Abstract
Objective:
To determine the relationship between transient neonatal hypoglycemia in at-risk infants and neurocognitive function at 6–7 years’ corrected age.
Study design:
The pre-hPOD Study involved children born with at least one risk factor for neonatal hypoglycemia. Hypoglycemia was defined as ≥1 consecutive blood glucose concentrations <2.6 mmol/L, severe as <2.0 mmol/L, mild as ≥2.0 mmol/L and <2.6 mmol/L, brief as 1 to 2 episodes, and recurrent as ≥3 episodes. At 6–7y children were assessed for cognitive and motor function (NIH-Toolbox), learning, visual perception and behavior. The primary outcome was neurocognitive impairment, defined as >1 SD below the normative mean in ≥1 Toolbox tests. The eight secondary outcomes covered children’s cognitive, motor, language, emotional-behavioral, and visual perceptual development. Primary and secondary outcomes were compared between children who did and did not experience neonatal hypoglycemia, adjusting for potential confounding by gestation, birthweight, sex and receipt of prophylactic dextrose gel (pre-hPOD intervention). Secondary analysis included assessment by severity and frequency of hypoglycemia.
Results:
Of 392 eligible children, 315(79%) wereassessed at school age (primary outcome 308), 47% of whom experienced hypoglycemia. Neurocognitive impairment was similar between exposure groups (hypoglycemia 51% vs. 50% no hypoglycemia; aRD −4%, 95%CI −15%, 7%). Children who experienced severe or recurrent hypoglycemia had worse visual motion perception and increased risk of emotional-behavioral difficulty.
Conclusion:
Exposure to neonatal hypoglycemia was not associated with risk of neurocognitive impairment at school-age in at-risk infants, but severe and recurrent episodes may have adverse impacts.
Trial registration:
Keywords: infant newborn, hypoglycemia, cognition
Introduction
Neonatal hypoglycemia is a common metabolic problem in neonates and a preventable cause of neurodevelopmental impairment and learning problems in childhood.1 Prolonged, symptomatic neonatal hypoglycemia is known to cause neonatal brain injury,2 but the long-term significance of asymptomatic transitional neonatal hypoglycemia is less clear due to a limited number of high-quality follow-up studies.3 Importantly, the effects of neonatal hypoglycemia may only become apparent as higher cognitive functions emerge in later childhood;4 thus, assessment at school age is critical. In a retrospective study of universal neonatal blood glucose screening, children exposed to brief transitional hypoglycemia (≥1 episodes <2.5 mmol/L [<45 mg/dL] in the first 48 hours), compared to those with an initial blood glucose concentration (BGC) ≥2.5 mmol/L [≥45 mg/dL], were 1.6 times more likely to have low literacy (below grade level) at 10 years of age.5 When the definition of hypoglycemia used in analysis was decreased to <1.9 mmol/L [<35 mg/dL], children exposed to transient hypoglycemia, compared to those with an initial BGC ≥1.9 mmol/L [≥35 mg/dL], were two times more likely to have low literacy and low numeracy, despite this being the typical clinical threshold for treatment. However, in a prospective study of infants born at risk of hypoglycemia who were carefully screened and treated to maintain BGC ≥2.6 mmol/L [≥47 mg/dL], no association was seen between hypoglycemia and low academic achievement at 9–10 years of age, measured using a standardized curriculum-based assessment, suggesting this was an adequate operational treatment threshold.6 Nevertheless, in this cohort, neonatal hypoglycemia, both mild and severe, was associated with impaired cortical and basal ganglia development at 9 to 10 years of age compared to children born at risk who did not develop hypoglycemia.7 Thus, further evidence is needed concerning the long-term effects of different approaches to screening and treatment of neonatal hypoglycemia on neurocognitive functioning at school age.
We assessed associations between neonatal hypoglycemia and neurocognitive function at 6 to 7 years of age in children who participated in the hypoglycemia Prevention with Oral Dextrose [pre-hPOD] Study, 47% of whom developed episodes of low blood glucose <2.6 mmol/L [<47 mg/dL] after birth. We hypothesized that children exposed to neonatal hypoglycemia would have worse neurocognitive function at 6 to 7 years’ corrected age than those not so exposed, with increased risk associated with severe or recurrent episodes.
Methods
Participants
This cohort analysis was performed as part of the pre-hPOD Early School-age Outcomes Study which followed children in the pre-hPOD Study at 6 to 7 years’ corrected age. Pre-hPOD was a randomized, double-blind, placebo-controlled dose-finding trial of buccal dextrose gel to prevent neonatal hypoglycemia, performed at two hospitals in New Zealand (ACTRN12613000322730).8 It recruited a total of 415 infants born at risk of neonatal hypoglycemia from August 2013 to November 2014. To be eligible, infants had to have a birthweight ≥2.2 kg with no apparent indication for neonatal intensive care unit (NICU) admission, their mother had to be intending to breastfeed, and they had to have one or more risk factors for hypoglycemia, including being the infant of a mother with diabetes or being born late preterm (35 or 36 weeks’ gestation), small (birthweight <10th centile on population or customized birthweight charts or <2.5 kg) or large (birthweight >90th centile on population or customized birthweight charts or >4.5 kg). All infants were enrolled and randomized within 1 hour of birth to one of eight treatment groups: buccal dextrose gel (40%) 200 mg/kg at 1 hour of age, 400 mg/kg at 1 hour of age, 200 mg/kg at 1 hour and then before three subsequent feeds (total 800 mg/kg), 400 mg/kg at 1 hour and 200 mg/kg before three subsequent feeds (total 1000 mg/kg), and four corresponding equivolume placebo groups.9 Each dose of gel was followed by a breastfeed.
