In‐flight glycaemic control using automated insulin delivery systems: A within‐subject comparative pilot study

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Background: Automated insulin delivery systems (AID) are increasingly used by people living with type 1 diabetes, but their functionality during air travel remains insufficiently studied. Some concerns have arisen regarding safety using AID systems under flight conditions, particularly related to potential insulin delivery errors induced by cabin pressure changes. Methods: This study assessed the safety and efficacy commercial AIDs in 20 adults with type 1 diabetes (age: 37.9 ± 12.8 years, HbA1c: 52.3 ± 4.9 mmol/mol, TIR: 74 ± 9.8%, diabetes duration: 19.25 ± 10.6 years) during mid‐haul commercial flights and matched ground‐based conditions (MiniMed780G; Omnipod 5; Control‐IQ; and CamAPS FX with YpsoPump, n = 5 per system). Outcomes included time in range (TIR, 3.9–10.0 mmol/L), time in tight range (TITR, 3.9–7.8 mmol/L), time above range 1 (TAR1, 10.1–13.9 mmol/L), time above range 2 (TAR2, >13.9 mmol/L), time below range 1 (TBR1, 3.0–3.8 mmol/L), time below range 2 (TBR2, <3.0 mmol/L), and total daily insulin doses. Results: Overall TIR was 88.2 ± 13.5% in‐flight versus 82.1 ± 17.7% on‐ground. The mean difference in TIR was 4.6%, (95% CI: −5.0‐ 14.2). Other overall metrics (in‐flight versus on‐ground) were: TITR = 67.7 ± 19.4% vs. 64.2 ± 18.1%, TAR = 7.1 ± 10.5% vs. 12.0 ± 13.8% and TBR = 4.7 ± 7.2% vs. 5.9 ± 15.7%. Insulin delivery was slightly increased in‐flight (25.4 ± 17.6 IU vs. 18.7 ± 11.6 IU), driven by bolus dosing (17.8 ± 13.3 IU vs. 11.9 ± 6.6 IU). No episodes of level 2 hypoglycaemia, diabetic ketoacidosis, or device failure occurred. Conclusions: AID systems maintained robust glycaemic control and safety during commercial flights, comparable to ground‐based conditions. These data support the feasibility of in‐flight AIDs use and provide a basis for future regulatory and clinical practice updates. Larger, multicenter studies are needed to generalize findings.

This publication has 18 references indexed in Scilit: