Neurologic Critical Care

Effect of cerebral perfusion pressure augmentation on regional

oxygenation and metabolism after head injury*
Andrew J. Johnston, FRCA; Luzius A. Steiner, PhD; Jonathan P. Coles, PhD; Doris A. Chatfield, BSc;
Tim D. Fryer, PhD; Peter Smielewski, PhD; Peter J. Hutchinson, PhD; Mark T. OConnell, PhD;
Pippa G. Al-Rawi, BSc; Franklin I. Aigbirihio, PhD; John C. Clark, DSc; John D. Pickard, FRCS;
Arun K. Gupta, FRCA; David K. Menon, PhD

Objective: In this study we have used 15O positron emission

tomography, brain tissue oxygen monitoring, and cerebral microdialysis to assess the effects of cerebral perfusion pressure
augmentation on regional physiology and metabolism in the setting of traumatic brain injury.
Design: Prospective interventional study.
Setting: Neurosciences critical care unit of a university hospital.
Patients: Eleven acutely head-injured patients requiring norepinephrine to maintain cerebral perfusion pressure.
Interventions: Using positron emission tomography, we have
quantified the response to an increase in cerebral perfusion pressure
in a region of interest around a brain tissue oxygen sensor (Neurotrend) and microdialysis catheter. Oxygen extraction fraction and
cerebral blood flow were measured with positron emission tomography at a cerebral perfusion pressure of 70 mm Hg and 90 mm
Hg using norepinephrine to control cerebral perfusion pressure. All
other aspects of physiology were kept stable.
Measurements and Main Results: Cerebral perfusion pressure
augmentation resulted in a significant increase in brain tissue
oxygen (17 8 vs. 22 8 mm Hg; 2.2 1.0 vs. 2.9 1.0 kPa,
p < .001) and cerebral blood flow (27.5 5.1 vs. 29.7 6.0
mL/100 mL/min, p < .05) and a significant decrease in oxygen

extraction fraction (33.4 5.9 vs. 30.3 4.6 %, p < .05). There
were no significant changes in any of the microdialysis variables
(glucose, lactate, pyruvate, lactate/pyruvate ratio, glycerol). There
was a significant linear relationship between brain tissue oxygen
and oxygen extraction fraction (r2 .21, p < .05); the brain tissue
oxygen value associated with an oxygen extraction fraction of
40% (the mean value for oxygen extraction fraction in normal
controls) was 14 mm Hg (1.8 kPa). The cerebral perfusion pressure intervention resulted in a greater percentage increase in
brain tissue oxygen than the percentage decrease in oxygen
extraction fraction; this suggests that the oxygen gradients between the vascular and tissue compartments were reduced by the
cerebral perfusion pressure intervention.
Conclusions: Cerebral perfusion pressure augmentation significantly increased levels of brain tissue oxygen and significantly
reduced regional oxygen extraction fraction. However, these
changes did not translate into predictable changes in regional
chemistry. Our results suggest that the ischemic level of brain
tissue oxygen may lie at a level below 14 mm Hg (1.8 kPa);
however, the data do not allow us to be more specific. (Crit Care
Med 2005; 33:189 195)
KEY WORDS: head injury; ischemia; oxygenation; cerebral perfusion pressure; brain tissue oxygen

upport of cerebral perfusion

pressure (CPP) is one of the
mainstays of critical care in
head-injured patients, and
maintaining CPP 60 70 mm Hg is
thought to correlate with improved brain
oxygenation, autoregulation, and out-

come (1). However, both higher and

lower CPP thresholds have been suggested (25). Although vasoactive agents
such as norepinephrine are frequently
used to support CPP in these critically ill
patients, these agents may themselves influence cerebral blood flow and metabo-

*See also p. 255.

From University Department of Anaesthetics (AJJ,
LAS, JPC, DAC, AKG, DKM), Academic Neurosurgery
(LAS, PS, PJH, MTO, PGA-R, JDP), and Wolfson Brain
Imaging Centre (AJJ, LAS, JPC, DAC, TDF, PS, PJH,
MTO, FIA, JCC, JDP, DKM), Addenbrookes Hospital,
Cambridge, UK.
Supported, in part, by an unrestricted grant from
Codman, Raynham, MA (AJJ); by a Myron B. Laver
Grant (Department of Anesthesia, University of Basel,
Switzerland) (LAS); by grants from the Margarete und
Walter Lichtenstein-Stiftung (Basel, Switzerland) (LAS)

and the Swiss National Science Foundation (LAS); by

an Overseas Research Student Award (Committee of
Vice-Chancellors and Principals of the Universities of
the United Kingdom) (LAS); by a Wellcome research
training fellowship (JPC); by an Academy of Medical
Sciences/The Health Foundation Senior Surgical Scientist Fellowship (PJH); and by grant G 9439390
ID56833 from the Medical Research Council.

