Intravenous Dextrose for the Treatment of Neonatal Hypoglycaemia: A Systematic Review

Neonatal hypoglycaemia, typically defined as blood glucose concentration <2.6 mmol/L [1], affects up to 39% of all newborn infants [2]. Among at-risk infants, including those born preterm, low or high birthweight, or to mothers with diabetes (IDM), the incidence is around 50% [3]. Symptoms include poor feeding, seizures, and hypotonia [4]. Adverse effects have also been observed in early- and mid-childhood, including neurodevelopmental impairment, visual-motor impairment, and executive dysfunction [5, 6].

The usual first-line treatment for asymptomatic hypoglycaemia is increasing feed frequency [7], with or without oral dextrose gel [8]. Infants whose low blood glucose concentrations are severe or persist, or those who are symptomatic, are often admitted to the neonatal intensive care unit (NICU) for treatment with intravenous (IV) dextrose [7, 9].

However, the evidence of the effectiveness of this widespread practice is limited, and the dose of dextrose required may vary in different groups of babies. For instance, in one study, a 200 mg/kg IV 10% dextrose bolus followed by 8 mg/kg/min constant IV 10% dextrose infusion led to varied blood glucose responses between and within groups at risk for hypoglycaemia [10]. There are also concerns regarding the rate at which infants’ blood glucose concentrations are increased, with rapid rises in glucose concentration after hypoglycaemia associated with neurosensory impairment at 2 years of age [11]. We aimed to systematically assess the benefits and risks of IV dextrose given to newborn infants for the treatment of hypoglycaemia, and whether these differed with different treatment schedules.

This review was prospectively registered with PROSPERO (registration number CRD42023461654) and is reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist [12] (online suppl. material 1; for all online suppl. material, see https://doi.org/10.1159/000541471).

The following electronic databases were searched from inception to October 5, 2023: MEDLINE (Ovid), Embase (Ovid), CINAHL Complete, and the Cochrane Central Register of Controlled Trials (CENTRAL). Registered trials were searched for in: Current Controlled Trials (www.controlled-trials.com), Clinical Trials (www.ClinicalTrials.gov), and the WHO ICTRP Search Portal (https://apps.who.int/trialsearch/). Conference abstracts that provided usable data were considered for inclusion. Reference lists of included studies were screened. For the full search strategy, see online supplementary material 2.

Inclusion criteria were newborn infants (term or preterm) diagnosed with neonatal hypoglycaemia in all care settings and treated with IV dextrose (including bolus and/or continuous infusions) compared to infants with hypoglycaemia treated with a placebo/other treatment/alternative IV dextrose administration protocols. We included published randomised controlled trials (RCTs), non-randomised studies of interventions, cohort studies, and before and after studies, without language or date restrictions.

The primary outcome was neurodevelopmental impairment at ≥18 months of age (including executive dysfunction; low language/literacy, low numeracy; highest education level, emotional-behavioural difficulty; and cognitive/visual-motor/visual/hearing/motor impairment). Secondary outcomes included: hypoglycaemia (investigator defined) after initial treatment until discharge home; adverse effects; hypoglycaemic injury on brain imaging; breastmilk feeding exclusively from birth to discharge; duration of initial hospital stay after birth; cost of neonatal care (for hypoglycaemia intervention and total hospital cost); time to correction of hypoglycaemia (time from initiation of IV treatment to achievement of blood glucose concentration above the threshold definition) before discharge home (minutes); hyperglycaemia (any, severity, and number of episodes, study defined); requirement for any medications for treatment of hypoglycaemia such as glucagon or corticosteroids before discharge home; measures of glycaemic stability e.g., proportion of glucose measurements within acceptable range (study defined); duration of NICU stay; and duration of treatment for hypoglycaemia.

Two review authors (L.F.R. and L.G.L.) independently screened titles and abstracts of identified records, then assessed potentially eligible full-text articles for inclusion using Covidence [13]. The same review authors then independently extracted data into a pre-specified data extraction form. The extracted data included the authors, study setting, study methodology, ethics approval, conflicts of interest, funding sources, information for the assessment of the risk of bias, participant characteristics, intervention and control details, and outcome data. Two authors (L.F.R. and L.G.L.) independently assessed the risk of bias of RCTs with the Cochrane Risk of Bias 2 tool (RoB 2) [14] and of non-randomised studies of interventions and observational studies with the Cochrane Risk of Bias in Non-Randomized Studies of Interventions (ROBINS-I) tool [15]. Discrepancies were resolved by discussion between reviewers. Non-English articles were translated by a colleague, DeepL [16], or Google Translate [17].

