Home > Health & Research > Health Education Campaigns & Programs > Cochrane Neonatal Review > Interventions for prevention of neonatal hyperglycemia in very low birth weight infants

Interventions for prevention of neonatal hyperglycemia in very low birth weight infants

Skip sharing on social media links
Share this:

Authors

John C Sinclair1, Marcela Bottino2, Richard M Cowett3

Background - Methods - Results - Characteristics of Included Studies - References - Data Tables & Graphs


1Departments of Pediatrics and Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Canada [top]
2(Formerly, Fellow, Division of Neonatology, Department of Pediatrics, McMaster University, Hamilton, Ontario, Canada), Rio de Janeiro, Brazil [top]
3Neonatal Resource Services, Complex Medical Conditions, Optum, UnitedHealth Group, Lisle, Illinois, USA [top]

Citation example: Sinclair JC, Bottino M, Cowett RM. Interventions for prevention of neonatal hyperglycemia in very low birth weight infants. Cochrane Database of Systematic Reviews 2011, Issue 10. Art. No.: CD007615. DOI: 10.1002/14651858.CD007615.pub3.

Contact person

Marcela Bottino

(Formerly, Fellow, Division of Neonatology, Department of Pediatrics
McMaster University, Hamilton, Ontario, Canada)
Avenida Senador Danton Jobim
150/404 Barra da Tijuca
22631-060 Rio de Janeiro
Brazil

E-mail: mabottino@hotmail.com

Dates

Assessed as Up-to-date: 17 May 2011
Date of Search: 30 April 2011
Next Stage Expected: 17 May 2013
Protocol First Published: Issue 1, 2009
Review First Published: Issue 3, 2009
Last Citation Issue: Issue 10, 2011

What's new

Date / Event Description
18 May 2011
Updated

This is an update of the review "Interventions for prevention of neonatal hyperglycemia in very low birth weight infants" first published in the Cochrane Database of Systematic Reviews, 2009 Issue 3 (Sinclair 2009).

Literature searches were updated to April 30 2011.

18 May 2011
New citation: conclusions not changed

No new eligible included trials were found. One possibly eligible trial (Hawk 2009) published only in abstract form was classed as Studies Awaiting Classification. Six possibly eligible studies were added to the list of excluded studies.

The results and conclusions of the review are unchanged.

History

Date / Event Description
16 June 2010
Amended

Minor edit to Outcome 2.9

Abstract

Background

Among very low birth weight (VLBW) infants, early neonatal hyperglycemia is common and is associated with increased risks for death and major morbidities. It is uncertain whether hyperglycemia per se is a cause of adverse clinical outcomes or whether outcomes can be improved by preventing hyperglycemia.

Objectives

To assess effects on clinical outcomes of interventions for preventing hyperglycemia in VLBW neonates receiving full or partial parenteral nutrition.

Search methods

We searched the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library, issue 4 of 12, 2011; MEDLINE (1966 to April 2011); EMBASE (1980 to April 2011); CINAHL (1982 to Nov 2008); abstracts of Pediatric Academic Societies 2000 to 2011 and European Society for Pediatric Research 2005 to 2010.

Selection criteria

Randomized or quasi-randomized controlled trials of interventions for prevention of hyperglycemia in neonates with birth weight < 1500 g or gestational age < 32 wk.

Data collection and analysis

Two review authors independently selected studies for eligibility and extracted data on study design, methods, clinical features, and treatment outcomes. Included trials were assessed for blinding of randomization, intervention and outcome measurement, and completeness of follow-up. Treatment effect measures for categorical outcomes were relative risk and risk difference, and for continuous outcomes, mean difference, each with their 95% confidence intervals.

Results

We detected four eligible trials. Two trials compared lower versus higher rates of glucose infusion in the early postnatal period. These trials were too small to assess effects on mortality or major morbidities. Two trials, one a moderately large multicentre trial (NIRTURE, Beardsall 2008), compared insulin infusion with standard care. Insulin infusion reduced hyperglycemia but increased death before 28 days and hypoglycemia. Reduction in hyperglycemia was not accompanied by significant effects on major morbidities; effects on neurodevelopment are awaited.

Authors' conclusions

Glucose infusion rate: There is insufficient evidence from trials comparing lower with higher glucose infusion rates to inform clinical practice. Large randomized trials are needed, powered on clinical outcomes including death, major morbidities and adverse neurodevelopment.

Insulin infusion: The evidence reviewed does not support the routine use of insulin infusions to prevent hyperglycemia in VLBW neonates. Further randomized trials of insulin infusion may be justified. They should enrol extremely low birth weight neonates at very high risk for hyperglycemia and neonatal death. They might use real time glucose monitors if these are validated for clinical use. Refinement of algorithms to guide insulin infusion is needed to enable tight control of glucose concentrations within the target range.

Plain language summary

Interventions for prevention of neonatal hyperglycemia in very low birth weight infants

Blood sugar levels higher than usually seen in full term infants are frequently seen in babies born very early (before 32 weeks gestation) or with very low birth weight (< 1500 grams) and who are fed totally or partially by vein. Several types of adverse outcomes have been associated with high blood sugar levels including increased risks for death, infections, vision problems, and bleeding into the brain. It is not known if prevention of high blood sugar levels improves those complications and, if so, which intervention is best. Possible options include restriction of the amount of sugar delivered by vein to nourish the baby or administration of insulin. Trials which compared lower with higher amounts of sugar delivered by vein were too small to determine effects on the health outcomes of the babies. Insulin was found to reduce the number of babies who developed high blood sugar levels, but the health outcomes of the babies were not improved. In fact, insulin infusion was associated with an increased risk of death before 28 days of age.

[top]

Background

Description of the condition

In the very low birth weight neonate (VLBW, < 1500g at birth), an elevated blood glucose concentration occurs frequently, especially during the first days after birth. Hays 2006 reported that among 82 parenterally fed extremely low birth weight infants (ELBW, < 1000 g at birth) nearly 60% had blood glucose concentrations > 8.3 mM/L (> 150 mg/dL) between day two and seven of life, with the peak prevalence on the second day. The risk of hyperglycemia is inversely related to gestational age and birth weight (Blanco 2006; Falcão 1998) and increases with the severity of accompanying illnesses (Louik 1985; Cowett 1997). Adverse clinical outcomes have been associated with neonatal hyperglycemia. These adverse outcomes include death (Hays 2006; Kao 2006; Heimann 2007), intraventricular hemorrhage (IVH) grades 3 and 4 (Hays 2006), late onset bacterial (Kao 2006) and fungal infection (Rowen 1995; Manzoni 2006), retinopathy of prematurity (ROP) (Garg 2003; Blanco 2006; Ertl 2006), necrotizing enterocolitis (NEC) (Kao 2006), bronchopulmonary dysplasia (BPD) (Hays 2006), and prolonged length of hospital stay (LOS) (Hays 2006).

There are probably several factors that contribute to the development of hyperglycemia in the parenterally fed VLBW/ELBW neonate. It is commonly recommended that VLBW neonates be given sufficient nutrient intake in the early postnatal period to maintain growth at the intrauterine rate (AAP 1985; Ziegler 2002). In order to provide energy and support growth until full enteral feeding is established, these infants are usually given a parenteral nutrition regimen that delivers glucose at a rate substantially higher than the endogenous glucose production rate of 4 to 7 mg/kg/min reported for VLBW infants (Bier 1977; Cowett 1983). It is currently recommended that glucose infusion rates for parenterally fed VLBW/ELBW neonates should be incrementally increased from about 6 to about 10 mg/kg/min over the first postnatal week (Ehrenkranz 2007) and a recent survey of parenteral feeding prescriptions in the UK shows that this practice is commonly followed (Grover 2008). Secondly, the preterm neonate does not completely suppress endogenous glucose production in response to a parenteral glucose infusion (Cowett 1983; Sunehag 1994). Thirdly, in comparison to older infants and children, the VLBW neonate may have a reduced tolerance for intravenous glucose administration due to a limited amount of insulin-dependent tissue (specifically fat and muscle) and a limited insulin secretory response to glucose (Mitanchez-Mokhtari 2004).

Description of the intervention

Interventions to be reviewed that might be used to prevent hyperglycemia in the parenterally fed VLBW/ELBW neonate include:

  1. restriction of rate of advancement of parenteral glucose intake
  2. insulin infusion

How the intervention might work

Restriction of rate of advancement of parenteral glucose intake: In VLBW neonates there is a direct linear association between the glucose infusion rate and plasma glucose concentration (Cowett 1979; Cowett 1997; Sunehag 2004). Thus, restriction of the rate of advancement of parenteral glucose intake might be effective in preventing hyperglycemia, but with the potential risk of causing hypoglycemia. However, during total parenteral nutrition the blood glucose concentration does not appear to be sustained only by the rate of glucose infusion. Blood glucose concentration may be supported also by an unsuppressed rate of endogenous glucose production. This concept was advanced by Cowett 1983, following studies in which he suggested that there is a lack of precise control of glucose homeostasis in the neonatal period in contrast to that noted in the adult. This concept was confirmed by Chacko 2010 who reported a short-term study in eight 3 to 5 day VLBW pre-term infants receiving routine parenteral nutrition providing glucose at 7.4 to 11.4 mg/kg/min, along with parenteral lipid and amino acids. Blood glucose concentration ranged from 5.2 to 14.3 mM/L (94 to 257 mg/dL). Even though the glucose infusion rate exceeded the normal glucose production rate for newborn babies, endogenous glucose production was not suppressed. Ongoing gluconeogenesis was the major contributor to the rate of glucose production.

However, the potential longer-term nutritional disadvantages of glucose restriction also need to be considered. In the early postnatal period, many VLBW/ELBW infants cannot be fed enterally and must rely on parenteral feeding, either in whole or in part, in order to grow. The parenterally-fed neonate needs an energy intake of at least 80 to 90 kcal/kg/day in order to grow at the intrauterine rate (Zlotkin 1981; Heird 1992). The energy requirement for adequate growth may not be met if there is a restricted parenteral glucose intake depending on the amount of energy provided by the other components of the parenteral feeding regimen, i.e., amino acids or lipids. Sunehag 1999 showed that when the glucose infusion rate is reduced, the VLBW neonate may utilize part of the energy supplied by non-carbohydrate sources to maintain the blood glucose concentration in the euglycemic range. Chacko 2011 reported a short-term study in seven ELBW infants, post-natal age 3 to 4 days, who were receiving parenteral glucose, lipid and amino acids. Glucose infusion rate was reduced while maintaining the supply of parenteral amino acids and lipid at pre-study rates. In response to a reduction in glucose infusion rate the blood glucose concentration was reduced but remained within the normal range, mainly due to ongoing gluconeogenesis. Moreover, a reduced exogenous glucose supply did not increase proteolysis, at least in the short term. It is possible that a restricted rate of advancement of parenteral glucose, when offset by an increase in the administration of lipid and amino acids, may result in the prevention of hyperglycemia and reduce the risk of energy insufficiency. Moreover, in both parenterally fed adults (Floyd 1966) and VLBW neonates (Grasso 1968; Thureen 2003), parenteral amino acids have been shown to stimulate insulin release.

Insulin infusion: Insulin acts on cells throughout the body to stimulate uptake, utilization and storage of glucose. In particular, insulin stimulates the liver to store glucose in the form of glycogen and facilitates the entry of glucose into muscle and adipose tissue. There is some evidence that VLBW infants may have limited capacity to increase mature insulin secretion in response to an increase in blood glucose concentration (Mitanchez-Mokhtari 2004). The same investigator found that ELBW neonates did respond to exogenous insulin infusion, but needed higher doses to achieve a satisfactory response. These data suggest the possibility of insulin resistance. Levels of insulin-like growth factor I (IGF-I) are regulated by insulin in the newborn (Ogilvy-Stuart 1998); thus, insulin deficiency may contribute to impaired IGF-I generation. Low IGF-I levels have been implicated in the pathogenesis of ROP and brain growth (Löfqvist 2006).

In adult ICU patients, most of whom were not hyperglycemic at study entry, Van den Berghe 2001 used insulin infusions to control blood glucose concentration. She reported decreased risks for death and sepsis in the group in which glucose was kept at lower concentrations as compared to the group in which higher blood glucose concentrations were tolerated. It is possible that in VLBW infants at risk for hyperglycemia, insulin infusion might allow the increased utilization of glucose for energy and growth, and corresponding reductions in adverse clinical outcomes. However, close and even continuous monitoring of blood glucose concentrations would be required (Beardsall 2005), especially if insulin were infused in infants who were not hyperglycemic at the time.

For both interventions, a high amino acid intake might serve as an effect modifier, leading to increased growth in lean body mass, indicated by an increased rate of nitrogen accretion. If low nitrogen intake was limiting, an increase in nitrogen accretion might occur as a direct effect of high parenteral amino acid intake. If non-protein energy intake was limiting, an increase in nitrogen accretion might occur as an indirect effect, mediated through an increased endogenous insulin release in response to a high parenteral amino acid intake.This might result in a subsequent increase in glucose utilization, facilitating an increase in nitrogen accretion (Zlotkin 1981; Heird 1992). Thus, we plan to assess effect on nitrogen accretion in this review, and we plan subgroup analyses according to higher or lower level of parenteral amino acid intake.

Why it is important to do this review

The association of neonatal hyperglycemia with adverse clinical outcomes in the VLBW neonate, described previously in observational studies, does not necessarily mean that the hyperglycemia causes these sequelae. It is possible that it is the sicker neonate at higher risk for adverse clinical outcomes who is more prone to hyperglycemia. Furthermore, it is uncertain whether more good than harm is done by attempts to prevent hyperglycemia, whether by restricting the early postnatal advancement of parenteral glucose intake to a level at which energy intake is likely to be insufficient to support growth at the intrauterine rate or by continuous insulin infusion. Therefore, it is important to systematically review the evidence from randomized controlled trials of interventions for the prevention of hyperglycemia in VLBW neonates at risk for the development of neonatal hyperglycemia.

In a separate review, the evidence from randomized trials of interventions for the treatment of hyperglycemia in VLBW neonates has been systematically reviewed (Bottino 2009).

Objectives

Primary Objective:

To assess the effects of interventions for prevention of neonatal hyperglycemia in VLBW neonates receiving or intended to receive total or partial parenteral nutrition. Specific interventions that were reviewed were:

  1. restriction of rate of advancement of parenteral glucose intake (compared with no restriction in the rate of advancement of parenteral glucose intake)
  2. insulin infusion compared with:
    1. no restriction in the rate of advancement of parenteral glucose intake
    2. restriction of rate of advancement of parenteral glucose intake

Secondary Objectives:

Data permitting, subgroup analyses were to be carried out:

According to population:

  1. BW 500 to 749 g; 750 to 999 g; 1000 to 1499 g
  2. associated morbidity at study entry: high/low morbidity score or as reported by authors

According to intervention:

  1. higher, lower parenteral amino acid intake

[top]

Methods

Criteria for considering studies for this review

Types of studies

Randomized or quasi-randomized trials; parallel groups; allocation of individual patients; publication in any language. Randomized cross-over trials were to be excluded. Completed but unpublished trials or completed trials reported only in abstract form were eligible.

Types of participants

Neonates with birth weight < 1500 g or gestational age < 32 weeks, postnatal age up to 24 hours, receiving or intended to receive full or partial parenteral nutrition, without hyperglycemia at study entry

Types of interventions

  1. Restriction of rate of advancement of parenteral glucose intake
    Partial enteral feedings were allowed if available to both arms.
    Any levels of parenteral intakes of amino acids and lipids were allowed.
    Restriction of rate of advancement of parenteral glucose intake was defined as parenteral glucose intake < 6 mg/kg/min on the first day of life with intention of advance to < 8 mg/kg/min by day seven or as defined by authors.
    The comparison was no restriction in the advancement of parenteral glucose intake. No restriction was defined as targets for initial amount, rate of advancement or final amount by day seven which were higher than stated above, or no targets for restriction specified in advance or as defined by authors.
    High amino acid intake was defined as greater than/or equal to 3 g/kg/day on the first day of life and maintained at this level during the first week or as defined by authors.
    Low amino acid intake was defined as less than/or equal to 1.5 g/kg/day on the first day of life and gradual advancement with the intention to reach less than/or equal to 3 g/kg/day by day seven or as defined by authors.
  1. Insulin infusion (any dose, any dose adjustment protocol, any duration). The comparison was:
    1. No restriction in the rate of advancement of parenteral glucose intake as defined above.
    2. Restriction of rate of advancement of parenteral glucose intake as defined above.
    Partial enteral feedings were allowed if available to both arms.
    Any levels of parenteral intakes of amino acids and lipids were allowed.

