Abstracts

Rationale: Bilateral absent cortical responses on somatosensory evoked potential studies is well established as a good predictor of poor prognosis in comatose patients(1). However, given that the prevalence of absent N20 is low, the sensitivity of this neurodiagnostic tool is low to moderate. We examine the clinical utility of SSEPs at an urban center and contrast to other prognostic tools. (1) Wijdicks EF, Hijdra A, Young GB, Bassetti CL, Wiebe S. Practice parameter: prediction of outcome in comatose survivors after cardiopulmonary resuscitation (an evidence-based review): report of the Quality Standards Subcommittee of the American Academy of Neurology. Neurology 2006;67:203-210. Methods: A retrospective review of inpatient EPs from 2009 to 2014 was done. Only those who have had SSEPs done in comatose patients for the purposes of prognostic indication were included. The chart was reviewed for etiology, GCS, hospital course, therapeutic hypothermia, latency to SSEP, the presence of EEG, and outcomes. SSEP results were reviewed for absence of N20 bilaterally, and EEG results were reviewed and classified as highly malignant, malignant, or benign as recently published by Westhall et al.(2). Highly malignant EEG was defined as having one of the following findings; suppressed background without discharges, suppressed background with continuous periodic discharges, or burst-suppression background with or without discharges. Glasgow Outcome Scale (GOS) was used to classify bad outcome (GOS 1-2) and good outcome (GOS 3-5). (2) Westhall E, Rossetti AO, van Rootselaar AF, et al. Standardized EEG interpretation accurately predicts prognosis after cardiac arrest. Neurology 2016;86:1482-1490. Results: We identified 67 patients with inpatients SSEPs. 63 patients had SSEPs for prognostication and were included in this study, of which 48 (76.19%) received therapeutic hypothermia. Overall, eight SSEP studies (12.69 %) were artifactual, 23 (33.33%) were normal, 13 (20.63%) revealed present but abnormal N20, and 19 (30.15%) showed bilaterally absent N20. Mean latency from admission to SSEP was 5.75 days while mean latency to SSEP from hypothermia was 2.16 days. Sensitivity and specificity of bilaterally absent SSEP in predicting poor outcome was 36.54% (95% CI 23.6-51) and 100% (95% CI 71.5-100) respectively. 29 patients (46.03 %) with post-anoxic brain injury had additional EEG testing. 14 had highly malignant EEG and 15 had malignant or benign EEG. Sensitivity and Specificity of a highly malignant EEG in predicting poor outcome was 56% (95% CI 34.9-75.6) and 100% (95% CI 39.8-100) respectively. 64% of very malignant EEGs (n=9) were done on patients with non-prognostic SSEP (present N20 or artifactual study). The combination of very malignant EEG and/or absent SSEP had a sensitivity of 53.85%. Conclusions: Highly malignant EEG has the same specificity in predicting absent N20 responses on SSEP in our cohort with better sensitivity and serves to improve sensitivity in predicting poor outcome when added to SSEP study without reaching statistical significance. Furthermore, technical difficulties in obtaining adequate SSEP in more than 10% of patients introduces additional challenges. This is likely related to portability of the SSEP machine and patient factors such as edema and adventitious movements in the ICU. Larger studies that includes a cost-benefit analysis will be useful in reproducing these results and in determining how to incorporate EEG in the decision tree of prognosis in post-anoxic brain injury Funding: NO EXTERNAL FUNDING