Résumé
Objective: Availability of closed-loop insulin delivery systems is expected by patients with type 1 diabetes. Our main objective was to assess whether patients using the Diabeloop artificial pancreas system could achieve a better glucose control compared to sensor-assisted pump therapy, as measured by percentage time in range 70-180 mg/dl.
Methods: A multicenter clinical trial was conducted in a home setting, using an MPC-based algorithm, a hybrid monohormonal system (CellNovo® pump and Dexcom G5® sensor) and centralised remote monitoring. Adult patients with T1D for ≥ 2 years, pump therapy for ≥ 6 months, HbA1c ≤ 10%, and preserved hypoglycemia awareness were included among 12 investigation centers. Design was a crossover trial with a 2-week run-in period, followed by 2 sequences of 12 weeks each separated with washout. Results of the first 12-week arm are reported.
Results: 68 patients (27 men, age 47.2±13.4 years, HbA1c 7.6±0.9%, duration of disease 27.9±13.2 years) were randomised between closed-loop (CL, n=33) and open-loop (OL, n=35) first. During the first 12-week period, time spent in the 70-180 mg/dl range was 69.5% [62.4;75.6] in CL vs. 58.7% [45.2;64.9] in OL. Time in hypoglycaemia <70 mg/dl was 2.1% [1.2;2.6] in CL vs. 4.1% [1.9;6.1] in OL.
Conclusion: The Diabeloop closed-loop system was efficient regarding metabolic outcomes.