Types of outcome measures

Primary outcomes
  1. All-cause mortality up to 36 weeks postmenstrual age (PMA) or as defined by authors
  2. Neurodevelopmental impairment defined as presence of one or more of the following: cerebral palsy, MDI or PDI < 70, blindness or deafness assessed either between 18 and 24 months postmenstrual age or at the latest assessment up to 24 months corrected age
  3. Growth assessed at 36 weeks PMA as measured by:
    1. Weight gain (g/kg/day) or as defined by authors
    2. Head circumference (cm/wk) or as defined by authors
    3. Length (cm/wk) or as defined by authors
  4. Blood glucose concentrations
    1. Mean glucose concentration (mM/L)
    2. Proportion of infants developing hyperglycemia by day seven of life defined as whole blood or plasma glucose concentration > 8.3 mM/L (> 150 mg/dL) or as defined by authors
Secondary outcomes
  1. Caloric intake assessed as kcal/kg/day in one week period or longer or as reported by authors
  2. Cumulative parenteral glucose intake (g/kg up to day seven of life) or as reported by authors
  3. Number of episodes of hypoglycemia per patient and/or proportion of neonates having one or more episodes; defined as whole blood or plasma glucose concentration < 2.5 mM/L (< 45 mg/dL) or as defined by authors
  4. Nitrogen accretion (mg/kg/day)
  5. Severe intraventricular hemorrhage (IVH) defined as grade 3 or 4 by Papile classification assessed at 36 weeks PMA (Papile 1978)
  6. Incidence of retinopathy of prematurity (ROP) to latest follow-up: a) any stage; b) requiring treatment
  7. Proportion of neonates with one or more episodes of sepsis (bacterial or fungal) defined as a positive culture in blood, urine or cerebrospinal fluid up to 36 weeks postmenstrual age or as defined by authors
  8. Incidence of necrotizing enterocolitis (NEC) defined as stage 2 or above by the Bell classification assessed at 36 weeks PMA (Bell 1978)
  9. Incidence of chronic lung disease (CLD) at 36 weeks postmenstrual age (Jobe 2001)
  10. Length of hospital stay defined as number of days until discharge home or as reported by authors
  11. Levels of IGF-I (ng/mL)

Search methods for identification of studies

The standard search strategy of the Neonatal Review Group as outlined in the Cochrane Library was used. The following sources were searched for eligible reports in any language:

Electronic searches

Electronic databases that were searched included: The Cochrane Central Register of Controlled Trials (CENTRAL), The Cochrane Library, Issue 4 of 12, 2011. MEDLINE (1966 to April 2011), EMBASE (1980 to April 2011) and CINAHL (1982 to November 2008).

The search string for searching CENTRAL, and MEDLINE via PubMed, included the following terms: (glucose/administration and dosage OR glucose infusion OR intravenous glucose OR energy intake OR insulin/administration and dosage OR insulin/therapeutic use) AND (infant, very low birth weight OR very low birth weight OR VLBW OR extremely low birth weight OR ELBW OR preterm). To attempt to limit to clinical trials, we used the maximum sensitivity methodologic filter for questions of therapy implemented in PubMed Clinical Queries (Haynes 2005).

We used a similar search string for searching EMBASE via Ovid and CINAHL via Ebsco, adapting the search terms to the structured vocabulary and syntax required for those databases; and adapting the limits according to what was available in those databases.

Searching other resources

Abstracts presented at the annual meetings of The European Society for Pediatric Research (ESPR) 2005 to 2010 and Pediatric Academic Societies (PAS) 2000 to 2011 were searched from the journal Pediatric Research and abstracts on line.

On-going trials were searched in April 2011 at the following websites: ClinicalTrials.gov, Controlled-Trials.com External Web Site Policy, and WHO International Clinical Trials Registry Platform (ICTRP) External Web Site Policy.

Using Web of Science, we did a citation search of Gilbertson 1991 in November 2008.

Data collection and analysis

Selection of studies

The titles and abstracts of reports that were detected by the described search strategies were assessed independently by two review authors (MB and JS) to determine their eligibility for inclusion in this review. Eligibility for inclusion was judged according to the criteria listed under Criteria for considering studies. If there was uncertainty as regards inclusion/exclusion, the full report was obtained in order to make a decision re eligibility. Any disagreement was resolved by discussion. Unresolved disagreements were to be referred to RC for arbitration.

Data extraction and management

For included studies, data were extracted concerning study design, methodology, clinical features of the population, interventions and outcomes, and treatment effects, using specially designed data collection forms. For studies that were initially considered possibly eligible for inclusion but which were excluded after reading the full report, the reason for exclusion was documented. All data were extracted independently by two review authors (MB and JS), compared, and any discrepancies resolved by discussion or, if necessary, through contact with the primary investigators. Unresolved disagreements were to be referred to RC for arbitration. Categorical outcomes were expressed as the negative outcome, e.g. death, not survival. We attempted to obtain data sets that were as complete as possible. We requested from the primary investigators any unreported data on study outcomes, as necessary. To the extent possible, we extracted outcome data on all patients randomized.

Assessment of risk of bias in included studies

Two review authors (JS and MB) independently assessed the risk of bias for each study, in general accordance with the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Any disagreement was resolved by discussion.

Each trial was assessed for blinding of randomization, blinding of intervention(s), complete follow-up, and blinding of outcome(s) measurement, using a 3-category scale: Yes (low risk of bias); No (high risk of bias); Can't tell (unclear risk of bias). The risk of bias arising from selective reporting was also assessed if studies prospectively registered and published the protocol.

Measures of treatment effect

For individual trials, effect measures for categorical outcomes included relative risk (RR) and absolute risk difference (RD), each with its 95% confidence interval (CI). For statistically significant effects, number needed to treat (NNT) or number needed to harm (NNH) were calculated. For continuous outcomes, the effect measure was mean difference (MD) with its 95% CI. For any meta-analyses (see below), for categorical outcomes we planned to calculate typical estimates of RR and RD, each with its 95% CI; for continuous outcomes we planned to calculate the weighted mean difference (WMD) with its 95% CI.

Unit of analysis issues

We did not anticipate any unit of analysis issues. Cross-over trials were not eligible. Cluster randomized trials were not expected in this field. For outcomes that can occur more than once in the same patient, such as episodes of hypoglycemia, we analyzed the data as proportion of neonates having one or more episodes.

Dealing with missing data

If some outcome data remained missing despite our attempts to obtain complete outcome data, we performed an available-case analysis, based on the numbers of patients for whom outcome data were known. For primary outcomes, if there were instances of statistically significant effects but with missing data, we planned to perform a worst/best case sensitivity analysis based on imputation to test whether the effect was sustained or overturned.

Assessment of heterogeneity

Before any meta-analysis was done, we planned to judge whether there was sufficient similarity between the eligible studies in their design features and clinical features (population, interventions) to make pooling for meta-analysis scientifically and clinically credible. If there was sufficient similarity, we planned to proceed to meta-analysis. If not, the results of individual trials were to be described separately.

We planned to estimate the amount of heterogeneity of treatment effect across trials in a meta-analysis using the I-squared statistic. We planned to test whether heterogeneity was statistically significant using the chi-squared statistic. If substantial heterogeneity was present, we planned to explore its source(s) post facto, considering differences in design or clinical features of the trials.

Assessment of reporting biases

We did not anticipate a sufficient number of included trials in this field to permit assessment of possible publication bias and other biases using symmetry/asymmetry of funnel plots.

For included trials, we planned to explore possible selective reporting of study outcomes by comparing the primary and secondary outcomes in the reports with the primary and secondary outcomes nominated at trial registration, using the websites ClinicalTrials.gov and Controlled-Trials.com External Web Site Policy. If such discrepancies were found, we planned to contact the primary investigators to try to obtain missing outcome data on outcomes pre-specified at trial registration.

Data synthesis

If meta-analysis was judged to be appropriate, it was to be done using RevMan 5, supplied by the Cochrane Collaboration. For estimates of typical relative risk and risk difference, we planned to use the Mantel-Haenszel method. For measured quantities, we planned to use the inverse variance method. All meta-analyses were to be done using the fixed effect model. When meta-analysis was judged to be inappropriate, individual trials were to be analyzed and interpreted separately.

Subgroup analysis and investigation of heterogeneity

Data permitting, pre-specified subgroup analyses were done according to population and intervention subgroups as defined in Secondary Objectives. If data were not available to permit a pre-planned subgroup analysis to be done, that was stated as a result in this review. Any post factosubgroup analyses, e.g. to explore unanticipated sources of heterogeneity, were to be labelled as such. It was not anticipated that a sufficient large number of eligible trials would be available to permit exploration of heterogeneity of treatment effect using meta-regression.

Sensitivity analysis

As noted previously, in the case of missing outcome data, a worst case/best case analysis was planned to test whether a significant result was sensitive to (i.e. overturned by) imputations according to worst/best case assumptions. There were no other pre-planned sensitivity analyses.

[top]

Results

Description of studies

Results of the search

We detected 27 possibly eligible studies. Of these, 22 were excluded from this review. Excluded studies are listed in the table 'Characteristics of excluded studies' along with the reason for exclusion. One study, Hawk 2009, reported only in abstract form at present, was categorized as awaiting classification. (See table 'Characteristics of studies awaiting classification'.) Four studies were eligible for inclusion in this review: Gilbertson 1991; Pappoe 2009; Beardsall 2007; Beardsall 2008.

Included studies

Two of the four included studies (Gilbertson 1991; Pappoe 2009) compared restriction versus no restriction in the rate of advancement of parenteral glucose intake. Two of the included studies (Beardsall 2007; Beardsall 2008) compared insulin infusion with standard care (each arm having no restriction of the rate of advancement of parenteral glucose intake). None of the included trials compared insulin infusion with restriction in the rate of advancement of parenteral glucose intake.

Restriction versus no restriction of rate of advancement of parenteral glucose intake
Early versus delayed introduction of parenteral lipid

Gilbertson 1991 compared early (day 1) versus delayed (day 8) introduction of lipid in parenterally fed VLBW neonates. In the early-lipid arm, there was a compensatory reduction in the prescribed parenteral glucose intake. This was intended to result in isocaloric non-protein energy intakes in the two groups over the first seven days. The details of this trial are given in the Table "Characteristics of Included Studies". Included in the trial were 29 ventilator-dependent VLBW neonates of < 6 hours postnatal age with an estimated need of total parenteral nutrition (TPN) for at least one week. In the study arm providing early lipid and restricted glucose (n = 16), glucose was to be infused at a rate of 2.8 mg/kg/min on day one with step-wise increases to a rate of 7.9 mg/kg/min by days 4 to 7. Intravenous lipid was started at 1 g/kg/day on day one and was increased progressively to 3 g/kg/day by days 4 to 7. In the study arm providing delayed lipid and more liberal glucose (n = 13), glucose was to be infused at a rate of 4.5 mg/kg/min on day one with step-wise increases to a rate of 13.2 mg/kg/min by days 4 to 7. Neonates in this arm received no parenteral lipid until day 8. Parenteral amino acids were provided in each arm of the trial and were increased progressively to reach 2.6 g/kg/day by day four. If hyperglycemia developed, it was treated by reducing the glucose concentration of the infusate. Hypoglycemia was treated with intravenous dextrose 0.6 g/kg. Outcomes reported in this trial included measures of intravenous lipid tolerance; measures of glucose homeostasis (mean blood glucose concentration, insulin concentration, hyperglycemia, hypoglycemia); gains in weight, length and head circumference; death; and major morbidities including bronchopulmonary dysplasia, septicemia, periventricular hemorrhage, necrotizing enterocolitis, patent ductus arteriosus and retinopathy of prematurity.

Slow versus rapid rate of advancement of parenteral nutrition

Pappoe 2009 randomized 42 neonates of birth weight 600 to 1200 g to slow or rapid advancement of parenteral glucose, lipids and amino acids over the first week of life. In the slow-advancement group (which provided restricted parenteral glucose intake), the goal was to reach a non-protein energy intake of 75 to 80 kcal/kg/day (glucose:lipid ratio approximately 60:40) by day six. In the rapid-advancement group (which provided a liberal parenteral glucose intake), this target was to be reached by day three. Only the 32 neonates randomized in the 600 to 800 g and 801 to 1000 g strata were eligible for inclusion in this review; we excluded the stratum 1001 to 1200 g because in this stratum the targets for initial glucose infusion rate and its advancement over the first four days did not indicate lower target glucose infusion rates in the slow-advancement group. Details of this trial are given in the table "Characteristics of Included Studies".

In the glucose-restricted arm of Pappoe 2009 (n = 15 included in this review) neonates were started on 5% dextrose/water (D/W) on day one, to be increased to 7.5% and 10% D/W on subsequent days only if blood glucose concentration remained below 8.3 mM/L (150 mg/dL). On 5% D/W the target glucose infusion rates were 2.6 to 2.8 mg/kg/min on day one, rising to 4.0 to 4.3 mg/kg/min by day four and thereafter. If the dextrose infusion was able to be advanced to a 7.5% or 10% concentration the target glucose infusion rates would have been correspondingly higher: e.g. using 7.5% D/W, about 6 mg/kg/min by day six. Parenteral lipids and amino acids were each provided at 1 g/kg/day on day one and increased by 0.5 g/kg/day to a maximum of 3.5 g/kg/day by day six. Glucose intolerance was treated initially by decreasing the glucose infusion and then, if it persisted, with an insulin infusion.

In the unrestricted arm (n = 17 included in this review) neonates were started on 10% D/W immediately on day one. The target glucose infusion rates were 4.9 to 5.3 mg/kg/min on day one, rising to 7.8 to 8.3 mg/kg/min by day four. Lipids and amino acids were each provided at 2 g/kg/day on day 1, 3.0 g/kg/day on day two and 3.5 g/kg/day on day three and thereafter. Glucose intolerance was treated with an insulin infusion.

Outcomes reported in this trial included postnatal weight loss, weight gain in seven days, days to regain birth weight, glucose intolerance treated with insulin infusion, hypertriglyceridemia, azotemia, intraventricular hemorrhage, retinopathy of prematurity, patent ductus arteriosus and death.

Insulin infusion versus standard care

Beardsall 2007 was a small pilot randomized trial, conducted at two centres, to test the hypothesis that elective insulin infusion with dextrose support begun on day one and continued for seven days would reduce hyperglycemia and increase IGF-1 levels in VLBW neonates requiring intensive care. The trial enrolled 17 VLBW neonates < 24 hours postnatal age. A continuous glucose monitoring sensor (CGMS) was inserted subcutaneously into the thigh in each participant to allow continuous monitoring of glucose concentrations throughout the first seven days of life. The details of this trial are given in the Table "Characteristics of Included Studies".

Neonates randomized to the insulin group (n = 8 analyzed) received a continuous fixed-dose insulin infusion at 0.025 U/kg/h. A separate IV infusion set containing 20% dextrose was prepared, to be started immediately in the event of hypoglycemia. Algorithms to guide the response to low or high glucose values included instructions for adjustment of the insulin infusion rate, the use of 20% dextrose support, and adjustment to the rate of dextrose infusion in the parenteral nutrition. Neonates randomized to the control group (n = 8) received standard care. In this group, insulin was used only to treat documented blood glucose levels > 10.0 mM/L (> 180 mg/dL), using a sliding-scale infusion. In both groups, parenteral nutrition (including amino acids and lipid as well as dextrose) was begun on the first day after birth. No targets for dextrose infusion rates were stated. We estimated that by day seven the mean dextrose infusion rate in the standard care group exceeded 8 mg/kg/min. Enteral feeds were started when possible within the first few days with either expressed own mother´s milk or donated breast milk. Outcomes reported in this trial included CGMS results [duration of time over the first seven days with hyperglycemia > 10.0 mM/L (> 180 mg/dL), with hypoglycemia < 2.6 mM/L (< 47 mg/dL), within the target range of 4 to 8 mM/L (72 to 144 mg/dL)]; blood glucose concentrations; dextrose infusion rate; total caloric intake; gains in weight, lower leg length and head circumference; blood levels of proinsulin, insulin, IGF-1, IGF-1 binding protein; and death before discharge.

Beardsall 2008 was a moderately large multicentre trial (Neonatal Insulin Replacement Therapy in Europe, NIRTURE) that was conducted in 8 neonatal intensive care centres. It aimed to determine whether early insulin infusions in VLBW neonates requiring intensive care begun within 24 h of birth and continued for seven days would reduce hyperglycemia during the first postnatal week and reduce the incidence of death and major morbidities associated with hyperglycemia. In order to assess the efficacy of glucose control, continuous monitoring of glucose levels throughout the first seven days was achieved by the use of a continuous glucose monitoring sensor (CGMS) inserted subcutaneously in the thigh of each participant. 389 participants were enrolled, allocated either to insulin infusion (n = 195) or standard care (n = 194) using minimization to achieve balance on major prognostic factors at study entry. The details of this trial are given in the table 'Characteristics of Included Studies'.

In the insulin group, a fixed-dose continuous infusion of insulin was used (0.05 U/kg/h) with intravenous glucose support as needed using a 20% dextrose solution. The goal was to maintain euglycemia (target range 4 to 8 mM/L, 72 to 144 mg/dL). Blood glucose determinations were performed hourly after insulin was started and subsequently every six hours after glucose concentrations had stabilized. Algorithms guided the responses to be taken in the event of high or low blood glucose values. These included instructions concerning the use of 20% dextrose support, adjustment of the insulin infusion, and adjustment of the glucose infused via the parenteral nutrition infusate. In the standard care group, blood glucose concentrations were measured at least 3 times daily, and more frequently if clinically indicated. If blood glucose was > 10.0 mM/L (> 180 mg/dL), the physician responsible for clinical care would determine whether the rate of infusion of glucose should be decreased, and/or if insulin treatment should be started. If blood glucose dropped below 2.6 mM/L (< 47 mg/dL), the physician would determine whether the rate of infusion of glucose via the parenteral nutrition infusate should be increased. In both groups, parenteral nutrition with dextrose/water was begun on day one and was increased over the first week. There were no specified targets for either initial dose or rate of advancement of parenteral nutrition that were imposed by the trial protocol. However, neonates in the standard care group received on average a glucose infusion of about 5.5 mg/kg/min on day one and about 9.7 mg/kg/min by days 4 to 7. The observed glucose infusion rates in the insulin group were similar to the standard care group on day one, but somewhat higher by days 4 to 7. In both groups, the observed infusion rates of amino acids and lipid via parenteral nutrition were similar, but low.

