Reduced Brain Tissue Oxygen in Traumatic Brain Injury: Are ......Key Words: Brain tissue oxygenation, Traumatic brain injury, Treatment interventions, PbtO 2-guided management, Clinical

ORIGINAL ARTICLE

Reduced Brain Tissue Oxygen in Traumatic Brain Injury: Are MostCommonly Used Interventions Successful?

Jose L. Pascual, MD, PhD, FRCPS(C), Patrick Georgoff, BS, Eileen Maloney-Wilensky, CRNP,Carrie Sims, MD, MS, FACS, Babak Sarani, MD, FACS, Michael F. Stiefel, MD, PhD,

Peter D. LeRoux, MD, FACS, and C. William Schwab, MD, FACS

Background: Brain tissue oxygenation (PbtO2)-guided management facili-tates treatment of reduced PbtO2 episodes potentially conferring survival andoutcome advantages in severe traumatic brain injury (TBI). To date, thenature and effectiveness of commonly used interventions in correctingcompromised PbtO2 in TBI remains unclear. We sought to identify the mostcommon interventions used in episodes of compromised PbtO2 and toanalyze which were effective.Methods: A retrospective 7-year review of consecutive severe TBI patientswith a PbtO2 monitor was conducted in a Level I trauma center’s intensivecare unit or neurosurgical registry. Episodes of compromised PbtO2 (definedas �20 mm Hg for 0.25–4 hours) were identified, and clinical interventionsconducted during these episodes were analyzed. Response to treatment wasgauged on how rapidly (�T) PbtO2 normalized (�20 mm Hg) and how greatthe PbtO2 increase was (�PbtO2). Intracranial pressure (�ICP) and cerebralperfusion pressure (�CPP) also were examined for these episodes.Results: Six hundred twenty-five episodes of reduced PbtO2 were identifiedin 92 patients. Patient characteristics were: age 41.2 years, 77.2% men, andInjury Severity Score and head or neck Abbreviated Injury Scale score of34.0 � 9.2 and 4.9 � 0.4, respectively. Five interventions: narcotics orsedation, pressors, repositioning, FIO2/PEEP increases, and combined seda-tion or narcotics � pressors were the most commonly used strategies.Increasing the number of interventions resulted in worsening the time toPbtO2 correction. Triple combinations resulted in the lowest �ICP and dualcombinations in the highest �CPP (p � 0.05).Conclusion: Clinicians use a limited number of interventions when correct-ing compromised PbtO2. Using strategies employing many interventionsadministered closely together may be less effective in correcting PbO2, ICP,and CPP deficits. Some PbtO2 deficits may be self-limited.Key Words: Brain tissue oxygenation, Traumatic brain injury, Treatmentinterventions, PbtO2-guided management, Clinical practice guidelines.

(J Trauma. 2011;70: 535–546)

Traumatic brain injury (TBI) remains a major cause ofmortality and morbidity in young people worldwide and

has a significant long-term socioeconomic impact. In partic-ular, severe TBI (Glasgow Comas Scale [GCS] �8) is asso-ciated with 30% mortality and significant disability amongsurvivors.1 To date, there is no effective drug treatment forTBI. Instead, management is centered on identifying andmanaging the secondary brain injury that evolves in the hoursand days after TBI. Secondary injury is known to occur withcerebral underperfusion but also may occur with dysfunc-tional cerebral metabolism, tissue hypoxia, and inflammationand contributes to further tissue destruction.2–4

Although there is no Level I evidence to suggest thatmanagement of intracranial pressure (ICP) is associated withbetter outcome, the use of an ICP monitor is endorsed by majormedical societies (The Brain Trauma Foundation [BTF], TheEuropean Brain Injury Consortium, The American Associationof Neurologic Surgeons, and The Congress of Neurologic Sur-geons Joint Section on Neurotrauma and Critical Care).5,6

Several lines of evidence suggest that reduced brainoxygen is not a benign event and that compromised brainoxygen (�20 mm Hg) or brain hypoxia (variably defined as�15 or 10 mm Hg) is associated with increased mortality andunfavorable outcome.7–10 Consistent with this, nonrandomizedclinical studies indicate that therapy based on both an ICP andbrain oxygenation (PbtO2) monitor is associated with betteroutcome than management with only an ICP monitor.11–16

There are, however, several unanswered questions aboutPbtO2-based care including what are available treatments toimprove brain oxygenation, how effective are they, and how dothey affect traditional management parameters such as ICP andcerebral perfusion pressure (CPP)? This retrospective study wasconducted to (1) identify the most common interventions ad-ministered by neurosurgeons and intensivists during short (�4hours) episodes of low PbtO2 and (2) whether these interven-tions (alone or in combination) resulted in significant benefits inrapidly correcting PbtO2 deficits and improving ICP and CPP.

METHODS

Patient PopulationAll blunt TBI patients admitted to the intensive care unit

(ICU) of an academic Level I trauma center who had a paren-chymal intracranial monitor able to measure partial brain tissueoxygen tension (PbtO2) between October 2001 and September

Submitted for publication September 20, 2010.Accepted for publication December 13, 2010.Copyright © 2011 by Lippincott Williams & WilkinsFrom the Division of Traumatology, Surgical Critical Care & Emergency Surgery

(J.L.P., P.G., C.S., B.S., C.W.S.); and Department of Neurosurgery (E.M.-W.,M.F.S., P.D.L.), University of Pennsylvania School of Medicine, Philadel-phia, Pennsylvania.

Supported, in part, by research grants from the Integra Foundation, IntegraNeurosciences, and the Mary Elisabeth Groff Surgical and Medical ResearchTrust (to P.D.L.). P.D.L. is a member of the Integra Speaker’s Bureau.

Presented at the 69th Annual Meeting of the American Association for the Surgeryof Trauma, September 22–25, 2010, Boston, Massachusetts.

Address for reprints: Jose L. Pascual, MD, PhD, FRCPS(C), Division of Traumatol-ogy, Surgical Critical Care and Emergency Surgery, Department of Surgery, 3400Spruce Street, Philadelphia, PA 19104; email: [email protected].

DOI: 10.1097/TA.0b013e31820b59de

The Journal of TRAUMA® Injury, Infection, and Critical Care • Volume 70, Number 3, March 2011 535

2008 were considered for study. Patients were retrospectivelyidentified from a prospective observational database, the BrainOxygen Monitoring Outcomes (BOMO) registry, which hasInstitutional Review Board approval. Patients with gunshotwounds or other penetrating cranial injuries, ongoing bloodloss, or whose postresuscitation systolic blood pressure was�90 mm Hg and arterial oxygen saturation (SaO2) �93%were excluded from analysis. Patients in whom pupils werebilaterally fixed and dilated or were brain dead or imminentlydead on admission also were excluded.

Brain Intraparenchymal MonitorsICP, brain temperature, and brain tissue oxygen

(PbtO2) were continuously monitored using a commerciallyavailable intracranial device (LICOX; Integra LifeSciences,Plainsboro, NJ). All three intracranial monitors were insertedat the bedside through the same burr-hole into the frontal lobeand secured with a triple-lumen bolt. The PbtO2 monitor wasplaced into white matter that appeared normal on the admis-sion head CT and on the side of maximal pathology. Braintissue oxygenation data were acquired after a stabilizationperiod of 2 hours after probe insertion. Probe function andlocation were confirmed by an appropriate increase in PbtO2after an hyperoxic FIO2 challenge (FIO2 � 100%)17 and ahead CT scan to verify correct placement of the variousmonitors, e.g., not in a contusion or infarct. CPP was calcu-lated from measured parameters (CPP � mean arterial pres-sure [MAP] � ICP). All intracranial monitors were removedwhen ICP was normal for 24 hours without specific treatmentother than sedation for ventilation, when the patient was ableto follow commands or when the patient was brain dead.

