вход по аккаунту


Cerebral blood flow requirement for brain viability in newborn infants is lower than in adults.

код для вставкиСкачать
Cerebral Blood Flow Requirement
for Brain Viability in Newborn Infants
Is Lower than in Adults
Denis 1. Altman, MB,*$ William J. Powers, MD,t$ll Jeffrey M. Perlman, MB," Peter Herscovitch, MD$$
Sara L. Volpe, MS," and Joseph J. Volpe, MD"I$
Measurements of regional cerebral blood flow (CBF) with positron emission tomography in adult humans with cerebrovascular disease have demonstrated consi:tently that values below 10 mU(100 gm . min) occur only in infarcted
brain. Although experimental data suggest that the newborn brain may be more resistant to ischemic injury than the
adult brain, the minimum CBF necessary tci sustain neuronal viability in newborn infants is unknown. We have
measured CBF with positron emission tomogipaphy in 16 preterm and 14 term newborn infants and have determined
the relationship between CBF and subsequent brain function as assessed by neurological examination and developmental assessment. The range of mean CBF in the preterm infants was 4.9 to 23 mV(l00 gm . min) and the range of
mean CBF in the term infants was 9.0 to 73 in1/(100 gm . min). Five preterm infants and one term infant with mean
CBF less than 10 mU(100 gm . min) survived. Three of these 5 preterm infants, with mean CBF of 4.9, 5.2, and 9.3 ml/
(100 gm . min), respectively, have normal neurological examinations and Bayley Scales of 80 or greater at 6,6, and 24
months of age, respectively. One (mean CBF 6.9) has normal cognitive development (Bayley 103) and a mild spastic
diplegia at age 19 months, and one infant (mean CBF 6.2) has a left hemiparesis and a Binet IQ score of 70 at age 33
months. The term infant, with a mean CBF of 9.0 m1/(100 gm . min), was developing normally when he died of sepsis at
age 5 months. These data indicate that mean CBF as low as 5 ml/(100 gm . min) in newborn infants can be associated
with normal subsequent neurological development and therefore with preservation of a substantial number of cortical
neurons. Since the minimal CBF below this vdue necessary to maintain neuronal viability in newborn infants remains
to be determined, measurements of CBF in the newborn period should not be used in determination of brain death.
Altman DI, Powers WJ, Perlman JM, Herscovitch P, Volpe SL, Volpe JJ Cerebral blood How requirement
for b r a n viabil ty in newborn infants is lower than in adults. Ann Neurol 10HX,L4 218-226
It is generally accepted that the newborn brain is more
resistant to hypoxic and ischemic injury than the adult
brain. This conclusion is based primarily on differences
in the duration of survival during systemic hypoxia in
animals of different postnatal ages 11-41. However, it
is not entirely clear if the longer survival of newborn
animals reflects only increased resistance of the immature cardiovascular system or if there is indeed increased resistance of the brain as well. Support for the
latter is provided by observations that the rate of decline of cerebral high-energy phosphate compounds
during both hypoxemia and global cerebral ischemia
(produced by decapitation) is slower in newborn animals than in adults [S-7). There are, however, no direct measurements of the tolerance of newborn brain
to reductions in either blood flow or substrate supply.
In adults, measurements of regional cerebral blood
flow (CBF) in patients with cerebrovascular disease
have demonstrated that CBF values below 17 to 18
mV(100 gm . min) are associated with evidence of
neuronal dysfunction (electroencephalographic [EEG]
slowing o r hemiparesis) 18-1 I] and that CBF below
10 mY(100 gm . min) occurs only in areas of cerebral
infarction [11-13]. In newborn infants, Lou and coworkers [ 141 have reported that CBF values below 20
mV( 100 gm . min) measured in the first few hours after
birth were associated consistently with later evidence
of brain damage. They therefore concluded that maintenance of flow above this level was essential to prevent permanent brain injury. Recently, Greisen and
Trojaberg 1151 have reported normal visual evoked
responses in preterm infants with mean CBF below I0
mV( 100 gm . min), an observation that suggests neuronal viability is preserved at these low CBF levels.
However, no neurological Eollow-up data o n these infants were provided.
From the Departments of "Pediatrics, tNeurology, $Biological
Chemistry, and the $Division of Radiation Sciences, Mallinckrodt
Institute of Radiology of the Washington Univers.ty School of
Medicine, St Louis, and the "Department of Neurolog). of the Jewish
Hospital of St Louis, St Louis, MO.
Received Sep 25, 1987, and in revised form Dec 14, 1987, and Feb
8, 1988. Accepted for publication Feh 14, 1988.
Address correspondence to Dr Altman, Children's Hospital at
Washington University School of Medicine, 400 South Kingshighway, St Louis, MO 63 110.
218 Copyright @ 1988 by the American Neurological Association
To obtain more data on the level of CBF necessary
to sustain neuronal viability in newborn infants, we
have measured CBF with positron emission tomography (PET) in 30 newborn infants with various neurological disorders and determined the relationship
between CBF and neurological outcome. We have
demonstrated that mean CBF of 5 to 10 mU( 100 gm .
min) can be associated with survival and normal neurological development at 6 to 24 months of age.
Patient Population
PET measurements of CBF were performed on 16 preterm
and 14 term infants between December 1982 and March
1987. The preterm infants were studied as part of several
different protocols designed to investigate CBF in infants
with periventricular/intraventricular hemorrhage, hypoxicischemic encephalopathy, patent ductus arteriosus (PDA), or
previous extracorporeal membrane oxygenation. The term
infants were studied during investigations of CBF in infants
with hypoxic-ischemic encephalopathy or previous extracorporeal membrane oxygenation. Because of the effect of large
ventricles on PET measurements of CBF (see Appendix),
infants with ventriculomegaly were excluded from the study.
