Comparison of continuous glucose monitoring with self-monitoring of blood glucose in type 1 diabetes in the changing atmospheric pressures in aviation: a hypobaric flight simulation.

AIM/HYPOTHESIS: Pilots with type 1 diabetes are required to perform capillary glucose monitoring regularly during flights. Continuous glucose monitoring (CGM) may be an effective and more practical alternative. This study aimed to assess the accuracy of CGM systems against self-monitoring of blood glucose (SMBG) during a hypobaric flight simulation.

METHODS

Twelve insulin pump users with type 1 diabetes were studied using two simulation protocols. Protocol A consisted of a ground phase, ascent, a 190 min cruise with ingestion of a liquid meal, descent and then ground. Protocol B consisted of a ground phase, ascent, a 60 min cruise while fasting, descent, a 20 min ground phase, ascent, a second flight of 120 min with ingestion of a meal, followed by descent and ground. Insulin was administered with or before the meal according to the participants' carbohydrate-counting regimen during both protocols. In Protocol A, capillary, interstitial and plasma glucose were measured during flight and at ground, while in Protocol B, glucose and oxygen were measured. Measurements from three CGM brands and two SMBG devices were recorded during the flight simulations. Findings at cabin pressures during flight (550 mmHg) and ground (750 mmHg) were compared. Fasted and postprandial glucose measurements were analysed using Spearman's correlations and mean absolute relative differences (MARDs).

RESULTS

Eleven men and one woman (n=6 men in Protocol A; n=5 men and n=1 woman in Protocol B) were studied. A total of 1533 data points were recorded. During flight vs ground level, Spearman's correlations for CGM system- and SMBG-derived glucose values were very strong in both Protocol A (r=0.96 during flight vs r=0.94 at ground) and Protocol B (r=0.85 during flight vs r=0.69 at ground). The differences in aggregated CGM MARDs during flight vs ground level were minimal across Protocol A (11.85%; 95% CI [9.78, 13.92] vs 9.08%; 95% CI [7.02, 11.14]) and Protocol B (12.01%; 95% CI [3.34, 20.69] vs 12.97%; 95% CI [4.30, 21.65]).

CONCLUSIONS/INTERPRETATION: The performance of CGM systems and SMBG are comparable during flight-associated atmospheric pressure changes. All tested measurement devices for CGM and SMBG were suitable for diabetes-care-based decisions during flight simulation.