Content uploaded by Sunit C Singhi
Author content
All content in this area was uploaded by Sunit C Singhi
Content may be subject to copyright.
http://jcn.sagepub.com
Journal of Child Neurology
DOI: 10.1177/0883073807308716
2008; 23; 192 J Child Neurol
Rakshay Shetty, Sunit Singhi, Pratibha Singhi and M. Jayashree Infections and Raised Intracranial Pressure: Is It Feasible?
Cerebral Perfusion PressureTargeted Approach in Children With Central Nervous System
http://jcn.sagepub.com/cgi/content/abstract/23/2/192
The online version of this article can be found at:
Published by:
http://www.sagepublications.com
can be found at:Journal of Child Neurology Additional services and information for
http://jcn.sagepub.com/cgi/alerts Email Alerts:
http://jcn.sagepub.com/subscriptions Subscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
http://jcn.sagepub.com/cgi/content/refs/23/2/192
SAGE Journals Online and HighWire Press platforms):
(this article cites 13 articles hosted on the Citations
© 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.
at WUHAN UNIV LIBRARY on June 18, 2008 http://jcn.sagepub.comDownloaded from
192
caused by raised intracranial pressure, using therapies aimed
at optimizing cerebral perfusion pressure.
Optimization of cerebral perfusion pressure has gained
recognition as the therapeutic end point in the management
of traumatic brain injury. In adult patients with severe trau-
matic brain injury, Rosner et al3showed that cerebral
perfusion pressure could be elevated by inducing sys-
temic hypertension with the help of vasopressors, without
worsening intracranial pressure, and that maintaining cere-
bral perfusion pressure greater than 70 mm Hg reduced mor-
tality rate and improved functional outcome across all
Glasgow Coma Scale score categories. These authors hypoth-
esized that the brain recognizes a change in its perfusion
rather than a change in blood pressure and autoregulates its
flow. Optimal perfusion decreases intracranial pressure by
cerebral vasoconstriction and also by decreasing ischemic
brain damage.3Thus, when cerebral perfusion pressure is
stabilized at a higher level, intracranial pressure can be better
controlled without the risk of cerebral ischemia.3Prabhakaran
et al4compared a cerebral perfusion pressure–targeted
approach with an intracranial pressure–targeted approach in
Central nervous system infections are a major cause of
mortality and morbidity in children. Case fatality in
patients with meningitis in developing countries is as
high as 30% despite adequate antibiotic management.1
Cerebral herniation has been reported in ~30% of deaths on
autopsy, reflecting a critical increase in intracranial pressure
in patients with meningitis.2In addition to carrying a high
mortality rate, meningitis carries a high risk of serious neu-
romorbidity in survivors. Further improvement in outcome
from central nervous system infections in children could
perhaps be achieved by preventing secondary brain injury
Original Article
Cerebral Perfusion Pressure–Targeted
Approach in Children With Central
Nervous System Infections and Raised
Intracranial Pressure: Is It Feasible?
Rakshay Shetty, MBBS, MD, Sunit Singhi, MBBS, MD, Pratibha Singhi, MBBS, MD,
and M. Jayashree, MBBS, MD
This study was conducted to evaluate the feasibility of cerebral
perfusion pressure–targeted therapy in children with raised
intracranial pressure caused by central nervous system infection.
A prospective observational pilot study was conducted in the
pediatric intensive care unit of a tertiary care teaching hospital.
Twenty children (ages 6 months to 12 years) with a clinical
diagnosis of meningitis or meningoencephalitis were included.
