Article Text

Download PDFPDF

Outcome at 4.5 years after dextrose gel treatment of hypoglycaemia: follow-up of the Sugar Babies randomised trial
  1. Deborah L Harris1,2,
  2. Greg D Gamble2,
  3. Jane E Harding2
  4. on behalf of the CHYLD Study Group
    1. 1 School of Nursing Midwifery and Health Practice, Victoria University of Wellington, Wellington, New Zealand
    2. 2 Liggins Institute, The University of Auckland, Auckland, New Zealand
    1. Correspondence to Professor Jane E Harding, Liggins Institute, The University of Auckland, Auckland, Auckland, New Zealand; j.harding{at}auckland.ac.nz

    Abstract

    Objective Dextrose gel is used to treat neonatal hypoglycaemia, but later effects are unknown.

    Design and setting Follow-up of participants in a randomised trial recruited in a tertiary centre and assessed in a research clinic.

    Patients Children who were hypoglycaemic (<2.6 mmol/L) recruited to the Sugar Babies Study (>35 weeks, <48 hours old) and randomised to treatment with 40% dextrose or placebo gel.

    Interventions Assessment of neurological status, cognitive ability (Weschler Preschool and Primary Scale of Intelligence), executive function (five tasks), motor function (Movement Assessment Battery for Children-2 (MABC-2)), vision, visual processing (Beery-Buktenica Development Test of Visual Motor Integration (Beery VMI) and motion coherence thresholds) and growth at 2 years.

    Main outcome measures Neurosensory impairment (cerebral palsy; visual impairment; deafness; intelligence quotient <85; Beery VMI <85; MABC-2 score <15th centile; low performance on executive function or motion coherence).

    Results Of 237 babies randomised, 185 (78%) were assessed; 96 randomised to dextrose and 89 to placebo gel. Neurosensory impairment was similar in both groups (dextrose 36/96 (38%) vs placebo 34/87 (39%), relative risk 0.96, 95% CI 0.66 to 1.34, p=0.83). Secondary outcomes were also similar, except children randomised to dextrose had worse visual processing scores (mean (SD) 94.5 (15.9) vs 99.8 (15.9), p=0.02) but no differences in the proportion with visual processing scores <85 or other visual test scores. Children randomised to dextrose gel were taller (z-scores 0.18 (0.97) vs −0.17 (1.01), p=0.001) and heavier (0.57 (1.07) vs 0.29 (0.92), p=0.01).

    Conclusions Treatment of neonatal hypoglycaemia (<2.6 mol/L) with dextrose gel does not alter neurosensory impairment at 4.5 years. However, further assessment of visual processing and growth may be warranted.

    Trial registration number ACTRN1260800062392.

    • neurodevelopment
    • neonatology
    • endocrinology

    Data availability statement

    Data are available on reasonable request. Data and associated documentation are available to other users under the data sharing arrangements provided by the Maternal and Perinatal Research Hub, based at the Liggins Institute, University of Auckland. The data dictionary and metadata will be 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.

    Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • Dextrose gel is now widely recommended as first-line treatment for neonatal hypoglycaemia, but later follow-up has not been reported.

    WHAT THIS STUDY ADDS

    • Treatment of hypoglycaemic babies (<2.6 mmol/L) with dextrose gel does not alter the rate of neurosensory impairment at 4.5 years; however, further assessment of visual processing and growth maybe warranted when the children are older.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • This study provides reassurance to parents and clinicans that treatment of neonatal hypoglycaemia with dextrose gel does not appear to cause harm at 4.5 years of age.