Crit Care Med 2005 Vol. 33, No. 1

Copyright 2005 by the Society of Critical Care

Medicine and Lippincott Williams & Wilkins
DOI: 10.1097/01.CCM.0000149837.09225.BD

lism. Furthermore, CPP interventions

may also result in cerebral (6) and systemic derangements (7). Recent changes
in the published guidelines for the management of traumatic brain injury (8)
suggest a lower CPP target (60 vs. 70 mm
Hg), since aggressive maintenance of
CPP levels above 70 mm Hg has been
associated with increased cardiovascular
and respiratory complications (7). Although there appears a clear overall detrimental effect on outcome, this appears
to be related to extracranial effects of
therapy. However, these data do not address the specific question of whether increases in CPP above 70 mm Hg provide
any benefit to the injured brain.
Indeed, despite a number of studies
reporting CPP thresholds for cerebral ox189

ygenation and metabolism (9 13), there

are fewer data on the cerebral responses
to further increases in CPP, with the literature reporting variable results (4, 10).
The objective of this study was to explore
whether CPP augmentation results in local metabolic benefits. Demonstration of
such benefits might prompt the search
for new approaches to CPP augmentation
that have fewer extracranial complications. Conversely, the absence of any
metabolic benefit would make such efforts less worthwhile.
Positron emission tomography (PET)
has recently been used to explore how
regional oxygenation (14) and tissue
chemistry (15) respond to hyperventilation in head-injured humans. Using Oxygen-15 to assess cerebral blood flow and
the cerebral metabolic rate of oxygen,
both studies demonstrated important relationships between regional physiology
and brain tissue oxygen/tissue chemistry.
However, the exact relationships between
tissue oxygen levels and tissue chemistry
are still not fully elucidated.
In this study, we have used 15O PET,
brain tissue oxygen monitoring, and cerebral microdialysis to assess the effects
of CPP augmentation on regional physiology and metabolism in the setting of
traumatic brain injury.

METHODS
Patient Selection. Following approval from
the local Research and Ethics Committee and
written informed assent from the patients
next of kin, 11 patients were studied. Patients
older than age 16 yrs with moderate to severe
head injury requiring sedation, ventilation, intracranial pressure monitoring, and norepinephrine infusions to support CPP were eligible for the study. Patients with coagulation
disorders or unstable physiology or who required a high (50%) inspired fraction of
oxygen were excluded.
Patients were managed according to Addenbrookes Neurosciences Critical Care Unit
protocols (16, 17), which include sedation
with propofol and fentanyl, paralysis with atracurium, and support of CPP to 70 mm Hg.
Monitored variables included electrocardiogram, peripheral oxygen saturation, end-tidal
carbon dioxide (Marquette Solar 8000M, GE
Medical Systems, UK), jugular venous oxygen
saturation, and intracranial pressure (Codman
MicroSensors ICP Transducer, Codman &
Shurtleff, Raynham, MA). Comprehensive
neurointensive care monitoring and treatment were maintained throughout the PET
scans in the adjacent Wolfson Brain Imaging
Centre. Apart from the CPP intervention, all
other aspects of physiology were kept as stable
as possible during the PET scans.

190

Intracranial Monitoring. Every patient had

a microdialysis catheter (Gold tipped CMA 70
10-mm membrane, CMA, Stockholm, Sweden)
and Neurotrend catheter (Codman, Raynham,
MA) inserted into the frontal cerebral parenchyma, in conjunction with the intracranial
pressure sensor using a triple-lumen cranial
access device (Technicam, Newton Abbot, UK)
(18). The Neurotrend sensor is a multivariable
sensor that measures parenchymal oxygen
(PbO2), parenchymal carbon dioxide, pH, and
temperature.
The microdialysis catheter was inserted to
a depth of 30 mm, and the Neurotrend sensor
was inserted to a depth of 45 mm; these depths
were calculated so that predominantly white
matter was being monitored and so that the
oxygen sensor on the Neurotrend catheter was
at approximately the same depth as the microdialysis membrane. The catheters were inserted 4 hrs before the PET scan to ensure
baseline stability.
Microdialysis. The microdialysis catheters
were perfused with Ringers solution (K 4
mmol/L, Na 147 mmol/L, Ca2 2 mmol/L,
Cl 155 mmol/L) at a rate of 0.3 L/min using
the CMA106 pump. Vials were changed at approximately 30-min intervals and were placed
onto dry ice and stored at 70C for later
analysis. The delay from the microdialysis
membrane to the collecting vial (the deadspace) is 17 mins, and this was accounted for
in the study design. Microdialysis samples
were analyzed for glucose, lactate, pyruvate,
and glycerol using a CMA600 bedside microdialysis analyzer.
PET Technique. Patients were positioned
in the PET scanner, and the infusion of norepinephrine was adjusted to maintain the CPP
at 70 mm Hg. PET scans were performed on
a GE Advance scanner (General Electric Medical Systems, Milwaukee, WI). Following a 10min transmission scan using rotating germanium-68 rods, performed to correct
subsequent emission scans for photon attenuation, emission data were acquired after, sequentially, a bolus inhalation of C15O (300
MBq over 60 secs, with data acquired during
the subsequent 5 mins in 3D mode); a steadystate inhalation of 15O (7200 MBq, with data
acquired in two 5-min frames in 2D mode
after a build-up period of 10 mins); and a
steady-state infusion of H215O (800 MBq, with
data acquired in two 5-min frames in 3D mode
after a build-up period of 10 mins). Arterial
blood gases were taken at regular intervals
throughout the scan, and if necessary minor
adjustments were made to the ventilator settings to keep arterial CO2 stable. Paired arterial
and jugular venous gases were taken at the
end of each 15O scan. CBF, cerebral blood
volume, oxygen extraction fraction (OEF), and
cerebral metabolic rate of oxygen were calculated from radioactivity concentrations in tissue (PET) and arterial blood (discrete samples)
using standard modeling techniques (19, 20)
with a volume of distribution set to 0.95.

