespread local edema
(affecting nearly all the localflows)would be expected
to show itself by raised ICP. Hyperemia may have
caused extravasation, rupture of weak endothelia, or
sloughing of intraluminal "blebs." Chiang et al.9 discovered diffuse intraluminal "blebs" which were more
frequent 30 minutes after ischemia. However, Fischer
et al.56 have reinvestigated the "bleb" hypothesis and
have concluded that "blebs" may have been artifacts
caused by an improper histological perfusion-fixation
technique. Also, Little et al.10 found only minor
capillary changes even after hours of ischemia and
concluded that these changes did not appear to be
producing significant obstruction.
As discussed in Part II, the local vasculature was
still highly reactive to increased arterial Pco, after the
ischemia. Of 70 electrode sites, 14 sites did not show
an increase in blood flow during hypercarbia; the
remaining 56 sites indicated a significant vasodilatation during hypercarbia. These increases in LCBF indicate that the post-ischemic vasculature was not
generally paralytic and was not blocked by various
cellular aggregates or fragments. The most likely explanation for the reduced blood flows is an increase in
cerebral vascular resistance, i.e. a reduction in the
VOL 11, No
5, SEPTEMBER-OCTOBER
1980
caliber of the microvessels. It is not certain whether
the resistance is increased by delayed perivascular
glial swelling" or by a sudden increased vessel tonus."
The precipitous reduction in flow at some electrode
sites (fig. 3) suggests that increased tonus must be a
contributing factor. This sudden flow decrease would
be more likely due to smooth muscle contraction
rather than to a sudden swelling of many endothelial
or perivascular glial cells.
An important consideration in these results is the
barbiturate anesthesia, which has been shown to have
protective effects in cerebral hypoxia and on ICP."
Anesthesia was kept at a relatively light level by giving
small doses of pentobarbitone only when the eyelid
reflex reappeared. Even though pentobarbitone may
reduce the vulnerability of brain cells to hypoxia, 4 to
5 minutes of complete global ischemia is sufficient to
cause motor and behavioral disabilities or death.80- "• M
Five minutes of total ischemia and isoelectric ECoG
was enough to seriously derange the local CBF's in the
present experiments.
It is highly unlikely that any heparin would have
been in the cerebral vasculature during ischemia. If
thrombi form during ischemia11- "• " heparin may inhibit thrombogenesis. Heparin was added after the
brachiocephalic and left subclavian arteries were
clamped and immediately prior to bypass from the
aorta to the external reservoir. After the ischemia,
heparinized blood may be beneficial by preventing
secondary clotting in areas of lowflowor by dissolving
thrombi, but others have shown7 that heparin did not
appear to improve post-ischemic recirculation.
These results show that relatively short periods of
global ischemia (but sufficient to incur permanent
damage) are followed by generally increased LCBFs.
This initial hyperemia may be accompanied by incipient ECoG activity, but when the LCBFs begin to
decrease to subnormal levels, ECoG activity becomes
markedly depressed. Two animals that maintained
normal LCBFs after the initial hyperemia recovered
ECoG activity quickly. Results suggest that subsequent hypoperfusion was due to microvascular
narrowing and that increased vascular tonus was at
least partly responsible for decrease vessel caliber.
This report supports the hypothesis that the brain may
not suffer severe neuronal death during ischemia, but
that recovery may be hindered by microcirculatory
deficits.
Acknowledgment
This investigation was funded by a grant from the National
Health and Medical Research Council of Australia.
References
1. Hossmann K-A, Lechtape-Gruter H, Hossmann V: The role of
the cerebral blood flow for the recovery of the brain after
prolonged ischemia. Z Neurol 204: 281-299, 1973
2. Hossmann K-A, Kleihues P: Reversibility of ischemic brain
damage. Arch Neurol 29: 375-384, 1973
3. Yatsu FM, Lee LW, Liao CL: Energy metabolism during brain
ischemia: stability during reversible and irreversible damage.
Stroke 6: 678-683, 1975
4. Siesjo BK, Ljunggren B: Cerebral energy reserves after
prolonged hypoxia and ischemia. Arch Neurol 29: 400-407,
CBF AFTER TIA. I. IMPAIRED REPERFUSION/A//7/er et al.
