Abstract 18009: Early Prognostic Assessment of Patients Who Underwent Target Temperature Management: A Multiparametric Model

Introduction: Targeted temperature management (TTM) after return of spontaneous circulation is considered beneficial in comatose patients; identification of prognostic factors to stratify patients with worse outcome, especially within 72 hours after cardiac arrest (CA), still remains an important task. Aim: To realize a multiparametric model of early prognostication (within 24 hours after CA) for post-cardiac-arrest resuscitated patients, treated with TTM. Methods: 120 consecutive patients (2011-2015) resuscitated from CA and treated with TTM (target temperature 32.5°C) were studied. In all subjects, records on continuous hemodynamic parameters, biochemical markers (NSE and S100-B protein), echocardiographic data, CA anamnestic features and TTM strategy were pooled. We also evaluated electroencephalogram parameters and somatosensory evoked potentials (N20) during normothermia and hypothermia. Outcome was defined according to the Cerebral Performance Categories (CPC), considering a CPC of ≥ 3 as an unfavorable outcome. Multivariate analysis found out parameters which predicted a worse outcome (permutation test was applied to increase the sample size if necessary). We elaborated a prognostic model for patients who underwent TTM, adopting a conditional inference tree analysis. Results: NSE and S100-B protein value at 24 hours, N20 and age seem to be the best predictors of patient outcome, according to the multivariate analysis. Absence of N20 responses in these patients was associated with a worse outcome, even if the NSE elevation accuracy turned out to be superior. We produced a prognostic model based on three nodes (reference values in the figure). ROC analysis showed an area under the curve of 0.8. Conclusions: This multiparametric model appears to be useful to predict outcome of patients who underwent TTM, thus supporting Intensivists to better tailoring the post cardiac arrest care to each patient, based on an early prognostic stratification.