We planned to determine certainty of evidence for key outcomes (neurodevelopmental impairment at ≥18 months of age, hypoglycaemia after initial treatment until discharge home, adverse effects, breastmilk feeding exclusively at discharge, hypoglycaemic injury on brain imaging, breastmilk feeding exclusively from birth to discharge, and duration of initial hospital stay) using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach [18]. A “Summary of Findings” table was compiled using the GRADEpro Guideline Development Tool (GDT) [19]. Findings were summarised according to guidance from Cochrane, whereby statements for conclusions are based on the effect estimate size (e.g., large effect or no effect) and certainty of evidence (very low, low, moderate, or high) [20].

We intended to carry out meta-analyses using RevMan 5.4.1 [21] but found insufficient data for each reported outcome. We used random-effects models in RevMan to calculate reported risk ratios (RRs) or odds ratios (ORs) with 95% confidence intervals (CIs) for dichotomous outcomes. For continuous outcomes, we calculated mean differences (MDs) or MDs with 95% CIs. A p value of <0.05 denoted statistical significance. We planned to calculate I² and χ² for each analysis to determine heterogeneity but found insufficient data. Blood glucose measurements were converted to mmol/L, and dextrose infusion rates were converted to mg/kg/min where possible. We planned to undertake subgroup analyses by ethnicity, gestational age, risk status, maternal diabetes status, symptomatic versus asymptomatic hypoglycaemia, dose of treatment, mode of dextrose delivery, and pre-treatment blood glucose concentration. All analyses were pre-specified unless stated otherwise.

In total, 2,014 records were identified by database searches (Fig. 1). After automatic removal of 480 duplicates, title and abstract screening was conducted for 1,534 records, followed by full-text screening of 21 records. All records were retrieved. In total, 6 studies (8 records) were included in the review.

Of the six included studies, two were RCTs and four were cohort studies (Table 1). In total, 711 participants were included (302 intervention; 409 comparison). The studies were published between 1980 and 2021; one (17%) was conducted in a lower middle-income country and five (83%) was conducted in a high-income country [22]. Three studies compared IV infusion protocols, whereas three compared IV dextrose infusion with other. Meta-analysis was not performed due to insufficient studies reporting the pre-specified outcomes.

Ten outcomes from two RCTs were assessed using Cochrane’s RoB 2 tool, seven (70%) of which were assessed as low risk of bias, with some concerns about the remaining three (30%) outcomes. These concerns arose from uncertainty as to whether bias arose from how outcomes were measured and, for one outcome, multiple eligible outcome measurements.

Ten outcomes from four observational studies were assessed using Cochrane’s ROBINS-I tool, four of which (40%) were deemed at moderate risk of bias, four (40%) at serious risk of bias, and two (20%) with insufficient information to assess risk of bias because participant selection information lacked. No outcome was considered at low risk of bias as confounding was likely in all studies (Fig. 2a, b).

One retrospective cohort study of large for gestational age (L.G.A.) infants found no difference between children treated for hypoglycaemia with IV dextrose infusion and enteral feeding (breastfeeding, bottle feeding, or tube feeding) to achieve a total glucose intake rate of 5 mg/kg/min, compared to those who received enteral feeding only (no target glucose intake rate), in psychological test scores at 4 years of age using the Dutch versions of the Denver Developmental Scale, the Child Behaviour Check List and the Snijders-Oomen non-verbal intelligence test (no effect sizes provided; p > 0.2; very low certainty evidence; 60 infants) [24].

One prospective cohort study by Burakevych et al. [25] of at-risk infants provided very uncertain evidence about the effect of IV dextrose treatment (200 mg/kg bolus 10% dextrose over 10 min, followed by 4–6 mg/kg/min infusion) on episodes of hypoglycaemia. The median number of hypoglycaemic episodes in the 6 h after initiation of treatment may be higher among infants treated with IV dextrose (2 episodes [IQR 1; 2], 19 infants) than among infants fed with breastmilk (1 episode [IQR 1; 1], 25 infants), formula (1 episode [IQR 1; 2], 23 infants), dextrose gel and breastmilk (1 episode [1;2], 32 infants), or dextrose gel and formula (1 episode [IQR 1;2], 29 infants) (very low certainty evidence; 128 infants).

