Population modelling approach for the study of insulin clearance in previous gestational diabetes.

Gestational diabetes is a high blood glucose condition that develops during pregnancy and although it usually disappears after giving birth, women who experienced gestational diabetes are more prone to develop type 2 diabetes later in their life. Insulin clearance, a physiological process representing the removal of insulin from the blood in the entire organism, is one of the main processes underlying the development of type 2 diabetes, together with insulin resistance and altered insulin secretion. Due to the role of insulin clearance in the development of type 2 diabetes, it is important to investigate this process also in women who experienced a history of gestational diabetes. Population modelling is a tool that allows to find correlations between heterogenic characteristics of subjects and to study the metabolism of certain molecules within the body. The aim of the present thesis was to exploit a population modelling approach for the study of insulin clearance in previous gestational diabetes. To this purpose, a mathematical model able to segregate hepatic and extrahepatic insulin clearance, previously proposed by Polidori et al., has been considered. The involved population consisted of 114 women with an history of gestational diabetes (pGDM) and a group of 41 healthy women as controls (CNT) who underwent an insulin modified intravenous glucose tolerance test. Data were processed with Monolix, a software providing simple solution for non-linear mixed effects modeling for pharmacometrics. To exploit the information about the heterogeneous characteristics among the population certain covariates, among those suggested by the software, were included into the model. Analyses were performed either on the complete dataset (OVP, overall population) and including the group (CNT, pGDM) as a categorical variable or considering the CNT and pGDM datasets separately. Population estimates of extrahepatic insulin clearance for pGDM women resulted almost three times smaller with respect to that of CNT group (0,32 L/min vs 0,91 L/min). Instead, concerning the hepatic insulin clearance population estimates, the pGDM population showed a higher value with respect to CNT (29,7% vs 44,7%). Individual estimates for FEL resulted significantly different for pGDM vs CNT, whereas CLP were found different only when considering the separate datasets. In conclusion, the proposed population modelling approach showed its capability to provide population parameter estimates related to hepatic and extrahepatic insulin clearance. Hepatic insulin clearance may be affected by the presence of a history of gestational diabetes, whereas extrahepatic insulin clearance requires further investigation in wider populations.