Intracranial pressure and blood pressure monitoring were initi-
ated soon after enrollment. Interventions to reduce intracranial
pressure and elevate blood pressure were used to achieve a tar-
get cerebral perfusion pressure of greater than 70 mm Hg in chil-
dren 2 years of age or older and greater than 60 mm Hg in
children less than 2 years. Therapies used to achieve target cere-
bral perfusion pressure were initial fluid boluses (in 14 patients),
vasopressors (in 8), and mannitol (in 10). The target cerebral
perfusion pressure was achieved in 6 patients, whereas a cerebral
perfusion pressure greater than 50 mm Hg was achieved in 16
patients. All 4 patients with mean cerebral perfusion pressure
less than 50 mm Hg died of intractable, raised intracranial pres-
sure. In contrast, only 3 of 16 patients with mean cerebral per-
fusion pressure more than 50 mm Hg died. In children with raised
intracranial pressure caused by central nervous system infection,
it was feasible to achieve a cerebral perfusion pressure greater
than 50 mm Hg, mainly by increasing the blood pressure within
the first 24 hours and by reducing intracranial pressure after the
first 24 hours.
Keywords: cerebral perfusion pressure; central nervous sys-
tem infections; intracranial pressure; intracranial pressure
monitoring
From the Department of Pediatrics, Advanced Pediatric Centre, Postgraduate
Institute of Medical Education and Research, Chandigarh, India.
Address correspondence to: Prof. Sunit Singhi, MBBS, MD, Pediatric
Emergency and Intensive Care Units, Advanced Pediatrics Centre, PGIMER,
Chandigarh–160 012, India; e-mail: sunit.singhi@gmail.com, dr_singhi
@yahoo.com.
Shetty R, Singhi S, Singhi P, Jayashree M. Cerebral perfusion pressure–
targeted approach in children with central nervous system infections and
raised intracranial pressure: is it feasible? J Child Neurol. 2008;23:192-198.
Journal of Child Neurology
Volume 23 Number 2
February 2008 192-198
© 2008 Sage Publications
10.1177/0883073807308716
http://jcn.sagepub.com
hosted at
http://online.sagepub.com
© 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.
at WUHAN UNIV LIBRARY on June 18, 2008 http://jcn.sagepub.comDownloaded from
Cerebral Perfusion Pressure–Targeted Approach / Shetty et al 193
pediatric patients with traumatic brain injury. They found this
approach safe and effective.
We have evaluated the feasibility of cerebral perfusion
pressure–targeted therapy in children with raised intracra-
nial pressure caused by central nervous system infection,
because there are few published data on this subject.
Patients and Methods
This observational pilot study was conducted in the pedi-
atric intensive care unit of a multispecialty tertiary care
teaching hospital in north India. Children with a clinical
diagnosis of meningitis or meningoencephalitis admitted
to the pediatric intensive care unit from January 2005 to
December 2005 were enrolled in the study if they fulfilled
all the following criteria:
1. Age 6 months to 12 years.
2. Glasgow Coma Scale score of 7 or less.
3. Evidence of increased intracranial pressure on computed
tomography or magnetic resonance imaging5: namely, loss
of sulci, slit-like ventricles, loss of grey and white matter
distinction, and obliteration of suprasellar and quadrigem-
inal cisterns or presence of brain herniation syndromes.
Diagnosis of meningitis or meningoencephalitis was
clinically suspected if a patient with fever had 1 or more of
the following6: impaired sensorium, convulsions, headache,
vomiting, neck stiffness and signs of meningeal irritation
such as neck rigidity, Kernig’s sign, and Brudzinski’s sign.
The diagnosis was accepted if the cerebrospinal fluid from
lumbar tap showed leukocytosis (cells >20 μL) and elevated
proteins (>40 mg/μL). Blood and cerebrospinal fluid cul-
tures for bacteria and serology for Japanese encephalitis and
herpes simplex virus were obtained in all cases.
Patients with traumatic brain injury, subarachnoid hem-
orrhage, epidural or subdural hematoma, hydrocephalus,
central nervous system tumors on neuroimaging, hyperten-
sive encephalopathy, and clinical brain death7were excluded.
The parents of eligible patients were explained the
purpose and the benefits of the study. Those who gave
written consent were enrolled. The study was approved by
the Ethics Committee of the Institute.