    Background

    Neonatal hypoglycaemia is common and linked to poor neurosensory outcomes.1 Prompt diagnosis and treatment is recommended for at-risk babies.2 3 Treatment with dextrose gel was shown in the Sugar Babies randomised trial to be effective for reversing hypoglycaemia, and also reduced admission to the newborn intensive care unit (NICU) for management of hypoglycaemia, and increased breast feeding.4 Treatment with dextrose gel does not impair subsequent feeding,5 can be used in many settings,6–8 is cost saving9 and is being increasingly recommended as first-line treatment for neonatal hypoglycaemia.10–15

    We have previously reported that treatment of neonatal hypoglycaemia with dextrose gel did not alter neurodevelopmental outcomes to 2 years’ corrected age.16 However, neurosensory outcomes related to neonatal hypoglycaemia may change over time.17 Therefore, we sought to determine the rate of neurosensory impairment or processing difficulties at 4.5 years’ corrected age in children who participated in the Sugar Babies Study.

    Patient and methods

    The Sugar Babies Study was a computer randomised double-blind, placebo-controlled trial performed at a tertiary referral centre (Waikato Women’s Hospital) in Hamilton, New Zealand between 1 December 2008 and 26 November 2010.4 In brief, eligible babies were born at ≥35 weeks’ gestation and at risk of neonatal hypoglycaemia (infant of a diabetic mother, small (<10th percentile or <2500 g), large (>90th percentile or >4500 g) or other). Blood glucose concentrations were measured using the glucose oxidase method 1 hour after birth, then 3–4 hourly before feeds for the first 24 hours, then 6–8 hourly for the next 24 hours. A continuous glucose monitor (CGMS system gold Medtronic, MiniMed, Northridge, California, USA) was placed as soon as possible after birth and remained in place for at least 48 hours.

    Babies who became hypoglycaemic were randomised to 40% dextrose or placebo gel 0.5 mL/kg massaged into the buccal mucosa, followed by a feed.

    Parents of all surviving participants who had not previously withdrawn were approached to participate in this follow-up study when the child was 4.5 years’ corrected age.

    Assessments

    Assessments of cognitive, motor and executive function, vision, hearing, social-emotional and behavioural problems were performed from September 2011 to June 2015 in a research clinic or the child’s home by a paediatrician, a psychologist and an optometrist. Assessors were blind to randomisation group and all neonatal glucose concentrations.

    Details of the assessments have been previously reported.17 In brief, cerebral palsy was diagnosed by neurological examination. Cognitive ability was assessed using the Weschler Preschool and Primary Scale of Intelligence, third edition (WPPSI-3) (standardised mean (SD) 100 (15)). Executive function was assessed using a battery of five tasks involving working memory, flexible attention, delay inhibition and complex or conflict inhibition (maximum 6 points per task, summed to give an Executive Function Composite Score, maximum 30 points).

    Parents completed the Behaviour Rating Inventory of Executive Function, Preschool Version (BRIEF-P, T-scores, mean 50, SD 10, T-scores >65 considered clinically significant).18 Motor function was assessed by the Movement Assessment Battery for Children-2 (MABC-2, standardised mean (SD) 10 (3)).19

    The Beery-Buktenica Development Test of Visual Motor Integration, sixth edition was used to assess visual motor integration (VMI) skills,20 including the visual processing subscale (standardised mean (SD) 100 (15)).20 Visual function composite score was assessed by awarding one point for each of: any internal or external ocular pathology; strabismus; absence of motor fusion on the 20Δ prism test; muscle restrictions, weakness or imbalance on standard tests of ocular motility, smooth pursuit and near point convergence; stereopsis not measurable on the Randot, Lang or Frisby Stero-tests or visual acuity worse than 0.3 logMAR (Snellen 20/40) in either eye or the difference between eyes worse than 0.1 logMAR (more than one line) (maximum 6 points, lower score means better vision).17 Visual impairment was logMAR ≥0.5 and blindness was logMAR ≥1.4 in both eyes.

