Early Blood Glucose Level Post-Admission Correlates with the Outcomes and Oxidative Stress in Neonatal Hypoxic-Ischemic Encephalopathy

The antioxidant defense system is involved in the pathogenesis of neonatal hypoxic-ischemic encephalopathy (HIE). To analyze the relationship between first serum blood glucose levels and outcomes in neonatal HIE, seventy-four patients were divided, based on the first glucose level, into group 1 (>0 mg/dL and <60 mg/dL, n =11), group 2 (≥60 mg/dL and <150 mg/dL, n = 49), and group 3 (≥150 mg/dL, n = 14). Abnormal glucose levels had poor outcomes among three groups in terms of the clinical stage (p = 0.001), brain parenchymal lesion (p = 0.004), and neurodevelopmental outcomes (p = 0.029). Hearing impairment was more common in group 3 than in group 1 (p = 0.062) and group 2 (p = 0.010). The MRI findings of group 3 exhibited more thalamus and basal ganglion lesions than those of group 1 (p = 0.012). The glucose level was significantly correlated with clinical staging (p < 0.001), parenchymal brain lesions (p = 0.044), hearing impairment (p = 0.003), and neurodevelopmental outcomes (p = 0.005) by Pearson’s test. The first blood glucose level in neonatal HIE is an important biomarker for clinical staging, MRI findings, as well as hearing and neurodevelopment outcomes. Hyperglycemic patients had a higher odds ratio for thalamus, basal ganglia, and brain stem lesions than hypoglycemic patients with white matter and focal ischemic injury. Hyperglycemia can be due to prolonged or intermittent hypoxia and can be associated with poor outcomes.


Introduction
Birth asphyxia is a physiological derangement seen in newborn infants due to a prolonged or profound mismatch between oxygen demand and oxygen delivery [1][2][3][4]. It can cause mild to severe neurodevelopmental disabilities. Moderate to severe asphyxia can cause irreversible cerebral cell damage, neonatal seizure, and death, leading to a syndrome of hypoxic-ischemic encephalopathy (HIE) that has multi-organ involvement. Rescue hypothermia has been proven effective and has few adverse effects on newborns with HIE [5][6][7], and used to reduce neurological injury; nevertheless, a 45-55% risk of death or moderate-severe disability remains in treated infants [5,6,8]. Rescued hypothermia therapy has brought pressure on clinicians to make an early and accurate assessment of neonatal HIE and predict the severity of encephalopathy that will ensue [9]. Although hypothermia

Patients
We retrospectively reviewed the patient charts of neonates diagnosed with HIE based on a clinical history of fetal distress, metabolic acidosis, or positive-pressure ventilation immediately after birth, at Chung Shan Medical University Hospital from 2015 to 2020. The clinical stages of HIE were classified as Sarnat stage I (mild), II (moderate), and III (severe) [5,6]. Blood glucose levels were measured at the time of admission.
Further examinations for HIE, including head ultrasound (HUS), MRI, automated electrocardiography (aEEG), continuous neonatal conventional EEG monitoring, hearing testing (automated auditory brainstem response (aABR)), and auditory brainstem response (ABR) testing, were performed before discharge. For stage I patients, HUS was performed at birth and at 1, 3, 7, and 14 days of age. An MRI was performed if the clinical condition was suspected to be a brain lesion among these patients.
An experienced pediatric neurologist and neonatologist consultant divided the patients into group 1, classified as Sernat stage I (mild) HIE, and group 2, classified as stage II (moderate) and III (severe). The differences in blood biomarker levels were compared between the two groups ( Figure 1). magnetic resonance imaging (MRI) findings, and neurodevelopmental outcomes for early diagnosis of neonatal HIE.

