Estimation of insulin sensitivity from continuous glucose monitoring and insulin pump data in type 1 diabetes

Diabetes mellitus is a disease characterized by chronic hyperglycemia either due to a lack of secretion of the hormone insulin (Type 1 Diabetes Mellitus, T1DM) or due to impaired action of this hormone (Type 2 Diabetes Mellitus, T2DM). Due to its short- and long-term complications, it is currently one of the major health problems of the economically developed countries and at the same time, one of the first items of healthcare spending. Diabetic patients therefore need regular blood glucose monitoring associated with adequate insulin therapy whose goal is to keep glucose concentration within the normal safe range (70  180 mg/dl), trying to limit excursions in hypoglycemic (20  70 mg/dl), due to short-term complications, and hyperglycemic range (180  600 mg/dl), due to long-term complications. In order to optimize insulin therapy, and then assess the correct amount of insulin to be administered to the patient, it is necessary to know its insulin sensitivity (SI), i.e. the ability of insulin to stimulate glucose utilization and inhibit its production, specific for each individual and changing during the day. The aim of this thesis is to estimate an index of insulin sensitivity in patients with type 1diabetes by using a recently proposed technique which exploits minimally invasive technologies used by diabetic patients for control therapy. This parameter will be estimated in correspondence of meals over the whole day and, in order to be able to estimate the index of insulin sensitivity even in the presence of meals close, a tool for the estimation of carbohydrates absorbed during the meal (Carbohydrates On Board, COB) will be proposed. In Chapter 1 the glucose-insulin regulatory system, diabetes and its complications, conventional therapy for its control and indices of insulin sensitivity in literature are introduced. In Chapter 2 the experimental protocols applied to the patients and the data available data are presented. In Chapter 3 the recently proposed method for the estimation of SI using minimally invasive technologies and the COB function for the estimation of carbohydrates absorbed during the meal are presented. In Chapter 4 the estimates of insulin sensitivity in different datasets with and without the COB function, which in turn was developed using simulated and real data, are presented