Factors affecting glucose homeostasis in premature neonates

Introduction After birth rapid metabolic adaptation is required to establish glucose homeostasis.Preterm birth disrupts normal timing of this physiological adaptation putting premature neonates at risk of hypo- and hyperglycaemia. A major barrier to conducting pharmacokinetic studies in neonates is the relatively large volume of blood required by most assays. The use of dried blood spots (DBS) has the potential to enable the use of smaller sample volumes and simplifies processing and handling. However, haematocrit in neonates varies significantly and can affect the accuracy of DBS assays. Insulin and C-peptide are secreted in equimolar amounts from the pancreas. C-peptide can be used as a biomarker for insulin secretion as, unlike insulin, it is not hepatically cleared and has a longer plasma half-life. Urinary C-peptide may give a non-invasive way of measuring insulin production in neonates. The insulin/C-peptide (I/CP) ratio gives an indication of relative insulin clearance in adults but normal ranges have not been reported in neonates. Premature neonates have significantly higher fasting levels of glucagon-like peptide-1 (GLP-1) than adults and are increased even further with feeding. Hyperglycaemia can often occur in extremely premature neonates and is associated with an increase in morbidity and mortality. Insulin resistance is a key issue in preterm neonates and it is thought that hyperglycaemia is a consequence of relative insulin deficiency. Aims The aims of this research were: to investigate the use of a DBS assay for measuring insulin concentrations to improve sampling for pharmacokinetic studies; to describe the relationship between C-peptide and insulin secretion in the neonatal population; to investigate the relationship between feeding, incretin (GLP-1) and insulin secretion in neonates. Finally, to compare these results with neonates treated with insulin for hyperglycaemia to investigate alterations in glucose homeostasis in hyperglycaemic neonates. Methods Blood samples, DBS and urine samples were obtained from euglycaemic and hyperglycaemic insulin-treated preterm neonates admitted into the Dunedin Hospital Neonatal Intensive Care Unit (NICU) for this research project. DBS of varying haematocrit (0.25-0.65) were prepared at three insulin plasma concentrations (10, 25 and 50mU/L) and were analysed for insulin using the method developed by Butter et al (2001). DBS and paired plasma samples were obtained from neonates. Standard addition and direct measurement from the kit standard curve were compared for measurement of urinary C-peptide. Insulin concentrations in plasma, along with C-peptide concentrations in plasma and urine (diluted to 40%) were measured in 20 neonates. Urine volume and time of collection were recorded and used to calculate C-peptide urinary excretion rate (UER) and clearance (Cl). Samples for 102 euglycaemic neonates were obtained from leftover blood from routine tests. Samples were obtained for 10 hyperglycaemic insulin-treated neonates and there were 9 for which additional blood was collected at the same time as the patients’ routine four hourly blood glucose test over a 24-hour period (6 samples). Data were collected from clinical notes for gestational age (GA), postmenstrual age (PMA), postnatal age, birth weight, current weight, birth length, birth head circumference, oral intake, and intravenous fluids at time of samples. Insulin and C-peptide concentrations from the hyperglycaemic insulin-treated neonates were compared with a group of age-matched (PMA <30 weeks) euglycaemic neonates. Plasma was analysed for insulin and C-peptide using chemiluminescent kits (Invitron, UK), GLP-1 and glucagon using ELISA kits (BioCore Pty Ltd, Australia). Statistical analyses were performed using Stata/IC (Version 11.2). Results Insulin chemiluminescence responses were significantly lower at higher haematocrit values (p 0.05). Direct measurement from kit standards produced a more precise and simple method for determining C-peptide concentration. The mean (SD) C-peptide Cl was 0.309 (0.329) mL/min/kg and UER was 0.0329 (0.0342) pmol/min/kg. Correlations between Cl or UER and gestational age were not significant (p >0.05). No significant correlation was shown between UER or Cl and blood glucose concentration (BGC). Insulin and C-peptide concentrations were elevated in very preterm infants (PMA ≤32 weeks) and decreased as the neonate approached term. Data were natural log normally distributed and linear regression of ln-transformed data was performed. Linear regressions showed that I/CP ratio and insulin/BGC (I/BGC) ratio decreased significantly with increasing PMA (p <0.001). ANOVA showed GLP-1 increased following initial oral feeds regardless of PMA (p <0.001). C-peptide concentrations were significantly lower in the hyperglycaemic neonates indicating lower insulin production in hyperglycaemic neonates. Insulin concentrations, blood glucose concentrations, C-peptide concentrations and I/CP ratios were all significantly affected by PMA (p <0.01) in the hyperglycaemic neonates. As expected insulin concentrations were higher in the hyperglycaemic neonates confirming that exogenous insulin is reaching these neonates intravenously and not retained in the i.v apparatus. Conclusions Measuring plasma insulin concentrations with DBS in neonates is not an ideal method due to high variability. Since GA and BGC did not appear to affect Cl and UER, urinary C-peptide may provide a non-invasive method of measuring insulin production in neonates. Preterm neonates exhibit insulin resistance irrespective of hyperglycaemia. The significantly higher GLP-1 concentrations in fed neonates confirm that enteral feeds result in an increase in GLP-1. These factors are likely to contribute to the increased risk of hyperglycaemia in premature neonates (PMA