Introduction: Dextrose gel is widely used as first-line treatment for neonatal hypoglycaemia given its cost-effectiveness and ease of use. The Sugar Babies randomized trial first showed that 40% dextrose gel was more effective in reversing hypoglycaemia than feeding alone. Follow-up of the Sugar Babies Trial cohort at 2 and 4.5 years of age reported that dextrose gel appeared safe, with similar rates of neurosensory impairment in babies randomized to dextrose or placebo gel. However, some effects of neonatal hypoglycaemia may not become apparent until school age. Methods: Follow-up of the Sugar Babies Trial cohort at 9–10 years of age was reported. The primary outcome was low educational achievement in reading or mathematics. Secondary outcomes included other aspects of educational achievement, executive function, visual-motor function, and psychosocial adaptation. Results: Of 227 eligible children, 184 (81%) were assessed at a mean (SD) age of 9.3 (0.2) years. Low educational achievement was similar in dextrose and placebo groups (36/86 [42%] vs. 42/94 [45%]; RR 1.04, 95% CI 0.76, 1.44; p = 0.79). Children allocated to dextrose gel had lower visual perception standard scores (95.2 vs. 100.6; MD −5.68, 95% CI −9.79, −1.57; p = 0.006) and a greater proportion had low (<85) visual perception scores (20/88 [23%] vs. 10/95 [11%]; RR 2.23, 95% CI 1.13, 4.37; p = 0.02). Other secondary outcomes, including other aspects of visual-motor function, were similar in both groups. Conclusion: Treatment dextrose gel does not appear to result in any clinically significant differences in educational achievement or other neurodevelopmental outcomes at mid-childhood.

© 2022 The Author(s). Published by S. Karger AG, Basel

Introduction

Neonatal hypoglycaemia is common, affecting up to 50% of at-risk infants, [1] and is associated with neurodevelopmental impairment [2, 3]. Since hypoglycaemia is usually without clinical signs, it is usually recommended that at-risk infants are screened and treated if hypoglycaemia is detected. The Sugar Babies randomized trial demonstrated that treatment with dextrose gel was more effective at reversing hypoglycaemia than feeding alone, resulting in reduced neonatal intensive care unit (NICU) admission for hypoglycaemia and improved breastfeeding rates [4]. Dextrose gel is now widely used as first-line treatment due to its cost-effectiveness and ease of administration [5].

However, particularly because of widespread use, it is important to determine whether management of hypoglycaemia using dextrose gel has any beneficial or adverse effects on later development. We have previously reported that treating neonatal hypoglycaemia with dextrose gel compared with placebo gel did not alter neurodevelopment at 2 years’ corrected age [6]. At 4.5 years, there were no differences in neurosensory impairment between randomized groups, but scores on the visual perception subscale of the Beery Visual-Motor Integration (VMI) test were lower in the dextrose group [7]. Therefore, to explore whether these differences persisted and if any other later outcomes were influenced by treatment with dextrose gel, we assessed neuropsychological and health outcomes at 9–10 years of children who participated in the Sugar Babies Study.

Materials and MethodsSugar Babies Cohort

The Sugar Babies Study was a randomized trial undertaken at a tertiary centre (Waikato Women’s Hospital) in Hamilton, New Zealand, between December 2008 and November 2010 (Australian New Zealand Clinical Trials Registry: ACTRN 126080000623392) [4]. In brief, eligible babies were born at ≥35 weeks’ gestation, <48 h old, and identified as at risk of hypoglycaemia (infant of diabetic mother, preterm, small (<10th centile or <2500 g), large (>90th centile or >4500 g), or other reason). Blood glucose concentrations were measured using the glucose oxidase method, and babies who became hypoglycaemic (<2.6 mmol/L) were randomized to 0.5 mL/kg of 40% dextrose or placebo gel massaged into the buccal mucosa, followed by a feed. The primary outcome was treatment failure, defined as a blood glucose concentration <2.6 mmol/L 30 min after two treatment attempts.

