Objective: To investigate clinical relevance and prognostic value of brain tissue oxygen response (TOR: response of brain tissue pO(2) to changes in arterial pO(2)) in traumatic brain injury (TBI).
Patients and methods: In a prospective cohort study TOR was investigated in 41 patients with severe TBI (Glasgow Coma Score < or =8) in whom continuous monitoring of brain tissue oxygen pressure (PbrO(2)) was performed.TOR was investigated each day over a five day period for 15 minutes by increasing FiO(2) on the ventilator setting. FiO(2) was increased directly from baseline to 1.0 for a period of 15 minutes under stable conditions (145 tests). In 34 patients the effect of decreasing PaCO(2) was evaluated on TOR by performing the same test after increasing inspiratory minute volume on the ventilator setting to 20% above baseline. Arterial blood gas analysis was performed before and after changing ventilator settings. Multimodality monitoring, including PbrO(2) was performed in all patients. Outcome at six months was evaluated according to the Glasgow Outcome Scale. For statistical analysis the Mann-whitney U-test was used for ordinally distributed variables, and the Chi-square test for categorical variables. Predictive value of TOR was analyzed in a multivariable model.
Results: 145 tests were available for analysis. Baseline PbrO(2) varied from 4.0 to 50 mmHg at PaO(2) values of 73-237 mmHg. At FiO(2) settings of 1.0, PbrO(2) varied from 9.1-200 mmHg and PaO(2) from 196-499 mmHg. Three distinct patterns of response were noted: response type A is characterized by a sharp increase in PbrO(2), reaching a plateau within several minutes; type B by the absence of a plateau, and type C by a short plateau phase followed by a subsequent further increase in PbrO(2). Patterns characterized by a stable plateau (type A), considered indicative of intact regulatory mechanisms, were seen more frequently from 48 hours after injury on. If present within the first 24 hours after injury such a response was related to more favorable outcome (p = 0.06). Mean TOR of all tests was 0.73 +/- 0.59 with an median TOR of 0.58. Patients with an unfavourable outcome had a higher TOR (1.03 +/- 0.60) during the first 24 hours, compared to patients with a favorable outcome (0.61 +/- 0.51; p = 0.02). Multiple logistic regression analysis supported the independent predictive value of tissue oxygen response for unfavorable outcome (odds ratio 4.8). During increased hyperventilation, mean TOR decreased substantially from 0.75 +/- 0.54 to 0.65 +/- 0.45 (p = 0.06; Wilcoxon test). Within the first 24 hours after injury a decrease in TOR following hyperventilation was significantly related to poorer outcome (p = 0.01).
Conclusions: Evaluation of TOR affords insight in (disturbances in) oxygen regulation after traumatic brain injury, is of prognostic value and may aid in identifying patients at (increased) risk for ischemia.