    Hearing impairment was the requirement for aids. Auditory processing was assessed by the auditory subscales of the Phelps Kindergarten Readiness Scale (standardised mean (SD) 10 (3)).21

    Height, weight and head circumference measured using standard techniques22 were converted to z-scores.23

    Parents completed questionnaires including the Strength and Difficulties Questionnaire (SDQ),24 Child Behaviour Checklist (CBCL)25 and Social Communication Questionnaire (SComQ).26

    Statistical analysis

    The sample size was limited by the size of the primary randomised trial.4 All outcomes and analyses were prespecified. The primary outcome was neurosensory impairment, defined as one or more of cerebral palsy; visual or hearing impairment; full-scale intelligence quotient (IQ) or VMI score >1 SD below the test mean; MABC-2 total score <15th centile; motion coherence threshold or executive function score worse than 1.5 SDs from the CHYLD (Children with Hypoglycaemia and their Later Development) cohort mean.17 Secondary outcomes included all components of the primary outcome, the related subscales and the proportion of children with low scores; the proportion of children with seizures, known developmental or neurological disorder or sensorimotor impairment (cerebral palsy, visual or hearing impairment or MABC-2 <5th centile) or social-emotional or behavioural problems (abnormal score on CBCL, BRIEF-P, SDQ or SComQ, or diagnosed with autistic spectrum disorder, attention deficit hyperactivity disorder or related disorder).

    For the primary analysis, outcomes were compared between children who were randomised to dextrose or placebo gel using generalised linear regression models (binomial or normal distribution, log or identity link function) adjusted for stratification at randomisation for reason for risk of hypoglycaemia (maternal diabetes and birth weight small, appropriate, large); socioeconomic status (New Zealand Deprivation Index)27 and sex.

    Secondary and exploratory analyses compared findings in the subset of children assessed at both 2 and 4.5 years, allowing increasing certainty of classification of children who had normal outcomes (no impairment at both assessments), or abnormal outcomes (impairments at both assessments).

    Sensitivity analyses excluded twins to test for any effect of clustering within the same mother on key outcomes, and excluded children judged by adjudicating paediatricians to have had a postnatal insult unrelated to neonatal hypoglycaemia and potentially influencing the primary outcome.

    Neonatal hypoglycaemia was defined as ≥1 episode of hypoglycaemia in the first week; an episode as ≥1 consecutive blood glucose concentrations <2.6 mmol/L; severe hypoglycaemia as ≥1 episodes <2 mmol/L; recurrent hypoglycaemia as ≥3 episodes. An episode of low interstitial glucose was ≥2 consecutive measurements (10 min) below these thresholds. A hypoglycaemic event was an episode of either hypoglycaemia or low interstitial glucose concentration occurring >20 min after a previous episode.

    Data were analysed using SAS V.9.4 (SAS Institute, Cary, North Carolina, USA) and are presented as unadjusted and adjusted median (range), mean (SD), risk ratios (binary outcomes, estimated from general linear modelling with a binomial distribution and a log link function), ORs from a multinomial distribution with a logit link function or mean differences (MDs) (continuous outcomes) with 95% CIs. Cell sizes of fewer than five participants are suppressed to maintain confidentiality. Two-sided tests with p<0.05 were considered statistically significant. No adjustments for multiplicity were performed, since in this safety study we wished to identify differences between groups in any outcomes. There was no imputation of missing data because the assumption of missing at random does not apply, as data are likely to be missing because of unmeasured factors potentially related to developmental or behavioural problems.

    Results

    Of the 237 babies randomised, 185 were assessed at 4.5 years (78% of those eligible, figure 1). Characteristics of mothers of children who were and were not assessed were similar, except that mothers of children who were assessed were more likely to use alcohol and tobacco during the pregnancy than mothers of those not assessed, and assessed children were more likely to be multiples, Māori, admitted to NICU and had lower minimum interstitial glucose concentrations. Neonatal characteristics were similar between randomisation groups, except that children allocated to dextrose gel had fewer episodes of low glucose concentrations (table 1). Age at assessment was similar in the two groups (dextrose 53.7 (1.5) vs placebo 53.6 (1.5) months, MD (95% CI) 0.15 (0.27 to 0.57), p=0.48).

