Journal of Cerebral Blood Flow and Metabolism
5:207-213 © 1985 Raven Press, New York
Regional Cerebral Blood Volume and Hematocrit
Measured in Normal Human Volunteers by Single-Photon
Emission Computed Tomography
Fumihiko Sakai, Keiji Nakazawa, Yoshiaki Tazaki, Katsumi Ishii, Hidetada Hino,
Hisaka Igarashi, and Tadashi Kanda
Department of Medicine, School of Medicine, Kitasato University, Kanagawa-ken, Japan
Summary: Single-photon emiSSIOn computed tomog­
raphy (SPECT) was used for the measurement of regional
cerebral blood volume (CBV) and hematocrit (Hct) in
normal healthy human volunteers (mean age 30 ± 8
years), Regional cerebral red blood cell (RBC) volume
and plasma volume were determined separately and their
responses to carbon dioxide were investigated, Ten right­
handed healthy volunteers were the subjects studied,
SPECT scans were performed following intravenous in­
jection of the RBC tracer (99mTc-Iabeled RBC) and plasma
tracer (99ffiTc-labeled human serum albumin) with an in­
terval of 48 h, Regional cerebral Hct was calculated as
the regional ratio between RBC and plasma volumes and
then was used for calculating CBV, Mean regional CBV
in the resting state was 4,81 ± 0,37 ml/IOO g brain, sig­
nificantly greater in the left hemisphere compared with
the right by 3,8% (p < 0,01), Mean regional RBC volumes
(1.50 ± 0,09 ml/lOO g brain) were less than mean regional
plasma volumes (3,34 ± 0,28 ml/IOO g brain), and mean
regional cerebral Hcts were 31.3 ± 1.8%, which was 75,9
± 2,1% of the large-vessel Hct. During 5% CO2 inhala­
tion, increases in plasma volume (2.48 ± 0,82%/mmHg
Paco2) were significantly greater than for RBC volume
(1.46 ± 0.48%/mmHg Paco2), Consequently, the cerebral­
to-large-vessel Hct ratio was reduced to 72.4 ± 2,2%,
Results emphasize the importance of cerebral Hct for the
measurement of CBV and indicate that regional cerebral
Hcts are not constant when shifted from one physiolog­
ical state to another. Key Words: Cerebral blood
volume-Cerebral hematocrit-Single-photon emission
computed tomography,
Blood volume and hematocrit (Hct) are known to
be important in the regulation of cerebral circula­
tion, Tomographic measurements of circulating red
blood cells (RBC) in the brain, labeled either with
positron- or single-photon-emitting tracers, have
been used to display a three-dimensional distribu­
tion of cerebral blood volume (CBV) in humans
(Kuhl et aI., 1975; Grubb et aI., 1978; Phelps et aI.,
1979). When the method utilizes labeled RBC as a
tracer for the determination of whole-blood volume
in the brain, obtained CBV values must be cor­
rected for the cerebral Hct. Although it is known
that the brain Hct is lower than the large-vessel
Hct, until the present study no tomographic studies
in humans had been undertaken to measure regional
cerebral Hct values. Previous authors have used a
value of 0.85 for the cerebral-to-Iarge-vessel Hct
ratio, This was suggested by Phelps et al. (1973) as
the standard value based on the best available evi­
dence in the literature.
The purpose of the present investigation was to
determine whether cerebral Hct measured by three­
dimensional techniques in normal human volun­
teers give comparable values to previously reported
values obtained either by global or two-dimensional
techniques in humans (Larsen and Lassen, 1964;
Oldendorf et aI. , 1965) or by tissue-sampling tech­
niques in experimental animals (Gibson et aI., 1946;
Everett et aI., 1956; Sklar et aI. , 1968; Levin and
Ausman, 1969; Cremer et aI. , 1983). The question
of whether or not regional cerebral Hct might show
variations during different physiological conditions
was also examined. Single-photon emission com-
Received August 31, 1984; accepted November I, 1984,
Address correspondence and reprint requests to Dr. F Sakai
at Department of Medicine, School of Medicine, Kitasato Uni­
versity, 1-15-1 Kitasato, Sagamihara, Kanagawa-ken 228, Japan,
Abbreviations used: CBV, cerebral blood volume; Hct, he-
matocrit; HSA, human serum albumin: RBC, red blood cells:
SPECT, single-photon emission computed tomography,
207
F.
