Brain tissue oxygen tension and its response to

laboratory investigationJ neurosurg 125:1217–1228, 2016

abbreviations ABG = arterial blood gas; CCI = controlled cortical impact; CPP = cerebral perfusion pressure; FiO2 = fraction of inspired oxygen; GLM = general linear model; ICP = intracranial pressure; LMM = linear mixed model; MAP = mean arterial pressure; PaCO2 = partial pressure of carbon dioxide in arterial blood; PaO2 = partial pressure of oxygen in arterial blood; PbtO2 = brain tissue oxygen tension; REML = restricted maximum likelihood modeling; TBI = traumatic brain injury.submitted April 16, 2015. accepted July 28, 2015.include when citing Published online February 5, 2016; DOI: 10.3171/2015.7.JNS15809.

Brain tissue oxygen tension and its response to physiological manipulations: influence of distance from injury site in a swine model of traumatic brain injurygregory w. J. hawryluk, md, phd, Frcsc,1–3 nicolas phan, md, Frcsc,3,4 adam r. Ferguson, phd,2,3 diane morabito, rn, mph,2,3 nikita derugin, ms,2,3 campbell l. stewart, md,5 m. margaret Knudson, md,6 geoffrey manley, md, phd,2,3 and guy rosenthal, md2,3,7

1Department of Neurological Surgery, University of Utah, Salt Lake City, Utah; 2Department of Neurological Surgery, 3Brain and Spinal Injury Center, and 6Department of General Surgery, University of California, San Francisco, California; 4Division of Neurological Surgery, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Ontario, Canada; 5Department of Dermatology, University of Pennsylvania, Philadelphia, Pennsylvania; and 7Department of Neurosurgery, Hadassah-Hebrew University Medical Center, Jerusalem, Israel

obJective The optimal site for placement of tissue oxygen probes following traumatic brain injury (TBI) remains unresolved. The authors used a previously described swine model of focal TBI and studied brain tissue oxygen tension (PbtO2) at the sites of contusion, proximal and distal to contusion, and in the contralateral hemisphere to determine the effect of probe location on PbtO2 and to assess the effects of physiological interventions on PbtO2 at these different sites.methods A controlled cortical impact device was used to generate a focal lesion in the right frontal lobe in 12 anes-thetized swine. PbtO2 was measured using Licox brain tissue oxygen probes placed at the site of contusion, in pericon-tusional tissue (proximal probe), in the right parietal region (distal probe), and in the contralateral hemisphere. PbtO2 was measured during normoxia, hyperoxia, hypoventilation, and hyperventilation.results Physiological interventions led to expected changes, including a large increase in partial pressure of oxygen in arterial blood with hyperoxia, increased intracranial pressure (ICP) with hypoventilation, and decreased ICP with hyperventilation. Importantly, PbtO2 decreased substantially with proximity to the focal injury (contusion and proximal probes), and this difference was maintained at different levels of fraction of inspired oxygen and partial pressure of car-bon dioxide in arterial blood. In the distal and contralateral probes, hypoventilation and hyperventilation were associated with expected increased and decreased PbtO2 values, respectively. However, in the contusion and proximal probes, these effects were diminished, consistent with loss of cerebrovascular CO2 reactivity at and near the injury site. Similarly, hy-peroxia led to the expected rise in PbtO2 only in the distal and contralateral probes, with little or no effect in the proximal and contusion probes, respectively.conclusions PbtO2 measurements are strongly influenced by the distance from the site of focal injury. Physiological alterations, including hyperoxia, hyperventilation, and hypoventilation substantially affect PbtO2 values distal to the site of injury but have little effect in and around the site of contusion. Clinical interpretations of brain tissue oxygen measure-ments should take into account the spatial relation of probe position to the site of injury. The decision of where to place a brain tissue oxygen probe in TBI patients should also take these factors into consideration.http://thejns.org/doi/abs/10.3171/2015.7.JNS15809Key words traumatic brain injury; oxygen monitor; optimal location; licox; in vivo

©AANS, 2016 J neurosurg Volume 125 • November 2016 1217

Unauthenticated | Downloaded 05/28/22 10:22 PM UTC

g. w. J. hawryluk et al.

J neurosurg Volume 125 • November 20161218

While tissues such as muscle can tolerate ische-mia for hours, cells of the brain die after only minutes of anoxia.8,12 Cerebral ischemic in-

jury is a significant problem after traumatic brain injury (TBI);44 ischemic changes are seen in 80% of brains of patients who die following TBI.19 Given the limited re-generative capacity of the CNS, as well as evidence that brain tissue hypoxia after TBI is predictive of poor out-come,3,4,16,23,26,32,35,48,49 efforts to minimize brain tissue hy-poxia following TBI are warranted.

To reliably prevent brain tissue hypoxia, it is important to be able to continuously measure cerebral oxygenation, either directly or indirectly.33 For over 2 decades brain tis-sue oxygen probes have been available, allowing practitio-ners of neurocritical care to continuously monitor brain tissue oxygen tension (PbtO2) directly.11 Importantly, these invasive brain tissue oxygen probes provide focal measure-ments reflecting oxygen tension within a small volume of brain near the sensor. The limited knowledge regarding optimal probe placement in the injured brain remains a major limitation when brain tissue oxygen probes are used to monitor TBI patients.

Several studies have investigated the relevance of brain tissue oxygen probe location in human patients; however, these investigations have studied data from different pa-tients, each with a single probe located at different distanc-es from a focal brain lesion. In these studies, the distance of the probe from focal injury was not planned, since the probes are placed in standard locations in the frontal lobes and the distance from focal injury varies.9,14,20,35 Therefore, it is difficult to draw definitive conclusions regarding opti-mal probe location based on these clinical studies. A labo-ratory study performed under controlled conditions that involves placement of multiple concurrent probes may be better able to provide data concerning the relative advan-tages of different probe locations. Our study used a previ-ously characterized porcine model of controlled cortical impact (CCI),24 which generates a focal contusion injury to determine how measurements of PbtO2 are affected by distance from a focal injury in the gyrencephalic brain. We also sought to determine how concurrent readings at distinct sites responded to physiological alterations includ-ing hyperoxia, hyperventilation, and hypoventilation in a controlled fashion that could not be easily achieved in hu-mans.20

methodsThis experiment employed our previously described

CCI model in swine that mimics the gross anatomical and histological features of human focal brain injury.24 The experimental protocol was approved by the Committee on Animal Research at the University of California, San Francisco.

animal preparationWe used male Yorkshire swine weighing from 35 to

45 kg. Surgeries and perioperative intensive care were performed in the large animal facility at the University of California, San Francisco.24 As previously described,24 animals were premedicated with ketamine (20 mg/kg)

and xylazine (2 mg/kg). They were paralyzed with 0.05 mg/kg of pancuronium and were mechanically ventilated following intubation (Narkovet 2; Drager). Anesthesia consisted of a fentanyl infusion (1 mg/kg/hr) and inhaled isoflurane (0.5%–2%). Inspired oxygen concentration and end-tidal CO2 were monitored continuously. An arterial line facilitated arterial blood gas (ABG) analysis. A Swan-Ganz catheter was employed to monitor cardiac indices. Body temperature (38.5°C ± 1.0°C) was maintained with the use of a forced air warmer (BAIR Hugger; Augustine Medical). Animals were additionally monitored by elec-trocardiography, pulse oximetry, and placement of a uri-nary catheter to measure urine output.

