Prognostication of neurological outcome after cardiac arrest using wavelet phase coherence analysis of cerebral oxygen.

Abstract

The prognosis for cardiac arrest (CA) is associated with the degree of cerebral ischemia. We investigated the relationship between the wavelet coherence of cerebral oxyhemoglobin (HbO₂) among different channels and outcomes after CA. Moreover, we aimed to develop a prognostication method after CA. Eighty-three post-resuscitation patients were included. The HbO₂ data were collected during the post-resuscitation period (median day, 1) using functional near-infrared spectroscopy. The coherence between sections of prefrontal HbO₂ oscillations in five frequency intervals (I, 0.6-2 Hz; II, 0.15-0.6 Hz; III, 0.05-0.15 Hz; IV, 0.02-0.05 Hz; and V, 0.0095-0.02 Hz) were analyzed. We evaluated the outcomes using cerebral performance category (CPC) scores (good outcome, CPC ≤ 2 and poor outcome, CPC ≥ 3) at 3 months after CA. Additionally, the predictive method was developed using the biomarker and coherence value after CA. Among the included patients, 19 patients (22.9%) had a good outcome. Poor outcome group had significantly lower phase coherence in the myogenic frequency interval III compared to good outcome group (0.36 ± 0.14 vs. 0.54 ± 0.18, P < 0.001). The predictive method using neuron-specific enolase (NSE) and interval III value demonstrated good discrimination (area under the curve 0.919; 95% confidence interval, 0.850-0.989). The predictive method using NSE and phase coherence of HbO₂ in the interval III from the vascular smooth muscle cells could be a useful tool for prognosticating after CA. This suggests that evaluating cerebral ischemia using phase coherence of HbO₂ might be a helpful outcome predictor following CA.