Mid-Childhood Follow-Up

All surviving participants who had not previously withdrawn were eligible for follow-up. At 9–10 years’ corrected age, children were evaluated by trained assessors who were blinded to the glycaemic history of the child including randomization group. Key domains assessed were academic achievement, executive function, vision and visual-motor function, psychosocial adaption, and general health [8].

In brief, academic achievement was assessed using the Assessment Tools for Teaching and Learning (e-asTTle), a standardized curriculum-based online achievement test for Reading Comprehension/Pānui and Mathematics/Pāngarau [9, 10]. The child’s performance was rated as at, above, below, or well below the normative range at that curriculum level. The child’s teacher also rated their academic achievement against the expected curriculum level for school year and term, and relative to their peers.

Executive function was assessed using tablet-based tests from the Cambridge Neuropsychological Test Automated Battery (CANTAB) system [11]. VMI was assessed using the VMI subscale of the Beery-Buktenica Developmental Test of Visual Motor Integration, Sixth Edition (BBVMI-6) [12]. Fine motor skills were assessed using the Manual Dexterity test from the Movement Assessment Battery for Children, Second Edition (MABC-2) [13] and the motor coordination subscale of the BBVMI-6. Visual processing was measured using laptop-based tests of motion and form coherence thresholds (MCT and FCT) [14] and the visual perception subscale of the BBVMI-6.

Parental questionnaires were completed to screen for emotional and behavioural issues (Strengths and Difficulties Questionnaire [SDQ] [15]; Behaviour Rating Inventory of Executive Function [BRIEF] [16]; Autistic Spectrum Quotient [AQ] [17]) and general health and wellbeing (Child Health Questionnaire [CHQ] [18]). Teachers also completed the appropriate versions of the SDQ and BRIEF.

Outcomes

The primary outcome of this study was low educational achievement, defined as e-asTTle score below or well below the normative curriculum level in Reading Comprehension/Pānui or Mathematics/Pāngarau. Secondary outcomes included other aspects of the primary outcome: e-asTTle z-score for year and term of school, receiving additional learning support or older than expected for year level (≥1 year older than 95% of children in their year level) or low educational achievement (e-asTTle score below or well below). Additional secondary outcomes included the proportion of children with abnormal scores and those in the clinical ranges from CANTAB, MABC-2, BBVMI-6, MCT, FCT, SDQ, BRIEF, AQ, and CHQ.

Statistical Analysis

Data were analysed using SAS version 9.4 (SAS Institute Inc) according to a prespecified statistical analysis plan. The primary analyses compared primary and secondary outcomes between children randomized to dextrose or placebo gel using generalized linear models (log-binomial for categorical and identity-normal for continuous) with adjustment for prespecified confounders of randomization stratification, socioeconomic status [19], and sex. Exposure effects are presented as adjusted risk ratio (RR), mean difference (MD), or count ratio with 95% confidence intervals (CI). Statistical tests were two-sided with a significance level of 5%. For secondary outcomes, there was no adjustment for multiple comparisons and thus results were considered exploratory. Sensitivity analyses carried out for the primary outcome were exclusion of children from multiple pregnancies and exclusion of children with a congenital or post-neonatal neurological problem that was adjudicated by study paediatricians as unlikely to result from neonatal hypoglycaemia but potentially influencing the study outcome. Exploratory analyses compared treatment groups using χ2 tests for categorical variables.

The sample size was limited by the size of the inception cohort. A retrospective power calculation estimated that our study had 90% power at the 5% significance level to detect absolute differences of at least 23% between the groups for the primary outcome.

Results

Of the 237 infants randomized in the Sugar Babies Study, 10 withdrew, leaving 227 children eligible for follow-up. Of the 184 children assessed at 9–10 years (81% of eligible), the primary outcome was available for 180 children (Fig. 1). Children who were and were not assessed had similar characteristics, except those who were assessed were more likely to be of Māori ethnicity and be admitted to the NICU (Table 1). Assessed children randomized to dextrose gel were more likely to be female but were otherwise similar to those randomized to placebo gel.

Table 1.

Baseline characteristics of participants

Fig. 1.

Flowchart of participants in the Sugar Babies mid-childhood study.