208
SAKAI ET AL.
puted tomography (SPECT) was used for measuring
regional CBV and Hct in humans. Both regional ce­
rebral RBC volume and plasma volume were deter­
mined in each subject, and any changes during 5%
CO2 inhalation were measured.
SUBJECTS AND METHOD
CBV and Hct were measured in 10 right-handed neu­
rologically normal healthy male volunteers aged 24-41
years (mean ± SD 30 ± 8 years). SPECT scans of the
head were performed on each subject following the intra­
venous injection of 99mTc-Iabeled RBC for measurement
of regional RBC volume and of 99mTc-Iabeled human
serum albumin (HSA) for measurement of regional ce­
rebral plasma volume. These two studies were performed
separately after an interval of 48 h. Serial scans were
made in each study in the resting state and repeated
during inhalation of 5% CO,. To maintain the subjects in
the same physiological con-ditions during the RBC and
the HSA study, they were allowed to lie quietly until the
blood pressure and the arterial CO, reached steady-state
condition. The first scan was made 15 min after the in­
travenous injection of the tracer. The second study was
performed 10 min after the completion of the first study.
The inhalation of 5% COc was started 2 min prior to and
was continued throughout the second study. At the time
of each scan, arterial blood samples were withdrawn
through an indwelling catheter for the measurement of
Paco2 and large-vessel Hct, and the blood pressure was
also measured.
after the injection of 99mTc-Iabeled HSA, revealed that
the mean label ratio for albumin in the circulating blood
was 96.1 ± 2.4% (n
5). The scan time for both RBC
and HSA studies was 10 min, and �2,500,000 counts was
collected in that interval. Arterial blood samples were
taken during each scan, and radioactivities for both RBC
and plasma were determined. Measured concentrations
of blood radioactivity were used for calculation of CBV
and brain Hct.
CBV was calculated by Eq. 1, which gives the ratio of
the 99mTc activity per weight of brain to the estimated
99mTc activity per milliliter of cerebral whole blood:
=
CBV
ex .
Hct
Hct
(I
-
ex
+
The SPECT system used for the present study was a
GE Maxi 400T camera connected to an Informatek Simis
3 computer system. The radius for the rotation of the
camera was set at 25 cm, and the data were collected
through 64 equal angular samplings. A low-energy, high­
resolution collimator was used. The blood volume images
were reconstructed in a series of horizontal sections of
the brain parallel to the anatomical line (Reid's line) for
every 1.2 cm from the vertex (Fig. 1). A filtered back
projection method was used, with a value of 0.15 for the
attenuation correction. The spatial resolution measured
in a skull phantom, expressed by the full width half-max­
imum of the profile, was 1.7 cm, and the cold lesion de­
tectability was a cylinder measuring 2 cm in diameter.
The brain scans were calibrated by matching to the scan
of a skull phantom filled with water and a pack of 99mTc_
labeled human blood of known radioactivity. For the RBC
volume study, a sample of the subject's RBC was labeled
with �20 mCi of 99mTc using stannous pyrophosphate as
a reducing agent, and was injected as the RBC tracer into
an antecubital vein. For plasma volume measurements,
20 mCi of 99mTc-Iabeled HSA was used as the plasma
tracer. The label ratio for RBC remained constant during
the first hour at 97.2 ± 1.2% (n
10). The radioactivity
of labeled albumin reached a plateau within 7 min fol­
lowing intravenous injection, then stayed constant there­
after for an interval of 45 min. A study by gel partition
chromatography for plasma samples, obtained 45 min
=
J
Cereh Blood Flow Metahol, Vol. 5, No.2, 1985
+
(I
-
ex .
Hct) .
Cplasma
Ch rainlHSAI
. Hc!)
ex
. CplasmalRBCI
--------------------------------------�---------. CrhclRBCI
. Crhc
. 100
(I)
where Cbrain is the activity per gram of brain determined
by scanning and the denominator is the activity of cere­
bral whole blood estimated from arterial blood counts.
Crbc and C plasm a are the activities per milliliter of RBC and
plasma in the sampled arterial blood, respectively; Hct is
the large-vessel arterial Hct measured from drawn arterial
blood samples. The constant ex represents the ratio of the
cerebral small-vessel Hct to the large-vessel Hct, and the
product of 100 converts CBV values to milliliters per 100
g brain.
The constant ex was determined by Eq. 2, which com­
pares the data between [99mTcjRBC and [99mTcjHSA
studies:
CbrainlRBCI
ex '
C ain
t
� �" = �__ ______ ____
______________
. Hct
. CrbclHSAI
+ (1
-
ex '
Hct)
. CplasmalHSAI
(2)
where the left term is the CBV value calculated by the
[99mTcjRBC study and the right term is the CBV value by
the [99mTcjHSA study. If Eq. 1 is applied to both
[99mTcjRBC and [99mTcjHSA studies for the calculation
of CBV, the results should give the same CBV value if
the correct ex is used in the equation. The coefficient of
ex can be calculated from Eq. 2, in which ex is the only
variable. The calculated ex value can then be put in Eq.
1 for the determination of corrected CBV, cerebral RBC
volume, and cerebral plasma volume. Figure 2 illustrates
the images for cerebral RBC volume, cerebral plasma
volume, calculated Hct, and CBV.
The coefficient of variation of hemispheric average
CBV values for repeated measurements of the same slice
was 2.6%, and the average variation from one measure­
ment to another was 0.12 ± 0.03 (mean ± SD) ml/100 g
brain. There was no tendency for increased accumulation
of radioactivity during the second scan.
Since all CO, runs were performed as second runs, we
examined the p ossibility of overestimations of the in­
creases in plasma volume during the second run owing
to an increase in unlabeled 99mTc that may have accu­
mulated in the vessel. This possibility, however, was con­
sidered to be unlikely for two reasons. First, the coeffi­
cient of variation for the repeated measurements with
plasma tracer was 2.9%, which was as good as that of
2.4% with the RBC tracer. Second, we measured the CO2
responsiveness in three additional subjects during hyper-
CEREBRAL BLOOD VOLUME AND HEMATOCRIT
209
FIG. 1. Cerebral blood volume images were reconstructed in a series of sections of the brain, parallel to the anatomical line
(Reid's line) for every 1.2 em from the vertex. Cerebral cortex with higher blood volume is shown with an increase in lightness,
and cerebral white matter with lower blood volume is shown with a decrease in lightness.
ventilation, in which the change (reduction) in the plasma
volume was also greater than that in the RBC volume
seen during 5% CO2 inhalation (see Results). The data
derived from the hyperventilation studies were not in­
cluded among the results of the present study for the
determination of CO2 responsiveness because of the dif­
ficulty in obtaining a constant Pac02 level over the IO-min
hyperventilation studies.
RESULTS
Steady-state values for regional CBV and
regional HCT
Table I summarizes mean hemispheric regional
CBV and Hct values among the 10 right-handed
healthy volunteers. The values represent the mean
of four horizontal slices between 4.2 and 7.8 cm
below the vertex. This includes large areas of ce­
rebral cortex, subcortex, white matter of the cen­
trum semiovale, gray matter of the basal ganglia,
thalamus, and brainstem structures above the mid­
brain. The mean of the regional CBV values was
4.81 ± 0.37 mlllOO g brain. The blood volume for
the left hemisphere (4.89 ± 0.37 mill00 g brain) was
significantly greater than for the right hemisphere
(4.71 ± 0.39 mlllOO g brain) by 3.8% at p < 0.01
(paired t test). The mean for regional RBC volumes
in the resting state was 1.50 ± 0.09 mlllOO g brain.
This was significantly smaller than the mean of the
plasma volumes, which was 3.34 ± 1.8 mlllOO g
brain. The mean cerebral Hct calculated from these
values was 31.3 ± 1.8%, and the mean ratio of the
cerebral to the large-vessel Hct was 75. 9 ± 2.1%.
Both RBC volume and plasma volume were greater
in the left hemisphere than in the right, but only the
difference in the plasma volume reached the level
of significance (p < 0.01).
Effect of hypercarbia on regional CBV
and MABP
Figure 3 illustrates the effect of 5% CO2 inhala­
tion on mean cerebral RBC volume and mean ce­
rebral plasma volume, and Table 2 summarizes
changes in regional CBY during 5% CO2 inhalation.
During 5% CO 2 inhalation, the volume was in­
creased in both RBC and plasma, but the increase
was greater for plasma volume than for RBC
volume. When expressed as the percentage in­
crease of volume per mmHg Paco2 increase, the
increase in the RBC volume was 1.45 ± 0.48 /1%/
mmHg Paco2' and the increase in the plasma
volume was 2.81 ± 1.04 /1%/mmHg Paco2. The dif­
ference in CO2 responsiveness between the RBC
volume and the plasma volume was significant at
the p < 0.01 level. Since changes in Paco2 and
MABP during 5% CO2 inhalation were identical for
both RBC and plasma studies for each individual,
the responses of CBY and Hct to the changes in
Paco2 were also calculated (Table 2). The increase
in CBY was 2.34 ± 0.94 /1%/mmHg Paco2, and the
J
Cereb Blood Flow Metahol. Vol. 5, No. 2, 1985
F.
210
SAKAI ET AL.
!-IG. 2. Upper two images are cerebral red blood cell (RBC)
volume measured by 99mTc-labeled RBC (upper left) and
plasma volume measured by 99mTc-labeled human serum al­
bumin (upper right). Cerebral hematocrit (Hct) (lower left)
was obtained from the ratio between RBC and plasma
volume and was used for the calculation of cerebral blood
volume (lower right). RBC volume and plasma volume are
�reater in c�rtical regions. shown with more lightness in this
figure. The Image for cerebral Hct did not show such a dis­
tinct difference between gray and white matter.
fiG. 3. Cerebral red blood cell (RBC) volume (left) and
plasma volume (right) were measured during steady state
(top) and during 5% CO2 inhalation (bottom). The increase
In blood volume during 5% CO2 inhalation, shown as an in­
crease in lightness in this figure, was more marked in cere·
bral plasma volume than in cerebral RBC volume.
ters as an independent variable was performed, and
Eq. 3 was obtained:
CBV
cerebral Hct was reduced significantly from the
steady-state value of 31.3 ± 1.8% to 30.3 ± 0. 9%
during 5% CO2 inhalation (p < 0.05 by paired t test).
Since CBV is known to respond not only to
changes in Paco2 but also to changes in MABP, a
bivariate analysis including both of these parame-
TABLE 1. Cerebral blood volume (CBV) and
hematocrit (Hct) in 10 right-handed healthy volunteers
Right
hemisphere
CBVb
4.71 ± 0. 39
Plasma volume
3.29 ± 0.28
RBC volume
Cerebral Hct (%)d
Hct ratio (%)'
Left
hemisphere
4.89 ± 0. 37'
1.49 ± 0.09
1.52 ± 0.11
31.9 ± 1.9
30.7 ± 1.8
77.4 ± 0.8
3.40 ± 0. 31'
74.4 ± 1.8I
Mean"
31. 3 ± 1.8
75.9 ± 2. 1
Mean values for bilateral hemispheres.
b Corrected by cerebral Hct.
cp<O.O I.
d Calculated as the ratio between the RBC and the plasma
e Ratio of cerebral-to-Iarge-vessel Hct.
I p < 0.05, paired t test of difference between right and left
J
Cereb Blood Flow Metabol. Vol. 5, No.2, 1985
0.10' Paco2
-
0.02' MABP
(3)
where CBV is in milliliters per 100 g brain.
The overall fit of Eq. 3 was significant at the p <
0.01 level. For the purpose of comparison with sim­
ilar equations in the literature, the same analysis
was applied to CBV values calculated only by the
data in the [99mTc]RBC study, using a standard ce­
rebral-to-large-vessel Hct ratio of 0.85 (85.0%), and
Eq. 4 was obtained:
CBV
=
3.85
+
0.05' Paco2
-
0.019' MABP
(4)
where CBV is in milliliters per 100 g brain.
DISCUSSION
3. 34 ± 0.28
volume, and plasma volume are in mlllOO g brain.
hemispheres.
+
4.81 ± 0. 37
Data are the means ± SD of four slices between 4.2 and 7.8
volume.
2.05
1.50 ± 0.09
cm below the vertex. Values for CBV, red blood cell (RBC)
a
=
Regional CBV
The present study emphasizes the importance of
measurements of cerebral Hct in the same subjects
for accurate calculation of CBV values, and cau­
tions against calculation of regional CBV using an
arbitrary and/or fixed cerebral-to-large-vessel Hct
ratio. Quantitative tomographic measurements of
regional CBV in humans by emission computed to­
mography was first reported by Kuhl et al. (1975)
utilizing 99mTc-Iabeled RBC as a single-photon­
emitting tracer. More recently, regional CBV has
CEREBRAL BLOOD VOLUME AND HEMATOCRIT
2Il
TABLE 2. Cerebral blood volume (CBV) responses to changes in P"co2 and MABP
Steady
state
CBV
4.81 ± 0. 37
Plasma volume
3. 34 ± 0.28
RBC volume
Cerebral Hct (%)"
Hct ratio (%)'
1.50 ± 0.09
31. 3 ± 1.8
75.9 ± 2.1
5% CO,
inhalation
5.52 ± 0.51
1.64 ± 0.11
3.87 ± 0.41
CO, responsiveness
2.16 ± 0.74
0.104 ± 0.021
2.48 ± 0.82
0.083 ± 0.018
1.46 ± 0.48
0.021 ± 0.008
30. 3 ± 0.9"
72.4 ± 2.2"
Data are means ± SD of four brain slices between 4.2 and 7.8 cm below the vertex in 10 right­
handed healthy volunteers. CBV values were corrected by cerebral hematocrit (Hc!). Values for
CBV . red blood cell (RBC) volume, and plasma volume are in mllIOO g brain. T he systemic
variables measured were: for RBC volume study. Paeo2 39.8 ± 2.6 mm Hg. MABP 87.8 ± 10.9
mm Hg during steady state and Paeo2 46.2 ± 2. 7 mm Hg, MABP 94.2 ± 11.4 mm Hg during 5%
CO2 inhalation; for plasma volume study, Paco) 40.4 ± 1.5 mm Hg, MABP 87.7 ± 8. 3 mm Hg
during steady state and Paeo2 46.6 ± 1.5 mm Hg. MABP 95.1 ± 10.4 mm Hg during 5% CO,-
inhalation.
a
Calculated as the ratio between the RBC and the plasma volume.
b p< 0.05, paired t test of difference between steady-state and CO,- runs.
C
Ratio of cerebral to large-vessel Hc!.
been determined by positron emission tomography
using lICO-labeled RBC (Grubb et al.. 1978; Phelps
et aI., 1979). Hitherto reported values for CBV mea­
sured by RBC tracers show only small variations
among authors: 4.2 ± 0.4 ml/lOO g brain by Phelps
et al. (1979) using [IICO]RBC, 4.3 ± 0.41 ml/lOO g
brain by Grubb et al. (1978) using [IICO]RBC, and
4.34 ± 0.5 ml/lOO g brain by Kuhl et al. (1980) using
[99mTc]RBc. On the other hand, a higher value of
5.7 ± 0. 6 mi/lOO g brain was reported in the
monkey by a method using the plasma tracer Re­
nographin 76 (Phelps et aI., 1973). All of these au­
thors used a cerebral-to-large-vessel Hct ratio of
0.85 for the correction of CBV, which was obtained
by averaging the cerebral Hct values in the litera­
ture (Phelps et aI., 1973). The mean CBV value of
4.81 ± 0.37 ml/lOO g brain reported here is higher
than any values measured by RBC tracer and lower
than values measured by plasma tracers. However,
if CBV were calculated by our data obtained in the
[99mTc]RBC study with the correction using a ce­
rebral-to-large-vessel Hct ratio of 0.85, the mean
CBV was 4.30 ± 0.32 ml/l00 g brain, which shows
closer agreement with previously reported values
measured by RBC tracers. The same situation ap­
plies to the data for the present plasma tracer mea­
surements. If CBV is calculated using our data ob­
tained in the plasma tracer study with a cerebral­
to-large-vessel Hct ratio of 0.85, a value of 5.13 ±
0.41 ml/IOO g brain is obtained. This is in good
agreement with the relatively high value measured
by a plasma tracer (Phelps et aI., 1979). Therefore,
the discrepancy between our CBV value and those
of previous authors may well be explained by dif­
ferences in the cerebral-to-Iarge-vessel Hct ratio,
which must be calculated for each method and
cannot be assumed to be fixed.
