Study Design
This single-center, retrospective cohort study was approved by the institutional review board at the University of South Alabama in Mobile, Alabama. We included healthy newborn infants who met the following criteria: gestational age of ≥ 34 weeks, birth weight > 2000 grams, born from April 2016 through March 2023, and considered to be at high-risk for transitional hypoglycemia. We excluded infants born in 2017 between February and December from the study due to data unavailability caused by our institution's electronic medical record system switch from Cribnote to Cerner that year. The implementation of the ODG protocol began in February 2017.
Consequently, we categorized the infants into two groups: the baseline group, consisting of historical control infants, born before the ODG implementation (April 2016 to January 2017) and managed according to an institutional protocol; the intervention group, including infants born from January 2018 through March 2023, during a period during which we implemented an ODG protocol which was incorporated into the 2011 AAP guidelines. (Fig. 1)
Protocols and Interventions
During the baseline period, the operational threshold for hypoglycemia was a blood glucose (BG) level of < 40 mg/dl within the first 24 hours. Asymptomatic infants with hypoglycemia received their initial feeding through either breastfeeding or bottle-feeding. If any subsequent confirmed BG levels remained below 35 mg/dL or if they had 3 consecutive BG levels ≥ 35 mg/dl and < 40 mg/dl, in accordance with our institutional protocol, the infants were subsequently transferred to the NICU.
During the intervention period, hypoglycemic infants received an initial supplementation of 0.5 ml/kg of 40% oral dextrose gel (Glutose-15 40% ODG, Perrigo, Minneapolis, MN) sublingually. This was followed by gently massaging the gel onto the buccal mucosa before feeding. However, in February 2021, the unit switched from using the Glutose-15 formulation to Sweet Cheeks (40% ODG, manufactured by DandleLion Medical in Danbury, CT) due to concerns regarding parabens 10,11. The threshold for transferring infants to the NICU is depicted in Fig. 1.
Glucose Measurements
Infants identified as at risk for hypoglycemia underwent serial blood glucose level screening tests as recommended by the 2011 AAP guidelines during both the baseline and intervention periods. Blood glucose (BG) levels were screened by obtaining a small blood sample from a heel stick using a point-of-care (POC) testing device (Accu-Chek Inform II, Roche Diagnostics, Abbott Park, IL). We ensured that all POC testing devices were used in accordance with the operator's manual. Any POC BG levels below 50 mg/dL were subsequently confirmed using the glucose oxidase method (RapidLab blood gas analyzer, Siemens Diagnostics, Deerfield, IL), conveniently located near the nursery.
We defined infants at risk for hypoglycemia as any infant who received a POC BG test in the newborn nursery. In order to ensure consistency, we maintained the definition of hypoglycemia as BG levels < 40 mg/dL throughout the entire study. Consequently, during the intervention period, infants with BG levels ranging from 30 mg/dL to 39 mg/dL within the first 4 hours of life were considered hypoglycemic but did not receive oral dextrose gel (ODG) supplementation, as they did not meet the criteria for hypoglycemia outlined in the guidelines established by the AAP.
Data Variables
In this study, we collected demographic data, including maternal diabetes, gestational age (GA), and birth weight (BW), to identify at-risk infants: infants born to diabetic mother (IDM), preterm infants (< 37 weeks GA), small for gestation (SGA, < 10th percentile), and large for gestation (LGA, > 90th percentile) based on Fenton's intrauterine growth chart criteria.12 We retrieved data on feeding practices during hospital stays, categorizing infants into two subgroups: exclusively breastfeeding or breastfeeding with exposure to bovine-based milk. Additionally, we compiled data on major daily hospital expenses, primarily consisting of room and professional charges for infants in both the Newborn Nursery (NBN) and the NICU. We also identified reasons for NICU transfers, specifically those transferred solely for hypoglycemia management.
Data Collection
We collected relevant clinical, laboratory, and treatment data for infants in our study. For infants born during the baseline period, data were acquired from individual charts created in Cribnote and later scanned into Cerner. For infants born after December 2017, data was retrieved from generated Cerner reports.
Outcome measures
The primary outcome was NICU admission, defined as any newborn admitted to the NBN for routine care who developed low BG levels necessitating interventions and was subsequently transferred to the NICU due to failure to achieve normoglycemia. Secondary outcomes included feeding practices at the time of hospital discharge, such as exclusive breastfeeding and feeding with bovine-based milk formula, as well as the cost of hospital care. All of these outcomes were compared between the two periods.
Statistical Analyses
Statistical differences between the baseline and intervention groups were analyzed using Pearson’s chi-square test in categorical data, and a two-sample Wilcoxon rank-sum (Mann- Whitney) test for continuous variables not normally distributed. We used multivariate logistic regression to adjust for any significant association between the primary outcome or any secondary outcome and the study groups, by including baseline characteristic variables that are different between the groups (p < 0.05). All baseline characteristics that were included in the multivariate regression analysis were sequentially excluded either for p values > 0.2 or when Bayesian Information Criteria of the depleted model were less than the prior model by an absolute value more than 2.13 Measures of association were expressed as odds ratio (OR) with 95% confidence intervals (95% CI). All statistical analyses were conducted using Stata 13 (StataCorp LP, Lakeway Drive, College Station, TX).