Electroencephalography for neurological prognostication after cardiac arrest

Sammanfattning: This thesis focuses on the prognostic value of electroencephalography(EEG) in comatose patients resuscitated after cardiac arrest (CA), using both simplified continuous EEG monitoring (cEEG) and routine EEG.

Background:
Comatose survivors are admitted to an intensive care unit (ICU) to support vital functions. Postresuscitation care includes target temperature management (TTM) for 24 hours. The degree of brain injury after CA varies among patients. Withdrawal of life-sustaining therapy due to presumed extensive brain injury is the most common cause of death during the hospital stay. Multiple prognostic tools are used to identify patients with a potential for recovery. Next to the neurological examination, EEG is the most commonly used tool to assess
prognosis. However, the value of EEG has been limited by varying classification systems, interrater variability and influence of sedation.

Methods:
In the “coma project” (2004-2008) consecutive patients at the general ICU in Lund were monitored with simplified cEEG from arrival until 120 hours after CA. Pre-defined cEEG patterns at different time points were
correlated to outcome.
In the TTM trial (2010-2013) where patients were randomized to 33ºC versus 36ºC, a routine EEG was performed in patients still comatose after rewarming. The EEGs were classified into highly malignant, malignant and benign patterns by four EEG specialists from different countries according to the standardized EEG terminology proposed by the American Clinical Neurophysiology Society. The rationale and study design for this EEG evaluation was published.

Results:
95 patients in the “coma project” were monitored with simplified cEEG. A continuous background at start of registration or at normothermia strongly predicted a good outcome. All patients with electrographic status epilepticus (ESE) evolving from a burst-suppression background died without regaining consciousness whereas ESE evolving late from an established continuous background was compatible with good outcome.
At 8 selected TTM trial sites, routine EEGs were recorded after rewarming in 103 comatose patients. A highly malignant EEG was identified with substantial interrater agreement and had a specificity of 100% to predict poor outcome for all four EEG specialists. Any malignant EEG feature was identified with moderate interrater agreement but had a low specificity to predict a poor outcome (48%). A benign EEG was found in 1% of the patients with a poor outcome.

Conclusions:
Simplified cEEG provides early positive and negative prognostic information in comatose patients after cardiac arrest.
A highly malignant routine EEG after rewarming reliably predicted a poor outcome. An isolated malignant routine EEG feature was not a reliable predictor whereas a benign routine EEG was highly predictive of good outcome.