After completion of the baseline scan, the

norepinephrine infusion was increased to
maintain a CPP 20 mm Hg above the baseline CPP. When the new CPP level had been
stable for 10 mins, the PET scans were repeated in reverse order. After the PET scan, a
radiograph computed tomography (CT) scan
was acquired to determine the location of the
monitoring probes. Physiologic data were recorded at regular intervals throughout the
PET scan.
Data Analysis. After reslicing CT images to
5 mm, PET and CT images were coregistered
and analyzed using custom-designed automated software (PETAN) (21) incorporating
elements of several different software packages, including Statistical Parametric Mapping (SPM99, Wellcome Department of Cognitive Neurology, London, UK), Matlab 5.2
(MathWorks, Natick, MA), and Analyze 5 (AnalyzeDirect, Lenexa, KS), and registration by
multiresolution optimization of mutual information (Department of Radiologic Sciences,
Guys Hospital, London, UK) (22, 23).
A 20-mm region of interest was defined on
the coregistered CT scan around the sensor on
three contiguous slices to represent the volume of brain measured with the microdialysis
catheter and the Neurotrend sensor. The size
of the region of interest has previously been
determined (14, 15), aiming to keep the region
as focal as possible while at the same time
limiting variable variance. By superimposing
this region of interest over the PET scans,
region of interest values for CBF, cerebral
blood volume, OEF, and cerebral metabolic
rate of oxygen were determined.
Data were analyzed using StatView 4 (SAS,
Cary, NC). Data were tested for normality of
distribution using the one-sample Kolmogorov-Smirnov test and were found to be distributed normally. Comparisons were made
using paired t-tests; correlations were defined
using linear regression analysis and analysis of
variance. These analyses have been designed as
hypothesis generating rather than hypothesis
testing; consequently, we have not provided
corrections for multiple comparisons.

RESULTS
Eleven patients with a major head injury were studied (nine males, two females). The mean age SD was 32 14
yrs; the median postresuscitation Glasgow Coma Score was 7 (range 39). Patients were studied on day 2.8 1.4 after
the injury. In one patient the catheters
were positioned in an area of brain that
appeared to be abnormal on radiograph
CT scan; in all other patients the probes
were located in areas of the brain that
appeared normal on CT.
Physiologic data during the course of
the study protocol are shown in Table 1.
Mean CPP (SD) for the first scan was
Crit Care Med 2005 Vol. 33, No. 1

67.9 3.4 mm Hg and for the second

scan 89.2 1.8 mm Hg. Baseline arterial
partial pressure of carbon dioxide was 33
3 mm Hg (4.23 0.33 kPa); this did
not change significantly between the
scans.
PET Data. PET data are presented in
Table 1. CPP augmentation resulted in a
significant increase in CBF and a significant reduction in OEF; however, even at
baseline these were not at levels that suggested ischemia.
Neurotrend Data. Neurotrend data are
presented in Table 2. Brain tissue oxygen
showed a high degree of variability. CPP
augmentation resulted in a significant increase in PbO2; other Neurotrend variables did not change significantly. In two
patients in whom the baseline PbO2 was
at an ischemic level (10 mm Hg, 1.3
kPa), the CPP intervention increased the
PbO2 to nonischemic levels; in one of
these patients the lactate/pyruvate (L/P)
ratio decreased from a high level to a
normal level (from 56 to 19). There was a
weak inverse relationship between the
baseline PbO2 and the change in PbO2
during the CPP intervention, with low
baseline levels of Pb O 2 exhibiting a
greater increase in PbO2 than high baseline levels (Fig. 1). However, this relationship did not reach statistical significance (r2 .35, p .06).
When the data from both CPP interventions were pooled, there was a significant linear relationship between PbO2
and OEF (r2 .21, p .05), and the PbO2
value associated with an OEF of 40% (the
mean value for OEF in normal controls)
(24) was 14 mm Hg (1.8 kPa, Fig. 2).
However, the spread of values was wide,
and patients with widely differing PbO2
values showed the same OEF (e.g., an
OEF of 32% was associated with PbO2
values that varied between 12 and 34 mm
Hg [1.6 and 4.5 kPa]). There was no correlation between PbO2 and CBF.
Microdialysis Data. Microdialysis variables are presented in Table 3. In two
patients we do not have complete micro-

dialysis data because there was insufficient dialysate to measure all the microdialysis variables. There were no
significant changes in any of the microdialysis variables (glucose, lactate, pyruvate, L/P ratio, glycerol) with CPP augmentation. Changes in microdialysis
variables did not correlate with other
changes in physiology.
Outcome Data. Outcome data are
available for nine of the 11 patients (two
patients were lost to follow up). Five patients had a favorable outcome (moderate
disability or good recovery) and four patients had a unfavorable outcome (dead,
vegetative state, or severe disability).
There was a significant difference in baseline L/P ratio between patients in these
two groups (17 5 vs. 36 17, p .05,
unpaired Students t-test). However, no
other significant differences were found.