At 6 to 7 years’ corrected age, all surviving participants who had not been previously withdrawn (N=392) were invited to take part in the pre-hPOD Early School-age Outcomes Study. Parents/caregivers provided written informed consent and the children gave written assent. Ethical approval was provided by the Southern Health and Disability Ethics Committee (19/STH/202). This report adheres to the STROBE Checklist (Supplementary Materials).
Assessments
As part of the pre-hPOD Early School-age Outcomes Study, children underwent a comprehensive assessment at 6 to 7 years’ corrected age by trained assessors at school, a research clinic or the child’s home. Assessors were blinded to the child’s neonatal glycemic history. Five tests from the tablet-based National Institutes of Health (NIH) Toolbox Cognition Battery were used to assess episodic memory (Picture Sequence Memory Test),9 general vocabulary knowledge (Picture Vocabulary Test), reading decoding ability (Oral Reading Recognition Test),10 attention and inhibitory control (Flanker Test), and cognitive flexibility (Dimensional Change Card Sort Test).11 The NIH Toolbox Motor Battery was used to assess manual dexterity (Nine-hole Pegboard Test) and static and dynamic standing balance (Standing Balance Test).12 Age-adjusted scale scores and an Early Childhood Composite Score (a global measure of cognitive function based on the Picture Sequence Memory, Picture Vocabulary, Flanker and Dimensional Change Card Sort Tests) were provided by the Toolbox software, with normative mean of 100 and standard deviation (SD) of 15 (higher scores indicate higher levels of ability).
Global motion perception, a reflection of the integrity of the dorsal visual stream of the brain, was assessed by the Motion Coherence Test,13 which generates a motion coherence threshold as a percentage of signal to noise, with lower percentages indicating better visual perception. Numeracy was assessed with the Checkout Game, which measures the ability to form sets, numeral recognition, pattern recognition, number sequence counting, and mental operations.14 The total maximum score is 32 and low numeracy was defined as a score <25. Emotional and behavioral development was assessed by the Strength and Difficulties Questionnaire (SDQ, Supplementary Methods).15 Teachers completed a questionnaire about children’s learning profiles and academic performance at school (Supplementary Methods).16
Outcomes
The primary outcome was neurocognitive impairment defined as a standard score <85 in one or more Toolbox tests (>1 SD below the normative mean). Eight critical secondary outcomes were defined: executive dysfunction (Flanker or Dimensional Change Card Sort Test standard score <85); motor impairment (Pegboard Dexterity or Standing Balance Test standard score <85); language impairment (Picture Vocabulary or Oral Reading Recognition Test standard score <85); episodic memory impairment (Picture Sequence Memory Test standard score <85); composite cognitive score (Early Childhood Composite Score, or if this was unavailable, the mean of the scores for Picture Vocabulary, Flanker, Dimensional Change Card Sort and Picture Sequence Memory Tests); low numeracy (Checkout Game total score <25); Motion Coherence Threshold (%); emotional-behavioral difficulty (SDQ Total Difficulties Score ≥14). A range of tertiary outcomes were reported for completeness (Supplementary Methods). For children who were unable to complete testing due to known or presumed developmental impairment, the following scores were assigned: Toolbox scale score 55 (3 SD below the normative mean); motion coherence threshold 100%; numeracy score 10.
Exposures
A hypoglycemic episode was defined as ≥1 consecutive blood glucose concentrations <2.6 mmol/L in the first 48 hours after birth and was further classified as severe (<2.0 mmol/L) or mild (≥2.0 mmol/L and <2.6 mmol/L). The frequency of hypoglycemia was defined as brief (1 to 2 episodes) or recurrent (≥3 episodes).
Statistical analysis
All analyses were performed according to a pre-specified statistical analysis plan using JMP version 17 or SAS version 9.4 (SAS Institute Inc, Cary NC, USA). In the primary analysis, outcomes for children exposed to any neonatal hypoglycemia were compared to those without neonatal hypoglycemia using generalized linear models, adjusted for potential confounding by gestation length, birthweight z-score, sex and receipt of dextrose gel as part of the pre-hPOD Study intervention (Model 1). Exposure effects are presented as adjusted risk difference (aRD), mean difference (aMD) or ratio of geometric means (aRGM), with 95% confidence intervals (CI). For categorical data, an adjusted risk ratio (aRR) was also estimated, with 95% CI. A hypothesis test was performed for the primary outcome (two-tailed α=0.05). We estimated that the study would have 80% power to detect an increase in the proportion of children with the primary outcome from 38% to 54%.
In secondary analyses of the primary and secondary outcomes, we planned two exploratory models. In the first (Model 2), the primary model was additionally adjusted for the New Zealand Deprivation Index (NZDep) an index of socioeconomic deprivation by geographical area divided by decile.17 Deprivation was not included in the primary model because it could act as a confounder or modifier of the relationship between neonatal hypoglycemia and neurocognitive function, with potentially opposing influences on effect estimates; therefore, it was judged that the covariate effect of deprivation should be assessed separately. In the second exploratory model (Model 3), the primary model was additionally adjusted for maternal diabetes. Maternal diabetes is likely to be a confounder of the relationship between neonatal hypoglycemia and neurocognitive function, as there is increasing evidence that exposure to a diabetic milieu in utero may adversely affect brain maturation and long-term neurodevelopment independent of any effects of neonatal hypoglycemia.18 However, among at-risk infants, those exposed to gestational diabetes, compared to other risk factors, are generally less likely to develop severe hypoglycemia.19 Therefore, the covariate effect of diabetes is uncertain and could contribute to colliding, so it was judged that this should also be examined separately.
Secondary analyses also included examination of the relationship between the primary and secondary outcomes and the severity of hypoglycemia (≥1 severe episode or only mild episodes vs. no hypoglycemia) and the frequency of hypoglycemia (recurrent [≥3 episodes] or brief [1 or 2 episodes] hypoglycemia vs. no hypoglycemia), using the primary model. In these analyses, family-wise error rate was controlled with Dunnett correction. Finally, the influence of sex on the association between neonatal hypoglycemia and neurocognitive function was explored by subgroup analysis and a test of interaction, also using the primary model.