The primary outcome in Beardsall 2008 was mortality before the expected day of delivery. Secondary outcomes were sepsis, somatic growth, necrotizing enterocolitis, retinopathy of prematurity, intracranial disease, mortality by 28 days, and number of days of intensive care. Other outcomes that were reported included several measures of glucose control over the first week: mean glucose concentrations, proportion of neonates with glucose > 10.0 mM/L (> 180 mg/dL) for > 10% of time during the first week, percent of time with hyperglycemia > 10 mM/L (> 180 mg/dL) by day, proportion with an episode of hypoglycemia < 2.6 mM/L (< 47 mg/dL) for > 60 min, percent of time in hypoglycemic range < 2.6 mM/L (< 47 mg/dL) by day.

Beardsall 2008 aimed to recruit 500 participants. However, recruitment was suspended after 389 participants had been enrolled, based on the recommendation of the trial data safety and monitoring committee. The trial steering committee subsequently recommended that the trial be discontinued. The reasons for these decisions are given in the table 'Characteristics of Included Studies', Notes column.

Excluded studies

See table, 'Characteristics of excluded studies'.

Risk of bias in included studies

The included trials were of varying methodologic quality. Our ratings of methodologic quality are given in the table, 'Characteristics of Included Studies'.

Gilbertson 1991 used alternate allocation, a quasi-random method of treatment group assignment. Completeness of follow-up suffered from the post-allocation exclusion from analysis of three participants who required TPN for less than a week. The numbers included in the reported analyses were 16 (early lipid, restricted glucose) and 13 (delayed lipid, unrestricted glucose). Since allocation was alternate, it is likely that all three exclusions were from the delayed lipid (unrestricted glucose) group.

Pappoe 2009 used random allocation with concealment by opaque envelopes. Caretakers were not blinded to the intervention, slow versus rapid advancement of parenteral nutrition. Outcome assessors were not blinded to treatment group. Follow-up was almost complete.

Beardsall 2007 used random allocation with concealment by opaque envelopes. Blinding of caretakers to the intervention, insulin infusion, was clinically not feasible. However, the continuous glucose monitoring system (CGMS) readings could not be viewed in real time and were downloaded only after completion of the seven day study period. Thus, that information did not influence clinical management during the first week. Follow-up was almost complete.

The multicentre trial of Beardsall 2008 allocated participants by minimization using an internet-based computer program in order to reduce variability at study entry arising from centre, birth weight and gestational age. Blinding of caretakers to the intervention, insulin infusion, was as described for Beardsall 2007 (see above). Follow-up was virtually complete: 386 of 389 entrants were included in intention-to-treat analyses. Blinding of outcome ascertainment was accomplished for intracranial disease, necrotizing enterocolitis and retinopathy of prematurity. For this and all trials, we did not regard blinding as an applicable quality criterion for the ascertainment of the outcome all-cause mortality. There was no evidence of selective reporting.

Effects of interventions

Restriction versus no restriction of rate of advancement of parenteral glucose intake

Two trials contributed to this comparison (Gilbertson 1991, Pappoe 2009). Each was of small size. Gilbertson 1991 included 29 VLBW neonates. Pappoe 2009 included 42 neonates of birth weight 600 to 1200 g; however, only the 32 neonates in the strata 600 to 800 g and 801 to 1000 g were eligible for inclusion in this review. Because of differences between the two trials in the interventions tested, meta-analysis was judged not to be clinically appropriate. Thus, their results are reported separately.

1. Early versus delayed introduction of parenteral lipid

Gilbertson 1991 compared parenteral lipid beginning on day one with no parenteral lipid until day eight. The early-lipid group received non-protein energy as both glucose and lipid from day one, whereas the late-lipid group was targeted to receive an approximately isocaloric amount of non-protein energy entirely as glucose during the first week. Thus, the early-lipid policy included a component of restriction of parenteral glucose intake during the first week.

Primary outcomes
1.1 Mean blood glucose, days 1-7 (mM/L)

Gilbertson 1991 found that the early-lipid group had significantly lower blood glucose concentrations over days 1 to 7 compared to the delayed-lipid group: overall mean 6.01 mM/L (108 mg/dL) in the early-lipid group versus 7.50 mM/L (135 mg/dL) in the delayed-lipid group. Mean difference (MD) -1.49 mM/L (95% CI -2.50 to -0.48).

1.2 Hyperglycemia > 8.0 mM/L

Gilbertson 1991 reported the number of neonates who developed hyperglycemia > 8.0 mM/L (> 144 mg/dL): 6/16 (37.5%) in the early-lipid group versus 8/13 (61.5%) in the delayed-lipid group. The reduction in hyperglycemia was not statistically significant: RR 0.61 (95% CI 0.28 to 1.31).

1.3 All-cause mortality

1.3.1 Death in neonatal period

In Gilbertson 1991, death in the neonatal period occurred in 1/16 (early lipid) versus 2/13 (delayed lipid). The effect was not statistically significant: RR 0.41 (95% CI 0.04 to 4.00).

Neurodevelopmental impairment

Gilbertson 1991 did not report this outcome.

1.4 Growth

1.4.1 Weight gain in first week (g)

Gilbertson 1991 reported that neonates in both groups lost weight during the first week, but that the early-lipid group lost significantly less weight than did the delayed-lipid group: MD 66 g cumulative weight loss (95% CI 44 to 87). The investigators interpreted the difference as likely reflecting a difference in fluid balance.

1.4.2 Length gain (cm/wk)

Gilbertson 1991 reported no significant difference between groups in length gain.

1.4.3 Head circumference gain (cm/wk)

Gilbertson 1991 reported no significant difference between groups in gain in head circumference.

1.5 Days to regain birth weight

Gilbertson 1991 reported no significant difference between groups in the number of days to regain birth weight, mean 10.1 days (early lipid) versus 11.4 days (delayed lipid). MD -1.3 days (95% CI -5.9 to 3.3).

Secondary outcomes
Caloric intake

Glucose infusion rate, days 1-7 (mg/kg/min)

Gilbertson 1991 reported that the glucose infusion rate actually achieved over days 1 to 7 averaged 5.1 mg/kg/min in the early-lipid group versus 8.4 mg/kg/min in the delayed-lipid group (data not shown). The glucose infusion rate was lower by 35 to 40% in the early-lipid group on each day throughout the whole of the first week.

Non-protein energy intake, days 1-7 (kcal/kg/day)

Gilbertson 1991 calculated average non-protein energy (NPE) intakes over days 1 to 7 as 51.4 kcal/kg/day in the early-lipid group versus 48.7 kcal/kg/day in the delayed-lipid group (data not shown). However, they used a factor of 4.0 kcal/g glucose. Corrected using a factor of 3.4 kcal/g for hydrated glucose, the corresponding NPE intakes averaged 46.6 and 41.0 kcal/kg/day in the early-lipid and delayed-lipid groups respectively. Although an isocaloric comparison was intended, a somewhat higher NPE intake was seen in the early-lipid group. This was due to the fact that in the early-lipid group the lesser energy intake from glucose was more than offset by a greater energy intake from lipid.

Nitrogen accretion

Gilbertson 1991 did not report this outcome.

1.6 Hypoglycemia < 2.2 mM/L

Gilbertson 1991 reported the number of neonates who developed hypoglycemia < 2.2 mM/L (< 40 mg/dL): 7/16 (43.8%) in the early-lipid group versus 5/13 (38.5%) in the delayed-lipid group. The effect was not statistically significant: RR 1.14 (95% CI 0.47 to 2.75).

Hypertriglyceridemia > 1.5 mM/L

Gilbertson 1991 reported hypertriglyceridemia > 1.5 mM/L (> 133 mg/dL) in 3/16 (19%) of the early-lipid group versus 1/13 (8%) in the delayed-lipid group (data not shown). The trend to higher incidence in the early-lipid group reflects the fact that only the early-lipid group received parenteral lipids during the first week. Hypertriglyceridemia was not a pre-specified outcome in our protocol.

1.7 Intraventricular hemorrhage

1.7.1 Grade not stated

In Gilbertson 1991, intraventricular hemorrhage, grade not stated, occurred in 5/16 (early lipid) versus 7/13 (delayed lipid). The effect was not statistically significant: RR 0.58 (95% CI 0.24 to 1.40).

1.8 Retinopathy of prematurity

1.8.1 Stage not stated

In Gilbertson 1991, retinopathy of prematurity, stage not stated, was reported in 0/16 (early lipid) versus 1/13 (delayed lipid). The effect was not statistically significant: RR 0.27 (95% CI 0.01 to 6.23).

1.9 Sepsis

1.9.1 Positive blood culture before discharge

Gilbertson 1991 reported a positive blood culture before discharge in 2/16 (early lipid) versus 2/13 (delayed lipid). The effect was not statistically significant: RR 0.81 (95% CI 0.13 to 5.01).

1.10 Necrotizing enterocolitis

Gilbertson 1991 reported necrotizing enterocolitis in 1/16 (early lipid) versus 1/13 (delayed lipid). The effect was not statistically significant: RR 0.81 (95% CI 0.06 to 11.77).

1.11 Chronic lung disease

Gilbertson 1991 reported chronic lung disease in 2/16 (early lipid) versus 3/13 (delayed lipid). The effect was not statistically significant: RR 0.54 (95% CI 0.11 to 2.77).

For Gilbertson 1991, our planned subgroup analyses according to birth weight subgroups, morbidity score or level of parenteral amino acid intake could not be done, for lack of available data.

2. Slow versus rapid rate of advancement of parenteral nutrition

Pappoe 2009 compared slow versus rapid advancement of parenteral glucose, lipids and amino acids in the first week of life. The goals were to reach a non-protein energy intake of 75 to 80 kcal/kg/day by either day six (slow advancement) or day three (rapid advancement). Thus, the policy of slow advancement of parenteral nutrition included a component of parenteral glucose restriction, particularly in the first 3 to 4 days. Although Pappoe 2009 enrolled neonates of 600 to 1200 g, only those of 600 to 1000 g were eligible for inclusion in this review (see table, 'Characteristics of Included Studies'). Thus, in the data tables and text of this review, Pappoe 2009 refers to neonates of birth weight 600 to 1000 g.

Primary outcomes

2.1.1 Hyperglycemia > 11.1 mM/L

In Pappoe 2009, the number of neonates who developed any episode of hyperglycemia > 11.1 mM/L (> 200 mg/dL) was 7/15 (47%) in the slow-advance group and 14/17 (82%) in the rapid-advance group (data supplied by Dr Pappoe). The reduction in hyperglycemia in the slow-advance group was of borderline statistical significance: RR 0.57 (95% CI 0.32 to 1.02; RD -0.36 (95% CI -0.67 to -0.05); number needed to treat (NNT) 3.

2.2.1 Hyperglycemia with insulin treatment

Pappoe 2009 found that hyperglycemia treated with insulin was significantly less frequent in the slow-advance group (2/15, 13%) compared to the rapid-advance group (11/17, 65%). RR 0.21 (95% CI 0.05 to 0.78); RD -0.51 (95% CI -0.80 to -0.23); NNT 2. The lower incidence of hyperglycemia with insulin treatment in the slow-advance group is at least in part a reflection of the fact that the study protocol specified that the first response to hyperglycemia in the slow-advance group was to decrease the glucose infusion rate, whereas in the rapid-advance group it was to give an insulin infusion.

2.3.1 All-cause mortality: death before discharge

In Pappoe 2009, death before discharge occurred in 1/15 in the slow-advance group and 1/17 in the rapid-advance group. The effect was not statistically significant: RR 1.13 (95% CI 0.08 to 16.59).

Neurodevelopmental impairment

Pappoe 2009 did not report this outcome.

Growth

2.4.1 Weight gain, birth to 7 days

Pappoe 2009 found that neonates in the slow-advance group had, on average, a cumulative weight loss of 27 g over the first week. Those in the rapid-advance group averaged a cumulative weight gain of 30 g. The difference was statistically significant: Mean difference (MD) -57 g (95% CI -100 to -14). Thus, the slow-advance group had on average 57 g less weight gain than the rapid-advance group over the first week.

2.5.1 Days to regain birth weight

Pappoe 2009 reported that the slow-advance group took on average 8.2 days to regain birth weight versus 6.4 days in the rapid-advance group. The difference did not reach statistical significance: MD 1.9 days (95% CI -0.8 to 4.5). We did not pre-specify this outcome in our protocol.

2.6.1 Percent weight loss

Pappoe 2009 found that percent weight loss averaged 8.5% in the slow-advance group and 5.0 % in the rapid-advance group. The difference did not reach statistical significance: MD 3.5% (95% CI -0.5 to 7.5). We did not pre-specify this outcome in our protocol.

2.7.1 Failure to reach birth weight by day 7

Pappoe 2009 found that failure to regain birth weight by day 7 occurred in 9/15 (60%) of the slow-advance group versus 7/17 (41%) of the rapid-advance group. The effect was not statistically significant: RR 1.46 (95% CI 0.72 to 2.94). We did not pre-specify this outcome in our protocol.

Secondary outcomes
Caloric intake
2.8.1 Glucose infusion rate, days 1 to 7, mg/kg/min

Pappoe 2009 found that the average glucose infusion rate over the first seven days was lower in the slow-advance group (mean 5.76 mg/kg/min) compared with the rapid-advance group (mean 6.54 mg/kg/min). The difference was statistically significant: MD -0.78 mg/kg/min (95% CI -1.53 to -0.03). The reduction in the glucose infusion rate over the first week in the slow-advance group was almost entirely accounted for by a reduction during the first three days (data not shown, supplied by Dr Pappoe).

2.9.1 Non-protein energy intake, days 1 to 7, kcal/kg/day

Pappoe 2009 found that the average parenteral non-protein energy intake over the first seven days was lower in the slow-advance group (mean 46 kcal/kg/day) than in the rapid-advance group (mean 58 kcal/kg/day). The difference was statistically significant: MD -12 kcal/kg/day (95% CI -17 to -7).

Nitrogen accretion

Pappoe 2009 did not report this outcome.

2.10.1 Hypoglycemia

In Pappoe 2009, hypoglycemia occurred in 3/15 (20%) of the slow-advance group versus 1/17 (6%) of the rapid-advance group. The effect was not statistically significant: RR 3.40 (95% CI 0.39 to 29.31).

2.11.1 Hypertriglyceridemia > 1.7 mM/L

In Pappoe 2009, hypertriglyceridemia > 1.7 mM/L (> 150 mg/dL) was reported in 3/15 (20%) of the slow-advance group versus 12/17 (59%) of the rapid-advance group. The effect was statistically significant: RR 0.28 (95% CI 0.10 to 0.82); RD -0.51 (95% CI -0.80 to -0.21); NNT 2. This outcome was not pre-specified in our protocol.

2.12.1 Azotemia BUN > 14.3 mM/L

In Pappoe 2009, azotemia (BUN > 14.3 mM/L, > 40 mg/dL) was reported in 1/15 (7%) in the slow-advance group versus 3/17 (18%) in the rapid-advance group. The effect was not statistically significant: RR 0.38 (95% CI 0.04 to 3.26). We did not pre-specify this outcome in our protocol.

2.13.1, 2.14.1 Intraventricular hemorrhage, Any grade; > Grade 2

In Pappoe 2009, intraventricular hemorrhage, any grade, occurred in 4/15 (27%) in the slow-advance group and 7/17 (41%) in the rapid-advance group. The effect was not statistically significant: RR 0.65 (95% CI 0.24 to 1.78). Intraventricular hemorrhage > grade 2 occurred in 1/15 (7%) of the slow-advance group and 4/17 (24%) of the rapid-advance group. Again, the effect was not statistically significant: RR 0.28 (95% CI 0.04 to 2.26).

2.15.1, 2.16.1 Retinopathy of prematurity, Any stage; Stage 2 or more with surgery

In Pappoe 2009, retinopathy of prematurity, any stage, occurred in 13/15 (87%) of the slow-advance group and 15/17 (88%) of the rapid-advance group. The effect was not statistically significant: RR 0.98 (95% CI 0.75 to 1.28). Retinopathy of prematurity, stage 2 or more with surgery, occurred in 3/15 (20%) of the slow-advance group and 2/17 (12%) of the rapid-advance group. Again, the effect was not statistically significant: RR 1.70 (95% CI 0.33 to 8.84).

Sepsis

This outcome was not reported by Pappoe 2009.

Necrotizing enterocolitis

This outcome was not reported by Pappoe 2009.

Chronic lung disease

This outcome was not reported by Pappoe 2009.

2.17 Days neonatal intensive care

Pappoe 2009 found that the duration of neonatal intensive care averaged 89 days in the slow-advance group and 78 days in the rapid-advance group. The difference was not statistically significant: MD 11 days (95% CI -7 to 29).

Slow versus rapid rate of advancement of parenteral nutrition: Subgroup analyses

Data tables 2.1 through 2.16 include subgroup analyses of Pappoe 2009 according to birth weight strata 600 to 800g, 801 to 1000 g. Statistically significant treatment effects within subgroup were found only for the following outcomes:

2.4.2 Weight gain, birth to 7 days

Weight gain from birth to 7 days was significantly lower in the slow-advance group compared to the rapid-advance group among neonates in the 600 to 800 g stratum: MD -76 g (95% CI -136 to -16). There was no significant effect on weight gain in the 801-1000 g stratum (2.4.3).