DISCUSSION
We have used 15O PET imaging, microdialysis, and tissue oxygen monitoring
to explore how regional oxygenation and
chemistry respond to an increase in CPP.
CPP Intervention. CPP intervention
led to a significant increase in PbO2 and
CBF and a significant reduction in OEF;
however, this was not accompanied by
significant changes in tissue chemistry.
Three patients had a baseline L/P ratio of
25 (25 being regarded as the upper
limit of normal) (25, 26); in one of these
patients the L/P ratio decreased to a normal level during CPP augmentation. Although the change in CBF was statistically significant, it may not represent a
clinically significant change. However,
we believe the corresponding oxygenation changes to be clinically significant.
It is interesting to note that the CPP
intervention resulted in a greater percentage increase in PbO2 than the percentage decrease in OEF. Both OEF and
PbO2 represent the balance between oxygen supply and demand; however, the
measurements are made from different

Table 1. Physiologic data during the course of the study protocol

Baseline CPP
Intervention CPP
p

CPP,
mm Hg

CBV,
mL/100 mL

CBF,
mL/100
mL/min

CMRO2
mol/100
mL/min

OEF,
%

67.9 3.4
89.2 1.8
.0001

2.8 0.6
3.0 0.5
NS

27.5 5.1
29.7 6.0
.05

50.0 10.0
49.1 10.7
NS

33.4 5.9
30.3 4.6
.05

CPP, cerebral perfusion pressure; CBV, cerebral blood volume; CBF, cerebral blood flow; CMRO2,
cerebral metabolic rate of oxygen; OEF, oxygen extraction fraction; NS, not significant.

Crit Care Med 2005 Vol. 33, No. 1

anatomical compartments. This suggests

that the oxygen gradients between the
vascular and tissue compartments were
reduced by the CPP intervention (27).
Below an autoregulatory threshold for
CPP, PbO2 is very dependent on CPP (10,
28, 29). Our data are consistent with others indicating that supranormal levels of
CPP may further increase PbO2 levels (4,
30). At least two studies have specifically
looked at the effects of CPP augmentation
on PbO2 levels. Kiening and colleagues
(10) increased CPP from 68 to 84 mm Hg
in 20 patients with a mixture of traumatic
brain injury and intracerebral hemorrhage. Dopamine was used to increase
the CPP, and an unspecified Clark-type
microcatheter was used to measure PbO2
in tissue that appeared normal on CT
scan. The authors demonstrated a small
increase in PbO2; however, this was not
statistically significant. There are a number of explanations for this discordance
between these data and our results. The
CPP intervention in Kiening et al.s study
was not as large as in our study; furthermore, the vasoactive agent used to
achieve the CPP intervention may have
influenced the results. There is increasing evidence that exogenous dopamine
and norepinephrine may directly influence cerebral blood flow and metabolism,
independent of any effects on systemic
hemodynamics. This may, in part, explain
the differences between the studies (31,
32). Stocchetti and colleagues (4) placed
a Licox sensor in abnormal brain tissue in
nine patients (seven subarachnoid hemorrhage, one traumatic brain injury, one
meningioma). The authors studied the
effect of both spontaneous and induced
(norepinephrine) increases in CPP on
PbO2. The authors found a positive correlation between the change in CPP and the
change in PbO2. Patients in whom baseline PbO2 values were at low levels demonstrated a greater increase in PbO2 in
response to CPP augmentation. These
data are in accordance with our results.
However, our results suggest that abnormality on radiograph CT does not reliably
predict the response to CPP intervention.
Exclusion of the one patient in whom the
oxygen sensor was positioned in abnormal tissue did not make any significant
difference in our results. In fact, exclusion of this patient resulted in a stronger
relationship between baseline PbO2 and
the percentage change in PbO2 in response to CPP augmentation (r2 .4, p
.05). Failure of autoregulation in hy191

Table 2. Neurotrend data

Baseline CPP
Intervention CPP
p

PaO2,
mm Hg (kPa)

PbO2,
mm Hg (kPa)

PbCO2,
mm Hg (kPa)

Brain pH

105 11 (13.8 1.5)

107 14 (14.0 1.8)
NS

17 8 (2.2 1.0)
22 8 (2.9 1.0)
.001

43 5 (5.6 0.7)
43 4 (5.6 0.5)
NS

7.2 0.1
7.2 0.1
NS

PbO2, parenchymal oxygen; PbCO2, parenchymal carbon dioxide; CPP, cerebral perfusion pressure; NS, not significant.

Figure 1. The relationship between the baseline brain tissue oxygen (PbO2 ) and the percentage change
in PbO2 during the cerebral perfusion pressure intervention.

Figure 2. The relationship between brain tissue oxygen (PbO2) and oxygen extraction fraction (OEF) in
a region of interest of the brain. The upper limit of normal OEF (40%) (24) corresponds to a PbO2
of 1.8 kPa (14 mm Hg).

poxic tissue may partly explain these results.

There are a number of possible reasons why reductions in OEF and in192

creases in PbO2 do not translate into predictable changes in tissue chemistry.

First, the relationships between PbO2 and
tissue chemistry are poorly understood.

