Objective To determine the contribution of total gluconeogenesis to glucose production in preterm infants receiving total parenteral nutrition (TPN) providing glucose exceeding normal infant glucose turnover rates.
Study Design Eight infants (0.955±0.066 kg, 26.5±0.5 weeks, 4±1 days) were studied while receiving routine TPN. The glucose appearance rate (the sum of rates of glucose infusion and residual glucose production) and gluconeogenesis were measured by stable isotope–gas chromatography–mass spectrometry techniques using deuterated water and applying the Chacko and Landau methods.
Results Blood glucose ranged from 5.2 to 14.3 mmol/l (94–257 mg/dl) and the glucose infusion rate from 7.4 to 11.4 mg/kg per min, thus exceeding the normal glucose production rates (GPR) of newborn infants in most of the babies. The glucose appearance rate was 12.4±0.6 and GPR 2.1±0.3 mg/kg per min. Gluconeogenesis as a fraction of the glucose appearance rate was 11.2±1.1% (Chacko) and 10.5±1.2% (Landau) (NS) and the rate of gluconeogenesis was 1.35±0.15 mg/kg per min (Chacko) and 1.29±0.14 mg/kg per min (Landau) (NS). Gluconeogenesis accounted for 73±11% and 68±10 (NS) of the GPR for the two methods, respectively. Gluconeogenesis and glycogenolysis were not affected by the total glucose infusion rate, glucose concentration, gestational age or birth weight. Glucose concentration correlated with the total glucose infusion rate and gestational age (combined R2=0.79, p=0.02).
Conclusions Gluconeogenesis is sustained in preterm infants receiving routine TPN providing glucose at rates exceeding normal infant glucose turnover rates and accounts for the major part of residual glucose production. Gluconeogenesis is not affected by the glucose infusion rate or blood glucose concentration.
Statistics from Altmetric.com
During their first week of life, very premature infants are at high risk of disturbed glucose homeostasis, which might result in increased morbidity and mortality.1,–,5 As very premature newborns have a low tolerance to enteral feeding, they are dependent on total parenteral nutrition. However, parenteral glucose frequently results in hyperglycaemia when the glucose infusion rate exceeds the glucose production rate (GPR) of healthy newborn infants (6–8 mg/kg per min).6,–,12 Therefore, providing an optimal energy intake while maintaining normal glucose homeostasis is a challenge requiring detailed knowledge about the physiology of the glucose metabolism of these infants.
We have previously demonstrated that very premature infants receiving TPN providing glucose at rates corresponding to only half the GPR of healthy newborn infants maintained normoglycaemia by glucose produced primarily by the gluconeogenic pathway.13 Under these conditions, the components of the parenteral lipid emulsion (particularly glycerol) were more important substrates for gluconeogenesis than the amino acids.14 15
What is already known on this topic
▶ Hyperglycaemia is associated with morbidity and mortality in critically ill neonates, children and adults.
▶ Hyperglycaemia is common in very premature infants receiving routine TPN.
▶ Glucose production is not completely suppressed even when very premature infants receive intravenous glucose that exceeds their requirements.
What this study adds
▶ Gluconeogenesis is sustained even when premature infants receive TPN at rates exceeding normal infant glucose turnover rates.
▶ Gluconeogenesis accounts for the major part of residual glucose production.
▶ Gluconeogenesis is not related to blood concentrations of glucose, infusion rates of glucose, lipid or amino acids, birth weight or gestational age.
Studies in adults have demonstrated that glucose production is turned off when glucose is infused at a rate corresponding to their normal glucose production.16 17 In contrast, during the same conditions, very premature infants do not completely suppress glucose production.10 16 18 In the referenced studies, the infants did not receive parenteral lipids and amino acids and the contribution from gluconeogenesis and glycogenolysis was not measured.
The question remains whether very premature infants suppress glucose production and its components during routine nutritional supply, which most often includes glucose at rates exceeding that of their normal glucose production plus lipids and amino acids. To our knowledge this issue has not previously been addressed in very premature infants.
The primary aim of the present study was to determine whether gluconeogenesis is sustained in very premature infants receiving standard nutritional care, and whether it correlates with the glucose infusion rate and/or blood glucose concentration. To ensure the accuracy of the measurements of gluconeogenesis under conditions of high exogenous glucose infusion rates, two different approaches to measure gluconeogenesis were applied.19 20 We hypothesised that in very premature infants receiving routine TPN, gluconeogenesis accounts for the major part of residual glucose production. Furthermore, that gluconeogenesis is an ongoing process that is not affected by either the glucose infusion rate or blood glucose concentration.