controlled cortical impactTo secure the heads of the swine, a modified animal

head frame was applied (David Kopf Instruments). The surgical sites were then cleansed, sterilized, and draped. A scalp incision was made to facilitate placement of a 15-mm bur hole 7 mm anterior to the coronal suture and 3 mm to the right side of the midline.24 The right fron-tal dura was exposed but kept intact. The head was then placed in the CCI device24 (Bioengineering Department, Medical College of Virginia). CCI was performed using the following parameters: velocity 3.5 m/sec, dwell time 400 msec, and depth of depression 11 mm. These param-eters lead to a well-characterized focal contusion associ-ated with subarachnoid hemorrhage.24

brain tissue oxygen monitor placementFour Licox brain tissue oxygen probes (Integra Life-

sciences) were placed in each animal (Fig. 1) in the same fashion and locations previously employed by our group.41 Before placement, all probes were tested to ensure proper functioning. Distal and contralateral probes were placed prior to the CCI while the probes placed in the contusion and the penumbra were placed through the right frontal bur hole subsequent to the CCI at that site. Probes were inserted to a depth of 12 mm using a stereotactic device (David Kopf Instruments) and micromanipulators. The distal probe was placed through a separate bur hole 3.5 mm posterior to the posterior edge of the bur hole at the right frontal impact site and 7 mm to the right of mid-line. The contralateral probe was placed 3.5 mm posterior to the posterior edge of the bur hole at the right frontal impact site and 7 mm to the left of midline. Following CCI, the contusion probe was inserted through the center of the bur hole used for the impact injury. The proximal probe was placed at the posterior margin of the impact injury bur hole. All bur holes were then sealed with dental impression material (CutterSil; Heraeus Kulzer GmbH) to reduce leakage of CSF and to secure the brain tissue probes in place.

data collectionPhysiological data were collected continuously through-

out the experiment using a computerized data acquisition system designed by Aristein Bioinformatics LLC. Data were collected at 1-min intervals. Measured parameters included heart rate; systolic, diastolic, and mean arterial


brain tissue oxygen probe location

J neurosurg Volume 125 • November 2016 1219

blood pressure (MAP); systolic, diastolic, and mean pul-monary artery pressures; intracranial pressure (ICP); and cerebral perfusion pressure (CPP). Respiratory rate, frac-tion of inspired oxygen (FiO2), end-tidal CO2, and oxygen saturation were also recorded once per minute. In addi-tion, PbtO2 and brain temperature measurements were re-corded from each of the 4 probes. Mean values at each physiological intervention were calculated as described below. ABG measurements including pH, partial pressure of carbon dioxide in arterial blood (PaCO2), partial pres-sure of oxygen in arterial blood (PaO2), and base excess and bicarbonate were performed to verify anticipated physiological responses to changes in ventilation or FiO2.

physiological interventionsFollowing CCI and insertion of all PbtO2 monitors,

animals were exposed to a total of 6 physiological states. These included baseline, hyperventilation, and hypoven-tilation under conditions of normoxia (FiO2 = 0.21) and hyperoxia (FiO2 = 1.00). After baseline was established at normoxia, animals were exposed to the remainder of

the conditions (Fig. 2). A stratified randomization was performed whereby animals were first randomized to normoxic or hyperoxic conditions. In the second stage animals were randomized to the 3 respiratory rates, which they completed before the FiO2 was changed. After the FiO2 change, the 3 different respiratory rates were repeat-ed in random order. At baseline, the respiratory rate was 10/min, during hyperventilation it was 22/min, and during hypoventilation it was 5/min. Vital signs and brain PbtO2 measures had to reach a steady state for a minimum of a 5-minute period before physiological conditions were al-tered. At each of the 6 physiological conditions (normoxic normocarbia, normoxic hypocarbia, normoxic hypercar-bia, hyperoxic normocarbia, hyperoxic hypocarbia, hyper-oxic hypercarbia) parameters including MAP, ICP, CPP, heart rate, and brain tissue oxygen were recorded. Mean values for each physiological parameter were calculated for each physiological state as follows: the time point at which PbtO2 values at the distal or contralateral probes first reached 90% of the maximal change induced by each particular physiological intervention were identified. Five

Fig. 1. Schematic representation of the dorsal surface of the swine skull depicting the injury impact site and the locations of the ICP and brain tissue oxygen monitors, which were inserted to a depth of 12 mm using a stereotactic device. Anterior, posterior, right, and left are denoted. The right frontal site of injury resulting from CCI is indicated by the red circle. The impact was con-ducted at the site of a 15-mm bur hole placed 7 mm anterior to the coronal suture and 3 mm to the right side of the midline. Sites at which the 4 PbtO2 monitors were placed are indicated by blue circles. The distal probe was placed through a separate bur hole 3.5 mm posterior to the posterior edge of the bur hole at the right frontal impact site and 7 mm to the right of midline. The contralateral probe was placed 3.5 mm posterior to the posterior edge of the bur hole at the right frontal impact site and 7 mm to the left of midline. After generating the contusion, the contusion probe was inserted through the center of the bur hole used for the impact injury. The proximal probe was placed at the posterior margin of the impact injury bur hole. All 4 probes were placed anterior to the coronal suture. The intended locations for the placement of the contusion and proximal probes were in the lesion and near the lesion, respectively, and both probes were placed through the small craniectomy used for the experimental injury. The distal and contralateral probes were inserted through separate incisions and bur holes. A parenchymal ICP monitor (green circle) was placed in the left parietal lobe.




measures of MAP, ICP, CPP, heart rate, and PbtO2 (mean minute values) extending from 2 minutes before to 2 min-utes after this point were averaged to calculate the mean values at that physiological state. The same number of measurements of MAP, ICP, CPP, heart rate, and PbtO2 were made at each condition and for each experimental animal. In addition, in the first 3 experimental animals an ABG was drawn at baseline and with each physiological intervention to verify the anticipated changes in PaO2 and PaCO2.

postoperative careAt the completion of the experimentation animals were

killed while under anesthesia. All probes were removed and a return to previously measured atmospheric oxygen values was verified.

imagingDuring model optimization prior to the experimenta-

tion described here, a Philips Intera I/T 1.5-T whole-body MR imager (Philips Medical Systems) was used to charac-terize the tissue injury associated with the employed CCI model in a fashion previously described by our group.41 An anesthetized, ventilated pig subjected to the same in-jury parameters employed in the experimental animals underwent T2-weighted MR imaging 8 hours after CCI, generating the image shown in Fig. 3 right. While in the MRI suite, the swine received maintenance intravenous fluids and analgesics. Oxygen saturation and heart rate were continuously monitored using an MRI-compatible pulse oximeter (In Vivo Research, Inc.).