Results
Of the 415 infants in the pre-hPOD Study, 392 were eligible for early school-age follow-up (22 withdrew, 1 death), 315 (80%) underwent assessment at a mean (SD) corrected age of 6.8 (0.3) years and were included in the final analysis (Supplementary Figure). Of these, 147 (47%) children experienced hypoglycemia (112 [76%] mild, 35 [24%] severe; 113 [77%] 1–2 episodes, 34 [23%] recurrent episodes; 16 [11%] severe and recurrent). Two hundred and twelve (67%) children were randomized to prophylactic buccal dextrose gel and 103 (23%) to placebo gel. In comparison to children who were not assessed (N=100), children who were assessed were born slightly earlier (mean [SD] 38.3 [1.1] vs. 38.6 [1.2] weeks’ gestation, P=0.01), their mothers were older (32.7 [5.5] vs. 30.8 [4.9] years, P<0.01) and were more likely to have tertiary education (P=0.02) at study entry (Supplementary Table S1). Among assessed children, those who experienced hypoglycemia, comparing to those who did not, were born slightly earlier (38.0 [1.0] vs. 38.5 [1.1] weeks’ gestation, P <0.001), were more likely to require NICU admission (number [percent] 21 [14%] vs. 2 [1%], P <0.01), receive intravenous dextrose (15 [10%] vs. 2 [1%], P <0.01), be born by caesarean (81 [55%] vs. 73 [44%], P=0.04) and have reduced intensity and duration of breastfeeding (Table 1).
Table 1.
Characteristics of children, and their mothers, in the pre-hPOD cohort who were assessed at school age.
| No hypoglycemia | Hypoglycemia | P | |
|---|---|---|---|
| Age—years | 32.3 (5.3) | 33.3 (5.7) | 0.10 |
| BMI at booking—kg/m2 | 28.4 (7.9) | 29.2 (8.0) | 0.45 |
| Nulliparous | 68 (41%) | 72 (49%) | 0.14 |
| Caesarean birth | 73 (44%) | 81 (55%) | 0.04 |
| Highest education level | 0.42 | ||
| Diabetes | 122 (73%) | 104 (71%) | 0.80 |
| Pre-eclampsia | 6 (4%) | 5 (3%) | 1.00 |
| Randomized to prophylactic buccal dextrose gel | 121 (72%) | 91 (62%) | 0.07 |
| Female | 89 (53%) | 67 (46%) | 0.21 |
| Gestation—weeks | 38.5 (1.1) | 38.0 (1.0) | <0.001 |
| Birthweight—g | 3287 (586) | 3177 (605) | 0.10 |
| Birthweight z-score | 0.19 (1.24) | 0.11 (1.28) | 0.56 |
| Multiple pregnancy | 10 (6%) | 18 (12%) | 0.07 |
| High deprivation | 64 (38%) | 50 (34%) | 0.48 |
| Prioritized ethnicity | 0.46 | ||
| Admission to NICU | 2 (1%) | 21 (14%) | <0.001 |
| Neonatal intravenous dextrose | 2 (1%) | 15 (10%) | <0.001 |
| Exclusively breastfed to 6 weeks | 64 (39%) | 18 (13%) | <0.001 |
| Continued breastfeeding at 6 weeks | 149 (91%) | 115 (82%) | <0.03 |
| Formula use by day 3 of age | 45 (27%) | 89 (61%) | <0.001 |
| Age at follow-up—years | 6.84 (0.30) | 6.81 (0.33) | 0.44 |
Data are number (percent) or mean (standard deviation). BMI, body mass index; NICU, neonatal intensive care unit. Missing data: maternal BMI/obesity 35 (no hypoglycemia 26; hypoglycemia 9); highest maternal education 41 (no hypoglycemia 23; hypoglycemia 18); exclusively breastfed to 6 weeks (no hypoglycemia 5; hypoglycemia 3); continued breastfeeding at 6 weeks (no hypoglycemia 4; hypoglycemia 7); formula use by day 3 of age (no hypoglycemia 1; hypoglycemia 1). P value is for the comparison of children exposed to neonatal hypoglycemia vs those not exposed to neonatal hypoglycemia (Fisher exact test or Student’s t test).
Primary analyses
Among 308 children for whom primary outcome data were available, the risk of neurocognitive impairment was not statistically significantly different between children who did and did not experience hypoglycemia, with no apparent clinically important difference between groups (51% vs. 50%; aRD −4%, 95% CI −15%, 7%; aRR 0.94, 95% CI 0.75, 1.17) (Table 2). Similarly, across the eight secondary outcomes (Table 2) and 15 tertiary outcomes (Supplementary Table S2), measures were broadly similar between groups, with no statistical evidence of differences between groups.
Table 2.
