Article Text
PDF
Abstract
Objectives To determine the impact of incorporating dextrose gel in the treatment of neonatal hypoglycaemia (NH) and the role of feeding type in NH outcomes.
Study design We conducted a retrospective analysis of 2688 infants >35 weeks’ gestation who were screened for NH before and after implementation of a clinical guideline for NH evaluation and treatment. We analysed the proportion of infants who required intravenous dextrose for NH before and after guideline implementation, the change in blood glucose concentrations with gel by feeding type and the odds of successful NH treatment with gel and feeding by feeding type.
Results Following implementation of the guideline, a lower proportion of infants required intravenous dextrose for NH treatment (8.6% (60 infants) before guideline vs. 5.6% (112 infants) after guideline (p=0.007)). The median rise in blood glucose concentration with gel administration in the entire cohort was 0.61 mmol/L (11 mg/dL) (IQR 0.28–1.06 mmol/L (5–19 mg/dL)). Blood glucose concentration of formula-fed infants rose more in response to feeding and gel than breastfed infants (p≤0.0001). Formula feeding was associated with a lower odds of recurrent hypoglycaemia, as defined by requiring a second gel, in a fully adjusted model. Specifically, in infants with a pregel blood glucose of 2.00–2.17 mmol/L (36–39 mg/dL), formula feeding with gel was associated with a lower odds of recurrent hypoglycaemia.
Conclusions Dextrose gel is an effective tool in the treatment of NH. An infant’s pregel blood glucose concentration may be helpful in guiding decisions around type of feeding provided.
- endocrinology
- infant feeding
- neonatology
Statistics from Altmetric.com
What is already known on this topic?
Neonatal hypoglycaemia is associated with neurodevelopmental sequelae.
Dextrose gel has been shown in one randomised controlled trial to decrease neonatal intensive care unit (NICU) admissions.
What this study adds?
Implementation of dextrose gel in a clinical setting decreases the need for intravenous dextrose and NICU admissions for hypoglycaemia.
Dextrose gel raises blood glucose in infants with hypoglycaemia, and the degree of the rise in blood glucose differs by feeding type provided with gel.
Pregel blood glucose concentration may be useful in guiding feeding type for infants with hypoglycaemia.
Introduction
Neonatal hypoglycaemia (NH) affects 15% of newborns and is estimated to account for $2.1 billion in annual healthcare costs in the USA.1 2 Because NH is a preventable cause of neurological injury and poor neurodevelopment outcome,3 much attention has been focused on defining threshold blood glucose values for NH that require treatment. To this end, both the American Academy of Pediatrics (AAP) and the Pediatric Endocrine Society (PES) have published guidelines that aim to define blood glucose goals and treatment guidelines for infants at risk for NH.4 5 Both PES and AAP guidelines suggest that feeding should be used to prevent and treat NH, and the AAP guidelines state that ‘efforts should not unnecessarily disrupt the mother-infant relationship and breastfeeding’.4 Neither of these guidelines provide guidance on when or how to introduce formula feeding, perhaps because minimal data exist on which to base such recommendations. Both guidelines further recommend intravenous dextrose if feeding does not achieve glycaemic control, necessitating separation of mother and infant at most institutions. Buccal administration of dextrose gel, which was found to decrease hypoglycaemia-related neonatal intensive care unit (NICU) admissions in a large randomised controlled trial,6 is not included in either guideline, possibly due to a paucity of published data on the use of dextrose gel in newborns with hypoglycaemia.
The Department of Pediatric Newborn Medicine at Brigham and Women’s Hospital in Boston, Massachusetts, implemented a new clinical practice guideline (CPG) in 2015 for screening and management of NH during the initial neonatal hospitalisation. This CPG, based on the AAP and PES guidelines, incorporated the use of dextrose gel. The specific objectives of this implementation research were to: (1) determine the impact of implementation of the CPG on need for intravenous dextrose; (2) quantify the effect of gel administration on rise in blood glucose concentrations stratified by infant feeding type; (3) understand whether feeding type was associated with need for repeat dextrose gel administration or intravenous dextrose; and (4) identify a specific population of infants for whom formula supplementation combined with dextrose gel may prevent the need for intravenous dextrose administration.
