Dextrose Gel for Neonates at Risk With Asymptomatic Hypoglycemia: A Randomized Clinical Trial

We conducted an open-label, parallel design, randomized controlled trial in a level 3B tertiary neonatal care unit in South India. The study period was from October 2017 to January 2019. The Institutional Ethics Committee approved the study (IEC/17/AUG135/27). The trial was registered with the clinical trial registry of India (CTRI/2017/11/010383).

We included all inborn babies ≥35 weeks of gestation who were at risk for hypoglycemia (SGA and IUGR, IDM and LGA and late preterm) and developed asymptomatic hypoglycemia in the first 48 hours of life. We excluded babies who required resuscitation at birth, needed intravenous fluids, or who were admitted to the NICU for any other reason within 48 hours of life.

We stratified babies into 3 categories: SGA and IUGR, IDM and LGA, and late preterm babies (≥ 35 weeks) to ensure equal distribution in each category.^11–13  If a neonate had more than 1 risk factor, the neonate was stratified in the order of SGA and IUGR, IDM and LGA, and late preterms.^14,15  Fenton’s 2013 growth chart was used for plotting the growth.¹⁶  Hypoglycemia was defined by the sugar levels as outlined by the 2011 American Association of Pediatrics (AAP) guidelines during the 3 periods or brackets; 0 to 4 hours of life, 4 to 24 hours of life, and 24 to 48 hours of life.

A computer generated variable block randomization sequence was used and allocation concealment was by serially numbered, opaque, sealed envelopes at a 1:1 allocation ratio. We stratified based on the at-risk factors into 3 categories, namely SGA and IUGR, IDM and LGA, and late preterm babies. The principle investigator and the primary care physician could not be blinded because of unavailability of a similar-looking placebo.

Informed written consent was obtained from mothers of eligible neonates, either in the antenatal period at the time of admission to hospital or soon after delivery. As per our unit protocol, arterialized capillary blood glucose (CBG) was monitored using a bedside glucometer (Freestyle Optium Neo). In the study, the same was chosen as a pragmatic option for blood glucose monitoring. CBG was conducted for all at-risk neonates, starting at 1 hour of life and continued every 3 hours until 48 hours of life. If hypoglycemia was present by 48 hours, then monitoring was continued until 2 consecutive values were ⋧ 60 mg%, whichever was later. Breast milk or breastfeeding was used for pain control during heel lancing. We used AAP 2011 guidelines for blood glucose monitoring and management of hypoglycemia (Fig 1). If blood glucose was less than the threshold and the neonate was asymptomatic, a sealed envelope from the corresponding category was opened by the neonatologist on call. In the DGG, 200mg/kg (0.5 mL/kg) of oral 40% dextrose gel taken in a 5 mL disposable syringe was slowly injected into the space between the inside of the cheek and lower gum (the buccal cavity), followed by breastfeeding. CBG was rechecked after 30 minutes. If there was a rising trend and the neonate continued to be asymptomatic, a similar process was repeated, and CBG was checked again after 30 minutes. Up to 2 consecutive attempts were made for an episode of hypoglycemia (Fig 1). Dextrose gel could be repeated up to 6 times if hypoglycemia recurred within 48 hours of life, defined as CBG < 45 mg% after successful treatment. In India, 40% oral dextrose gel is currently not being manufactured. It was imported for the purpose of this study. It can also be procured online.

If the neonate came under the CG, breastfeeding was provided, and CBG was checked after 30 minutes. If there was an improvement in sugar level and the neonate continued to be asymptomatic, breastfeeding was repeated, and blood sugar was rechecked after 30 minutes. As a unit policy, we practiced exclusive human milk feeding during the first 48 hours in all babies.

Successful treatment was defined as a blood sugar of > 45 mg/dL 30 minutes after the first or the second attempt. Treatment failure was defined as the need for intravenous fluids for stabilizing blood glucose levels, that is, blood sugar < 45 mg/dL after 30 minutes of the second attempt in the DGG or after a second attempt of breastfeeding alone in the CG. Rebound hypoglycemia was defined as any episode of hypoglycemia, CBG < 45 mg/dL, within 6 hours after successful treatment.

In case of treatment failure, venous samples were drawn for blood glucose estimation (enzyme method) and the neonate was provided intravenous dextrose with a glucose infusion rate of 6 to 8 mg/kg per min (Fig 1).

At any point during the study, if a neonate developed symptoms secondary to hypoglycemia, a statim mini bolus (2 mL/kg of 10% dextrose IV) was administered and an intravenous dextrose infusion was initiated at a glucose infusion rate of 6 to 8 mg/kg per min.

IUGR was defined as a fall in centiles or Doppler changes on antenatal scans or a clinical assessment of nutrition score < 25 in the neonate at birth.¹⁷  IDM was defined as infants born to mothers who had pregestational diabetes (type 1 or type2) or those who developed diabetes during pregnancy (gestational).