Primary and secondary outcomes at school age
| No hypoglycemia | N | Hypoglycemia | N | Model 1 aRD, aMD, aRGM (95% CI) [aRR (95%CI)] |
Model 2 aRD, aMD, aRGM (95% CI) [aRR (95%CI)] |
Model 3 aRD, aMD, aRGM (95% CI) [aRR (95%CI)] |
|
|---|---|---|---|---|---|---|---|
| Neurocognitive impairment | 82 (51%) | 161 | 73 (50%) | 147 | −4% (−15, 7) [0.94 (0.75, 1.17)] P=0.49 |
−3% (−14, 8) [0.97 (0.78, 1.20)] |
−4% (−15, 7) [0.95 (0.76, 1.17)] |
| Executive dysfunction | 19 (12%) | 161 | 18 (12%) | 147 | −4% (−12, 4) [0.89 (0.49, 1.61)] |
−2% (−11, 8) [0.87 (0.48, 1.56)] |
−2% (−11, 6) [0.84 (0.47, 1.52)] |
| Motor impairment | 13 (8%) | 161 | 9 (6%) | 147 | −4% (−13, 6) [0.65 (0.29, 1.43)] |
−3% (−13, 6) [0.64 (0.29, 1.41)] |
−4% (−14, 7) [0.64 (0.29, 1.41)] |
| Language impairment | 23 (14%) | 161 | 25 (17%) | 147 | −4% (−12, 5) [1.10 (0.65, 1.86)] |
−2% (−10, 6) [1.12 (0.67, 1.87)] |
−3% (−11, 5) [1.10 (0.66, 1.83)] |
| Episodic memory impairment | 18 (11%) | 160 | 10 (7%) | 147 | −5% (−11, 2) [0.60 (0.28, 1.30)] |
−5% (−12, 2) [0.61 (0.28, 1.31)] |
−5% (−12, 3) [0.61 (0.28, 1.32)] |
| Composite cognitive score | 105 (15) | 161 | 104 (16) | 147 | 1 (−2, 5)aMD | 1 (−2, 4) aMD | 1 (−2, 5) aMD |
| Low numeracy | 16 (10%) | 161 | 21 (14%) | 147 | 2% (−8, 11) [1.15 (0.61, 2.14)] |
−1% (−11, 9) [1.20 (0.65, 2.22)] |
1% (−9, 12) [1.13 (0.61, 2.08)] |
| Motion Coherence Threshold | 15 (9, 24) | 155 | 16 (11, 31) | 146 | 1.14 (0.94, 1.39)aRGM | 1.14 (0.94, 1.39)aRGM | 1.14 (0.94, 1.39)aRGM |
| Emotional-behavioral difficulty | 14 (12%) | 119 | 18 (17%) | 109 | 0% (−10, 11) [1.34 (0.69, 2.59)] |
8% (−5, 21) [1.24 (0.65, 2.37)] |
0% (−10, 11) [1.36 (0.70, 2.64)] |
Data are number (percent), mean (standard deviation) or median (interquartile range). aRD, adjusted risk difference; aMD, adjusted mean difference; aRGM, adjusted ratio of geometric means; aRR, adjusted risk ratio (relative risk). Exposure effect estimates are for hypoglycemia vs. no hypoglycemia. Model 1 adjusted for sex, gestation length, birthweight z-score and prophylactic dextrose gel as trial intervention. Model 2 adjusted for sex, gestation length, birthweight z-score, prophylactic dextrose gel as trial intervention and deprivation index. Model 3 adjusted for sex, gestation length and birthweight z-score, prophylactic dextrose gel as trial intervention and maternal diabetes.
Secondary analyses
Additional adjustment for NZDep (Model 2) and maternal diabetes (Model 3) did not alter results for the primary and secondary outcomes (Table 2).
The Motion Coherence Thresholds of children who experienced severe or recurrent hypoglycemia were higher than those did not experience hypoglycemia (48% and 45% increase, respectively; Table 3 & Table 4), indicating poorer visual perception. The risk of emotional-behavioral difficulty was very likely higher in children who were experienced recurrent hypoglycemia than in those who did not experience hypoglycaemia, particularly given the large absolute risk difference (33% vs. 12%, aRD 18%, 95% CI −7, 43; aRR 2.57, 95% CI 1.07, 6.17; Table 4). The severity and frequency of hypoglycemia was not associated with the primary outcome or other secondary outcomes.
Table 3.
Primary and secondary outcomes and severity of neonatal hypoglycemia
| No hypoglycemia | N | Mild hypoglycemia | N | aRD or aMD (95% CI) [aRR 95%CI] |
Severe hypoglycemia | N | aRD or aMD (95% CI) [aRR 95%CI] |
|
|---|---|---|---|---|---|---|---|---|
| Neurocognitive impairment | 82 (51%) | 161 | 53 (47%) | 112 | −6% (−20, 8) [0.90 (0.68, 1.19)] |
20 (57%) | 35 | 4% (−17, 26) [1.18 (0.82, 1.71)] |
| Executive dysfunction | 19 (12%) | 161 | 11 (10%) | 112 | −6% (−18, 4) [0.75 (0.34, 1.64)] |
7 (20%) | 35 | 3% (−14, 19) [1.46 (0.61, 3.48)] |
| Motor impairment | 13 (8%) | 161 | 3 (3%) | 112 | −7% (−27, 14) [0.30 (0.07, 1.17)] |
6 (17%) | 35 | 7% (−14, 28) [1.44 (0.49, 4.18)] |
| Language impairment | 23 (14%) | 161 | 16 (15%) | 112 | −5% (−14, 6) [0.66 (0.48, 1.88)] |
9 (26%) | 35 | 4% (−15, 24) [1.77 (0.82, 3.48)] |
| Episodic memory impairment | 18 (11%) | 160 | 7 (6%) | 112 | −4% (−14, 5) [0.58 (0.22, 1.55)] |
3 (9%) | 35 | −4% (−16, 9) [0.85 (0.22, 3.32)] |
| Composite cognitive score | 105 (15) | 161 | 106 (15) | 112 | 2 (−2, 6) aMD | 100 (18) | 35 | −4 (−10, 3) aMD |
| Low numeracy | 16 (10%) | 161 | 15 (14%) | 112 | 1% (−10, 13) [1.11 (0.52, 2.38)] |
6 (17%) | 35 | 5% (−15, 24) [1.42 (0.53, 3.76)] |
| Motion Coherence Threshold | 15 (9, 24) | 155 | 16 (10, 25) | 111 | 1.08 (0.85, 1.36)aRGM | 19 (15, 56) | 35 | 1.48 (1.04, 2.10) aRGM |
| Emotional-behavioral difficulty | 14 (12%) | 119 | 12 (14%) | 85 | −3% (−14, 9) [1.01 (0.44, 2.32)] |
6 (25%) | 24 | 9% (−13, 31) [2.03 (0.76, 5.42)] |
Data are number (percent), mean (standard deviation) or median (interquartile range). aRD, adjusted risk difference; aMD, adjusted mean difference; aRGM, adjusted ratio of geometric means; aRR, adjusted risk ratio (relative risk). Exposure effect estimates are for mild/severe hypoglycemia vs. no hypoglycemia, adjusted for sex, gestation length, birthweight z-score and prophylactic dextrose gel as trial intervention, with Dunnett correction of family-wise error. Higher Motion Coherence Threshold indicates worse visual processing. A higher cognitive score indicates better cognitive function.