2.6.2 Percent weight loss

This was significantly increased in the slow-advance group among neonates in the 600 to 800 g stratum: MD 5.9% (95% CI 1.6 to 10.2). There was no significant effect on percent weight loss in the 801 to 1000 g stratum (2.6.3)

2.8.2 Glucose infusion rate, days 1 to 7, mg/kg/min

This was significantly lower in the slow-advance group among neonates in the 600 to 800 g stratum: MD -1.23 mg/kg/min (95% CI -2.19 to -0.27). There was no significant difference in glucose infusion rate in the 801 to 1000 g stratum (2.8.3).

2.9.2, 2.9.3 Non-protein energy intake, days 1-7, kcal/kg/day

This was significantly lower in the slow-advance group in each stratum: 600 to 800 g, MD -16 kcal/kg/day (95% CI -23 to -9); 801 to 1000 g, MD -7 kcal/kg/day (95% CI -13 to -1). Substantial heterogeneity of treatment effect was noted: the size of reduction in non-protein energy intake in the slow-advance group was considerably greater in the 600 to 800 g than the 801 to 1000g stratum: I squared was 74% (data not shown).

2.11.2, 2.11.3 Hypertriglyceridemia > 1.7 mM/L

This was significantly reduced in the slow-advance group in the 600 to 800 g stratum: RR 0.42 (95% CI 0.18 to 0.97); RD -0.63 (95% CI -1.00 to -0.25). In the 801 to 1000 g stratum, the reduction of hypertriglyceridemia in the slow-advance group was of borderline statistical significance.

For Pappoe 2009, our planned subgroup analyses according to morbidity score or level of parenteral amino acid intake could not be done, for lack of available data.

Insulin infusion versus no restriction of rate of advancement of parenteral glucose intake

3. Insulin infusion versus standard care

Two trials contributed to this comparison (Beardsall 2007; Beardsall 2008). Beardsall 2007 was a small pilot study in preparation for the multicentre trial, NIRTURE (Beardsall 2008). There were no opportunities for meta-analysis of these two trials, so their results are reported separately.

Primary Outcomes
3.1.1 Mean glucose concentration, first week (mM/L)

Beardsall 2007 found the mean value for glucose concentration measured by continuous sensor recordings to be significantly lower in the insulin group compared to the standard care group over the study period, days 1 to 7. However, for this pilot study, the data were not reported in a form permitting quantitative analysis in this review.

Beardsall 2008 reported that the mean sensor glucose value during days 2 to 7 was significantly lower in the insulin group (6.2 mM/L, 112 mg/dL) compared to the standard care group (6.7 mM/L, 121 mg/dL). Mean difference (MD) -0.50 mmol/L (95% CI -0.87 to -0.13).

3.2.1 Hyperglycemia (> 10 mM/L) for > 10% of first week

Beardsall 2007 reported that insulin infusion resulted in a significant reduction in the percentage of time during days 1 to 7 that sensor glucose values were > 10 mM/L (> 180 mg/dL): median 7.6% of time in the insulin group versus 35.9% of time in the standard care group. These data could not be analyzed quantitatively in this review. The proportion of neonates who developed hyperglycemia > 10 mM/L (> 180 mg/dL) requiring insulin treatment fell from 6/8 in the standard care group to 3/8 in the insulin group.

Beardsall 2008 found that significantly fewer neonates in the insulin group compared to the standard care group had hyperglycemia > 10 mM/L (> 180 mg/dL) for more the 10% of time over the first seven days, 21% versus 33%. RR 0.64 (95% CI 0.46 to 0.90); RD -0.12 (95% CI -0.21 to -0.03); NNT 8. The size of reduction in this measure of hyperglycemia was not as great as had been observed in the pilot study. Sixty-nine of 192 neonates in the standard care group (36%) required insulin for treatment of hyperglycemia in the first week.

All-cause mortality
3.3.1 Death before discharge

Beardsall 2007 found no significant difference between insulin infusion and standard care in death before discharge (1/8 versus 2/8): RR 0.50 (95% CI 0.06 to 4.47).

3.4.1 Death by 28 days

Beardsall 2008 found a significant increase in death by 28 days in the group receiving insulin infusion (23/194) versus the group receiving standard care (11/192). Relative risk (RR) 2.07 (95% CI 1.04 to 4.13); risk difference (RD) 0.06 (95% CI 0.01 to 0.12); number needed to harm (NNH) 17. Death by 28 days was a secondary outcome in Beardsall 2008.

3.5.1 Death before expected date of delivery (EDD)

Beardsall 2008 found no significant difference between the insulin and standard care groups in death before EDD (28/194 versus 18/192). RR 1.54 (95% CI 0.88 to 2.69); RD 0.05 (95% CI -0.01 to 0.11). Death before expected date of delivery was the primary outcome in Beardsall 2008.

Neurodevelopmental impairment

Neither Beardsall 2007 nor Beardsall 2008 reported this outcome.

Growth

Beardsall 2007 reported trends toward increased gains in weight and leg length in the insulin group over the study period of the first seven days. These data were not reported in a form permitting quantitative analysis in this review.

3.6.1 Weight gain (g), birth - 28 days

Beardsall 2008 found no significant difference between the insulin and standard care groups in cumulative weight gain from birth to 28 days. This averaged 302 g in the insulin group and 284 g in the standard care group. Mean difference (MD) 18 g (95% CI -11.8 to 47.8)

3.7.1 Length gain (cm), birth to 28 days

Beardsall 2008 found no significant difference in length gain from birth to 28 days between the groups receiving insulin (mean 3.1 cm) versus standard care (mean 3.2 cm). MD -0.10 cm (95% CI -0.53 to 0.33)

3.8.1 Head circumference gain (cm), birth to 28 days

Beardsall 2008 found no significant difference in head circumference gain from birth to 28 days between the groups receiving insulin (mean 2.0 cm) versus standard care (mean 1.9 cm). MD 0.10 cm (95% CI -0.14 to 0.34)

Secondary outcomes
Caloric intake

In Beardsall 2007, mean caloric intake over days 1 to 7 was similar in the insulin and standard care groups. In Beardsall 2008, both caloric and non-protein energy intakes over days 1 to 7 were higher in the insulin as compared to the standard care group. However, the statistical significance of these differences was not reported, nor were the data provided in a form permitting quantitative analysis in this review.

3.9.1 Carbohydrate infused (kcal/kg/day)

In Beardsall 2007, the amount of dextrose infused was higher, but not significantly so, in the insulin group compared to the standard care group: median 9.27 versus 7.64 mg/kg/min which would correspond to 45.4 versus 37.4 kcal/kg/day. In Beardsall 2008, significantly more intravenous carbohydrate was infused over days 1 to 7 in the insulin group (mean 51 kcal/kg/day) than in the standard care group (mean 43 kcal/kg/day). MD 8.0 kcal/kg/day (95% CI 5.7 to 10.3).

Hypoglycemia (< 2.6 mM/L)

3.10.1 Hypoglycemia, sensor recording

In Beardsall 2007, the proportion of sensor readings in the hypoglycemic range was very low in both the insulin and standard care groups. The medians for occurrence of readings < 2.6 mM/L (< 47 mg/dL) were 0.2% in the insulin group, 0.4% in the standard care group, difference not significant. In Beardsall 2008, the percent of time in the hypoglycemic range was slightly higher, although these data were not reported in a form permitting quantitative analysis in this review. At least one occurrence of hypoglycemia was documented by sensor recording in an important number of neonates, and the risk was significantly increased in the insulin group (29%) compared with the control group (17%). RR 1.74 (95% CI 1.19 to 2.55); RD 0.13 (95% CI 0.04 to 0.21); NNH 8. No episodes of hypoglycemia were reported to be associated with clinical signs.

3.11.1 Hypoglycemia, report by clinician

Based on blood glucose measurements reported by clinicians, at least one episode of hypoglycemia occurred in 8.8% of neonates in the insulin group and 1.6% in the standard care group. RR 5.61 (95% CI 1.67 to 18.8); RD 0.07 (95% CI 0.03 to 0.12); NNH 14.

Nitrogen accretion

Neither Beardsall 2007 nor Beardsall 2008 reported this outcome.

Intraventricular hemorrhage
Intracranial disease, grades 1-4

Beardsall 2007 did not report this outcome. Beardsall 2008 assessed "intracranial disease" on cranial ultrasound examination, and reported the proportions of neonates who had any grade of abnormality. For this outcome, as for retinopathy of prematurity and chronic lung disease (see below), the results are presented as the proportion with the outcome among all randomized, and also the proportion with the outcome among those who survived to the time of ascertainment and were actually examined.

3.12.1 Intracranial disease, grades 1-4, among randomized

In Beardsall 2008, there was no significant difference between the insulin and standard care groups: RR 1.09 (95% CI 0.81 to 1.47)

3.13.1 Intracranial disease, grades 1-4, among examined

In Beardsall 2008, there was no significant difference between the insulin (36%) and standard care (32%) groups. RR 1.13 (95% CI 0.85 to 1.51); RD 0.04 (95% CI -0.06 to 0.14). The outcome intracranial disease stage 1-4 was not pre-specified in our protocol.

Retinopathy of prematurity, stage 3 or greater

Beardsall 2007 did not report this outcome.

3.14.1 Retinopathy of prematurity, stage 3 or greater, among randomized

In Beardsall 2008, there was no significant difference between the insulin and standard care groups: RR 1.06 (95% CI 0.54 to 2.07).

3.15.1 Retinopathy of prematurity, stage 3 or greater, among examined

Beardsall 2008 reported the proportions of neonates examined who had retinopathy of prematurity stage 3 or greater: 9.7% (insulin) versus 8.7% (standard care). The difference was not significant. RR 1.12 (95% CI 0.57 to 2.19); RD 0.01 (95% CI -0.05 to 0.07)

Sepsis in the first 2 weeks

Beardsall 2007 did not report this outcome. Beardsall 2008 reported sepsis in the first two weeks of life - both culture positive infection and culture negative infection in neonates with clinical signs which warranted > 48 hours of antibiotics.

3.16.1 Culture positive systemic infection

Beardsall 2008 reported culture positive infection in 21% in the insulin group and 23% in the standard care group. The difference was not significant. RR 0.92 (95% CI 0.63 to 1.34); RD -0.02 (95% CI -0.10 to 0.06).

3.17.1 Presumed infection with negative cultures

Beardsall 2008 reported culture negative infection in 27% in the insulin group and 29% in the standard care group. The difference was not significant. RR 0.95 (95% CI 0.69 to 1.31); RD -0.01 (95% CI -0.10 to 0.08).

3.18.1 Necrotizing enterocolitis, first 28 days

Beardsall 2007 did not report this outcome. Beardsall 2008 reported necrotizing enterocolitis (NEC) defined as radiologic evidence of NEC assessed by a consultant radiologist blinded to treatment group. NEC was diagnosed in 11.9% of the insulin group and 11.5% of the standard care group. The difference was not significant. RR 1.03 (95% CI 0.60 to 1.79); RD 0.00 (95% CI -0.06 to 0.07)

Chronic lung disease

Beardsall 2007 did not report this outcome.

3.19.1 Chronic lung disease, among randomized

Beardsall 2008 reported chronic lung disease defined as respiratory support or oxygen dependency at 36 weeks postmenstrual gestational age. Among all randomized, there was no significant difference in chronic lung disease between the insulin and standard care groups: RR 1.05 (95% CI 0.76 to 1.44)

3.20.1 Chronic lung disease, among examined

In Beardsall 2008 chronic lung disease among those examined was diagnosed in 33% and 30% of the insulin and standard care groups respectively. The difference was not significant. RR 1.11 (95% CI 0.81 to 1.52); RD 0.03 (95% CI -0.07 to 0.13)

Length of hospital stay

Beardsall 2007 did not report this outcome.

3.21 Number of days of neonatal intensive care

Beardsall 2008 reported mean number of days of neonatal intensive care: 16.9 (insulin) and 19.2 (standard care). The difference was not significant. MD -2.3 days (95% CI -5.9 to 1.3).

Other outcomes

Beardsall 2007 reported effects on several other outcomes. These included serum insulin and pro-insulin concentrations, IGF-1 bioactivity (KIRA ug/L). IGF-I concentration (ng/mL), and IGFBP-1 concentration (ng/mL). Significant effects were reported only for IGF-1 bioactivity, which increased 2.4 fold in the insulin group; and IGFBP-1 concentration, which was reduced in the insulin group. Of these outcomes, only concentration of IGF-1 was pre-specified in the protocol for this review.

Insulin infusion versus standard care: Subgroup analyses

Subgroup analyses of Beardsall 2008 according to birth weight strata < 1000 g, 1000 to 1499 g, are included in data tables 3.1 through 3.20. Statistically significant treatment effects within subgroup were found only for the following outcomes.

3.1.2 Mean glucose concentration, first week (mM/L)

Among neonates in the < 1000 g stratum, mean glucose was significantly lower in the insulin infusion group (6.8 mM/L, 122 mg/dL) compared to the standard care group (7.6 mM/L, 137 mg/dL): MD -0.80 mM/L (95% CI -1.37 to -0.23). Among neonates in the 1000 to 1499 g stratum there was no significant difference (3.1.3). Substantial heterogeneity of treatment effect across the two subgroups was noted: I squared was 67% (data not shown).

3.2.2 Hyperglycemia (> 10 mM/L) for > 10% of first week

Among neonates in the < 1000 g stratum, hyperglycemia was significantly reduced: RR 0.69 (95% CI 0.50 to 0.95); RD -0.17 (95% CI -0.31 to -0.03). Among neonates in the 1000 to 1499 g stratum, the reduction did not reach statistical significance (3.2.3).

3.9.2, 3.9.3 Infused carbohydrate, kcal/kg/day

Among neonates in the 1000 to 1499 g stratum, significantly more carbohydrate was infused in the first week in the insulin group (mean 59 kcal/kg/day) than in the standard care group (mean 46 kcal/kg/day): MD 13.0 kcal/kg/day (95% CI 9.6 to 16.5). Among neonates in the < 1000 g stratum the size of increase was much smaller (insulin group, mean 50 kcal/kg/day; standard care group, mean 47 kcal/kg/day) but still statistically significant: MD 3.0 kcal/kg/day (95% CI 0.1 to 5.9). Substantial heterogeneity for this effect between the two subgroups was noted: I squared 95% (data not shown).

3.10.3 Hypoglycemia < 2.6 mM/L (sensor recording)

Hypoglycemia ascertained by sensor recording was significantly increased within the 1000 to 1499 g subgroup: RR 2.97 (95% CI 1.59 to 5.54); RD 0.23 (95% CI 0.11 to 0.34); NNH 4. There was no significant increase in hypoglycemia (sensor recording) within the < 1000 g subgroup (3.10.2). There was substantial heterogeneity for this treatment effect across the two birth-weight subgroups: I squared 83% (data not shown). In interpreting this result, Beardsall et al noted that the 1000 to 1499 g subgroup had a lower risk of hypoglycemia and their clinical care was less likely to have been tightly monitored for hypoglycemia.

In summary, subgroup analyses of Beardsall 2008 according to birth weight strata, like the overall analyses, did not show that reductions in mean glucose levels and in hyperglycemia resulting from insulin infusion were associated with improvements in clinical outcomes. In the < 1000 g stratum, although not in the 1000 to 1499 g stratum, insulin infusion resulted in statistically significant reductions in mean glucose levels and risk of hyperglycemia. But within the < 1000 g stratum, as within the 1000 to 1499 g stratum, there was no evidence of benefit as regards any of the clinical outcomes examined, including death, intracranial disease, retinopathy of prematurity, sepsis, necrotizing enterocolitis or chronic lung disease.

For Beardsall 2008, our planned subgroup analyses according to morbidity score or level of parenteral amino acid intake could not be done, for lack of available data.

Insulin infusion versus restriction of rate of advancement of parenteral glucose intake

We found no trials which compared these alternatives.

Discussion

Only four trials were eligible for inclusion in this review of interventions for the prevention of hyperglycemia in VLBW neonates. Two trials evaluated strategies which included a restriction of the glucose infusion rate during the first week, although neither trial was designed primarily to test the effects of glucose restriction. One of these trials (Gilbertson 1991) reduced the glucose infusion rate by substituting parenteral lipid for a portion of the parenteral glucose intake. The other (Pappoe 2009) compared slow versus rapid advancement over the first week of all components of parenteral nutrition, including glucose. Each of these trials was of small size. Two trials compared insulin infusion with no restriction of glucose infusion rate (standard care). One of these trials (Beardsall 2007) was a pilot study in preparation for the multicentre trial, NIRTURE (Beardsall 2008).

The two small trials which compared lower versus higher glucose infusion rates (Gilbertson 1991; Pappoe 2009) provided insufficient evidence to meet the major objectives of this review - to determine the effect of interventions to prevent hyperglycemia on death and major morbidities. These trials provided some evidence that a lower glucose infusion rate reduced mean blood glucose concentrations and reduced the risk of hyperglycemia. However, they were of such small size that they had insufficient power to test for significant effects on death or major morbidities.

On the other hand, the multicentre trial of insulin infusion (NIRTURE, Beardsall 2008) was designed to have sufficient power to show effects of prevention of hyperglycemia on clinical outcomes. However, the significant clinical effects which were demonstrated were in the adverse direction. Although insulin infusion reduced mean glucose concentrations and reduced hyperglycemia, it resulted in an increase in the risk of death before 28 days and an increase in the proportion of neonates with a hypoglycemic episode. It should be noted that death before 28 days was a secondary outcome in this trial. There was no statistically significant effect on the primary outcome, death before expected date of delivery, although there was a trend towards an increased risk in the insulin group. All-cause mortality was reported in this trial; no data were presented concerning cause of death or the mechanism(s) by which insulin infusion would result in an increase in mortality.

