Context: The added benefit of glucagon in artificial pancreas systems for overnight glucose control in type 1 diabetes has not been fully explored.
Objective: The objective of the study was to compare the efficacy of dual-hormone (insulin and glucagon) artificial pancreas, single-hormone (insulin alone) artificial pancreas, and conventional insulin pump therapy.
Design: This study was a three-center, three-arm, open-label, randomized, crossover controlled trial involving three interventions, each applied over a night after a high carbohydrate/high fat meal and a second after exercise to mimic real-life glycemic excursions.
Setting: The study was conducted in a home setting.
Patients: Twenty-eight type 1 diabetes participants (21 adults and seven adolescents) participated in the study.
Interventions: Dual-hormone artificial pancreas, single-hormone artificial pancreas, and conventional pump therapy was activated from 9:00 PM to 7:00 AM.
Main outcome: The main outcome was a proportion of time in target (4-8 mmol/L) by continuous glucose monitoring from 11:00 PM to 7:00 AM. Analysis was by intention to treat.
Results: The median (interquartile range) percentage of time-in-target glucose range was 47% (36%-71%) for conventional therapy, higher on both single-hormone (76% [65%-91%], P < .001) and dual-hormone artificial pancreas (81 [68%-93%], P < .001). The median (interquartile range) time spent below 4 mmol/L was 14% (4%-28%) for conventional therapy, lower on both single-hormone (5% [0%-13%], P = .004) and dual-hormone artificial pancreas (1% [0%-8%], P < .001). There were 14 hypoglycemic events on conventional therapy compared with six incidences on the single-hormone artificial pancreas (P = .059) and three incidences on the dual-hormone artificial pancreas (P = .017). None of these outcomes differed significantly between single- and dual-hormone configurations.
Conclusions: Single- and dual-hormone artificial pancreas systems both provided better glucose control than conventional therapy. Although the dual-hormone configuration did not increase overnight time-in-target glucose levels, an effect on lowering hypoglycemia risk cannot be ruled out.
Trial registration: ClinicalTrials.gov NCT01905020.