Infants were transported to the PET scanner in an Ohio
Transport Incubator fully equipped with life-support equipment and were attended by both an experienced neonatal
transport nurse and a neonatologist (J.M. P.). All infants had
stable neurological, cardiovascular, and respiratory functions
at the time the scan was performed. A newborn neurological
evaluation was performed on all infants immediately following the scan, and each infant's vital signs and ventilatory
requirements were recorded either at the time of the scan or
in the neonatal intensive care unit (immediately before leaving or immediately on returning).
Written informed consent was obtained from the next of
kin of all infants. All of these studies were approved by the
Human Studies Committee (Institutional Review Board) and
the Radioactive Drug Research Committee of Washington
University School of Medicine.
Measurements of regional CBF were performed on the
PETT VI tomograph in the low-resolution mode [ 161. Data
were collected simultaneously from 7 parallel slices with a
center-to-center separation of 1.4 cm. The reconstructed resolution was 18 mm full width at half maximum (FWHM) in
the transverse tomographic plane. Infants were positioned so
that the PET slices were approximately parallel to the orbitomeatal line. A radiopaque wire was used to record the
position of the lowest PET slice on a lateral skull radiograph,
thus providing a permanent record of the relationship between PET slices and the head. A transmission scan with a
"Ge/68Ga ring source was performed on each subject for
attenuation correction of the emission scans.
CBF was measured with a 40-second emission scan after
0.7 mCi/kg of 150-labeled water in 0.5 ml of saline was hand
injected over 2 to 5 seconds into a peripheral vein. Arterial
blood samples of 0.1 to 0.2 ml were drawn every 3 to 5
seconds from either an umbilical, radial, or tibial arterial
catheter placed previously for intensive care. The blood
radioactivity and scan data were corrected for physical decay
of " 0 and CBF was calculated using an adaptation to PET of
the Kety autoradiographic method 117, 181. We have previously described details of this method 119, 201. In the current study we used a value of 1.10 mVgm for the partition
coefficient of water based on the ratio of the water contents
of neonatal brain and blood {21].
For each subject the PET slices through the cerebral hemispheres were identified from the skull x-ray. The edge of the
brain on the CBF scan for each of these slices was determined as follows: the anteroposterior length of the skull for
each slice was measured on the skull x-ray and corrected for
magnification. An edge-finding algorithm was then used on
the corresponding slice of the transmission scan to produce a
template with this anteroposterior dimension and thus define
the outline of the skull. This outline was then applied to the
corresponding emission scan to eliminate extracerebral structures in the scan plane. In most newborn infants the subarachnoid space is 0 to 3 mm 122) and thus the position of
the skull closely identifies the edge of the cortex. In a minority of preterm infants the subarachnoid space in the posterior
parietal region may be 12 to 13 mm 1221. Accordingly, in 5
infants, the transmission scan template was adjusted to conform more closely to the apparent edge of the brain as
defined by the emission scan. Because of the small size of the
newborn brain relative to the axial resolution (slice thickness), PET slices near the base of the brain contain substantial contribution from extracerebral and brainstem structures
and cannot be used for determining mean hemispheric CBF.
To obtain a more accurate measurement of mean hemispheric CBF by including cerebral tissue at the base, singlepixel values for each templated slice were linearly interpolated with corresponding pixels in slices above and below
to produce 4 interpolated slices between each original slice.
Final mean CBF was determined from all original and interpolated slices encompassing the hemispheres.
Given the 18-mm FWHM resolution of P E T I VI and the
small head size of newborn infants, accurate regional measurements of gray matter and white matter CBF were not
possible (see Discussion). Gray matter flows are underestimated and white matter flows are overestimated. Since
many of the experimental data on CBF thresholds of ischemia deal with regional values in small gray and white matter
regions, estimates of mean gray matter and mean white matter CBF were obtained for comparative purposes as follows.
In each infant, the maximum CBF value in any single pixel
from all original hemispheric slices was taken as the minimum possible mean gray matter flow for combined cortical
and subcortical structures. This value together with a hemispheric gray/white matter ratio of 47:53 (determined
planimetrically from an atlas of the newborn brain [23]) and
the mean CBF was used to calculate the maximum possible
mean white matter blood flow according to the formula:
Mean white matter blood flow = (mean CBF - 0.47 mean
gray matter blood flow)/0.53.
The absorbed dose of radiation for a 3,500-gm infant receiving 0.7 mCi/kg of H2150intravenously is 57 mrads to
the whole body, 201 mrads to the organ that receives the
largest radiation exposure (spleen), and 130 mrads to the
Altman et al: CBF in Newborns
Data Analysis
gonads. An additional 75 mrads to the head is provided by
the lateral skull x-ray and transmission scan.
Neurological and Cognitive Follow-Up
The statistical significance of relationships between CBF and
neurological outcome was determined by one-way analysis of
Surviving infants were evaluated by neurological and cognitive testing specifically for the purposes of this study. All
clinical neurological examinations were performed by the
same examiner (D. I. A.).
Cognitive function was evaluated in infants less than 2
years old with the Bayley Scales of Infant Developrnent and
in infants 2 years or older with the Stanford-Binet IQ Test or
with the Cattel Infant Intelligence Scale. In children less than
2 years old, chronological age was corrected for prematurity,
i.e., postterm age was used. All cognitive studies were carried out by the same examiner (S. L. V.). Two infants who
underwent PET measurements of CBF in the newborn period are not included in this report because of inadequate
follow-up. The family of one infant had moved and could not
be traced, and the parents of the second infant declined to
bring their child for the follow-up evaluation.
Neurological outcome in all surviving infants W Z i classified
as: ( 1) normal, development quotient (DQ) greater than 80,
no motor or other neurological abnormality; (2) mild deficit,
DQ greater than 80, but with one or more slightly abnormal
motor findings, e.g., slight hypotonia or hypertonix; (3) moderate deficit, DQ between 50 and 80, with or without moderate or marked neurological deficits, e.g., quadriparesis,
hemiparesis, spastic diplegia; or ( 4 ) severe deficir, DQ less
than 50 with moderate to marked neurological deficits as
described above.
Clinical details and outcome of the 16 preterm infants
are shown in Table 1. The range of mean CBF in these
infants was 4.9 to 23 mV(100 gm . min) and peak pixel
(minimum mean gray matter) CBF was 7.6 to 39 mV
(100 gm * min). The infant with mean CBF of 4.9 mV
(100 gm . min) had a repeat measurement 10 minutes
later of 5.4 mV(100 gm . mirr). There were 0 infants
with mean CBF beIow 10 rrtI/(lOO g m . min), 5 of
whom survived the newborn period. These 5 included
3 infants who are normal (mean CBF 4.9, 5.2, 9.3,
respectively), one infant with normal cognitive function and a mild spastic diplegia (mean CBF 6.9), and
one infant with a left hemiparesis and DQ of 70 (mean
CBF 6.2).
Clinical details of the 14 term infants and their
neurological outcome are shown in Table 2. The range
of mean CBF was 9.0 to 73 mV(100 g m . min) and the
range of peak pixel (minimum mean gray matter) CBF
was 14 to 134 mV(100 gm . rnin). One infant with
mean CBF of 9.0 mV(100 gm min) who was developing normally (by report from the infant's physician)
died at 5 months from peritonitis and septicemla.
Table 1 . Cerebral Blood F~OUJ
in Preterm Infants
Age at
PET Scan
"Measured in ml/( 100 gm
PDA closed
2 days
L hemiparesis
Spastic d iplegia
Spastic quadriparesis
R hemiparesis
Spastic diplegia
138 Bayley
90 Hayley
7 0 Binet
103 Bayley
101 Binet
91 Binet
<50 Baykey
84 Bayley
121 Bayley
81 Bayley
0 1 Bayley
86 Binet
GA = gestational age; PET = positron emission tomography; IVH = inrraventricular hemorrhage, Grade I t o 11; Grade I, subependymal
hemorrhage with minimal or no IVH; Grade 11, I V H but neither ventricle completely filled with blood; IPE = intraparenchymal (pcriventriculac) echodensity; HIE = hypoxic-ischemic encephaloparhy; ECMO = status after common carotid ligation and extracorporeal membranc
oxygenation; PDA = patent ducrus arreriosus; CBF = cerebral blood flow; L = left; R = right.
220 Annals of Neurology Vol 24 No 2
August 1988
Table 2. Cerebral Blood Flow and Neurological Outcome in Term Infants
Age at
PET Scan
7 days
Neurological Examination
Normal by MD report
Microcephaly, spastic quadriparesis,
seizure disorder
Microcephaly, spastic quadriparesis,
seizure disorder
Mild spastic diplegia
Microcephaly, spastic quadriparesis,
seizure disorder
Microcephaly, spastic quadriparesis,
seizure disorder
<50 Bayley
71 Bayley
64 Bayley
<50 Bayley
51 Cattel
"Measured in mU(100 gm . min).
bDied at age 5 mo from peritonitis.
= status after common carotid ligation and extracorporeal membrane oxygenation; HIE
cerebral blood flow; PET = positron emission tomography.
Estimated minimum mean gray matter CBF and
maximum mean white matter CBF for all surviving
infants with mean CBF less than 10 mV( 100 gm . min)
are shown in Table 3. Minimum mean gray CBF
ranged from 7.6 to 16.7 mV(100 gm . min) whereas all
white matter values were below 5 mV(100 gm * min).
Details of the newborn neurological examination,
vital signs, and ventilatory requirements of the infants
who survived with mean CBF less than 10 mV(100 gm
min) are as follows. At the time of the scan, the
neurological examination was normal in 3 infants (Preterm Infants 2 and 6, and Term Infant 1); the other 3
infants exhibited mild flexor hypotonia and reduced
spontaneous movements. All infants had normal alerting responses, intact brainstem reflexes, and normal
motor responses to sensory stimulation. All 5 preterm
infants required assisted ventilation with an F I ~ of
0.25 to 0.30, positive airway pressures of 15 to 23 mm
Hg, positive end-expiratory pressures of 3 to 5 mm
Hg, and ventilatory rates of 10 to 30/min. These infants' blood pressures were 60 to 80128 to 48 mm Hg.
The term infant was ventilated at a rate of 20/min,
positive airway pressure of 31 mm Hg, and positive
end-expiratory pressure of 6 mm Hg with an F I ~ of
0.8. The blood pressure was 100140 mm Hg. None of
these infants had received phenobarbital.
No significant relationship between mean CBF or
hypoxic-ischemic encephalopathy; CBF =
Tabie 3. Minimum Mean Gray and Maximum Mean
White Matter CBF" in Infants Surviving with Mean
CBF Less than 1C mll(l00 gm . min)
Minimum Mean
Gray CBF
Maximum Mean
White CBF
"Measured in mV(100 gm . min).
cerebral blood flow.
peak pixel (minimum mean gray matter) CBF and
neurological outcome was demonstrated for the 30 infants (mean CBF: F = 0.858, p = 0.481; peak pixel
C B F F = 0.966, p = 0.431).
We have demonstrated that mean hemispheric CBF as
low as 4.9 mV(100 gm . min) in newborn infants can be
associated with normal subsequent neurological development. The method that we have used to measure
Altman et al: CBF in Newborns 221
CBF is an adaptation to PET of the tissue ;autoradiographic method of Kety and co-workers [17, 181. We
have previously presented an analysis of this technique
and experimental validation of its accuracy rl9, 20). A
detailed analysis of its specific application to the measurement of CBF in newborn infants is presented in
the Appendix.
The principal CBF measurement we have used is a
mean value for blood flow to both ceretiral hemispheres. In adult humans with cerebrovascular disease,
multiple PET measurements of CBF from regions of
noninfarcted brain containing both gray and white matter in approximately equal proportions have demonstrated that CBF is always greater than 10 mL'( 100 gm .
min). CBF below 10 mV( 100 gm . min) has been observed only in brain regions that corresponded to areas
of lucency on x-ray computed tomographic 1:CT) scanning indicative of necrosis of normal brain cdlular elements [ll-13). These measurements are subject to
the same technological constraints as those we report
here, and thus are directly comparable. We observed
mean hemispheric CBF below 10 mV( 100 gm . min) in
4 infants who had normal neurological development
subsequently. These infants have not undergone follow-up neuroimaging procedures, e.g., CT and magnetic resonance imaging. Since it is well known that
infants may have apparently normal neurological examinations and normal cognitive function in the presence of neuroradiological evidence of brain injury, we
do not preclude the possibility that these infants suffered some brain injury. Rather, the point of our observation is that the low mean CBF values we observed
in these infants are invariably associated with destruction of all neuronal tissue in adults. The latter is clearly
not consistent with the normal developmental outcome we observed, which requires preservation of
substantial numbers of both neurons and :supporting
glial and vascular elements.
Experimental data from adult animals describe CBF
thresholds for neuronal death of 10 to 12 ml/( 100 gm .
min) based primarily on direct measurements of gray
matter flow in small areas r24-28). The CBEi threshold
for white matter infarction is slightly below i:his at 5 to
8 mU(100 gm . min) C24, 25, 291. As noted in the
Appendix, the limited spatial resolution of PET makes
accurate measurement of gray and white matter CBF
impossible; gray matter CBF is underestimated and
white matter CBF is overestimated, and :hus these
experimental values are not directly comparable to our
values. We have used the peak CBF value in any pixel
as an estimate of the minimum possible mean gray flow
(Table 3). Although the accuracy of CBF measurement
in a single pixel is limited by statistical variations in
radioactive decay rate, use of this small region allowed
us to minimize the effect of surrounding white matter
areas on the underestimation of mean gray flow. With
Annals of Neurology
Vol 24
No 2
August 1988
this value for mean gray flow, we were in turn able ro
estimate a maximum possible mean white flow from
the mean CBF and the gray/white matter ratio 1231.
These mean white matter values (shown in Table 3)
are all below 5 mU(100 gm . min), the minimum
threshold for white matter viability in adult experimental animals. Although very low, these values still ovctrestimate the true white matter flow to the extent that
the gray flow is underestimated. Furthermore, if white
matter flow is heterogeneous, flow to some regions
would be higher and flow to other regions would be
even lower.
Experimental observations in primates subjected to
total cerebral ischemia for 1, hour have demonstrated
the return of cerebral electrical activity and metabolic
function if recirculation is restored under carefully
controlled conditions 130-32). In such animals, although the EEG returned to nearly normal after 24
hours and partial recovery of somatosensory potentials
was observed, no animals, regained consciousness,
demonstrated EEG activation to pain, or had pupillary
light reflexes. While these observations have been of
crucial importance in the understanding of the pathophysiology of ischemic neuronal death, they have little
relevance to clinical conditions of global ischemia
when the ultimate concern is preservation of integrated cognitive brain function.
Experimental models of cerebral ischemia have
demonstrated that the relationship between CBF and
the occurrence of cerebral infarction is more complex
than a simple threshold phenomenon. The development of irreversible neuronal damage depends not
only on the magnitude of the blood flow reduction but
also on its duration. Since all of the infants we studied
had stable neurological, carcliovascular, and respiratory
functions, we do not believe that the flows we measured below 10 mU(100 gm . min) reflected brief, fortuitously recorded episodes of reduced cerebral perfusion, but rather a stable condition. This conclusion is
supported by findings in the infant who had 2 mean
CBF measurements 10 minutes apart of 4.9 and 5.4
mV( 100 gm . min).
Although several other investigators have reported
CBF below 10 mY(100 gm . rnin) in newborn infants
E14, 15, 33, 341, none has provided data on neurological outcome of these infants, with the exception of
Lou and co-workers [14}. They reported CBF measurements made a few hours after birth in 1 5 preterm
and 4 term newborn infants with varying degrees of
respiratory distress [ 141. They found that infants with
initial CBF below 20 mV(100 gm . min) invariably
demonstrated either cerebral atrophy or abnormal
neurological signs at follow-up 9 t o 12% months later.
Only one infant had CBF below 10 mU( 100 gm . min),
i.e., 3 rnli(100 gm . rnin). A.t follow-up, this infant had
focal cortical atrophy seen on CT, two or more abnor-
mal neurological signs, and a Cattell score between 90
and 100. Lou and associates concluded that the critical
value for CBF of 20 mV(100 gm * min) in newborn
infants is similar to that for adults and that the maintenance of CBF at higher levels is critical in preventing
permanent cerebral damage. Our data suggest that the
critical value for CBF is very much lower than 20 mV
(100 gm . rnin); in fact, we have demonstrated that
CBF as low as 5 mY(100 gm . min) can be associated
with normal subsequent neurological development.
The difference between the critical CBF values derived in the present study and in that of Lou and coworkers 1351 cannot be explained by methodological
considerations alone. Lou and associates 1351 used the
initial slope method of Olesen and associates 136) to
calculate CBF after injection of xenon 133 into the
common carotid artery. While this method tends to
produce overestimation of mean CBF at higher flows,
at flows below 30 mV(100 gm . min), the clearance
curve becomes monoexponential 1361. Its initial slope
is therefore more representative of whole brain blood
flow. Since the partition coefficient of xenon for both
gray matter and white matter in newborn brain is similar to that for gray matter in adult brain 1341, conventional application of the adult gray matter partition
coefficient to the initial slope as described by Olesen
and co-workers will yield an accurate mean value for
CBF in newborn brain when CBF is low. Thus measurements of CBF below 20 ml/( 100 gm . min) by Lou
and co-workers should be accurate measurements of
mean CBF within the limits of estimating the initial
slope in the presence of recirculating tracer 1373. Their
values for CBF therefore can be compared to our own
values for mean CBF, despite the slight underestimation of their measurement due to contamination by
low blood flow in extracranial structures supplied by
branches of the common carotid artery.
Moreover, differences in timing of CBF measurements between the study of Lou and co-workers 114)
and our own are unlikely to underlie the differences in
critical values for CBF. Lou and co-workers performed
their measurements within a few hours of birth,
whereas ours were carried out between 2 and 54 days
of age, the majority before 14 days. However, it is not
apparent why lower CBF should be tolerated more
poorly immediately after birth than a few weeks later.
In experimental animals, both CBF and cerebral metabolic rate for oxygen have been shown to increase in
the first few weeks after birth 138-401; this increase
raises the possibility that low levels of CBF might be
tolerated even better in the first few days of life.
Our best explanation of the discrepancy between
our data and those of Lou and associates is twofold.
First, with the study of more subjects, we may simply
have encountered greater biological variability. Second, the CBF measured by Lou and co-workers im-
mediately post partum may have reflected the severiry
of both systemic cardiorespiratory and neurological
dysfunction. Infants with CBF below 20 mY(100 gm .
min) may have been more liable to develop subsequent cardiorespiratory abnormalities in the postnatal
period that produced neurological damage.
This study was designed to investigate the relationship of CBF and neurological outcome and was not
designed to investigate the numerous poorly understood factors that may regulate CBF in newborn infants. Such a study would require large numbers of
infants with a homogeneous clinical condition, measurements of CBF on specific days after birth, and
manipulation of such factors as Pco~, arterial oxygen
content, hemoglobin affinity for oxygen, mean arterial
blood pressure, intracranial pressure, hematocrit, and
drug therapy. Such a study would provide invaluable
information, but this was not our purpose. As the infants in this study formed a clinically heterogeneous
group, studied at different gestational and postnatal
ages with multiple potentially interacting factors, no
valid analysis of factors controlling CBF in newborn
infants can be made from our data.
The value of single CBF measurements in the newborn period as an indicator of the degree of brain
injury appears to be minimal. Thus, although Lou and
associates { 141 have demonstrated a poor prognosis
for infants with CBF measurements below 20 mV( 100
gm . min) obtained in the first few hours of life, CBF
above this value did not necessarily predict a good
outcome. Three of their 8 patients with CBF greater
than 20 mV( 100 g m . min) subsequently developed
abnormal neurological signs. Our data demonstrate
that a single low CBF measurement taken after the first
day of life does not predict a poor outcome, and, like
that of Lou and associates, our data show that higher
CBF values do not assure normal neurological development. This failure of a single CBF measurement to
correlate with the apparent degree of brain injury is
not surprising. In the acute period after brain injury,
normal mechanisms of CBF control are impaired.
Thus, CBF is often normal or elevated in damaged
tissue with reduced metabolic rate [41, 42). Poor correlations between CBF and clinical outcome have been
observed after cardiac arrest, head trauma, and stroke
in adults C43-457, and we now show the same poor
correlation for newborn infants.
The minimal CBF necessary to maintain neuronal
viability in newborn infants remains to be determined.
The results we have reported here indicate that the
minimum sufficient value must lie below a mean CBF
of 5 d ( 1 0 0 gm * min). Our data are consistent with
the recent demonstration by Greisen and Trojaberg
1151that mean CBF as low as 4.3 mV(100 g m . min) is
sufficient to maintain neuronal function (as measured
by visual evoked responses) in preterm infants. The
Altman et al: CBF in Newborns
mechanism by which newborn brain can tolerate lower
CBF remains to be determined. The most obvious hypothesis is that the metabolic requirements of the tissue are lower [5, 71. Other differences such as the
increased ability to excrete such potentially toxic substances as lactate 146) or to utilize lactate for energy
production 1471 may be important as well.
The observation that a mean CBF of 5 mY(100 gm .
min) can be associated not only with survival but also
with normal neurological development has important
clinical implications for the determination of brain
death in newborn infants. Because the diagnosis of
brain death in these infants is extremely difficult to
establish by clinical and EEG criteria 1481, the use of
tests of cerebral perfusion such as radionuclide angiography has been proposed 1491. The sensitivil:y of radionuclide angiography for detecting CBF below 10 mY
(100 gm . min) is unknown. Furthermore, the capability of such methods as cerebral radionuclide angiography and stable xenon C T to distinguish between totally absent CBF and CBF of 5 mY( 100 gm . min) has
not been established. Recent data on the use of xenon
CT to diagnose brain death in adults suggescs that this
distinction cannot be made accurately [SO]. TJntil there
is additional documentation of the value of these tests
in the newborn, we agree with the Task Force on Brain
Death in Children that such tests should not ,3e used to
determine brain death in the newborn 1491. Furthermore, since we have demonstrated previously that
pathological evidence of brain death can b,: demonstrated in newborn infants with CBF above 20 mY( 100
gm . rnin), cerebral perfusion tests cannot be used to
rule out brain death 1511.
