Validation of non-invasive cerebrovascular pressure reactivity and pulse amplitude reactivity indices in traumatic brain injury

Background Two transcranial Doppler (TCD) estimators of cerebral arterial blood volume (CaBV) coexist: continuous outflow of arterial blood outside the cranium through a low-pulsatile venous system (continuous flow forward, CFF) and pulsatile outflow through regulating arterioles (pulsatile flow forward, PFF). We calculated non-invasive equivalents of the pressure reactivity index (PRx) and the pulse amplitude index PAx with slow waves of mean CaBV and its pulse amplitude. Methods About 273 individual TBI patients were retrospectively reviewed. PRx is the correlation coefficient between 30 samples of 10-second averages of ICP and mean ABP. PAx is the correlation coefficient between 30 samples of 10-second averages of the amplitude of ICP (AMP, derived from Fourier analysis of the raw full waveform ICP tracing) and mean ABP. nPRx is calculated with CaBV instead of ICP and nPAx with the pulse amplitude of CaBV instead of AMP (calculated using both the CFF and PFF models). All reactivity indices were additionally compared with Glasgow Outcome Score (GOS) to verify potential outcome-predictive strength. Results When correlated, slow waves of ICP demonstrated good coherence between slow waves in CaBV (>0.75); slow waves of AMP showed good coherence with slow waves of the pulse amplitude of CaBV (>0.67) in both the CFF and PFF models. nPRx was moderately correlated with PRx (R = 0.42 for CFF and R = 0.38 for PFF; p < 0.0001). nPAx correlated with PAx with slightly better strength (R = 0.56 for CFF and R = 0.41 for PFF; p < 0.0001). nPAx_CFF showed the strongest association with outcomes. Conclusions Non-invasive estimators (nPRx and nPAx) are associated with their invasive counterparts and can provide meaningful associations with outcome after TBI. The CFF model is slightly superior to the PFF model.