Physiologic MonitorsHeart rate, arterial line blood pressures, and arterial oxy-

gen saturations (SaO2) were monitored continuously (Compo-nent Monitoring System M1046-9090C: Hewlett Packard,Andover, MA). The MAP was derived from arterial lines withtransducers leveled at the phlebostatic axis. Central venouspressures and pulmonary artery pressures were followed in patientswith intravascular depletion or cardiopulmonary compromise.

General Clinical Management of TBITrauma surgeons resuscitated all patients according to

Advanced Trauma Life Support (ATLS) protocols (AmericanCollege of Surgeons Committee on Trauma: AdvancedTrauma Life Support Course for Doctors. Chicago: AmericanCollege of Surgeons, 1997). Patients then were managedaccording to a local algorithm based on the BTF Guidelinesfor Severe TBI5 in the Neurosurgical Intensive Care Unitor the Surgical and Trauma Intensive Care Unit. Thisincluded (1) early identification and evacuation of traumaticspace-occupying intracranial hematomas, (2) intubation andventilation with low-volume pressure-limited ventilation tomaintain PaCO2 between 30 and 40 mm Hg and SaO2 �93%, (3)sedation using propofol during the first 24 hours followed bysedation and analgesia using lorazepam, morphine, or fenta-nyl, (4) bedrest with head elevation of �30 degrees, (5)normothermia �35°C to 37°C, (6) euvolemia using a base-line crystalloid infusion (0.9% normal saline, 20 mEq/L KCl;

80–100 mL/h), (7) anticonvulsant prophylaxis with phenyt-oin for 1 week or longer if seizures occurred, and (8) packedred blood cell transfusion if their Hgb was �7.

Management of Intracranial HypertensionIf ICP remained persistently elevated (�20 mm Hg

�10 min) despite baseline initial measures, osmotherapy wasadministered using repeated boluses of mannitol (1 gm/kg,25% solution) provided that serum osmolar gap �20. Secondtier therapies for refractory intracranial hypertension (�20mm Hg �15 minutes in a 1-hour period despite therapy)included optimized hyperventilation (PaCO2 30–35 mm Hg),decompressive craniectomy, or pharmacological coma (withpropofol or pentobarbital). Induced hypothermia and hyper-tonic saline for ICP control were used to manage ICP in thepatients included during the last 2.5 years of this study.

Evaluation of Brain Oxygen TreatmentDuring the study period patients received “cause” directed

therapy at the intensivist’s discretion to maintain PbtO2 �20mm Hg according to our local protocol (Fig. 1). The BOMOregistry records and codes every event noted by a bedside nursein the chart and also records hemodynamics, cerebral parame-ters, and other nursing entries every 10 minutes to 20 minutes.There was �8000 hours of PbtO2 monitoring in eligible patientsavailable for review. Episodes where PbtO2 was �20 mm Hgfor �15 minutes but �4 hours were abstracted. There have beenseveral important described thresholds for PbtO2 that identifywhen cell death or ischemia may be evident and at what level totreat. We chose a PbtO2 threshold of 20 mm Hg because thiscorresponds to the minimal necessary oxygen tension for mito-chondria, where the majority of cellular oxygen metabolismoccurs to maintain aerobic metabolism.18 In addition, it is thethreshold being used in a current NIH-funded trial to prospec-tively examine PbtO2 in TBI. We chose a 15-minute minimumtime window to eliminate incidental, self-limited episodes ofcompromised PbtO2 and to allow time for a 2-minute oxygenchallenge as required by protocol to test monitor function. Thus,this brief FIO2 challenge was not considered therapeutic and sowas excluded from analysis. We chose a 4-hour maximum toavoid inclusion of patients who no longer were receiving activetreatment for PbtO2 deficits by bedside clinicians. Also this wasdone to avoid the inclusion of episodes where clinicians wereaggressively using all possible interventions in the setting of aresistant compromised PbtO2.

We consulted seasoned intensivists, neurologists, traumasurgeons, and neurosurgeons who cared for brain injured pa-tients to achieve a consensus of what therapies were considereduseful to correct PbtO2 deficits. Eleven interventions were se-lected by the panel and thereafter identified from the codedtreatments in the BOMO registry. They are presented in Table 1.

During each episode of PbtO2 compromise, various reg-istry parameters were recorded before the decrease and oncorrection of the PbtO2 deficit (�20 mm Hg). Collected param-eters included patient hemodynamics (systolic blood pressure,MAP, and heart rate [HR]), ventilation parameters (positive endexpiratory pressure [PEEP], PaO2 or PaCO2, SaO2, and FIO2), andcerebral parameters (ICP, CPP, and PbtO2). All coded interven-tions thought to potentially affect PbtO2 (Table 1) that were

Pascual et al. The Journal of TRAUMA® Injury, Infection, and Critical Care • Volume 70, Number 3, March 2011

© 2011 Lippincott Williams & Wilkins536

administered during this time period were collected. Basic de-mographics (age, sex, and race) and illness scores (Injury Se-verity Score, Acute Physiology and Chronic Health Evaluation,Abbreviated Injury Scale, and Glasgow Coma Score [GCS])were collected for each patient.

Outcome AssessmentDischarge disposition was recorded for all patients. In

patients who survived past discharge, a functional outpatientevaluation interview 3- to 6-month postdischarge was obtainedusing the Glasgow Outcome Score-extended (GOS-e)19 andmodified Rankin Scale (mRS)20 scores. These data are acquired

routinely and entered into the BOMO registry by an outpatientnurse.

Analysis and StatisticsComparison between treatments, combination of treat-

ments, or no treatment was evaluated with analysis of varianceand post hoc analysis (Tukey) to examine their effect on time tonormalization of PbtO2 (�T), the magnitude of PbtO2 change(�PbtO2) as well as on �ICP and �CPP. To analyze whichtreatment or combination of treatments, if any, was superiorlinear regression analysis was used. SPSS software (SPSS,Chicago, IL) was used for analysis. Continuous data are pre-

Figure 1. Institutional algorithm for severe TBI management. ET tube, endotracheal tube; ABG, arterial blood gas; SjvO2, jug-ular venous bulb oxygenation; Hgb, hemoglobin; PaCO2, partial carbon dioxide tension of blood.

The Journal of TRAUMA® Injury, Infection, and Critical Care • Volume 70, Number 3, March 2011 Interventions for PbtO2-Guided TBI Management

© 2011 Lippincott Williams & Wilkins 537

sented as mean � SD unless otherwise specified and a two-tailed p value of �0.05 was considered statistically significant.

