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.
Keywords: AID, aviation, CGM, diabetes, HCL, insulin, medicine, pilots, pump, technology
1. INTRODUCTION
Professions that demand physical or cognitive precision, such as piloting an aircraft, are subject to specific regulatory restrictions. 1 , 2 , 3 , 4 , 5 Individuals with insulin‐treated diabetes are permitted to engage in aviation activities, including private and commercial piloting, under stringent guidelines. 6 , 7
Canada led a regulatory shift in 2002 by becoming the first country to allow carefully selected individuals with insulin‐treated diabetes to operate commercial aircraft. 8 In 2010, the UK Civil Aviation Authority convened an expert panel that developed the ARA.MED.330 protocol, implemented in 2012 and later adopted by Ireland and Austria. This protocol is based on capillary blood glucose self‐monitoring. 9 In contrast, the USA and Australia permit insulin‐treated pilots to fly while using continuous glucose monitoring (CGM) systems during flights. 10 , 11
Despite these advances, use of other technologies such as insulin pumps remains debatable. The issue stems not from device malfunction but from fluid‐gas dynamics described by Boyle's Law: as altitude increases and cabin pressure decreases, air bubbles within pump cartridges or tubing can expand, unintentionally delivering insulin. 12 On the ground, however, robust evidence supports the effectiveness of insulin pumps and automated insulin delivery (AID) systems in improving glycaemic control and preventing hypoglycaemia. 13 , 14 This is particularly relevant not only for pilots but also for travellers with diabetes, who must ensure stable glycaemic control. 15
Recent evidence indicates that both automated insulin delivery systems and continuous subcutaneous insulin infusion (CSII) therapy can maintain effective glycaemic control and safety under simulated hypobaric conditions and retrospective observations. However, hypoglycaemia remains a notable risk during prolonged fasting periods. 16 , 17
Despite these advancements, important knowledge gaps remain. First, the safety and performance of AID systems during flight have not been evaluated in real‐world aviation environments, Second, existing studies are underpowered due to small sample sizes. 16 , 17 To address these limitations, we conducted a pilot study comparing four widely used AID systems under commercial mid‐haul flight and ground‐based control conditions, in both fed (standardized meal) and fasting states. This study aims to provide foundational evidence on the feasibility, safety, and glycaemic control of AID systems at flight settings.
2. METHODS
This was a monocentric, within‐subject, comparative pilot trial designed to evaluate the performance and safety of four commercially available automated insulin delivery AID systems under real‐world commercial flight conditions compared with ground‐based conditions. Each participant completed all four experimental conditions: fed (after a standardized meal consisting of convenience noodles and an apple totalling 54 g of carbohydrates) and fasted states during both mid‐haul commercial flights (Vienna, Austria—Reykjavik, Iceland—Vienna) and ground‐based sessions.
The first flight took place on 24 October 2024 (evening), and the second flight occurred around midnight on 27 October 2024. Ground‐phase assessments were subsequently conducted between 5 and 21 November 2024 for all participants. Each participant completed two ground sessions (fasting and fed) scheduled 1–4 days apart, designed to mimic the short interval used during the flight phase. Ground assessments were performed either in the clinic or in participants' homes, according to participant preference, and started at 5:00 pm. For the flight phase, participants fasted for at least 5 h before departure and on the non‐fasting flight received the same standardized meal 1 h after departure. These conditions were replicated during the ground phase. As a within‐subject study, each individual served as their own control; no separate control group was used.
Inclusion criteria were adults aged ≥18 years with a diagnosis of type 1 diabetes mellitus (T1D) for ≥12 months, HbA1c <8.0% (64 mmol/mol), ≥70% time in range (TIR; 3.9–10 mmol/L) during the 2 weeks preceding enrolment, and ≥90% usage of an AID system in the prior 3 months. Eligible participants had been using one of the following systems for at least 6 months: Medtronic MiniMed 780G, Insulet Omnipod® 5 AID System, Tandem Control‐IQ, or CamAPS FX with Ypsomed YpsoPump. Exclusion criteria included recent severe hypoglycaemia or diabetic ketoacidosis (DKA), pregnancy, and any contraindication to air travel. Written informed consent was obtained from all participants. Omnipod 5 and Tandem participants were recruited from a center in Germany, as these devices were not available in the Austrian market.
