No abstract

Aim To compare the efficacy of three strategies for real‐time continuous glucose monitoring (RT‐CGM) over 12 months in children and adolescents with type 1 diabetes. Methods A French multicenter trial (NCT00949221) with a randomized, controlled, prospective, open, and parallel‐group design was conducted. After 3 months of RT‐CGM, patients were allocated to one of three groups: return to self‐monitoring of blood glucose, continuous CGM (80% of the time), or discontinuous CGM (40% of the time). The primary outcome was hemoglobin A1c (HbA1c) levels from 3 to 12 months. The secondary outcomes were acute metabolic events, hypoglycemia, satisfaction with CGM and cost. Results We included 151 subjects, aged 2 to 17 years, with a mean HbA1c level of 8.5% (SD0.7; 69 mmol/mol). The longitudinal change in HbA1c levels was similar in all three groups, at 3, 6, 9 and 12 months. The medical secondary endpoints did not differ between groups. The rate of severe hypoglycemia was significantly lower than that for the pretreatment year for the entire study population. Subjects reported consistent use and good tolerance of the device, regardless of age or insulin treatment. The use of full‐time RT‐CGM for 3 months costs the national medical insurance system €2629 per patient. Conclusion None of the three long‐term RT‐CGM strategies evaluated in pediatric type 1 diabetes was superior to the others in terms of HbA1c levels. CGM‐use for 3 months decreased rates of severe hypoglycemia. Our results confirm the feasibility of long‐term RT‐CGM‐use and the need to improve educational support for patients and caregivers.

This randomized control trial investigated glucose control with closed-loop (CL) versus threshold-low-glucose-suspend (TLGS) insulin pump delivery in pre-pubertal children with type 1 diabetes in supervised hotel conditions. The patients [n = 24, age range: 7-12, HbA1c: 7.5 ± 0.5% (58 ± 5 mmol/mol)] and their parents were admitted twice at a 3-week interval. CL control to range or TLGS set at 3.9 mmoL/L were assessed for 48 hour in randomized order. Admissions included three meals and one snack, and physical exercise. Meal boluses followed individual insulin/carb ratios. While overnight (22:00-08:00) per cent continuous glucose monitoring (CGM) time below 3.9 mmol/L (primary outcome) was similar, time in ranges 3.9 to 10.0 and 3.9 to 7.8 mmoL/L and mean CGM were all significantly improved with CL (P < 0.001). These results were confirmed over the whole 48 hour. Disconnections between devices and limited accuracy of glucose sensors in the hypoglycaemic range appeared as limiting factors for optimal control. CL mode was well accepted while fear of hypoglycaemia was unchanged. CL did not minimize nocturnal hypoglycaemia exposure but improved time in target range compared to TLGS. Although safe and well-accepted, CL systems would benefit from more integrated devices.