There is limited literature evidence; however, one study examining the relationship between PbO2 and tissue chemistry
found that large increases in microdialysis variables normally occurred only
when tissue oxygen levels fell to very low
levels (33). Second, despite the increase
in tissue oxygen levels, the ischemic burden may not actually decrease, either because there is minimal ischemic burden
at baseline (as the OEF figures suggest)
or because mitochondrial dysfunction is a
barrier to oxidative glucose use (34, 35).
Third, changes in tissue chemistry may
occur over a longer time frame than we
followed. The absence of change could
represent a type 2 statistical error, and
larger studies may give different results;
however, the heterogeneity in neurochemistry (Table 3) suggests that such
studies would have to recruit many patients.
Correlations Between PET and Neurotrend. In line with an earlier PET and
Neurotrend study (14), we did not find a
correlation between PbO2 and regional
CBF. Basic physiology would suggest that
as CBF decreased then PbO2 would decrease. Indeed, in studies examining the
effects of temporary artery clipping during intracerebral aneurysm surgery, application of the clip does lead to a reduction in PbO2 (36, 37). A similar effect has
also been shown in a swine model where
a stepwise reduction in CBF resulted in a
stepwise reduction in PbO2 (38). In human TBI studies there is conflicting evidence regarding the relationship between
PbO2 and CBF. Menzel and colleagues
(39) and Doppenberg and colleagues (40)
used a single stable Xenon-CT scan to
measure CBF in a region of interest
around a Paratrend probe and found reasonable correlations with PbO2. However,
in both this current study and a previous
study from our group (14), no such correlation was found. There are a number
of possible reasons for this discordance.
First, both Menzel et al. (39) and Doppenberg et al. (40) studied a number of patients in whom regional CBF was at levels
Crit Care Med 2005 Vol. 33, No. 1

Table 3. Microdialysis variables

Baseline CPP
Intervention CPP
p

Lactate, mmol/L

Pyruvate, mol/L

Lactate/Pyruvate Ratio

Glucose, mmol/L

Glycerol, mol/L

2.1 1.6
1.5 0.7
NS

70.3 26.1
68.7 23.4
NS

29.0 17.7
23.3 16.1
NS

1.2 0.8
1.2 0.9
NS

96.5 87.2
84.1 57.5
NS

CPP, cerebral perfusion pressure; NS, not significant.

much lower than in the current study. In

the patients with very low levels of CBF,
PbO2 values were found to be low and this
meant that significant polynomial regression curves could be fitted; however,
across a range of other CBF levels, PbO2
appeared to be very heterogeneous. Second, our results may arise from differing
methods of monitoring brain tissue oxygen. Both Menzel et al. and Doppenberg
et al. used Paratrend sensors to measure
brain tissue oxygen. Paratrend sensors
were designed to measure arterial partial
pressures of oxygen, and they may not be
accurate at low oxygen tensions. Indeed,
Paratrend sensors have been compared
with Licox oxygen sensors and have been
shown to be potentially inaccurate at low
oxygen tensions (41). The Neurotrend
sensor is calibrated to work at the lower
oxygen partial pressures that are found in
the brain. To our knowledge, Neurotrend
and Licox have not been directly compared. Differences between the oxygensensing technology used may result in
differences in results. Third, there may be
a relationship between CBF and PbO2 in
individual patients. However, the varying
degrees of metabolic suppression, edema,
and mitochondrial disruption seen across
a range of patients may mean that a
meaningful relationship between PbO2
and CBF is less easy to demonstrate. Finally, we may be looking at the wrong
physiologic variable in these correlations;
consideration of the physiology would
suggest that a relationship between PbO2
and OEF would be more likely than one
between PbO2 and CBF.
Brain tissue oxygen is the partial pressure of oxygen in the brain extracellular
space, and a consideration of basic physiology implies that PbO2 is in equilibrium
with end-capillary or venous PO2 (PO2). At
least in health, this appears to be the case
(42). Normal metabolic activity may be
maintained in the brain, in the face of
reductions in cerebral blood flow, by increasing OEF. However, when this capacity is exhausted, the cerebral metabolic
rate for oxygen becomes compromised
and neurologic impairment may follow
Crit Care Med 2005 Vol. 33, No. 1

(43). The OEF at which metabolic derangement occurs has not been fully elucidated; however, it probably lies in the
region of 75% (44). A consideration of
these physiologic premises would suggest
that increases in OEF will result in reductions in PO2 and hence reductions in
PbO2. Indeed, our results show that CPP
intervention leads to a significant reduction in OEF and a significant increase in
PbO2. Interpretation of these changes is
complicated, but examination of the relevant literature provides some insight
into the pathophysiology. Most cells are
only a few micrometers from a capillary,
and so there is only a small distance for
oxygen to diffuse (45). However, the volume of tissue that contributes to oxygen
readings derived from a tissue sensor is
likely to be a few cubic millimeters (14).
This means that tissue oxygen measurements are the average value from a small
capillary bed. Within this capillary bed
there are likely to be areas of patchy microvascular collapse that would result in
significant increases in tissue path
lengths for oxygen (42). Local capillaries
may be able to unload more oxygen, but
other capillaries are unlikely to be able do
so as the diffusion distances will be too
large. Furthermore, perivascular edema
will contribute to further increases in diffusion barriers for oxygen delivery. Overall, this may result in a situation in which
PbO2 is low while the OEF and endcapillary oxygen tension are within normal limits; that is, there is a capillarytissue oxygen gradient across the
capillary bed. CPP augmentation leads to
a number of physiologic changes. Cerebral blood flow increases as collapsed,
and partially obstructed microvessels are
recruited. As vessels are recruited and
oxygen delivery increases, the difference
between the end-capillary oxygen and the
tissue oxygen will reduce; that is, the
gradient will reduce. These changes will
result in an increase in PbO2 and a reduction in OEF; however, there will be a
greater relative change in PbO2 than in
OEF. This pathophysiological model is in
accordance with our results; however,