Patients and methods
The study was approved by the Institutional Review Board for Human Research at Baylor College of Medicine and the Advisory Board of the General Clinical Research Center at Texas Children's Hospital, Houston, Texas, USA. The infants were recruited from the neonatal intensive care unit at Texas Children's Hospital and they were enrolled after at least one parent had provided written informed consent.
Eight consecutive premature infants (four boys and four girls) fulfilling our inclusion criteria, that is, gestational age 29 weeks or less; weight appropriate for gestational age; and absence of syndromes, anomalies and sepsis were studied. Furthermore, the infants must be clinically stable on intermittent positive pressure ventilation or continuous positive airway pressure and not have a mother with diabetes or substance abuse.
All mothers had received antenatal steroids (Celestone). Three infants were delivered vaginally and five by Caesarean section. The average Apgar score at 5 min was 7, with no infant having less than 5. Subject characteristics are shown in table 1.
At time of the study, six out of eight infants were receiving caffeine and three out of eight were receiving dopamine (two at 6.4 µg/kg per min and one at 2.8 µg/kg per min). Two infants were still receiving antibiotics (ampicillin plus gentamicin). In the remaining six infants, the antibiotics had been discontinued before the start of the study. No infant had any positive cultures or clinical signs of sepsis. None of the infants had received insulin. The infants were clinically stable, with oxygen saturation between 85% and 95% either on intermittent positive pressure ventilation (n=5) (19±1/5±1 cm water and fractional inspired oxygen 0.25±0.03) or continuous positive airway pressure (n=3) (9±1 cm water and fractional inspired oxygen averaging 0.29±0.04).
None of the infants had received any enteral feedings. During the study, the TPN was continued at the prestudy rates as ordered by the attending physician. These rates had been maintained for approximately 12 h at start of the study. When the measurements of metabolic parameters were obtained, that is, at study hours 9.5 and 10, the TPN had been administered at unchanged rates for more than 20 h. Therefore, the TPN provided glucose at 10.12±0.50 mg/kg per min (range 7.37–11.44 mg/kg per min=10.61–16.48 g/kg per day). In addition, the infants received a tracer dose of [6, 6-2H2] glucose to measure the glucose appearance rate7 11 21 22 (see details below under ‘Compounds labelled with stable isotopes’) resulting in a total glucose infusion rate of 10.29±0.50 mg/kg per min (range 7.54– 11.60 mg/kg per min=10.86–16.70 g/kg per day). Lipids (represented by 20% Intralipid; Kabivitrum, Stockholm, Sweden) were given at a rate of 1.91±0.25 mg/kg per min (range 0.81–3.85 g/kg per day) and amino acids (TrophAmine; Braun Medical Inc, Bethlehem, Pennsylvania, USA) at 2.06±0.05 mg/kg per min (range 1.83–2.18 mg/kg per min=2.63–3.13 g/kg per day). The fluid volume averaged 119±4 ml/kg per day (range 100–130 ml/kg per day).
Compounds labelled with stable isotopes
Sterile and pyrogen-free deuterium oxide (2H2O), 99 atom per cent 2H and [6, 6-2H2] glucose, 99 atom% 2H were purchased from Cambridge Isotopes Laboratories (Andover, Massachusetts, USA). The labelled compounds were tested again for sterility and pyrogenicity, dissolved in isotonic saline, and prepared for intravenous infusion by the Investigation Pharmacy at Texas Children's Hospital, Houston, Texas, USA.
The infusions of glucose, Intralipid and TrophAmine were continued at the prestudy rates throughout the 10-h study period. In addition, during the first two study hours, a total of 4 g/kg of sterile deuterated water (1 g/ml, made isotonic by the addition of sodium chloride) was given intravenously at a constant rate of 0.033±0.000 g/kg per min. At completion of the infusion of deuterated water (at study hour 2), a constant rate infusion of [6, 6-2H2] glucose (10 mg/ml) was initiated and continued for 8 h at 0.165±0.005 mg/kg per min to measure the glucose appearance rate.7 11 21 22 The TPN solution and the compounds labelled with stable isotopes were infused by means of umbilical venous catheters already in place for clinical care purposes.
Three blood samples (a total of 3 ml/kg) were obtained during the study: one just before the start of the infusion of deuterated water, one at study hour 9.5 and one at study hour 10. The blood samples were obtained by means of umbilical artery catheters, already in place for clinical care purposes.
