Introduction: We evaluated the concordance of the Neurological pupil Index (NPi) with other predictors of outcome after cardiac arrest (CA).
Methods: Post hoc analysis of a prospective, international, multicenter study including adult CA patients. Predictors of unfavorable outcome (UO, Cerebral Performance Category of 3-5 at 3 months) included: a) worst NPi ≤ 2; b) presence of discontinuous encephalography (EEG) background; c) bilateral absence of N20 waves on somatosensory evoked potentials (N20ABS); d) peak neuron-specific enolase (NSE) blood levels > 60 mcg/L; e) myoclonus, which were all tested in a subset of patients who underwent complete multimodal assessment (MMM).
Results: A total of 269/456 (59 %) patients had UO and 186 (41 %) underwent MMM. The presence of myoclonus was assessed in all patients, EEG in 358 (78 %), N20 in 186 (41 %) and NSE measurement in 228 (50 %). Patients with discontinuous EEG, N20ABS or high NSE had a higher proportion of worst NPi ≤ 2. The accuracy for NPi to predict a discontinuous EEG, N20ABS, high NSE and the presence of myoclonus was moderate. Concordance with NPi ≤ 2 was high for NSE, and moderate for discontinuous EEG and N20ABS. Also, the higher the number of concordant predictors of poor outcome, the lower the observed NPi.
Conclusions: In this study, NPi ≤ 2 had moderate to high concordance with other unfavorable outcome prognosticators of hypoxic-ischemic brain injury. This indicates that NPi measurement could be considered as a valid tool for coma prognostication after cardiac arrest.
Keywords: Automated pupillometer; Automated pupillometry; Brain injury; Concordance; Heart arrest; Prognosis.
Copyright © 2022 Elsevier B.V. All rights reserved.