our data do not allow us to be more

specific about the mechanisms involved.
A further explanation could be that a failure of autoregulation would lead to an
increase in CBF in response to an increase in CPP. This would lead to increases in PbO2 and a reduction in OEF,
but it would not explain why PbO2 increases to a greater extent than OEF falls.
Although OEF was related to PbO2,
this correlation was poor, with an r2 value
of .21. This poor correlation is important
and suggests that although OEF may be a
significant predictor of PbO2, factors
other than the adequacy of cerebrovascular oxygen delivery have an important
effect on PbO2 in this setting. One major
cause of the variable relationships between OEF and PbO2 may be the presence
of microvascular heterogeneity, which we
have previously described (14, 27, 46).
Another reason may be the position of the
oxygen sensor in relation to regional vasculature (45); however, our data do not
allow us to be more specific about this.
Although close proximity to arterial vessel may result in elevated levels of PbO2, a
recent study comparing PbO2 with regional end-capillary oxygen (PO2) did not
find any values of PbO2 that were higher
than PO2 (14). Although absolute values
for PbO2 correlate poorly with absolute
values of PO2 after traumatic brain injury,
changes in the two compartments correlate well. These findings are consistent
with PbO2 being a marker of tissue and
venous levels of oxygen rather than arterial levels and are also consistent with the
presence of oxygen diffusion gradients
between the capillary and tissue compartments (42).
As far as we are aware, we are the first
to try to define ischemic thresholds for
PbO2 using PET imaging and regional oxygen extraction fraction. Although we
found wide interindividual variability in
PbO2, we found a significant relationship
between OEF and PbO2, and using this
relationship we used normal OEF levels
(24) to suggest that the lower limit of
normal PbO2 is of the order of 14 mm
Hg (1.8 kPa). However, considerable in193

erebral perfusion
pressure augmentation significantly

increased levels of brain tissue oxygen and significantly

reduced regional oxygen extraction fraction.

terpatient heterogeneity, the fact that

none of our patients had OEF levels that
are classically thought to be consistent
with ischemia, and the relatively poor
relationship between OEF and PbO2 mean
that this lower limit of normal PbO2 must
be interpreted with considerable caution.
The ischemic threshold for PbO2 may lie
below this level, although our data do not
allow us to be more specific.
Ischemic thresholds for PbO2 have
been variably described using a number
of different approaches such as outcome
analysis after head injury (10, 39 41, 47
50), relating PbO2 to recognized threshold limits for CBF (39, 48), relating PbO2
to jugular bulb oxygen saturation limits
(51), assessing PbO2 in patients with a
compromised cerebral circulation (52),
and assessing thresholds for infarction
during cerebral aneurysm clipping (37).
Although there is a large variation in
reported ischemic thresholds, the most
commonly accepted threshold is approximately 10 mmHg (1.3 kPa); our data are
consistent with a threshold that lies at
this level. Attempts to define an ischemic
threshold for PbO2 are beset by a number
of problems. The most important of these
is the difficulty in identifying whether the
brain was actually ischemic. For example,
if CBF and SjO2 are used to identify ischemic thresholds, then it is assumed that
the sensitivity and specificity of CBF and
SjO2 for ischemia are 100%. When outcome analysis is used to define ischemic
thresholds, then the results are very dependent on whether the thresholds were
defined prospectively or retrospectively
and on how a good outcome was defined. Terminal events are frequently included in outcome analysis and may skew
results. The monitoring technology used
to measure PbO2 varies widely between
different studies, and this may well influence the results. For these reasons, there
194

remains considerable controversy about

where the ischemic threshold for PbO2
actually lies and for how long brain tissue
must lie at or below this threshold before
irreversible damage is done. Although
our data give further insights into ischemic thresholds for PbO2, we do not believe that this threshold has yet been determined. Cellular metabolism depends
on cellular oxygen levels, and as yet, despite the potential of neurochemical and
tissue oxygen monitoring, we do not have
the means to determine when cellular
oxygen levels have fallen to levels where
irreversible damage will be done unless
therapeutic interventions are instigated.

CONCLUSIONS
In the injured brain, considerable interpatient heterogeneity exists in regional oxygenation and chemistry. Cerebral perfusion pressure augmentation
from a baseline of 70 mm Hg significantly increased levels of brain tissue oxygen and significantly reduced regional
oxygen extraction fraction. However,
these changes did not translate into predictable changes in regional chemistry.
The observed increases in PbO2 may, in
part, result from reductions in capillarytissue oxygen gradients.
Regional oxygen extraction fraction is
a predictor of brain tissue oxygen; however, the association is weak. This suggests that factors other than the adequacy
of cerebrovascular oxygen delivery have
an important effect on PbO2. Our results
suggest that the ischemic level of PbO2
may lie at a level below 14 mm Hg (1.8
kPa); however, the data do not allow us to
be more specific.

5.

6.

7.

8.

9.

10.

11.

12.

ACKNOWLEDGMENTS
We thank Dr. Raymond Salvador for
statistical advice.

13.

REFERENCES
1. The Brain Trauma Foundation. The American Association of Neurological Surgeons.
The Joint Section on Neurotrauma and Critical Care: Guidelines for cerebral perfusion
pressure. J Neurotrauma 2000; 17:507511
2. Changaris DG, McGraw CP, Richardson JD,
et al: Correlation of cerebral perfusion pressure and Glasgow Coma Scale to outcome.
J Trauma 1987; 27:10071013
3. Rosner MJ, Rosner SD, Johnson AH: Cerebral
perfusion pressure: Management protocol
and clinical results. J Neurosurg 1995; 83:
949 962
4. Stocchetti N, Chieregato A, De Marchi M, et

14.

