Abstract PR117

Abstract

Background & Objectives: Bilateral absence of cortical response to somatosensory evoked potentials (SSEPs) after cardiac arrest (CA)is an established marker of poor neurological outcome. However, the presence of these responses (N20PRES) has failed to reliably predict good neurological outcome. In the setting of multimodal monitoring, few data are available on the clinical characteristics of N20PRES patients with poor neurological outcome. Materials & Methods: Retrospective analysis of an institutional database (January 2010 to January 2013) including all adult patients admitted to the Intensive Care Unit (ICU) after CA and undergoing SSEPs between 48 and 72 hours after cardiac arrest. All patients underwent targeted temperature management (TTM; 32–34°C) for 24 hours. We collected the absence of pupillary reflexes, absent or posturing motor response and status myoclonus on day 2–3 and electroencephalography (EEG) data (e.g. absence of reactivity to painful stimuli; presence of a malignant pattern, such as burst-suppression or flat tracings; electroencephalographic status epilepticus) after return to normothermia. Poor neurological outcome (PNO) was defined as a Cerebral Performance Categories of 3–5 at assessed at 3 months Results: We studied 126 patients, including 81 with N20PRES; 55 had PNO. Demographics, comorbidities and CA characteristics were similar between groups except for non-cardiac origin of arrest and non-shockable rhythm which were more frequent in patients with PNO than in the others. PNO Patients with N20PRES were more likely to have absent pupillary reflexes (31% vs. 0%; p=0.003), absent or posturing motor response (51% vs. 11%; p=0.001) on day 2–3 than the other patients. They were also more likely to have a malignant EEG pattern, particularly burst suppression (18% vs. 0%; p=0.02) and a non-reactive EEG (75% vs. 27%; p<0.001). All patients with absent pupillary reflexes on day 2–3 had a PNO (p=0.003). Conclusion: In patients with preserved cortical responses to painful stimuli prognostication may be reliably guided by clinical and EEG signs. The presence of early malignant EEG patterns and the absence of pupillary reflexes on day 2–3 are associated with poor neurological outcome The use of a multimodal assessment may be helpful to guide neuroprognostication in this setting.