Methods
Subjects
This was a retrospective analysis of asymptomatic infants screened for hypoglycaemia during two time periods: pre-CPG, a 5-month period before guideline implementation (June–October 2015), and post-CPG, a 13-month period (January 2016–January 2017) after guideline implementation. Using a data extract report from the electronic health record (EHR), we first identified all infants born during these time periods who had a blood glucose measurement within 48 hours of birth. This included infants who did not require any treatment, received dextrose gel and/or received intravenous dextrose for treatment of NH. Timing of all variables, including infant feedings, blood glucose measurements and gel administration were provided by a date and time stamp in this data extract report. Infants were excluded from analysis if they were less than 35 weeks gestation at birth, a multiple gestation or received their first gel after 48 hours of life. This project was undertaken as a quality improvement initiative at Brigham and Women’s Hospital and, according to institutional policy, was not formally supervised by the Institutional Review Board.
Changes in clinical care as a result of CPG implementation
Before implementation of this CPG, our hospital followed the AAP guideline (2011). The target prefeed blood glucose was >2.22 mmol/L from birth to 4 hours of age. After 4 hours of age, the target prefeed blood glucose increased to >2.50 mmol/L. Intravenous dextrose was provided if the the blood glucose concentration was <1.39 mmol/L between 0 and 4 hours of life and <1.94 mmol/L between 4 and 24 hours of life.
In November 2015, a CPG for neonatal glucose assessment and management was developed and implemented, integrating the most recent AAP and PES recommendations. Per the CPG, nurses measured blood glucose concentrations with the Precision Xceed Pro (Abbott) point-of-care device within 1 hour of delivery for infants at risk for NH. Ongoing management of NH was determined based on infant’s point-of-care blood glucose concentration as follows: if an infant’s blood glucose was between 1.11 mmol/L and 2.50 mmol/L, bedside nurses administered dextrose gel, 40%, 0.5 mL/kg via buccal mucosal massage; if blood glucose was below 1.11 mmol/L, the infant was transferred to the NICU for intravenous dextrose; if blood glucose was ≥2.50 mmol/L, the value would be rechecked at an interval defined in the CPG. Feeding was encouraged within 1 hour of birth, after gel administration (if applicable) and based on infant cues, with type of feeding (breast feeding or formula) according to parental preference and clinician recommendation. For infants who received gel, a follow-up blood glucose measurement was performed 30 min after the initial gel administration. Gel could be administered a total of two times before intravenous dextrose was initiated.
Statistical approach
Proportion of infants requiring intravenous dextrose before and after implementation of CPG
From the EHR data abstract, we determined the number and the proportion of infants per month who were screened for NH and received intravenous dextrose for hypoglycaemia in: (1) the pre-CPG and (2) the post-CPG epochs. We used SPC macros in Microsoft Excel to create an empiric p-chart and determine the p-bar (mean) proportion of infants who received intravenous dextrose as a proportion of the infants who were screened for NH in both epochs.
Quantifying the effect of dextrose gel administration on blood glucose concentrations by infant feeding type (post-CPG period only)
To determine the effect of gel administration on blood glucose concentration, we identified all infants in the post-CPG period who received one or two gel treatments. We compared: (1) pregel, (2) postgel and (3) difference between pregel and postgel blood glucose concentrations for first and second gel administration separately. We stratified each set of preglucose and postglucose values by the type of feeding (formula only, maternal breast milk (MBM) only, mixed feeding and no feeding) the infant received with that gel administration. For this reason, it was possible that an infant could have been categorised under one feeding type for gel 1 and a different feeding type for gel 2. We used the Wilcoxon signed-rank test to determine the significance of the change from pregel to postgel blood glucose values by feeding group and the Kruskal-Wallis test to determine whether the blood glucose changes were different by feeding group.
Feeding type and odds of recurrent NH (post-CPG period only)
We fitted logistic regression models with feeding group as the exposure (breastfed infants served as the reference group) and need for second gel as the outcome. For this analysis, only infants who were fed breast milk or formula were included (infants in the mixed feed or no feed groups were excluded.) Models sequentially adjusted for potential confounders included: M1: maternal race and parity; M2: NH risk factor (gestational diabetes, small for gestational age/large for gestational age and gestational age); M3: maternal BMI; and M4: lowest infant blood glucose. To better delineate whether associations between feeding group and recurrent NH were specific to a strata of pregel blood glucose, we then examined the association of pregel blood glucose quartile (1.11–1.94, 2.00–2.17, 2.22–2.33 and 2.39–2.44 mmol/L, with the latter group as the reference) on odds of requiring a second dose of gel stratified by feeding group.