The blood samples were analysed for blood glucose concentrations using a glucose analyser (YSI 2300 Stat Plus; YSI Inc, Yellow Springs, Ohio, USA). Isotopic enrichment of [6, 6-2H2] glucose was measured by gas chromatography–mass spectrometry (GCMS) (6890/5973; Agilent Technologies, Wilmington, Delaware, USA) using the penta-acetate derivative.21 22 The incorporation of deuterium in glucose was determined according to two different methods: (1) using the average deuterium enrichment in glucose carbons 1, 3, 4, 5 and 6 according to Chacko et al19 and (2) using the hexamethylenetetramine (HMT) derivative followed by analysis by GCMS in the electron impact mode according to Landau et al.20 23
Briefly, the average deuterium method19 involves preparation of the penta-acetate derivative of glucose, followed by sample analysis using GCMS in the positive chemical ionisation mode. Selective ion monitoring of m/z 170/169 was performed to determine the M+1 enrichment of deuterium in the circulating glucose carbons (C-1, 3, 4, 5, 6, 6) (M is the base mass, 169, representing unlabelled glucose).19 After subtracting the enrichment of M+1 resulting from the natural abundance, the average enrichment of deuterium on a gluconeogenic carbon was calculated from these M+1 data.19 Deuterium enrichment in plasma water was determined by isotope ratio mass spectrometry (Delta+XL IRMS; Thermo Finnigan, Bremen, Germany). For simplicity, from now on we refer to the average deuterium enrichment method as the ‘Chacko method’ and the C5/HMT method as the ‘Landau method’.
All kinetic measurements were performed under steady state conditions and based on the mean isotopic enrichments obtained at study hours 9.5 and 10 (no values for isotopic enrichments or blood glucose concentration differed by more than 5% between the two time points). Glucose appearance rates were calculated using established isotope dilution equations.7 22
Fractional gluconeogenesis (ie, gluconeogenesis as a fraction of the glucose appearance rate) was calculated according to Chacko et al19 as follows:
Where (M+1)(2H)(m/z170/169) is the M+1 enrichment of deuterium in glucose measured using m/z 170/169 and ‘6’ is the number of 2H labelling sites on the m/z 170/169 fragment of glucose (ie, the average M+1 enrichment derived from deuterated water) and E2H2O is the deuterium enrichment in plasma water. Fractional gluconeogenesis was also calculated from the deuterium enrichment in glucose carbon 5 derived from deuterated water (product) and the deuterium enrichment in plasma water (precursor) according to the Landau method as previously described.20 23
Where E2HC5 is the deuterium enrichment in glucose carbon 5 by the deuterium incorporation from deuterated water using the HMT derivative20 and E2H2O deuterium enrichment in plasma water. For both methods, rates of gluconeogenesis were calculated as the product of the total glucose appearance rate (Ra)19 20 23 and fractional gluconeogenesis (GNG). Glycogenolysis was calculated by subtracting the rate of gluconeogenesis from the GPR.
Where GNG% Ra is gluconeogenesis as a fraction of the appearance rate.
All results are provided as mean±SE. A p value less than 0.05 was used to define significance. Linear regression analysis was applied to analyse relationships between measured variables, that is, glucose appearance rate, glucose concentration, gluconeogenesis, gestational age and birth weight. Multiple regression analysis was used to assess the effects of individual substrate infusion rates (glucose, lipids and amino acids) and interactions among them on gluconeogenesis and blood glucose concentration. The two methods to measure gluconeogenesis19 20 were compared using paired t test and Bland and Altman's test.
The blood glucose concentration averaged 8.9±1.2 mmol/l (160±21 mg/dl) (range 5.2–14.3 mmol/l; 94–257 mg/dl).
The glucose appearance rate averaged 12.4±0.6 mg/kg per min (68.8±3.3 µmol/kg per min).
GPR averaged 2.1±0.3 mg/kg per min (11.7±1.7 µmol/kg per min).
Gluconeogenesis and glycogenolysis: fractional gluconeogenesis (ie, gluconeogenesis as a fraction of glucose appearance rate) was 11.0±1.1% (Chacko method)19 and 10.5±1.2% (Landau method)20 23 (NS, p=0.46) (figure 1). The rate of gluconeogenesis was 1.35±0.15 mg/kg per min (7.44±0.78 µmol/kg per min) (Chacko method)19 and 1.29±0.14 mg/kg per min (7.17±0.78 µmol/kg per min) (Landau method),20 respectively (NS, p=0.33) (table 2). Gluconeogenesis accounted for 73±11% (Chacko method)19 and 68±10% of glucose production (Landau method)20 (NS, p=0.36). The rate of glycogenolysis was 0.71±0.27 and 0.81±0.27 mg/kg per min for the Chacko19 and Landau methods,20 respectively (NS, p=0.32).