    Table 1

    Maternal and birth characteristics of children who were and were not assessed at 4.5 years and randomised to dextrose or placebo gel

    Figure 1

    Study profile. ABC, Assessment Battery for Children; BRIEF-P, Behaviour Rating Inventory of Executive Function, Preschool Version; NZ, New Zealand; WPPSI, Weschler Preschool and Primary Scale of Intelligence.

    Overall, 70 (38%) children had neurosensory impairment (moderate/severe 10, mild 60). Rates of impairment were similar in the dextrose (36/96 (38%)) and placebo groups (34/87 (39%), risk ratio (RR) 0.96 (95% CI 0.66 to 1.39), p=0.83) and the degree of impairment was also similar (table 2). Fewer than five children had cerebral palsy or visual impairment. No child was deaf.

    Table 2

    Outcomes at 4.5 years in babies randomised to dextrose or placebo gel

    Children randomised to dextrose gel had worse Beery scores for visual processing (dextrose 94.5 (15.9) vs placebo gel 99.8 (15.9), MD (95% CI) −5.34 (−9.93 to 0.74), p=0.02), but there were no significant differences between groups in the proportion of children with visual processing scores <85 or scores on any other Beery subscale (table 2). All other secondary outcomes were similar between groups. Adjustment for hypoglycaemia risk factors, socioeconomic status and sex did not alter the findings

    Children allocated to dextrose gel compared with placebo were taller (z-scores dextrose 0.18 (0.97) vs placebo −0.17 (1.01), adjusted MD (aMD) (95% CI) 0.43 (0.17 to 0.69), p=0.001) and heavier (z-scores dextrose 0.57 (1.07) vs placebo 0.29 (0.92), aMD (95% CI) 0.33 (0.07 to 0.60), p=0.01) (table 2). There were no differences in any other body size measurements, including body mass index.

    There were 166 children assessed at both 2 and 4.5 years, and data were available for 164 (98%). The proportion of children allocated to dextrose or placebo gel was similar in children who were or were not impaired at both ages or changed categories between ages (table 3). The relative risk of neurosensory impairment at either age compared with not impaired at either age for the dextrose compared with placebo groups was 1.02 (95% CI 0.77 to 1.36), p=0.88 (adjusted relative risk (aRR) 0.98 (95% CI 0.74 to 1.31), p=0.89), and was similar after excluding children who were impaired at only one time point (RR 1.12 (95% CI 0.61 to 2.07), p=0.71; aRR 1.03 (95% CI 0.56 to 1.88), p=0.93).

    Table 3

    Neurosensory outcomes of children assessed at both 2 and 4.5 years randomised to dextrose or placebo gel

    Findings were not changed after excluding twins (n=35, 32/78 dextrose, 30/71 placebo gel, aRR=0.99 (95% CI 0.69 to 1.43), p=0.99) or excluding children with a postnatal diagnosis likely to affect the primary outcome (aRR=0.97 (95% CI 0.68 to 1.39), p=0.88).

    Discussion

    Dextrose gel is widely recommended as first-line treatment for neonatal hypoglycaemia, and is not associated with adverse effects up to 2 years’ corrected age.16 This report extends those findings to 4.5 years. Specifically, treatment with dextrose gel did not alter the incidence of neurosensory impairment, poor cognitive, executive or motor function, hearing or overall visual impairment, nor did it alter social and emotional behaviour. However, the children randomised to dextrose gel had poorer scores for visual processing, and were taller and heavier than those randomised to placebo.

