Evaluation of insulin secretion and action in normal and small for gestational age children

Abstract

This thesis examines three sequentially linked themes that encompass the broader field of insulin secretion and action in normal children and those born small for gestational age (SGA). In section A, in vivo and in vitro methods to evaluate insulin secretion and action in children are examined. The establishment and assessment of techniques to assess insulin action in children enabled selection of the most appropriate tools to study insulin action and related metabolic changes in SGA children, as addressed in section B. In section C, the effect of growth hormone (GH) therapy on insulin action is examined. The first phase insulin response (FPIR) is a widely used method to evaluate β-cell function during the prediabetic phase of evolving type 1 diabetes mellitus. In a large cohort of normal children and adolescents we found that the following clinical and procedural parameters influenced the FPIR: puberty, age, relative adiposity and peak blood glucose level achieved during the FPIR. Despite correcting for the influence of puberty and obesity, there remained considerable unaccounted variability in FPIR. Wide intersubject variability and the need to develop detailed standards that incorporate factors that influence FPIR limit the applicability of this measure of insulin secretory capacity, especially in children in whom preventive measures to delay or prevent the appearance of Type 1 diabetes mellitus may be considered. The modified minimal model (MMM) is a mathematical model developed in adults to calculate insulin sensitivity (SI) from plasma glucose and insulin data drawn from a 180 min frequently sampled iv glucose tolerance test (FSIGT) with tolbutamide. The efficacy and safety of this model had not been evaluated in children. We demonstrated that the FSIGT is a safe procedure to perform in children, and that variables known to influence clamp derived SI in children such as puberty and obesity have a major influence on minimal model derived SI, supporting the efficacy of this technique in children. In addition we validated a shortened 90 min FSIGT and a modification of the minimal model that is more appropriate for children. A non-invasive in vitro test of insulin action that correlates with in vivo insulin sensitivity had not been developed. We demonstrated in vitro insulin mediated glucose uptake from cultured monocytes correlated with minimal model derived insulin sensitivity (SI) in children. we also showed that SGA children had reduced insulin mediated glucose uptake in monocytes. This labour intensive method may be an alternative to assess insulin action in very young children. Homeostasis model assessment (RHOMA) and quantitative insulin check index (QUICKI) are popular methods to assess insulin resistance/sensitivity in adults based upon a single fasting blood sample, but had never previously been validated in children. In a large cohort of prepubertal children, RHOMA was no better correlated than fasting plasma insulin with the gold standard, minimal model derived SI (SIMM). QUICKI was very poorly correlated with SIMM. Neither RHOMA nor QUICKI were able to detect changes in SI associated with obesity or growth hormone (GH) treatment. RHOMA and QUICKI were shown to be imprecise and no better than fasting insulin at detecting changes in SI and are not recommended as tools to accurately determine SI in prepubertal children. Epidemiological studies have found an association between reducing birth weight and increasing risk of type 2 diabetes mellitus in late adult life. Evidence of Metabolic abnormalities in early life in SGA children had not been explored. We found that short prepubertal SGA children had reduced insulin sensitivity when compared to an appropriate control group, which we speculate is a risk factor for the future development of type 2 diabetes mellitus. In addition short, SGA children had elevated plasma IGF-I and IGFBP-3 levels, which were likely secondary to hyperinsulinism, as plasma insulin and IGF-I values were correlated. GH exhibits anti-insulin effects. The impact of GH therapy on the development of diabetes mellitus had never previously been explored in a large study In a study of over 23,000 children with growth disorders enrolled within the KIGS database, comprising 52,000 years of GH treatment, the incidence of type 2 diabetes mellitus was six-fold greater than reported in children and adolescents not treated with GH. Following complex incidence modelling, no increase in the incidence of type 1 diabetes mellitus was found during GH therapy. We postulate that the higher than expected incidence of type 2 diabetes mellitus with GH treatment may be due to the acceleration of evolving diabetes mellitus in predisposed individuals. Reduction in insulin sensitivity has been shown during GH therapy in growth disorders such as Turner Syndrome and inconsistently in children with normal variant short stature. Formal evaluation of insulin sensitivity during GH therapy in SGA and following treatment in any growth disorder had never been conducted. We found that SGA children who have pre-existing reduced insulin sensitivity have a further reduction in insulin sensitivity during GH treatment. Furthermore, insulin sensitivity did not improve three months after GH treatment was discontinued. The adverse effect of GH treatment on insulin sensitivity in SGA children raises an issue about the potential long term effect of this treatment on future type 2 diabetes mellitus risk. To assess whether cotreatment with insulin sensitisers could overcome GH induced insulin resistance seen in our SGA children, we evaluated the impact of one month of troglitazone and metformin during GH therapy in SGA juvenile rats. Troglitazone and metformin corrected GH induced hyperinsulinaemia and led to a reduction in systolic blood pressure and free fatty acids. Therefore insulin sensitiser cotreatment with GH may ameliorate the adverse metabolic effects of GH therapy in children, but further studies are required to confirm this.