statistical analysisSPSS version 21.0.0.1 was employed for statistical anal-

ysis. GraphPad Prism version 6.03 was used to generate graphs. ANOVA was performed with a 3-way repeated-measures general linear model (GLM) to assess effects complying with the sphericity assumption. In cases in which the Mauchly test revealed a violation of spheric-ity, linear mixed model (LMM) regression was applied using restricted maximum likelihood modeling (REML). Significant main effects and interactions were followed up with Tukey and Bonferroni post hoc testing (for GLM pro-cedures) and with fixed-effect tests REML procedures. In all cases a p value < 0.05 was employed as the threshold for statistical significance. Error bars in graphs (Figs. 4 and 5) represent the standard error in all cases.

resultsWe studied 12 swine, all of which survived to comple-

Fig. 2. Schematic showing the sequence of physiological interventions carried out after animals were anesthetized (segments are represented by double-headed arrows). Distal and contralateral PbtO2 probes were placed at locations denoted in Fig. 1. CCI was then performed. Contusion and proximal probes were then placed. After a stable postinjury baseline at normoxia was achieved, animals were exposed to 5 additional physiological conditions in random order. A steady state in vital signs and brain PbtO2 mea-sures had to be achieved for a minimum of a 5-minute period before physiological conditions were altered. When PbtO2 measures at the distal or contralateral probes first reached 90% of the maximal change induced by each particular physiological intervention, values representative of each of the physiological states were calculated. This was done by averaging 5 measures from 2 minutes before to 2 minutes after the 90% change values. The segment delineated by the blue arrow involved normoxia (FiO2 = 0.21), and the red arrow denotes hyperoxia (FiO2 = 1.00).

Fig. 3. Representative images showing a porcine brain injured with the CCI device. left: Photograph showing the gross pathology of a typical swine brain following CCI. This lesion produces a consistent right frontal contusion with subarachnoid hemorrhage. right: A coronal T2-weight-ed MR image showing a section of swine brain 8 hours after injury with the CCI device. Contusion and subarachnoid hemorrhage are evident at the site of the right frontal lobe injury.




tion of the protocol. Typical gross pathology that demon-strates the extent of focal injury is shown in Fig. 3 left. Figure 3 right demonstrates a T2-weighted MR image of the swine brain with a focal contusion and surrounding edema.

On average, the time spent from anesthesia induction to CCI was 322.1 ± 59.2 minutes (Fig. 2, Segment a) and the time from CCI to normoxic baseline was 124.8 ± 83.5 minutes (Fig. 2, Segment b). For normoxic hyperventila-tion, it took a mean of 27.4 ± 21.9 minutes to reach 90%

Fig. 4. Box plots showing that ICP is associated with PaCO2 change. ICP at baseline as well as changes related to hypoventilation and hyperventilation are presented at normoxia (FiO2 = 0.21 [left]) and hyperoxia (FiO2 = 1.00 [right]). A 3-way within-subject GLM revealed a significant effect of ventilation on ICP (p < 0.0001) but no significant main effect of FiO2 or higher-order interactions. An LMM confirmed this result. Post hoc fixed-effects testing indicated hypoventilation increased ICP relative to baseline (p < 0.0001). Mean ICP values during hypoventilation were substantially greater than those measured at baseline and during hyperventilation (p < 0.001 on Bonferroni post hoc comparisons); however, baseline values were not significantly different from those measured during hyperventilation (p = 0.508 on Bonferroni post hoc comparison). The same results were observed at normoxia (FiO2 of 0.21 [left]) and hyperoxia (FiO2 of 1.0 [right]). In each of the 12 swine, mean ICP values were calculated at each of the 6 physiologi-cal states. Boxes indicate interquartile range, horizontal rules indicate the median, and whiskers demonstrate the maximum and minimum values.

Fig. 5. Graphs showing PbtO2 measurements with different PaCO2 values. PbtO2 values measured in 4 probes located in different positions relative to cerebral contusion are compared at baseline and during hypoventilation and hyperventilation. Hypoventila-tion and hyperventilation were associated with hypercarbia and hypocarbia, respectively. This physiological manipulation was performed under conditions of normoxia (FiO2 = 0.21 [left]) and hyperoxia (FiO2 = 1.00 [right]). LMM analysis revealed significant main effects of probe location, ventilation, and FiO2 on PbtO2 (all p < 0.001). In addition there was a probe-ventilation interaction, (p < 0.001). Post hoc fixed-effects testing by probe location revealed that the contusion and proximal probes were unresponsive to the ventilation condition (both p > 0.05), whereas the distal and contralateral probes were highly responsive to the ventilation condition and FiO2 (both p < 0.001). Fixed-effects testing of the main effects also revealed that the contusion probe registered PbtO2 values significantly lower than the proximal probe (p < 0.05). The distal and contralateral probes both had significantly higher PbtO2 than the proximal probe (p < 0.001) but did not differ from each other (p > 0.05). In each of the 12 swine, mean PbtO2 values were calculated at each of the 6 physiological states. Error bars represent standard error.




of the maximal change (Fig. 2, Segment c); animals re-mained in this condition for a mean of 10.1 ± 6.7 addi-tional minutes before changing conditions. For normoxic hypoventilation, it took a mean of 22.2 ± 9.4 minutes to reach 90% of maximal change (Fig. 2, Segment e); ani-mals remained in this condition for a mean of 17.2 ± 21.1 additional minutes before changing conditions. The mean time for physiological normalization between normoxic conditions was 28.9 ± 13.1 minutes (Fig. 2, Segment d). The mean time for physiological normalization between normoxic and hyperoxic conditions was 62.5 ± 35.4 min-utes (Fig. 2, Segment f). From normoxia to hyperoxic baseline, the mean time to reach 90% of maximal change was 21.3 ± 8.9 minutes (Fig. 2, Segment g), with a mean of 22.9 ± 27.5 additional minutes in the condition prior to change. For hyperoxic hyperventilation, it took a mean of 18.9 ± 7.6 minutes to reach 90% of maximal change (Fig. 2, Segment h), with 16.0 ± 12.2 additional minutes before change in conditions. It took a mean of 22.5 ± 7.3 minutes to reach 90% of maximal change for hyperoxic hypoventi-lation (Fig. 2, Segment j), and 11.1 ± 6.9 additional minutes were spent in the condition prior to change. A mean of 37.4 ± 15.6 minutes were spent to reach physiological normal-ization between hyperoxic conditions (Fig. 2, Segment i). Total mean time for completion of the experimental proto-col was 15.2 hours.

Mean values for heart rate, MAP, ICP, CPP, PaO2, and PaCO2 during each of the 6 physiological states are pre-sented in Table 1. These data demonstrate significant differences between physiological states with respect to MAP, ICP, PaO2, and PaCO2 values. Hyperoxia led to the expected marked increase in PaO2; hypoventilation was as-sociated with expected hypercarbia and higher ICP values. In contrast, hyperventilation was associated with expected hypocarbia and lower ICP values.