Bland and Altman's test showed a mean difference of 0.46% between the methods for fractional gluconeogenesis (figure 2) and 0.1 mg/kg per min for the rate of gluconeogenesis. All data points were within ±2 SD.
Gluconeogenesis did not correlate with the glucose infusion rate (R2=0.06) (figure 3), glucose concentration (R2=0.02), gestational age (R2=0.03) or birth weight (R2=0.05). After controlling for the glucose infusion rate, gluconeogenesis did also not correlate with infusion rates of lipids or amino acids. Similarly, these factors did not affect glycogenolysis. The blood glucose concentration was directly correlated with the glucose appearance rate, R2=0.51, p=0.047. Of the two components of the glucose appearance rate (exogenous glucose plus glucose production), the glucose infusion rate accounted for the major part of the variance in blood glucose concentration, R2=0.43 compared with glucose production, R2=0.09. Adding gestational age to the regression analysis increased the R2 value to 0.79 (p=0.02) (individual p values were 0.036 for glucose appearance rate and 0.047 for gestational age). Therefore, collectively, the glucose appearance rate and gestational age explained 79% of the variance in blood glucose concentration. After controlling for the effect of the glucose infusion rate, the amino acid and/or lipid infusion rate did not have any significant effect on the blood glucose concentration.
In contrast to healthy adults,16 17 premature infants do not completely turn off their glucose production when receiving parenteral glucose corresponding to their normal GPR.10 16 18 The present study demonstrates that in very premature infants, glucose production is not completely suppressed even when glucose is supplied at rates exceeding their normal GPR as part of TPN. Furthermore, gluconeogenesis is sustained and accounts for the major part of residual glucose production in these infants. Rates of gluconeogenesis were not affected by infusion rates of glucose, lipids and amino acids or blood glucose concentration, gestational age and birth weight.
We and others demonstrated several years ago that premature infants receiving no or very small amounts of intravenous glucose are capable of producing glucose at rates comparable to those of term newborns within a few hours of birth.7 11 12 24 25 It was also shown that the gluconeogenic pathway was activated in the immediate neonatal period in both term and preterm infants.26,–,29 Despite their minimal body fuel stores, very premature infants were capable of using endogenous glycerol for glucose production to maintain normoglycaemia at least for a shorter period of time.28 As the activity of gluconeogenic enzymes is very low during fetal life, this indicates that it is the birth process itself rather than gestational age that activates key gluconeogenic enzymes.
In a subsequent series of studies, we explored whether very preterm infants were also capable of producing glucose from parenterally supplied lipids and amino acids.13,–,15 We demonstrated that newborn very premature infants receiving parenteral nutrition with the glucose supply reduced to half their normal GPR (3 mg/kg per min) maintained normoglycaemia over periods of at least 10–12 h by producing glucose via the gluconeogenic pathway.13 Furthermore, the parenteral lipid emulsion (primarily the glycerol part) was more important than the amino acids as substrate for gluconeogenesis.13,–,15
These previous results led up to the question addressed in the present study, that is, whether gluconeogenesis is suppressed when very premature infants receive parenteral nutrition providing glucose at rates exceeding normal glucose turnover rates. Under these conditions, the infants' glucose and energy needs are supplied by the parenteral nutrition and, theoretically, there would be no need for glucose produced by gluconeogenesis. However, rates of gluconeogenesis and the proportion of glucose production found in the present study (during the infusion of parenteral nutrition providing glucose at an average of 10.3 mg/kg per min) were virtually identical to those obtained in infants receiving parenteral nutrition providing glucose at only 3 mg/kg per min.13 Furthermore, rates of gluconeogenesis were not affected by the blood glucose concentration (8.9 mmol/l=160 mg/dl in the present study compared with 3 mmol/l=54 mg/dl in previous studies).13 These results clearly demonstrate that in newborn very premature infants receiving routine TPN, gluconeogenesis is an ongoing process that is not affected by either the glucose infusion rate or blood glucose concentration. The incomplete suppression of glucose production is thus primarily due to the contribution from gluconeogenesis.