Our primary outcome was to analyze the rate of treatment failure in the DGG compared with CG. We also assessed a failure rate in each category under each of the time brackets. Our secondary outcomes were to compare the proportion of neonates discharged from the hospital on exclusive breastfeeding. We were vigilant of potential side effects, like rebound hypoglycemia.

Through an audit conducted in our unit, we observed the rate of asymptomatic hypoglycemia among neonates at-risk of hypoglycemia was 45%. Of these, 33% neonates needed ICU admission for intravenous fluids to stabilize blood sugar levels. This number is similar to the data in other studies. We aimed to reduce the incidence of NICU admission from 33% to 20%. To achieve this, while ensuring 80% power and 5% precision, we needed to recruit 284 infants (142 in each group).

Descriptive data were expressed as a number and a percentage for categorical variables, mean and standard deviation (SD) were used for normally distributed continuous variables, and a median and interquartile range was used for abnormally distributed continuous variables. Results of categorical variables were compared using the χ² test. Independent sample t test, and the Mann Whitney test were used for normally and abnormally distributed continuous variables, respectively. All statistical analysis was performed using SPSS 23. A P value <.05 was considered significant.

The consort flow diagram is shown in Fig 2. Maternal and neonatal baseline characteristics were similar between the 2 groups (Table 1). The rate of treatment failure (need for intravenous fluids) in the DGG and CG was 17(11.5%) and 58(40.2%) respectively, (RR 0.28 [0.17–0.46]) P < .001(Table 2). The mean random blood glucose of the study group and the control group were 21.4 ± 6.9 and 21.3 ± 6 mg/dL, respectively. Absolute risk reduction was 0.29 and the number needed to treat was 3.5. The mean SD number of doses used in each stratum was 1.5 (0.4) in SGA and IUGR, 1.3 (0.4) in IDM and LGA, and 1.35 (0.5) in late preterm babies. Treatment failure was significantly lower in the DGG in all 3 categories: SGA and IUGR, IDM and LGA, and late preterm infants with a RR of 0.29 (95% CI 0.13–0.67), 0.31 (95% CI 0.14–0.66) and 0.24(95% CI 0.09–0.66), respectively (Table 2).

Under each time bracket, it was observed that during 0 to 4 hours of life, 131 neonates developed asymptomatic hypoglycemia: 66 in the DGG and 65 in the CG. The failure rate was 6 (9.1%) and 24(36.9%), respectively (RR 0.33 [0.15–0.78]). During 4 to 24 hours of life, 150 neonates developed asymptomatic hypoglycemia, 75 in the DGG and CG each. There was treatment failure in 11 (15%) neonates in the DGG and 34 (45.3%) neonates in the CG (RR 0.33 [0.19–0.61]). The treatment failure rate was significantly lower among the DGG in the IDM and LGA group between 0 and 4 hours of life and in SGA and IUGR and late preterm babies between 4 and 24 hours of life. During 24 to 48 hours of life, because of a small number of neonates, analysis could not be done.

More neonates were discharged on exclusive breastfeeds in the DGG (132) than in the CG (122). However, this was not statistically significant (RR 1.06 [0.98–1.15] P = .19) (Table 3). No neonate developed side effects with 40% oral dextrose gel, like rebound hypoglycemia, hyperglycemia, or vomiting. None of the neonates in our study developed symptomatic hypoglycemia after randomization.

Our study shows that 40% oral dextrose gel followed by breastfeeding, compared with only breastfeeding, reduces the need for NICU admission for intravenous fluids in at-risk neonates with asymptomatic hypoglycemia within 48 hours of life. This was consistent across other randomized and observational trials conducted in our study; 1 NICU admission was prevented for every 3.5 babies treated with dextrose gel. Troughton et al¹⁸  did not find a beneficial effect on recovery from hypoglycemia at 15 or 30 minutes following 40% oral dextrose gel administration.

We included neonates with risk factors for hypoglycemia as quoted by the AAP and the Pediatric Endocrine Society,^19,20  similar to other studies.^21–25  There is no single value to define hypoglycemia with difference in the thresholds recommended by various committees.^26–31  Hence, we used a widely accepted value of ≤ 45 mg%, similar to other studies.^{21,23–25,32–38}  Blood sugar was tested by a point of care glucometer, as in previous trials.^24,25,36  Consistent with prior studies, we administered oral 40% dextrose gel at 200 mg/kg (0.5 mL/kg) in DGG.^23,25,39,40  In a study conducted in 2000, Troughton et al⁴¹  used a higher dose of 40% dextrose gel (400 mg/kg), and it did not show increased benefit.