ICP values for each of the 6 physiological states are pre-sented in Fig. 4. A 3-way within-subject GLM revealed a significant effect of ventilation on ICP (p < 0.0001) but no significant main effect of FiO2 or higher-order interactions. The same effect was independently confirmed by LMM.

Post hoc fixed-effects testing indicated hypoventilation in-creased ICP relative to baseline (p < 0.0001). Mean ICP values during hypoventilation were substantially greater than those measured at baseline and during hyperventi-lation (p < 0.001 in Bonferroni post hoc comparisons); however, baseline values were not significantly different from those measured during hyperventilation (p = 0.508 in Bonferroni post hoc comparison). The same results were observed at an FiO2 of 0.21 (Fig. 4 left) and at an FiO2 of 1.00 (Fig. 4 right).

As demonstrated in Fig. 5, the probe location had an important influence on PbtO2 values and their responsive-ness to physiological interventions. At both normoxia and hyperoxia, PbtO2 decreased substantially with proximity to the focal injury (i.e., at contusion and proximal probes). LMM confirmed this, revealing significant main effects of probe location, ventilation, and FiO2 on PbtO2 (all p < 0.001). In addition, there was a probe-ventilation interac-tion (p < 0.001). Post hoc fixed-effects testing by probe location revealed that the contusion probe and proximal probes were unresponsive to ventilation condition (both p > 0.05), whereas the distal and contralateral probes were highly responsive to ventilation condition and FiO2 (both p < 0.001). Fixed-effects testing of the main effects also revealed that the contusion probe registered significantly lower PbtO2 values than did the proximal probe (p < 0.05). The distal and contralateral probes both measured signifi-cantly higher PbtO2 than the proximal probe (p < 0.001) but did not differ from each other (p > 0.05).

Figure 6 demonstrates the relationship between PaO2 and PbtO2 at different locations in relation to the injury site. The rise in PbtO2 with increasing arterial oxygen tension in the contralateral and distal probes is clearly evident (Fig. 6A and B). In the distal and contralateral probes, both PaO2 and PbtO2 values clearly and uniformly segregate into clus-ters obtained at arterial normoxia and hyperoxia, indicat-ing a strong relationship between arterial and brain tissue oxygen tension. Interestingly, while PaO2 values increased more than 5-fold as FiO2 was increased from 0.21 to 1.00, PbtO2 values increased only slightly more than 2-fold. In

table 1. systemic and cerebral physiological parameters at baseline and following physiological manipulations*

Parameter BaselineNormoxia (FiO2 = 0.21) Hyperoxia (FiO2 = 1.0) ANOVA

p Value†Hypoventilation Hyperventilation Hyperoxia Hypoventilation Hyperventilation

HR, beats/min 106.2 ± 27.2 131.2 ± 35.3 106.2 ± 19.9 104.8 ± 23.7 120.2 ± 34.8 103.9 ± 21.2 0.13MAP 92.0 ± 12.0 82.4 ± 18.3 71.9 ± 15.7

(p = 0.031)88.4 ± 11.7 92.6 ± 17.6 81.5 ± 10.6 0.01

ICP 11.5 ± 2.9 16.6 ± 3.8(p = 0.002)

8.3 ± 1.8 11.1 ± 2.5 17.2 ± 4.6(p < 0.0001)

9.2 ± 2.3 <0.01

CPP 78.3 ± 14.6 64.4 ± 19.6 63.1 ± 15.6 75.4 ± 13.1 73.1 ± 17.3 71.1 ± 11.8 0.12PaO2 125.7 ± 26.6 85.0 ± 39.6 151.5 ± 36.1 555.0 ± 36.4

(p < 0.0001)506.0 ± 51.6(p < 0.0001)

565.0 ± 53.8(p < 0.0001)

<0.01

PaCO2 40.0 ± 3.5 51.0 ± 1.4 34.5 ± 7.8 37.7 ± 2.5 60.0 ± 7.0(p = 0.004)

29.7 ± 1.5 <0.01

HR = heart rate.* Values are mean mm Hg ± SD unless otherwise indicated. Values typifying the physiological state in question were selected for analysis when PbtO2 measures at the distal or contralateral probes first reached 90% of the maximal change. Boldface indicates values that significantly differ from baseline values on Bonferroni post hoc testing.† ANOVA reflects statistical comparison between all 6 physiological states.




contrast, the lack of response of PbtO2 to a rise in PaO2 is clearly evident for the proximal and contusion probes (Fig. 6C and D), indicating that the relationship between arterial and brain tissue oxygen tension is lost at the site of injury (contusion probe), and nearly lost in its close vi-cinity (proximal probe). These data were analyzed with a GLM as well as an LMM, which demonstrated a signifi-cant 2-way interaction of probe location X PaO2 level (p < 0.001). This effect was observed whether PaO2 was dichot-omized into normoxia and hyperoxia or treated as a con-tinuous covariate (all p < 0.001). Post hoc 1-way ANOVA on dichotomized PaO2 and Bonferroni’s tests revealed that PaO2 influenced PbtO2 readouts at the distal and contralat-eral probes (p < 0.001), but not the proximal and contu-sion probes (both p > 0.05). Together the data reinforce the findings that the distal and contralateral probes respond to physiological changes whereas the contusion and proximal probes do not.

discussionCurrently, a major limitation of brain tissue oxygen

monitoring is a lack of clear data about how probe location

relative to a focal lesion affects measured values, and how these values respond differently to physiological events such as decreased MAP, impaired systemic oxygenation, or alterations in ventilation based on their distance from a focal injury. How the response of brain tissue oxygenation to treatment interventions may differ due to distance of the probe from injury site is also uncertain.

Several clinical reports have studied the effect of probe location on brain tissue oxygen values. Previous studies have documented different brain tissue oxygen values within white matter, gray matter, and the ventricular sys-tem.22 Dings et al. found that PbtO2 values decreased as probe depth increased.9 Longhi et al. found that PbtO2 val-ues were lower in pericontusional tissue than in apparently normal brain and that episodes of hypoxia were of longer duration in pericontusional tissue.20 Similarly, Ponce et al. found that PbtO2 values were lower near focal traumatic lesions than in apparently normal brain.35 PbtO2 measures in pericontusional tissue were predictive of outcome while those in normal brain were not. Hlatky et al. found that PbtO2 values change minimally in response to systemic hyperoxia in areas of low regional cerebral blood flow as

