Transitional Neonatal Hypoglycemia and Adverse Neurodevelopment in Midchildhood

Key Points Question Are episodes of severe transitional neonatal hypoglycemia associated with adverse neurodevelopment in midchildhood? Findings In this cohort study of 140 children aged 7 to 11 years, a history of severe transitional neonatal hypoglycemia was associated with a 4.8 points lower full-scale IQ and significantly lower performance in visual-motor and fine motor functions. Meaning These results suggest that severe transitional hypoglycemia is associated with adverse neurodevelopmental outcomes; treatment thresholds and goals for neonatal hypoglycemia should be high enough to prevent hypoglycemia at such levels from occurring.


Introduction
Hypoglycemia is the most common metabolic condition in neonates.The reported incidence is 15% in all neonates 1 and 50% in babies born with risk factors, including infants of mothers with diabetes, those large or small for their gestational age, preterm births, or births with perinatal stress. 2The most common form is transitional neonatal hypoglycemia, which occurs during the metabolic transition from intrauterine to extrauterine life.4][5][6] In children with risk factors, hypoglycemia may be more severe and prolonged, or may occur later in the first days of life, eg, in case of lower glycogen reserves, increased glucose consumption, or increased insulin secretion. 7Therefore, neonates with risk factors usually undergo blood glucose screening to identify hypoglycemia early and treat it appropriately.However, guideline recommendations and treatment thresholds vary widely, particularly because it remains unclear what blood glucose levels are considered physiological in the first days of life and under what circumstances hypoglycemia can damage the neonatal brain. 80][11] However, data on long-term neurodevelopmental outcomes beyond childhood for children with transitional neonatal hypoglycemia are limited.This is important because specific developmental consequences may not manifest until midchildhood or late childhood and may only be detected through developmental assessments.Therefore, the aim of this study was to evaluate whether transitional neonatal hypoglycemia with blood glucose measured at 30 mg/dL or lower (to convert glucose concentrations to millimoles per liter, multiply by 0.0555) is associated with adverse neurodevelopment in midchildhood compared with a control group.

Study Design
The ProBrain-D 7-11 Study is a matched cohort study that examined the neurodevelopmental outcomes of 140 children aged 7 to 11 years who were born in a tertiary hospital in Düsseldorf, Germany.Neurodevelopmental testing was prospectively conducted.The study was performed between March 2022 and February 2023.
Two groups were recruited based on retrospective data: 70 children (50%) with severe neonatal hypoglycemia, defined as a history of at least 1 recorded blood glucose level of 30 mg/dL or below (exposed group) and 70 children (50%) with all blood glucose levels above 30 mg/dL (unexposed group).3][14][15][16][17][18][19][20][21] However, none of these thresholds are currently supported by evidence.Additionally, studies using higher cutoffs, ie, between 36 and 47 mg/dL, have conflicting and inconclusive results regarding the neurodevelopmental consequences of hypoglycemia.Therefore, the cutoff of 30 mg/dL was intentionally selected for this study to provide evidence of neurological consequences at a lower threshold that usually triggers immediate action and intravenous glucose treatment.
To identify eligible patients, the database of all births in our hospital from 2010 to 2015 was first screened using International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) codes (Figure).Of these, medical records were reviewed.The inclusion criterion was existing data on neonatal blood glucose screening, eg, due to suspected hypoglycemia or risk factors for hypoglycemia as defined by the hospital's screening protocol (eAppendix in Supplement 1).Children with persistent hypoglycemia or a history of any risk factor or condition that could potentially cause adverse neurodevelopment other than hypoglycemia were excluded (Figure).All children were screened and treated for hypoglycemia according to the hospital's Recruitment was performed via telephone and email.Participants exposed to severe hypoglycemia were contacted in random order from the list of eligible individuals until the target number of 70 was recruited.In parallel, participants for the unexposed group were also contacted in random order from the eligible candidates, frequency matched to the exposure group for sex, birth weight, gestational age, socioeconomic status (SES) according to the SES Index, 23 and primary risk factors for neonatal hypoglycemia.Parents were not informed of the group allocation or blood glucose cutoffs for group assignment at recruitment to avoid voluntary bias depending on exposure status.No incentives were offered for participation.Written informed consent was obtained from all parents along with childrens' assent.The study was approved by the institutional review board of the Medical Faculty of the Heinrich-Heine-University Düsseldorf, Germany, and was performed  following the Declaration of Helsinki. 24This report follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for observational studies.

Measures
Neurodevelopmental assessment was conducted by trained pediatricians or psychologists who were masked to the neonatal glycemic history of the children.Neurodevelopmental outcome tests were selected based on previously published studies to allow comparison of results.The predefined primary outcome parameter was the assessment of cognitive function (full-scale IQ) using Wechsler's Intelligence Scale for Children-Fifth Edition (WISC-V) (standardized mean [SD], 100 [15]; below 85 points [ie, 15th percentile] was the cutoff for IQ below average). 25 50 [10]), 29 and a questionnaire providing additional information regarding the child's medical history, development and education, and the family's SES.
Data were recorded in a digital database (Claris International Inc).The medical and birth history was obtained from the medical records of both the participants and their mothers.Furthermore, we reviewed data from the child's last 2 school report cards and from the German child examination booklet, which contains all information on pediatric check-ups and developmental assessments.All data not required for recruitment or matching were collected after completion of developmental testing to avoid information and selection bias.

Statistical Analysis
Sample size estimation was performed using G*Power version 3.1 (Heinrich-Heine-University). 30For the primary outcome parameter, we estimated a total sample size of at least 140 children (70 per group) to detect a true mean difference in IQ scores of 0.5 SD with a power of 0.9 at an α level of .05using univariate analysis.Data were controlled for prespecified potential confounders by frequency matching the groups.Post hoc analysis of covariance was conducted to adjust the data for the remaining possible explanatory confounders that were assessed in the study but not already controlled for by matching, excluded by sensitivity analysis, or related to hypoglycemia.Unadjusted data are reported in eTable 1 in Supplement analyze associations between categorical variables.Mean differences (MD) or odds ratios (OR) are reported with a 95% CI for continuous or binary data, respectively.The effect size was evaluated using partial η p 2 (0.01 = small; 0.06 = medium; 0.14 = large effect). 31A 2-sided P < .05 was considered statistically significant.Adjustment for multiple comparisons was not applied because the study was only powered to test for significance of the primary end point at uncorrected P < .05.
Regarding the subscores of the WISC, the main objective was to analyze which specific subscales are primarily affected by hypoglycemia and therefore contribute most to the observed effect in our primary end point.Similarly, secondary end points were assessed as exploratory.[95% CI, 1.5 to 15.5]) (Table 3).There were no differences in balance, but both groups had difficulty with the gross motor tests (

Discussion
Here, we compare the standardized neurodevelopmental assessment of 140 children aged 7 to 11 years, with and without a history of severe transitional neonatal hypoglycemia.Children exposed to severe hypoglycemia had a 4.8 points lower full-scale IQ on the WISC-V when compared with a matched group of unexposed children.Furthermore, exposed children had significantly lower scores for visual-motor function, general visual perception, fine motor function, as well as total motor Few studies have examined the neurodevelopmental outcomes of school-aged children after neonatal hypoglycemia using standardized measures.One retrospective study 32 found no significant difference in WISC-IV full-scale IQ at ages 6 to 9 years between 71 children after severe neonatal hypoglycemia and 32 control siblings.However, the study was not powered to detect differences of less than 8.4 IQ points, and blood glucose values were not reported for the control group because the only inclusion criterion was "no hospital report of neonatal hypoglycemia."Consistent with our findings, there were lower scores for fine motor function on the MABC-2 in the hypoglycemia group, and no significant differences for gross motor function or balance.
Two studies have evaluated academic performance in midchildhood after neonatal hypoglycemia. 33,36However, academic performance may be confounded by various factors, such as the quality and conditions of education.After controlling for confounders, test scores on a fourthgrade achievement test in 1395 ten-year-old children revealed an association of neonatal hypoglycemia with glucose levels between 40 mg/dL and 45 mg/dL with decreased proficiency in mathematics and literacy, respectively. 33However, glycemic and developmental data were retrospective, many childhood characteristics were unknown (eg, primary household language, disabilities), only the first 2 blood glucose values were evaluated, and the lowest blood glucose value or treatment strategies were not reported.
5][36] All children underwent blood glucose screening after birth and were treated to maintain blood glucose levels above 47 mg/dL.There were no differences in neurodevelopmental outcomes between children with neonatal hypoglycemia who had glucose levels below 47 mg/dL and euglycemic infants at 2 years of age. 35However, consistent with our findings, neonatal hypoglycemia b Recurrent hypoglycemia was defined as Ն2 episodes; severe hypoglycemia defined as blood glucose measurements of 30 mg/dL or below.
c Sensitivity analysis for the item did not alter the primary conclusion.
d Score range: 3.2 to 21.0.Higher scores indicate higher socioeconomic status. 23t that glucose level correlated with an increased risk of poor visual-motor function and poor executive function at 4.5 years, especially in those with recurrent hypoglycemia or severe episodes with glucose measured below 36 mg/dL.34 At 9 and 10 years of age, educational achievements and other neurodevelopmental domains were not significantly different between the groups.In contrast to our cohort, both groups showed low overall performance, academic achievements, fine-and visual-motor functions, emotional behavior regulation, and executive functions.36 It was hypothesized that the underlying risk factor for hypoglycemia may have a greater impact on neurodevelopmental trajectories than hypoglycemia itself.In our cohort, 61.4% of neonates had their lowest blood glucose level within the first 2 hours of life.This suggests that severe hypoglycemia may result in adverse outcomes even at the time of  ) only reported for significant values (.01 = small; .06= medium; .14 = arge effect).