Participants continued to use their usual AID systems throughout all phases. All AID systems were used with participants' usual clinical settings; no study‐specific changes to pump or algorithm settings were mandated or implemented. Sensor calibration and insulin cartridge replacement followed standardized protocols according to manufactures guidelines. Participants used their usual rapid‐acting insulin analogue within the AID systems; no study‐specific changes to insulin type or formulation were mandated. In total, 12 participants used insulin aspart (NovoRapid), three used insulin lispro (Humalog), four used ultra‐rapid lispro (Lyumjev), and one used fast‐aspart (Fiasp). Cabin pressure during flight conditions was continuously monitored by a study member stationed in the cockpit.
During the study, all participants wore an additional Dexcom G7 which was used as reference to calculate range metrics for the outcomes. The primary outcome of the study was the percentage of time that participants spent within the target glycaemic range TIR of 3.9–10 mmol/L, as measured by continuous glucose monitoring. This metric served as the main indicator of overall glycaemic control across the different flight and ground conditions.
In addition to TIR, several secondary outcomes were evaluated to provide a more detailed assessment of glycaemic control and insulin delivery patterns. These included the percentage of time in tight range (TITR 3.9–7.8 mmol/L), time above range 1 (TAR1 10.1–13.9 mmol/L), and time above range 2 (TAR2 >13.9 mmol/L). Time below range was also stratified into time below range 1 (TBR1 3.0–3.8 mmol/L) and time below range 2 (TBR2 <3.0 mmol/L). Each of these metrics was assessed across multiple phases of air travel (ascent, cruise, and descent) and during both fasted and fed sessions. Total insulin dose was recorded, along with the breakdown into basal and bolus components.
This exploratory pilot study included 20 participants, a number deemed sufficient to evaluate feasibility and generate preliminary performance and safety data. The study did not employ randomization. Allocation concealment was not applicable. Blinding was not feasible due to the nature of the intervention. No formal sample size calculation was performed. Descriptive statistics (mean ± SD, 95%CI) were used to summarize glycaemic metrics and insulin delivery outcomes. An equivalence margin of ±5% was pre‐defined for the primary outcome (TIR); if the 95% confidence interval for the in‐flight versus ground difference fell within this margin, performance was considered equivalent. All analyses were conducted using R.
The study was approved by the Medical University of Graz’ Ethics Committee (Reference 36‐147 ex 23/24) and registered with the Austrian AGES (Nr: 10353599). The study protocol was registered with the German Clinical Trial Register (DRKS‐ID: DRKS00037565).
3. RESULTS
Twenty participants (age: 37.9 ± 12.8 years, HbA1c: 52.3 ± 6.2 mmol/mol, TIR: 74 ± 9.8%, diabetes duration 19.3 ± 10.6 years) were included in the analysis.
On the Vienna–Reykjavík leg, airborne time was 232 min, comprising 22 min of ascent, 189 min of cruise, and 21 min of descent, plus 28 min of ground parking and taxiing (25 min before take‐off and 3 min after landing). On the return Reykjavík–Vienna flight, airborne time was 214 min, with 24 min spent in ascent, 171 min at cruise, and 19 min in descent, and an additional 26 min of parking and taxiing (11 min before take‐off and 15 min after landing). During cruise, cabin pressure stabilized at a cabin altitude of 7072 ft. (≈585 mmHg) on the outbound flight and 6848 ft. (≈590 mmHg) on the inbound flight.