The PET autoradiographic method for measuring CBF consists of three components: measurement of the input of
radioactivity to the brain by arterial blood sampling, measurement of radioactivity present in the brain by PET, and
application of the mathematical model relating CBF to these
two measurements. W e analyze each of these with respect to
its accuracy in newborn infants.
Arterial Input Function to the Brain
The arterial input function to the brain is derived from blood
samples collected from a peripheral artery. Sampling intervals between 1 and 5 seconds have little effect on the CBF
measurement [ 191. Adequate counts, typically lo4 to 10' per
sample, are recorded to ensure statistical counting accuracy.
An assumption is made that the peripheral arterial timeradioactivity curve is identical to that of the cerebral arteries,
save for the difference in arrival time of the radioxtivity at
the peripheral and cerebral sites. This difference in arrival
time is corrected for by recording with PET the coincidence
events at I-second intervals to time the arrival of the radioactive bolus in the head. The peripheral arterial curve is shifted
appropriately [l9, 207. In all infants reported here, even
Annals of Neurology
Vol 24
No 2 August 1988
those with mean CBF below 10 mV( 100 gm . rnin), an abrupt
sustained increase in coincidence events was readily apparent. The shape of the input function (and therefore thc rapidity of injection) has minimal effect on the calculation of
CBF 1191. Any dispersion in the arterial time-radioactivity
curve will be similar for peripheral and cerebral ,arteries 1521.
The cerebral arterial input will not be reflected accurately by
peripheral arterial sampling when there is a right-to-left
shunt via a PDA. Under these circumstances, arterial sampling at sites distal to the origin of the ductus arteriosus from
the aorta will overestimate the cerebral input and result in
underestimation of CBF. Although many of the infants we
studied had arterial samples drawn from sites distal to the
origin of a PDA, this cannot be the explanation for the low
mean CBF values we have measured for several reasons. All
infants were acyanotic and resting quietly; thus, they were
unlikely to generate the high intrathoracic pressures necessary to produce right-to-left shunts. Furthermore, o f the 6
surviving infants with mean (3RF below 10 mV( 100 grn .
min), one had no clinical evidence of a PDA (Term Infant l ) ,
one had no evidence of PDA by echocardiography (Prcterm
Infant 5 ) , and one had the ductus arteriosus surgically ligated
3 days before the PET scan (Preterm Infant 2). Onc infant
with clinical evidence of a PDA (Preterm Infant 1) had a
right radial arterial catheter for arterial sampling (preductal
arterial blood), thus precluding a contribution of right-to-left
shunted blood to the arterial input curve.
Accuraly of PET
The accuracy of PET as a quantitative measurement of regional radioactivity in vivo when proper care is given to
technical considerations has been amply demonstrated { 5 31.
We have ensured the accuracy of our measurement by performing individual transmission scans o n each patient t o provide accurate attenuation correction, and by collecting a
sufficient number of total coincidence events (minimum
150,000 for slices used in calculation of mean CBF in spite o f
the small brain site) to ensure statistical counting accuracy.
Cross-calibration of PET measurements of radioactivity and
the well-counter used to count blood samples is performed
routinely [20]. The accuracy of PE?T VI, the Scanner used
in this study, has been demonstrated experimentally [ 16].
Given the effects of partial volume averaging caused by limited spatial resolution (18 mm FWHM in the present study),
accurate measurements of gray and white matter CBF are not
possible with PET [ 5 3 ] . W e have therefore chosen to calculate mean hemispheric CBF based on all coincidence events
detected for the appropriate PET slices. Since PET measures
radioactivity per unit volume of intracranial contents, the
presence of cerebrospinal fluid in ventricles and subarachnoid space will lead to an underestimation of true CBF per
100gm of the brain tissue. This error is approximately >(% in
normal adults C541 and is even less in newborn infants because the sulci and ventricles are smaller [ 5 5 ] . Infants with
ventriculomegaly were excluded :from our study for this reason.
Mathematical Model
The mathematical model used to measure CBF is an adaptation to PET of the tissue autoradiographic method of Kety
and co-workers E17, 18). The relationship between tissue
radioactivity and CBF described by this model is almost
linear. This has several advantageous consequences [19]. In
regions with heterogeneous flow, the mean flow measured
by PET is very close to the true average flow. Furthermore,
changes in the value of the partition coefficient have little
effect on the measured value for CBF. W e have used a partition coefficient of 1.10 mVgm based on the normal water
contents of newborn brain and blood [2 11. While this value
may be slightly different in damaged cerebral tissue, this
difference will have minimal effect on the CBF values we
have obtained. The primary disadvantages of the method we
have used stem from the fact that H2150 does not freely
diffuse from blood into brain. While this leads to underestimation of CBF at high flows, the method is extremely accurate at lower flows, especially those below 20 mU(100 gm .
min) 120). Although the quasi-one-compartment model used
to calculate CBF by this method does not completely describe the behavior of H2I50 in the brain [56],it does provide accurate measurements of CBF with PET when scan
time is limited to 40 seconds [20, 56).
