Abstract 177: Optimal Selection Of Predictors To Enhance Prognostic Performance In Post-resuscitation Care: A Single Center Experience

Abstract

Objectives: The aim of this study was to investigate the optimal combinations that would enhance prognostic accuracy in post-resuscitation care after out-of-hospital cardiac arrest (OHCA). Methods: Prospectively registry data collected between May 2018 and January 2022 were analyzed. We analyzed the prognostic performance of all single predictors suggested from current guidelines: no ocular reflexes (OR) at 72 h, highly malignant pattern in electroencephalography at ≥ 24 h, neuron-specific enolase (NSE) at 72 h, and diffuse anoxic injury at 72 h on image studies of computed tomography (CT) and diffusion-weighted image (DWI). Gray-to-white matter ratio and high-signal intensity on CT and DWI, respectively, were used as a tool to estimate abnormal findings. We also analyzed the prognostic accuracy of combining strategies using 2 to 3 abnormal results from predictors. All prognostic accuracies were analyzed as sensitivities with false positive ratios (FPR) of < 2%. The primary outcome was poor neurological outcome after 3 months from OHCA. Results: We assessed 119 comatose adult survivors of OHCA, of whom 62 (52.1%) had poor neurological outcomes. The highest sensitivity among single prognostic modalities was found in abnormal DWI, with an FPR of < 2% (86.3%, 95% confidence interval [CI] 73.7 - 94.3). In multimodal assessments, combining NSE and DWI (87.8%, 95% CI 75.2 - 95.4) and combining NSE, OR, and DWI (87.8%, 95% CI 75.2 - 95.4) showed the highest sensitivities in dual and triple combining prognostic strategies, respectively. Combining strategies with image study at 72 h, regardless of whether CT or DWI, improved the sensitivities of every non-imaging study modality. Conclusions: This study emphasizes that combining prognostic strategies for multimodal assessment can enhance prognostic reliability. In particular, this combining strategy, along with the image study at 72 h, can be valuable to ensure the reliability of prognostication in post-resusciation care. 1