Fig. 6. Scatter plots showing PaO2 values obtained from ABG measurements along with concurrent PbtO2 measurements obtained from probes at 4 different locations relative to the focal contusion site. The data reflect at least 1 PaO2 blood gas measurement in each of the 6 physiological states (normoventilation, hyperventilation, and hypoventilation at both normoxia and hyperoxia) along with the corresponding PbtO2 measurements in 3 animals. Data were analyzed with a GLM, as well as an LMM, which demonstrat-ed a significant 2-way interaction of probe location and PaO2 level (p < 0.001). This effect was observed whether PaO2 was dichoto-mized into normoxia and hyperoxia or treated as a continuous covariate (all p < 0.001). Post hoc 1-way ANOVA on dichotomized PaO2 and Bonferroni tests revealed that PaO2 influenced PbtO2 readouts at the contralateral (a) and distal (b) probes (p < 0.001) but not the proximal (c) and contusion (d) probes (both p > 0.05).




determined by xenon CT.14 However, these data have not settled debate among clinicians regarding the optimal lo-cation for probe placement, because multiple probes were not studied concurrently in individual patients. To our knowledge, only 2 human studies have investigated con-current placement of 2 PbtO2 probes. Sarrafzadeh et al., using Licox and Paratrend monitors concurrently, found that PbtO2 values were lower near the contusion and rec-ommended that PbtO2 monitor placement be remote from focal injuries.43 Kiening et al. employed only Licox probes and reported similar findings in a sample of only 7 pa-tients.18

Several human studies have associated low PbtO2 val-ues with poor outcome following TBI.3,6,48,49,51 Maloney-Wilensky et al. performed a systematic review of such studies in 2009;23 3 met their criteria.3,17,49 In these stud-ies, an association between low PbtO2 and poor outcome was demonstrated, and it was concluded that PbtO2 values below 10 mm Hg may portend a worse prognosis. The same group published a similar systematic review of the literature in 2012, this time pooling available evidence on whether PbtO2-guided therapy improves outcome.31 They identified 7 relevant studies,1,25,26,28,32,46,47 and based on a pooled analysis of 4 studies26,28,32,46 that they felt could be combined, they concluded that PbtO2-guided therapy was associated with improved outcomes from severe TBI.

An ongoing multicenter randomized clinical trial is comparing treatment directed by ICP values alone ver-sus treatment directed by ICP and PbtO2 values (BOOST 2, NCT00974259). It is possible that it will be easier to interpret the results of therapeutic trials if the influence of probe location is better understood. Of interest, the PbtO2 monitors in the BOOST 2 study are being placed in the more normal frontal lobe.

Our results indicate that probe location in relation to a focal injury is an important determinant of brain tissue oxygen tension. Basic physiological parameters including arterial oxygen tension, hypocarbia, and hypercarbia influ-ence PbtO2 in a fashion that is dependent on the distance of the probe from the injury site. This finding has impor-tant clinical implications. The 2007 edition of the Brain Trauma Foundation’s Guidelines for the Management of Severe Traumatic Brain Injury recommends a PbtO2 treat-ment threshold of 15 mm Hg.5 This recommendation is made without reference to probe location, however.5 In-deed, the guidelines recognize this limitation of the pres-ent literature and recommend studying the site of probe location in relation to focal injury as a “key issue for fur-ther investigation.”5

Both in this experimental study and in the clinical set-ting, we prefer to place brain tissue oxygen probes within the white matter. Placement of probes within gray matter is technically challenging, making proper positioning dif-ficult both to achieve and to verify. In addition, cerebral blood flow, which strongly influences brain tissue oxygen tension, varies over a greater range in gray matter than in white matter and is also influenced by the functional acti-vation of specific brain regions,2,50,53 making interpretation of measured values more difficult. Placement of probes within the ventricular system also does not appear to be very clinically useful since it is most strongly influenced

by arterial partial pressure of oxygen and is less depen-dent on cerebral physiological parameters.22 Placement of brain tissue oxygen probes within the white matter has advantages including technical feasibility, relatively low variability in readings compared with other sites, and responsiveness to changes in both systemic oxygenation (PaO2) and cerebral physiological parameters such as ce-rebral blood flow.10,14,21,30,36,39,40,51,52 However, our study in-dicates that varying location within the white matter itself in relation to the site of injury strongly influences brain tissue oxygen tension measurements, making the decision of where to place the probe clinically important.

Not surprisingly, in this experimental model, the clos-er the probe was to the site of focal contusion, the lower the PbtO2 measurements and the less responsive measures were to actions that augment PbtO2 in other brain regions. Several pathophysiological phenomena may account for this observation. First, an anatomically disturbed vascular tree at the site of focal contusion may impair blood flow and, thus, oxygen delivery at the site of focal brain injury.37 Second, since PbtO2 measurements are strongly influenced by oxygen diffusion rather than total oxygen delivery,29 the edema around focal injury may impair oxygen diffusion into the injured region, thereby decreasing measured tis-sue oxygen tension. In this regard, it is important to note that the brain tissue oxygen monitor is not truly an “is-chemia monitor,” since it directly measures cerebral tis-sue oxygen tension and not the balance between oxygen delivery and metabolism. Even so, measuring brain tissue oxygen tension can still be very clinically useful since low values will reflect either poor systemic oxygenation or low cerebral blood flow39 and can assist in the early detection of both issues.13,34,40,42 Our study implies that local changes around a site of cerebral injury strongly impact brain tis-sue oxygen measurements and, therefore, need to be taken into account when interpreting PbtO2 values.

An important goal of management in the neurointen-sive care unit is to control intracranial pressure while maintaining adequate cerebral blood flow and oxygen-ation. Physiological manipulations such as adjusting frac-tion of inspired oxygenation and moderate hyperventila-tion to lower elevated ICP are routinely used in patients with severe TBI. Advanced cerebral monitors are often used to monitor the effect of these manipulations and one of the goals of monitoring brain tissue oxygen tension after TBI is to assess the effect of interventions on PbtO2. Our data indicate that brain tissue oxygenation is affected in a markedly different fashion depending on the distance of the probe from the site of a cerebral contusion. Brain tissue oxygen increases more than 2-fold during hyperoxia when the probe is distal to the injury site or in the contralateral hemisphere (Fig. 6A and B) but only negligibly when the probe is located at the site of contusion or near it (Fig. 6C and D). This has important clinical implications since ad-justing FiO2 to optimize cerebral tissue oxygenation may be less efficacious in the areas around cerebral contusions. Similarly, changes in ventilation that alter cerebral blood flow influence PbtO2 in a manner that is strongly dependent on the location of the probe in relation to injury (Fig. 5). Hypoventilation increases PbtO2 at sites distal to the contu-sion and in the contralateral hemisphere but not at the site




of injury or near it. Conversely, hyperventilation decreases PbtO2 distal to a contusion and in the contralateral hemi-sphere but not within or near a contusion. These results imply that cerebral CO2 vasoreactivity is impaired in and around contused cerebral tissue, hindering or making less effective the alterations in ventilation that influence perfu-sion or lower ICP in and around the site of the focal contu-sion.