Each of the four trials included in this review provided rather low levels of early postnatal parenteral amino acid intake. None reported the effect of the intervention on nitrogen accretion. We were unable to perform planned subgroup analyses according to higher or lower levels of parenteral amino acid intake because of a lack of such data from the trials. Whether higher levels of parenteral amino acid intake would increase endogenous insulin response, reduce blood glucose concentrations and thereby modify the effects of the interventions reviewed here remains to be determined in future randomized trials.

It is unlikely that we overlooked a relevant trial in this review. We performed extensive searches of several databases for reports in any language of published studies and studies reported only as abstracts. The primary investigators of Pappoe 2009 and Beardsall 2008 were contacted and responded with additional information concerning the outcome data for his/her trial.

Having noted published recommendations concerning parenteral intakes of glucose and other nutrients for VLBW neonates in the early postnatal period (eg AAP 1985; Ziegler 2002; Ehrenkranz 2007), we were surprised at the paucity of randomized trials in this population that examined the clinical effects of lower versus higher glucose infusion rates in the neonatal period. Recommendations supporting aggressive early parenteral nutrition for VLBW neonates often base their rationale largely on the promotion of brain development. Yet, we found no randomized trials of lower versus higher parenteral glucose intakes in VLBW neonates in the early postnatal period which actually examined the effect on neurodevelopment. The findings of the two trials reviewed here (Gilbertson 1991; Pappoe 2009) suggest that aggressively increasing the provision of parenteral glucose during the first postnatal week may cause both positive and negative short-term effects. Much larger trials are needed in VLBW neonates who are at risk of neonatal hyperglycemia in order to determine whether lower or higher rates of parenteral glucose administration are of benefit or harm with respect to death, major morbidities and neurodevelopment of survivors.

As regards insulin infusion as a strategy to prevent hyperglycemia and reduce adverse clinical outcomes, the current state of the evidence is more solid, given the moderately large trial of Beardsall 2008. Assessment of the neurodevelopment status at two years of survivors in this trial is in progress, using a validated parent-report measure, PARCA (Johnson 2004, personal communication from Dr Beardsall). The major adverse effect associated with insulin infusions that was observed in Beardsall 2008, an increased risk of death before 28 days, argues strongly against the routine use of insulin infusions as tested in this trial. Still, caution is needed before rejecting outright the possible benefits of insulin to prevent hyperglycemia and improve clinical outcomes in VLBW neonates receiving intensive care and at risk for hyperglycemia. Early stopping of Beardsall 2008 triggered by an interim look at the accumulated results might have detected a random high as regards the incidence of adverse effects. Despite the relatively large size of this trial, it may actually have been underpowered as regards its ability to be able to show any real benefits from the prevention of hyperglycemia. The planned sample size of 500 was not attained due to early stopping. The observed risk in the control group for the primary outcome, death before the expected date of delivery (9.4%), was less than half that (20%) which was hypothesized at the trial design phase. The potency of the intervention was limited. By one measure, the proportion of neonates having hyperglycemia for more than 10% of time during the first week, insulin infusion neither abolished that risk nor markedly reduced it. The proportion of neonates with hyperglycemia for >10% of time over the first seven days was reduced only moderately, from 33% to 21% (RD -0.13, number needed to treat, 8). Thus, by this measure, only one of eight neonates allocated to the insulin group actually stood to benefit. However, all in that group were exposed to the possible hazards of insulin infusion.

If further trials of insulin infusion to achieve tight glucose control in VLBW neonates are carried out, they should seek to enrol neonates who are at very high risk for hyperglycemia and death. Real time continuous monitoring of glucose concentrations might be used if such systems are validated for clinical use in VLBW neonates. In order to maximize possible benefits and minimize risks, further refinement and testing of algorithms to guide insulin infusion in this population are required.

However, it should be noted that in the field of adult intensive care, the accumulating evidence does not support the use of insulin infusions and tight glucose control as a strategy to improve clinical outcomes. The most recent meta-analysis (Griesdale 2009) of 26 randomized trials involving 13, 567 adult ICU patients found that intensive glucose control using insulin infusions significantly increased the risk of hypoglycemia and conferred no overall mortality benefit. In fact, the largest trial (NICE-SUGAR 2009) which randomized over 6000 patients found that intensive glucose control slightly but significantly increased mortality (risk ratio 1.10, 95% CI 1.01, 1.20).

We did not find any trials concerning the prevention of hyperglycemia in VLBW neonates that compared insulin infusion with restriction of glucose intake. Since neither insulin infusion nor glucose restriction has been shown to be superior to no glucose restriction (standard care) in improving clinical outcomes, there is no rationale at present to support a head-to-head comparison.

Randomized trials of interventions that can prevent neonatal hyperglycemia in VLBW neonates have so far not shown that those interventions reduce death or major morbidities. It remains uncertain whether the prevention of hyperglycemia reduces adverse clinical outcomes which were associated with hyperglycemia in observational studies in this population. The evidence to date warns that routine insulin infusions in VLBW infants as a strategy to prevent hyperglycemia can be harmful.

Authors' conclusions

Implications for practice

Glucose infusion rate: There is insufficient evidence from randomized trials in VLBW neonates to determine whether lower or higher glucose infusion rates in the early postnatal period reduce mortality, morbidity, or adverse neurodevelopment.

Insulin infusion: One moderately large randomized trial, NIRTURE (Beardsall 2008) has been reported. It found that insulin infusion reduced hyperglycemia but increased the risk of death before 28 days and increased the risk of hypoglycemia. There were no significant effects on major morbidities; effects on neurodevelopment have not yet been reported. The evidence from this trial does not support the routine use of insulin infusions to prevent the development of hyperglycemia in VLBW neonates.

Implications for research

Glucose infusion rate: Large randomized trials which compare lower versus higher glucose infusion rates in the early neonatal period in VLBW neonates are needed. These should have sufficient power to determine effects on death, major morbidities and neurodevelopmental impairment at greater than/or equal to 18 months adjusted age.

Insulin infusion: There may be a role for further randomized trials of insulin infusion in ELBW neonates who are at very high risk for hyperglycemia and neonatal death. Accurate and tight control of glucose levels within the target range might be facilitated by real time monitoring once such systems are validated for clinical use in VLBW neonates. There is a need for further refinement and testing of algorithms for insulin dosing, insulin dose adjustment and glucose support in this population. Future trials should have sufficient power to determine effects on death, major morbidities and neurodevelopmental impairment at greater than/or equal to 18 months adjusted age.

Acknowledgements

We gratefully acknowledge the assistance of Dr Kathryn Beardsall and Dr Timothy Pappoe who provided additional information on study outcomes for their trials. Work on the protocol for this review was begun while Dr Marcela Bottino was a Neonatal Fellow in the Dept. of Pediatrics, McMaster University.

Contributions of authors

All authors contributed to the review.

Declarations of interest

  • None noted.

Differences between protocol and review

  • None noted.

Additional tables

  • None noted.

Potential conflict of interest

  • None noted.

[top]

Characteristics of studies

Characteristics of Included Studies

Beardsall 2007

Methods

Randomized controlled trial, single centre

Participants

17 neonates of birth weight < 1500 g, < 24 hours postnatal age, requiring intensive care. The Clinical Risk Index for Babies (CRIB) score was calculated for each infant as a marker of illness severity on day 1. A Continuous Glucose Monitoring Sensor (CGMS, Medtronic Diabetes, Northbridge CA) was inserted subcutaneously in the thigh of each participant, to allow continuous monitoring of glucose levels throughout the first 7 days of life

Interventions

Insulin infusion (n = 8 analyzed)

A fixed-dose insulin infusion (0.025 U/kg/h) was started within 24 hours of birth, after a blood glucose concentration > 3.5 mM/L (> 63 mg/dL) had been confirmed. A separate IV infusion set containing 20% dextrose was also prepared, to be started only if blood glucose concentration fell to < 4.0 mM/L (< 72 mg/dL), and to be discontinued once the blood glucose level reached > 4.0 mM/L (> 73 mg/dL). If the blood glucose concentration fell to < 2.6 mM/L (< 47 mg/dL), the fixed-dose insulin infusion was stopped and the 20% dextrose infusion rate was increased. Once the blood glucose concentration was stable at > 4.0 mM/L (> 72 mg/dL), the fixed-dose insulin infusion was restarted. If the blood glucose level exceeded 10.0 mM/L (180 mg/dL), the dextrose infusion rate in the parenteral nutrition was reduced if it was > 5 mg/kg/min; then, if needed, a sliding-scale insulin infusion was given in addition to the fixed-dose insulin infusion. Initially, blood glucose concentrations were measured at least every 4 hours; as the infant became more stable, this was reduced to at least 4 times a day.

Standard care (n = 8)

If the blood glucose concentration fell to < 2.6 mM/L (< 47 mg/dL), a 2-3 ml IV bolus of 10% dextrose was administered, and the dextrose infusion rate in the parenteral nutrition was increased. If the blood glucose concentration exceeded 10.0 mM/L (> 180 mg/dL), the dextrose infusion rate in the parenteral nutrition was reduced if > 5 mg/kg/min; then, if needed, a sliding-scale insulin infusion was started, initially at 0.05 U/kg/h, and subsequently titrated to maintain blood glucose at 2.6 - 10 mmol/L. Blood glucose concentrations were measured at least 4 times a day.

In both the insulin and standard care groups, parenteral nutrition including dextrose, amino acids and lipid was started within the first 24 hours after birth. We estimated that by day 7 the mean dextrose infusion rate was greater than 8 mg/kg/min (see Notes, below). Enteral feeds were started when possible within the first few days of life with either expressed own mother's milk or donated breast milk

Outcomes

Duration of time with hyperglycemia (CGMS > 10 mM/L, > 180 mg/dL)

Duration of time with hypoglycemia (CGMS < 2.6 mM/L, < 47 mg/dL)

Duration of time in target range (CGMS 4-8 mM/L, 72-144 mg/dL)

Blood glucose concentrations

Dextrose infusion rate

Total caloric intake

Gains in weight, head circumference, lower leg length

Blood concentrations of proinsulin, insulin, IGF-1 bioactivity, IGF-1 binding protein

Death before discharge

Notes

No targets for dextrose infusion rates were specified in advance. Mean dextrose infusion rate over the first 7 days in the standard care group was reported as 7.64 mg/kg/min. Assuming that the dextrose infusion rate increased over the first week, we estimated that by day 7 the mean dextrose infusion rate was at least 8 mg/kg/min in the standard care group. We classified this as no restriction in the rate of advancement of dextrose infusion.

Risk of bias table
Bias Authors' judgement Support for judgement
Allocation concealment (selection bias) Low risk

Randomization by computer-generated random numbers; group allocation concealed in opaque envelopes

Blinding of participants and personnel (performance bias) Unclear risk

Caretakers not blinded to the intervention. However, continuous glucose monitoring sensor readings could not be viewed in real time, and were downloaded only after completion of the 7-day study period. Thus, that information did not affect clinical management

Blinding of outcome assessment (detection bias) High risk

No blinding of outcome assessment

Incomplete outcome data (attrition bias) Low risk

Almost complete follow-up. One of 9 participants randomized to the insulin arm was excluded from analysis because of significant protocol violations involving the use of insulin and 20% dextrose

Selective reporting (reporting bias) Unclear risk

Not assessed. The protocol for the trial was not published

Beardsall 2008

Methods

Randomized controlled trial using minimization, multi-centre

Participants

389 neonates of birth weight < 1500 g, < 24 hours postnatal age, requiring intensive care, were recruited from 8 centres in UK, Belgium, The Netherlands and Spain. The Clinical Risk Index for Babies (CRIB) score was calculated for each neonate as a marker of illness severity at study entry. A continuous glucose monitoring system (CGMS Gold, Medtronic) was inserted in the lateral thigh of each participant to permit continuous subcutaneous glucose monitoring throughout the first 7 days of life

Interventions

Insulin infusion (n = 195)

A fixed-dose continuous infusion of insulin, at a dose of 0.05 U/kg/h, with support as needed using an intravenous 20% dextrose infusion to maintain euglycemia (target range 4-8 mM/L, 72-144 mg/dL), was started within 24 h of birth and continued through 7 days of age. Blood glucose concentrations were measured hourly after insulin was started, and every 6 hours once glucose concentrations had stabilized. If blood glucose concentrations dropped to < 4 mM/L (< 72 mg/dL), the 20% dextrose infusion was started, beginning at 1 ml/kg/h (3.3 mg/kg/min). If, despite this, blood glucose concentrations continued to drift downwards towards 2.6 mM/L, 47 mg/dL (the threshold for hypoglycemia), the insulin infusion was discontinued. If blood glucose concentrations rose so as to exceed 10 mM/L (180 mg/dL) the rate of glucose infusion in the parenteral nutrition infusate was reduced or additional insulin was infused.

Standard care (n = 194)

Blood glucose concentrations were measured at least 3 times daily, and as clinically indicated. The physician responsible for clinical care reviewed glucose concentrations. If blood glucose was > 10 mM/L (> 180 mg/dL), the physician would determine whether the rate of infusion of dextrose should be decreased, and/or if insulin should be started. Insulin was to be started only after 2 glucose concentrations > 10 mM/L (> 180 mg/dL), using a sliding-scale with an initial insulin infusion rate of 0.05 U/kg/h. If blood glucose dropped below 2.6 mM/L (47 mg/dL), the physician would determine whether the rate of infusion of dextrose should be increased.

In both groups, parenteral nutrition with dextrose/water was begun on day 1, with or without the provision of parenteral amino acids and lipid. There were no specified target amounts for either initial dose or rate of advancement of parenteral nutrition. However, the observed rates of carbohydrate infusion indicate that in the standard care group, neonates received on average a glucose infusion rate of about 5.5 mg/kg/min on day 1, which was increased stepwise over the first few days to reach on average glucose infusion rate of about 9.7 mg/kg/min on days 4-7. In the insulin group, the observed glucose infusion rates were similar to the standard care group on day 1, but somewhat higher by days 4-7. In both groups, the observed infusion rates of amino acids and lipid were relatively low.

Outcomes

Primary outcome:

Mortality before expected date of delivery

Secondary outcomes:

Sepsis in the first 2 weeks of life

Somatic growth at 28 days

Necrotizing enterocolitis

Retinopathy of prematurity (stages 3 through 5)

Intracranial disease

Mortality at 28 days

Number of days of intensive care

Other outcomes reported:

Mean glucose concentrations during 1st week

Neonates with glucose > 10 mM/L (> 180 mg/dL) for > 10% of time during the 1st week

Percent of time in hyperglycemic range (> 10 mM/L, > 180 mg/dL), by day

Neonates with an episode of hypoglycemia (< 2.6 mM/L, < 47 mg/dL) for > 60 minutes during the 1st week

Percent of time in hypoglycemic range (< 2.6 mM/L, < 47 mg/dL) by day

Carbohydrate infused during the 1st week, and by day

Notes

This trial aimed to recruit 500 participants. However, recruitment was suspended after 389 patients had been enrolled, after the data and safety and monitoring committee, on an interim analysis, suggested that analysis of the centralized cranial ultrasound images revealed an excess of ventricular hemorrhage and parenchymal lesions and a trend to more deaths in the early-insulin group. Although these were not statistically significant differences, the recommendation to suspend enrollment was based on futility associated with the primary outcome and concern about potential harm. After this suspension of enrollment, the trial steering committee recommended that the trial be discontinued on the grounds of futility.

Dr Beardsall provided us with the following information and data which we could not derive from the published report: i) Clarification that the subgroup data reported in Figure 4 of Beardsall 2008 were based on analyses of neonates as randomized, not as treated ii) For all participants, the N per arm for mean glucose concentration, and the denominators per arm for the proportions with hyperglycemia > 10% of time during the first week and the proportions with hypoglycemia iii) For each of the birth weight subgroups, the mean, SD and N per arm for mean glucose concentration, and the numerators and denominators per arm for the proportions with hyperglycemia > 10% of time during the first week and the proportions with hypoglycemia iv) For each of the birth weight subgroups, the following additional data: the mean, SD and N for each arm for gains in weight, length and head circumference, infused carbohydrate, and days of intensive care; and denominators for the number examined in each arm for intracranial disease, retinopathy of prematurity and chronic lung disease.

Risk of bias table
Bias Authors' judgement Support for judgement
Allocation concealment (selection bias) Low risk

Group allocation was by minimization in order to reduce variability due to centre, birth weight (< 1000 or 1000 - 1500 g) and gestational age (< 25 or 25 weeks or more), using an internet-based computer program. Documentation of entry to the trial was required before the treatment group was allocated.

Blinding of participants and personnel (performance bias) Unclear risk

Caretakers were not blinded to the intervention (not feasible). However, continuous glucose monitoring sensor readings could not be viewed in real time, and were downloaded only after completion of the 7-day study period; thus, that information did not affect clinical management

Blinding of outcome assessment (detection bias) Low risk

Not relevant for all-cause mortality. Intracranial hemorrhage, necrotizing enterocolitis, retinopathy of prematurity were assessed, blinded to treatment group, by a neonatologist, radiologist or ophthalmologist respectively. Unclear for other outcomes

Incomplete outcome data (attrition bias) Low risk

Follow-up almost complete. 386 of 389 trial participants were included in intention-to-treat analyses; 374 were included in as-treated analyses (not presented in this review)

Selective reporting (reporting bias) Low risk

No evidence of selective reporting. Outcomes that were stated in published trial protocol were all reported.