3.1. Primary endpoint
The overall TIR during flights was 88.2 ± 13.5% (95% CI: 82.1–94.2%) compared to 82.1 ± 17.7% (95% CI: 74.4–89.9%) on the ground. The mean difference in TIR between in‐flight and ground phases was 4.6 ± 21.9% (95% CI: −5.0 to 14.2). During fasting conditions, TIR was similarly high: 93.1 ± 8.2% in‐flight (95% CI: 89.4–96.8%) and 92.6 ± 11.3% on the ground (95% CI: 87.5–97.7%), with a mean difference of 0.5 ± 12.5% (95% CI: −5.0 to 6.0). Under non‐fasting conditions, TIR averaged 82.3 ± 23.8% in‐flight (95% CI: 71.4–93.3%) and 75.0 ± 24.1% on‐ground (95% CI: 64.4–85.5%), with a corresponding difference of 5.0 ± 30.5% (95% CI: −8.4 to 18.4).
3.2. Secondary endpoints
3.2.1. Time in tight range (3.9–10.0 mmol/L)
In‐flight TITR averaged 67.7 ± 19.4%, while on‐ground TITR was 64.2 ± 18.1%. In fasting conditions, TITR was 73.9 ± 23.6% in‐flight vs. 79.0 ± 20.4% on‐ground. During non‐fasting conditions, the in‐flight TITR was 59.3 ± 26.8% vs. 52.0 ± 26.3% on‐ground.
3.2.2. Time below range (<3.8 mmol/L)
For the system reported 4.7 ± 7.2 in‐flight and 5.9 ± 15.7% on‐ground.
In fasting flights, TBR1 was 3.9 ± 6.6% versus 2.7 ± 5.0% on‐ground. During non‐fasting conditions, in‐flight TBR1 rose to 5.8 ± 14.2% compared to 5.7 ± 16.2% on‐ground. TBR2 remained rare. In‐flight values were 1.5 ± 5.6%, and ground‐based values were 2.0 ± 8.5%.
3.2.3. Time above range (>10.1 mmol/L)
Hyperglycaemic excursions were generally infrequent. For TAR1, the systems showed a mean of 7.1 ± 10.5% in‐flight and 12.0 ± 13.8% on‐ground. During fasting, TAR1 averaged 3.0 ± 6.9% in‐flight and 4.6 ± 9.4% on‐ground. Under non‐fasting conditions, TAR1 increased to 11.8 ± 21.7% in‐flight and 19.3 ± 23.5% on‐ground. TAR2 remained low throughout, with 0.2 ± 0.7% in‐flight and 2.1 ± 4.8% on‐ground.
A summary of the results can be found in Table 1.
TABLE 1.
In‐flight vs. on‐ground metrics.
| Metric | In‐flight mean ± SD (95% CI) (N) | On‐ground mean ± SD (95% CI) (N) |
|---|---|---|
| Overall metrics (in‐flight vs. on‐ground) | ||
| TIR | 88.2 ± 13.5 (82.1–94.2) (19) | 82.1 ± 17.7 (74.4–89.9) (20) |
| TITR | 67.7 ± 19.4 (58.9–76.4) (19) | 64.2 ± 18.1 (56.2–72.1) (20) |
| TBR1 | 4.7 ± 7.2 (1.5–8.0) (19) | 5.9 ± 15.7 (−1.0 to 12.8) (20) |
| TBR2 | 1.5 ± 5.6 (−1.0 to 4.0) (19) | 2.0 ± 8.5 (−1.8 to 5.7) (20) |
| TAR1 | 7.1 ± 10.5 (2.4–11.8) (19) | 12.0 ± 13.8 (5.9–18.0) (20) |
| TAR2 | 0.2 ± 0.7 (−0.1 to 0.5) (19) | 2.1 ± 4.8 (0.0–4.2) (20) |
| Overall fasting (in‐flight vs. on‐ground conditions) | ||
| TIR | 93.1 ± 8.2 (89.4–96.8) (19) | 92.6 ± 11.3 (87.5–97.7) (19) |
| TITR | 73.9 ± 23.6 (63.3–84.5) (19) | 79.0 ± 20.4 (69.8–88.2) (19) |
| TBR1 | 3.9 ± 6.6 (1.0–6.9) (19) | 2.7 ± 5.0 (0.5–5.0) (19) |
| TBR2 | 0.3 ± 1.4 (−0.3 to 0.9) (19) | 0.0 ± 0.0 (0.0–0.0) (19) |
