Electroencephalography and optical neuromonitoring predict short-term outcomes in neonates undergoing therapeutic hypothermia for hypoxic-ischemic encephalopathy

Abstract

Electroencephalography (EEG) and optical neuromonitoring were used to predict short-term outcomes in neonates undergoing therapeutic hypothermia (TH) for hypoxic-ischemic encephalopathy (HIE). Fifty-two neonates undergoing TH for HIE were prospectively recruited. Continuous EEG monitoring was initiated within 24 h of life and a quantitative discontinuity index was calculated. Combined frequency-domain near infrared spectroscopy (FDNIRS) and diffuse correlation spectroscopy (DCS) were initiated within 48 h of life and used to measure cerebral hemoglobin oxygen saturation (SO₂) and a cerebral blood flow index. Using these parameters and hemoglobin concentration measurements, cerebral oxygen extraction fraction (OEF), indices of cerebral oxygen delivery and metabolism (CMRO_2i) as well as cerebral oxygen reserve (CRO₂) were derived. Short-term outcome was classified based on brain injury pattern on magnetic resonance imaging and/or death; as normal-mild, moderate or severe outcome. Results showed that EEG discontinuity index, SO₂ and CRO₂ were higher and OEF lower in neonates with severe compared to normal-mild and moderate outcomes during TH. EEG discontinuity index was the most accurate and earliest parameter to identify moderate vs. severe outcomes while CMRO_2i identified normal-mild vs. moderate outcomes as early as day 2 of TH. Combining EEG and FDNIRS-DCS parameters improved area-under-the-curve, sensitivity and specificity for most of the predictive models.

Keywords: Diffuse correlation spectroscopy (DCS); Electroencephalography (EEG); Frequency-domain near infrared spectroscopy (FDNIRS); Hypoxic-ischemic encephalopathy (HIE); Magnetic resonance imaging (MRI); Predictive models.

Fig. 1

Boxplot representation of the EEG discontinuity index in patients with normal-mild (white), moderate (light gray) and severe (dark gray) short-term outcomes during therapeutic hypothermia day 1 (TH1), day 2 (TH2), day 3 (TH3), rewarming (Rewarm) and post-therapy (Post-TH). Mean, median and outliers are represented by squares (■), lines (▬) and plus ( +) symbols, respectively. *p < 0.05; **p < 0.01; ***p < 0.001.

Fig. 2

Boxplot representation of (A) cerebral oxygen hemoglobin saturation (SO₂), (B) cerebral oxygen extraction fraction (OEF), (C) cerebral oxygen reserve (CRO₂), and (D) cerebral metabolic rate of oxygen consumption (CMRO_2i) in HIE neonates with normal-mild (white), moderate (light gray) and severe (dark gray) short-term outcomes during therapeutic hypothermia day 2 (TH2), day 3 (TH3), rewarming (Rewarm), post-therapy (Post-TH) and before discharge (Pre-DC). Mean, median and outliers are represented by squares (■), lines (▬) and plus ( +) symbols, respectively. *p < 0.05; **p < 0.01; ***p < 0.001.