Patient Care and Monitoring
All patients were in a supine position and were continuously
monitored for heart rate, respiratory rate, mean blood pres-
sure, intracranial pressure, central venous pressure, and
oxygen saturation. Adequate airway using endotracheal
intubation and breathing using mechanical ventilation were
ensured in all patients. The ventilator settings were aimed at
maintaining a PaO2of more than 90 mm Hg and PaCO2of 35
mm Hg (normal ventilation). The goal of fluid therapy was to
maintain euvolemia. Isotonic fluids were preferred. Blood
sugar and serum electrolytes were monitored and maintained
within the normal range by appropriate adjustment. Fever
was managed promptly with acetaminophen and, if necessary,
hydrotherapy. Seizures were managed as per unit protocol
with intravenous diazepam and phenytoin. All ventilated
patients were sedated with midazolam as per unit protocol.
Outcome Measure
The goal was to maintain cerebral perfusion pressure greater
than 70 mm Hg in children older than 2 yrs of age and
greater than 60 mm Hg in children younger than 2 years.
An intraparenchymal pressure transducer (CODMAN
Microsensor basic kit) with continuous displays of intracra-
nial pressure (ICP EXPRESS Monitor, Johnson and Johnson
India Ltd) was used for intracranial pressure monitoring. A
record of intracranial pressure, measured every 2 hours, was
kept. Cerebral perfusion pressure was calculated as the
arithmetic difference of mean blood pressure and mean
intracranial pressure. If at any point cerebral perfusion pres-
sure was less than the target values, the intervention, as per
protocol, was carried out.
Cerebral Perfusion Pressure–Targeted Protocol
and Monitoring
The protocol centered on volume expansion and use of
vasopressors to support blood pressure and use of manni-
tol to reduce intracranial pressure. Acute hyperventilation
with manual bag ventilation was used for periods of
acutely raised intracranial pressure and was titrated
against cerebral perfusion pressure. Continuous hyper-
ventilation was not used.
Isotonic fluid (0.9% saline or lactated Ringer’s solution) in
boluses of 10 to 20 mL/kg was given if cerebral perfusion
pressure was below the target and central venous pressure
was less than 10 mm Hg. The goal was to maintain central
venous pressure at 10 to 12 mm Hg. If cerebral perfusion
pressure was still low and central venous pressure was within
target range, dopamine was started at 15 μg/kg/min and
titrated up or down until the target cerebral perfusion pressure
was achieved or the maximum infusion rate was 20 μg/kg/min.
If cerebral perfusion pressure was still low, epinephrine was
started at 0.05 μg/kg/min and titrated to a maximum of 0.4
μg/kg/min to raise mean arterial blood pressure. Mannitol,
0.25 to 0.5 g/kg, was given if intracranial pressure was more
than 20 mm Hg, target cerebral perfusion pressure was not
achieved with the help of inotropes alone, and the patient was
already hypertensive. Mannitol was repeated as needed, usu-
ally every 4 to 6 hours, if target cerebral perfusion pressure
was not achieved. Isotonic fluids were used to replace urinary
loss. The intracranial pressure transducer was removed once
the intracranial pressure was under control.
© 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.
at WUHAN UNIV LIBRARY on June 18, 2008 http://jcn.sagepub.comDownloaded from
194 Journal of Child Neurology / Vol. 23, No. 2, February 2008
Data Analysis
Descriptive statistics (mean, standard deviation, median,
range, percentage) were used to describe the patients’
characteristics. Regression analysis was used to analyze
the relationship between mean arterial blood pressure
and intracranial pressure, mean arterial blood pressure
and cerebral perfusion pressure, and cerebral perfusion
pressure and intracranial pressure. Analysis of variance
and independent ttest were used to analyze the differ-
ences between those who survived without sequelae,
those who survived with sequelae, or those who died. P≤
.05 was considered significant. All data analysis was done
using SPSS 10 software (SPSS Inc, Chicago, Illinois).
Results
Clinical characteristics of the 20 patients included in the
study are shown in Table 1. Fourteen (70%) had encephali-
tis and 6 (30%) had meningitis. All the patients were in deep
coma at admission (median Glasgow Coma Scale score 5,
range 3-7). Two patients had hypotension at admission. The
target cerebral perfusion pressure could be achieved in 6
(30%) patients only. However, a cerebral perfusion pressure
greater than 50 was maintained in 16 (80%) patients.