Results
Clinical characteristics and outcomes of infants pre-CPG and post-CPG period
During the pre-CPG period, we identified 694 infants, and during the post-CPG period, we identified 1994 infants who met inclusion criteria for NH screening. These groups differed based on the following maternal–infant characteristics: primiparity, race, gestational diabetes, feeding type, birth weight, gestational age, weight for age centile and intravenous dextrose administered (table 1). It should be noted that overall, a greater proportion of infants from our hospital population were screened during the post-CPG period (ie, pre-CPG average: 29% of infants born per month vs post-CPG average: 34% of infants born per month), reflecting the inclusion of more risk factors in the CPG. During the post-CPG period, dextrose gel was administered to 32.5% (n=649) of at-risk infants. The proportion of infants requiring intravenous dextrose decreased from 8.6% to 5.6% (p=0.005) as shown in table 1 and figure 1.
Proportion of infants who received intravenous dextrose pre-CPG and post-CPG implementation. The first five months (June–October 2015) represent the preguideline period. The dashed black vertical line denotes implementation of the guideline in November 2015, and the subsequent 12 months represent the postguideline period (January–December 2016). P-bar is the mean proportion of infants who received intravenous dextrose during both periods. P-hat is the proportion of infants born who received intravenous dextrose for hypoglycaemia. LCL and UCL are the lower and upper control limits. CPG, clinical practice guideline.
Characteristics of mothers and infants screened for neonatal hypoglycaemia at Brigham and Women’s Hospital, preclinical and postclinical practice guideline (CPG) implementation
The effect of dextrose gel administration on blood glucose concentration by infant feeding type
On average, the first gel was administered 4.3 hours after delivery. With administration of the first dose of gel, median blood glucose values increased in infants in all four feeding groups (table 2). The median change (or difference) in blood glucose was 0.61 (IQR 0.28–1.06) mmol/L, and the median postfeed blood glucose (2.72 mmol/L) in all four groups was greater than the 2.50 mmol/L threshold defined by our CPG. Infants who received formula at the time of first gel administration had the greatest increase in blood glucose levels, with a median increase of 0.83 (IQR 0.44–1.33) mmol/L. In contrast, infants who were breast fed only or those who were not fed had a median difference in blood glucose of 0.56 (MBM fed IQR 0.25–0.94; not fed IQR 0.17–1.06) mmol/L, and the mixed feeding group had a median difference of 0.67 (IQR 0.33–0.83) mmol/L. At the time of the second gel administration, at a mean of 6.5 hours after delivery, formula-fed and breastfed infants had increases in blood glucose concentration of 0.89 and 0.83 mmol/L, respectively, compared with a median increase of 0.55 mmol/L in infants who were not fed at the time of second gel administration.
Pregel and postgel blood glucose (BG) concentrations in infants who received dextrose gel stratified by feeding group
The mean volume of formula taken by the infants was 15.5 mL with the first gel administration and 12.9 mL with the second gel administration. The volume of formula taken did not correlate with the rise in blood glucose at either time point (first gel: Pearson’s r=0.13, p=0.13, second gel: Pearson’s r=−0.002, p=0.99).
Feeding type and odds of recurrent NH
Infants who received formula with their first gel administration were less likely to require a second gel in a fully adjusted model (OR 0.4, 95%, CI 0.24 to 0.73) (table 3). This association was strongest in a subset of infants who had moderate NH, with a pregel blood glucose of 2.00–2.17 mmol/L (table 4). Within this subgroup, infants who were breast fed were 3.5 times more likely than the reference group to require a second gel (95% CI 1.3 to 9.3), compared with formula-fed infants with the same pregel blood glucose (OR 0.86, 95% CI 0.24 to 3.0).
Odds of recurrent NH (receiving second gel or receiving intravenous dextrose) for formula-fed (vs breast fed=referent) infants
Odds of requiring second gel administration by pregel blood glucose, stratified by feeding group
Discussion
Dextrose gel is increasingly used as a treatment for NH. Here, we report that implementation of an NH evaluation and management guideline that included use of dextrose gel decreased the need for intravenous dextrose in a population of infants at risk for NH. In addition, we report that dextrose gel (alone and in combination with feeding) is effective in raising neonatal blood glucose concentrations and identified a specific group of infants for whom supplementation with formula (in addition to gel) may prevent recurrent hypoglycaemia.