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
1973
5. Schutz H, Silverstein PR, Vapolahti M: Brain mitochondrial
function after ischemia and hypoxia. I. Ischemia induced by increased intracranial pressure. II. Normotensive systemic
hypoxemia. Arch Neurol 29: 408-419, 1973
6. Ames A, Wright RL, Kowada M et al: Cerebral ischemia. II.
The no-reflow phenomenon. Am J Pathol 52: 437-453, 1968
7. Fischer EG, Ames A: Studies on mechanisms of impairment of
cerebral circulation following ischemia: effect of hemodilution
and perfusion pressure. Stroke 3: 538-542, 1972
8. Fischer EG: Impaired perfusion following cerebrovascular
stasis: a review. Arch Neurol 29: 361-366, 1973
9. Chiang J, Kowada M, Ames A et al: Cerebral Ischemia. III.
Vascular changes. Am J Pathol 52: 455-476, 1968
10. Little JR, Kerr FWL, Sundt TM: Microcirculatory obstruction
in focal cerebral ischemia: an electron microscopic investigation
in monkeys. Stroke 7: 25-30, 1976
11. Cuypers J, Matakas F: The effect of post ischemic hyperemia
on intracranial pressure and the no-reflow phenomenon. Acta
Neuropathol 29: 73-84, 1974
12. Wade JG, Amtorp O, Sorensen SC: No-reflow state following
cerebral ischemia. Arch Neurol 32: 381-384, 1975
13. Klatzo I: Experimental studies on brain ischemia. Fourth
Danubia Symposium on Neuropathology, May, 1973 (Vienna)
14. Hekmatpanah J: Cerebral blood flow dynamics in hypotension
and cardiac arrest. Neurology (Minneap) 23: 174-180, 1973
15. Hossmann K-A, Hossmann V: Coagulopathy following experimental cerebral ischemia. Stroke 8: 249-254, 1977
16. Hallenbeck JM: Prevention of postischemic impairment of
microvascular perfusion. Neurology (Minneap) 27: 3-10, 1977
17. Snyder JV, Nemoto EM, Carrol RG, et al: Global ischemia in
dogs: intracranial pressures, brain blood flow and metabolism.
Stroke 6: 21-27, 1975
18. Nemoto EM: Pathogenesis of cerebral ischemia-anoxia. Crit
Care Mcd 6: 203-214, 1978
19. Levy DE, Duffy TE: Cerebral metabolism during transient
ischemia and recovery in the gerbil. J Neurochem 28: 63-70,
1977
20. Singer RB, Hernandez T: Acid-base balance. In Altman PL,
Dittmer DS, (eds) Respiration and Circulation Handbook,
Federation of American Societies for Experimental Biology,
Bethesda, Md, 1971. Section 86, p 225
21. Lampard DG, Coles JR, Brown WA: Electronic digital computer control of ventilation and anaesthesia. Anaes Int Care 1:
382-392, 1973
22. Aukland K, Bower BF, Berliner RW: Measurement of local
blood flow with hydrogen gas. Circ Res 14: 164-187, 1964
23. Fieschi C, Bozzao L, Agnoli A: Regional clearance of hydrogen
as a measure of cerebral blood flow. Acta Neurol Scand 41
(Suppl 14): 46-52, 1965
24. Donaldson PEK: The encapsulation of microelectronic devices
for long term surgical implantation. IEEE Trans Bio-mcd Engr
23: 281-285, 1976
25. Aukland K: Hydrogen polarography in measurement of local
blood flow: theoretical and empirical basis. Acta Neurol Scand
41 (Suppl 14): 42-45, 1965
26. Haining JL, Turner MD, Pantall RM: Measurement of the
local cerebral blood flow in the unanesthetized rat using a
hydrogen clearance method. Circ Res 23: 313-324, 1968
27. Meyer JS, Fukuuchi Y, Kanda Y et al: Regional cerebral blood
flow measured by intracarotid injection of hydrogen. Neurology
(Minneap) 22: 571-584, 1972
28. Miller CL, Alexander K, Lampard DG et al: The effect of carbon dioxide on local cerebral blood flow during surface
hypothermia in dogs. Cryobiol 16: 240-258, 1979
29. Reynolds M: Blood volumes of vertebrates. In Altman PL,
Dittmer DS, (eds) Respiration and Circulation Handbook.
Federation of American Societies for Experimental Biology,
Bethesda, Md, 1971. Section 147, p 281
30. Markovitz J, Archibald J, Downie HG: The mechanical heart:
cardiac arrest. In Experimental Surgery. Fifth edition.
Williams and Wilkins, Baltimore, Md, 1964, Chapter 13, p 394
31. Brierly JB, Meldrum BS, Brown AW: The threshold and
neuropathology of cerebral "anoxic-ischemic" cell change.
Arch Neurol 29: 367-374, 1973
541
32. Moossy J, Rao R, Nemoto EM, et al: Post-ischemic
histopathologic brain changes in treated and untreated animals.