Strengths and Limitations
Strengths of this study include the matched cohort design to adjust for potential confounders, and the prospective neurodevelopmental assessment of cognition, motor function, visual function, executive function, and child behavior using standardized measures by investigators masked to neonatal glycemia, powered to allow the detection of small, but still clinically significant, betweengroup differences.Treatment and screening for hypoglycemia was similar for all children because a single-center protocol was used.
This study has some limitations.First, because neonatal glycemia data are retrospective, granularity and accuracy may be compromised compared with a study design capturing neonatal data prospectively.Second, despite matching and adjustment for confounders of the primary outcome, we cannot rule out residual confounders that were not accounted for in this study.Third, the study was only powered to test for significance of the primary end point at uncorrected P < .05.
Therefore, all secondary exploratory end points need to be interpreted cautiously and require further evaluation in confirmatory studies.
Fourth, high SES, above average full-scale IQ, and a high parental educational level are overrepresented in both groups, which limits generalizability of the results.However, after careful group matching and data adjustment, this aspect should not bias the between-group differences in terms of neurodevelopmental outcomes.Fifth, the low overall performance on the gross motor function test could be attributed to the extended cancellation of physical education classes and limited social and interactive play throughout the COVID-19 pandemic.

Conclusions
Our data suggest that the group of neonates with severe hypoglycemia measured at 30 mg/dL or below was associated with an increased blood glucose-related risk of suboptimal neurodevelopmental outcomes.Therefore, treatment strategies should aim to prevent episodes of neonatal hypoglycemia at these levels until further data are available on intervention thresholds and neurodevelopmental outcomes in school-aged children from prospective and randomized multicenter trials.

Table 1 .
Characteristics of Study Participants (continued) a Including births from 35 weeks, 0 days to 36 weeks, 6 days.

Table 3 .
Association Between Neonatal Hypoglycemia and Underachievement on Neurodevelopmental Tests in Midchildhood

Table 2 .
Association of Neonatal Hypoglycemia and Neurodevelopmental Outcomes in Midchildhood a Results from analyses of covariance adjusting for potential confounding by parental highest educational level, maternal smoking in pregnancy and breastfeeding are reported.Missing data were not imputed: 3 missing data on breastfeeding, 3 on BRIEF score, and 4 on CBCL score.
a b Partial eta-squared (η p 2 Transitional Neonatal Hypoglycemia and Adverse Neurodevelopment in Midchildhood Preventive measures, such as early breastfeeding and prevention of hypothermia, should therefore be considered.Our study supports that visual-motor functions are particularly affected by neonatal hypoglycemia.These are largely represented in the primary visual cortex in the occipital and parietal lobes, where structural damage following neonatal hypoglycemia has been described frequently, although the underlying mechanism for this spatial association remains unidentified.39Notably,although we show that neonatal blood glucose levels of 30 mg/dL or below are associated with adverse effects on neurodevelopment, this does not imply that values above 30 mg/dL are completely safe, or that the threshold for potential brain damage is exactly 30 mg/dL.Brain injury depends on several factors, including the duration of low blood glucose, the concentration of other energy-providing metabolites such as ketone bodies and lactate, and the energy reserves in the brain.Therefore, multicenter studies are necessary to progressively evaluate the results of different intervention thresholds. JAMA Network Open.2024;7(3):e243683.doi:10.1001/jamanetworkopen.2024.3683(Reprinted) March 26, 2024 9/13 Downloaded from jamanetwork.comby guest on 03/31/2024 physiological nadir.