Patients
We retrospectively reviewed the patient charts of neonates diagnosed with HIE based on a clinical history of fetal distress, metabolic acidosis, or positive-pressure ventilation immediately after birth, at Chung Shan Medical University Hospital from 2015 to 2020. The clinical stages of HIE were classified as Sarnat stage I (mild), II (moderate), and III (severe) [5,6]. Blood glucose levels were measured at the time of admission.
Further examinations for HIE, including head ultrasound (HUS), MRI, automated electrocardiography (aEEG), continuous neonatal conventional EEG monitoring, hearing testing (automated auditory brainstem response (aABR)), and auditory brainstem response (ABR) testing, were performed before discharge. For stage I patients, HUS was performed at birth and at 1, 3, 7, and 14 days of age. An MRI was performed if the clinical condition was suspected to be a brain lesion among these patients.
An experienced pediatric neurologist and neonatologist consultant divided the patients into group 1, classified as Sernat stage I (mild) HIE, and group 2, classified as stage II (moderate) and III (severe). The differences in blood biomarker levels were compared between the two groups ( Figure 1).   The HIE patients were then divided into three groups according to the first glucose level after the first admission after birth: group 1 (>0 mg/dL and <60 mg/dL, n = 11), group 2 (≥60 mg/dL and <150 mg/dL, n = 49), and group 3 (≥150 mg/dL, n = 14). The analysis of the three-group outcomes were based on short-term (clinical staging, hearing test, and MRI findings) and long-term neurodevelopmental changes at 1 year.

MRI Classification
The MRI findings were divided into two groups to study the correlation between the biomarkers and MRI changes. The first group showed no brain lesions in the parenchyma, while the second group showed brain lesions in the parenchyma. In the second group, brain MRI was classified into two subgroups based on the location of the lesions: one of the basal ganglia, thalamus, or brain stem (midbrain, pons, and lower brain stem), and the other involved areas other than the basal ganglion, thalamus, and brain stem.

Hearing Tests before First Discharge
For patients who failed the aABR test twice during universal newborn screening, ABR testing, otoacoustic emissions, and steady-state evoked potentials were performed [47]. The ABR waveforms were analyzed, and the latency of peak V was defined and adjusted by an experienced pediatric neurologist or otolaryngologist. The degree of hearing loss was classified as normal (<25 and ≤35 dB nHL) and abnormal, including mild (>35 and ≤45 dB nHL), moderate (>45 and ≤65 dB nHL), severe (>65 and ≤90 dB nHL), or profound (>90 dB nHL) [48,49].

Measurement of Neurodevelopmental Outcome at >1 Year of Age
The third edition of the Bayley Scales of Infant and Toddler Development (Bayley-III) was used to evaluate the neurodevelopmental outcomes at >1 year of age. Cognitive and motor subscales were used to interpret the neurodevelopmental outcomes. The Bayley-III scores were defined as follows: normal if both cognitive and motor subscale scores were ≥85, and abnormal if one of the cognitive and motor subscale scores was <85 [17,18].

Statistical Analysis
The independent t-test was performed to compare the means of two independent groups for the significant differences between groups, and the categorical variables were analyzed using the chi-square test. The Fisher's exact test was performed when the sample size was small. The odds ratio (OR) was calculated by dividing the odds of the first group by the odds of the second group. Furthermore, the Mann-Whitney U test was performed if the sample distribution was nonparametric, and the statistical significance was set at a p-value of < 0.05. For correlation analyses, Pearson's test was performed to measure the strength of the linear association between the two variables. All statistical tests were performed using SPSS (version 14.0; SPSS Institute, Chicago, IL, USA).
Ethical approval for the study was provided by Chung Shan Medical University Hospital's internal review board (IRB #: CS14003) and was performed in accordance with the relevant guidelines.

Informed Consent Statement
Since this is a retrospective study, informed consent was not required.

Demographic Data in Newborns with HIEs
After excluding 18 patients due to congenital anomalies (n = 7), preterm with a gestational age less than 36 weeks (n = 10), or with a confirmed genetic defect later on (n = 1), 74 patients with HIE were enrolled. Eleven belonged to group 1, 49 belonged to group 2, and 14 belonged to group 3 ( Figure 1). Among the three groups, factors, including birth weight, sex, age, and inborn or outborn method of delivery (cesarean section or vaginal delivery), were not significant ( Table 1). The 1 min and 5 min Apgar scores were not significantly different in the three groups ( Table 1). The initial blood glucose level is shown in Figure 2. Table 1. Seventy-four neonatal hypoxic-ischemic encephalopathy cases were classified into three groups, according to the first serum glucose level taken before 6 h of birth.  weight, sex, age, and inborn or outborn method of delivery (cesarean section or vaginal delivery), were not significant ( Table 1). The 1 min and 5 min Apgar scores were not significantly different in the three groups ( Table 1). The initial blood glucose level is shown in Figure 2.  First Glucose

