OBJECTIVE: To compare brain perfusion-computed tomography (CT) results with invasive cerebral perfusion pressure (CPP) monitoring in severe head trauma patients.
DESIGN: Prospective cohort study.
SETTING: Emergency room and surgical intensive care unit of our hospital.
PATIENTS: Sixty-one severe head trauma patients.
INTERVENTIONS: We prospectively collected 103 perfusion-CT examinations with simultaneous measurement of mean arterial pressure and intracranial pressure, affording calculation of CPP. The statistical relationship between perfusion-CT results and the corresponding CPP values was evaluated using Wilcoxon (Mann-Whitney) and generalized F-tests. The functional outcome of the 61 patients was evaluated 3 months after trauma on the basis of the Glasgow Outcome Scale score and compared between groups using Fisher's exact tests.
MEASUREMENTS AND MAIN RESULTS: Perfusion-CT enabled us to distinguish between two groups of patients. Within each group, a significant correlation (p <.001) between the CPP values and the corresponding perfusion-CT results was demonstrated. There was also a significant correlation (p <.001) between the CPP values and the extent of the abnormal perfusion-CT areas (R up to.817). The first group was characterized by a weak dependence of perfusion-CT results on the corresponding CPP values (low slope) and the second group by a strong dependence (steep slope). These groups were interpreted as having preserved (or pseudo) and impaired cerebral vascular autoregulation, respectively. The functional outcome was better in the second group of patients.
CONCLUSIONS: Intermittent perfusion-CT measurements plus continuous CPP measurement provide more information than continuous CPP alone. Perfusion-CT gives unique information regarding regional heterogeneity of brain perfusion. It might allow clinicians to distinguish between patients with preserved auto-regulation (or pseudoautoregulation) and those with impaired autoregulation and could therefore guide interpretation of CPP measurements and therapy.