RESULTS

Study PopulationFour hundred sixty-two (462) patients who had 1,601

episodes of compromised PbtO2 (�20 mm Hg) were identi-

fied in the BOMO registry from September 30, 2001, toOctober 1, 2008 (Fig. 2). From these, 92 patients who had625 episodes of compromised PbtO2 (between 15 minutesand 4 hours) were selected once non-TBI patients and entrieswith insufficient data were removed. Two hundred eighty(280) compromised PbtO2 episodes received no interventionand normalized within 4 hours, whereas 345 episodes wereidentified in patients that received some form of interventionbefore normalization of PbtO2. The age and ethnic break-down of the patients included in this study are illustrated inFigures 3 and 4. Each patient had an admission head CT scan,and the findings are illustrated in Figure 5. Table 2 de-scribes the demographic and clinical characteristics oftreated patients.

Clinical Course and OutcomeOf the 92 patient cohort, 26% had more than one ICP

monitor placed, 9% had placement of a jugular venous bulbmonitor and 34% underwent an operative neurosurgical pro-cedure (evacuation of hematoma, decompressive craniec-tomy). Median hospital length of stay for all cohort patientswas 25 (0–146) days and in hospital mortality was 27.2%.The mean number of episodes of compromised PbtO2 inpatients who survived to discharge was 8 � 7 compared with5 � 5 in those who died in hospital, although those who diedgenerally had a shorter hospital stay and duration of PbtO2monitoring. Advanced age, especially �70 years old wasassociated with unfavorable outcome using both the GOS-e(p � 0.03) and mRS (p � 0.03). In addition, female sex wasassociated with better outcome (GOS-e, p � 0.02; mRS, p �0.01).

TABLE 1. Description of Interventions

Intervention BOMO-Coded Treatments

Paralytics Any addition/increase in dose of neuromuscularblocking agents

Cooling Any event where body temperature waspurposefully reduced below normothermia

Pressors Any addition/increase in dose of norepinephrine,phenylephrine, epinephrine, or vasopressin

FIO2/PEEP increases Any net increase of inspired oxygen or positiveend expiratory pressure

Narcotics/sedation Any addition/increase in dose of opiods,benzodiazepines, or propofol

Repositioning Any turn to right, left, head-of-bed elevation,lowering

Fluids Any bolus of crystalloids or colloids (excludingtransfusion therapy)

Osmotherapy Any boluses of mannitol or hypertonic saline

PRBCs Any transfusion of packed red blood cells

Ionotropes Any addition/increase in dose of milrinone,dopamine, or dobutamine

FFP Any transfusion of fresh frozen plasma

PRBC, packed red blood cells; FFP, fresh frozen plasma; PEEP, positive endexpiratory pressure.

Figure 2. Abstraction of study patients and low PbtO2 episodes.



Brain Oxygen TreatmentWe examined 92 patients who had 625 episodes of

compromised PbtO2 (280 episodes normalized without re-ceiving an intervention and 345 received some form ofintervention). Among episodes of compromised PbtO2 thatreceived one or more interventions, mean time to PbtO2normalization (�T) was greater (p � 0.05) but of similarmagnitude (�PbtO2) to that observed in episodes that cor-rected without treatment (Table 3). One hundred eighty-seven(54% of treated episodes) episodes of compromised PbtO2were treated successfully with a single intervention and 101(30% of treated episodes) with 2 interventions (Table 4; Fig.6). No episode of compromised PbtO2 required more thanfive interventions for PbtO2 correction. The time to PbtO2correction (�T) became longer (p � 0.0002) as the number of Figure 6. Number of interventions used in treated patients.

TABLE 2. Demographics and Patient Characteristics

Characteristic Mean � SD (%) (Range)

Sex (male/female) 72/20 (78.2/21.8%)

Age 41 � 19

ISS 34 � 12

Head AIS 4.88 � 0.42

ED arrival GCS 6 � 4

Patients receiving �1 ICP monitor 24 (26%)

Patients receiving and SjO2 monitor 8 (9.1%)

Patients with craniotomy/craniectomy 31 (34%)

Hospital LOS (days) 24.6 (0–146)

Hospital mortality 25 (27.2%)

Post discharge GOS-e 3 � 3

Post discharge mRS 4 � 2

AIS, Abbreviated Injury Scale; ED, emergency department; ISS, Injury SeverityScore; SjO2, jugular bulb venous oxygen; LOS, length of stay; GOS-e, GlasgowOutcome Score-extended; mRS, modified Rankin Scale.

TABLE 3. Attributes of Reduced PbtO2 Episodes

N �T (h) �PbtO2 (mm Hg)

No treatment 280 0.84 � 0.63 9.12 � 9.14

Any treatment 345 1.15 � 0.85 9.77 � 10.43

p �0.05 1.0

�T (h), duration of hypoxic episode (PbtO2 �20 mm Hg) in hours; �PbtO2 (mmHg), difference in mm Hg between first �20 mm Hg PbO2 recording and firstsubsequent �20 mm Hg PbtO2 recording.

TABLE 4. Comparison of Interventions Number on Time toPbtO2 Normalization

Interventions N

Episode Duration (�T)

Percentageof Treated Mean (h)* Range SD

No intervention 280 — 0.84 0.25–3.58 0.63

Any 1 intervention 187 54.2 1.03 0.25–3.97 0.79

Any 2 interventions 101 30.0 1.13 0.25–3.83 0.82




* ANOVA, analysis of variance: p � 0.0002 comparison among groups.

Figure 3. Age distribution.

Figure 4. Racial distribution.

Figure 5. TBI diagnoses by CT and clinical evaluation. SAH,subarachnoid hemorrhage; SDH, subdural hemorrhage; NOS,not otherwise specified; EDH, epidural hemorrhage; DAI, dif-fuse axonal injury; IPH, intraparenchymal hemorrhage.



interventions increased. The number of patients treated withmore than two interventions was small; however, �T wasgreater in patients that had more than two interventions thanthose with one or two interventions (p � 0.05; Table 5). The

magnitude of PbtO2 (�PbtO2) increase averaged 9.5 mmHg � 9.9 mm Hg and was not associated with number ofinterventions used (p � 0.53). The average (SD) �T when nointervention was administered was 0.84 hours � 0.63 hours.

What Interventions Were Used to CorrectCompromised PbtO2?

Only one or two interventions were used to correctcompromised PbtO2 in the majority of episodes and sosubsequent analyses were limited to these groups. The com-monest strategies used to treat compromised PbtO2 are sum-marized in Table 6. The most common single interventionswere sedation or analgesia (N � 60), pressors (N � 51),increase in FIO2/PEEP (N � 24), and patient repositioning(N � 27; Table 7). Of these, pressors were associated withthe lowest �T (fastest correction). Single interventions asso-ciated with the greatest �PbtO2 were osmotherapy (14 mmHg � 11 mm Hg) and red cell transfusion (27 mm Hg) but thefrequency of these interventions was low. No one interven-tion appeared superior using simple linear regression analysis(R2 � 0.2; p � 0.05). One hundred one episodes were treatedwith two interventions; the most common was sedation or

TABLE 5. Comparison Between Number of Interventions asDescribed in Table 4

No. of Intervention(s) p

1 vs. 3 0.00

1 vs. 4 0.01

1 vs. 5 0.01

2 vs. 5 0.01

2 vs. 4 0.03

2 vs. 3 0.03

5 vs. 3 0.07

5 vs. 4 0.24

2 vs. 1 0.33

4 vs. 3 0.38

Tukey posthoc analysis comparing the number of interventions used on time toPbtO

2correction (�T). One or two interventions always correct PbtO2 significantly

faster than 3, 4, or 5 interventions. There is no significance difference when using 1 or2 interventions to correct PbtO2.