The interventions used to achieve target cerebral perfu-
sion pressure were fluid boluses in 14 (70%) patients, vaso-
pressors in 8 (40%) patients, and mannitol in 10 patients.
Table 1. Clinical Characteristics of Enrolled Patients
PRISM GCS Mean CPP, Minimum CPP, Maximum ICP,
Patient Age Diagnosis Score Score mm Hg mm Hg mm Hg Trend of CPP Outcome
1 3.5 y Acute bacterial meningitis 21 4 61 43 43 Stable Minor neurological
deficits
2 3.5 y Viral encephalitis 28 6 69 35 27 Stable Normal
3 8 y Viral encephalitis 28 5 51 35 90 Frequent Major neurological
decreases sequelae
4 4 y Herpes encephalitis 10 4 77 64 19 Stable Major neurological
sequelae
5 4.5 y Viral encephalitis 17 3 58 38 30 Infrequent Major neurological
decreases sequelae
6 10 y Varicella encephalitis 15 4 80 62 20 Stable Major neurological
sequelae
7 7 mo Viral encephalitis 19 3 52 21 33 Infrequent Died on day 4a
decreases
8 5 y Viral encephalitis 31 3 46 18 68 Infrequent Died on day 5
decreases
9 7 y Japanese encephalitis 15 4 53 26 43 Infrequent Major neurological
decreases sequelae
10 8 y Cysticercus encephalitis 13 7 77 49 16 Stable Normal
11 8 y Viral encephalitis 10 7 55 56 30 Stable Normal
12 6 mo Haemophilus influenzae 10 7 61 56 11 Stable Normal
meningitis
13 6 y Viral encephalitis 26 6 46 11 81 Frequent Died on day 1
decreases
14 5 y Viral encephalitis 22 3 49 0 84 Frequent Died on day 2
decreases
15 6 y Viral encephalitis 32 3 16 0 41 Frequent Died on day 1
decreases
16 5 y Japanese encephalitis 18 3 58 14 70 Infrequent Died on day 6a
decreases
17 4.5 y Pneumococcal meningitis 18 6 71 40 40 Stable Died on day 16a
18 6 mo Acute bacterial meningitis 15 6 50 30 22 Infrequent Normal
decreases
19 8 y Acute bacterial meningitis 21 3 60 14 62 Frequent Major neurological
decreases sequelae
20 6 mo Pneumococcal meningitis 18 6 55 37 20 Infrequent Minor neurological
decreases deficits
NOTE: PRISM =Pediatric Risk of Mortality; GCS =Glasgow Coma Scale; CPP =cerebral perfusion pressure; ICP =intracranial pressure; stable =up to 1 acute
decrease; infrequent decreases =up to 2 decreases; frequent decreases =3 or more acute decreases in CPP within first 48 hours.
a . Death attributable to hospital-acquired sepsis.
© 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.
at WUHAN UNIV LIBRARY on June 18, 2008 http://jcn.sagepub.comDownloaded from
Cerebral Perfusion Pressure–Targeted Approach / Shetty et al 195
Mean numbers of fluid (20 mL/kg) boluses (±SD) and man-
nitol doses given over a period of 48 hours were 4 ±3 and
3 ±2, respectively. Mean ± SD concentrations of dopamine
and epinephrine used during the first 48 hours were 17 ±
2 μg/kg/min and 0.2 ±0.2 μg/kg/min, respectively.
Study of variation in mean arterial blood pressure,
intracranial pressure, and cerebral perfusion pressure over
baseline during the first 48 hours revealed that the cerebral
perfusion pressure curve closely followed that of mean arte-
rial blood pressure, particularly during the first 24 hours
(Figure 1). A significant decrease in intracranial pressure
associated with an increase in cerebral perfusion pressure
was noted only after 24 hours of application of protocol.
The patients who died within in the first 24 hours showed a
steady increase in intracranial pressure (Figure 1); it was
not possible to improve cerebral perfusion pressure in these
patients by either raising mean arterial blood pressure or
using mannitol and acute hyperventilation.
