Abstract
INTRODUCTION: Real-time continuous glucose monitoring (rtCGM) has become an integral component of modern diabetes management. However, sensor performance data of adolescents, especially during exercise, are scarce. Therefore, this investigation aimed to determine the performance of a rtCGM-system before, during, and after exercise around a 4-week track and field training intervention in adolescents with type 1 diabetes (T1D). MATERIAL AND METHODS: Eight adolescents (2 females, 15.5 ±1.0 years, BMI: 21.0 ±2.3 kg/m2) with T1D and an HbA1c of 7.0 ±0.6% (53.0 ±6.6 mmol/mol) participated in a 4-week track and field intervention consisting of three 60-minute exercise sessions per week. Sensor glucose levels (Dexcom G6; Dexcom Inc., San Diego, CA) were obtained at rest and after 15, 30, 45, and 60 minutes during, as well as 5 minutes after exercise and compared to a reference capillary blood glucose value (Biosen S-Line, EKF Diagnostics, GER). rtCGM performance was assessed using the median absolute relative difference (MedARD) and interquartile range [IQR] as well as the Bland-Altman method. RESULTS: 443 points of comparison were available for analysis. The overall MedARD (IQR) was 19.1% (12.2-27.3). MedARDs for time in range (70-180 mg/dl), below range ( 180 mg/dl) were 22.0% (17.8-29.2), 28.4% (18.8-36.3), and 15.9% (9.4-20.7), respectively. The in-exercise MedARD was 19.4% (12.3-27.8), and pre- and post-exercise MedARDs were 21.6% (14.5-28.3) and 14.9% (9.0-21.6), respectively. CONCLUSIONS: In structured track and field training involving adolescents with T1D, the accuracy of the Dexcom G6 system is limited. This supports the importance of considering glucose trend arrows and, when in doubt, blood measurements during exercise.