TABLE 6. Most Popular Combinations of Interventions for Reduced PbtO2 Deficits

Intervention 1 Intervention 2 N Episode Duration �T (h) Magnitude �PbO2 (mm Hg)

Narcotics/sedation 60 1.02 � 0.77 10.6 � 12.1

Pressors 51 0.94 � 0.87 7.9 � 8.4

Repositioning 27 1.11 � 0.67 9.1 � 11.8

FIO2/PEEP increases 24 0.99 � 0.51 9.5 � 8.8

Sedation/narcotics Pressors 19 1.11 � 0.86 9.5 � 7.9

Fluids 9 1.22 � 0.89 11.9 � 12.7

Sedation/narcotics FIO2/PEEP increases 9 1.71 � 1.18 9 � 5.3

Sedation/narcotics Reposition 9 1.11 � 0.86 10.1 � 9.8

Sedation/narcotics Fluids 8 0.98 � 0.35 7.2 � 4.8

Pressors Fluids 7 1.25 � 0.81 7.6 � 6.9

Sedation/narcotics Osmotherapy 5 0.25 � 0.53 16.9 � 14.5

PEEP, positive end expiratory pressure.

TABLE 7. Single Treatment Combinations on �T and �PbtO2

Intervention

Episode Duration (�T) Magnitude of Change (�PbtO2)

N Mean (h) Range SD Mean (mm Hg) Range SD

Paralytics 4 0.4 0.25–0.5 0.12 8.4 5.7–14 4.01

Cooling 4 0.91 0.47–1.5 0.51 6.7 1.7–18 7.8

Pressors 51 0.94 0.25–3.97 0.88 7.9 1.3–46.6 8.35

FIO2/PEEP increases 24 0.99 0.27–2.5 0.52 9.5 0.3–40 8.83

Narcotics/sedation 60 1.02 0.25–3.0 0.77 10.6 �6.8 to 62.7 12.16

Repositioning 27 1.11 0.25–3.0 0.67 9.1 0.4–57 11.81

Fluids 9 1.22 0.25–3.2 0.89 11.9 1.1–40 12.68

Osmotherapy 7 1.33 0.33–3.7 1.07 14.4 3.2–50 16.85

PRBCs 1 3.65 3.65–3.65 0 27.5 28–28 0

Ionotropes 0 N/A N/A N/A N/A N/A N/A

FFP 0 N/A N/A N/A N/A N/A N/A

Any single treatment 187 1.03 0.25–4.0 0.68 11.8 �6.8–63 10.31

PEEP, positive end expiratory pressure; PRBC, packed red blood cells; FFP, fresh frozen plasma.



analgesia � pressors (N � 19; Table 8). Cooling � paralytics(n � 1 episodes) was associated with the lowest �T, andosmotherapy � FIO2/PEEP increase (n � 2 episodes) withthe greatest �PbtO2. An osmotherapy � sedation or narcoticscombination appeared to demonstrate an optimal �PbtO2/�Tcombination (16.9 mm Hg � 14.5 mm Hg/0.82 hours � 0.5hours) but was only used in five episodes. When controllingfor age, sex, Injury Severity Score, and arrival GCS nointervention or combination of interventions appeared to bebetter than others in correction of compromised PbtO2.

PbtO2 Treatment and the Effect on ICP andCPP

We next examined how PbtO2-related treatment influ-enced ICP and CPP during the same analyzed episodes (Table9). As with �T and �PbtO2, use of osmotherapy � sedationor narcotics seemed to best manage ICP but did not translateinto an important CPP increase. Pressors, fluids and reposi-tioning were associated with the greatest increases on CPP.When evaluating the number of interventions on �ICP, anythree-intervention regimen (�6.40 mm Hg � 2.5 mm Hg)had greater ICP reduction than no intervention (�1.19 mmHg � 0.47 mm Hg, p � 0.009), or one intervention (�1.62mm Hg � 0.61 mm Hg, p � 0.03). CPP was most increasedwith combinations of two interventions (11.1 mm Hg � 2.4

mm Hg), which was significantly greater than that with nointervention (3.8 mm Hg � 1.2 mm Hg, p � 0.02). We alsoevaluated all compromised PbtO2 episodes that underwentone or more interventions and found a mean decrease of ICPof 1.25 mm Hg � 5.9 mm Hg and CPP increase of 7.02 �17.7 in the interval evaluated.

TABLE 8. Double Treatment Combinations on �T and �PbtO2

Interventions

Event Duration (�T)Magnitude of Change

(�PbtO2)

N Mean (h) (Range) SD Mean (mm Hg) (Range) SD

Cooling Paralytics 1 0.33 (0.33–0.33) N/A 11.9 (11.9–11.9) N/A

Osmotherapy Paralytics 2 0.54 (0.42–0.7) 0.18 6.8 (3.2–10.4) 5.09

Paralytics Sedation/narcotics 3 0.63 (0.5–0.75) 0.13 5.3 (1.8–9.7) 4.03

Pressors Reposition 4 0.67 (0.42–1.0) 0.26 11.0 (1.2–32) 14.27

Reposition PRBCs 4 0.67 (0.42–1.0) 0.26 11.0 (1.2–32) 14.27

Cooling Sedation/narcotics 5 0.78 (0.25–1.2) 0.46 11.4 (0.6–37.4) 15.27

Osmotherapy Sedation/narcotics 5 0.82 (0.25–1.5) 0.53 16.9 (4.8–37) 14.48

Pressors FFP 2 0.94 (0.88–1.0) 0.08 10.0 (6.2–138) 5.37

Pressors Sedation/narcotics 19 0.99 (0.25–3.2) 0.83 8.4 (1.6–17.5) 4.92

Fluids Sedation/narcotics 8 0.99 (0.50–1.5) 0.35 7.3 (3.1–17.5) 4.84

Cooling FIO2/PEEP1 1 1.00 (1–1) N/A 14.1 (14.1–14.1) N/A

FIO2/PEEP 1 Reposition 4 1.01 (0.57–1.7) 0.48 4.9 (3.2–6.6) 1.63

Cooling Pressors 1 1.03 (1.03–1.03) N/A 1.4 (1.4–1.4) N/A

FIO2/PEEP 1 Pressors 6 1.06 (0.33–3.5) 1.22 9.6 (1.8–17.8) 5.92

Reposition Sedation/narcotics 9 1.11 (0.37–2.9) 0.86 10.1 (�2.2 to 33.6) 9.75

FIO2/PEEP 1 Osmotherapy 2 1.13 (0.5–1.8) 0.88 42.4 (5.8–79.0) 51.76

Fluids Pressors 7 1.25 (0.25–2.8) 0.81 7.6 (1.7–21.3) 6.94

FIO2/PEEP 1 Fluids 3 1.31 (0.5–2.6) 1.1 6.0 (4.7–8.0) 1.74

FIO2/PEEP 1 Sedation/narcotics 9 1.74 (0.33–3.8) 1.26 8.2 (2.2–14.6) 4.88

Cooling Reposition 1 2.00 (2–2) N/A 4.4 (4.4–4.4) N/A

Fluids Osmotherapy 2 2.08 (1.2–3.0) 1.3 3.9 (2.6–5.2) 1.84

FIO2/PEEP 1 PRBCs 1 2.58 (2.6–2.6) N/A 13.2 (13.2–13.2) N/A

Osmotherapy Pressors 1 3.00 (3–3) N/A 2.8 (2.8–2.8) N/A

Osmotherapy Reposition 1 3.00 (3–3) N/A 2.8 (2.8–2.8) N/A

Any two treatments 101 1.28 (0.25–3.8) 0.65 9.6 (�2.2 to 79) 9.82

PEEP, positive end expiratory pressure; PRBC, packed red blood cells; FFP, fresh frozen plasma.