On regression analysis, a positive correlation between
cerebral perfusion pressure and mean arterial blood pressure
(r =0.64, P=.0006) and a negative correlation between
intracranial pressure and cerebral perfusion pressure (r =
0.65, P=.0005) were noted. The correlation between
intracranial pressure and mean arterial blood pressure was
not significant (r =0.11; P=.60).
Of the 20 patients, 7 died and 13 survived, 8 with seque-
lae. Four deaths were attributed to refractory intracranial
hypertension. These 4 patients had a mean cerebral perfu-
sion pressure less than 50 mm Hg and had episodes of
increased intracranial pressure greater than 40 mm Hg last-
ing for 6 hours or longer (Table 1). They had viral encephali-
tis, and the Glasgow Coma Scale score at admission was 3 in
three of them. One patient had hypotension at admission. Of
the 9 patients who had intracranial pressure greater than 40
mm Hg, 5 died; of the 4 survivors, 2 had major sequelae, 1
had minor sequelae, and only 1 survived without sequelae.
A comparison of data of patients who survived without
sequelae, those who survived with sequelae, and those who
died is shown in Table 2. The cerebral perfusion pressure
and intracranial pressure at admission and mean intracranial
pressure and mean cerebral perfusion pressure over the first
48 hours were not significantly different across the 3 groups
(Table 2). Glasgow Coma Scale score at admission (P=.01)
and the lowest recorded cerebral perfusion pressure (P=.01)
were significantly lower in those who died (Table 2). The
maximum recorded intracranial pressure showed a trend
Figure 1. (A) Variations in intracranial pressure (squares), cerebral per-
fusion pressure (triangles), and mean arterial blood pressure (diamonds)
from baseline during 48 hours (excludes 3 patients who died within 48
hours). (B) Variations in intracranial pressure (squares), cerebral perfusion
pressure (triangles), and mean arterial blood pressure (diamonds) from the
baseline among 3 patients who died within 48 hours.
Figure 2. Mean ±1 SD cerebral perfusion pressures of study patients
grouped according to outcome—survived with no sequelae (n =5), sur-
vived with sequelae (n =8), and died (n =7)—measured every 2 hours
during the first 24 hours.
© 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.
at WUHAN UNIV LIBRARY on June 18, 2008 http://jcn.sagepub.comDownloaded from
196 Journal of Child Neurology / Vol. 23, No. 2, February 2008
toward the higher side in children who survived with seque-
lae and those who died. The frequency of therapies required
to achieve target cerebral perfusion pressure did not vary sig-
nificantly across the 3 groups (Table 2). There was no statis-
tically significant difference in serum electrolytes and blood
sugar in the 3 groups. Patients who died or survived with
sequelae showed wide swings in cerebral perfusion pressure
during their monitoring period (Figure 1B). The mean cere-
bral perfusion pressure was generally greater than 60 mm Hg
in survivors without sequelae, in contrast to those who had
sequelae or who died (Figure 2). The cerebral perfusion
pressure remained greater than 50 mm Hg for a significantly
higher proportion of time among survivors compared with
those who died (78.8% vs 53.8%, P=.029).
Of the 13 survivors, 8 had neurologic sequelae at the
time of discharge. Six had major neurological sequelae (3
hemiparesis, 2 paraparesis, and 1 loss of vision) and 2
patients had minor sequelae (1 had extrapyramidal move-
ments, 1 had mild monoparesis of left upper limb). These
sequelae persisted at 1-month follow-up.
Discussion
Management of intracranial hypertension in patients with
central nervous system infections remains a formidable
challenge. We studied the feasibility of using a cerebral per-
fusion pressure–targeted approach in comatose children
(median Glasgow Coma Scale score of 5) with underlying
central nervous system infections with a hope that it would
prevent secondary brain injury.