Our results are in general agreement with those of McLaughlin and Marion who demonstrated, using xenon CT, low cerebral blood flow and variable cerebral vaso-reactivity in pericontusional tissue in patients with severe TBI.27 Our findings are also similar to those previously described in work by Hlatky et al.14 In general, our results indicate that PbtO2 measurements will be less sensitive to physiological alterations when placed in or near a contu-sion, limiting the ability of the brain tissue oxygen monitor to guide therapeutic interventions. It is interesting that in some instances hyperoxia did not increase PbtO2, even in the distal and contralateral probes (Fig. 6). This may relate to focal brain injury induced during placement of some of the probes and is consistent with variability noted in humans.14

Our findings are relevant to the important question of where to place the probe in patients with severe TBI. How-ever, the answer to this simple question is not necessarily clear-cut. The advantages and disadvantages of different placement strategies need to be carefully weighed. One school of thought advocates placement of PbtO2 probes in pericontusional tissue to help save cerebral “tissue at risk.”35 In the aforementioned paper by Ponce and col-leagues, a multivariate analysis found that probe location was an important factor when looking at the relationship between mean PbtO2 and patient outcome in severe TBI.35 When the probe was placed in “vulnerable” or abnormal brain tissue defined on postinsertion CT as being within a contusion, near a contusion, or under a surgically evacuat-ed hematoma, mean PbtO2 values were substantially higher in patients with a favorable outcome (28.8 ± 12.0 mm Hg) compared with those with a poor outcome (19.5 ± 13.7 mm Hg, p = 0.01). Conversely, when the probe was placed in normal-appearing brain, PbtO2 was not related to outcome in the multivariate model. On the one hand, our data could be viewed as supporting placement of brain tissue oxygen monitors in pericontusional tissue since values will be lowest in and around focal injuries. Therapy would then be geared to maximizing values in this “tissue at risk.” However, the PbtO2 threshold to maintain tissue viability is not known, either clinically or in controlled experimental models. In this regard, it is important to remember that the brain tissue oxygen monitor does not convey information about oxygen utilization or metabolism.38,41 In addition, it is well known that mitochondrial dysfunction can lead to cellular energy failure, even when sufficient oxygen is pro-vided to tissue.7,45 Nevertheless, an argument can be made that maintaining PbtO2 as high as possible around focal injury may be advantageous. In our view, the data more strongly support the placement of brain tissue oxygen probes distal to the site of focal injury or in the contralat-eral hemisphere since cerebral tissue in these regions be-haves in a more “physiological” way and therefore can be

used to assess the effect of interventions such as changing FiO2, changing ventilation, or altering CPP. Such use of the brain tissue oxygen monitor allows the clinician to “fine tune” systemic and cerebral physiological parameters to optimize cerebral tissue oxygenation.

It is important to consider that an additional probe site between our proximal and distal probes might better char-acterize lesion penumbra and exhibit greater responsive-ness. Such a probe site could have optimal intermediate properties. We strongly discourage readers from abandon-ing efforts to characterize or study penumbral PbtO2 on the basis of these results, especially because our proximal probes are likely located nearer the focal lesion than peri-contusional probes investigated in other studies.

In clinical practice, considerations other than purely physiological ones often affect the decision on where to place monitors. If a brain tissue oxygen monitor is placed via a bolt, it can only be inserted where the cranium is in-tact. Therefore, patients undergoing decompressive hemi-craniectomy must have bolts placed on the side contralat-eral to decompression, which is usually the least-injured hemisphere. In addition, bolt placement is generally lim-ited to the anterior frontal lobes to achieve safe localiza-tion within the large white matter tracts without injury to descending motor fibers. Given these practical consider-ations, placement of probes in truly pericontusional cere-bral tissue is often difficult without imaging-based stereo-tactic guidance. Placement in the contralateral frontal lobe may thus be the best option until more precise placement of brain tissue oxygen probes becomes possible in human patients.15 In interpreting the results of our study it is im-portant to consider that the precise location of proximal probes in relation to injury was not determined by ana-tomical imaging studies. Previous work by our group in a swine model of CCI used MRI loaded to a Stealth neuro-navigation station to precisely determine the site of proxi-mal probes in relation to the injury site.41 In that study, we found that proximal probes were actually located within the contusion 20% of the time, indicating the difficulty of precise probe placement even under controlled experi-mental conditions. This difficulty is likely to apply to the clinical setting as well, making the placement of probes in precise pericontusional regions that reflect “tissue at risk” or penumbral tissue challenging. The values measured by the proximal probe likely reflect some bias introduced by proximal probe location within contusional tissue, and for this reason, the proximal probe should not be viewed as providing information about penumbral tissue and should not discourage future efforts to study probes located in a formally defined penumbra. Despite this limitation, we found a subtle but statistically significant difference in the physiological responses between proximal and contusion probes. Our data would likely reflect the clinical reality of imprecise placement in any attempt to place probes near a site of focal injury.

limitationsOur study has several limitations. First, it analyzed the

characteristics of PbtO2 measurements in relation to the dis-tance of a brain tissue oxygen monitor from a focal contu-sion, but it did not aim to define a treatment threshold that




would preserve tissue viability. Neither does it provide in-formation on histological or functional outcomes that could be used for this purpose. Defining the threshold and dura-tion of low brain tissue oxygen values that lead to irrevers-ible cellular damage is an important goal for future studies. However, it does demonstrate that probe location should be a key consideration when such studies are performed and whenever probe readings are interpreted. A second limi-tation of our model is that the experiment was performed in the period immediately following injury, extending to a mean of 15.2 hours postinjury. Since many physiological alterations, including increased brain edema, alterations in cerebral blood flow, and autoregulation, can occur days after injury, caution must be used in extrapolating our re-sults to a period of days after injury. Our study design may not exactly duplicate the clinical situation during the pe-riod of maximal brain swelling. Both ethical concerns re-garding the humane treatment of the experimental animals and practical issues of resource availability prevented us from extending the experiment to several days postinjury to more closely parallel the typical clinical scenario. Lastly, we did not verify probe location in this study by imaging studies. We, therefore, cannot be certain whether the proxi-mal probe sensors were located in penumbra. A previous study by our group did image the swine brain with MRI following CCI and probe placement, demonstrating the ca-pabilities and limitations of probe placement in the swine CCI model.41 Our goal in the present study was to assess the impact of distance from a focal injury and not presence within tissue formally defined as penumbra. Because of this, our findings related to the proximal probe should not discourage needed research into PbtO2 monitoring in the injury penumbra nor should it prompt cessation of clinical efforts to monitor the penumbra. We did not endeavor to define the location of the proximal probe employed in our study relative to penumbra, as penumbra is not routinely characterized in the course of TBI patient care, and brain oxygen probes cannot currently be placed into such a re-gion with stereotactic precision.

conclusionsBrain tissue oxygen measurements are strongly influ-

enced by the distance from focal brain injury. Physiologi-cal alterations including hyperoxia, hyperventilation, and hypoventilation more readily affect PbtO2 values distal to the site of injury but have little effect on measures near the contusion. Clinical interpretations of brain tissue oxygen measurements should take into account the spatial relation of probe position to sites of focal brain injury. Likewise, decisions regarding optimal placement of brain tissue oxy-gen probes in patients with severe TBI should take these factors into consideration.

