1Linton Evans, MD., 1Atman Desai, MD., 1Kimon Bekelis, MD., 1William Spire, MD., 1George Kakoulides, MD., 1Wesley Whitson, MD.
1Division of Neurological Surgery, Dartmouth-Hitchock Medical Center, Lebanon, N
Journal Club Article: Oddo M, Levine J, Mackenzie L, et al. Brain hypoxia is associated with short-term outcome after severe traumatic brain injury independently of intracranial hypertension and low cerebral perfusion pressure. Neurosurgery. 2011;69:1037-1045
Significance/context and importance of study
Traumatic brain injury is a leading cause of death and disability in the United States with significant economic costs to society. Timely recognition and prevention of the physiologic derangements that develop in the hours and days following injury have been shown to be important in the management of TBI. Current brain trauma foundation guidelines recommend monitoring and treatment of abnormal intracranial and cerebral perfusion pressure. Other variables, however, such as cerebral blood flow, brain oxygenation, and cerebral metabolism have been increasingly investigated and may be of use in monitoring and therapy of TBI. Brain hypoxia in particular has been associated with increased mortality and poor neurologic outcomes. Recently the use of monitors to measure the pressure of oxygen in brain tissue (PbtO2) has become widely available and effective in identifying cerebral hypoxia defined as a PbtO2 < 15mm Hg2.
Several studies have demonstrated a relationship between PbtO2 and neurologic outcome3,4,5. It remains unknown, however, whether brain hypoxia is simply a marker for injury severity or in fact has an effect on outcome that is independent of other variables such as ICP and CPP. There is little data on whether augmentation of brain oxygenation leads to reduced secondary neurologic injury and improved neurologic outcomes following trauma. In light of this the authors seek to investigate whether monitoring PbtO2 may be of use in the treatment of TBI.
Originality of the work
Brain oxygenation after TBI and its correlation with outcome have previously been investigated by several groups. This study builds on the current literature although its methodology and key conclusions replicate those from previous work. In this study and only one other, however, the authors establish that brain hypoxia was associated with poor outcomes independent of ICP.
Appropriateness of study design or experimental approach
In this retrospective study, intracranial pressure, brain temperature, and brain oxygenation were measured with a Licox monitor in 103 patients with a nonpenetrating traumatic brain injury. All subjects had a GCS of < 9 during admission and underwent at least 24 hours of continued monitoring with the Licox. Those with fixed and dilated pupils, < 24 hours of monitoring, PbtO2 of 0mmHg for > 3 hours, or met brain death requirements within 48 hours of monitoring were excluded. When possible the Licox probe was placed within normal white matter on the side of maximal pathology. All patients were treated according to standard treatment protocols for TBI and intracranial hypertension if applicable. In addition, however, the patients enrolled in the study also received directed therapy to maintain PbtO2 > 20mmHg that included optimization of ICP, elevation of CPP with pressors, or treatment of systemic and metabolic etiologies. If the initial interventions were unsuccessful patients were transfused to maintain a hemoglobin > 10mg/dL, and if brain hypoxia was refractory to these interventions a decompressive craniectomy was considered.
The variables measured included age, admission GCS, APACHE II score, Marshall CT classification, PbtO2, ICP, CPP, and MAP. Thresholds for intervention were defined as PbtO2 < 15mm Hg, ICP > 20mm Hg, and CPP < 60mmHg. Single abnormal episodes for each physiologic parameter were recorded and linear interpolation was used to define the total duration of derangement for each variable. For each episode of brain hypoxia the suspected etiology was recorded as either related to ICP elevation, decrease in CPP, reduction in MAP, systemic hypoxia, hyperthermia or cooling, Hgb < 9 mg/dL, or unknown cause. Outcomes were assessed by a neurointensivist and neurocritical care nurse independently within 30-days of injury and dichotomized as favorable (GOS 4 or 5) or unfavorable (GOS < 4). The relationship of and factors associated with outcome with respect to PbtO2, ICP, and CPP were then analyzed. A fundamental limitation of this design, however, is the simultaneous assessment of brain hypoxia and outcomes while concurrently using brain hypoxia as a marker to direct therapy, potentially confounding results. In addition there was no control group receiving standard therapy for elevated ICP or low CPP for comparison of outcomes. Therefore one cannot exclude that based upon this data poor outcomes following brain hypoxia were related to the interventions implemented to obtain a PbtO2 of > 20 mmHg rather than the presence of hypoxia alone.
Adequacy of experimental techniques
The authors investigate the role of brain hypoxia in traumatic brain injury using their series of consecutive patients from a single institution receiving standardized therapy and assessment of outcomes. There are, however, several limitations inherent to studies of traumatic brain injury and in this analysis. Patients with head injury constitute a heterogeneous population with wide variations in demographics, mechanisms of injury, prehospital care, injury type, injury location, and comorbidities. Without subgroup analysis these are all potentially confounding variables. In this study the authors used the APACHE II score to characterize patients. This scoring system is extensive but arguably less informative than other scoring systems used in trauma patients such as the Injury Severity Score. A further potential limitation in studies examining the effect of intracranial pressure and brain oxygenation in TBI is the tendency to “drift” in these systems. In this study the Licox monitor was allowed to equilibrate for 3 hours, but unfortunately there is little discussion of validation of the monitors and accuracy of the measurements.