Correlation of Hearing Impairments and Glucose Level
Hearing impairment with neonatal HIE were 1 (8.1%) in group 1, 6 (12.2%) in group 2, and 6 (42.9%) in group 3. The hearing impairment across the three groups were significantly different (χ2 (2, n = 74) = 7.7, p = 0.021). Hearing impairment was more common in group 3 than in group 1 (χ2 (1, n = 25) = 3.5, p = 0.062) and group 2 (χ2 (1, n = 63) = 6.6, p = The imaging study in group 1 showed that 1 (9.1%) out of 11 cases exhibited thalamus or basal ganglion lesions. Regarding the remaining 10 cases in group 1, 4 (36.4%) had normal brain findings and 6 (54.5%) exhibited white matter lesions or focal brain lesions that did not involve the thalamus and basal ganglia. The findings were compared with those of group 3 patients, in whom 9 (64.3 %) out of 14 had a thalamus, basal ganglion, or brain stem lesion. The difference in involving the thalamus, basal ganglion, and brain stem between groups 1 and 3 was significant (χ2 (1, n = 16) = 8.9, p = 0.002) ( Table 2). Of the 28 patients with abnormal lesions on the brain MRI scans, the glucose level was significantly correlated with the locations of the brain lesions on MRI scans (r(26) = 0.698, p < 0.001).