We found that an increase in cerebral perfusion pressure
paralleled elevation of mean arterial blood pressure in the
first 24 hours; a decrease in intracranial pressure during
the first 24 hours was not significant. However, on day 2
there was an appreciable decrease in intracranial pressure
associated with a concomitant increase in cerebral perfu-
sion pressure. This indicated that maintenance of mean
arterial blood pressure in the first 24 hours was critical
in maintaining cerebral perfusion pressure; the increase in
cerebral perfusion pressure as a result of the decrease in
intracranial pressure was more evident after the first 24
hours. Marmarou et al,8in a retrospective analysis of 139
patients with head injury, found that blood pressure was the
chief determinant of cerebral perfusion pressure in those
not undergoing cerebral perfusion pressure–targeted ther-
apy, whereas intracranial pressure was the determinant of
cerebral perfusion pressure in those undergoing cerebral
perfusion pressure–targeted therapy.
Most of the patients who died or had major neurological
sequelae in our study showed wide fluctuations in cerebral
perfusion pressure. Similar findings were reported by
Table 2. Characteristics of Patients Who Survived Without Sequelae, Those Who
Survived With Sequelae, and Those Who Died
Survived Without Survived With PValue by
Parameters Sequelae (n =5) Sequelae (n =8) Died (n =7) ANOVA
Median (range) age, y 3.5 (0.5-8) 5.75 (0.5-10) 5 (0.7-6) .74
Sex ratio, male:female 4:1 6:2 5:2 .94
Diagnosis, meningitis:encephalitis 2:3 3:5 1:6 .53
Median (range) GCS score at admission 6 (6-7) 4 (3-6) 3 (3-6) .01a
CVP, mm Hg 9 ±0.83 9.5 ±2.8 10 ±0.96 .74
MBP, mm Hg 75 ±9.5 82 ±9.3 75 ±11.3 .30
Opening ICP, mm Hg 18 ±6 34 ±13 26 ±17 .15
Opening CPP, mm Hg 62 ±17 46 ±17 51 ±25 .38
Lowest CPP, mm Hg 40 ±12 36 ±22 15 ±25 .01b
Maximum ICP, mm Hg 29 ±23 41 ±25 60 ±21 .09
Mean CPP, mm Hg 63 ±11 62 ±11 48 ±17 .11
Mean ICP, mm Hg 15 ±14 23 ±8 29 ±17 .15
Patients receiving fluid boluses, n 3 6 5 —
Patients receiving vasopressors, n 3 4 5 —
Patients receiving mannitol, n 2 4 4 —
Sodium, mol/L 136 ±3.2 136 ±4.3 132 ±2.5 .95
Potassium, mol/L 4.3 ±0.43 4 ±0.44 4.2 ±0.3 .5
Blood sugar, mg/dL 124 ±20 132 ±20 140 ±24 .51
PaO2, mm Hg 124 ±34 123 ±29 134 ±21 .81
PaCO2, mm Hg 36 ±9 37 ±9 29 ±9 .22
NOTE: ANOVA =analysis of variance; GCS =Glasgow Coma Scale; CVP =central venous pressure; MBP = mean arterial blood pressure; ICP = intracranial pressure;
CPP = cerebral perfusion pressure. All values expressed as mean ±SD unless specified otherwise.
a . GCS score: between those who survived without sequelae versus those who survived with sequelae (P=.006); those who survived without sequelae versus those who
died (P=.01).
b . Lowest CPP: between those who survived without sequelae versus those who died (P=.03); those who survived with sequelae versus those who died (P=.01).
© 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.
at WUHAN UNIV LIBRARY on June 18, 2008 http://jcn.sagepub.comDownloaded from
Cerebral Perfusion Pressure–Targeted Approach / Shetty et al 197
Kirkness et al,9who found that episodes of low cerebral per-
fusion pressure had a strong correlation with adverse out-
come in patients with severe primary brain injury. Those with
less cumulative percentage of time below specific cerebral
perfusion pressure thresholds were more likely to have a bet-
ter outcome at discharge.9
We did not find any correlation between intracranial
pressures and mean arterial blood pressure. This indicates
that intracranial pressure remained stable over a mean arte-
rial blood pressure of 73 to 87 mm Hg—the range that was
recorded in our patients. This pressure-stable response10
suggests that autoregulation was unaffected in our patients.