Significantly greater CBV values were measured
in the left cerebral hemisphere compared with the
right in our volunteers, all of whom were right­
handed and may be assumed to have predominantly
left cerebral dominance. The present results are in
agreement with those of Grubb et al. (1978), but
Kuhl et al. (1980) did not find consistent CBV dif­
ferences between left and right hemispheres. No
conclusive explanation can be offered for this dis­
crepancy, but it is possible that differences in the
steady-state conditions in different laboratories
may account for the different results reported (e.g.,
speaking to the patients, asking them to move their
right hands, etc.). It is worthy of note that the hemi­
spheric difference was more marked for plasma
volume than for RBC volume. It seems reasonable
to suggest that an increase in CBV in physiological
conditions may be associated with more marked in­
creases in the plasma component.
Regional cerebral Hct
The present results demonstrate the feasibility of
quantitative measurements of regional cerebral Hct
in humans using SPECT. They also suggest that the
method can detect variations of cerebral Hct during
different physiological conditions. The cerebral-to­
large-vessel Hct ratio in the present study was 75.9
± 2.1%, as summarized in Table 1. As this is the
first report in the literature of three-dimensional
measurements of cerebral Hct in a series of normal
human volunteers, the present results were com­
pared with earlier reports on global or two-dimen­
sional measurements of cerebral Hct in humans and
J
Cereb Blood Flow Metabol, Vol. 5, No.2, 1985
212
F.
SAKAI ET AL.
with measurements in experimental animals. De­
spite the fact that most of the previous studies used
similar types of tracers, namely, labeled RBC and
albumin, there are wide variations among reported
values concerning the ratio of cerebral-to-large­
vessel Hct, ranging from 0.41 to 0.92 (Gibson et aI.,
1946; Everett et aI., 1956; Larsen and Lassen, 1964;
Oldendorf et aI., 1965; Sklar et aI., 1968; Levin and
Ausman, 1969; Cremer et aI., 1983). The present
values for the cerebral Hct ratio were lower than
two reported values of global cerebral measure­
ments in humans, i.e., 0.92 from Larsen and Lassen
(1964) and 0.84 from Oldendorf et al. (1965). The
present results agree well with those of Cremer et
al. (1983) and Sklar et al. (1968), who employed
direct tissue-sampling techniques in experimental
animals and obtained cerebral Hct ratios of 0.70 and
0.74, respectively. Our value also compares well
with that of 0.69 measured by positron emission
tomography in eight patients with various neurolog­
ical disorders and one normal subject (Lammertsma
et aI., 1984). Since the average Hct of the capillary
blood is known to be less than that of blood in large­
diameter vessels as reported by Fahraeus (1929),
lower Hct values should be expected if tissue sam­
ples include more capillaries and arterioles. It is
likely that measurements of cerebral Hct with
three-dimensional resolution by the present method
or by positron emission tomography (Lammertsma
et aI., 1984) are representing more regional values
than previously reported by two-dimensional or
whole-brain measurements in humans. Since three­
dimensional values for cerebral Hct are obviously
essential for determining three-dimensional regional
CBV values in humans, accurate measurements of
regional CBV with emission computed tomography
can best be accomplished by sequential measure­
ments of RBC and plasma volume in the same pa­
tients.
Testing changes of CBV and Hct by hypercarbia
The measurement of the responsiveness of CBV
to alterations in Paco2 and blood pressure provides
a potential method for quantitating cerebral vas­
cular reactivity. Bivariate analysis in the present
study revealed a positive correlation between CBV
and Paco2 and an inverse correlation between CBV
and MABP. This indicates normal cerebral vasodi­
latory responses to increases in Paco2 with pre­
served cerebral autoregulatory responses to in­
creases in MABP.
Greenberg et al. (1978), using [99mTc]RBC as a
tracer, performed tomographic measurements of
local CBV responses to carbon dioxide in humans
and reported a value of 0.0495 ml/l00 g brain/mmHg
J
Cereb Blood Flow Metabol. Vol. 5, No.2, 1985
Pac02 or a 1.225% change per mmHg Paco2. Their
values were in good agreement with previous re­
ports of CBV responses to carbon dioxide inhala­
tion in experimental studies, which ranged from
0.041 to 0.049 mlllOO g brain/mmHg Paco2 (Smith
et aI., 1971; Phelps et aI., 1973; Grubb et aI., 1974;
Kuhl et aI., 1975). The CO2 responsiveness of CBV
in the present study, as calculated in Eq. 3, was
0.10 mlll00 g/mmHg Paco2, and was greater than
the previously reported results. However, if CBV
was calculated using only the data from the RBC
tracer studies as did previous authors, then CO2
responsiveness was 0.05 mlil00 g brain/mmHg
Paco2, as derived from Eq. 4. This value is in good
agreement with those of Greenberg et al. (1978) and
others. Thus, the greater responsiveness of CBV to
CO2 in the present study is attributed to greater
increases in plasma than in RBC volume during 5%
CO2 inhalation. As a result of this difference be­
tween RBC and plasma volume, cerebral Hct was
reduced during 5% CO2 inhalation.
A good agreement between our values on CO2
responsiveness by either RBC or plasma tracer and
previously reported values by three-dimensional
measurements may be taken as additional evidence
for validation of the present measurement. The
exact isolation of the intracerebral from the extra­
cranial tissue by selecting the region of interest in
the present method may not be as precise as that
in studies using positron emission tomography in
which the superposition of the brain image by
l18F]fluorodeoxyglucose in the same subject is pos­
sible (Kuhl et al., 1980). However, cerebral gray
and white matter were considered to be the main
constituents of the blood volume measured in the
present method, since blood volume of extracranial
structures should show little or no reaction to
changes in Paco2 (Heistad et aI., 1977).
Cerebral Hct was found to be reduced further
during 5% CO2 inhalation. To the best of our knowl­
edge, there has been no observation reported in the
literature of changes in cerebral tissue Hct during
vasodilatation associated with an increase in Paco2•
The interpretation of the mechanism for this Hct
change should then be based on previous hemor­
rheological studies. Studies on microcirculation
have shown that if flow rates increase in small ves­
sels and capillaries during vasodilatation, increased
shear on cellular orientation and deformation lead
to more cellular migration toward the tube axis and
faster RBC flow, which consequently results in a
reduced Hct (Gaehtgens et aI., 1978). Other pos­
sible mechanisms for the reduction of Hct include
a disproportional inflow of fluid from the interstitial
fluid space during vasodilatation. The most likely
CEREBRAL BLOOD VOLUME AND HEMATOCRIT
explanation is the possibility that arterial vasodila­
tation may increase the number of open capillary
channels available for plasma passage, with more
plasma skimming and a decrease in tissue Hct (Ro­
senblum, 1972).
Acknowledgment: Mr. K. Yoda, Mr. J. Suzuki, and Mr.
T. Takamatsu provided technical assistance. Prof. J. S.
Meyer, Baylor College of Medicine (Houston, TX,
U.S.A.), kindly assisted in preparing the manuscript.