Crit Care Med 4: 133, 1976
33. Dennis C, Kabat H: Behaviour of dogs after complete temporary arrest of the cephalic circulation. Proc Soc Exp Biol
Med 40: 559-561, 1939
34. Safar P, Stezoski W, Nemoto EM: Amelioration of brain
damage after 12 minutes cardiac arrest in dogs. Arch Neurol
33: 91-95, 1976
35. Wolin LR, Massopust LC, Taslitz N: Tolerance to arrest of
cerebral circulation in the rhesus monkey. Exp Neurol 30:
103-115, 1971
36. Nemoto EM, Bleyaert AL, Stezoski SW, et al: Global brain
ischemia: a reproducible monkey model. Stroke 8: 558-564,
1977
37. Nilsson B, Norberg K, Siesj5 BK: Biochemical events in
cerebral ischemia. Br J Anaesth 47: 751-760, 1975
38. Fisk GC, Wright JS, Hicks RG et al: The influence of duration
of circulatory arrest at 20°C on cerebral changes. Anaesth Int
Care 4: 126-234, 1976
39. Fisk GC, Wright JS, Turner BB et al: Cerebral effects of circulatory arrest at 20°C in the infant pig. Anaesth Int Care 2:
33-42, 1974
40. Hossmann K-A, Sakaki S, Katsuharu K: Cerebral uptake of
glucose and oxygen in the cat brain after prolonged ischemia.
Stroke 7: 301-305, 1976
41. Lang R, Zimmer R, Oberdorster G: Post-ischemic O,
availability and O, consumption of the isolated perfused brain
of the dog. PflQgers Arch 334: 103-113, 1972
42. Ginsberg MO, Myers RE: The topography of impaired microvascular perfusion in the primate brain following total circulatory arrest. Neurology (Minneap) 22: 998-1011, 1972
43. Cameron IR, Segal M B: The effect on pial arteriole diameter of
local changes in potassium concentration. Eur Neurol 6:
100-106, 1971-1972
44. Osburne RC, Halsey JH: Cerebral blood flow. A predictor of
recovery from ischemia in the gerbil. Arch Neurol 32:457-461,
1975
45. Takagi S, Cocito L, Hossmann K-A: Blood recirculation and
pharmacological responsiveness of the cerebral vasculature
following prolonged ischemia of cat brain. Stroke 8: 707-712,
1977
46. Mchedlishvili G, Kapuscinski A, Nikolaishvili L: Mechanisms
of post-ischemic brain edema: contribution of circulatory factors. Stroke 7: 410-416, 1976
47. Hciss WD, Hayakawa T, Waltz AG: Patterns of changes of
blood flow and relationship to infarction in experimental
cerebral ischemia. Stroke 7: 454-459, 1976
48. Zaren HA, Weinstein JD, Langfitt TW: Experimental ischemic
brain swelling. J Neurosurg 32: 227-235, 1970
49. Meining G, Reulen MJ, Hadjinimos A et al: Induction of filtration edema by extreme reduction of cerebrovascular resistance
associated with hypertension. Eur Neurol 87: 97-103, 1972
50. Dinsdale HB, Robertson DM: Experimental hypertensive
cerebrovascular disease. Excerpta Med 205: 193, 1969
51. Molinari GF, Laurent JP: A classification of experimental
models of brain ischemia. Stroke 7: 14-17, 1976
52. Miller JR, Myers RE: Neuropathology of systemic circulatory
arrest in adult monkeys. Neurology (Minneap) 22: 888-904,
1972
53. Hills CP: Ultrastructural changes in the capillary bed of the rat.
Am J Pathol 44: 531-551, 1964
54. Hirsch HM, Breuer HP, Kdnzel E et al: Uber die bildung von
thrombocytenaggregaten und die anderung des hSmatokrits
durch komplette gehirnischamie. Deutsch Z Nervenheilk 186:
58-66, 1964
55. Braasch D, Gossling G: Erythrocytendeformierung und
quellung durch plasmafaktoren nach schweren verbrennungen.
PflQgers Arch 289: 1-11, 1966
56. Fischer EG, Ames A, Hcdley-White ER et al: Reassessment of
cerebral capillary changes in acute global ischemia and their
relationship to the no-reflow phenomenon. Stroke 8: 36-39,
1977
57. Smith AL: Barbiturate protection in cerebral hypoxia.
Anesthesiol 47: 285-293, 1977
Local cerebral blood flow following transient cerebral ischemia. I. Onset of impaired
reperfusion within the first hour following global ischemia.
C L Miller, D G Lampard, K Alexander and W A Brown
Stroke. 1980;11:534-541
doi: 10.1161/01.STR.11.5.534
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1980 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/11/5/534
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click Request
Permissions in the middle column of the Web page under Services. Further information about this process is
available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Stroke is online at:
http://stroke.ahajournals.org//subscriptions/