TABLE 9. Most Popular Intervention Combinations on ICPand CPP

N Intervention 1 Intervention 2 �ICP �CPP

60 Narcotics/sedation �2.6 � 11.2 4.0 � 16.6

51 Pressors �1.4 � 7.5 8.3 � 18.9

27 Repositioning �0.8 � 7.8 10.2 � 12.8

24 FIO2/PEEPincreases

�0.1 � 4.8 5.3 � 15.4

19 Sedation/narcotics Pressors �1.2 � 6.8 9.6 � 22.7

9 Fluids �0.8 � 2.8 6.8 � 16.9

9 Sedation/narcotics FIO2/PEEP increases 2.1 � 5.3 �7.1 � 10.2

9 Sedation/narcotics Reposition 0.5 � 4.4 20 � 35.2

8 Sedation/narcotics Fluids �2.0 � 4.2 4.7 � 7.7

7 Pressors Fluids 0.3 � 2.5 13.9 � 26.2

5 Sedation/narcotics Osmotherapy �7.8 � 7.9 1.5 � 12.3



DISCUSSIONIn this study, we examined 92 severe TBI patients with

625 episodes of compromised PbtO2 (�20 mm Hg for0.25–4 hours) and how these episodes were treated. Weobserved the following: (1) 44% of episodes were correctedwithout specific treatment; (2) most episodes of compromisedPbtO2 could be successfully treated with one or two inter-ventions and none required more than five interventions; (3)no single intervention or combination of interventions ap-peared better than any other to reduce time to PbtO2 correc-tion or to increase magnitude of PbtO2 change; (4) use ofmore interventions was associated with a greater time toPbtO2 normalization; (5) many episodes underwent PbtO2

normalization without interventions, and (6) an increasing thenumber of interventions to a maximum of three only had afavorable effect on ICP or CPP. These data may be used toguide therapy of compromised PbtO2 and suggest what effectcan be expected.

Methodological LimitationsThis study has several potential limitations. First, it is a

retrospective analysis of interventions recorded in a prospec-tive observational database. Second, the sample size of 92patients is relatively small. However, we examined �600episodes of compromised PbtO2. Third, this is a singleinstitution series. Our results, therefore, may lack externalvalidity and should be considered preliminary. Fourth, wedefined an episode of compromised PbtO2 as 15 minutes to240 minutes in duration. The lower time limit was arbitrarilychosen to prevent transient self-limited changes in PbtO2, i.e.,episodes of coughing, straining, or moving that bedsideclinicians do not routinely treat. Our upper time limit alsowas arbitrarily chosen. However, we felt it reasonable toconclude that if no treatment was initiated within 4 hours, thatcare had been deliberately withheld. Fifth, although all spe-cific treatments were prospectively coded in our BOMOregistry, we chose, for the purposes of this study, to examinetreatment classes or interventions (see Methods), thus, anyeffects of the specific drug or fluid administered are beyondthe scope of this study. Sixth, we did not control for thenumber of times a specific intervention was used to correct asingle episode of compromised PbtO2 or in which ordertherapies were given. It was not our goal to describe a preciserecipe to treat compromised PbtO2 because such a recipe isunlikely to exist. Seventh, the intention behind the clinician’suse of a given intervention was not examined and the use ofa given intervention could have occurred for reasons otherthan to correct PbtO2 deficits. Finally, use of �T as a primaryoutcome may have lead to intrinsic bias because the greaterduration of compromised PbtO2 would inherently allow cli-nicians more time to administer more interventions. In addi-tion, defining the episode endpoint as normalized PbtO2 (�20mm Hg) may have underestimated any intervention’s maxi-mal effect on �PbtO2. Despite these limitations, the results ofthis study from an institution with several years experienceusing and studying PbtO2 monitoring by a group of interdis-ciplinary clinicians provide an in-depth description of what

interventions are usually administered to patients with com-promised PbtO2 and what results can be expected.

Significance of Reduced PbtO2Increased ICP and reduced CPP are associated with

mortality and poor outcome in TBI.21–24 Consequently, anICP monitor is recommended in current severe TBI guide-lines in part to also maintain CPP.5,6 Although it may appearphysiologically plausible, there is no Level I evidence tosupport the role of an ICP monitor (or any monitor) in TBIcare. However, some recent observational cohort studiescontinue to question the use of ICP monitors,25,26 and a recentmeta-analysis of the literature suggests that use of an ICPmonitor is associated with better outcome.27

In a separate set of recent studies, the concept thatcellular hypoxia or dysfunction may occur when ICP andCPP are normal has emerged.8,28–30 In particular, positronemission tomography (PET) and microdialysis studies havefound that after TBI, cellular hypoxia or anerobic metabolismoften is associated with defects in oxygen diffusion and maybe independent of perfusion,2,3,28,31,32 and therefore not cou-pled with ICP variations. Consistent with this, several obser-vational clinical studies found that mortality and pooroutcome can be associated with brain hypoxia particularlyof greater duration,10,13–15,33,34 magnitude (�15 mmHg),10,13–15,33,34 or frequency.13–15 Consequently, the mostrecent edition of the BTF Guidelines recommended the useof a brain tissue oxygen monitor.35

PbtO2-Based Care of Severe TBISeveral groups have described the use of a PbtO2

monitor and PbtO2-based care to supplement ICP and CPP-based care of severe TBI.11–15 Management strategies includeprotocols to correct CPP when reduced PbtO2 is observed12 ortiered approaches based on physiologic targets although thespecific individual interventions often are not well detailed.13

Our usual protocol is illustrated in Figure 1 and described inprevious publications.14 There are seven published reports, allnonrandomized, that compare PbtO2 and ICP or CPP-guidedTBI management strategies.11–16,36 Six of the studies suggesta potential benefit to PbtO2-based care and a pooled analysisindicates that PbtO2-based care is associated with a twofoldincrease chance of favorable outcome.37 However, Martiniet al.11 observed increased hospital mortality associated withPbtO2-based care although this was no longer a significantrelationship when adjusted for variables such as age, headAbbreviated Injury Scale, Marshall CT classification, andGCS. In addition, these authors found that PbtO2-guided carewas associated with more use of osmotherapy, vasopressors,and prolonged sedation. However, we have found that amongpatients treated with PbtO2-based care that mortality is asso-ciated with longer periods of compromised PbtO2 deficits(�T), and with compromised PbtO2 that is less responsive totreatment.14 This balance between effective treatment andovertreatment is crucial because every therapy has potentialside-effects and although brain physiology may be improved,this result may not always translate into better outcome ifother organ systems are harmed.38



Therapeutic Interventions for CompromisedPbtO2

Although clinical series suggest that there may be abenefit to PbtO2-based care, it remains unclear what consti-tutes this care as the specific therapies for compromisedPbtO2 are only beginning to be defined. In this study, wedescribe what interventions may be used and their expectedefficacy when PbtO2 is compromised. Our registry (BOMO)codes �80 clinician interventions in the ICU. This includesspecific items such as “mouthcare,” “proning,” “abdominaloperation,” or “patient reposition” that are recorded in thenursing record at time intervals as short as 10 minutes to 20minutes. We conducted a detailed evaluation of these codedtreatments and, with expert consensus, grouped them in 11classes of interventions (Table 6). Although many of theseinterventions are used by clinicians caring for TBI patientsworldwide and often in the setting of reduced ICP or MAP,there is limited study of how these interventions affect com-promised PbtO2. We thus sought to first delineate whatinterventions were used most frequently in TBI patients withcompromised PbtO2. Greater than one third of such episodesPbtO2 resolved without any apparent intervention. Theseepisodes may simply have been self-limited and thus raise thequestion on how to identify reductions in PbtO2 that do notrequire treatment. However, they also may represent anepisode where some other trigger (e.g., low ICP, high CPP)resulted in treatment initiation before the identified decreasein PbtO2 and the effect of these intervention(s) eventuallyalso normalized PbtO2. Yet, when evaluating the effect of anyintervention or combination of interventions on ICP and CPPeffects were at most modest.