| TAR1 | 3.0 ± 6.9 (−0.1 to 6.1) (19) | 4.6 ± 9.4 (0.4–8.9) (19) |
| TAR2 | 0.0 ± 0.0 (0.0–0.0) (19) | 0.0 ± 0.0 (0.0–0.0) (19) |
| Overall non‐fasting (in‐flight vs. on‐ground conditions) | ||
| TIR | 82.3 ± 23.8 (71.4–93.3) (18) | 75.0 ± 24.1 (64.4–85.5) (20) |
| TITR | 59.3 ± 26.8 (46.9–71.7) (18) | 52.0 ± 26.3 (40.4–63.5) (20) |
| TBR1 | 5.8 ± 14.2 (−0.8 to 12.4) (18) | 5.7 ± 16.2 (−1.3 to 12.8) (20) |
| TBR2 | 2.8 ± 11.6 (−2.5 to 8.2) (18) | 2.1 ± 8.5 (−1.7 to 5.8) (20) |
| TAR1 | 11.8 ± 21.7 (1.8–21.9) (18) | 19.3 ± 23.5 (9.0–26.6) (20) |
| TAR2 | 0.3 ± 1.4 (−0.3 to 1.0) (18) | 4.1 ± 9.5 (−0.1–8.3) (20) |
3.2.4. Flight phase analysis
The system showed strong glycaemic control during all flight phases:
Ascent: TIR was 92.3 ± 15.8%.
Cruising: TIR was 86.1 ± 14.1%.
Descent: TIR was 92.0 ± 21.0%.
Differences between phases showed modest variability, with a slight drop in TIR during cruising compared to ascent, accompanied with slight increases in both TBR and TAR. A detailed summary of the results for flight phases can be found in Table 2.
TABLE 2.
Flight phase analysis metrics.
| Metric | Ascent Mean ± SD (95% CI) (n = 19) | Cruise Mean ± SD (95% CI) (n = 19) | Descent Mean ± SD (95% CI) (n = 19) |
|---|---|---|---|
| TIR | 92.3 ± 15.8 (85.2–99.4) | 86.1 ± 14.1 (79.7–92.4) | 92.0 ± 21.0 (82.5–101.4) |
| TITR | 70.4 ± 31.4 (56.3–84.4) | 65.4 ± 19.5 (56.6–74.1) | 75.3 ± 28.2 (62.6–88.0) |
| TBR1 | 2.2 ± 7.2 (−1.1–5.4) | 5.6 ± 7.8 (2.1–9.1) | 4.4 ± 13.4 (−1.7–10.4) |
| TBR2 | 0.5 ± 2.3 (−0.5–1.6) | 1.5 ± 6.1 (−1.3–4.2) | 0.6 ± 2.5 (−0.6–1.7) |
| TAR1 | 5.5 ± 14.9 (−1.2–12.2) | 8.3 ± 13.4 (2.3–14.3) | 3.7 ± 11.4 (−1.5–8.8) |
| TAR2 | 0.0 ± 0.0 (0.0–0.0) | 0.3 ± 1.3 (−0.3–0.9) | 0.0 ± 0.0 (0.0–0.0) |
3.2.5. Sensitivity analysis
A post‐hoc linear mixed‐effects model with a random intercept for participant, adjusting for prandial state and interaction between phase and prandial state was used to estimate mean (95% C.I.) TIR for flight vs. ground. Overall, TIR was 86.62% (95% CI: 81.67–91.57%) in flight and 83.52% (95% CI: 78.58–88.45%) on ground, corresponding to a difference of 3.10% (95% CI: 1.27–4.93%). During fasting conditions, no meaningful difference between flight and ground was found, 91.74% (95% CI: 86.63–96.85%) vs. 92.12% (95% CI: 87.01–97.23%) respectively. In contrast, during non‐fasting conditions, in‐flight TIR was 81.49% (95% CI: 76.36–86.62) while TIR on ground was 74.91 (95% CI: 69.83–79.99%). The intraclass correlation coefficient was 0.10, indicating that approximately 10% of the variance in TIR was attributable to between‐participant differences rather than within‐participant changes across conditions. Model based estimations were in line with the original descriptive analysis. Detailed results can be found in Table 3.