Gilbertson 1991

Methods

Quasi-randomized controlled trial (alternate assignment), single centre

Participants

29 neonates, birth weight < 1500 g, < 6 hr postnatal age on admission to NICU, with ventilator dependence and estimated need of total parenteral nutrition (TPN) for at least 1 week

Interventions

Restricted rate of advancement of parenteral glucose (n = 16)

From day 1. parenteral non-protein energy was supplied as both glucose and lipid. Glucose was to be infused at a rate of 2.8 mg/kg/min on day 1, with step-wise increases to a rate of 7.9 mg/kg/min by days 4-7. (Actual rates of glucose infusion averaged 2.4 mg/kg/min on day 1, increasing to 6.2 mg/kg/min by days 4-7). Intravenous lipid was administered as Intralipid 20%, to be started on day 1 at 1 g/kg/day and increased step-wise to 3 g/kg/day by days 4-7.

"Unrestricted" rate of advancement of parenteral glucose (n = 13)

From day 1 through day 7, parenteral non-protein energy was supplied only by glucose; no parenteral lipid was administered until the 8th day. Parenteral glucose was administered at a higher dose so as to provide an isocaloric amount of non-protein energy. Glucose was to be infused at a rate of 4.5 mg/kg/min on day 1, with step-wise increases to a rate of 13.2 mg/kg/min by days 4-7. (Actual rates of glucose infusion were 3.7 mg/kg/min on day 1, increasing to 10.3 mg/kg/min by days 4-7)

In each arm of the trial, parenteral amino acid intake (provided as Vamin Infant; KabiVitrum) was to be increased progressively to reach a total of 2.6 g/kg/day by day 4. None of the trial participants on whom outcome data were reported received any enteral fluids during the first week. Hyperglycemia, defined as blood glucose level > 8.0 mM/L (> 144 mg/dL) was managed by reducing the glucose concentration of the infusate. Hypoglycemia (< 2.2 mM/L, < 40 mg/dL) was treated with intravenous dextrose, 0.5 g/kg.

Outcomes

Neonatal death, death to latest follow-up

Measures of intravenous lipid tolerance (blood gas tensions, respiratory morbidity, thrombocytopenia, lipid levels)

Measures of glucose homeostasis (mean glucose concentrations, insulin concentrations, hyperglycemia, hypoglycemia)

Growth (gains in weight, length, head circumference)

Clinical outcomes (bronchopulmonary dysplasia, septicemia, periventricular hemorrhage, necrotizing enterocolitis, patent ductus arteriosus, retinopathy of prematurity)

Notes
Risk of bias table
Bias Authors' judgement Support for judgement
Allocation concealment (selection bias) High risk

Alternate allocation

Blinding of participants and personnel (performance bias) High risk

Caretakers not blinded to the intervention

Blinding of outcome assessment (detection bias) Unclear risk

Uncertain if outcome assessors were blinded

Incomplete outcome data (attrition bias) Unclear risk

Outcome data not reported for 3 trial participants who required TPN for < 1 week

Selective reporting (reporting bias) Unclear risk

Not assessed. The protocol for the trial was not published

Pappoe 2009

Methods

Randomized controlled trial, single centre

Participants

32 neonates, birth weight 600-1000 g, entered into the trial on the day of admission to the NICU. Participants were stratified by birth weight 600-800 g (14 neonates), 801-1000 g (18 neonates). A third stratum in this trial, 1001-1200 g (10 neonates), was not eligible for inclusion in this review (see Notes, below).

Interventions

Restricted rate of advancement of parenteral glucose (n = 15)

Neonates were started on 5% dextrose/water on day 1, which was to be increased to 7.5% and 10% on subsequent days only if blood glucose remained below 150 mg/dl. Total daily fluid intake was to be 100 ml/kg on day 1, to be increased by 20 ml/kg each day to a maximum of 150 ml/kg/day. The resulting target glucose infusion rates depended on the total daily fluid intake, minus the amounts of dextrose-free fluid that were administered for perfusion of the umbilical artery catheter, and for Intralipid administration. The target glucose infusion rates, using 5% dextrose/water, were 2.6-2.8 mg/kg/min on day 1, rising to 4.0-4.3 mg/kg/day by day 4. If the dextrose infusion was able to be advanced to a 7.5% or 10% concentration, the target glucose infusion rates would be correspondingly higher. Observed glucose infusion rates averaged 4.4 mg/kg/min mg/kg/min over days 1-3 and 5.7 mg/kg/min over days 1-7. Glucose intolerance was first treated by decreasing the glucose infusion rate to a minimum of 3 mg/kg/min. If hyperglycemia persisted, an insulin infusion at 0.05 U/kg/h was started. The management of insulin infusion was similar to that in the "unrestricted" group (see below).

Parenteral lipids were to be provided as Intralipid 20% at 1 g/kg/day on day 1 and increased by 0.5 g/kg/day to a maximum of 3.5 g/kg/day on day 6. Parenteral amino acids were to be provided as TrophAmine 10% at a rate of 1 g/kg/day on day 1 and increased by 0.5 g/kg/day to a maximum of 3.5 g/kg/day on day 6.

"Unrestricted" rate of advancement of parenteral glucose (n = 17)

Neonates were started on 10% dextrose/water on day 1. As in the restricted group, the total fluid intake was to be 100 ml/kg/day on day 1, to be increased by 20 ml/kg/day to a maximum of 150 ml/kg/day. As in the restricted group, the resulting target glucose infusion rates depended on the total daily fluid intake, minus the amounts of dextrose-free fluid that were administered for perfusion of the umbilical artery catheter, and for Intralipid administration. The target glucose infusion rates were 4.9-5.3 mg/kg/min on day 1, rising to 7.8-8.3 mg/kg/min by day 4. Observed glucose infusion rates averaged 5.8 mg/kg/min over days 1-3 and 6.5 mg/kg/min over days 1-7. Hyperglycemia (blood glucose > 11.1 mM/L, > 200 mg/dL) was treated with insulin infusion, started at 0.05 U/kg/h, which was discontinued when blood glucose dropped to 5.6 mM/L, 100 mg/dL or below. If hyperglycemia persisted at an insulin infusion rate of 0.1 U/kg/h, the glucose infusion rate was decreased until blood glucose dropped below 11.1 mM/L (200 mg/dL).

Lipids (Intralipid 20%) were to be provided as Intralipid 20% at 2 g/kg/day on day 1, 3 g/kg/day on day 2, and 3.5 g/kg/day on day 3 and thereafter. Amino acids (TrophAmine 10%) were to be provided at 2 g/kg/day on day 1, 3 g/kg/day on day 2, and 3.5 g/kg/day on day 3 and thereafter.

Outcomes

Postnatal weight loss, time to regain birth weight, weight gain

Glucose intolerance treated with insulin infusion

Hypertriglyceridemia > 1.7 mM/L (>150 mg/dL)

Azotemia (BUN > 14.3 mM/L (> 40 mg/dL)

Clinical outcomes: IVH, ROP, PDA, death

Notes

Pappoe 2009 randomized 42 neonates of birth weight 600-1200 g. Only the strata 600-800 g (14 neonates) and 801-1000 g (18 neonates) were eligible for inclusion in this review. The 32 neonates in these two strata were randomized to contrasting policies of slow vs. rapid advancement of glucose infusion rate over the first 4 days, as a component of slow vs. rapid advancement of parenteral nutrition by day 4. We excluded the stratum 1001-1200 g from this review because the targets for initial glucose infusion rate and rate of advancement of glucose infusion rate over the first 4 days did not indicate lower target glucose infusion rates in the slow-advancement group.

We included the 600-1000 g data in this review with the permission of Dr Pappoe. Dr Pappoe provided us with the following additional data relevant to this review which we could not derive from the full report of Pappoe 2009: For all participants 600-1000 g and for the subgroups 600-800 g and 801-1000 g: any hyperglycemia, death before discharge, weight gain variables, glucose infusion rate, non-protein energy intake, hypoglycemia, intraventricular hemorrhage, retinopathy of prematurity, days of intensive care.

Risk of bias table
Bias Authors' judgement Support for judgement
Allocation concealment (selection bias) Low risk

Random allocation (random number generator) with group allocation concealed in opaque envelopes

Blinding of participants and personnel (performance bias) High risk

Caretakers not blinded to the intervention

Blinding of outcome assessment (detection bias) High risk

Outcome assessors not blinded

Incomplete outcome data (attrition bias) Low risk

Almost complete follow-up. One neonate randomized to rapid-advance group was diagnosed with trisomy 21 and excluded from analysis

Selective reporting (reporting bias) Unclear risk

Not assessed. The protocol for the trial was not published

Alwaidh 1996

Reason for exclusion

RCT of late vs. earlier initiation of lipid infusion in VLBW neonates, 14 days vs. 5 days. No randomized comparison of lower vs. higher glucose infusion rates, and comparative timing too late to be eligible for this review

Bellagamba 2010

Reason for exclusion

Random allocation to higher vs lower parenteral amino acid intake in VLBW infants, but no mention of randomly allocated difference in glucose infusion rates, and non-protein energy intakes were similar by study design

Blanco 2008

Reason for exclusion

RCT comparing higher vs. lower parenteral amino acid intake in ELBW neonates. No randomly allocated difference in target glucose infusion rates; observed glucose infusion rates similar in the two arms

Brownlee 1993

Reason for exclusion

RCT of early vs. late initiation of parenteral lipid; no randomized comparison of lower vs. higher glucose infusion rates

Clark 2007

Reason for exclusion

RCT comparing higher vs. lower parenteral amino acid intakes in preterm neonates of 23-29 weeks gestational age. No randomly allocated difference in target glucose infusion rates; observed glucose infusion rates almost identical in the two arms

Collins 1991

Reason for exclusion

RCT of insulin infusion for treatment of hyperglycemia in ELBW neonates; participants were hyperglycemic at entry to trial

da Silva 2009

Reason for exclusion

RCT in VLBW infants, but intervention not eligible for this review

Drenckpohl 2008

Reason for exclusion

RCT of higher vs. lower lipid infusion rates in VLBW neonates; no randomly allocated difference in target glucose infusion rates in the two arms

Ekblad 1987

Reason for exclusion

RCT of 5% vs. 10% glucose/water infusate in 1150-2610 g neonates. Only 7 of the 24 participants were < 1500 g birth weight, and their outcome data were not separately reported. Parenteral glucose infusion rate in each group was 3 mg/kg/min or less, ie both groups received restricted parenteral glucose infusion rate by our definition

Hammerman 1988

Reason for exclusion

RCT of early vs. delayed initiation of parenteral lipid. No randomized comparison of lower vs. higher glucose infusion rate

Ibrahim 2004

Reason for exclusion

RCT of early aggressive amino acids in 501-1250 g neonates; no randomized comparison of lower vs. higher glucose infusion rates

Kashyap 2007

Reason for exclusion

RCT, reported only in a PAS abstract at present, of higher vs. lower parenteral amino acid intakes in neonates with birth weight < 1250 g; no randomized comparison of lower vs. higher parenteral glucose intakes (personal communication from Dr Sudha Kashyap, principal investigator)

Kirsten 1997

Reason for exclusion

RCT comparing four intravenous fat emulsions in VLBW neonates. No randomized comparison of different target glucose infusion rates

Macwan 2003

Reason for exclusion

Quasi-randomized trial, with allocation by day of admission, which compared restricted vs. unrestricted fluid intake in VLBW neonates. This trial was reported only as a PAS abstract. The glucose concentration of the infusate was not prescribed by the study protocol and was left to the judgement of the attending physicians (personal communication from the lead investigator, Dr Macwan). Thus, the allocated difference in fluid intake did not necessarily comprise a similar allocated difference in glucose infusion rate

Meetze 1998

Reason for exclusion

Nested RCT of insulin infusion vs. glucose restriction for treatment of hyperglycemia in ELBW neonates; participants hyperglycemic at entry to the trial

Sosenko 1997

Reason for exclusion

RCT of early vs. late initiation of lipid infusion in ELBW neonates; no randomized comparison of lower vs. higher glucose infusion rates

Tang 2009

Reason for exclusion

RCT comparing 3 levels of parenteral amino acid intake, but no randomly allocated difference in target glucose infusion rates. Babies of 1000-2000 g included, < 1500 g not separately reported

te Braake 2005

Reason for exclusion

RCT of early high-dose parenteral amino acids in VLBW neonates; identical glucose infusion rates prescribed in the two arms

Thureen 2003

Reason for exclusion

RCT of higher vs. lower parenteral amino acid intake in VLBW neonates; no difference in target glucose infusion rates in the two arms

van den Akker 2010

Reason for exclusion

Late follow-up of excluded study, te Braacke 2005

Vlaardingerbroek 2009

Reason for exclusion

Random allocation to varying parenteral amino acid and lipid intakes, but no randomly allocated difference in target glucose intakes

Wilson 1997

Reason for exclusion

RCT of aggressive enteral/parenteral nutrition in VLBW neonates, which included a difference between the two arms in the rate of advancement of parenteral glucose as part of the randomized intervention, but neither arm used restriction of parenteral glucose infusion rate as defined in this review

Characteristics of studies awaiting classification

Hawk 2009

Methods

Randomized controlled trial, single centre

Participants

VLBW infants, 24 hours postnatal age at study entry, abstract contains no mention of whether or not participants were hyperglycemic at entry

Interventions

Experimental: continuous infusion of insulin 1 U/kg/day titrated to target blood glucose range of 4.4 - 6.1 mM/L (80 - 100 mg/dL), N = 19. Control: conventional management with target blood glucose range < 10.0 mM/L (< 180 mg/dL), N = 16.

Outcomes

Blood glucose concentrations, episodes of hypoglycemia, growth, death, severe ROP, severe IVH, CLD, NEC

Notes

Insufficient information in abstract to judge whether eligible as included study for this review. Awaiting full report.

  • None noted.

[top]

References to studies

Included studies

Beardsall 2007

Beardsall K, Ogilvy-Stuart AL, Frystyk J, Chen J-W, Thompson M, Ahluwalia J et al. Early elective insulin therapy can reduce hyperglycemia and increase insulin -like growth factor-1 levels in very low birth weight infants. The Journal of Pediatrics 2007;151:611-7.

Beardsall 2008

Beardsall K, Vanhaesebrouck S, Ogilvie-Stuart AL, Ahluwalia JS, Vanhole C, Palmer C et al. A randomised controlled trial of early insulin therapy in very low birth weight infants, "NIRTURE" (neonatal insulin replacement therapy in Europe). BMC Pediatrics 2007;7:29.

Beardsall K, Vanhaesebrouck S, Ogilvy-Stuart AL, Vanhole C, Palmer CR, Ong K et al. Prevalence and determinants of hyperglycemia in very low birth weight infants: cohort analyses of the NIRTURE study. The Journal of Pediatrics 2010;157:715-9.

* Beardsall K, Vanhaesebrouck S, Ogilvy-Stuart AL, Vanhole C, Palmer CR, van Weissenbruck M et al. Early insulin therapy in very-low-birth-weight infants. New England Journal of Medicine 2008;359:1873-84.

Gilbertson 1991

Gilbertson N, Kovar IZ, Cox DJ, Crowe L, Palmer NT. Introduction of intravenous lipid administration on the first day of life in the very low birth weight neonate. The Journal of Pediatrics 1991;119:615-23.

Pappoe 2009

* Pappoe TA, Wu S-Y, Pyati S. A randomized controlled trial comparing an aggressive and a conventional parenteral nutrition regimen in very low birth weight infants. Journal of Neonatal-Perinatal Medicine 2009;2:149-56 (600-1000 g only; >1000 g excluded from this review).

Pappoe TA, Wu S-Y, Pyati S. Rapid vs slow advancement of parenteral nutrition in preterm infants with birth weight less than 1250 g. In: Pediatric Academic Societies Conference Proceedings. 2006:Abstract 5572.441.

Excluded studies

Alwaidh 1996

Alwaidh MH, Bowden L, Shaw B, Ryan SW. Randomised trial of effect of delayed intravenous lipid administration on chronic lung disease in preterm neonates. Journal of Pediatric Gastroenterology and Nutrition 1996;22:303-6.

Bellagamba 2010

Bellagamba MP, Burattini I, Manna M, D'Ascenzo R, Spagnoli C, Malatesta M, Cogo P, Carnielli VP. High vs standard parenteral aminoacid (AA) intake in preterm infants (PI): a randomized clinical trial. In: Pediatric Academic Societies Conference Proceedings. 2010:Abstract 1665.1.

Blanco 2008

Blanco CL, Falck A, Green BK, Cornell JE, Gong AK. Metabolic response to early and high protein supplementation in a randomized trial evaluating the prevention of hyperkalemia in extremely low birth weight infants. The Journal of Pediatrics 2008;153:535-40.

Brownlee 1993

Brownlee KG, Kelly EJ, Ng PC, Kendall Smith SC, Dear PRF. Early or late parenteral nutrition for the sick preterm infant. Archives of Disease in Childhood 1993;69:281-3.

Clark 2007

Clark RH, Chace DH, Spitzer AR, for the Pediatrix Amino Acid Study Group. Effects of two different doses of amino acid supplementation on growth and blood amino acid levels in premature neonates admitted to the neonatal intensive care unit: a randomized controlled trial. Pediatrics 2007;120:1286-1296.