acknowledgmentsWe would like to thank Larry Carbone, DVM, PhD, and Ron

Bairuther for their assistance with the large-animal studies. Kotaro Oshio MD, PhD, and Ken Monson, PhD, also made important contributions to the technical aspects of this study. This work was supported by CDC grant R49, CCR903697, and ONR grant N00014002-1-0203.

references 1. Adamides AA, Cooper DJ, Rosenfeldt FL, Bailey MJ,

Pratt N, Tippett N, et al: Focal cerebral oxygenation and neurological outcome with or without brain tissue oxygen-guided therapy in patients with traumatic brain injury. Acta Neurochir (Wien) 151:1399–1409, 2009

2. Andrade KC, Pontes-Neto OM, Leite JP, Santos AC, Baffa O, de Araujo DB: Quantitative aspects of brain perfusion dynamic induced by BOLD fMRI. Arq Neuropsiquiatr 64:895–898, 2006

3. Bardt TF, Unterberg AW, Härtl R, Kiening KL, Schneider GH, Lanksch WR: Monitoring of brain tissue PO2 in traumatic brain injury: effect of cerebral hypoxia on outcome. Acta Neurochir Suppl 71:153–156, 1998

4. Bouma GJ, Muizelaar JP: Relationship between cardiac output and cerebral blood flow in patients with intact and with impaired autoregulation. J Neurosurg 73:368–374, 1990

5. Bratton SL, Chestnut RM, Ghajar J, McConnell Hammond FF, Harris OA, Hartl R, et al: Guidelines for the management of severe traumatic brain injury. X. Brain oxygen monitoring and thresholds. J Neurotrauma 24 (Suppl 1):S65–S70, 2007 (Erratum in J Neurotrauma 25:276–278, 2008)

6. Chang JJ, Youn TS, Benson D, Mattick H, Andrade N, Harper CR, et al: Physiologic and functional outcome correlates of brain tissue hypoxia in traumatic brain injury. Crit Care Med 37:283–290, 2009

7. Cheng G, Kong RH, Zhang LM, Zhang JN: Mitochondria in traumatic brain injury and mitochondrial-targeted multipotential therapeutic strategies. Br J Pharmacol 167:699–719, 2012

8. Cole SL, Corday E: Four-minute limit for cardiac resuscitation. J Am Med Assoc 161:1454–1458, 1956

9. Dings J, Meixensberger J, Jäger A, Roosen K: Clinical experience with 118 brain tissue oxygen partial pressure catheter probes. Neurosurgery 43:1082–1095, 1998

10. Fandino J, Stocker R, Prokop S, Imhof HG: Correlation between jugular bulb oxygen saturation and partial pressure of brain tissue oxygen during CO2 and O2 reactivity tests in severely head-injured patients. Acta Neurochir (Wien) 141:825–834, 1999

11. Haitsma I, Rosenthal G, Morabito D, Rollins M, Maas AI, Manley GT: In vitro comparison of two generations of Licox and Neurotrend catheters. Acta Neurochir Suppl 102:197–202, 2008

12. Hansen AJ: Effect of anoxia on ion distribution in the brain. Physiol Rev 65:101–148, 1985

13. Hemphill JC III, Knudson MM, Derugin N, Morabito D, Manley GT: Carbon dioxide reactivity and pressure autoregulation of brain tissue oxygen. Neurosurgery 48:377–384, 2001

14. Hlatky R, Valadka AB, Gopinath SP, Robertson CS: Brain tissue oxygen tension response to induced hyperoxia reduced in hypoperfused brain. J Neurosurg 108:53–58, 2008

15. Holzschuh M, Woertgen C, Metz C, Brawanski A: Clinical evaluation of the InnerSpace fibreoptic intracranial pressure monitoring device. Brain Inj 12:191–198, 1998

16. Hutchinson PJ, Menon DK, Czosnyka M, Kirkpatrick PJ: Monitoring cerebral blood flow and metabolism, in Reilly P, Bullock MR (eds): Head Injury: Pathophysiology and Management. Boca Raton, FL: CRC Press, 2005, pp 216–245

17. Kiening KL, Härtl R, Unterberg AW, Schneider GH, Bardt T, Lanksch WR: Brain tissue pO2-monitoring in comatose patients: implications for therapy. Neurol Res 19:233–240, 1997

18. Kiening KL, Schneider GH, Bardt TF, Unterberg AW, Lanksch WR: Bifrontal measurements of brain tissue-PO2 in comatose patients. Acta Neurochir Suppl 71:172–173, 1998




19. Kirkpatrick PJ, Smielewski P, Piechnik S, Pickard JD, Czosnyka M: Early effects of mannitol in patients with head injuries assessed using bedside multimodality monitoring. Neurosurgery 39:714–721, 1996

20. Longhi L, Pagan F, Valeriani V, Magnoni S, Zanier ER, Conte V, et al: Monitoring brain tissue oxygen tension in brain-injured patients reveals hypoxic episodes in normal-appearing and in peri-focal tissue. Intensive Care Med 33:2136–2142, 2007

21. Longhi L, Valeriani V, Rossi S, De Marchi M, Egidi M, Stocchetti N: Effects of hyperoxia on brain tissue oxygen tension in cerebral focal lesions. Acta Neurochir Suppl 81:315–317, 2002

22. Maas AI, Fleckenstein W, de Jong DA, van Santbrink H: Monitoring cerebral oxygenation: experimental studies and preliminary clinical results of continuous monitoring of cerebrospinal fluid and brain tissue oxygen tension. Acta Neurochir Suppl (Wien) 59:50–57, 1993

23. Maloney-Wilensky E, Gracias V, Itkin A, Hoffman K, Bloom S, Yang W, et al: Brain tissue oxygen and outcome after severe traumatic brain injury: a systematic review. Crit Care Med 37:2057–2063, 2009

24. Manley GT, Rosenthal G, Lam M, Morabito D, Yan D, Derugin N, et al: Controlled cortical impact in swine: pathophysiology and biomechanics. J Neurotrauma 23:128–139, 2006

25. Martini RP, Deem S, Yanez ND, Chesnut RM, Weiss NS, Daniel S, et al: Management guided by brain tissue oxygen monitoring and outcome following severe traumatic brain injury. J Neurosurg 111:644–649, 2009

26. McCarthy MC, Moncrief H, Sands JM, Markert RJ, Hall LC, Wenker IC, et al: Neurologic outcomes with cerebral oxygen monitoring in traumatic brain injury. Surgery 146:585–591, 2009

27. McLaughlin MR, Marion DW: Cerebral blood flow and vasoresponsivity within and around cerebral contusions. J Neurosurg 85:871–876, 1996

28. Meixensberger J, Jaeger M, Väth A, Dings J, Kunze E, Roosen K: Brain tissue oxygen guided treatment supplementing ICP/CPP therapy after traumatic brain injury. J Neurol Neurosurg Psychiatry 74:760–764, 2003