Table 4.
Primary and secondary outcomes and frequency of neonatal hypoglycemia
| No hypoglycemia | N | Brief hypoglycemia | N | aRD or aMD (95% CI) [aRR 95%CI] |
Recurrent hypoglycemia | N | aRD or aMD (95% CI) [aRR 95%CI] |
|
|---|---|---|---|---|---|---|---|---|
| Neurocognitive impairment | 83 (51%) | 164 | 55 (50%) | 110 | −3% (−17, 11) [0.96 (0.73, 1.26)] |
17 (50%) | 34 | −5% (−27, 17) [0.96 (0.63, 1.46)] |
| Executive dysfunction | 19 (12%) | 164 | 10 (9%) | 110 | −6% (−18, 6) [0.68 (0.30, 1.55] |
8 (24%) | 34 | 6% (−12, 24) [1.63 (0.71, 3.72)] |
| Motor impairment | 14 (9%) | 164 | 5 (5%) | 110 | −5% (−19, 9) [0.51 (0.17, 1.57)] |
3 (9%) | 34 | −3% (−22, 17) [0.68 (0.18, 2.53)] |
| Language impairment | 23 (14%) | 164 | 18 (16%) | 110 | −3% (−14, 7) [1.10 (0.58, 2.10)] |
7 (21%) | 34 | 0% (−18, 18) [1.28 (0.59, 2.80)] |
| Episodic memory impairment | 18 (11%) | 163 | 9 (8%) | 110 | −2% (−11, 7) [0.76 (0.31, 1.84)] |
1 (3%) | 34 | −9% (−28, 9) [0.26 (0.03, 2.53)] |
| Composite cognitive score | 105 (15) | 164 | 106 (15) | 110 | 2 (−2, 7) aMD | 99 (17) | 34 | −4 (−11, 3) aMD |
| Low numeracy | 16 (10%) | 164 | 13 (12%) | 110 | 0% (−12, 11) [1.01 (0.46, 2.21)] |
8 (24%) | 34 | 8% (−9, 28) [1.72 (0.70, 4.20)] |
| Motion Coherence Threshold | 15 (9, 24) | 158 | 16 (11, 24) | 109 | 1.08 (0.86, 1.37)aRGM | 22 (22, 54) | 34 | 1.45 (1.01, 2.07)aRGM |
| Emotional-behavioral difficulty | 15 (12%) | 121 | 10 (12%) | 86 | −4% (−16, 7) [0.88 (0.37, 2.09)] |
7 (33%) | 21 | 18% (−7, 43) [2.57 (1.07, 6.17)] |
Data are number (percent), mean (standard deviation) or median (interquartile range). aRD, adjusted risk difference; aMD, adjusted mean difference; aRGM, adjusted ratio of geometric means; aRR, adjusted risk ratio (relative risk). Exposure effect estimates are for brief/recurrent hypoglycemia vs. no hypoglycemia, adjusted for sex, gestation length, birthweight z-score and prophylactic dextrose gel as trial intervention, with Dunnett correction of family-wise error. Higher Motion Coherence Threshold indicates worse visual processing. A higher cognitive score indicates better cognitive function.
There was some evidence that the association between hypoglycemia and language impairment may be influenced by sex (P=0.02 for interaction), with a higher rate in boys who experienced hypoglycemia, compared to boys who did not, although there was insufficient statistical evidence to be confident of a difference between exposure groups in either sex (Supplementary Table S3). Infant sex did not influence the association between neonatal hypoglycemia and the primary outcome or other secondary outcomes.
In post hoc analysis, additional adjustment of the primary model for continued breastfeeding at 6 weeks did not alter results for the primary and secondary outcomes.
Discussion
In this cohort of children born at risk, transitional neonatal hypoglycemia was not associated with neurocognitive impairment at school age, regardless of severity and frequency of hypoglycemic episodes. However, compared to at-risk children with did not experience neonatal hypoglycemia, those who experienced severe or recurrent episodes had worse global motion perception and those who experienced recurrent episodes very likely had increased emotional-behavioral difficulty.
Transitional hypoglycemia is a common neonatal metabolic problem, affecting up to 15% of normal newborns and 50% of infants with risk factors, including preterm birth, maternal diabetes, and small or large size for gestational age.20 Neonatal hypoglycemia can cause brain injury as glucose is a critical energy source for the brain after birth,1 and this may have long-lasting adverse effects on children’s neurodevelopment, including visual-motor impairment, executive dysfunction, cognitive impairment and reduced academic achievement.3 Notably, these effects may not be apparent until later in childhood when higher cognitive functions and skills emerge. For example, neonatal hypoglycemia has been associated with executive dysfunction and visual-motor problems at 4.5 years’ corrected age4 but not at 2 years.21 Thus, assessment of the impact of neonatal hypoglycemia and it’s management requires evaluation of children through to school age.
