Optimal cerebrovascular reactivity thresholds for the determination of individualized intracranial pressure thresholds in traumatic brain injury: a CAHR-TBI cohort study

Abstract It has been demonstrated that patient-specific intracranial pressure (ICP) thresholds are possible to derive using the function intersectionality between ICP and cerebrovascular reactivity (CVR). Such individualized ICP (iICP) thresholds represent a potential personalized medicine approach to neurocritical care management. However, it is currently unknown how various CVR thresholds compare in regard to deriving iICP. Here we attempt to identify the CVR thresholds that are best suited for iICP derivation. Leveraging 365 patient data sets from the CAnadian High-Resolution TBI (CAHR-TBI) Research Collaborative, iICP was derived using three ICP-based CVR indices: the pressure reactivity index (PRx); the pulse amplitude index (PAx); and the RAC index, and thresholds ranging from − 1 to + 1, in 0.05 increments. Patients were dichotomized based on 6-month outcome scores into Alive vs. Dead and Favorable vs. Unfavorable outcome. 2 × 2 tables were created for each threshold, grouping patients by outcome and whether their mean ICP was greater or less than their calculated iICP. Chi-squares were calculated for each table and subsequently plotted. The thresholds that produced the largest Chi-square values were identified as those able to derive the iICP with the greatest ability to predict outcomes. Next, Spearman rank correlation testing was used to evaluate associations between iICP, for each threshold, and measures of cerebral physiologic insult burden. With consideration of yield data, ability to predict outcome, and association with cerebral physiologic insult burden, a threshold of + 0.05 was identified for PRx. No optimal threshold could be identified for PAx or RAC.