Children randomized to dextrose gel had similar rates of low educational achievement to those randomized to placebo (36/86 [42%] vs. 42/94 [45%]; RR 1.04, 95% CI 0.76, 1.44; p = 0.79) (Table 2). There were no significant differences between groups in other educational achievement outcomes, including low achievement in Reading Comprehension/Pānui or Mathematics/Pāngarau, asTTle z-scores, learning support, and teacher-rated performance relative to peers and the curriculum.

Table 2.

On sensitivity analysis, results for the primary outcome were not altered when 31 children from a multiple pregnancy were excluded (26/71 [37%] vs. 33/79 [42%]; RR 0.97, 95% CI 0.66, 1.43; p = 0.88) or when 3 children with a congenital or postnatal neurological problem were excluded (34/84 [40%] vs. 42/93 [45%], RR 1.01, 95% CI 0.73, 1.40; p = 0.95).

Children randomized to dextrose gel had similar scores in most executive function tests compared to those who received placebo (Table 2). Psychosocial adaptation scores were also similar between dextrose and placebo groups (Table 3).

Table 3.

Psychosocial adaptation outcomes

Children randomized to dextrose gel had lower visual perception standard scores (95.2 vs. 100.6; MD −5.5, 95% CI −9.4, −1.7; p = 0.005) and a higher proportion of scores <85 on the visual perception subscale compared to the placebo group (20/88 [23%] vs. 10/95 [11%]; RR 2.23, 95% CI 1.13, 4.37; p = 0.02). Other components of vision and visual-motor function were similar between groups.

In post hoc exploratory analyses, children with low (<85) visual perception scores were more likely to be of low socioeconomic status (NZDPI 8–10, 17/30 [57%]) than those with higher scores (50/153 [33%]; difference 24%, 95% CI 4.8–32%; p = 0.02), and also had lower mean birthweight (2,729 g vs. 3,127 g; MD 398 g, 95% CI 86–706; p = 0.01) and birthweight z-score (−0.68 vs. 0.26; MD 0.94, 95% CI 0.33–1.54; p = 0.003). Adjusting for socioeconomic status and birthweight did not alter the association between treatment group and lower visual perception scores (adjusted RR 2.37; 95% CI 1.20, 4.70; p = 0.014). Additionally, there was no evidence of an interaction between the relationship between treatment group and low visual perception with birthweight (p = 0.11) or socioeconomic status (p = 0.61). There were no other differences in baseline characteristics between children with or without low visual perception scores.

Because lower visual perception scores were also found in the dextrose gel group at 4.5 years [7], we further examined the subgroup of 163 children who had BBVMI-6 visual perception scores at both 4.5 years and 9–10 years. Only 6 children in each of the dextrose (7.1%) and placebo (7.6%) groups had low visual perception scores at both assessments, whereas 15 children had normal visual perception scores at 4.5 years but low scores at 9–10 years (12 [14%] in the dextrose gel and 3 [3.8%] in the placebo group), while 23 children had low visual perception scores at 4.5 years but not at 9–10 years (16 [19%] in the dextrose gel and 7 [8.9%] in the placebo group). More children from the dextrose group than the placebo group changed their visual perception result from 4.5 years to mid-childhood (33%; 28/84; 95% CI 24, 44 vs. 13%; 10/79; 95% CI 7, 22; p = 0.0018). Thus, there was a significant time-group interaction, with children in the dextrose group more likely to worsen (RR 4.24; 95% CI 1.24, 14.52; p = 0.02) but also to improve between assessments (RR 2.41; 95% CI 1.05, 5.56; p = 0.039).

Discussion

Dextrose gel is widely used as the first-line treatment for neonatal hypoglycaemia. We have previously reported that dextrose gel, compared with placebo, combined with feeding for the initial treatment of neonatal hypoglycaemia, is safe with no effect on composite neurosensory impairment at 2 or 4.5 years. This report extends these findings, showing that dextrose gel treatment did not appear to alter educational achievement at mid-childhood. Thus, the short-term benefits of reduced NICU admission for hypoglycaemia, improved breastfeeding, and reduced costs do not appear to be outweighed by any long-term adverse effects.

