OBJECTIVES: The three centers in this study have different policies regarding cerebral perfusion pressure targets and use of vasopressors in traumatic brain injury patients. The aim was to determine if the different policies affected the estimation of cerebral perfusion pressure which optimizes the strength of cerebral autoregulation, termed "optimal cerebral perfusion pressure." DESIGN: Retrospective analysis of prospectively collected data. SETTING: Three neurocritical care units at university hospitals in Cambridge, United Kingdom, Groningen, the Netherlands, and Uppsala, Sweden. PATIENTS: A total of 104 traumatic brain injury patients were included: 35 each from Cambridge and Groningen, and 34 from Uppsala. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: In Groningen, the cerebral perfusion pressure target was greater than or equal to 50 and less than 70 mm Hg, in Uppsala greater than or equal to 60, and in Cambridge greater than or equal to 60 or preferably greater than or equal to 70. Despite protocol differences, median cerebral perfusion pressure for each center was above 70 mm Hg. Optimal cerebral perfusion pressure was calculated as previously published and implemented in the Intensive Care Monitoring+ software by the Cambridge group, now replicated in the Odin software in Uppsala. Periods with cerebral perfusion pressure above and below optimal cerebral perfusion pressure were analyzed, as were absolute difference between cerebral perfusion pressure and optimal cerebral perfusion pressure and percentage of monitoring time with a valid optimal cerebral perfusion pressure. Uppsala had the highest cerebral perfusion pressure/optimal cerebral perfusion pressure difference. Uppsala patients were older than the other centers, and age is positively correlated with cerebral perfusion pressure/optimal cerebral perfusion pressure difference. Optimal cerebral perfusion pressure was significantly lower in Groningen than in Cambridge. There were no significant differences in percentage of monitoring time with valid optimal cerebral perfusion pressure. Summary optimal cerebral perfusion pressure curves were generated for the combined patient data for each center. These summary curves could be generated for Groningen and Cambridge, but not Uppsala. The older age of the Uppsala patient cohort may explain the absence of a summary curve. CONCLUSIONS: Differences in optimal cerebral perfusion pressure calculation were found between centers due to demographics (age) and treatment (cerebral perfusion pressure targets). These factors should be considered in the design of trials to determine the efficacy of autoregulation-guided treatment.