Specific Interventions for CompromisedPbtO2—Alone or in Combination?

The use of a single intervention for compromised PbtO2was the most frequent strategy, and among these, narcoticanalgesics or sedatives, often used in TBI to control ICP,39–42

was most common (Table 4). Pressors were the next mostfrequent therapy. These agents are also frequently used inTBI care, but which specific agent (phenylephrine, dopamine,norepinephrine) is preferable is not well defined.43,44 In ad-dition, overuse of pressors may exacerbate lung function.5 Ofinterest, we observed that, among the most popular treat-ments, pressors appeared to be associated with the shortesttime interval for PbtO2 correction (�T). However, this maymean pressors were only used late or after other interventionsfailed. Patient repositioning was the third most commonintervention we observed and often increased PbtO2 by 10mm Hg or more. Head and neck repositioning, elevating thehead of the bed, turning a patient or loosening a cervicalcollar are common practice in TBI patients, but the directeffects and durability of these interventions may be moreopinion than fact. These maneuvers, however, are recom-mended as good clinical care and are likely to benefit thepatient without significant risk. It is conceivable that weunderestimated their use and effect because they most likelywere applied as part of general care rather than when aspecific monitored abnormality occurred. Efforts to increase

available oxygen through increases in FIO2, PEEP, or pul-monary dynamics (prostacyclin in other studies) were thefourth most common intervention in our study.45–49 Thedefinitive benefits and, more importantly, risks (oxygentoxicity) of hyperoxic treatment in TBI, however, remainunknown.

Combined use of sedation or narcotics � osmotherapyresulted in the most rapid correction of compromised PbtO2(about 15 minutes) and also appeared to increase PbtO2and reduce ICP by the greatest margins. It must be noted,however, that the occurrence of this combination wasexceeding low and therefore no durable conclusion can bemade. Both mannitol and hypertonic saline were included inthe osmotherapy intervention class and are well known toreduce ICP.50,51 Hypertonic saline in particular has beenshown to improve PbtO2 in TBI patients52 refractive tomannitol therapy. Efforts to enhance oxygen delivery (DO2)through transfusion of red blood cells is commonly usedthough an improvement in PbtO2 is not always present.53,54 Inaddition, there remains controversy on what is the optimalhemoglobin level for TBI patients.7,55 There is limited data onhow other blood products (e.g., plasma) affect PbtO2.56 Acombination of pressors and fluid boluses may also improveDO2 although studies in subarachnoid hemorrhage suggestonly induced hypertension has a positive effect on PbtO2.57 Inour study, combined pressors and fluids was associated withthe greatest increase of CPP (�14 mm Hg). Although infu-sions of crystalloids or colloids are commonplace in the ICUand thought to benefit TBI patients by raising MAP and CPP,their effect on PbtO2 remains intuitive with limited scientificevaluation.58 Indeed, some studies suggests PbtO2 monitoringmay lead to overuse of fluids.59 Therapies such as inducedhypothermia and cerebral spinal fluid (CSF) drainage wereused infrequently in our patient cohort and so we cannotmake specific conclusions about their use.

CONCLUSIONSAccumulating evidence suggests that PbtO2-directed

therapy may provide an advantage over ICP or CPP onlyguided management. However, there is little data on whatinterventions should be included in such a PbtO2 treatment“bundle.” Bundled treatments have improved the care ofcritically ill patients in other fields such as sepsis.60 However,the individual components may not demonstrate this advan-tage in the absence of the remaining components and theenvironment where they are administered. This study doesnot provide a recipe for the best ICU intervention or combi-nation of interventions to treat compromised PbtO2 in TBIpatients. However, we have learnt several important points.First, clinicians tend to use a small number of treatments andprefer to use combinations of at most one or two to correctcompromised PbtO2. Second, the most common interventionsmay improve ICP or CPP but may not always increase PbtO2.Third, use of more treatments does not mean more rapidcorrection of compromised PbtO2. This is particularlyimportant because some interventions to correct intracra-nial abnormalities are known to exacerbate injury in otherorgan systems of critically ill patients. Finally, some PbtO2



compromises may correct on their own and how to identifysuch self-resolving deficits remains unclear. In summary, thisstudy provides an insight in what interventions are most usedin a center familiar with PbtO2-directed TBI care. The effectsof these popular interventions will need further evaluationand comparisons to establish their efficacy, safety, timing,and sequencing for future severe TBI management.

REFERENCES1. Murray GD, Teasdale GM, Braakman R, et al. The European Brain

Injury Consortium survey of head injuries. Acta Neurochir (Wien).1999;141:223–236.

2. Menon DK, Coles JP, Gupta AK, et al. Diffusion limited oxygendelivery following head injury. Crit Care Med. 2004;32:1384–1390.

3. Vespa P, Bergsneider M, Hattori N, et al. Metabolic crisis without brainischemia is common after traumatic brain injury: a combined microdi-alysis and positron emission tomography study. J Cereb Blood FlowMetab. 2005;25:763–774.

4. Ross DT, Graham DI, Adams JH. Selective loss of neurons from thethalamic reticular nucleus following severe human head injury.J Neurotrauma. 1993;10:151–165.

5. Brain Trauma Foundation; American Association of Neurological Sur-geons; Congress of Neurological Surgeons. Guidelines for the managementof severe traumatic brain injury. J Neurotrauma. 2007;24(suppl 1):S1–S106.

6. Brain Trauma Foundation; American Association of Neurological Sur-geons; Congress of Neurological Surgeons; Joint Section on Neu-rotrauma and Critical Care, AANS/CNS, Carney NA, Ghajar J. Guide-lines for the management of severe traumatic brain injury. Introduction.J Neurotrauma. 2007;24(suppl 1):S1–S2.

7. Chang JJ, Youn TS, Benson D, et al. Physiologic and functionaloutcome correlates of brain tissue hypoxia in traumatic brain injury. CritCare Med. 2009;37:283–290.

8. Gopinath SP, Robertson CS, Contant CF, et al. Jugular venous desatu-ration and outcome after head injury. J Neurol Neurosurg Psychiatry.1994;57:717–723.

9. Maloney-Wilensky E, Gracias V, Itkin A, et al. Brain tissue oxygen andoutcome after severe traumatic brain injury: a systematic review. CritCare Med. 2009;37:2057–2063.

