RESUMEN
BACKGROUND: In patients with severe traumatic brain injury (TBI), early CT perfusion (CTP) provides additional information beyond the non-contrast CT (NCCT) and may alter clinical management. We hypothesized that this information may prognosticate functional outcome. METHODS: Five-year prospective observational study was performed in a level-1 trauma centre on consecutive severe TBI patients. CTP (obtained in conjunction with first routine NCCT) was interpreted as: abnormal, area of altered perfusion more extensive than on NCCT, and the presence of ischaemia. Six months Glasgow Outcome Scale-Extended of four or less was considered an unfavourable outcome. Logistic regression analysis of CTP findings and core variables [preintubation Glasgow Coma Scale (GCS), Rotterdam score, base deficit, age] was conducted using Bayesian model averaging to identify the best predicting model for unfavourable outcome. RESULTS: Fifty patients were investigated with CTP (one excluded for the absence of TBI) [male: 80%, median age: 35 (23-55), prehospital intubation: 7 (14.2%); median GCS: 5 (3-7); median injury severity score: 29 (20-36); median head and neck abbreviated injury scale: 4 (4-5); median days in ICU: 10 (5-15)]. Thirty (50.8%) patients had an unfavourable outcome. GCS was a moderate predictor of unfavourable outcome (AUC = 0.74), while CTP variables showed greater predictive ability (AUC for abnormal CTP = 0.92; AUC for area of altered perfusion more extensive than NCCT = 0.83; AUC for the presence of ischaemia = 0.81). CONCLUSION: Following severe TBI, CTP performed at the time of the first follow-up NCCT, is a non-invasive and extremely valuable tool for early outcome prediction. The potential impact on management and its cost effectiveness deserves to be evaluated in large-scale studies. LEVEL OF EVIDENCE III: Prospective study.