We undertook this analysis because of conflicting results in two previous school-age outcome cohort studies in neonatal hypoglycemia,5,6 although this may have been due to differences in how the infants were screened and treated. In the Children With Hypoglycemia and Their Later Development (CHYLD) Study, at-risk infants were carefully screened for hypoglycemia for at least the first 24 hours and treated to maintain BGC >2.6 mmol/L [>47 mg/dL], whereas in the study by Kaiser et al., infants did not have ongoing screening if the first or second BCG was normal, potentially missing subsequent episodes, and infants were generally only treated if the hypoglycemia was severe (<1.9 mmol/l [<35 mg/dL]). Taken together, these cohorts and the present study, suggest that 2.6 mmol/L [47 mg/dL] is an adequate operational blood glucose threshold for initiating treatment among at-risk late preterm and term infants. However, given the secondary findings in this study, some caution is required and exposure to a high burden of hypoglycemia, either severity or frequency, should be avoided. Moreover, two recent large cohort studies found that severe hypoglycemia, especially of early onset, was associated with neurodevelopmental impairment in preschool children.22,23 Thus, further research is needed to optimize screening, prevention and treatment approaches in at-risk infants to minimize exposure to hypoglycemia while also avoiding iatrogenic harm.
Avoiding adverse effects of neonatal hypoglycemia, and its management, on long-term development is important, particularly given the high prevalence of neurocognitive impairment at school age (~50%, moderate-severe ~ 25%) in the children in this study, reflecting difficulties across a range of outcomes, including executive dysfunction, low language and numeracy, and emotional-behavioral difficulty. Consequently, these children are at risk of low educational achievement with ≥30% already judged by their teachers at 6 to 7 years of age to be below or well below the expected curriculum level. This risk increased to nearly 50% by 9 to 10 years of age in the CHYLD cohort, both by teacher judgement and standardized curriculum-based assessments, indicating that the reasons infants are considered at risk of hypoglycemia has substantial impact on their long-term development, aside from any problems with glucose regulation after birth.18,24,25 Understanding the effect of different fetal exposures on later development and learning, including potential interventions, such as preterm birth prevention and earlier detection and management of maternal diabetes, is an equally important research priority in neonatal hypoglycemia.
The finding of worse global motion perception in children who experienced severe or recurrent episodes is consistent with previous reports of occipital-parietal lobe injuries following neonatal hypoglycemia,26–28 although there has been debate about the extent to which mild to moderate neonatal hypoglycemia affects visual function, as well as the degree to which associations may be confounded by comorbidities.29 Neonatal hypoglycemia has been linked to injury to the optic radiations,26 which connect the lateral geniculate nucleus (LGN) to the primary visual cortex. The optic radiations include projections from the magnocellular layers of LGN that innervate the dorsal visual cortical processing pathway and support global motion perception. Global motion perception in 7-year old children born very preterm is also associated with neonatal nutrition suggesting that the dorsal processing stream is sensitive to the perinatal environment.30 It is unclear whether the differences in Motion Coherence Threshold seen at 6 to 7 years of age in this study represent an impairment of the dorsal stream or a delay in maturation, as no association was seen between neonatal hypoglycemia and global motion perception at 9 to 10 years of age in children in the CHYLD Study.6 Nevertheless, given the concomitant finding of increased behavioral difficulties in children exposed to recurrent hypoglycemia seen in this study, it is interesting that reductions in dorsal stream function are correlated with autistic traits in children born at risk of neonatal hypoglycemia.31 Moreover, neonatal hypoglycemia has been associated with smaller caudate volume in school-age children exposed to neonatal hypoglycemia, which, in turn, was associated greater parent-reported emotional and behavioral difficulties, and poorer prosocial behavior.32
The increased risk of language impairment in boys following neonatal hypoglycemia, in contrast to girls, could potentially be attributed to the well-established sex differences in language development, with girls acquiring language faster than males,33,34 and having advantage in several basic linguistic domains by mid-childhood.35,36 Boys had greater variability in language development and a higher overrepresentation at the lower end of the language proficiency spectrum.36,37 Therefore, any disruptions in language development caused by exposure to hypoglycemia may be more apparent in boys.
The strengths of our study include prospective design, adjustment for potential confounding, a comprehensive school-based assessment across multiple domains, high follow-up rate and the near-complete data collection for most measures. Limitations of this study include being a secondary analysis of a trial cohort with a modest sample size and imprecise exposure effect estimates for some outcomes, such that smaller but potentially clinically important differences cannot be excluded. The findings for visual perception and behavior could reflect type 1 error, given these were secondary analyses, but we sought to reduce type 1 error by prespecifying a limited number of secondary outcomes considered important based on prior literature,1 and by taking account of effect size, precision and consistency of findings. We did not employ methods to correct for multiple analyses due to increased risk of type 2 error, which is equally important to avoid when evaluating long-term safety. Ongoing school-age follow-up of the larger hPOD Trial,38 which employs the same assessment, may help to clarify the relationship between neonatal hypoglycemia and visual and behavioral development. Finally, this study focused only on children born at risk of hypoglycemia, with a high proportion at risk due to maternal diabetes, and the findings may not apply to the general neonatal population.
Conclusion
Among children born at risk of neonatal hypoglycemia, who were screened and treated to maintain BGC ≥2.6 mmol/L, the rate of neurocognitive impairment at 6 to 7 years’ corrected age was not different between those who did and did not experience neonatal hypoglycemia. Exposure to severe and recurrent episodes may be associated with specific adverse neurocognitive effects.
Supplementary Material
Acknowledgements
Our gratitude goes out to the children and families who participated in the pre-hPOD Study. We also acknowledge the following members of the pre-hPOD Early School-age Outcomes Study Group as non-author contributors: Coila Bevan, Frank Bloomfield, Nataliia Burakevych, J Geoffrey Chase, Caroline Crowther, Darren Dai, Richard Edlin, Rebecca Griffiths, Jo Hegarty, Olga Ivashkova, Peter Kegan, Rachel Lamdin, Jocelyn Ledger, Stephanie Macdonald, Anna Mikaelian, David Nyakotey, Hannah Park and Rajesh Shah.