TABLE 3.
Random‐intercept model (participant): TIR (%) by flight phase and prandial state.
| Null model | Model 1 (Main effects): Phase + Prandial State | Model 2 (Interaction term): Main effects + phase * prandial state | ||||
|---|---|---|---|---|---|---|
| Fixed Effects | unstd. Estimates (%) | 95% CI | unstd. Estimates (%) | 95% CI | unstd. Estimates (%) | 95% CI |
| Intercept | 84.89 | (79.83–89.94) | 93.46 | (88.43–98.50) | 91.74 | (86.63–96.85) |
| Phase | Reference: flight | Reference: flight | Reference: flight | |||
| Ground | −3.11 | (−4.94 – −1.27) | 0.38 | (−2.20–2.96) | ||
| Prandial state | Reference: fasting | Reference: fasting | Reference: fasting | |||
| Non‐fasting | −13.81 | (−15.63 – −12.00) | −10.25 | (−12.84 – −7.66) | ||
| Interaction Phase * Prandial state | −6.96 | (−10.58 – −3.34) | ||||
| Random effects | ||||||
| ICC | 0.10 | 0.10 | 0.10 | |||
| N | 20sid | 20sid | 20sid | |||
| Observations | 5077 | 5077 | 5077 | |||
| Marginal R 2/Conditional R 2 | 0.000/0.102 | 0.041/0.136 | 0.043/0.138 | |||
3.2.6. Insulin delivery
The mean total insulin dose administered during in‐flight conditions was 25.4 ± 17.6 IU (95% CI: 17.2–33.5, N = 18), compared to 18.7 ± 11.6 IU (95% CI: 13.3–24.1, N = 18) on‐ground. This increase was largely driven by bolus insulin, which averaged 17.8 ± 13.3 IU (95% CI: 11.7–24.0, N = 17) in‐flight versus 11.9 ± 6.6 IU (95% CI: 8.8–14.9, N = 17) on‐ground. Basal insulin delivery was similar across conditions: 11.0 ± 4.8 IU (95% CI: 8.8–13.2, N = 14) in‐flight and 10.4 ± 8.2 IU on‐ground (95% CI: 6.6–14.1, N = 13).
3.2.7. Hypoglycaemic and hyperglycaemic events
Capillary glucose data showed no episodes of level 2 hypoglycaemia (SMBG <3.0 mmol/L) during either flight or ground phases. Across the entire study, 28 mild hypoglycaemic events (3.0–3.9 mmol/L) were documented, of which 15 (6 during flight, 9 during ground settings) required dextrose treatment.
Sensor‐detected hypoglycaemia (SDH) episodes revealed modestly higher per‐person events in‐flight, 13 vs. 8. Mean episodes per participant in‐flight versus on‐ground were 0.68 vs. 0.40. There were no SDH episodes under 3.0 mmol/L, but only individual or few consecutive values. Overall, episodes under 3.9 mmol/L lasted 30.7 min during flight vs. 41.1 min on ground.
