Background: Utilization of brain tissue oxygenation (pBtO(2)) is an important but controversial variable in the treatment of traumatic brain injury. We hypothesize that pBtO(2) values over the first 72 hours of monitoring are predictive of mortality.
Methods: Consecutive, adult patients with severe traumatic brain injury and pBtO(2) monitors were retrospectively identified. Time-indexed measurements of pBtO(2), cerebral perfusion pressure (CPP), and intracranial pressure (ICP) were collected, and average values over 4-hour blocks were determined. Patients were stratified according to survival, and repeated measures analysis of variance was used to compare pBtO(2), CPP, and ICP. The pBtO(2) threshold most predictive for survival was determined.
Results: There were 8,759 time-indexed data points in 32 patients. The mean age was 39 years ± 16.5 years, injury severity score was 27.7 ± 10.7, and Glasgow Coma Scale score was 6.6 ± 3.4. Survival was 68%. Survivors consistently demonstrated higher pBtO(2) values compared with nonsurvivors including age as a covariate (F = 12.898, p < 0.001). Individual pBtO(2) was higher at the time points 8 hours, 12 hours, 20 hours to 44 hours, 52 hours to 60 hours, and 72 hours of monitoring (p < 0.05). There was no difference in ICP (F = 1.690, p = 0.204) and CPP (F = 0.764, p = 0.389) values between survivors and nonsurvivors including age as a covariate. Classification and regression tree analysis identified 29 mm Hg as the threshold at which pBtO(2) was most predictive for mortality.
Conclusion: The first 72 hours of pBtO(2) neurologic monitoring predicts mortality. When the pBtO(2) monitor remains below 29 mm Hg in the first 72 hours of monitoring, mortality is increased. This study challenges the brain oxygenation threshold of 20 mm Hg that has been used conventionally and delineates a time for monitoring pBtO(2) that is predictive of outcome.
Level of evidence: III, prognostic study.