Funding:
The pre-hPOD Study was funded by Lottery Health Research (241266), Cure Kids (3561), philanthropic donations to the University of Auckland Foundation (F-ILG- LRSR), Health Research Council of New Zealand (15-216), Gravida, National Centre for Growth and Development (SCH-14-14), Auckland Medical Research Foundation (1113012), A+ Trust (5696) and Eunice Kennedy Shriver National Institute of Child Health & Human Development, National Institutes of Health (R01HD091075). The pre-hPOD Early School-age Outcomes Study was funded by the Health Research Council of New Zealand (19/960) and the Eunice Kennedy Shriver National Institute of Child Health & Human Development, National Institutes of Health (R01HD091075). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Abbreviations:
- aMD
adjusted Mean Difference
- aRD
adjusted Risk Difference
- aRGM
adjusted ratio of geometric means
- aRR
adjusted Relative Risk
- BGC
blood glucose concentration
- CI
Confidence Interval
- NICU
Neonatal Intensive Care Unit
- Pre-hPOD Study
the hypoglycemia Prevention with Oral Dextrose Study
- SD
Standard Deviation
Footnotes
Disclosures: The authors have no conflicts of interest relevant to this article to disclose.
Clinical Trial registry name: Hypoglycemia Prevention in Newborns with Oral Dextrose: the Dosage Trial (pre-hPOD Study).
Registration number: ACTRN12613000322730.
Data sharing statement:
Deidentified individual participant data (including data dictionaries) will be made available, in addition to study protocols, the statistical analysis plan, and the informed consent form. The data will be made available upon publication to researchers who provide a methodologically sound proposal for use in achieving the goals of the approved proposal, according to the data sharing protocol of the University of Auckland (https://research-hub.auckland.ac.nz/subhub/human-health-research-services-platform). Proposals should be submitted tohumanhealth@auckland.ac.nz
References
- 1.Alsweiler JM, Harris DL, Harding JE, McKinlay CJD. Strategies to improve neurodevelopmental outcomes in babies at risk of neonatal hypoglycaemia. Lancet Child Adolesc Health. 2021;5(7):513–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Burns CM, Rutherford MA, Boardman JP, Cowan FM. Patterns of cerebral injury and neurodevelopmental outcomes after symptomatic neonatal hypoglycemia. Pediatrics. 2008;122(1):65–74. [DOI] [PubMed] [Google Scholar]
- 3.Shah R, Harding J, Brown J, McKinlay C. Neonatal glycaemia and neurodevelopmental outcomes: a systematic review and meta-analysis. Neonatology. 2019;115(2):116–26. [DOI] [PubMed] [Google Scholar]
- 4.McKinlay CJD, Alsweiler JM, Anstice NS, Burakevych N, Chakraborty A, Chase JG, et al. Association of neonatal glycemia with neurodevelopmental outcomes at 4.5 years. JAMA Pediatr. 2017;171(10):972–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kaiser JR, Bai S, Gibson N, Holland G, Lin TM, Swearingen CJ, et al. Association between transient newborn hypoglycemia and fourth-grade achievement test proficiency: a population-based study. JAMA Pediatr. 2015;169(10):913–21. [DOI] [PubMed] [Google Scholar]
- 6.Shah R, Dai DWT, Alsweiler JM, Brown GTL, Chase JG, Gamble GD, et al. Association of neonatal hypoglycemia with academic performance in mid-childhood. JAMA. 2022;327(12):1158–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Nivins S, Kennedy E, Thompson B, Gamble GD, Alsweiler JM, Metcalfe R, et al. Associations between neonatal hypoglycaemia and brain volumes, cortical thickness and white matter microstructure in mid-childhood: An MRI study. Neuroimage Clin. 2022;33:102943. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Hegarty JE, Harding JE, Gamble GD, Crowther CA, Edlin R, Alsweiler JM. Prophylactic oral dextrose gel for newborn babies at risk of neonatal hypoglycaemia: a randomised controlled dose-finding trial (the Pre-hPOD Study). PLoS Med. 2016;13(10):e1002155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Dikmen SS, Bauer PJ, Weintraub S, Mungas D, Slotkin J, Beaumont JL, et al. Measuring episodic memory across the lifespan: NIH Toolbox Picture Sequence Memory Test. J Int Neuropsychol Soc. 2014;20(6):611–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Gershon RC, Cook KF, Mungas D, Manly JJ, Slotkin J, Beaumont JL, et al. Language measures of the NIH Toolbox Cognition Battery. J Int Neuropsychol Soc. 2014;20(6):642–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Zelazo PD, Anderson JE, Richler J, Wallner-Allen K, Beaumont JL, Weintraub S. II. NIH Toolbox Cognition Battery (CB): measuring executive function and attention. Monogr Soc Res Child Dev. 2013;78(4):16–33. [DOI] [PubMed] [Google Scholar]
- 12.Reuben DB, Magasi S, McCreath HE, Bohannon RW, Wang Y-C, Bubela DJ, et al. Motor assessment using the NIH toolbox. Neurology. 2013;80(10):S65–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Thompson B, McKinlay CJD, Chakraborty A, Anstice NS, Jacobs RJ, Paudel N, et al. Global motion perception is associated with motor function in 2-year-old children. Neurosci Lett. 2017;658:177–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hattie J, Brown G, Earl Irving S. An analysis of an assessment tool for 5-year old students entering elementary school: the School Entry Assessment Kit. N Z J Educ Stud. 2015;50(1):87–105. [Google Scholar]
- 15.Goodman R The Strengths and Difficulties Questionnaire. J Child Psych Psychiat. 1997;38:581–6. [DOI] [PubMed] [Google Scholar]
