Does a combination of ≥2 abnormal tests vs. the ERC-ESICM stepwise algorithm improve prediction of poor neurological outcome after cardiac arrest? A post-hoc analysis of the ProNeCA multicentre study

Abstract Background bilaterally absent pupillary light reflexes (PLR) or N20 waves of short-latency evoked potentials (SSEPs) are recommended by the 2015 ERC-ESICM guidelines as robust, first-line predictors of poor neurological outcome after cardiac arrest. However, recent evidence shows that the false positive rates (FPRs) of these tests may be higher than previously reported. We investigated if testing accuracy is improved when combining PLR/SSEPs with malignant electroencephalogram (EEG), oedema on brain computed tomography (CT), or early status myoclonus (SM). Methods post-hoc analysis of ProNeCA multicentre prognostication study. We compared the prognostic accuracy of the ERC-ESICM prognostication strategy vs. that of a new strategy combining ≥2 abnormal results from any of PLR, SSEPs, EEG, CT and SM. We also investigated if using alternative classifications for abnormal SSEPs (absent-pathological vs. bilaterally-absent N20) or malignant EEG (ACNS-defined suppression or burst-suppression vs. unreactive burst-suppression or status epilepticus) improved test sensitivity. Results we assessed 210 adult comatose resuscitated patients of whom 164 (78%) had poor neurological outcome (CPC 3-5) at six months. FPRs and sensitivities of the ≥2 abnormal test strategy vs. the ERC-ESICM algorithm were 0[0-8]% vs. 7 [1–18]% and 49[41-57]% vs. 63[56-71]%, respectively (p  Conclusions in comatose resuscitated patients, a prognostication strategy combining ≥2 among PLR, SSEPs, EEG, CT and SM was more specific than the 2015 ERC-ESICM prognostication algorithm for predicting 6-month poor neurological outcome.