However, evidence from one RCT by Bora et al. [23] in a lower middle-income country in small for gestational age (SGA) infants born at 32–36 weeks and with symptomatic hypoglycaemia suggested that infants receiving IV dextrose treatment (200 mg/kg bolus 10% dextrose followed by 6 mg/kg/min infusion) may have a reduced risk of hypoglycaemic episodes (blood glucose concentration <2.2 mmol/L) compared to infants given 200 mg oral sucrose dissolved in expressed breast milk then sugar-fortified breastmilk every 2 h, calculated as 6 mg/kg/min sucrose and divided into 12 feeds (RR: 0.67 [95% CI: 0.20, 2.18], p = 0.5; very low certainty evidence; 80 infants). Adverse effects were reported by Bora et al. [23] who found IV dextrose treatment may reduce neonatal mortality (RR: 0.75 [95% CI: 0.18, 3.14], p = 0.69; very low certainty evidence; 80 infants) and feeding intolerance (RR: 0.67 [95% CI: 0.20, 2.18], p = 0.50; very low certainty evidence; 80 infants) but increase necrotising enterocolitis (RR: 5.00 [95% CI: 0.25, 100.97], p = 0.29; very low certainty evidence; 80 infants).

The same RCT reported that IV dextrose treatment may reduce the likelihood of infants fully breastmilk feeding at discharge (RR: 0.68 [95% CI: 0.44, 1.05], p = 0.08; very low certainty evidence; 80 infants) [23]. The same RCT reported that IV dextrose treatment may reduce duration of initial hospital stay (MD: −1.48 days [95% CI: −4.36, 1.40], p = 0.31; very low certainty evidence; 80 infants) [23].

Burakevych et al. [25] reported that the duration of hypoglycaemia (<2.6 mmol/L) after initial treatment may increase among infants treated with IV dextrose compared to infants treated with breastmilk (MD: 0.91 h [95% CI: 0.15, 1.67], p = 0.02; 44 infants), dextrose gel and breastmilk (MD: 0.74 h [95% CI: −0.06, 1.54], p = 0.07; 51 infants), dextrose gel and formula (MD: 0.67 h [95% CI: −0.14, 1.48], p = 0.10; 48 infants), or formula alone (MD: 0.54 h [95% CI: −0.29, 1.37], p = 0.20; 42 infants), but the evidence is very uncertain. Bora et al. [23] reported that IV dextrose may increase the risk of a hyperglycaemic episode (blood glucose concentration >4.4 mmol/L) 6 h after initiating treatment (RR: 2.33 [95% CI: 0.65, 8.39], p = 0.19; 80 infants).

Burakevych et al. [25] reported that treatment with IV dextrose may be associated with a higher proportion of interstitial glucose measurements outside a central band of 3–4 mmol/L compared to treatment with dextrose gel and breastmilk (MD: 0.17 [95% CI: 0.01, 0.33], p = 0.03; 51 infants), dextrose gel and formula (MD: 0.17 [95% CI: 0.01, 0.33], p = 0.04; 48 infants), breastmilk alone (MD: 0.14 [95% CI: −0.03, 0.31], p = 0.12; 44 infants), or formula alone (MD: 0.09 [95% CI: −0.07, 0.25], p = 0.28, 42 infants), but the evidence is very uncertain.

Evidence from one RCT by Vanhatalo et al. [26] suggested that treatment with 8 mg/kg/min 20% dextrose infusion may result in little to no difference in the risk of hypoglycaemia (plasma glucose concentration <2.6 mmol/L) after treatment initiation, compared to 15% dextrose infusion delivering the same dose of dextrose (RR: 0.87 [95% CI: 0.68, 1.13], p = 0.31; low certainty evidence; 121 infants). It also likely resulted in little to no difference in phlebitis risk (RR: 0.99 [95% CI: 0.74, 1.33], p = 0.94; moderate certainty evidence; 119 infants) [26].

One cohort study by Sen et al. [27] compared the cost of neonatal care for infants treated with a standard infusion (200 mg/kg bolus 10% dextrose followed by 4.2 mg/kg/min) versus infants treated with an adjustable infusion protocol based on baseline blood glucose concentration (if baseline <1.1 mmol/L: 200 mg/kg bolus followed by 4.2 mg/kg/min infusion; if baseline 1.1–1.7 mmol/L: 4.2 mg/kg/min infusion; if baseline 1.7–2.4 mmol/L: 2.1 mg/kg/min infusion). Evidence from this study suggested that treatment with the standard infusion may be associated with a higher cost of NICU stay, compared to treatment with the adjustable infusion protocol (MD: USD 5,441 [95% CI: USD 1,111, 9,772], p = 0.001; aMD: USD 4,417 [95% CI: USD 571, 8,263], p = 0.03; 277 infants). Evidence from the same study suggested that treatment with a standard infusion protocol may result in little to no difference in the time to achieve normoglycaemia (2.5 mmol/L) from the start of IV dextrose treatment when compared to an adjustable infusion protocol (MD: 1 min [95% CI: −12.1, 14.1], p = 0.9; aMD: 0 min [95% CI: −13.3, 13.3], p = 1.0; 277 infants) [27]. Evidence from one cohort study by Lilien et al. [10] suggested that treatment with a 200 mg/kg bolus 10% dextrose followed by 8 mg/kg/min infusion may be associated with a large increase in the likelihood of correction of hypoglycaemia (plasma glucose concentration ≥1.9 mmol/L in infants ≥2,500 g or ≥1.4 mmol/L in infants <2,500 g) within 10 min of infusion, compared to 8 mg/kg/min infusion only (OR: 11.43 [95% CI: 0.58, 226.11], p = 0.11; 45 infants), but evidence is very uncertain.