    Since treatment for neonatal hypoglycaemia with dextrose gel increases plasma glucose concentrations4 and is associated with improved breast feeding, we may have expected to see an improved neurosensory outcome in the children who received dextrose gel. However, babies randomised to the placebo gel group had additional treatment delayed only by approximately 1 hour if they remained hypoglycaemic, and the proportion of time babies were hypoglycaemic in the first 48 hours was not significantly prolonged.4 A retrospective power calculation showed that our study had 92% power to detect a 5-point difference between groups in any WPPSI subscale composite scores, so it is unlikely clinically important differences were missed.

    Children treated with dextrose gel had lower mean scores for visual processing. The clinical significance of this difference is unclear, particularly as the proportions of children with poor visual processing and poor motion coherence scores were similar between groups, as were scores for all other Beery subtests. There were also no differences in visual processing between the treatment groups at 2 years of age. It is possible this finding is a type 1 error, given that we compared 43 secondary outcomes between groups. Nevertheless, neonatal hypoglycaemia has been previously associated with MRI changes in the occipital cortex28 and later cortical visual impairments, although evidence is limited,29 so the differences in visual processing that we observed, if real, may be of clinical relevance, and later assessment of visual processing is warranted.

    Children treated with dextrose gel were also taller and heavier than those treated with placebo, but body mass index was similar. It is difficult to understand why dextrose gel may influence growth to 4.5 years, particularly when body size was unaffected at 2 years.16 It is unlikely that this apparently faster growth between 2 and 4.5 years is related to the higher rate of breast feeding in the dextrose gel group, since children who were breast fed generally have slower early childhood growth than those fed formula.30 It seems likely that these findings are also type 1 errors, although additional measures of growth would also be appropriate.

    Repeated assessment of children provides increasing certainty about earlier neurosensory classifications. Neurosensory assessment at 2 years’ corrected may not be predictive of later outcome.31 By 4.5 years children have increased skills for problem solving, planning, attention and goal orientation, so difficulties in these skills are more likely to be evident. Overall, nearly half of the children in this cohort had no impairment at either time point. Although children with impairment at 2 years were more likely to have impairment at 4.5 years, one-third of children had changed classification between the two ages, with more children having impairment at 4.5 years than at 2 years. Nevertheless, the proportion of children who changed classification was similar in children randomised to dextrose or placebo gel, providing further evidence that treatment with dextrose gel does not increase the risk of later neurosensory impairment.

    A major strength of this study is it is a prospective longitudinal cohort study of a randomised trial. Children underwent comprehensive validated neurodevelopmental assessments, and the follow-up rate was similar to the rate at 2 years (78 %), providing robust data for comparison between the two age groups.16 The baseline characteristics of those who were and were not assessed were largely similar: although slightly more children living in higher deprivation areas were not assessed, adjustment for this did not change our findings. Limitations of the study included limited power to detect small differences between the groups.

    Conclusion

    We have previously shown that dextrose gel is an effective treatment for neonatal hypoglycaemia (<2.6 mmol/L) in late-preterm and term babies and is safe up to 2 years’ corrected age. We now provide further reassurance for clinicians and families that dextrose gel does not alter neurosensory impairment at 4.5 years. However, further assessment of visual processing and growth is warranted when the children are older.

    Data availability statement

    Data are available on reasonable request. Data and associated documentation are available to other users under the data sharing arrangements provided by the Maternal and Perinatal Research Hub, based at the Liggins Institute, University of Auckland. The data dictionary and metadata will be 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.

    Ethics statements

    Patient consent for publication

    Ethics approval

    This study was approved by Health and Disability Ethics Committee of New Zealand Northern Y Ethics Committee (NTY/08/03/025 and NTY/10/03/021). Parents of participants gave informed consent for their child to participate in the study before taking part.

    Acknowledgments

    We are grateful to the children and families who participated in this study. We sincerely thank our International Advisory Group: Heidi Feldman (MD), Stanford University School of Medicine, USA; William Hay (MD), University of Colorado School of Medicine, USA; Darrell Wilson (MD), Stanford University School of Medicine, USA and Robert Hess (DSc), McGill Vision Research Unit, Department of Ophthalmology, McGill University, Canada. We also acknowledge the contribution of the following members: Sugar Babies Steering Committee—University of Auckland (Auckland, New Zealand): Jane Harding DPhil, (chair), Waikato District Health Board (Hamilton New Zealand): Philip J Weston MBChB, Deborah L Harris PhD.