Neurologic outcomes were reported as being assessed within thirty days of monitoring. In addition to this being a relatively short follow-up period there was no stratification of when within the 30 days the assessment was performed. This may potentially confound the findings if there was significant variation of when this evaluation occurred, as is likely. Finally for those patients that underwent decompressive craniotomy little further information is given. It would be of value for example to know the number of these patients and timing of surgery.
Soundness of conclusions and interpretation
The authors conclude that brain hypoxia defined by a PbtO2 of less than 15 mmHg is a frequent occurrence in TBI, is associated with raised ICP and reduced CPP, and is an independent predictor of poor short-term outcomes. Based on their data this is a reasonable conclusion. Their data shows that brain hypoxia may occur despite a normal ICP or adequate CPP. From this they also conclude that PbtO2 may be an “important physiological target after severe TBI to complement ICP and CPP and to help guide the management of TBI patients”. This statement implies causality between hypoxia and outcome. Given the multiple confounding variables at play as described above, such a causal relationship cannot be concluded.
Relevance of Discussion
The authors discuss potential future uses of brain oxygenation as a variable in the management of traumatic brain injury. They discuss that brain hypoxia is associated with poor neurologic outcomes, but in individuals in which hypoxia persists despite a normal ICP and CPP, it may be an important parameter to identify patients at risk of further neurologic injury and augment therapy. These are important findings in this study and are well discussed. Figure 2, however, outlines the varying etiologies of brain hypoxia in this group. Many of these etiologies may be potentially identified and corrected in the absence of PbtO2 monitoring. The authors unfortunately do not discuss this important issue.
Clarity of writing, strength, and organization of the paper
This is a well written manuscript. The criteria for patient inclusion, placement of Licox monitor, and treatment protocols were well described and clear. Additionally, the results and discussion points were presented in a logical order and progression with only relevant details offered.
Economy of words
The authors concisely describe the use of monitors for brain oxygenation in traumatic brain injury. They effectively incorporate discussion of prior studies into their manuscript offering insights into the impetus for their work as well as the unique elements of their results and implications for future studies. They offer a cogent explanation of the results and statistical methods that incorporates a number of variables that is concise but well developed.
Relevance, accuracy and completeness of bibliography
The authors reference studies from both the neurosurgical and critical care literature detailing the physiologic variables measured in the setting of traumatic brain injury such as ICP, CPP, and PbtO2 and their influence on outcome. Included within the references are the brain trauma foundation guidelines for management of TBI. The studies listed cross a decade and include recent reports of therapy directed towards maintaining a specific PbtO2.
Number and quality of figures, tables and illustrations
The table and figures were well constructed and each is an effective means of displaying their data. The content was clear and readily interpretable. Scoring systems, however, are not clarified and it would have been beneficial to have a table with features of each system. In addition, an illustration of the Licox monitor would have been of value to the readership.
Future/next steps the paper logically leads to
This study is written on an important topic and its findings lend itself to several future investigations. The authors clearly demonstrate that brain hypoxia is a frequent occurrence in traumatic brain injury, is related to poor outcomes independently of ICP and CPP, and may occur in the setting of a normal ICP and CPP. To further establish PbtO2 as an important physiologic variable in the management of traumatic brain injury, a prospective trial with a control group in which brain oxygenation was reported but not used to direct therapy would be valuable. These outcomes may be further investigated using subgroup analysis according to type of brain injury and by extending follow-up periods. Finally, as health care costs become increasingly scrutinized the cost-effectiveness of this intervention may be addressed.
Robertson CS. Critical are management of traumatic brain injury. In: Winn HR, ed. Youmans Neurological Surgery.5th ed. Philadelphia, PA: Saunders; 2004: 5103-5144
Braton SL, Chestnut RM, Ghajar J, et al. Guidelines for the management of severe traumatic brain injury, X: brain oxygen monitoring and thresholds. J Neurotrauma. 2007;24(suppl 1):S65-70
Valadka AB, Gopinath SP, Contant CF, Uzura M, Robertson CS. Relationship of brain tissue PbtO2 to outcome after severe head injury. Crit Care Med. 1998;26(9):1576-1581
Bardt TF, Unterberg AW, Hartl R, et al. Monitoring of brain tissue PO2 in traumatic brain injury: effect of cerebral hypoxia on outcome. Acta Neurochir Suppl 1998; 71: 153-156
Van den bring WA, van santbrin H, Steyerberg EW, et al. Brain oxygen tension in severe head injury. Neurosurgery 2000; 46:868-878.