10. van den Brink WA, van Santbrink H, Steyerberg EW, et al. Brainoxygen tension in severe head injury. Neurosurgery. 2000;46:868–876;discussion 876–868.

11. Martini RP, Deem S, Yanez ND, et al. Management guided by braintissue oxygen monitoring and outcome following severe traumatic braininjury. J Neurosurg. 2009;111:644–649.

12. Meixensberger J, Jaeger M, Vath A, Dings J, Kunze E, Roosen K. Braintissue oxygen guided treatment supplementing ICP/CPP therapy after trau-matic brain injury. J Neurol Neurosurg Psychiatry. 2003;74:760–764.

13. Narotam PK, Morrison JF, Nathoo N. Brain tissue oxygen monitoring intraumatic brain injury and major trauma: outcome analysis of a braintissue oxygen-directed therapy. J Neurosurg. 2009;111:672–682.

14. Spiotta AM, Stiefel MF, Gracias VH, et al. Brain tissue oxygen-directedmanagement and outcome in patients with severe traumatic brain injury.J Neurosurg. 2010;113:571–580.

15. Stiefel MF, Spiotta A, Gracias VH, et al. Reduced mortality rate inpatients with severe traumatic brain injury treated with brain tissueoxygen monitoring. J Neurosurg. 2005;103:805–811.

16. Adamides AA, Cooper DJ, Rosenfeldt FL, et al. Focal cerebral oxygen-ation and neurological outcome with or without brain tissue oxygen-guided therapy in patients with traumatic brain injury. Acta Neurochir(Wien). 2009;151:1399–1409.

17. Wilensky EM, Bloom S, Leichter D, et al. Brain tissue oxygen practiceguidelines using the LICOX CMP monitoring system. J Neurosci Nurs.2005;37:278–288.

18. Siesjo BK, Siesjo P. Mechanisms of secondary brain injury. Eur JAnaesthesiol. 1996;13:247–268.

19. Wilson JT, Pettigrew LE, Teasdale GM. Structured interviews for theGlasgow Outcome Scale and the extended Glasgow Outcome Scale:guidelines for their use. J Neurotrauma. 1998;15:573–585.

20. van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van GijnJ. Interobserver agreement for the assessment of handicap in strokepatients. Stroke. 1988;19:604–607.

21. Marshall LF, Smith RW, Shapiro HM. The outcome with aggressivetreatment in severe head injuries. Part II: acute and chronic barbiturateadministration in the management of head injury. J Neurosurg. 1979;50:26–30.

22. Marshall LF, Smith RW, Shapiro HM. The outcome with aggressivetreatment in severe head injuries. Part I: the significance of intracranialpressure monitoring. J Neurosurg. 1979;50:20–25.

23. Miller JD, Butterworth JF, Gudeman SK, et al. Further experience in themanagement of severe head injury. J Neurosurg. 1981;54:289–299.

24. Narayan RK, Kishore PR, Becker DP, et al. Intracranial pressure: tomonitor or not to monitor? A review of our experience with severe headinjury. J Neurosurg. 1982;56:650–659.

25. Cremer OL, van Dijk GW, van Wensen E, et al. Effect of intracranialpressure monitoring and targeted intensive care on functional outcomeafter severe head injury. Crit Care Med. 2005;33:2207–2213.

26. Shafi S, Diaz-Arrastia R, Madden C, Gentilello L. Intracranial pressuremonitoring in brain-injured patients is associated with worsening ofsurvival. J Trauma. 2008;64:335–340.

27. Stein SC, Georgoff P, Meghan S, Mirza KL, El Falaky OM. Relationshipof aggressive monitoring and treatment to improved outcomes in severetraumatic brain injury. J Neurosurg. 2010;112:1105–1112.

28. Le Roux PD, Newell DW, Lam AM, Grady MS, Winn HR. Cerebralarteriovenous oxygen difference: a predictor of cerebral infarction andoutcome in patients with severe head injury. J Neurosurg. 1997;87:1–8.

29. Stein SC, Graham DI, Chen XH, Smith DH. Association betweenintravascular microthrombosis and cerebral ischemia in traumatic braininjury. Neurosurgery. 2004;54:687–691; discussion 691.

30. van Santbrink H, vd Brink WA, Steyerberg EW, Carmona Suazo JA,Avezaat CJ, Maas AI. Brain tissue oxygen response in severe traumaticbrain injury. Acta Neurochir (Wien). 2003;145:429–438; discussion438.

31. Stiefel MF, Udoetuk JD, Spiotta AM, et al. Conventional neurocriticalcare and cerebral oxygenation after traumatic brain injury. J Neurosurg.2006;105:568–575.

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

33. Bardt TF, Unterberg AW, Hartl R, Kiening KL, Schneider GH, LankschWR. Monitoring of brain tissue PO2 in traumatic brain injury: effect ofcerebral hypoxia on outcome. Acta Neurochir Suppl. 1998;71:153–156.

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

35. Brain Trauma Foundation; American Association of Neurological Sur-geons; Congress of Neurological Surgeons; Joint Section on Neu-rotrauma and Critical Care, AANS/CNS; Bratton SL, Chestnut RM,Ghajar J, et al. Guidelines for the management of severe traumatic braininjury. X. Brain oxygen monitoring and thresholds. J Neurotrauma.2007;24(suppl 1):S65–S70.

36. Murray GD, Butcher I, McHugh GS, et al. Multivariable prognosticanalysis in traumatic brain injury: results from the IMPACT study.J Neurotrauma. 2007;24:329–337.

37. Nangunoori RM-WE, Stiefel M, Park S, et al. Brain Tissue OxygenBased Therapy and Outcome After Severe Traumatic Brain Injury: ASystematic Literature Review. Las Vegas, Nevada: National Neu-rotrauma Annual Meeting; 2010.

38. Contant CF, Valadka AB, Gopinath SP, Hannay HJ, Robertson CS.Adult respiratory distress syndrome: a complication of induced hyper-tension after severe head injury. J Neurosurg. 2001;95:560–568.

39. Gremmelt A, Braun U. Analgesia and sedation in patients with head-brain trauma. Anaesthesist. 1995;44(suppl 3):S559–S565.

40. Jagannathan J, Okonkwo DO, Yeoh HK, et al. Long-term outcomes andprognostic factors in pediatric patients with severe traumatic brain injuryand elevated intracranial pressure. J Neurosurg Pediatr. 2008;2:240–249.

41. Sabsovich I, Rehman Z, Yunen J, Coritsidis G. Propofol infusionsyndrome: a case of increasing morbidity with traumatic brain injury.Am J Crit Care. 2007;16:82–85.

42. Schmittner MD, Vajkoczy SL, Horn P, et al. Effects of fentanyl andS(�)-ketamine on cerebral hemodynamics, gastrointestinal motility, and



need of vasopressors in patients with intracranial pathologies: a pilotstudy. J Neurosurg Anesthesiol. 2007;19:257–262.

43. Pfister D, Strebel SP, Steiner LA. Effects of catecholamines on cerebralblood vessels in patients with traumatic brain injury. Eur J AnaesthesiolSuppl. 2008;42:98–103.

44. Myburgh JA. Driving cerebral perfusion pressure with pressors: how,which, when? Crit Care Resusc. 2005;7:200–205.

45. Tolias CM, Reinert M, Seiler R, Gilman C, Scharf A, Bullock MR.Normobaric hyperoxia-induced improvement in cerebral metabolismand reduction in intracranial pressure in patients with severe head injury:a prospective historical cohort-matched study. J Neurosurg. 2004;101:435–444.