Collins 1991

Collins JW, Hoppe M, Brown K, Edidin DV, Padbury J, Ogata ES. A controlled trial of insulin infusion and parenteral nutrition in extremely low birth weight infants with glucose intolerance. The Journal of Pediatrics 1991;118:921-7.

da Silva 2009

da Silva FS, Miyaki M, Guimaraes R, da Silva VC, Soliz A. The effect of enteral administration of insulin on feeding tolerance and weight gain in low birth weight preterm infants: a randomized control study. In: Pediatric Academic Societies Conference Proceedings. 2009:Abstract 5503.57.

Drenckpohl 2008

Drenckpohl D, McConnell C, Gaffney S, Niehaus M, Macwan KS. Randomized trial of very low birth weight infants receiving higher rates of infusion of intravenous fat emulsions during the first week of life. Pediatrics 2008;122:743-51.

Ekblad 1987

Ekblad H, Kero P, Takala J. Stable glucose balance in premature infants with fluid restriction and early enteral feeding. Acta Paediatrica Scandinavica 1987;76:438-43.

Hammerman 1988

Hammerman C, Aramburo MJ. Decreased lipid intake reduces mortality in sick premature neonates. The Journal of Pediatrics 1988;113:1083-8.

Ibrahim 2004

Ibrahim HM, Jeroudi MA, Baier RJ, Dhanireddy R, Krouskop RW. Aggressive total early parenteral nutrition in low-birth-weight infants. Journal of Perinatology 2004;24:482-6.

Kashyap 2007

Kashyap S, Abildskov K, Holleran SF, Ramakrishnan R, Towers HM, Sahni R effects of early aggressive nutrition in infants with birth weight < 1250 g: a randomized controlled trial. In: Pediatric Academic Societies Conference Proceedings. 2007:Abstract 5912.2.

Kirsten 1997

Kirsten GF, Smuts CM, Smith J, Pieper C, Kirsten CL, van der Riet M, Tichelaar HY, Faber M, Dhansay MA. Plasma cholesterol and triglyceride profiles and prevalence of essential fatty acid deficiency in very-low-birth-weight infants infused with a 10% or 20% lipid emulsion. South African Medical Journal 1997;87:1229-32.

Macwan 2003

Macwan KS, Hocker JR, Clark SE, Tolentino S, Drenckpohl D, McConnell C, Buss K. Restricted fluid intake in very low birth weight (750-1500 g) neonates. In: Pediatric Academic Societies Conference Proceedings. 2003:Abstract 2774.

Meetze 1998

Meetze W, Bowsher R, Compton J, Moorehead H. Hyperglycemia in extremely-low-birth-weight infants. Biology of the Neonate 1998;74:214-21.

Sosenko 1997

Sosenko IRS, Rodriguez-Pierce M, Bancalari E. Effect of early intravenous lipid administration on the incidence and severity of chronic lung disease in premature infants. The Journal of Pediatrics 1993;123:975-82.

Tang 2009

Tang ZF, Huang Y, Zhang R, Chen C. Intensive early amino acid supplementation is efficacious and safe in the management of preterm infants. Zhonghua Er Ke Za Zhi 2009;47 (3):209-15.

te Braake 2005

te Braake FWJ, van den Akker CHP, Wattimena DJL, Huijmans JGM, van Goudoever JB. Amino acid administration to premature infants directly after birth. The Journal of Pediatrics 2005;147:457-61.

Thureen 2003

Thureen PJ, Melara D, Fennessey PV, Hay WW Jr. Effect of low versus high intravenous amino acid intake on very low birth weight infants in the early neonatal period. Pediatric Research 2003;53:24-32.

van den Akker 2010

van den Akker CH, te Braake FW, Weisglas-Kuperus NW, van Goudoever JB. Two-year follow-up of early postnatal amino acid administration to premature infants. In: Pediatric Academic Societies Conference Proceedings. 2010:Abstract 2745.7.

Vlaardingerbroek 2009

Vlaardingerbroek H, van den Akker CH, van Goudoever JB. Is anabolism possible in very low birth weight infants during the first few days of life? Nutritional intervention for preterm infants-2. In: European Society for Pediatric Research Conference Proceedings. 2009:Abstract 161.

Wilson 1997

Wilson DC, Cairns P, Halliday H, Reid M, McClure G, Dodge JA. Randomised controlled trial of an aggressive nutritional regimen in sick very low birth weight infants. Archives of Disease in Childhood 1997;77:F4-F11.

Studies awaiting classification

Hawk 2009

Hawk BJ, Ternes RL, Wilson BJ. Tiny infant glycemic hypothesis trial: TIGHT glucose control in VLBW infants. In: Pediatric Academic Societies Conference Proceedings. 2009:Abstract 5508.149.

Ongoing studies

  • None noted.

Other references

Additional references

AAP 1985

American Academy of Pediatrics Committee on Nutrition. Nutritional needs of low-birth-weight infants. Pediatrics 1985;75:976-86.

Beardsall 2005

Beardsall K, Ogilvy-Stuart AL, Ahluwalia J, Thompson M, Dunger DB. The continuous glucose monitoring sensor in neonatal intensive care. Archives of Disease in Childhood. Fetal and Neonatal Edition 2005;90:F307-1.

Bell 1978

Bell MJ, Ternberg JL, Feigin RD, Keating JP, Marshall R, Barton L, Brotherton T. Neonatal necrotizing enterocolitis. Therapeutic decisions based upon clinical staging. Annals of Surgery 1978;187:1-7.

Bier 1977

Bier DM, Leake RD, Haymond MW, Arnold KJ, Gruenke LD, Sperling MA, Kipnis DM. Measurement of "true" glucose production rates in infancy and childhood with 6, 6-dideuteroglucose. Diabetes 1977;26:1016-23.

Blanco 2006

Blanco CL, Baillargeon JG, Morrison RL, Gong AK. Hyperglycemia in extremely low birth weight infants in a predominantly Hispanic population and related morbidities. Journal of Perinatology 2006;26:737-41.

Bottino 2009

Bottino M, Cowett RM, Sinclair JC. Interventions for treatment of neonatal hyperglycemia in very low birth weight infants. Cochrane Database of Systematic Reviews 2009, Issue 1. Art. No.: CD007453. DOI: 10.1002/14651858.CD007453.pub2.

Chacko 2010

Chacko SK, Sunehag AL. Gluconeogenesis continues in premature infants receiving total parenteral nutrition. Archives of Disease in Childhood Fetal and Neonatal Edition 2010;45:F413-8.

Chacko 2011

Chacko SK, Ordonez J, Sauer PJJ, Sunehag AL. Gluconeogenesis is not regulated by either glucose or insulin in extremely low birth weight infants receiving total parenteral nutrition. The Journal of Pediatrics 2011;158:891-6.

Cowett 1979

Cowett RM, Oh W, Pollack A, Schwartz R, Stonestreet BS. Glucose disposal of low birth weight infants: steady state hyperglycemia produced by constant intravenous glucose infusion. Pediatrics 1979;63:389-96.

Cowett 1983

Cowett RM, Oh W, Schwartz R. Persistent glucose production during glucose infusion in the neonate. Journal of Clinical Investigation 1983;71:467-75.

Cowett 1997

Cowett AA, Farrag HM, Gelardi NL, Cowett RM. Hyperglycemia in the micropremie: evaluation of the metabolic disequilibrium during the neonatal period. Prenatal and Neonatal Medicine 1997;2:360-5.

Ehrenkranz 2007

Ehrenkranz RA. Early, aggressive nutritional management for very low birth weight infants: what is the evidence? Seminars in Perinatology 2007;31:48-55.

Ertl 2006

Ertl T, Gyarmati J, Gaal V, Szabo I. Relationship between hyperglycemia and retinopathy of prematurity in very low birth weight infants. Biology of the Neonate 2006;89:56-9.

Falcão 1998

Falcão MC, Ramos JLA. Hyperglycemia and glucosuria in preterm infants receiving parenteral glucose: influence of birth weight, gestational age and infusion rate. Jornal de Pediatria 1998;74:389-96.

Floyd 1966

Floyd JC, Fajans SS, Conn JW, Knopf RF, Rull J. Stimulation of insulin secretion by amino acids. Journal of Clinical Investigation 1966;45:1487-1502.

Garg 2003

Garg R, Agthe AG, Donohue PK, Lehmann CU. Hyperglycemia and retinopathy of prematurity in very low birth weight infants. Journal of Perinatology 2003;23:186-94.

Grasso 1968

Grasso S, Messina A, Saporito N, Reitano G. Serum-insulin response to glucose and aminoacids in the premature infant. The Lancet 1968;2:755-57.

Griesdale 2009

Griesdale DE, de Sousa RJ, van Dam RM, Heyland DK, Cook DJ, Malhotra A, Dhaliwal R, Henderson WR, Chittock DR, Finfer S, Talmor D. Intensive insulin therapy and mortality among critically ill patients: a meta-analysis including NICE-SUGAR study data. CMAJ 2009;180:821-7.

Grover 2008

Grover A, Khashu M, Murkherjee A, Kairamkonda V. Iatrogenic malnutrition in neonatal intensive care units: urgent need to modify practice. Journal of Parenteral and Enteral Nutrition 2008;32:140-144.

Haynes 2005

Haynes RB, McKibbon KA, Wilczynski NL, Walter SD, Werre SD, for the Hedges Team. Optimal search strategies for retrieving scientifically strong studies of treatment from Medline: analytical survey. BMJ 2005;330:1179.

Hays 2006

Hays SP, Smith EO, Sunehag AL. Hyperglycemia is a risk factor for early death and morbidity in extremely low birth weight infants. Pediatrics 2006;118:1811-8.

Heimann 2007

Heimann K, Peschgens T, Kwiecien R, Stanzel S, Hoernchen H, Merz U. Are recurrent hyperglycemic episodes and median blood glucose level a prognostic factor for increased morbidity and mortality in premature infants less than/or equal to 1500g? Journal of Perinatal Medicine 2007;35:245-8.

Heird 1992

Heird WC. Parenteral feeding. In: JC Sinclair, MB Bracken, editor(s). Effective Care of the Newborn Infant. Oxford University Press, UK, 1992:Chapter 8;141-60.

Higgins 2011

Higgins JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.0 [updated March 2011]. The Cochrane Collaboration, 2011, Available from www.cochrane-handbook.org.

Jobe 2001

Jobe AH, Bancalari E. Bronchpulmonary dysplasia. American Journal of Respiratory and Critical Care Medicine 2001;163:1723-9.

Johnson 2004

Johnson S, Marlow N, Wolke D et al. Validation of a parent report measure of cognitive development in very preterm infants. Developmental Medicine and Child Neurology 2004;46:389-97.

Kao 2006

Kao LS, Morris BH, Lally KP, Stewart CD, Huseby V, Kennedy KA. Hyperglycemia and morbidity in extremely low birth weight infants. Journal of Perinatology 2006;26:730-6.

Louik 1985

Louik C, Mitchell AA, Epstein MF, Shapiro S. Risk factors for neonatal hyperglycemia associated with 10% dextrose infusion. American Journal of Diseases of Children 1985;139:783-6.

Löfqvist 2006

Löfqvist C, Engström E, Sigurdsson J, Hård A-L, Niklasson A, Ewald U, Holmström, Smith LEH, Hellström A. Postnatal head growth deficit among premature infants parallels retinopathy of prematurity and insulin-like growth factor-I deficit. Pediatrics 2006;117:1930-1938.

Manzoni 2006

Manzoni P, Castagnola E, Mostert M, Sala U, Galletto P, Gomirato G. Hyperglycemia as a possible marker of invasive fungal infection in preterm neonates. Acta Paediatrica 2006;95:486-93.

Mitanchez-Mokhtari 2004

Mitanchez-Mokhtari D, Lahlou N, Kieffer F, Magny J-F, Roger M, Voyer M. Both relative insulin resistance and defective islet beta-cell processing of proinsulin are responsible for transient hyperglycemia in extremely preterm infants. Pediatrics 2004;113:537-41.

NICE-SUGAR 2009

The NICE-SUGAR Study Investigators. Intensive versus conventional glucose control in critically ill patients. New England Journal of Medicine 2009;360:1283-97.

Ogilvy-Stuart 1998

Ogilvy-Stuart AL, Hands SJ, Adcock CJ, Holly JMP, Matthews DR, Mohamed-Ali V, Yudkin JS, Wilkinson AR, Dunger DB. Insulin, insulin-like growth factor I (IGF-I), IGF-binding protein-I, growth hormone, and feeding in the newborn. J Clin Endocrinol Metab 1998;83:3550-57.

Papile 1978

Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1, 500 gm. Journal of Pediatrics 1978;92:529-34.

Rowen 1995

Rowen JL, Atkins JT, Levy ML, Baer SC, Baker CJ. Invasive fungal dermatitis in the less than/or equal to 1000-gram neonate. Pediatrics 1995;95:682-7.

Sunehag 1994

Sunehag A, Gustaffson J, Ewald U. Very immature infants (less than/or equal to 30 Wk) respond to glucose infusion with incomplete suppression of glucose production. Pediatric Research 1994;36:550-5.

Sunehag 1999

Sunehag AL, Haymond MW, Schanler RJ, Reeds PJ, Bier DJ. Gluconeogenesis in very low birth weight infants receiving total parenteral nutrition. Diabetes 1999;48:791-800.

Sunehag 2004

Sunehag AL. Infusion rate of glucose is the major predictor of hyperglycemia in very premature infants receiving total parenteral nutrition. In: Pediatric Academic Societies Conference Proceedings. 2004:Abstract 2170.

Thureen 2003

Thureen PJ, Melara D, Fennessey PV, Hay WW Jr. Effect of low versus high intravenous amino acid intake on very low birth weight infants in the early neonatal period. Pediatric Research 2003;53:24-32.

Van den Berghe 2001

Van den Berghe G, Wouters P, Weekers F, Verwaest C, Bruyninckx F, Schetz M, Vlasselaers D, Ferdinande P, Lauwers P, Bouillon R. Intensive insulin therapy in critically ill patients. The New England Journal of Medicine 2001;345:1359-67.

Ziegler 2002

Ziegler EE, Thureen PJ, Carlson SJ. Aggressive nutrition of the very low birthweight infant. Clinics in Perinatology 2002;29:225-244.

Zlotkin 1981

Zlotkin SH, Bryan MH, Anderson GH. Intravenous nitrogen and energy intakes required to duplicate in utero nitrogen accretion in prematurely born human infants. The Journal of Pediatrics 1981;99:115-20.

Other published versions of this review

Sinclair 2009

Sinclair JC, Bottino M, Cowett RM. Interventions for prevention of neonatal hyperglycemia in very low birth weight infants. Cochrane Database of Systematic Reviews 2009, Issue 3. Art. No.: CD007615. DOI: 10.1002/14651858.CD007615.pub2.

Classification pending references

  • None noted.

[top]

Data and analyses

1 Early vs delayed introduction of parenteral lipid

For graphical representations of the data/results in this table, please use link under "Outcome or Subgroup".

Outcome or Subgroup Studies Participants Statistical Method Effect Estimate
1.1 Mean blood glucose, days 1-7 (mM/L) 1 29 Mean Difference (IV, Fixed, 95% CI) -1.49 [-2.50, -0.48]
1.2 Hyperglycemia > 8.0 mM/L 1 29 Risk Ratio (M-H, Fixed, 95% CI) 0.61 [0.28, 1.31]
1.3 All-cause mortality 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
1.3.1 Death in neonatal period 1 29 Risk Ratio (M-H, Fixed, 95% CI) 0.41 [0.04, 4.00]
1.4 Growth 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only
1.4.1 Weight gain in first week (g) 1 29 Mean Difference (IV, Fixed, 95% CI) 65.60 [44.45, 86.75]
1.4.2 Length gain (cm/wk) 1 29 Mean Difference (IV, Fixed, 95% CI) 0.10 [-0.18, 0.38]
1.4.3 Head circumference gain (cm/wk) 1 29 Mean Difference (IV, Fixed, 95% CI) 0.00 [-0.28, 0.28]
1.5 Days to regain birth weight 1 29 Mean Difference (IV, Fixed, 95% CI) -1.30 [-5.88, 3.28]
1.6 Hypoglycemia < 2.2 mM/L 1 29 Risk Ratio (M-H, Fixed, 95% CI) 1.14 [0.47, 2.75]
1.7 Intraventricular hemorrhage 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
1.7.1 Grade not stated 1 29 Risk Ratio (M-H, Fixed, 95% CI) 0.58 [0.24, 1.40]
1.8 Retinopathy of prematurity 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
1.8.1 Stage not stated 1 29 Risk Ratio (M-H, Fixed, 95% CI) 0.27 [0.01, 6.23]
1.9 Sepsis 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
1.9.1 Positive blood culture before discharge 1 29 Risk Ratio (M-H, Fixed, 95% CI) 0.81 [0.13, 5.01]
1.10 Necrotizing enterocolitis 1 29 Risk Ratio (M-H, Fixed, 95% CI) 0.81 [0.06, 11.77]
1.11 Chronic lung disease 1 29 Risk Ratio (M-H, Fixed, 95% CI) 0.54 [0.11, 2.77]

2 Slow vs rapid rate of advancement of parenteral nutrition

For graphical representations of the data/results in this table, please use link under "Outcome or Subgroup".

