Factors Influencing the Likelihood of Seizure Identification on Quantitative Electroencephalography Displays in the Intensive Care Unit
CCCF ePoster library. Lalgudi Ganesan S.
Nov 7, 2018; 233414
Disclosure(s): No conflicts of interest
Dr. Saptharishi (Rishi) Lalgudi Ganesan
Dr. Saptharishi (Rishi) Lalgudi Ganesan
Login now to access Regular content available to all registered users.

You may also access this content "anytime, anywhere" with the Free MULTILEARNING App for iOS and Android
Abstract
Rate & Comment (0)

INTRODUCTION: Non-convulsive seizures are common in critically ill and, are associated with worse short- and long-term outcomes1-5. Non-convulsive seizures are detectable only by continuous electroencephalography (cEEG). During such cEEG monitoring, quantitative electroencephalography (QEEG) trends can help EEG experts and critical care providers identify seizures efficiently with acceptable sensitivity6-15. However, our previous works show that accuracy of seizure identification using QEEG is lower in some EEG recordings, even for EEG experts6,7

OBJECTIVES: To identify the patient, recording and seizure characteristics that decrease the likelihood of seizure identification on QEEG displays - amplitude-integrated EEG (aEEG) and color density spectral array (CDSA)

METHODS: Neurophysiologists with at least 5 years clinical experience independently marked all 'suspected' seizures in 27 continuous critical care EEG recordings, that had been transformed into aEEG and CDSA displays. During this testing, there was no access to raw EEG or clinical data. Seizures confirmed by a different board-certified neurophysiologist on review of the raw EEG constituted the comparison standard. We analyzed the following factors for their influence on the likelihood of seizure identification: age of the patient, seizure burden, seizure duration, anatomical distribution, and spectral characteristics of seizures such as power and frequency (both absolute and relative to the immediate pre-ictal background). We then performed multivariable reduced regression modelling to identify independent predictors of poor seizure detectability on a per-recording and per-seizure basis.

RESULTS: Likelihood of seizure identification declined with decreasing age, seizure burden and duration. Median sensitivity was significantly lower in children 3 years or younger for both aEEG (p=0.045) and CDSA (p=0.045) with a wider spread in sensitivity for younger children (figure 1). There was a trend towards lower sensitivity in recordings with fewer seizures (aEEG: R2=0.05 & CDSA: R2=0.04) and lower seizure burden (aEEG: R2=0.06 & CDSA: R2=0.06). Using aEEG, seizure identification rates were 81% for seizures lasting >180s, 67% for 60-179s, 38% for 30-59s and 13% for less than 30s. Using CDSA, seizure identification rates were 65% for seizure durations of >180s, 62% for 60-179s, 44% for 30-59s and 16% for less than 30s (figure 2). Focal seizures were more difficult to detect on aEEG/CDSA (33%/26%) compared to both hemispheric (53%/44%) and generalized seizures (64%/69%). With aEEG, other independent predictors of poor seizure detectability were lower number of seizures during the recording, lower ictal-interictal ratio of frequency at 50% power (Ratio SEF 50) and lower ictal-interictal ratio of total power (Ratio TTLPower). With CDSA, additional independent predictors of poor seizure detectability were: younger age, lower seizure burden, lower Ratio TTLPower, lower Ratio SEF50 and lower ictal-interictal ratio of frequency at 95% power (Ratio SEF95) (figure 3). 

CONCLUSIONS: aEEG and CDSA-based screening of continuous EEG should be used with caution in critically ill children with short, infrequent, and focal seizures with lower ictal-interictal power and frequency. Also, CDSA-based screening may be less sensitive in younger critically ill population. These insights will inform the optimal use of these displays and aid the design of newer QEEG trends that enhance the contrast between ictal and interictal periods, to achieve more accurate seizure identification. 

 


    This eLearning portal is powered by:
    This eLearning portal is powered by MULTIEPORTAL
Anonymous User Privacy Preferences

Strictly Necessary Cookies (Always Active)

MULTILEARNING platforms and tools hereinafter referred as “MLG SOFTWARE” are provided to you as pure educational platforms/services requiring cookies to operate. In the case of the MLG SOFTWARE, cookies are essential for the Platform to function properly for the provision of education. If these cookies are disabled, a large subset of the functionality provided by the Platform will either be unavailable or cease to work as expected. The MLG SOFTWARE do not capture non-essential activities such as menu items and listings you click on or pages viewed.


Performance Cookies

Performance cookies are used to analyse how visitors use a website in order to provide a better user experience.


Save Settings