Abstract
The contemporary monitoring of cerebrovascular reactivity (CVR) relies on invasive intracranial pressure (ICP) monitoring which limits its application. Interest is shifting towards near-infrared spectroscopic regional cerebral oxygen saturation (rSO(2))-based indices of CVR which are less invasive and have improved spatial resolution. This study aims to examine and model the relationship between ICP and rSO(2)-based indices of CVR. Through a retrospective cohort study of prospectively collected physiologic data in moderate to severe traumatic brain injury (TBI) patients, linear mixed effects modeling techniques, augmented with time-series analysis, were utilized to evaluate the ability of rSO(2)-based indices of CVR to model ICP-based indices. It was found that rSO(2)-based indices of CVR had a statistically significant linear relationship with ICP-based indices, even when the hierarchical and autocorrelative nature of the data was accounted for. This strengthens the body of literature indicating the validity of rSO(2)-based indices of CVR and potential greatly expands the scope of CVR monitoring.