46. Menzel M, Doppenberg EM, Zauner A, Soukup J, Reinert MM, BullockR. Increased inspired oxygen concentration as a factor in improved braintissue oxygenation and tissue lactate levels after severe human headinjury. J Neurosurg. 1999;91:1–10.

47. Menzel M, Doppenberg EM, Zauner A, et al. Cerebral oxygenation inpatients after severe head injury: monitoring and effects of arterialhyperoxia on cerebral blood flow, metabolism and intracranial pressure.J Neurosurg Anesthesiol. 1999;11:240–251.

48. Diringer MN. Hyperoxia: good or bad for the injured brain? Curr OpinCrit Care. 2008;14:167–171.

49. Stiefel MF, Zaghloul KA, Bloom S, Gracias VH, LeRoux PD. Improvedcerebral oxygenation after high-dose inhaled aerosolized prostacyclin ther-apy for acute lung injury: a case report. J Trauma. 2007;63:1155–1158.

50. Wakai A, Roberts I, Schierhout G. Mannitol for acute traumatic braininjury. Cochrane Database Syst Rev. 2007;1:CD001049.

51. Ogden AT, Mayer SA, Connolly ES Jr. Hyperosmolar agents in neuro-surgical practice: the evolving role of hypertonic saline. Neurosurgery.2005;57:207–215; discussion 207–215.

52. Pascual JL, Maloney-Wilensky E, Reilly PM, et al. Resuscitation ofhypotensive head-injured patients: is hypertonic saline the answer? AmSurg. 2008;74:253–259.

53. Smith MJ, Stiefel MF, Magge S, et al. Packed red blood cell transfusionincreases local cerebral oxygenation. Crit Care Med. 2005;33:1104–1108.

54. Zygun DA, Nortje J, Hutchinson PJ, Timofeev I, Menon DK, Gupta AK.The effect of red blood cell transfusion on cerebral oxygenation andmetabolism after severe traumatic brain injury. Crit Care Med. 2009;37:1074–1078.

55. Fluckiger C, Bechir M, Brenni M, et al. Increasing hematocrit above28% during early resuscitative phase is not associated with decreasedmortality following severe traumatic brain injury. Acta Neurochir(Wien). 2010;152:627–636.

56. Senft C, Schuster T, Forster MT, Seifert V, Gerlach R. Management andoutcome of patients with acute traumatic subdural hematomas andpre-injury oral anticoagulation therapy. Neurol Res. 2009;31:1012–1018.

57. Muench E, Horn P, Bauhuf C, et al. Effects of hypervolemia andhypertension on regional cerebral blood flow, intracranial pressure, andbrain tissue oxygenation after subarachnoid hemorrhage. Crit Care Med.2007;35:1844–1851; quiz 1852.

58. Schmoker JD, Shackford SR, Wald SL, Pietropaoli JA. An analysis ofthe relationship between fluid and sodium administration and intracra-nial pressure after head injury. J Trauma. 1992;33:476–481.

59. Fletcher JJ, Bergman K, Blostein PA, Kramer AH. Fluid balance,complications, and brain tissue oxygen tension monitoring followingsevere traumatic brain injury. Neurocrit Care. 2010;13:47–56.

60. Rivers E, Nguyen B, Havstad S, et al; Early Goal-Directed TherapyCollaborative Group. Early goal-directed therapy in the treatment ofsevere sepsis and septic shock. N Engl J Med. 2001;345:1368–1377.

DISCUSSIONDr. William C. Chiu (Baltimore, Maryland): Dr. Pas-

cual and his coauthors from the University of Pennsylvaniahave presented another study investigating the utility of braintissue oxygen-directed management of traumatic brain injury.

Secondary brain injury is associated with episodes ofcerebral ischemia and hypoxia. Most current algorithms for

neurologic monitoring use intracranial and cerebral perfusionpressure monitoring-based therapy.

Previous work from this group has suggested that braintissue oxygen monitoring is safe and that low brain tissueoxygen can be corrected, and may be associated with reducedmortality.

This study has confirmed their previous findings thatcommon interventions employed for TBI management suc-cessfully improve episodes of low brain tissue oxygen andthat just one or two interventions normalize brain tissueoxygen in the majority of patients.

There are two particular study results that I found mostintriguing: one, the combination of sedation and osmotherapyresulted in the most rapid correction of compromised braintissue oxygen and reduction of ICP.

Change in CPP with this combination was not soimpressive. Instead, the combination of pressors and fluidsresulted in the greatest increase in CPP but the change inbrain tissue oxygen and ICP were only mild.

These findings suggest that interventions that reduceICP also improve brain tissue oxygen but not CPP. It alsoreminds us that interventions that improve CPP may simplyimprove mean arterial pressure without improving ICP.

Two, the proportion of reduced brain tissue oxygenthat normalized without treatment was 45 percent. Further-more, normalization of brain tissue oxygen was more rapidin those patients without intervention compared to thosewith intervention.

Patients requiring an incremental increase in numberof interventions required an associated increased time fornormalization of brain tissue oxygen. This finding remindsus of the difficulty in establishing causality in retrospectiveresearch.

Patients requiring five interventions to normalize braintissue oxygen may have had more resistant hypoxia. There-fore, it is probably unfair to conclude that increasing thenumber of interventions resulted in worsening the time tobrain tissue oxygen correction.

I just have two questions for Dr. Pascual, one, in thegroup of 280 instances of low brain tissue oxygen with nointervention the time to normalization was a mean of 50minutes. Please speculate on the justification for no interven-tions during 50 minutes of brain tissue hypoxia.

And, two, based upon your institution’s experience,which patients, then, do you recommend should have braintissue oxygen-directed management?

Dr. Jose L. Pascual (Philadelphia, Pennsylvania):Thank you, Dr. Chiu for your insightful comments.

I have to say that while interesting, the sedation/narcoticsplus osmotherapy combination is difficult to interpret in thesetting of such few occurrences in the whole sample size.

To address the specific questions, we also were won-dering two questions: how is it that no intervention for a meantime of 50 minutes results still in correction of brain tissueoxygen?

We speculated a few answers. One, that brain tissueoxygen in some cases does correct on its own without anyintervention.



Another is that maybe some interventions occurredbefore the drop in brain tissue oxygen because ICP or CPPtook a poor turn. ICP was increased; CPP was decreasedand interventions were performed just before the drop inbrain tissue oxygen and then affected brain oxygen.

Also, perhaps we didn’t capture interventions such as abedside nurse repositioning the patient and not recording it inthe record.

Why would people not treat a patient with 50 minutes ofbrain tissue decreases in oxygenation? Perhaps because they

were being moved or the thought was that this would resolve onits own and interventions had occurred that were not recorded bythe bedside nurse they waited for the effect to occur.

Who should we monitor brain tissue oxygen on remainsa difficult question. If greater than a third of patients correcttheir brain tissue oxygenation on their own, this may besomething very important in the decision tree of what type ofintracranial monitor to insert.

I’d like to thank the association, Dr. Jurkovich and Dr.Cioffi. Thank you.



Documents

Reduced Brain Tissue Oxygen in Traumatic Brain Injury: Are ......Key Words: Brain tissue oxygenation, Traumatic brain injury, Treatment interventions, PbtO 2-guided management, Clinical