Combination of initial neurologic examination, quantitative brain imaging and electroencephalography to predict outcome after cardiac arrest.

BACKGROUND: Prognosticating outcome following cardiac arrest is challenging and requires a multimodal approach. We tested the hypothesis that the combination of initial neurologic examination, quantitative analysis of head computed tomography (CT) and continuous EEG (cEEG) improve outcome prediction after cardiac arrest.
METHODS: Review of consecutive patients receiving head CT within 24h and cEEG monitoring between April 2010 and May 2013. Initial neurologic examination (Full Outline of UnResponsiveness_Brainstem reflexes (FOUR_B) score and initial Pittsburgh Post-Cardiac Arrest Category (PCAC)), gray matter to white matter attenuation ratio (GWR) on head CT and cEEG patterns were evaluated. The primary outcome was in-hospital mortality.
RESULTS: Of 240 subjects, 70 (29%) survived and 22 (9%) had a good neurologic outcome at hospital discharge. Combined determination of GW ratio and malignant cEEG had an incremental predictive value (AUC: 0.776 for mortality and 0.792 for poor neurologic outcome), with 0% false positive rate when compared with either test alone (AUC of GW ratio: 0.683 for mortality and 0.726 for poor outcome, AUC of malignant cEEG: 0.650 for mortality and 0.647 for poor outcome). Addition of FOUR_B or PCAC to this model improved prediction of mortality (p=0.014 for FOUR_B and 0.001 for PCAC) but not of poor outcome (p=0.786 for FOUR_B and 0.099 for PCAC).
CONCLUSIONS: Combining GWR with cEEG was superior to any individual test for predicting mortality and neurologic outcome. Addition of clinical variables further improved prognostication for mortality but not neurologic outcome. These preliminary data support a multi-modal prognostic workup in this population.