Estimating Pressure Reactivity Using Noninvasive Doppler-Based Systolic Flow Index.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Medical Statistics

Background:

• Cerebral autoregulation is crucial for managing blood flow in the brain, especially after traumatic brain injury (TBI).
• Estimating the pressure reactivity index (PRx) traditionally requires invasive intracranial pressure (ICP) monitoring.
• Noninvasive methods for assessing PRx are highly desirable in TBI management.

Purpose of the Study:

• To derive and validate models for estimating PRx using noninvasive transcranial Doppler (TCD) derived indices.
• To explore the relationship between TCD-based systolic (Sx_a) and mean (Mx_a) flow indices and PRx in TBI patients.
• To establish a link between cerebral blood flow dynamics and pulsatile cerebral blood volume using noninvasive measures.

Main Methods:

• Utilized a retrospective database of 347 TBI patients with ICP and TCD recordings.
• Applied autoregressive integrative moving average (ARIMA) and linear mixed-effects (LME) models to analyze PRx.
• Developed sequential LME models incorporating Sx_a and Mx_a to estimate PRx, assessing model adequacy and superiority using statistical criteria (AIC, BIC, LL, ANOVA).

Main Results:

• Identified an optimal ARIMA structure (2,0,2) for PRx in the TBI population.
• Two LME models demonstrated superior performance in estimating PRx: PRx ∼ Sx_a (R=0.794) and PRx ∼ Sx_a + Mx_a (R=0.814).
• Observed and estimated PRx showed acceptable agreement via Bland-Altman analysis, confirming model validity.

Conclusions:

• Noninvasive TCD-based indices (Sx_a, Mx_a) can effectively estimate PRx in TBI patients when analyzed with LME and ARIMA modeling.
• These findings establish a significant link between noninvasive measures of cerebral blood flow and pulsatile cerebral blood volume.
• Further research is warranted to validate these noninvasive PRx estimation models in clinical settings.