The overall accuracy of this method has been proven by
experimental verification comparing PET measurements of
mean hemispheric CBF to those obtained by direct intracarotid injection and residue detection 120). These validation studies were performed with PE’JT VI on primates with
brains smaller than those of the infants we studied and demonstrated excellent agreement between the two methods
Supported by N I H Grants NS06833, HL1385 1 , NS07027, GCRC
RR00036, Teacher Investigator Development Award NS00647 (Dr
Powers), The McDonnell Center for Studies of Higher Brain Function, and the Jewish Hospital of St Louis.
The authors thank the staff of the Division of Radiation Sciences,
the Edward Mallinckrodt Institute of Radiology, and the Neonatal
Transport Team of Children’s Hospital for their assistance.
I. Kabat H. The greater resistance of very young animals to arrest
of the brain circulation. Am J Physiol 1940;130:588-599
2. Himwich HE, Alexander FAD, Fazekas JF. Tolerance of the
newborn to hypoxia and anoxia. Am J Physiol 1941;133:327328
3. Fazekas JR, Alexander FAD, Himwich HE. Tolerance of the
newborn to anoxia. Am J Physiol 1941;134:281-287
4. Glass HG, Snyder FF, Webster E. The rate of decline in resistance to anoxia of rabbits, dogs, and guinea pigs from the onset
of viability to adult life. Am J Physiol 1944;140:609-615
5. Duffy TE, Kohle SJ, Vannuci RC. Carbohydrate and energy
metabolism in perinatal rat brain and relation to survival in
anoxia. J Neurochem 1975;28:271-276
6. Gonzalez-Mendez R, McNeill A, Gregory GA, et al. Effect of
hypoxic hypoxia on cerebral phosphate metabolites and p H in
the anesthetized infant rabbit. J Cereb Blood Flow Metab
7. Holowach-Thurston J, McDougal DB Jr. Effect of ischemia on
metabolism of the brain of the newborn mouse. Am J Physiol
8. Leech PJ+ Miller JD. Fitch W, Barker J. Cerebral blood flow,
internal carotid artery pressure and the EEG as a guide to the
safety of catotid ligation. J Neurol Neurosurg Psychiatry 1973;
9. Trojaberg W, Boysen G. Relation between EEG, regional cerebral blood flow and internal carotid artery pressure during
carotid endarterectomy. EEG Clin Neurophysiol 1973;35:62-
10. Sundt TM Jr, Sharbrough FS, Anderson RE, Michenfelder JD.
Cerebral blood flow measurements and electroencephalograms
during carotid endarterectomy. J Neurosurg 1974;41.3 10-320
11. Powers WJ, Grubb RL Jr. Darriet D, Raichle ME. Cerebral
blood flow and cerebral metabolic rate of oxygen requirements
for cerebral function and viability in humans. J Cereb Blood
Flow Metab 1985;5:600-608
12. Baron JC, Rougemont D, Bousser MG, et al. Local CBF, oxygen extraction and CMR02: prognostic value in recent supratentorial infarction. J Cereb Blood Flow Metab 1983;3(suppl
3. Baron JC, Rougemont D, Lebrun-Grandie P, et al. Measurement of local blood flow and oxygen consumption in evolving
cerebral infarction: an in vivo study in man. In: Meyer JS, Lechner H , Reivich M, Ott EO, eds. Cerebral vascular disease 4.
Princeton: Excerpta Medica, 1983
4. Lou HC, Skov H, Pederson H . Low cerebral blood flow, a risk
factor in the neonate. J Pediatr 1979;95:606-609
5. Greisen G, Trojaberg W. Cerebral blood flow, PaC02 changes,
and visual evoked potentials in mechanically ventilated, preterm
infants. Acta Pediatr Scand 1987;76:394-400
16. Ter-Pogossian MM, Ficke DC, Hood JT Sr, et al. PETT VI: a
positron emission tomograph utilizing cesium fluoride scintillation detectors. J Comput Assist Tomogr 1982;6:125-133
17. Landau WM, Freygang W H Jr, Rowland LP. The local circulation of the living brain, values in the unanesthetized and anesthetized cat. Trans Am Neurol Assoc 1955;80:125-129
18. Kery SS. Measurement of local blood flow by the exchange of
an inert diffusible substance. Meth Med Res 1960;8:228-236
19. Herscovitch P, Markham J, Raichle ME. Brain blood flow measured with intravenous H2”0. I. Theory and error analysis. J