Hyperglycaemia (SMBG >13.9 mmol/L) was rare. One such episode was recorded during flight, and three on‐ground.
3.2.8. Safety and device function
No cases of diabetic ketoacidosis, pump failure, or CGM failure occurred during any study phase.
4. DISCUSSION
This study is the first to systematically evaluate the real‐world performance and safety of four widely used automated insulin delivery systems during commercial air travel, directly comparing glycaemic outcomes in‐flight with ground‐based conditions within the same cohort of people living with T1D. The findings provide novel, clinically relevant evidence on the feasibility of AID‐supported glycaemic management in the unique physiological and operational environment of commercial aviation. Overall, the data demonstrates that glycaemic control, measured by percentage time in range, remained high during all phases of commercial flight, with no evidence of deterioration compared to ground‐based control.
The observed mean difference in TIR between in‐flight and on‐ground conditions was 4.6%, with absolute values well above ranges recommended by international consensus for people with T1D. 18 , 19 Although the confidence interval narrowly exceeded the pre‐defined equivalence margin of ±5%, the direction and magnitude suggest no clinically relevant impairment or reduced efficacy in‐flight performance. Importantly, fasting and non‐fasting conditions were both examined, capturing a range of metabolic states relevant to aviation medicine. During fasting, TIR was exceptionally high and virtually identical in‐flight and on‐ground (93.1% vs. 92.6%). In non‐fasting conditions, a slightly larger in‐flight advantage was observed (82.3% vs. 75.0%), though with greater inter‐individual variability as reflected by higher standard deviations (Table 1). Assessment of secondary glycaemic endpoints demonstrated again the robustness of AID performance at altitude. Total insulin delivery was higher in‐flight, largely due to increased bolus dosing, but this did not translate into increased risk of hypoglycaemic excursions. It is worth noting that ascent and descent were much shorter than the cruise phase; phase‐specific results for ascent and descent should therefore be interpreted with caution.
The study's strengths include its within‐subject design, real‐world operational setting, and comprehensive evaluation of both primary and secondary glycaemic outcomes. Continuous monitoring of cabin pressure enhanced the reliability of findings.
Several limitations must be acknowledged. First, this was a single‐centre, non‐randomized, pilot study with a modest sample size, limiting statistical power and precluding robust subgroup analysis or generalization to all individuals with diabetes. One insulin study dataset was not available due to malfunctioning of the readout device for own pump and due to connectivity and synchronization issues on the smartphones there were missing data from CGM reported in the sample size number across analysis. Only participants with established AID use and good glycaemic control were included. We did not systematically document individual glucose target set‐points for each AID system, so we cannot account for between‐participant variability in target settings, which may influence insulin delivery patterns and inter‐system differences. Finally, flight duration, route, and conditions were standardized, which may limit generalizability to other flight scenarios.
Despite these limitations, the findings support the feasibility and safety of in‐flight use of AID systems. When it comes to the pilots population, they can prepare for a flight, so glucose excursions can be limited. It should also be noted that pilots with diabetes are typically well‐trained in the use of their devices and highly motivated to maintain glucose within target ranges, as failure to do so could result in being declared unfit to fly. Importantly, this data provides a foundation for further regulatory discussions on updating protocols and regulations.
FUNDING INFORMATION
This study was supported by grants from ÖDG ‐ Österreichische Diabetesgesellschaft on behalf of Daniel A Hochfellner and Horizon Europe Research and Innovation Program on behalf of EASA ‐ European Union Aviation Safety Association. Insulet supported the travel of Omnipod 5 users.