The incomplete suppression of glucose production and its components in premature newborns might be a result of hepatic insulin resistance and/or insufficient insulin secretion. A potential mechanism might also be hyperglucagonaemia, which has been shown to contribute to increased glucose production from gluconeogenesis and glycogenolysis in individuals with type 2 diabetes.30 After controlling for the glucose infusion rate, infusion rates of lipids and amino acids did not affect rates of gluconeogenesis. In this study, the infusion rates of lipids varied within a wide range (0.8– 3.9 g/kg per day), while amino acid infusion rates were tight (2.6–3.1 g/kg per day). Multiple regression analyses did not show any significant interaction among substrates (glucose, lipids and amino acids) that could potentially affect gluconeogenesis. We also investigated which factors had an impact on blood glucose concentration in these infants. The glucose appearance rate, that is, exogenous glucose plus glucose production, explained 51% of the variation in blood glucose concentration. The glucose infusion rate was the major component contributing 43% of this variation, whereas residual glucose production played a minor role. The blood glucose concentration also correlated with gestational age. Collectively, the glucose appearance rate and gestational age explained 79% of the variation in blood glucose concentration. After controlling for the glucose infusion rate, lipid and amino acid infusion rates had no effect on blood glucose concentration.
In very preterm infants, the blood volume that can be safely withdrawn is limited. The method of Chacko et al19 requires only 25 µl plasma per sample, thus providing a valuable tool to study glucose metabolism and its regulation in newborn infants, particularly those born prematurely. As this method was not available when we designed the study and performed the initial measurements of gluconeogenesis, we used the Landau method,20 which requires large sample volumes (∼500 µl/sample). This limited the number of blood samples that could be obtained and precluded analyses of glucose-regulating hormones, for example, insulin and glucagon. In the present study, the infants received substantial amounts of intravenous glucose, resulting in low fractional gluconeogenesis (∼11%). To confirm the accuracy of our measurements of gluconeogenesis, we used both the Landau method20 and the method of Chacko et al.19 We previously demonstrated that this method compared very well with the Landau method in overnight and 3-day fasting adults, in whom gluconeogenesis as a fraction of the glucose appearance rate (fractional gluconeogenesis) ranged between 40% and 100%. The results of the present study demonstrated that the Chacko method compares very well with the Landau method also when fractional gluconeogenesis is low. The mean difference between the methods was only 0.46% for fractional gluconeogenesis corresponding to 0.1 mg/kg per min for the gluconeogenic rate. In conclusion, in very preterm infants receiving routine TPN providing glucose at rates exceeding their normal GPR, gluconeogenesis accounts for the major part of residual glucose production. Rates of gluconeogenesis were not different from those reported earlier in similar infants receiving parenteral nutrition with the glucose supply reduced to half their normal GPR, with resultant decreased blood glucose concentrations. These results demonstrate that gluconeogenesis is an ongoing process that is independent of both the glucose infusion rate and blood glucose concentration in very premature infants. Collectively our present and previous results thus suggest that a potential strategy to prevent hyperglycaemia without increasing the risk of hypoglycaemia or insufficient energy intake would be to provide a TPN solution supplying glucose at a rate corresponding to the normal GPR of newborn infants and lipids and amino acids according to current clinical routines during the first days of life.
Future studies using the Chacko method will allow us to determine whether insulin, glucagon and/or other factors regulate gluconeogenesis.
This work is a publication of the USDA/ARS Children's Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, Houston, Texas, USA. The contents of this publication do not necessarily reflect the views or policies of the US Department of Agriculture, nor does mention of trade names, commercial products, or organisations imply endorsement from the US government. The authors would like to thank Drs Dennis M Bier, Morey W Haymond and E O'Brian Smith (biostatistician) at the Children's Nutrition Research Center, Houston, Texas, USA, for invaluable help and advice; Cindy Bryant, Pamela Gordon, Geneva Shores, Cindy Clarke and Patricia Langley for excellent assistance; the pharmacist at Texas Children's Hospital for preparing isotope solutions; and the staff of the neonatal intensive care unit for professional collaboration.
Funding This study received support from NIH RO1 HD 37857, USDA Cooperative Agreement #58-6250-6-001, and the General Clinical Research Center, National Center for Research Resources, NIH MO1-RR-001888.
Ethics approval This study was conducted with the approval of the Baylor College of Medicine, Institutional Review Board for Human Research.
Competing interests None.
Patient consent Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.