- 16.Shah R, Brown G, Keegan P, Burakevych N, Harding JE, McKinlay CJ, et al. Teacher rating versus measured academic achievement: Implications for paediatric research. J Paed Child Health. 2020;56(7):1090–6. [DOI] [PubMed] [Google Scholar]
- 17.Atkinson J, Salmond C, Crampton P. NZDep2013 index of deprivation. Wellington: Department of Public Health, University of Otago; 2014. [Google Scholar]
- 18.Adane AA, Mishra GD, Tooth LR. Diabetes in pregnancy and childhood cognitive development: a systematic review. Pediatrics. 2016;137(5). [DOI] [PubMed] [Google Scholar]
- 19.Bailey MJ, Rout A, Harding JE, Alsweiler JM, Cutfield WS, McKinlay CJD. Prolonged transitional neonatal hypoglycaemia: characterisation of a clinical syndrome. J Perinatol. 2020;41(5):1149–57. [DOI] [PubMed] [Google Scholar]
- 20.Harding JE, Harris DL, Hegarty JE, Alsweiler JM, McKinlay CJD. An emerging evidence base for the management of neonatal hypoglycaemia. Early Hum Dev. 2017;104:51–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.McKinlay CJD, Alsweiler JA, Ansell JM, Anstice NS, Chase JG, Gamble GD, et al. Neonatal glycemia and neurodevelopmental outcomes at two years. N Engl J Med. 2015;373:1507–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Edwards T, Alsweiler JM, Gamble GD, Griffith R, Lin L, McKinlay CJD, et al. Neurocognitive outcomes at age 2 years after neonatal hypoglycemia in a cohort of participants from the hPOD randomized trial. JAMA Netw Open. 2022;5(10):e2235989. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Wickström R, Skiöld B, Petersson G, Stephansson O, Altman M. Moderate neonatal hypoglycemia and adverse neurological development at 2–6 years of age. Eur J Epidemiol. 2018;33(10):1011–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Levine TA, Grunau RE, McAuliffe FM, Pinnamaneni R, Foran A, Alderdice FA. Early childhood neurodevelopment after intrauterine growth restriction: a systematic review. Pediatrics. 2015;135(1):126–41. [DOI] [PubMed] [Google Scholar]
- 25.Chan E, Leong P, Malouf R, Quigley MA. Long-term cognitive and school outcomes of late-preterm and early-term births: a systematic review. Child Care Health Dev. 2016;42(3):297–312. [DOI] [PubMed] [Google Scholar]
- 26.Filan PM, Inder TE, Cameron FJ, Kean MJ, Hunt RW. Neonatal hypoglycemia and occipital cerebral injury. J Pediatr. 2006;148(4):552–5. [DOI] [PubMed] [Google Scholar]
- 27.Murakami Y, Yamashita Y, Matsuishi T, Utsunomiya H, Okudera T, Hashimoto T. Cranial MRI of neurologically impaired children suffering from neonatal hypoglycaemia. Pediatr Radiol. 1999;29(1):23–7. [DOI] [PubMed] [Google Scholar]
- 28.Alkalay AL, Flores-Sarnat L, Sarnat HB, Moser FG, Simmons CF. Brain imaging findings in neonatal hypoglycemia: case report and review of 23 cases. Clin Pediatr. 2005;44(9):783–90. [DOI] [PubMed] [Google Scholar]
- 29.Paudel N, Chakraborty A, Anstice N, Jacobs RJ, Hegarty JE, Harding JE, et al. Neonatal hypoglycaemia and visual development: A review. Neonatology. 2017;112(1):47–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Braddick O, Atkinson J, Wattam-Bell J. Normal and anomalous development of visual motion processing: motion coherence and ‘dorsal-stream vulnerability’. Neuropsychologia. 2003;41(13):1769–84. [DOI] [PubMed] [Google Scholar]
- 31.Silva AE, Harding JE, Chakraborty A, Dai DW, Gamble GD, McKinlay CJD, et al. Associations between autism spectrum quotient and integration of visual stimuli in 9-year-old children: preliminary evidence of sex differences. J Autism Dev Disord. 2023. [DOI] [PubMed] [Google Scholar]
- 32.Kennedy E, Nivins S, Thompson B, McKinlay CJD, Harding J, for the Chyld Study Team. Neurodevelopmental correlates of caudate volume in children born at risk of neonatal hypoglycaemia. Pediatr Res. 2022;DOI: 10.1038/s41390-022-02410-3. [DOI] [PubMed] [Google Scholar]
- 33.Galsworthy MJ, Dionne G, Dale PS, Plomin R. Sex differences in early verbal and non-verbal cognitive development. Dev Sci. 2000;3(2):1467–7687. [Google Scholar]
- 34.Wallentin M Putative sex differences in verbal abilities and language cortex: a critical review. Brain Lang. 2009;108(3):175–83. [DOI] [PubMed] [Google Scholar]
- 35.Bornstein MH, Hahn C-S, Haynes OM. Specific and general language performance across early childhood: Stability and gender considerations. First Language. 2004;24(3):267–304. [Google Scholar]
- 36.Euler HA, Lange BP, Zaretsky E. Sex differences in language competence of 3- to 6-year-old children. Appl Psycholinguist. 2016;37(6):1417–38. [Google Scholar]
- 37.Zubrick SR, Taylor CL, Rice ML, Slegers DW. Late language emergence at 24 months: an epidemiological study of prevalence, predictors, and covariates. J Speech Lang Hear Res. 2007;50(6):1562–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Harding JE, Hegarty JE, Crowther CA, Edlin RP, Gamble GD, Alsweiler JM. Evaluation of oral dextrose gel for prevention of neonatal hypoglycemia (hPOD): A multicenter, double-blind randomized controlled trial. PLoS Med. 2021;18(1):e1003411. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Deidentified individual participant data (including data dictionaries) will be made available, in addition to study protocols, the statistical analysis plan, and the informed consent form. The data will be made available upon publication to researchers who provide a methodologically sound proposal for use in achieving the goals of the approved proposal, according to the data sharing protocol of the University of Auckland (https://research-hub.auckland.ac.nz/subhub/human-health-research-services-platform). Proposals should be submitted tohumanhealth@auckland.ac.nz