Optical monitoring of cerebral perfusion and metabolism in adults during cardiac surgery with cardiopulmonary bypass

Abstract

During cardiac surgery with cardiopulmonary bypass (CPB), adequate maintenance of cerebral blood flow (CBF) is vital in preventing postoperative neurological injury - i.e. stroke, delirium, cognitive impairment. Reductions in CBF large enough to impact cerebral energy metabolism can lead to tissue damage and subsequent brain injury. Current methods for neuromonitoring during surgery are limited. This study presents the clinical translation of a hybrid optical neuromonitor for continuous intraoperative monitoring of cerebral perfusion and metabolism in ten patients undergoing non-emergent cardiac surgery with non-pulsatile CPB. The optical system combines broadband near-infrared spectroscopy (B-NIRS) to measure changes in the oxidation state of cytochrome c oxidase (oxCCO) - a direct marker of cellular energy metabolism - and diffuse correlation spectroscopy (DCS) to provide an index of cerebral blood flow (CBFi). As the heart was arrested and the CPB-pump started, increases in CBFi (88.5 ± 125.7%) and significant decreases in oxCCO (-0.5 ± 0.2 µM) were observed; no changes were noted during transitions off CPB. Fifteen hypoperfusion events, defined as large and sustained reductions in CPB-pump flow rate, were identified across all patients and resulted in significant decreases in perfusion and metabolism when mean arterial pressure dropped to 30 mmHg or below. The maximum reduction in cerebral blood flow preceded the corresponding metabolic reduction by 18.2 ± 15.0 s. Optical neuromonitoring provides a safe and non-invasive approach for assessing intraoperative perfusion and metabolism and has potential in guiding patient management to prevent adverse clinical outcomes.

Conflict of interest statement

The authors declare no conflicts of interest.

© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement.

Figures

Fig. 1.

A) Optical fiber probes and 3D-printed probe holder (5 × 2 × 1 cm) B) Schematic of the probe holder showing near-infrared spectroscopy (NIRS, 3-cm SDD) and diffuse correlation spectroscopy (DCS, 2-cm SDD) emission probes with a common detection probe.

Fig. 2.

Cerebral hemodynamics during transition A) on cardiopulmonary bypass (CPB) and B) off CPB. Figure shows a continuous measure of CPB-pump flowrate (FRCPB), mean arterial pressure (MAP), change in the cerebral blood flow index (ΔCBFi), cerebral tissue saturation (StO2), and changes in the oxidation state of cytochrome c oxidase (ΔoxCCO). ΔCBFi scales are adjusted across plots for data visualization.

Fig. 3.

Box plots showing average change across subjects during transitions onto the CPB pump (ON CPB) and off of the pump (OFF CPB). Absolute StO2 values are shown for periods immediately before and after each transition. Data were averaged for 5-minute durations immediately surrounding the transition period; + indicates statistical outliers and * indicates a statistically significant difference across the transition. Average transition time on CPB was 202 ± 111 s, and off CPB was 230 ± 160 s.

Fig. 4.

Pulsatile waveforms observed in cerebral microcirculation measured at 20 Hz A) pre and B) post aortic valve replacement.

Fig. 5.

Cerebral hemodynamics in subject 4 during A) a stable period and B) successive FRCPB-driven hypoperfusion events. Each graph includes a continuous measure of CPB-pump flow rate (FRCCPB), mean arterial pressure (MAP), change in cerebral blood flow (ΔCBFi), cerebral tissue saturation (StO2), and changes in the oxidation state of cytochrome c oxidase (ΔoxCCO). Data were temporally averaged (7-s window) for visualization.

Fig. 6.

Correlation boxplots of ΔCBF, StO2, and ΔoxCCO as a function of MAP; + indicates statistical outliers and * indicates a statistically significant difference from baseline values.

Fig. 7.

Correlation plot showing the relationship between the maximum reductions in ΔCBFi and ΔoxCCO. On average, the ΔoxCCO nadir occurred 18.2 ± 15.0 s after the corresponding ΔCBFi nadir.