Oral dextrose gel is stable at room temperature and can be manufactured at a reasonable cost.^37,42  The current cost is 5 US dollars (362 rupees) for a 15 g tube and 1 dose has an average cost of 20 rupees per 1.5 mL. It does not require clearance from the Drugs Controller General of India as it is does not fall into the definition of a “new” drug.⁴³ 

The Cochrane 2016 systematic review included 2 trials involving 589 neonates to evaluate oral dextrose gel for treatment of hypoglycemia. It concluded that dextrose gel did not alter the need for intravenous treatment. The risk factors used in the studies included were the same as in our study. However, none of the studies analyzed the results under 3 different time brackets.¹¹ 

Romald et al³³  conducted a study under 2 time brackets, 0 to 4 hours and 4 to 24 hours of life. They had similar results as our study. The impact of oral 40% dextrose gel in our study was clinically and statistically significant between 0 and 4 hours in IDM and LGA neonates, and between 4 and 24 hours in SGA and IUGR and late preterm neonates. This may be explained by hyperinsulinemia seen in IDM babies in the early hours of life, rendering the effect of dextrose gel most effective between 0 and 4 hours.⁴¹  Whereas babies in the other 2 categories have multifactorial etiology for hypoglycemia, namely suboptimal glycogen and fat stores, delayed appearance of gluconeogenic enzymes, and hyperinsulinism. Up to 94% of occurrences of hypoglycemia in these babies take place in the first 24 hours of life.⁴⁴  Pertaining to low incidence of hypoglycemia between 24 and 48 hours, a larger sample size is needed to assess the effect of oral dextrose gel during this time bracket.⁴⁵ 

As observed in our study, Neonatal Network 2017, and the meta-analysis by Egan et al⁴⁶  in 2019, a reduction in the rate of mother-baby separation by incorporation of dextrose gel in treating asymptomatic hypoglycemia among at-risk neonates was found. Recent Cochrane review in 2021 concluded little or no effect on mother-neonate separation.³⁸ 

The World Health Organization, the European Society for Paediatric Gastroenterology, Hepatology, and Nutrition, and the AAP recommend donor milk as the best alternative to mother’s own milk.^47,48  In recent years, an increasing number of milk banks are being developed in low- and middle-income countries. However, even these numbers are insufficient to meet the total demand for donor milk in low- and middle-income countries.⁴⁹  Use of formula milk is not always feasible because of financial constraints and its potential side effects.⁵⁰  Dextrose gel can bridge this gap by improving breastfeeding rates as demonstrated in previous studies.^11,22,25,33 We did not observe a statistically significant difference in regard to exclusive breastfeeding rate at the time of discharge, possibly because of a high baseline breastfeeding rate at our center with the help of a lactation support program.⁵¹ 

Although, there are concerns that 40% oral dextrose gel could potentially cause hyperglycemia, choking, or vomiting, we did not observe any adverse effects. Previous studies, and Cochrane review by Edward et al in 2021³⁸ did not reveal any adverse effects.^{11,21,23–25,33,52} 

Rawat et al²⁵  employed the Consumer Assessment of Healthcare Providers and Systems score to assess maternal satisfaction with the intervention. Mothers reported higher acceptance and satisfaction with dextrose gel. In our study, dextrose gel was readily accepted by mothers. However, we did not objectively study satisfaction rates among parents.

This is the first randomized control trial from India on the use of dextrose gel for asymptomatic hypoglycemia in at-risk neonates within 48 hours of life. We used 3 time-brackets for each category based on hours of life as recommended by the 2011 AAP Guidelines.

The limitations of our study are that we had only a few neonates in the time bracket of 24 to 48 hours. We could not blind the doctor who administered 40% oral dextrose gel because of an inability to procure a similar looking placebo, like 2% carboxymethyl cellulose. We did not make use of continuous glucose monitoring, which has the potential to provide more accurate and real-time measurements of blood glucose levels.

Dextrose gel reduces the need for intravenous fluids (NICU admission) for the treatment of asymptomatic hypoglycemia in at-risk neonates in the first 48 hours of life, enabling mothers to practice exclusive breast milk feeding. Dextrose gel was found to be effective in reducing NICU admissions of asymptomatic hypoglycemia, individually across 3 categories, namely SGA and IUGR, IDM and LGA, and late preterm babies. It is well tolerated, avoids mother-neonate separation, and can be used as a first line treatment modality in neonates with asymptomatic hypoglycemia.

Additional larger trials on the management of asymptomatic hypoglycemia occurring between 24 and 48 hours of life using dextrose gel and studies to evaluate the impact of this intervention on long term neurologic outcome are needed.

We thank the entire neonatal team for their support throughout this study.

AAP

American Academy of Pediatrics

CG

control group

DCGI

Drugs Controller General of India

DGG

dextrose gel group

IDM

infant of diabetic mother

IUGR

intrauterine growth restriction

LGA

large for gestational age

RR

relative risk

SGA

small for gestational age