The implementation of our CPG, which included the risk factors for NH highlighted by AAP and PES, resulted in a greater proportion of infants screened for NH but a lower proportion requiring intravenous dextrose to treat NH. This decrease in need for intravenous dextrose was despite an increase in the operational threshold for NH in the first 4 hours of life and was temporally related to institution of dextrose gel. At our institution, intravenous dextrose administration requires admission to the NICU, resulting in separation of mothers and infants, which may interfere with the establishment of breast feeding and result in lower patient satisfaction with the postpartum/newborn hospitalisation. Admission to the NICU also has implications for cost of care. We estimate that with implementation of our CPG at our institution, we have prevented three NICU admissions for every 100 infants at risk for NH or approximately 54 NICU admissions annually. The cost savings and associated patient-level impact of preventing NICU admissions warrant further investigation.
Our findings that infants who were exclusively breast fed and those who were not fed had similar increases in blood glucose concentrations after the first but not second gel administration was surprising. These findings are similar to those of the Sugar Babies post hoc analysis, which is the only other study that investigated the influence of feeding and gel on blood glucose outcome.7 In the first few hours following delivery, when minimal amounts of colostrum are produced, breast feeding may be a surrogate for skin-to-skin contact, which also has been linked with raising blood glucose levels.8 By the time of the second gel administration, it is likely that more colostrum is being produced, which supports the finding that infants who were breast feeding had a larger and more sustained increase in blood glucose levels (similar to formula-fed infants) compared with those who were not fed.
We also observed that dextrose gel led to a clinically and statistically significant increase in blood glucose levels in all four feeding groups (ie, not fed, breast fed, infant formula fed and mixed fed prior to blood glucose measurement). This finding is supported by one clinical trial and three observational studies that report reduced risk of NICU admission and mother–infant separation when dextrose gel was provided for treatment of NH.6 9–11
In contrast to a previous report,7 in our cohort, formula feeding was associated with a decreased requirement for a second gel. These findings refute early data that suggest glucose homeostasis is hormonally driven and not related to feeding.12 Our findings support the role of formula feeding as a potential adjunct treatment for NH in a specific subset of infants with moderate NH (blood glucose 2.00–2.17 mmol/L). Given the availability of dextrose gel and its effectiveness in raising blood glucose concentrations, our data suggest that pregel blood glucose concentration may guide decision making as to whether additional treatment of NH with formula supplementation is needed. These data might allow identification of at-risk groups for whom formula supplementation may prevent NICU admission, allay concerns from parents and staff regarding ‘unnecessary’ formula use and permit a subset of dyads (those with pregel blood glucose ≥2.22 mmol/L) to continue to attempt exclusive breast feeding. Our data also suggest that exclusive breast feeding may not raise the initial blood glucose level to the same degree as feeding infant formula, particularly in those infants with moderate NH (pregel blood glucose 2.00–2.17 mmol/L). In developing guidelines, institutions must incorporate these data and the goal of establishing early breast feeding.
Our observational study had limitations. Covariate data were extracted from an EHR, and therefore, relied on accurate clinician documentation. There may have been a bias in selecting infants who received formula because those infants had a lower pregel blood glucose concentration. The demographics and exclusive breast milk feeding rates at hospital discharge of our study population are representative of the Boston, Massachusetts area, which may limit generalisability to other populations. The goal blood glucoses per our CPG were higher in the first 4 hours of life than the 2011 AAP recommendations. It is possible that a proportion of infants who received gel in the first 4 hours of life would have had spontaneous resolution of NH without intervention, but given the study design, we were not able to evaluate this. We also do not have information on which infants were symptomatic in this cohort. We were not able to measure the association of feeding type with the need for intravenous dextrose due to the small number of infants with this outcome. Finally, all aspects of our guideline were implemented together as a new bundle of care, preventing our ability to determine the influence of any one aspect of the guideline on clinical outcomes.
Despite these limitations, our observations suggest that a practice guideline that includes use of dextrose gel to treat hypoglycaemia was effective in decreasing the need for intravenous dextrose, and supplemental formula feeding may be considered as an adjunct treatment for NH in a subset of infants.
Acknowledgments
The authors wish to acknowledge Drs Linda Van Marter and Ellice Lieberman for their thoughtful review and feedback of this manuscript. The authors also wish to acknowledge the multidisciplinary clinical practice council within the Department of Pediatric Newborn Medicine at Brigham and Women’s Hospital for their support in the development and implementation of the clinical practice guideline discussed here.
Footnotes
Contributors KG, DT, CM-D and SS were responsible for the design, conduct, analysis of the study and wrote the first draft of the manuscript. SH and JW contributed to the analysis and manuscript revisions.
Funding SS is supported by K23 HD 074648.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement The data for this study have not yet been published or shared with anyone outside of the authorship team.
Patient consent for publication Not required.