Neuromonitoring in the neonatal ECMO patient

Abstract

Utilization of extraocorporeal membrane oxygenation (ECMO) has become increasingly widespread as a bridging therapy for neonates with severe, reversible respiratory or cardiac diseases. While significant risks remain, due to advances in medical and surgical management, overall mortality has decreased. However, short and long-term neurological morbidity has remained high. Therefore, increasing attention has been focused on multimodal neuromonitoring to track and optimally, minimize or prevent intracranial injury. This review will explore the the indications, advantages, disadvantages, timing, frequency, duration, and any known correlation with neurodevelopmental outcomes of common types of neuromonitoring in the neonatal ECMO population. Investigational monitoring techniques such as NIRS will be briefly reviewed.

Keywords: Cerebral blood flow; ElectroEncephalography; Magnetic Resonance Imaging; Optical Technologies; head CT; head ultrasound; intracranial hemorrhage; stroke.

Figure 1

Patient and ECMO factors that may contribute to brain injury. Source: John Flibotte, MD. Illustrations by Eo Trueblood, © Stream Studios at the Children’s Hospital of Philadelphia.

Figure 2

Images of two different neonates demonstrating discrepancy between at the last normal HUS followed by subsequent MRI brain. A) MRI axial DWI sequence obtained 3 days post ECMO decannulation demonstrating a multifocal stroke. All HUS obtained during ECMO were normal. B) MRI GRE axial sequence, 5 days post ECMO decannulation. All HUS on ECMO were read as normal. However, MRI GRE sequence shows presence of hemosiderin staining (arrows) in the lateral horns of the ventricles suggesting prior intraventricular hemorrhage.

Figure 3

A left MCA stroke on the same neonate with different imaging modalities. A,B) Head CT on a term neonate. Note the decreased gray-white matter differentiation throughout the left hemisphere with increased hyperdensity in the left middle cerebral artery territory concerning for an ischemic stroke. C, D) In the same neonate, 2 days later, the MRI brain of DWI and ADC sequences, respectively, demonstrating concordance with prior HCT findings of an acute left MCA ischemic stroke.

Figure 4

Examples of white matter injury (WMI) in different neonates. A) MRI T1-weighted axial slice showing left periventricular injury as well as increased extra-axial space B) Another MRI T1-weighted axial slice demonstrating periventricular white matter injury in the right lateral ventricle, and separately, an old acute ischemic stroke (AIS) injury with associated encephalomalacia seen in the left parietal lobe.

Figure 5

Images all obtained from the same neonate demonstrating findings of intraparenchymal (IPH) and intraventricular hemorrhage (IVH) across different imaging modalities. A,B) HUS with LT sagittal and RT coronal views, respectively, demonstrating hyperechogenicity periventricularly and in parenchymal tissue. C,D) Axial HCT on same day as HUS showing hyperdensity in the same areas. E,F) MRI T1-weighted axial and coronal images, respectively, obtained 2 days later from HCT.

Figure 6

Top panel - continuous EEG neonatal montage showing evolution of an electrographic seizure (box) in a term neonate. Bottom panel – multiple seizures are captured in the 24 hour amplitude EEG as shown by the presence of the epochs (smaller box) This has been set up with the default mode of left (red) and right (blue) hemispheric detection. (Figure courtesy of Katherine Holland-Bouley, MD, PhD, with permission)

Figure 7

Top panel - In the same neonate as above, another seizure (box) arising from a different location that is missed by the default parameter of aEEG detection. Bottom panel – one can tailor the aEEG by adjusting the setting from left hemisphere to left mid-central (C3Cz, red) and right hemisphere to right mid-central (C4Cz, blue) seizures. Note the increase in the number of epochs (small box) detected by aEEG that were previously not seen in the same 24 hour aEEG montage. aEEG can miss a significant number of neonatal seizures if they are focal and outside of the conventional 1–2 channel set up. (Figure courtesy of Katherine Holland-Bouley, MD, PhD, with permission)