Abstract 272: Identifying Vascular Determinants Of Optimal Cerebral Perfusion Pressure In Traumatic Brain Injury Using A Continuous Measurement Algorithm.

Introduction: Individualized therapy targeting certain cerebral perfusion pressure (CPP) values in patients with traumatic brain injury may improve outcomes. One method to obtain “optimal” CPP (CPPopt) uses a measurement algorithm with continuous arterial blood pressure (ABP) and intracranial pressure (ICP) to estimate CPPopt where the pressure reactivity index is lowest. Yet, there is a knowledge gap regarding the cerebrovascular conditions that determine the algorithm’s success and validity. Here, we use a validated brain vasculature model to assess the algorithm’s performance at varying simulated vascular conditions. Methods: We used a convenience sample of 12 patient records with 4 hours of high-resolution ABP from the MIMIC III database. Each ABP set generated a synthetic ICP from our model. The algorithm then performed 38,815 CPPopt estimates across 12 patient records based on given input parameters. We examined “G”, an autoregulation parameter representing the sensitivity of ICP response to ABP change, and “Tau”, a time parameter representing the delay in ICP response to ABP change. We varied G and Tau independently and measured changes in frequency of algorithm success in finding CPPopt (CPPopt coverage), pressure reactivity index (PRx), and CPPopt itself. Results: 1.As G increases, CPPopt coverage, and mean minimum PRx decrease.2.As Tau increases, CPPopt coverage, and mean minimum PRx increase.3.CPPopt coverage had a median success rate of 27.111% and IQR of 29.118% across G and median success rate of 45.166% and IQR of 35.3362% across Tau for 38,815 algorithm estimates across the 12 patient records. Conclusion: The CPPopt algorithm may have a higher likelihood of finding CPPopt in pathological conditions where G is expected to be low and Tau to be large. The low and variable algorithm success rate poses a challenge for clinical use. More work is needed to understand the algorithm’s success and validity to determine its clinical viability.