15.

16.

al: High cerebral perfusion pressure improves low values of local brain tissue O2
tension (PtiO2) in focal lesions. Acta Neurochir Suppl 1998; 71:162165
Asgeirsson B, Grande PO, Nordstrom CH: A
new therapy of post-trauma brain oedema
based on haemodynamic principles for brain
volume regulation. Intensive Care Med 1994;
20:260 267
Durward QJ, Del Maestro RF, Amacher AL, et
al: The influence of systemic arterial pressure and intracranial pressure on the development of cerebral vasogenic edema. J Neurosurg 1983; 59:803 809
Robertson CS, Valadka AB, Hannay HJ, et al:
Prevention of secondary ischemic insults after severe head injury. Crit Care Med 1999;
27:2086 2095
The Brain Trauma Foundation. The American Association of Neurological Surgeons.
The Joint Section on Neurotrauma and Critical Care: Update notice. Guidelines for the
Management of Severe Traumatic Brain Injury: Cerebral Perfusion Pressure, 2003.
Available at: http://www2.braintrauma.
org/guidelines/downloads/btf guidelines
cpp 1.pdf?BrainTrauma Session78d
84ca82e7f00d5eec2b57925d070f6
Vespa P, Prins M, Ronne-Engstrom E, et al:
Increase in extracellular glutamate caused by
reduced cerebral perfusion pressure and seizures after human traumatic brain injury: A
microdialysis study. J Neurosurg 1998; 89:
971982
Kiening KL, Hartl R, Unterberg AW, et al:
Brain tissue pO2-monitoring in comatose patients: implications for therapy. Neurol Res
1997; 19:233240
Bruzzone P, Dionigi R, Bellinzona G, et al:
Effects of cerebral perfusion pressure on
brain tissue PO2 in patients with severe head
injury. Acta Neurochir Suppl 1998; 71:
111113
Chan KH, Miller JD, Dearden NM, et al: The
effect of changes in cerebral perfusion pressure upon middle cerebral artery blood flow
velocity and jugular bulb venous oxygen saturation after severe brain injury. J Neurosurg 1992; 77:55 61
Zauner A, Doppenberg E, Woodward JJ, et al:
Multiparametric continuous monitoring of
brain metabolism and substrate delivery in
neurosurgical patients. Neurol Res 1997; 19:
265273
Gupta AK, Hutchinson PJ, Fryer T, et al:
Measurement of brain tissue oxygenation
performed using positron emission tomography scanning to validate a novel monitoring
method. J Neurosurg 2002; 96:263268
Hutchinson PJ, Gupta AK, Fryer TF, et al:
Correlation between cerebral blood flow,
substrate delivery, and metabolism in head
injury: A combined microdialysis and triple
oxygen positron emission tomography study.
J Cereb Blood Flow Metab 2002; 22:735745
Menon DK: Cerebral protection in severe
brain injury: Physiological determinants of

Crit Care Med 2005 Vol. 33, No. 1

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

outcome and their optimisation. Br Med Bull

1999; 55:226 258
Patel HC, Menon DK, Tebbs S, et al: Specialist neurocritical care and outcome from head
injury. Intensive Care Med 2002; 28:547553
Hutchinson PJ, Hutchinson DB, Barr RH, et
al: A new cranial access device for cerebral
monitoring. Br J Neurosurg 2000; 14:46 48
Lammertsma AA, Baron JC, Jones T: Correction for intravascular activity in the oxygen-15 steady-state technique is independent
of the regional hematocrit. J Cereb Blood
Flow Metab 1987; 7:372374
Frackowiak RS, Lenzi GL, Jones T, et al:
Quantitative measurement of regional cerebral blood flow and oxygen metabolism in
man using 15O and positron emission tomography: Theory, procedure, and normal
values. J Comput Assist Tomogr 1980;
4:727736
Smielewski P, Coles JP, Fryer TD, et al: Integrated image analysis solutions for PET
datasets in damaged brain. J Clin Monit
Comput 2002; 17:427 440
Studholme C, Hill DL, Hawkes DJ: Automated 3-D registration of MR and CT images
of the head. Med Image Anal 1996;
1:163175
Studholme C, Hill DL, Hawkes DJ: Automated three-dimensional registration of
magnetic resonance and positron emission
tomography brain images by multiresolution
optimization of voxel similarity measures.
Med Phys 1997; 24:2535
Leenders KL, Perani D, Lammertsma AA, et
al: Cerebral blood flow, blood volume and
oxygen utilization. Normal values and effect
of age. Brain 1990; 113:27 47
Persson L, Valtysson J, Enblad P, et al: Neurochemical monitoring using intracerebral
microdialysis in patients with subarachnoid
hemorrhage. J Neurosurg 1996; 84:606 616
Reinstrup P, Stahl N, Mellergard P, et al:
Intracerebral microdialysis in clinical practice: Baseline values for chemical markers
during wakefulness, anesthesia, and neurosurgery. Neurosurgery 2000; 47:701710
Johnston AJ, Steiner LA, Coles JP, et al: Effects of cerebral perfusion pressure augmentation on capillary-tissue oxygen gradients
after acute brain injury. J Neurotrauma
2002; 19:P212
Lang EW, Czosnyka M, Mehdorn HM: Tissue
oxygen reactivity and cerebral autoregulation after severe traumatic brain injury. Crit
Care Med 2003; 31:267271
Soehle M, Jaeger M, Meixensberger J: Online