REFERENCES
Cremer JE, Cunningham V J, Seville MP (1983) Relationships
between extraction and metabolism of glucose, blood flow,
and tissue blood volume in regions of rat brain. J Cereb
Blood Flow Metabol 3:291-302
Everett NB, Simmons B, Lasher EP (1956) Distribution of blood
(Fe59) and plasma (1131) volumes of rats determined by liquid
nitrogen freezing. Cire Res 4:419-424
Fahraeus R (1929) T he suspension stability of blood. Physiol Rei'
9:241-274
Gaehtgens P, Albrecht K H, Kreutz F (1978) Fahraeus effect and
cell screening during tube flow of human blood. I. Effect of
variation of flow rate. BiorheoloRY 15:147-154
Gibson JG II, Seligman A M, Peacock WC, Aub JC, Fine J,
Evans RD (1946) T he distribution of red cells and plasma in
large and minute vessels of the normal dog, determined by
radioactive isotopes of iron and iodine. J Clin/m'est 25:848857
Greenberg JH, Alavi A, Reivich M. Kuhl D, Uzzell B (1978)
Local cerebral blood volume response to carbon dioxide in
man. Cire Res 43:324-331
Grubb RL, Raichle ME, Eichling JO, Ter-Pogossian MM (1974)
The effects of changes in PaCO, on cerebral blood volume,
�
blood flow, and vascular mean transit lime. Stroke 5:630639
Grubb RL, Raichle ME. Higgins C S, Eichling .10 (1978) Mea-
213
surement of regional cerebral blood volume by emission to­
mography. Ann Neurol 4:322-328
Heistad DD, Marcus ML, Sandberg S, Abond FM (1977) Effect
of sympathetic nerve stimulation on cerebral blood flow and
on large cerebral arteries of dogs. Cire Res 41 :342-349
Kuhl DE, Reivich M, Alavi A, Nyary I. Slaum MM (1975) Local
cerebral blood volume determined by three-dimensional re­
construction of radionuclide scan data. Cire Res 36:610-619
Kuhl DE, Alavi A, Hoffman EJ, Phelps ME, Zimmerman RA,
Obrist WD, Bruce DA, Greenberg J H, Uzzell B (1980) Local
cerebral blood volume in head-injured patients. J NeurosurR
52:309-320
Lammertsma AA, Brooks DJ, Beaney RP, Turton DR, Kensett
MJ, Heather JD, Marshall J, Jones T (1984) In vivo mea­
surement of regional cerebral haematocrit using positron
emission tomography. J Cereb Blood Flow Metabol 4:317322
Larsen OA, Lassen NA (1964) Cerebral hematocrit in normal
man. J Appl PhysioI19:571-574
Levin VA, Ausman JI (1969) Relationship of peripheral venous
hematocrit to brain hematocrit. J Appl Physiol 26:433-437
Oldendorf WH, Kitano M, Shimizu S, Oldendorf SZ (1965) He­
matocrit of the human cranial blood pool. Cire Res 17:532539
Phelps ME, Grubb R L, Ter-Pogossian MM (1973) Correlation
between PaCO, and regional cerebral blood volume by x­
�
ray fluorescence. J Appl Physiol 35:274-280
Phelps ME, Huang SC, Hoffman EJ, Kuhl DE (1979) Validation
of tomographic measurement of cerebral blood volume with
C- II-Iabeled carboxyhemoglobin. J Nile! Med 20:328-334
Rosenblum WI (1972) Can plasma skimming or inconstancy of
regional hematocrit introduce serious errors in regional ce­
rebral blood flow measurements or their interpretation '?
Stroke 3:248-254
Sklar F H, Burke EF Jr, Langfitt TW (1968) Cerebral blood
volume: values obtained with 5lCr-Iabeled red blood cells
and RISA. J Appl Physiol 24:79-82
Smith AL, Neufeld GR, Ominsky AJ, Wollman H (1971) Effect
of arterial CO, tension on cerebral blood flow, mean transit
time, and vascular volume. J Appl PhysioI31:701-707
J
Cereb Blood Flo\\' Metabol, Vol. 5, No. 2, 1985