CONFLICT OF INTEREST STATEMENT
Thomas R. Pieber has received research support (paid to the University) from Novo Nordisk, AstraZeneca and Eli Lilly and has served as board member or in an advisory panel for Adocia, Arecor, AstraZeneca, Eli Lilly, Novo Nordisk, Roche, and Sanofi. Insulet supported the travel of Omnipod 5 users. Chantal Mathieu serves or has served on the advisory panel for Novo Nordisk, Sanofi, Eli Lilly and Company, Novartis, Dexcom, Boehringer Ingelheim, Bayer, Roche, Abbott, Medtronic, Insulet, Biomea Fusion, SAB Bio and Vertex. Financial compensation for these activities has been received by KU Leuven; KU Leuven has received research support for CM from Medtronic, Novo Nordisk and Sanofi; CM serves or has served on the speakers bureau for Novo Nordisk, Sanofi, Eli Lilly and Company, Medtronic, Dexcom, Insulet, Abbott, Vertex and Boehringer Ingelheim. Financial compensation for these activities has been received by KU Leuven. Chantal Mathieu is president of EASD. All external support of EASD is to be found on www.easd.org. Gerd Koehler is a member of advisory boards of Eli Lilly, Novo Nordisk A/S, Roche Diabetes Care, Viatris and received speaker honoraria from Daiichi Sankyo, Dexcom, Eli Lilly, Novo Nordisk A/S, Roche Diabetes Care, Sanofi, Viatris. Julia K. Mader is a member of advisory boards of Abbott Diabetes Care, Becton‐Dickinson, Biomea Fusion, Dexcom, Eli Lilly, Embecta, Medtronic, myLife, Novo Nordisk A/S, Pharmasens, Roche Diabetes Care, Sanofi‐Aventis, Tandem, Viatris and received speaker honoraria from A. Menarini Diagnostics, Abbott Diabetes Care, Dexcom, Eli Lilly, Medtrust, MSD, Novo Nordisk A/S, Roche Diabetes Care, Sanofi, Viatris and Ypsomed. She is a shareholder of decide Clinical Software GmbH and elyte Diagnostics and serves as CMO of elyte Diagnostics. All other authors declare no conflicts of interest. Manufacturers of the devices evaluated in this study were not involved in the study design, data collection, access to study data, statistical analysis, interpretation of results, or review of the manuscript.
ACKNOWLEDGEMENTS
We thank Austrian Airlines and the Austrian Airlines crews of OS327 on October 24, 2024 and OS328 on October 27, 2024 for facilitating our experiments on their flights. We also thank Insulet for providing travel support for study participants. Open Access funding provided by Medizinische Universitat Graz/KEMÖ.
Meçani R, Basta S, Baumann PM, et al. In‐flight glycaemic control using automated insulin delivery systems: A within‐subject comparative pilot study. Diabetes Obes Metab. 2026;28(3):2159‐2165. doi: 10.1111/dom.70404
DATA AVAILABILITY STATEMENT
Data available on request from the authors.
REFERENCES
- 1. Gentile S, Furia A, Strollo F. Aircraft pilot licence and diabetes. Diabetes Res Clin Pract. 2020;161:108047. doi: 10.1016/j.diabres.2020.108047 [DOI] [PubMed] [Google Scholar]
- 2. Blaho‐Owens K. Aviation Medicine: Illness and Limitations for Flying. In: Encyclopedia of Forensic and Legal Medicine. 2nd ed. Elsevier; 2016:394‐400. [Google Scholar]
- 3. Kohrman DB. Driving with diabetes: precaution, not prohibition, is the proper approach. J Diabetes Sci Technol. 2013;7(2):350‐355. doi: 10.1177/193229681300700210 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Raveendran AV. Impact of diabetes on driving: the current needs and updates. Int J Diabetes Technol. 2024;3(3):91‐99. doi: 10.4103/ijdt.ijdt_26_24 [DOI] [Google Scholar]
- 5. American Diabetes Association . Diabetes and employment. Diabetes Care. 2009;32(Suppl 1):S80‐S84. doi: 10.2337/dc09-S080 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Russell‐Jones DL, Hutchison EJ, Roberts GA. Pilots flying with insulin‐treated diabetes. Diabetes Obes Metab. 2021;23(7):1439‐1444. doi: 10.1111/dom.14375 [DOI] [PubMed] [Google Scholar]