This is similar to the findings of Ashwal et al,11 who studied
20 critically ill children with acute bacterial meningitis and
found that autoregulation was intact within a range of mean
arterial blood pressure from 56 to 102 mm Hg. This is in
contrast to the findings of Moller et al,12 who in a study of
16 adult patients found that autoregulation was impaired in
the early phase of acute bacterial meningitis.
Our study showed a significant positive correlation
between cerebral perfusion pressure and mean arterial blood
pressure (r =0.64) and a negative correlation between cere-
bral perfusion pressure and intracranial pressure (r =–0.65).
This finds support in the concept that cerebral perfusion
pressure is the predominant stimulus for cerebral autoregula-
tion.3If cerebral perfusion pressure is inadequate, intracra-
nial pressure will progressively increase. An increase in
cerebral perfusion pressure above threshold will decrease
intracranial pressure by initiating autoregulatory cerebral
vasoconstriction with a resultant reduction in cerebral blood
volume. This important physiological response is clinically
demonstrated by a decrease in intracranial pressure with
improvement in cerebral perfusion pressure.3Hence, in
patients with intact autoregulation, optimizing cerebral per-
fusion pressure controls intracranial pressure indirectly by
autoregulating cerebral blood flow. Our findings are similar
to those of Rosner et al,3who, in a study of 150 adult
patients, found that intracranial pressure decreased as cere-
bral perfusion pressure increased.
Prabhakaran et al4demonstrated an effect known as
“hysteresis” in the setting of traumatic brain injury. The
basis of this effect is that an injured brain requires higher
cerebral perfusion pressure than the normal brain to
achieve a relatively normal cerebral blood flow because of
an increase in cerebrovascular resistance. As a result, the
critical closing pressures (the cerebral perfusion pressure at
which cerebral blood flow ceases) and lower autoregulatory
limit (the cerebral perfusion pressure below which vascula-
ture constricts or dilates passively and intracranial pressure
varies directly with cerebral perfusion pressure) are raised.
Prabhakaran et al4were able to achieve a mean cerebral
perfusion pressure of 74.6 ±12.2 mm Hg (target of >60 mm
Hg in children younger than 2 years, >70 mm Hg in those
older than 2 years) in children with traumatic brain injury.
Although we planned the same target cerebral perfusion
pressures, we failed to achieve supranormal cerebral perfu-
sion pressure. This was possibly attributable to the global
nature of brain swelling as well as differences in underlying
mechanisms of raised intracranial pressure in central nerv-
ous system infections. Our patients were also considerably
younger (average age 4.9 years vs 8.1 years) and had lower
mean Glasgow Coma Scale score in comparison to the
cohort studied by Prabhakaran et al.4However, we were
able to maintain cerebral perfusion pressure greater than
50 mm Hg in 16 patients. We found intact autoregulation
in our study population with cerebral perfusion pressure
greater than 50 mm Hg and a decrease in intracranial pres-
sure with increasing cerebral perfusion pressure. We did not
find the presence of hysteresis in our study population, pos-
sibly indicating that the lower autoregulatory limit remains
unchanged in setting of meningitis and meningoencephalitis.
There were 7 deaths in our study population; 5 of these
patients had a Glasgow Coma Scale score of 3. All 4
patients in whom intracranial pressure ranged between 40
and 100 mm Hg and the mean cerebral perfusion pressure
remained less than 50 mm Hg (despite aggressive volume
resuscitation, mannitol, and the use of vasopressors) died.
The inability to attain target cerebral perfusion pressure in
these patients probably reflects a greater severity of illness.
In an earlier study of 100 patients with nontraumatic coma
between the ages of 2 months and 12 years, we found that
modified Glasgow Coma Scale score at admission had a sig-
nificant association with outcome; mortality rates progres-
sively increased with decreasing Glasgow Coma Scale
score.13 Nayana Prabha et al14 found a significant associa-
tion between death and modified Glasgow Coma Scale
score scores on admission with a posttest probability for dis-
charge being only 10% with a score of less than 5 and 99%
with a score of more than 10. It is likely that cause of illness
also had an impact on final outcome. All 4 patients who died
of intractable intracranial hypertension had viral encephali-
tis. However, this being a pilot study, detailed analysis of
determinants and predictors of outcome was not con-
ducted. Further studies should address the benefit of cere-
bral perfusion pressure–targeted therapy in specific central
nervous system infection such as bacterial meningitis and
viral encephalitis.
