Overnight closed-loop insulin delivery in young people with type 1 diabetes: a free-living, randomized clinical trial

Abstract

Objective: To evaluate feasibility, safety, and efficacy of overnight closed-loop insulin delivery in free-living youth with type 1 diabetes.

Research design and methods: Overnight closed loop was evaluated at home by 16 pump-treated adolescents with type 1 diabetes aged 12-18 years. Over a 3-week period, overnight insulin delivery was directed by a closed-loop system, and on another 3-week period sensor-augmented therapy was applied. The order of interventions was random. The primary end point was time when adjusted sensor glucose was between 3.9 and 8.0 mmol/L from 2300 to 0700 h.

Results: Closed loop was constantly applied over at least 4 h on 269 nights (80%); sensor data were collected over at least 4 h on 282 control nights (84%). Closed loop increased time spent with glucose in target by a median 15% (interquartile range -9 to 43; P < 0.001). Mean overnight glucose was reduced by a mean 14 (SD 58) mg/dL (P < 0.001). Time when glucose was <70 mg/dL was low in both groups, but nights with glucose <63 mg/dL for at least 20 min were less frequent during closed loop (10 vs. 17%; P = 0.01). Despite lower total daily insulin doses by a median 2.3 (interquartile range -4.7 to 9.3) units (P = 0.009), overall 24-h glucose was reduced by a mean 9 (SD 41) mg/dL (P = 0.006) during closed loop.

Conclusions: Unsupervised home use of overnight closed loop in adolescents with type 1 diabetes is safe and feasible. Glucose control was improved during the day and night with fewer episodes of nocturnal hypoglycemia.

Trial registration: ClinicalTrials.gov NCT01221467.

Figure 1

Design of the study evaluating overnight closed-loop delivery.

Figure 2

Sensor glucose (top panel) and insulin delivery (bottom panel) from 2300 to 0700 h. Median overnight profiles are represented by red solid (closed loop) and black dashed (control therapy) lines. The interquartile range is shown as red (closed loop) and gray (control therapy) regions. The plots demonstrate progressively tighter overnight glucose levels during closed-loop therapy brought about by more varying insulin delivery, the typical tradeoff of insulin delivery variability for glucose consistency. The target glucose range 70–144 mg/dL is denoted by short dashed lines in the top panel.

Figure 3

Mean overnight glucose levels during closed-loop and control therapy. Individual mean values were calculated from 2300 to 0700 h. Percentages at the top represent the proportion of nights mean overnight glucose was below, at, and above the target; for example, on 70% of nights, mean overnight glucose was between 70 and 144 mg/dL (vertical dashed lines).