There were high rates of low educational achievement in this cohort. A systematic review found neonatal hypoglycaemia was associated with neurodevelopmental impairment, low literacy, and numeracy skills at mid-childhood, although with low evidence certainty [20]. However, the CHYLD study, which included the Sugar Babies cohort, found that transitional neonatal hypoglycaemia was not associated with these outcomes [8]. Our findings suggest that treatment of hypoglycaemia with dextrose gel compared with feeding alone also does not significantly alter this risk.

Treatment with dextrose gel also did not affect psychosocial outcomes, executive function, visual processing, and most aspects of visual-motor function. However, visual perception standard scores were lower in children randomized to dextrose gel, and a greater proportion had low visual perception scores. There were no differences between groups in other vision tests, particularly the other Beery-Buktenica subscales assessing visual-motor function. The visual perception test involves matching a target shape to an identical shape within an array of distractor shapes, requiring the child to have accurate perceptual representation of both the global and local features of the target and distractor objects. Accurate scanning and higher level encoding of each object is also required to identify the correct match. The other tests involving vision used in these assessments are designed to assess low to mid-level visual processes. Thus, it is possible that a child could perform normally on the other vision tasks but poorly on the visual perception task. However, children who had difficulty with the visual perception task would be expected to also have difficulty with visual motor function, and we did not detect differences between groups in these scores.

It appears unlikely that impaired visual acuity contributed to the poorer visual perception scores in the dextrose gel group. We did not specifically measure visual acuity at this age because there was no evidence of any relationship with hypoglycaemia or dextrose gel treatment at younger ages [6, 7]. Although there was a weak association between visual acuity and visual perception scores at 4.5 years [14], if visual acuity was impaired at mid-childhood we would expect to see poorer scores across the assessments involving vision, including visual perception, VMI, and both motion and form coherence.

In exploratory analyses, children who had low visual perception scores were more likely to be of low socioeconomic status and lower birthweight. However, these characteristics did not differ between the dextrose and placebo gel groups nor did adjusting for these change our findings, so they are unlikely to explain the lower visual perception scores in the dextrose gel group. In addition, in the subgroup of children whose visual perception was assessed both at 4.5 years and 9–10 years, only 7% of each group had low visual perception at both ages, while in each group similar proportions of children moved from low to normal and normal to low scores between assessments. Thus, the children with low scores in mid-childhood were generally not the same children as those who had low scores at 4.5 years, and there was no evidence that those in the dextrose gel group were more likely to have worsening visual perception with age.

Because it has been proposed that neonatal hypoglycaemia may preferentially affect the occipital and parietal lobes that play a central role in vision processing, the 2- and 4.5-year assessments included extensive eye and vision examinations including monocular and binocular vision, ocular motility, and visual processing. There were no group differences in any of these measures. There were no differences between treatment groups in other visual function tests at 2 years [6], but at 4.5 years, the visual perception standard score on the BBVMI-6 was lower for those who received dextrose compared to placebo [7]. It was suggested that this could have been a type 1 error. This may also be true for these findings at mid-childhood, given that 52 secondary outcomes were compared between dextrose and placebo groups.

If there is a true difference between groups, it does not appear to have a large impact on educational achievement or other neurodevelopmental outcomes at age 9–10 years. Treatment with dextrose gel effectively reversed hypoglycaemia in the initial randomized trial [4], which should have minimized hypoglycaemic insult to the visual pathways. Mechanistic studies suggest that glucose reperfusion, rather than hypoglycaemia itself, drives neuronal death [21]. However, in at-risk neonates, including the babies in this cohort, dextrose gel treatment increased glucose concentrations but not to high levels [22]. It therefore seems unlikely that normalizing blood glucose concentrations with dextrose gel would damage the visual pathways. With the other visual-motor function and visual processing tests indicating children in both groups have similar functional vision, it appears the visual perception finding, if not a type 1 error, is in any case not clinically significant in mid-childhood, although future implications remain uncertain.

Our study had multiple strengths, including comprehensive assessment of educational and neurocognitive outcomes known to be affected by neonatal hypoglycaemia [20]. Further strengths include prospective randomized design, relatively high follow-up rate, blinded assessments, and adjustment for potential confounders. A limitation of our study was limited power to detect small but potentially clinically significant differences in low educational achievement between groups.