In children with raised intracranial pressure caused by
central nervous system infections, it was feasible to main-
tain cerebral perfusion pressure greater than 50 mm Hg
using a cerebral perfusion pressure–based protocol, even
when patients had intracranial pressure exceeding 40 mm
Hg. Within the first 24 hours cerebral perfusion pressure
was achieved mainly by elevating blood pressure and later
by a reduction of intracranial pressure. Maintaining
the cerebral perfusion pressure greater than 50 mm Hg
was associated with both a decrease in intracranial pres-
sure and a better survival rate. However, larger studies
are needed to confirm the benefits of cerebral perfusion
pressure–targeted therapy in children with central nervous
© 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.
at WUHAN UNIV LIBRARY on June 18, 2008 http://jcn.sagepub.comDownloaded from
198 Journal of Child Neurology / Vol. 23, No. 2, February 2008
system infections. Future studies should focus on cere-
bral perfusion pressure–targeted therapy in patients with
different etiologies of raised intracranial pressure.
References
1. Saez-Llorens X, McCracken GH Jr. Bacterial meningitis in chil-
dren. Lancet. 2003;361:2139-2148.
2. Durand ML, Calderwood SB, Weber DJ, et al. Acute bacterial
meningitis in adults: a review of 493 episodes. N Engl J Med.
1993;328:21-28.
3. Rosner MJ, Rosner SD, Johnson AH. Cerebral perfusion pres-
sure: management protocol and clinical results. J Neurosurg.
1995;83:949-962.
4. Prabhakaran P, Reddy AT, Oakes WJ, et al. Pilot trial comparing
cerebral perfusion pressure–targeted therapy to intracranial pres-
sure targeted therapy in children with severe traumatic brain
injury. J Neurosurg. 2004; 100(suppl 5):454-459.
5. Taveras JM, Spellman JP. General pathological conditions. In:
Taveras JM, Spellman JP, eds. Neuroradiology. 3rd ed. Baltimore,
MD: Williams & Wilkins; 1996:120-121.
6. Dunn LT. Raised intracranial pressure. J Neurol Neurosurg
Psychiatry. 2002;73(suppl 1):123-127.
7. American Academy of Pediatrics. Report of special task force.
Guidelines for the determination of brain death in children.
Pediatrics. 1987;80:298-300.
8. Marmarou A, Saad A, Aygok G, Rigsbee M. Contribution of
raised ICP and hypotension to CPP reduction in severe brain
injury: correlation to outcome. Acta Neurochir Suppl. 2005;95:
277-280.
9. Kirkness CJ, Burr RL, Cain KC, et al. Relationship of cerebral
perfusion pressure levels to outcome in traumatic brain injury.
Acta Neurochir Suppl. 2005;95:13-16.
10. Howells T, Elf K, Jones PA, et al. Pressure reactivity as a guide
in the treatment of cerebral perfusion pressure in patients with
brain trauma. J Neurosurg. 2005;102:311-317.
11. Ashwal S, Stringer W, Tomasi L, et al. Cerebral blood flow and
carbon dioxide reactivity in children with bacterial meningitis.
J Pediatr. 1990;117:523-530.
12. Moller K, Larsen FS, Qvist J, et al. Dependency of cerebral
blood flow on mean arterial pressure in patients with acute bac-
terial meningitis. Crit Care Med. 2000;28:1027-1032.
13. Bansal A, Singhi SC, Singhi PD, et al. Nontraumatic coma.
Indian J Pediatr. 2005;72:467-473.
14. Nayana Prabha PC, Nalini P, Tiroumourougane Serane V. Role
of Glasgow Coma Scale in pediatric nontraumatic coma. Indian
Pediatr. 2003;40:620-625.
© 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.
at WUHAN UNIV LIBRARY on June 18, 2008 http://jcn.sagepub.comDownloaded from