Antiseizure medications in critical care: an update

Rapid response electroencephalography decreases time to seizure diagnosis in pediatric acute care patients

OBJECTIVES

Pediatric status epilepticus (SE) carries a high risk of morbidity and mortality and can result in neurologic injury. Establishing seizure activity on conventional EEG (cEEG) is essential but can delay treatment of seizures due to technician limitations. Rapid response EEG (rrEEG) device Ceribell and its Brain Stethoscope function can be used and interpreted rapidly by bedside providers with minimal training. This retrospective pilot study examines the impact of rrEEG introduction at a quaternary care children's hospital on time to definitive diagnosis and treatment, as well as the accuracy of the Brain Stethoscope.

METHODS

This was a single center retrospective observational cohort study that analyzed data from patients 2-18 years old who presented with concerns for SE. For rrEEG patients, the bedside physician used the Brain Stethoscope at four discrete points. TDEA (time to diagnosis of electrographic activity) and setup time were recorded and compared using Welch's T-test. Diagnostic specificity and sensitivity for SE using the Brain Stethoscope were calculated against the epileptologist's assessment.

RESULTS

Data were collected from 30 pediatric patients, 15 on each EEG modality. RrEEG decreased the average TDEA (132 min vs. 22 min, p < 0.001) and setup time (22 min vs. 9 min, p < 0.001), compared to the cEEG. Bedside physicians diagnosed electrographic activity using the Brain Stethoscope with 100% sensitivity (95% CI 63%-100%) and 92% specificity (95% CI 81%-97%). RrEEG ruled out seizures in 11 patients and changed clinical decision-making in five patients.

SIGNIFICANCE

RrEEG allowed for earlier diagnosis of brain electrographic activity in pediatric patients when compared to cEEG. The bedside provider was able to initiate EEG monitoring, successfully diagnose patients using the Brain Stethoscope, and decrease delays associated with technician availability. This promising rrEEG technology can facilitate faster assessment of SE in pediatric acute care settings, potentially reducing ongoing neurologic injury.

PLAIN LANGUAGE SUMMARY

Prolonged seizures in pediatric patients can cause death. Children can have seizures that are happening in the brain, but cannot be seen physically. They can be diagnosed by a machine that records the brain's electrical activity and the data can be interpreted by a specialized neurologist, but the process of getting the machine connected to the patient and getting the data often takes hours and delays diagnosis. This study evaluated a new machine called the Ceribell® and determined that it facilitated a much faster diagnosis than the conventional machine in children and helped the bedside physician interpret the data quickly.

Using artificial intelligence to optimize anti-seizure treatment and EEG-guided decisions in severe brain injury

Electroencephalography (EEG) is invaluable in the management of acute neurological emergencies. Characteristic EEG changes have been identified in diverse neurologic conditions including stroke, trauma, and anoxia, and the increased utilization of continuous EEG (cEEG) has identified potentially harmful activity even in patients without overt clinical signs or neurologic diagnoses. Manual annotation by expert neurophysiologists is a major resource limitation in investigating the prognostic and therapeutic implications of these EEG patterns and in expanding EEG use to a broader set of patients who are likely to benefit. Artificial intelligence (AI) has already demonstrated clinical success in guiding cEEG allocation for patients at risk for seizures, and its potential uses in neurocritical care are expanding alongside improvements in AI itself. We review both current clinical uses of AI for EEG-guided management as well as ongoing research directions in automated seizure and ischemia detection, neurologic prognostication, and guidance of medical and surgical treatment.

Therapeutic drug monitoring of phenytoin and valproic acid in critically ill patients at Windhoek Central Hospital, Namibia

Background

Phenytoin and valproic acid, anticonvulsants, have a low therapeutic index and are highly plasma protein bound, mainly to albumin. Hypoalbuminaemia is common in critically ill patients and increases the unbound drug concentration. Thus, monitoring unbound rather than total plasma drug concentrations is recommended to optimise the dosing of these drugs.

Objective

This retrospective study determined unbound plasma concentrations of phenytoin and valproic as a more accurate value of drug levels than total plasma drug concentrations.

Methods

Total plasma concentrations were retrieved for 56 Intensive Care Unit patients for phenytoin and 93 for valproic acid. Total drug concentrations were converted to unbound concentrations using a serum albumin-based normalising equation.

Results

Total phenytoin plasma concentration was below (41.1% of patients), within (46.4%) or above (12.5%) the therapeutic range (10 μg/mL – 20 μg/mL). However, the predicted unbound plasma concentration of phenytoin was above the therapeutic range (1 μg/mL – 2 μg/mL) in the majority of patients (57.1%). For valproic acid, the total plasma concentration of most patients (87.1%) was below the therapeutic range (50 μg/mL – 100 μg/mL); among remaining patients (12.9%), it was within the therapeutic range. In the majority of patients (91.4%), the predicted unbound plasma concentration of valproic acid was between 2.5 μg/mL and 20 μg/mL.

Conclusion

The usefulness of monitoring the total phenytoin or valproic acid levels for dose optimisation is limited as it is an inaccurate indicator of a patient’s drug therapeutic state. Thus, the unbound plasma drug concentrations should be quantified experimentally or predicted in resource-limited settings.