The Use of Different Components of Brain Oxygenation for the Assessment of Cerebral Haemodynamics: A Prospective Observational Study on COVID-19 Patients

Abstract

Introduction: The role of near-infrared spectroscopy (NIRS) for the evaluation of cerebral haemodynamics is gaining increasing popularity because of its noninvasive nature. The aim of this study was to evaluate the role of the integral components of regional cerebral oxygenation (rSO₂) measured by NIRS [i.e., arterial-oxyhemoglobin (O₂Hbi) and venous-deoxyhemoglobin (HHbi)-components], as indirect surrogates of cerebral blood flow (CBF) in a cohort of critically ill patients with coronavirus disease 2019 (COVID-19). We compared these findings to the gold standard technique for noninvasive CBF assessment, Transcranial Doppler (TCD). Methods: Mechanically ventilated patients with COVID-19 admitted to the Intensive Care Unit (ICU) of Policlinico San Martino Hospital, Genova, Italy, who underwent multimodal neuromonitoring (including NIRS and TCD), were included. rSO₂ and its components [relative changes in O₂Hbi, HHbi, and total haemoglobin (cHbi)] were compared with TCD (cerebral blood flow velocity, CBFV). Changes (Δ) in CBFV and rSO₂, ΔO₂Hbi, ΔHHbi, and ΔcHbi after systemic arterial blood pressure (MAP) modifications induced by different manoeuvres (e.g., rescue therapies and haemodynamic manipulation) were assessed using mixed-effect linear regression analysis and repeated measures correlation coefficients. All values were normalised as percentage changes from the baseline (Δ%). Results: One hundred and four measurements from 25 patients were included. Significant effects of Δ%MAP on Δ%CBF were observed after rescue manoeuvres for CBFV, ΔcHbi, and ΔO₂Hbi. The highest correlation was found between ΔCBFV and ΔΔO₂Hbi (R = 0.88, p < 0.0001), and the poorest between ΔCBFV and ΔΔHHbi (R = 0.34, p = 0.002). Conclusions: ΔO₂Hbi had the highest accuracy to assess CBF changes, reflecting its role as the main component for vasomotor response after changes in MAP. The use of indexes derived from the different components of rSO₂ can be useful for the bedside evaluation of cerebral haemodynamics in mechanically ventilated patients with COVID-19.

Keywords: NIRS (near infrared reflectance spectroscopy); autoregulation dysfunction; brain injury; cerebral oxygenation; intensive care.

Figure 1

Scatter plots showing the linear association and correlation (R) between changes in cerebral blood flow velocity (ΔCBFV) vs. total cerebral oxygenation (ΔrSO₂ - plot A) (A), sum of arterial and venous components of cerebral oxygenation (ΔΔcHbi - plot B) (B), venous component (ΔΔHHbi - plot C) (C), and arterial component (ΔΔO₂Hbi - plot D) (D). Repeated measurements for each patient are plotted in the same colour pattern. Linear regression lines are correspondent to repeated measurements within patients.

Figure 2

Individual data of cerebral blood flow relative response changes to MAP from T0 to T1 and linear mixed-effects models marginal mean values of CBF for cerebral blood flow velocity (ΔCBFV–plot A) (A), the arterial component of cerebral oxygenation (ΔO₂Hbi–plot B) (B), and arterial plus venous components (ΔcHbi–plot C) (C).