- 7. Stanwyck LK, DeVoll JR, Pastore J, Gamble Z, Poe A, Gui GV. Medical certification of pilots through the insulin‐treated diabetes mellitus protocol at the FAA. Aerosp Med Hum Perform. 2022;93(8):627‐632. doi: 10.3357/AMHP.6107.2022 [DOI] [PubMed] [Google Scholar]
- 8. Wientjens W, Cairns D. Fighting discrimination. Diabetes Res Clin Pract. 2012;98(1):33‐37. doi: 10.1016/j.diabres.2012.05.021 [DOI] [PubMed] [Google Scholar]
- 9. Garden GL, Hine JL, Mitchell SJ, et al. An evaluation of the safety of pilots with insulin‐treated diabetes in Europe flying commercial and noncommercial aircraft. Diabetes Care. 2020;43(12):2923‐2929. doi: 10.2337/dc20-0277 [DOI] [PubMed] [Google Scholar]
- 10. Strollo F, Simons R, Mambro A, Strollo G, Gentile S. Continuous glucose monitoring for in‐flight measurement of glucose levels of insulin‐treated pilots. Aerosp Med Hum Perform. 2019;90(8):735‐737. doi: 10.3357/AMHP.5315.2019 [DOI] [PubMed] [Google Scholar]
- 11. Authority CAS . Protocol for pilots with type 1 diabetes. 2021. https://www.casa.gov.au/licences‐and‐certificates/medical‐professionals/protocol‐pilots‐type‐1‐diabetes
- 12. King BR, Goss PW, Paterson MA, Crock PA, Anderson DG. Changes in altitude cause unintended insulin delivery from insulin pumps. Diabetes Care. 2011;34(9):1932‐1933. doi: 10.2337/dc11-0139 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Bassi M, Franzone D, Dufour F, et al. Automated insulin delivery (AID) systems: use and efficacy in children and adults with type 1 diabetes and other forms of diabetes in Europe in early 2023. Life. 2023;13(3):783. doi: 10.3390/life13030783 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Fan W, Deng C, Xu R, et al. Efficacy and safety of automated insulin delivery systems in patients with type 1 diabetes mellitus: a systematic review and meta‐analysis. Diabetes Metab J. 2024;49(2):235‐251. doi: 10.4093/dmj.2024.0130 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Uğurlu M. Flight and diabetes. Anatol J Cardiol. 2021;25(Suppl 1):S20‐S23. doi: 10.5152/AnatolJCardiol.2021.S108 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Garden GL, Fan KS, Paterson M, et al. Effects of atmospheric pressure change during flight on insulin pump delivery and glycaemic control of pilots with insulin‐treated diabetes: an in vitro simulation and a retrospective observational real‐world study. Diabetologia. 2025;68(1):52‐68. doi: 10.1007/s00125-024-06295-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Fan KS, Shojaee‐Moradie F, Manoli A, et al. The feasibility of an experimental hypobaric simulation to evaluate the safety of closed‐loop insulin delivery systems in flight‐related atmospheric pressure changes. Diabetes Technol Ther. 2025;27(2):128‐133. doi:10.1089/dia.2024.0380. [DOI] [PubMed] [Google Scholar]
- 18. Battelino T, Danne T, Bergenstal RM, et al. Clinical targets for continuous glucose monitoring data interpretation: recommendations from the international consensus on time in range. Diabetes Care. 2019;42(8):1593‐1603. doi: 10.2337/dci19-0028 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Holt RIG, DeVries JH, Hess‐Fischl A, et al. The management of type 1 diabetes in adults. A consensus report by the American Diabetes Association (ADA) and the European Association for the Study of diabetes (EASD). Diabetologia. 2021;64(12):2609‐2652. doi: 10.1007/s00125-021-05568-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data available on request from the authors.