Theta-Alpha Variability on EEG Is Associated With Acute Brain Injury in Children and Young Adults With Liver Failure.

Background and objectives: Patients with liver failure experience long hospitalizations and acute neurologic complications. Encephalopathy limits the bedside examination, rendering presenting signs of acute brain injury less specific. Seizures are common. Brain MRI is the gold standard for detecting acute brain injury, but intensive medical needs may preclude immediate transfer for imaging. EEG is a bedside test applied in cases of seizure or encephalopathy. We hypothesized that EEG variables can predict MRI signs of acute brain injury in children hospitalized with liver failure.

Methods: In this retrospective cohort analysis, records were collected for patients admitted to a MedStar hospital between 2014 and 2022 with ICD-9/10 codes related to liver failure, who underwent brain MRI and EEG testing during the same admission. Exclusion criteria included age older than 24 years and >7 days elapsing between EEG and MRI testing. Clinical data of interest from chart review, reinterpreted MRI scans, reinterpreted EEG tracings, and quantitative EEG variables were compiled into a database. Quantitative EEG variables were processed using MNE-Python.

Results: Of 746 records screened, 52 patients met inclusion criteria comprising 63 EEG-MRI pairs. Univariate analysis of all quantitative EEG variables of interest showed depressed theta-alpha variability (TAV) when paired MRI involved abnormal restricted diffusivity in cortical or deep gray matter structures (TAV 0.705, SD 0.310; p < 0.001) compared with MRI with no abnormal restricted diffusivity (TAV 0.895, SD 0.095). Multilinear regression analysis including potential confounders demonstrated independent association of depressed TAV with this MRI finding, with an odds ratio of 4.0317 (95% CI 1.3868-11.7165; AUROC 0.83).

Discussion: Depressed TAV on EEG is associated with increased odds of abnormal restricted diffusivity in gray matter on brain MRI in children and young adults hospitalized with liver failure. This MRI finding is seen in scenarios where changes to medical management are time-sensitive (i.e., acute stroke and PRES) or where prognostic discussion may be influenced by MRI findings (hypoxic-ischemic injury). TAV thus has a potential role as an automated, bedside decision support tool for clinicians deciding on the urgency of brain MRI in critically ill patients.