Outcome or Subgroup Studies Participants Statistical Method Effect Estimate
2.1 Hyperglycemia > 11.1 mM/L 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
2.1.1 All neonates 1 32 Risk Ratio (M-H, Fixed, 95% CI) 0.57 [0.32, 1.02]
2.1.2 Neonates 600-800g 1 14 Risk Ratio (M-H, Fixed, 95% CI) 0.60 [0.28, 1.31]
2.1.3 Neonates 801-1000g 1 18 Risk Ratio (M-H, Fixed, 95% CI) 0.52 [0.21, 1.29]
2.2 Hyperglycemia with insulin treatment 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
2.2.1 All neonates 1 32 Risk Ratio (M-H, Fixed, 95% CI) 0.21 [0.05, 0.78]
2.2.2 Neonates 600-800g 1 14 Risk Ratio (M-H, Fixed, 95% CI) 0.38 [0.10, 1.41]
2.2.3 Neonates 801-1000g 1 18 Risk Ratio (M-H, Fixed, 95% CI) 0.10 [0.01, 1.52]
2.3 Death before discharge, all-cause 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
2.3.1 All neonates 1 32 Risk Ratio (M-H, Fixed, 95% CI) 1.13 [0.08, 16.59]
2.3.2 Neonates 600-800g 1 14 Risk Ratio (M-H, Fixed, 95% CI) 2.33 [0.11, 48.99]
2.3.3 Neonates 801-1000g 1 18 Risk Ratio (M-H, Fixed, 95% CI) 0.50 [0.02, 10.80]
2.4 Weight gain, birth-7 days, g 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only
2.4.1 All neonates 1 32 Mean Difference (IV, Fixed, 95% CI) -56.97 [-100.13, -13.81]
2.4.2 Neonates 600-800g 1 14 Mean Difference (IV, Fixed, 95% CI) -76.00 [-135.92, -16.08]
2.4.3 Neonates 801-1000g 1 18 Mean Difference (IV, Fixed, 95% CI) -43.50 [-101.90, 14.90]
2.5 Days to regain birth weight 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only
2.5.1 All neonates 1 32 Mean Difference (IV, Fixed, 95% CI) 1.86 [-0.76, 4.48]
2.5.2 Neonates 600-800g 1 14 Mean Difference (IV, Fixed, 95% CI) 2.60 [-1.17, 6.37]
2.5.3 Neonates 801-1000g 1 18 Mean Difference (IV, Fixed, 95% CI) 0.50 [-2.83, 3.83]
2.6 Percent weight loss 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only
2.6.1 All neonates 1 32 Mean Difference (IV, Fixed, 95% CI) 3.48 [-0.49, 7.45]
2.6.2 Neonates 600-800g 1 14 Mean Difference (IV, Fixed, 95% CI) 5.90 [1.58, 10.22]
2.6.3 Neonates 801-1000g 1 18 Mean Difference (IV, Fixed, 95% CI) 1.60 [-3.17, 6.37]
2.7 Failure to reach birth weight by day 7 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
2.7.1 All neonates 1 32 Risk Ratio (M-H, Fixed, 95% CI) 1.46 [0.72, 2.94]
2.7.2 Neonates 600-800g 1 14 Risk Ratio (M-H, Fixed, 95% CI) 3.75 [0.58, 24.28]
2.7.3 Neonates 801-1000g 1 18 Risk Ratio (M-H, Fixed, 95% CI) 1.05 [0.45, 2.42]
2.8 Glucose infusion rate, days 1-7, mg/kg/min 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only
2.8.1 All neonates 1 32 Mean Difference (IV, Fixed, 95% CI) -0.78 [-1.53, -0.03]
2.8.2 Neonates 600-800g 1 14 Mean Difference (IV, Fixed, 95% CI) -1.23 [-2.19, -0.27]
2.8.3 Neonates 801-1000g 1 18 Mean Difference (IV, Fixed, 95% CI) -0.17 [-1.29, 0.95]
2.9 Non-protein energy intake, days 1-7, kcal/kg/day 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only
2.9.1 All neonates 1 32 Mean Difference (IV, Fixed, 95% CI) -11.97 [-16.65, -7.29]
2.9.2 Neonates 600-800g 1 14 Mean Difference (IV, Fixed, 95% CI) -16.23 [-23.04, -9.42]
2.9.3 Neonates 801-1000g 1 18 Mean Difference (IV, Fixed, 95% CI) -7.16 [-13.26, -1.06]
2.10 Hypoglycemia 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
2.10.1 All neonates 1 32 Risk Ratio (M-H, Fixed, 95% CI) 3.40 [0.39, 29.31]
2.10.2 Neonates 600-800g 1 14 Risk Ratio (M-H, Fixed, 95% CI) 2.25 [0.30, 16.63]
2.10.3 Neonates 801-1000g 1 18 Risk Ratio (M-H, Fixed, 95% CI) Not estimable
2.11 Hypertriglyceridemia > 1.7 mM/L 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
2.11.1 All neonates 1 32 Risk Ratio (M-H, Fixed, 95% CI) 0.28 [0.10, 0.82]
2.11.2 Neonates 600-800g 1 14 Risk Ratio (M-H, Fixed, 95% CI) 0.42 [0.18, 0.97]
2.11.3 Neonates 801-1000g 1 18 Risk Ratio (M-H, Fixed, 95% CI) 0.12 [0.01, 1.78]
2.12 Azotemia, BUN > 14.3 mM/L 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
2.12.1 All neonates 1 32 Risk Ratio (M-H, Fixed, 95% CI) 0.38 [0.04, 3.26]
2.12.2 Neonates 600-800g 1 14 Risk Ratio (M-H, Fixed, 95% CI) 0.75 [0.06, 9.72]
2.12.3 Neonates 801-1000g 1 18 Risk Ratio (M-H, Fixed, 95% CI) 0.30 [0.02, 5.46]
2.13 Intraventricular hemorrhage, any grade 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
2.13.1 All neonates 1 32 Risk Ratio (M-H, Fixed, 95% CI) 0.65 [0.24, 1.78]
2.13.2 Neonates 600-800g 1 14 Risk Ratio (M-H, Fixed, 95% CI) 1.00 [0.35, 2.88]
2.13.3 Neonates 801-1000g 1 18 Risk Ratio (M-H, Fixed, 95% CI) 0.17 [0.01, 2.69]
2.14 Intraventricular hemorrhage, > grade 2 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
2.14.1 All neonates 1 32 Risk Ratio (M-H, Fixed, 95% CI) 0.28 [0.04, 2.26]
2.14.2 Neonates 600-800g 1 14 Risk Ratio (M-H, Fixed, 95% CI) 0.38 [0.04, 3.23]
2.14.3 Neonates 801-1000g 1 18 Risk Ratio (M-H, Fixed, 95% CI) 0.30 [0.02, 5.46]
2.15 Retinopathy of prematurity, any stage 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
2.15.1 All neonates 1 32 Risk Ratio (M-H, Fixed, 95% CI) 0.98 [0.75, 1.28]
2.15.2 Neonates 600-800g 1 14 Risk Ratio (M-H, Fixed, 95% CI) 1.00 [0.77, 1.30]
2.15.3 Neonates 801-1000g 1 18 Risk Ratio (M-H, Fixed, 95% CI) 0.87 [0.51, 1.51]
2.16 Retinopathy of prematurity, stage 2 or more with surgery 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
2.16.1 All neonates 1 32 Risk Ratio (M-H, Fixed, 95% CI) 1.70 [0.33, 8.84]
2.16.2 Neonates 600-800g 1 14 Risk Ratio (M-H, Fixed, 95% CI) 1.13 [0.27, 4.76]
2.16.3 Neonates 801-1000g 1 18 Risk Ratio (M-H, Fixed, 95% CI) Not estimable
2.17 Days neonatal intensive care 1 31 Mean Difference (IV, Fixed, 95% CI) 10.69 [-7.46, 28.84]

3 Insulin infusion vs standard care

For graphical representations of the data/results in this table, please use link under "Outcome or Subgroup".

Outcome or Subgroup Studies Participants Statistical Method Effect Estimate
3.1 Mean glucose concentration, first week, mM/L 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only
3.1.1 All neonates 1 377 Mean Difference (IV, Fixed, 95% CI) -0.50 [-0.87, -0.13]
3.1.2 Neonates < 1000g 1 189 Mean Difference (IV, Fixed, 95% CI) -0.80 [-1.37, -0.23]
3.1.3 Neonates 1000-1499g 1 188 Mean Difference (IV, Fixed, 95% CI) -0.20 [-0.56, 0.16]
3.2 Hyperglycemia (> 10 mM/L) for > 10% of first week 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
3.2.1 All neonates 1 377 Risk Ratio (M-H, Fixed, 95% CI) 0.64 [0.46, 0.90]
3.2.2 Neonates < 1000g 1 189 Risk Ratio (M-H, Fixed, 95% CI) 0.69 [0.50, 0.95]
3.2.3 Neonates 1000-1499g 1 188 Risk Ratio (M-H, Fixed, 95% CI) 0.47 [0.19, 1.19]
3.3 Death before discharge 1 16 Risk Ratio (M-H, Fixed, 95% CI) 0.50 [0.06, 4.47]
3.4 Death by 28 days 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
3.4.1 All neonates 1 386 Risk Ratio (M-H, Fixed, 95% CI) 2.07 [1.04, 4.13]
3.4.2 Neonates < 1000g 1 194 Risk Ratio (M-H, Fixed, 95% CI) 1.78 [0.83, 3.83]
3.4.3 Neonates 1000-1499g 1 192 Risk Ratio (M-H, Fixed, 95% CI) 3.43 [0.73, 16.08]
3.5Weight gain (g), birth - 28 days 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
3.5.1 All neonates 1 386 Risk Ratio (M-H, Fixed, 95% CI) 1.54 [0.88, 2.69]
3.5.2 Neonates < 1000g 1 194 Risk Ratio (M-H, Fixed, 95% CI) 1.31 [0.73, 2.36]
3.5.3 Neonates 1000-1499g 1 192 Risk Ratio (M-H, Fixed, 95% CI) 3.43 [0.73, 16.08]
3.6 Weight gain (g), birth - 28 days 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only
3.6.1 All neonates 1 349 Mean Difference (IV, Fixed, 95% CI) 18.00 [-11.84, 47.84]
3.6.2 Neonates < 1000 g 1 166 Mean Difference (IV, Fixed, 95% CI) 23.00 [-10.01, 56.01]
3.6.3 Neonates 1000-1499g 1 183 Mean Difference (IV, Fixed, 95% CI) -2.00 [-47.10, 43.10]
3.7 Length gain (cm), birth 28 days 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only
3.7.1 All neonates 1 304 Mean Difference (IV, Fixed, 95% CI) -0.10 [-0.53, 0.33]
3.7.2 Neonates < 1000g 1 138 Mean Difference (IV, Fixed, 95% CI) 0.46 [-0.07, 1.00]
3.7.3 Neonates 1000-1499g 1 166 Mean Difference (IV, Fixed, 95% CI) -0.42 [-1.06, 0.22]
3.8 Head circumference gain (cm), birth - 28 days 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only
3.8.1 All neonates 1 328 Mean Difference (IV, Fixed, 95% CI) 0.10 [-0.14, 0.34]
3.8.2 Neonates < 1000g 1 153 Mean Difference (IV, Fixed, 95% CI) 0.29 [-0.08, 0.67]
3.8.3 Neonates 1000-1499g 1 175 Mean Difference (IV, Fixed, 95% CI) 0.04 [-0.25, 0.33]
3.9 Infused carbohydrate (kcal/kg/day) 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only
3.9.1 All neonates 1 386 Mean Difference (IV, Fixed, 95% CI) 8.00 [5.69, 10.31]
3.9.2 Neonates < 1000g 1 192 Mean Difference (IV, Fixed, 95% CI) 3.00 [0.10, 5.90]
3.9.3 Neonates 1000-1499g 1 191 Mean Difference (IV, Fixed, 95% CI) 13.00 [9.55, 16.45]
3.10 Hypoglycemia, sensor recording 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
3.10.1 All neonates 1 377 Risk Ratio (M-H, Fixed, 95% CI) 1.74 [1.19, 2.55]
3.10.2 Neonates < 1000g 1 189 Risk Ratio (M-H, Fixed, 95% CI) 1.13 [0.68, 1.86]
3.10.3 Neonates 1000-1499g 1 188 Risk Ratio (M-H, Fixed, 95% CI) 2.97 [1.59, 5.54]
3.11 Hypoglycemia, report by clinician 1 386 Risk Ratio (M-H, Fixed, 95% CI) 5.61 [1.67, 18.83]
3.12 Intracranial disease, grades 1-4, among randomized 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
3.12.1 All neonates 1 386 Risk Ratio (M-H, Fixed, 95% CI) 1.09 [0.81, 1.47]
3.12.2 Neonates < 1000g 1 194 Risk Ratio (M-H, Fixed, 95% CI) 0.97 [0.66, 1.43]
3.12.3 Neonates 1000-1499g 1 192 Risk Ratio (M-H, Fixed, 95% CI) 1.27 [0.81, 1.99]
3.13 Intracranial disease, grades 1-4, among examined 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
3.13.1 All neonates 1 357 Risk Ratio (M-H, Fixed, 95% CI) 1.13 [0.85, 1.51]
3.13.2 Neonates < 1000g 1 180 Risk Ratio (M-H, Fixed, 95% CI) 0.99 [0.68, 1.45]
3.13.3 Neonates 10000-1499g 1 177 Risk Ratio (M-H, Fixed, 95% CI) 1.34 [0.86, 2.08]
3.14 Retinopathy of prematurity, stage 3 or greater, among randomized 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
3.14.1 All neonates 1 386 Risk Ratio (M-H, Fixed, 95% CI) 1.06 [0.54, 2.07]
3.14.2 Neonates < 1000g 1 194 Risk Ratio (M-H, Fixed, 95% CI) 1.00 [0.47, 2.11]
3.14.3 Neonates 1000-1499g 1 192 Risk Ratio (M-H, Fixed, 95% CI) 1.31 [0.30, 5.68]
3.15 Retinopathy of prematurity, stage 3 or greater, among examined 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
3.15.1 All neonates 1 338 Risk Ratio (M-H, Fixed, 95% CI) 1.12 [0.57, 2.19]
3.15.2 Neonates < 1000g 1 158 Risk Ratio (M-H, Fixed, 95% CI) 1.08 [0.52, 2.25]
3.15.3 Neonates 1000-1499g 1 180 Risk Ratio (M-H, Fixed, 95% CI) 1.36 [0.31, 5.92]
3.16 Culture positive systemic infection 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
3.16.1 All neonates 1 386 Risk Ratio (M-H, Fixed, 95% CI) 0.92 [0.63, 1.34]
3.16.2 Neonates < 1000g 1 194 Risk Ratio (M-H, Fixed, 95% CI) 0.81 [0.52, 1.26]
3.16.3 Neonates 1000-1499g 1 192 Risk Ratio (M-H, Fixed, 95% CI) 1.21 [0.61, 2.37]
3.17 Presumed infection with negative cultures 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
3.17.1 All neonates 1 386 Risk Ratio (M-H, Fixed, 95% CI) 0.95 [0.69, 1.31]
3.17.2 Neonates < 1000g 1 194 Risk Ratio (M-H, Fixed, 95% CI) 1.09 [0.74, 1.61]
3.17.3 Neonates 1000-1499g 1 192 Risk Ratio (M-H, Fixed, 95% CI) 0.77 [0.44, 1.33]
3.18 Necrotizing enterocolitis, first 28 days 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
3.18.1 All neonates 1 386 Risk Ratio (M-H, Fixed, 95% CI) 1.03 [0.60, 1.79]
3.18.2 Neonates < 1000g 1 194 Risk Ratio (M-H, Fixed, 95% CI) 0.71 [0.36, 1.40]
3.18.3 Neonates 1000-1499g 1 192 Risk Ratio (M-H, Fixed, 95% CI) 2.15 [0.78, 5.97]
3.19 Chronic lung disease, among randomized 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
3.19.1 All neonates 1 386 Risk Ratio (M-H, Fixed, 95% CI) 1.05 [0.76, 1.44]
3.19.2 Neonates < 1000g 1 194 Risk Ratio (M-H, Fixed, 95% CI) 1.03 [0.72, 1.46]
3.19.3 Neonates 1000-1499g 1 192 Risk Ratio (M-H, Fixed, 95% CI) 1.11 [0.59, 2.09]
3.20 Chronic lung disease, among examined 1 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
3.20.1 All neonates 1 340 Risk Ratio (M-H, Fixed, 95% CI) 1.11 [0.81, 1.52]
3.20.2 Neonates < 1000g 1 156 Risk Ratio (M-H, Fixed, 95% CI) 1.11 [0.80, 1.54]
3.20.3 Neonates 1000-1499g 1 184 Risk Ratio (M-H, Fixed, 95% CI) 1.16 [0.62, 2.18]
3.21 Number of days neonatal intensive care 1 Mean Difference (IV, Fixed, 95% CI) Subtotals only
3.21.1 All neonates 1 324 Mean Difference (IV, Fixed, 95% CI) -2.30 [-5.88, 1.28]
3.21.2 Neonates < 1000g 1 147 Mean Difference (IV, Fixed, 95% CI) -2.00 [-7.01, 3.01]
3.21.3 Neonates 1000-1499g 1 177 Mean Difference (IV, Fixed, 95% CI) -3.00 [-6.26, 0.26]

[top]

Figures

  • None noted.

[top]

Sources of support

Internal sources

  • No sources of support provided.

External sources

  • No sources of support provided.

This review is published as a Cochrane review in The Cochrane Library, Issue 10, 2011 (see http://www.thecochranelibrary.com External Web Site Policy for information). Cochrane reviews are regularly updated as new evidence emerges and in response to feedback. The Cochrane Library should be consulted for the most recent version of the review.