A fully artificial pancreas versus a hybrid artificial pancreas for type 1 diabetes: a single-centre, open-label, randomised controlled, crossover, non-inferiority trial.

BACKGROUND

For people with type 1 diabetes, there is currently no automated insulin delivery system that does not require meal input. We aimed to assess the efficacy of a novel faster-acting insulin aspart (Fiasp) plus pramlintide fully closed-loop system that does not require meal input.

METHODS

In this open-label, randomised controlled, crossover, non-inferiority trial we compared the Fiasp (Novo Nordisk, Bagsværd, Denmark) plus pramlintide closed-loop system with no meal input (fully artificial pancreas) and the Fiasp-alone closed-loop system with precise carbohydrate counting (hybrid artificial pancreas). Adults (≥18 years) who had a clinical diagnosis of type 1 diabetes for at least 12 months, had glycated haemoglobin 12% or lower, and had been on insulin pump therapy for at least 6 months were enrolled at McGill University Health Centre, Montreal, QC, Canada. The Fiasp plus pramlintide fully closed-loop system delivered pramlintide in a basal-bolus manner with a fixed ratio of 10 μg:U relative to insulin. A research staff member counted the carbohydrate content of meals to input in the hybrid closed-loop system. Participants completed the two full-day crossover interventions in a random order allocated by a computer-generated code implementing a blocked randomisation (block size of four). The primary outcome was the percentage of time spent within the glucose target range (3·9-10·0 mmol/L), with a 6% non-inferiority margin, assessed in all participants who completed both interventions. This trial is registered with ClinicalTrials.gov, NCT03800875.

FINDINGS

Between Feb 8, 2019, and Sept 19, 2020, we enrolled 28 adults, of whom 24 completed both interventions and were included in analyses. The percentage of time spent in the target range was 74·3% (IQR 61·5-82·8) with the fully closed-loop system versus 78·1% (66·3-87·5) with the hybrid Fiasp-alone closed-loop system (paired difference 2·6%, 95% CI -2·4 to 12·2; non-inferiority p=0·28). Eight (33%) participants had at least one hypoglycaemia event (<3·3 mmol/L) with the fully closed-loop system compared with 14 (58%) participants with the hybrid closed-loop system (2200-2200 h). Non-mild nausea was reported by three (13%) participants and non-mild bloating by one (4%) participant with the fully closed-loop system compared with zero participants with the hybrid closed-loop system.

INTERPRETATION

The Fiasp plus pramlintide fully closed-loop system was not non-inferior to the Fiasp-alone hybrid closed-loop system for the overall percentage of time in the glucose target range. However, participants still spent a high percentage of time within the target range with the fully-closed loop system. Outpatient studies comparing the fully closed-loop hybrid systems with patient-estimated, rather than precise, carbohydrate counting are warranted.

FUNDING

Diabetes Canada.

Informations de copyright

Copyright © 2021 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. Published by Elsevier Ltd.. All rights reserved.

Déclaration de conflit d'intérêts

Declaration of interests AH received research support or grants from Eli Lilly, Adocia, AgaMatrix, and Diabetes Canada; consulting fees and speaker fees or honoraria from Eli Lilly; insulin pumps, glucose sensors, and monitors from Dexcom, Tandem, and Medtronic; and pending patents in the artificial pancreas area (US patent application 16/893,084; US patent application 16/955,191; US patent application 16/608,054). MAT received research support from AgaMatrix, and speaker honoraria from Eli Lilly, Novo Nordisk, Boeringher Ingelheim, Janssen, and AstraZeneca. J-FY received research support or grants from Bayer, Sanofi, and Novo Nordisk and honoraria from Merck, AstraZeneca, Janssen, Boehringer Ingelheim, Eli Lilly, Novo Nordisk, Medtronic, Abbott, Dexcom, and Sanofi; and participated in advisory boards for Merck, AstraZeneca, Janssen, Boehringer Ingelheim, Eli Lilly, Novo Nordisk, Medtronic, Abbott, Dexcom, and Sanofi. LL received consulting fees from Dexcom, Insulet, and Novo Nordisk, grants or contracts from Merck and AstraZeneca; participated in advisory boards for the US National Institutes of Health, Novo Nordisk, Insulet, and Dexcom; and participated in data safety monitoring boards for the US National Institutes of Health. All other authors declare no competing interests.