Evidence from the same cohort suggested that treatment with bolus followed by infusion may be associated with a large increase in the risk of hyperglycaemia (blood glucose concentration >6.9 mmol/L) during infusion, compared to infusion only (OR: 5.23 [95% CI: 0.24, 115.38], p = 0.29; 45 infants), but evidence is very uncertain. Visual representation of glucose concentrations over the first 20 min demonstrated glycaemic response to dextrose infusion varied markedly between and within at-risk groups: appropriate for gestational age, SGA, and LGA/IDM [10].

Evidence from Vanhatalo et al. [26] (2010) suggested that treatment with 8 mg/kg/min infusion 20% dextrose likely results in a slight reduction in risk of hyperglycaemia (plasma glucose concentration >7.7 mmol/L) during the infusion period, compared to infusion of 15% dextrose, but confidence intervals were wide (RR: 0.83 [95% CI: 0.37, 1.86], p = 0.65; 121 infants). Evidence from Sen et al. [27] suggested that a standard infusion protocol may be associated with increased duration of NICU stay compared to an adjustable infusion protocol with dose graded according to baseline glucose concentration (MD: 1.5 days [95% CI: 0.1, 2.9], p = 0.04; aMD: 1.9 days [95% CI: 0.7, 3.0], p = 0.002; 277 infants).

Evidence from Vanhatalo et al. [26] suggested that treatment with 8 mg/kg/min infusion 20% dextrose likely results in little to no difference in the duration of treatment for hypoglycaemia compared to 15% dextrose infusion (median difference 0 days [95% CI: −0.46, 0.46], p = 1.00; 121 infants). Subgroup analyses could not be conducted due to lack of data.

The certainty of evidence was assessed as very low for neurological impairment, hypoglycaemia after initial treatment until discharge home (IV dextrose vs. other treatment), adverse effects (one RCT comparing IV dextrose to other treatment), breastmilk feeding at discharge, and duration of initial hospital stay (Table 2). The certainty of evidence was assessed as low for: hypoglycaemia after initial treatment until discharge home (one RCT comparing alternative IV dextrose treatment protocols); and adverse effects – phlebitis (Table 3). There were no data for hypoglycaemic injury on brain imaging or breastmilk feeding exclusively from birth to discharge.

Our review aimed to systematically assess benefits and risks of IV dextrose given to newborn infants for the treatment of hypoglycaemia, and whether these differed with different treatment protocols. Despite a comprehensive search we found little evidence, predominantly from high-income countries. Hypoglycaemic infants are at higher risk of long-term neurodevelopmental impairment [5] and treatment aims to reduce this risk. This review found insufficient evidence to determine effects of IV dextrose on long-term neurodevelopmental outcomes.

We found limited evidence of other potential effects of IV dextrose as treatment for hypoglycaemia, including a possible reduction in hypoglycaemic episodes and duration of initial hospital stay. This is consistent with the use of IV dextrose for treatment of symptomatic, persistent, or severe hypoglycaemia after other interventions have failed [7, 9]. Also, compared to oral sucrose, IV dextrose may reduce the risk of adverse effects including neonatal mortality and feeding intolerance. However, these possible beneficial effects are reported in only one study, and the evidence is very uncertain.

We found limited evidence about potential harms of this treatment. Firstly, IV dextrose treatment may reduce full breastmilk feeding at discharge from hospital, perhaps because IV dextrose treatment commonly takes place in NICU, requiring separation of baby and mother potentially impeding breastfeeding. IV dextrose treatment may also increase the risk of necrotising enterocolitis, characterised by severe inflammation of the intestinal mucosa. Further, our review found that IV dextrose treatment may be associated with poorer neonatal glycaemic control, with an increased risk of hyperglycaemia, hypoglycaemia, and blood glucose concentrations outside a central range. Neuronal injury has been observed in the brains of developing rats that experienced hyperglycaemia after hypoglycaemia [28], and rapid increases in blood glucose concentration after neonatal hypoglycaemia have been linked to poorer neurodevelopmental outcomes in 2-year-old children [11]. This is of particular concern for at-risk infants born preterm who, owing to their metabolic immaturity, are likely highly susceptible to glycaemic instability [25].

