Combination of initial neurologic examination and continuous EEG to predict survival after cardiac arrest

Abstract

Background: Prognosticating outcome following cardiac arrest requires a multimodal approach. We tested whether the combination of initial neurologic examination combined with continuous EEG was superior to either test alone for predicting survival after cardiac arrest.

Methods: Review of consecutive patients receiving continuous EEG monitoring between April 2010 and June 2013. Initial neurologic examination was evaluated using the Full Outline of UnResponsiveness (FOUR) score and organ system dysfunction determined using the SOFA score. We defined four categories of initial post-cardiac arrest illness severity (PCAC): (I) awake, (II) coma (not following commands but intact brainstem responses) + mild cardiopulmonary dysfunction (SOFA cardiac + respiratory score < 4), (III) coma + moderate-severe cardiopulmonary dysfunction (SOFA cardiac + respiratory score ≥ 4), and (IV) coma without brainstem reflexes. A second analysis focusing on neurologic injury divided subjects into three groups according to initial FOUR_B score; FOUR_B = 0-1, FOUR_B = 2 and FOUR_B = 4. A blinded rater dichotomized continuous EEG patterns during the first 48h into malignant patterns (non-convulsive status epilepticus, convulsive status epilepticus, myoclonic status epilepticus and generalized periodic epileptiform discharges). The primary outcome was survival to hospital discharge.

Results: Of 331 subjects, mean age was 58 (SD 17) years and 206 (62.2%) subjects were male. Ventricular fibrillation or tachycardia (VF/VT) was the initial rhythm for 93 (28.1%) subjects. Among subjects with malignant cEEG, survival to hospital discharge rate was 0% for FOUR_B 0-1, 8.1% for FOUR_B 2 and 12.5% for FOUR_B 4, respectively. In one multivariate analysis, survival was independently associated with VF/VT, FOUR_B of 2, FOUR_B of 4, and non-malignant cEEG. In a separate model, survival was associated with VF/VT, PCAC < 4 and non-malignant cEEG. The AUCs of FOUR_B, cEEG and the combination of FOUR_B and cEEG are 0.740 (95% C.I. 0.684-0.797), 0.674 (95% C.I. 0.615-0.732) and 0.820 (95% C.I. 0.773-0.868) respectively. The AUCs of PCAC, cEEG and the combination of PCAC and cEEG are 0.779 (95% C.I. 0.721-0.838), 0.672 (95% C.I. 0.612-0.7321) and 0.846 (95% C.I. 0.798-0.894) respectively.

Conclusion: Combining the initial neurologic examination using either FOUR_B or PCAC, with cEEG was superior to any individual test for predicting survival after cardiac arrest. We caution against using these findings to speed prognostication until they are externally validated.

Keywords: Cardiac arrest; Examination; Hypothermia; Outcomes; Prognostication.