29. Menon DK, Coles JP, Gupta AK, Fryer TD, Smielewski P, Chatfield DA, et al: Diffusion limited oxygen delivery following head injury. Crit Care Med 32:1384–1390, 2004

30. Menzel M, Doppenberg EM, Zauner A, Soukup J, Reinert MM, Bullock R: Increased inspired oxygen concentration as a factor in improved brain tissue oxygenation and tissue lactate levels after severe human head injury. J Neurosurg 91:1–10, 1999

31. Nangunoori R, Maloney-Wilensky E, Stiefel M, Park S, Andrew Kofke W, Levine JM, et al: Brain tissue oxygen-based therapy and outcome after severe traumatic brain injury: a systematic literature review. Neurocrit Care 17:131–138, 2012

32. Narotam PK, Burjonrappa SC, Raynor SC, Rao M, Taylon C: Cerebral oxygenation in major pediatric trauma: its relevance to trauma severity and outcome. J Pediatr Surg 41:505–513, 2006

33. Oddo M, Bösel J: Monitoring of brain and systemic oxygenation in neurocritical care patients. Neurocrit Care 21 (Suppl 2):S103–S120, 2014

34. Oddo M, Nduom E, Frangos S, MacKenzie L, Chen I, Maloney-Wilensky E, et al: Acute lung injury is an independent risk factor for brain hypoxia after severe traumatic brain injury. Neurosurgery 67:338–344, 2010

35. Ponce LL, Pillai S, Cruz J, Li X, Julia H, Gopinath S, et al: Position of probe determines prognostic information of brain tissue PO2 in severe traumatic brain injury. Neurosurgery 70:1492–1503, 2012

36. Reinert M, Barth A, Rothen HU, Schaller B, Takala J, Seiler RW: Effects of cerebral perfusion pressure and increased fraction of inspired oxygen on brain tissue oxygen, lactate and glucose in patients with severe head injury. Acta Neurochir (Wien) 145:341–350, 2003

37. Rodríguez-Baeza A, Reina-de la Torre F, Poca A, Martí M, Garnacho A: Morphological features in human cortical brain microvessels after head injury: a three-dimensional and immunocytochemical study. Anat Rec A Discov Mol Cell Evol Biol 273:583–593, 2003

38. Rohlwink UK, Figaji AA: Methods of monitoring brain oxygenation. Childs Nerv Syst 26:453–464, 2010

39. Rosenthal G, Hemphill JC III, Sorani M, Martin C, Morabito D, Obrist WD, et al: Brain tissue oxygen tension is more indicative of oxygen diffusion than oxygen delivery and metabolism in patients with traumatic brain injury. Crit Care Med 36:1917–1924, 2008

40. Rosenthal G, Hemphill JC, Sorani M, Martin C, Morabito D, Meeker M, et al: The role of lung function in brain tissue oxygenation following traumatic brain injury. J Neurosurg 108:59–65, 2008

41. Rosenthal G, Morabito D, Cohen M, Roeytenberg A, Derugin N, Panter SS, et al: Use of hemoglobin-based oxygen-carrying solution-201 to improve resuscitation parameters and prevent secondary brain injury in a swine model of traumatic brain injury and hemorrhage: laboratory investigation. J Neurosurg 108:575–587, 2008

42. Rossi S, Longhi L, Balestreri M, Spagnoli D, deLeo A, Stocchetti N: Brain oxygen tension during hyperoxia in a swine model of cerebral ischaemia. Acta Neurochir Suppl 76:243–245, 2000

43. Sarrafzadeh AS, Kiening KL, Bardt TF, Schneider GH, Unterberg AW, Lanksch WR: Cerebral oxygenation in contusioned vs. nonlesioned brain tissue: monitoring of PtiO2 with Licox and Paratrend. Acta Neurochir Suppl 71:186–189, 1998

44. Schneider GH, von Helden A, Lanksch WR, Unterberg A: Continuous monitoring of jugular bulb oxygen saturation in comatose patients—therapeutic implications. Acta Neurochir (Wien) 134:71–75, 1995

45. Soustiel JF, Zaaroor M: Mitochondrial targeting for development of novel drug strategies in brain injury. Cent Nerv Syst Agents Med Chem 12:131–145, 2012

46. Spiotta AM, Stiefel MF, Gracias VH, Garuffe AM, Kofke WA, Maloney-Wilensky E, et al: Brain tissue oxygen-directed management and outcome in patients with severe traumatic brain injury. J Neurosurg 113:571–580, 2010

47. Stiefel MF, Spiotta A, Gracias VH, Garuffe AM, Guillamondegui O, Maloney-Wilensky E, et al: Reduced mortality rate in patients with severe traumatic brain injury treated with brain tissue oxygen monitoring. J Neurosurg 103:805–811, 2005

48. Valadka AB, Gopinath SP, Contant CF, Uzura M, Robertson CS: Relationship of brain tissue PO2 to outcome after severe head injury. Crit Care Med 26:1576–1581, 1998

49. van den Brink WA, van Santbrink H, Steyerberg EW, Avezaat CJ, Suazo JA, Hogesteeger C, et al: Brain oxygen tension in severe head injury. Neurosurgery 46:868–878, 2000

50. van der Zande FH, Hofman PA, Backes WH: Mapping hypercapnia-induced cerebrovascular reactivity using BOLD MRI. Neuroradiology 47:114–120, 2005

51. van Santbrink H, Maas AI, Avezaat CJ: Continuous monitoring of partial pressure of brain tissue oxygen in patients with severe head injury. Neurosurgery 38:21–31, 1996

52. van Santbrink H, van den Brink WA, Steyerberg EW, Carmona Suazo JA, Avezaat CJ, Maas AI: Brain tissue oxygen response in severe traumatic brain injury. Acta Neurochir (Wien) 145:429–438, 2003




53. Wise RG, Ide K, Poulin MJ, Tracey I: Resting fluctuations in arterial carbon dioxide induce significant low frequency variations in BOLD signal. Neuroimage 21:1652–1664, 2004

disclosuresThe authors report no conflict of interest concerning the materi-als or methods used in this study or the findings specified in this paper.

author contributionsConception and design: Manley, Morabito, Derugin, Knudson. Acquisition of data: Manley, Phan, Morabito, Derugin, Stewart,

Rosenthal. Analysis and interpretation of data: Manley, Hawry-luk, Ferguson, Rosenthal. Drafting the article: Manley, Hawryluk, Rosenthal. Critically revising the article: Manley, Hawryluk, Fer-guson, Rosenthal. Reviewed submitted version of manuscript: all authors. Statistical analysis: Hawryluk, Ferguson. Administrative/technical/material support: Manley, Hawryluk, Phan, Ferguson, Morabito, Derugin, Stewart, Knudson. Study supervision: Man-ley, Rosenthal.

correspondenceGeoffrey Manley, Department of Neurological Surgery, Universi-ty of California San Francisco, 1001 Potrero Ave., Rm. 101, San Francisco, CA 94110. email: [email protected].


Documents

Brain tissue oxygen tension and its response to