Discussion
A significant contribution of this study is the correlation of the first glucose level of neonates with HIE with clinical staging, findings of brain MRI, hearing outcomes, and neurodevelopmental outcomes at 1 year. Hypoglycemia and hyperglycemia are associated with advanced staging, brain parenchymal lesions, hearing impairment, and abnormal neurodevelopmental outcomes. Hyperglycemia was strongly related to hearing loss and thalamus, basal ganglia, and brain stem lesions; however, hypoglycemia was closely related to white matter lesions. This finding is also compatible with high systemic biomarker levels indicating liver injury (LDH, SGPT, SGPT, and platelets) in the hypoglycemic group. This finding supports the hypothesis that hyperglycemia is caused by prolonged or intermittent hypoxia. In addition, cases with increased blood glucose levels can be more severe than those with hypoglycemia due to the involvement of the thalamus, basal ganglia, and brain stem, especially the 8th nuclei with hearing impairments.
As hypothermia therapy needs to be performed in a timely manner, the use of a simple and convenient method, such as blood glucose measurement, can be useful in the early prediction of the staging of neonatal HIE and the need for the initiation of treatment. Although a combination of other biomarkers, such as lactate and LDH levels, can help predict the severity of HIE, obtaining the glucose level is a rapid and convenient method. This is beneficial for management, as it allows early rescue hypothermia performed 6 h after birth and permits the use of neuroprotective drugs.
Despite similar clinical staging and MRI findings, hearing impairment in patients with hyperglycemia with a first blood glucose level >150 mg/dL is worse than that in patients with hypoglycemia, with a first blood glucose level <60 mg/dL. This finding can be explained by several hypotheses. First, hypoglycemia can further induce fatty acid oxidation and cause ketosis, which can have a protective effect in the brain; hyperglycemia does not have this effect. Second, hypoxia demonstrated significant increases in plasma glucose and insulin [38]; however, intermittent or prolonged hypoxia can increase insulin resistance in genetically obese mice [35] that causes reflex hyperglycemia. Therefore, further studies on the effects of glucose abnormalities to neonatal HIE outcomes are warranted. Third, hypoglycemia can be induced by hyperinsulinemia due to neonatal HIE; however, the compensation mechanism of hyperglycemia in HIE can be exhausted, reflecting the poor condition seen in newborns.
In neonates with encephalopathy, periods of hyperglycemia were common and temporally associated with worse aEEG background scores, reduced sleep-wake cycling, and increased electrographic seizures, including after adjusting for clinical markers of hypoxiaischemia. Hyperglycemia epochs were also associated with poor aEEG background scores, including after adjusting for hypoxia-ischemia severity. Our data support the hypothesis that the proactive avoidance of hyperglycemia can be a neuroprotective strategy for infants with neonatal encephalopathy [50]. Hypoglycemic or hyperglycemic blood levels can affect MRI findings. In hypoglycemia, the watershed or focal multifocal infarcts were observed on MRI scans [43], whereas hyperglycemia was more associated with the basal ganglia or global injury findings on MRI scans [43]. In 56 neonatal HIE, all of whom died, we studied their neurodevelopmental outcomes and first 24 h glucose level and highlighted 9 patients with first glucose levels over 200 mg/dL. The findings of the aforementioned studies [28,29] were compatible with our findings that hypoglycemia and hyperglycemia can increase the risk of poor outcomes in neonatal HIE based on MRI findings. However, in our study, we highlighted that hyperglycemia was associated with a high risk of hearing impairment, which is crucial for childhood neurodevelopment. In hypothermia-treated neonates with HIE for 42 babies, 4 (9.5%) had hearing impairments. The development of hearing loss was associated with abnormal blood glucose levels, low Apgar scores, and evidence of multi-organ dysfunction and increased SGPT and SGPT levels [46], which are compatible with our findings. In addition, we also highlighted that the hyperglycemic patients had more thalamic and basal ganglion injuries than those with hypoglycemia before the first 6 h. These findings suggest that hyperglycemia can cause selective neuronal necrosis that causes injury to susceptible brain tissue, including the basal ganglia, thalamus, and brain stem. We hypothesized that the mechanism of neonatal HIE is related to glucose and clinical staging (Figure 4). Hyperglycemia in the reoxygenation and reperfusion stage can lead to further brain injury due to the consequence of oxidation stress. Hypoglycemia can cause ketogenesis by acting as an alternative cerebral fuel and as antioxidants. This can explain why the hypoglycemia group had better outcomes than the hyperglycemia group in the study. Hyperglycemia caused by insulin resistance can contribute to further brain injury as the consequence of oxidation stress that can be a useful biomarker of poor neurological outcomes and worse neurological consequences [51]. Thus, avoiding hyperglycemia after admission is mandatory in the clinical management of neonatal HIE.
creased SGPT and SGPT levels [46], which are compatible with our findings. In addition, we also highlighted that the hyperglycemic patients had more thalamic and basal ganglion injuries than those with hypoglycemia before the first 6 h. These findings suggest that hyperglycemia can cause selective neuronal necrosis that causes injury to susceptible brain tissue, including the basal ganglia, thalamus, and brain stem. We hypothesized that the mechanism of neonatal HIE is related to glucose and clinical staging (Figure 4). Hyperglycemia in the reoxygenation and reperfusion stage can lead to further brain injury due to the consequence of oxidation stress. Hypoglycemia can cause ketogenesis by acting as an alternative cerebral fuel and as antioxidants. This can explain why the hypoglycemia group had better outcomes than the hyperglycemia group in the study. Hyperglycemia caused by insulin resistance can contribute to further brain injury as the consequence of oxidation stress that can be a useful biomarker of poor neurological outcomes and worse neurological consequences [51]. Thus, avoiding hyperglycemia after admission is mandatory in the clinical management of neonatal HIE. However, this study has some limitations. We presented a limited number of HIE cases. Our findings can be biased and comprised owing to the fewer cases than needed for reliable results. Therefore, further studies with an increased number of cases are warranted. Furthermore, in the HIE stage I group with favorable outcomes, an aggressive image study is not available from the national insurance agency in Taiwan. However, a series of HUS can support the imaging findings, and a clinical follow-up of up to 1 year can diagnose the patients without significant brain parenchymal lesions.

Conclusions
The first blood glucose level is an important biomarker for clinical staging, MRI findings, hearing impairment, and neurodevelopmental outcomes in neonatal HIE. Hyperglycemic patients had higher ORs in the thalamus, basal ganglia, and brain stem lesions than hypoglycemic patients who were often related to white matter and focal ischemic injury. This finding supports the fact that hyperglycemia possibly occurred due to prolonged or intermittent hypoxia and oxidation stress, and led to worse outcomes due to the involvement of the thalamus and basal ganglia. In neonatal HIE, early glucose levels after the first admission can be a rapid and convenient biomarker for a timely diagnosis and early treatment administration.