Nucl Med 1983;24:782-789
20. Raichle ME, Martin WRW, Herscovitch P, et al. Brain blood
flow measured with intravenous H2150.11. Implementation and
validation. J Nucl Med 1983;24:790-798
21. Herscovitch P, Raichle ME. What is the correct value for the
brain-blood partition coefficient of water? J Cereb Blood Flow
Metab 1985;5:65-69
22. McArdle CB, Richardson CJ, Nicholas DA, et al. Developmental features of the neonatal brain: MR imaging. Part II.
Ventricular size and extracerebral space. Radiology 1987;162:
23. h s a u JP, Bastian D, Carbanis EA, Pourcelot L. Sections of the
cranium. In: Atlas of neonatal anatomy. New York: Masson,
24. Marcoux bW, Morawetz RB, Crowell RM, er al. Differential
regional vulnerability in transient focal cerebral ischemia. Stroke
25. Jones TH, Morawetz RE3, Crowell RM, et al. Thresholds of
focal cerebral ischemia in awake monkeys. J Neurosurg 1981;
26. Astrup J. Energy requiring cell functions in the ischemic brain. J
Neurosurg 1982;56:482-497
27. Heiss WD, Rosner G. Functional recovery of cortical neurons
as related to degree and duration of ischemia. Ann Neurol
28. Branston NM, Ladds A, Symon A, Wang AD. Comparison of
the effects of ischemia on early components of the somatosensory evoked potential in brainstem, thalamus, and cerebral cortex. J Cereb Blood Flow Metab 1984;4:68-81
29. Morawetz RB, Marcoux FW, Crowell RM, et al. Identical
Altman et al: CBF in Newborns
thresholds for cerebral ischemia in white and gray matter. Acra
Neurol Scand 1979;60(suppl 2):282-283
30. Hossman K-A, Zimmermaii V. Resuscitation of the monkey
brain after 1 h complete ischemia. 1. Physiological and morphological observations. Brain Res 1974;81:59-74
3 I . Hossmann K-A, Grosse Ophoff B. Recovery of monkey brain
after prolonged ischemia. I . Electrophysiology and brain electrolytes. J Cereb Blood Flow Metab 1986;6:15-21
32. Bodsch W, Barbier A, Oehmichen M, et al. Recovery of monkey brain after prolonged ischemia. 11. Protein synthesis and
marphologic alterarions. J Cereb Blood Flow Metab 1986;6:
33. Ment LR, Duncan CC, Ehrenkranz RA, et al. Intravenrricular
hemorrhage in the prererm neonate: timing and cerebral blooci
flow changes. J Pediatr 1984;104:413-425
34. Greisen G. Cerebral blood flow in preterm infants during the
first week of life. Acta Paediarr Scand 1986;75:43-5 1
35. Lou HC, Lassen NA, Fries Hansen B. Impaired autoregulation
of cerebral blood flow in the distressed newborn infant. J
Pediatr 1979;94:118-12 1
36. Olesen J , Paulson OB, Lassen NA. Regional cerebral blood
flow in man determined by the initial slope of the clearance of
rhe intra-arterially injected "'Xe. Stroke 1971;2:519-540
17. Larson KB. Models for dynamic tracer studies. in: Iarson KB,
Cox JR, eds. Computer processing of dynamic images from an
Anger scintillation camera. New York: Society o f Nuclear
Medicine. 197 1: 152-168
38. Kennedy C, Grave GD, JehleJW, SokoloffL. Changes in blood
flow in the component structures of the dog brain during posrnatal maturation. J Neurochem 1972;19:2423-2433
S9. Tyler CB, Van Harreveld A. The respiration of the developing
bran. Am J Physiol 1942;136:600-603
40. Himwich HE, Fazekas JF. Comparative studies of the metabolism of the brain of infant and adult dogs. Am J Physiol 1941;
41. Powers WJ, Raichle ME. Posirron emission tomography and its
applications to the srudy of cerebrovascular disease in man.
Stroke 1985;16:361-376
42. Obrisr WD, LangfmTW, JaggiJL,et al. Cerebral blood flow and
metabolism in comatose patients with acute head injury. J
Neurosurg 1984;61:241-253
43. BecksteadJE, Tweed WA, Lee J, MacKeen WL. Cerebral blood
Annals of Neurology
Vol 24
No 2
August 1988
flow and metabolism in man following cardiac arrest. Stroke
44. Overgaard J, Mosdal C, Tweed WA. Cerebral circulation atrer
head injury. Part 3: Does reduced regional cerebral blood flow
determine recovery of brain function after blunt head injury! J
Neurosurg 1981;55:63-74
45. Burke AM, Younkin D, Gordon J , et al. Changes in cerchral
blood flow and recovery from acute stroke. Stroke 1980.17:
173-1 78
46. Raichle ME. The pathophysiolo,e of bran ischemia. Ann
Neurol 1983;13:2-10
47. Volpe JJ. Hypoglycemia and brain injury. In. Neurology of the
Newborn. ed 2. Philadelphia: Saunders, 1087:369-370
48. Volpe JJ. Brain death determinarion in rhc newborn. Pediatrics
49.Task Force for the Determination of Bran Dedth in Children.
Guidelines for the determination of bran death in chilrlren.
Ann Neurol 1987;22:616-617
50. Darby JM, Yonas H , Gur D, latchaw RE. Xenon-enhanced
computed tomography in brain-cleath. Arch Neurol 1% ';44:
51. Altman DI, P e r h a n JM, Powers WJ* et d. Preservarion of
brainstem and cerebral blood flow (CBF) in two asphyxiated
newborn infants with clinical brain dearh. Ped Rcs 1W7;21
52. Herscovitch P, Raichle ME, Kilbourn MK, Welch MJ. Positron
emission tomographic measurement of cerebral blood How and
permeability-surface area product o f watcr using 150-waterand
"C-butanol. J Cereb Blood Flow Metah 1087;7:527-542
53. Hoffman EJ, Phelps ME. Positron emission tomography. principles and quantitarion. In: Phelps ME, Mazziotta J , Schelherg H ,
eds. Positron emission tomography and auioradiography principles and applications for the brain and heart. New York Raver.,
54. Herscovitch P, Auchus AP, Gaclo M, er al. Correction of posltron emission tomography data for cerebral atrophy. J Cereb
Blood Flow Metab 1986;6:120--124
55. Hanvood-Nash DC, Fitz CR. In: Neuroradiology in infanrs and
children. St Louis: Mosby, 1976:225-2%
56. h s o n KB, Markham J. Raichlc ME. Tracer-kinetic models for
measuring cerebral blood flow using exrernally derectcd radiotracers. J Cereb Blood Flow Mrtab 1987;7.443-463
Без категории
Размер файла
986 Кб
flow, adults, infant, brain, requirements, lowe, newborn, viability, cerebral, blood
Пожаловаться на содержимое документа