We found no differences in educational achievement at 9–10 years between children who were randomized to dextrose gel or placebo for treatment of neonatal hypoglycaemia. Clinicians and consumers should be reassured that the short-term benefits of dextrose gel for the initial treatment of neonatal hypoglycaemia do not appear to be outweighed by any clinically significant effects on learning and cognition at mid-childhood.

Acknowledgments

The authors would like to acknowledge the generosity of all families who participated in this study; the CHYLD Study Steering group: Jane Alsweiler, Gavin Brown, J. Geoffrey Chase, Gregory D. Gamble, Jane E. Harding, Deborah L. Harris, Peter Keegan, Christopher J.D. McKinlay, Benjamin Thompson, and Trecia Wouldes; the International Advisory Committee: Heidi Feldman, William Hay, Robert Hess, and Darrell Wilson; the assessors: Darren Dai, Jocelyn Ledger, Stephanie Macdonald, Alecia McNeill, Samson Nivins, and Rajesh K. Shah; the study coordinators: Coila Bevan, Nataliia Burakevych, Jenny Rogers, and Eleanor Kennedy; and the data managers: Safayet Hossin, Grace McKnight, and Rashedul Hasan.

Statement of Ethics

This study protocol was reviewed and approved by the Health and Disability Ethics Committee, approval number (16/NTB/208). Written informed consent was obtained from the participant’s caregiver, and the participants provided written assent.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

This study was funded by project grants from the Health Research Council of New Zealand (17/240) and Maurice and Phyllis Paykel Trust. Sophie L. St Clair was funded by a clinical research internship from the Aotearoa Foundation (9909494). Christopher J.D. McKinlay, Gregory D. Gamble, and Jane E. Harding were funded in part by the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the National Institutes of Health (R01HD0692201 and 1R01HD091075-01A1). The content is solely the responsibility of the authors and does not necessarily represent the official views of the Eunice Kennedy Shriver National Institute of Child Health and Human Development or the National Institutes of Health. These funding sources had no role in the data or manuscript preparation.

Author Contributions

Study conceptualization and funding acquisition: Deborah L. Harris, Christopher J.D. McKinlay, Jane E. Harding, and Benjamin Thompson. Data collection: Darren W.T. Dai, Samson Nivins, and Rajesh K. Shah. Data analysis and interpretation: Sophie L. St Clair, Gregory D. Gamble, and Jane E. Harding. Drafting the initial manuscript: Sophie L. St Clair. Critical revisions and manuscript approval: all authors. Jane E. Harding was the principal investigator and took overall responsibility for the work as guarantor.

Data Availability Statement

Data and associated documentation are available to other users under the data sharing arrangements provided by the Clinical Data Research Hub, based at the Liggins Institute, University of Auckland (https://wiki.auckland.ac.nz/researchhub). The data dictionary and metadata are published on the University of Auckland’s data repository Figshare, which allocates a DOI and thus makes these details searchable and available indefinitely. Researchers are able to use this information and the provided contact address (researchhub@auckland.ac.nz) to request a de-identified dataset through the Data Access Committee of the Liggins Institute. Data will be shared with researchers who provide a methodologically sound proposal and have appropriate ethical approval, where necessary, to achieve the research aims in the approved proposal. Data requestors are required to sign a Data Access Agreement that includes a commitment to using the data only for the specified proposal, not to attempt to identify any individual participant, a commitment to secure storage and use of the data, and to destroy or return the data after completion of the project. The Liggins Institute reserves the right to charge a fee to cover the costs of making data available, if needed, for data requests that require additional work to prepare. Further data enquiries may be directed to the corresponding author.

This article is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC). Usage and distribution for commercial purposes requires written permission. Drug Dosage: The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any changes in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new and/or infrequently employed drug. Disclaimer: The statements, opinions and data contained in this publication are solely those of the individual authors and contributors and not of the publishers and the editor(s). The appearance of advertisements or/and product references in the publication is not a warranty, endorsement, or approval of the products or services advertised or of their effectiveness, quality or safety. The publisher and the editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products referred to in the content or advertisements.