Crit Care Med 2005 Vol. 33, No. 1

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

assessment of brain tissue oxygen autoregulation in traumatic brain injury and subarachnoid hemorrhage. Neurol Res 2003; 25:
411 417
Meixensberger J, Jaeger M, Vath A, et al:
Brain tissue oxygen guided treatment supplementing ICP/CPP therapy after traumatic
brain injury. J Neurol Neurosurg Psychiatry
2003; 74:760 764
Johnston AJ, Steiner LA, Chatfield DA, et al:
Effect of cerebral perfusion pressure augmentation with dopamine and norepinephrine on global and focal brain oxygenation
after traumatic brain injury. Intensive Care
Med 2004; 30:791797
Steiner LA, Johnston AJ, Czosnyka M, et al:
Direct comparison of cerebrovascular effects
of norepinephrine and dopamine in headinjured patients. Crit Care Med 2004; 32:
1049 1054
Valadka AB, Goodman JC, Gopinath SP, et al:
Comparison of brain tissue oxygen tension to
microdialysis-based measures of cerebral
ischemia in fatally head-injured humans.
J Neurotrauma 1998; 15:509 519
Verweij BH, Muizelaar JP, Vinas FC, et al:
Impaired cerebral mitochondrial function after traumatic brain injury in humans. J Neurosurg 2000; 93:815 820
Clausen T, Zauner A, Levasseur JE, et al:
Induced mitochondrial failure in the feline
brain: Implications for understanding acute
post-traumatic metabolic events. Brain Res
2001; 908:35 48
Doppenberg EM, Watson JC, Broaddus WC,
et al: Intraoperative monitoring of substrate
delivery during aneurysm and hematoma
surgery: Initial experience in 16 patients.
J Neurosurg 1997; 87:809 816
Kett-White R, Hutchinson PJ, Al-Rawi PG, et
al: Cerebral oxygen and microdialysis monitoring during aneurysm surgery: Effects of
blood pressure, cerebrospinal fluid drainage,
and temporary clipping on infarction. J Neurosurg 2002; 96:10131019
Rossi S, Stocchetti N, Longhi L, et al: Brain
oxygen tension, oxygen supply, and oxygen
consumption during arterial hyperoxia in a
model of progressive cerebral ischemia.
J Neurotrauma 2001; 18:163174
Menzel M, Doppenberg EM, Zauner A, et al:
Cerebral oxygenation in patients after severe
head injury: Monitoring and effects of arterial hyperoxia on cerebral blood flow, metabolism and intracranial pressure. J Neurosurg
Anesthesiol 1999; 11:240 251
Doppenberg EM, Zauner A, Bullock R, et al:

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

Correlations between brain tissue oxygen

tension, carbon dioxide tension, pH, and cerebral blood flowA better way of monitoring the severely injured brain? Surg Neurol
1998; 49:650 654
Valadka AB, Gopinath SP, Contant CF, et al:
Relationship of brain tissue PO2 to outcome
after severe head injury. Crit Care Med 1998;
26:1576 1581
Menon DK, Coles JP, Gupta AK, et al: Diffusion limited oxygen delivery following head
injury. Crit Care Med 2004; 32:1384 1390
Yonas H, Pindzola RR: Physiological determination of cerebrovascular reserves and its
use in clinical management. Cerebrovasc
Brain Metab Rev 1994; 6:325340
Coles JP, Fryer TD, Smielewski P, et al: Incidence and mechanisms of cerebral ischemia in early clinical head injury. J Cereb
Blood Flow Metab 2004; 24:202211
Scheufler KM, Rohrborn HJ, Zentner J: Does
tissue oxygen-tension reliably reflect cerebral oxygen delivery and consumption?
Anesth Analg 2002; 95:10421048
Coles JP, Fryer TD, Gupta AK, et al: Diffusion
limited oxygen delivery following head injury. J Neurotrauma 2001; 18:L11
Bardt TF, Unterberg AW, Hartl R, et al: Monitoring of brain tissue PO2 in traumatic brain
injury: Effect of cerebral hypoxia on outcome. Acta Neurochir Suppl 1998; 71:
153156
Doppenberg EM, Zauner A, Watson JC, et al:
Determination of the ischemic threshold for
brain oxygen tension. Acta Neurochir Suppl
1998; 71:166 169
van den Brink WA, van Santbrink H, Steyerberg EW, et al: Brain oxygen tension in severe head injury. Neurosurgery 2000; 46:
868 868
Zauner A, Doppenberg EM, Woodward JJ,
et al: Continuous monitoring of cerebral
substrate delivery and clearance: Initial experience in 24 patients with severe acute
brain injuries. Neurosurgery 1997; 41:
10821083
Kiening KL, Unterberg AW, Bardt TF, et al:
Monitoring of cerebral oxygenation in patients with severe head injuries: Brain tissue
PO2 versus jugular vein oxygen saturation.
J Neurosurg 1996; 85:751757
Hoffman WE, Charbel FT, Edelman G: Brain
tissue oxygen, carbon dioxide, and pH in
neurosurgical patients at risk for ischemia.
Anesth Analg 1996; 82:582586

195