It is possible that the harm associated with IV dextrose treatment is linked to the treatment protocol. Evidence from one study suggests an IV dextrose bolus preceding continuous infusion may quickly correct hypoglycaemia but may increase the risk of hyperglycaemia, compared to continuous infusion only [10]. Another study found that a standard infusion protocol (including bolus) resulted in a steeper rise in blood glucose concentration compared to an adjustable infusion protocol with dose graded according to infant’s baseline glucose concentration [27]. Due to the link between glycaemic instability and poorer neurodevelopmental outcomes later in childhood, there have been concerns that bolus administration may risk harm to the neonate, and recent recommendations have suggested minimisation of bolus use in treating asymptomatic neonates [29]. However, additional RCTs are required on the effectiveness and safety of the bolus treatment protocol.

The IV dextrose treatment protocol may impact on the length of initial hospital stay and cost of treatment. Sen et al. [27] found that the adjustable infusion protocol may be linked to reduced length of stay in NICU and reduced cost of treatment, compared to the standard protocol. While infants on the adjustable infusion protocol achieved normoglycaemia in a similar timeframe to those on the standard infusion protocol, the less steep blood glucose concentration increase resulted in improved glycaemic stability and earlier weaning from treatment, permitting earlier discharge from NICU. Reducing time in NICU is beneficial for cost-savings – authors estimated a US nationwide saving of USD 36 million with the adjustable infusion protocol. Reducing time in NICU is also beneficial for minimising separation of mother and infant, potentially facilitating breastfeeding and enhancing care satisfaction [30].

To our knowledge, this is the only systematic review to synthesise the evidence on the effectiveness of IV dextrose for neonatal hypoglycaemia treatment. Strengths of the review include a broad search strategy and no restrictions of language, date, gestational age of infants, or geographic location of included studies to ensure all relevant studies were identified.

Limitations of the review include lack of data for some pre-specified outcomes (including the primary outcome) and only one to two studies reporting on other outcomes, so we were unable to conduct the planned meta-analyses or subgroup analyses. The evidence base is not geographically representative, with all but one of our included studies conducted in a high-income country, so the findings may not be directly applicable to lower income settings. While two RCTs were assessed, most of the included studies were observational, reducing the certainty of the evidence.

Although there is currently no alternative to IV dextrose for the treatment of persistent, severe, or symptomatic neonatal hypoglycaemia, the limited available evidence about the effects of different infusion protocols suggests it may not be appropriate to treat all infants the same. However, with just a few studies on which the evidence is based, it would be inappropriate to make recommendations for clinical practice. High-quality RCTs are needed to examine the benefits and risks of different IV dextrose treatment protocols, including their acceptability and practicality as well as their effects on later neurodevelopment, in different groups of infants. Outcomes should also be reported for subgroups, such babies of different ethnicity, birthweight, gestational age, and risk factors for hypoglycaemia.

We found very uncertain evidence from only one study that IV dextrose may have little to no effect on neurodevelopmental impairment at ≥18 months of age. It is unclear whether IV dextrose treatment reduces the risk of hypoglycaemia and adverse effects due to limited included studies with conflicting and very uncertain findings. Further studies are needed to elucidate the most promising treatment protocol to maximise benefits and minimise harms for specific groups of infants.

We thank librarian Rayna Dewar at University of Auckland Library for supporting development of the search strategy.

An ethics statement is not applicable since this review is based exclusively on published literature.

The authors declare that they have no competing interests.

This work was funded in part by grants from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) of the National Institutes of Health (NIH R01HD091075, L.F.R.), the Health Research Council of New Zealand (19/690, J.E.H., C.A.C.), and the Aotearoa Foundation (9909494, L.L.). The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NICHD or the NIH.

J.E.H., C.A.C., and L.L. planned the systematic review. L.F.R. designed the search strategy. L.F.R. and L.G.L. conducted title and abstract screening, full-text assessment, data extraction, quality and bias assessment, and assessment of certainty of the evidence. L.F.R., L.L., and J.E.H. drafted the manuscript. All authors read and approved the final manuscript.

All data generated or analysed during this study are included in this article online supplementary material files. Further enquiries can be directed to the corresponding author.