    References

    Footnotes

    • Twitter @DeborahHarrisNP

    • Collaborators CHYLD Study Group: Steering Committee—University of Auckland (Auckland, New Zealand): Jane Harding DPhil, (chair), Jane M Alsweiler PhD, Trecia A Wouldes PhD, Gavin T L Brown PhD, Christopher J D McKinlay; University of Waterloo (Ontario, Canada): Benjamin Thompson PhD; University of Canterbury (Christchurch, New Zealand): J. Geoffrey Chase PhD; Victoria University of Wellington (Wellington New Zealand): Deborah L Harris, PhD.Data collection: Judith Ansell (PhD),1 Anne Jaquiery (PhD),1 Kelly Jones (PhD),1 Sapphire Martin (BNurs),1 Christina McQuoid (DipEdPsych),1 Jenny Rogers,1 Heather Stewart,1 Anna Timmings (MBChB),2 Anna Tottman (MBBS),1 Kate Williamson (MBBS),1 Arun Nair (MD),2 Alexandra Wallace (PhD),2 Phil Weston (MBChB),2 Nicola Austin (DM),4 Jeremy Armishaw (MBChB),5 Nicola Webster (MBBS),6 Ross Haslam (MBBS),7 Pat Ashwood (BSc),7 Lex Doyle (MD),8 Kate Callanan,8 Ian Wright (MBChB)9. Study coordination: Jessica Brosnahan (MHSc),1 Ellen Campbell (PhD),1 Coila Bevan (BA),1 Tineke Crawford,1 Kelly Fredell (BNurs),1 Kate Sommers.1Data management: Greg D Gamble (MSc),1 Claire Hahnhaussen (BSc),1 Safayet Hossin (MSc),1 Karen Frost (BSc),1 Grace McKnight,1 Janine Paynter (PhD),1 Jess Wilson (MSc),1 Rebecca Young (BEd),1 Anna Gsell (PhD),1 Aaron Le Compte (PhD),3 Matthew Signal (PhD)3 Yannah Jiang (PhD),1 Tzu-Ying (Sandy) Yu PhD,1.

      1Liggins Institute, University of Auckland, Auckland, New Zealand. 2Waikato Hospital, Hamilton, New Zealand. 3University of Canterbury, Christchurch, New Zealand. 4Canterbury District Health Board, Christchurch, New Zealand. 5Bay of Plenty District Health Board, Tauranga, New Zealand. 6Mid-Central District Health Board, Palmerston North, New Zealand. 7Women’s and Children’s Hospital, Adelaide, Australia. 8Royal Women’s Hospital, Melbourne, Australia. 9John Hunter Children’s Hospital, Newcastle, Australia.

    • Contributors DLH contributed to the study design, data collection and interpretation, drafted the initial manuscript and subsequent revisions of the manuscript. GDG contributed to the study design, data analysis and interpretation and revision of the manuscripts. JEH contributed to the study design, data interpretation and revision of manuscripts, in addition to taking responsibility for the overall content as guarantor. All authors approved the final manuscript and agreed to be accountable for all aspects of the work.

    • Funding The Sugar Babies Study was funded by the Waikato Medical Research Foundation, the Auckland Medical Research Foundation, the Maurice and Phyllis Paykel Trust, the Health Research Council of New Zealand and the Rebecca Roberts Scholarship. Follow-up was funded by the Health Research Council of New Zealand and the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the National Institutes of Health under award number R01HD0692201.

    • Competing interests None declared.

    • Provenance and peer review Not commissioned; externally peer reviewed.