Performance of the Flash Glucose Monitoring System during exercise in youth with Type 1 diabetes

Aim: Metabolic changes during exercise may affect the accuracy of glucose sensors impacting on Type 1 diabetes (T1D) management. The present study aimed at assessing the performance of the Flash Glucose Monitoring system (isCGM) during exercise and in free-living condition in youth with T1D. Methods: Seventeen youth (53% male), aged 13.7 \ub1 3.8 years, with T1D for 5.4 \ub1 3.8 years, HbA1c 7.4 \ub1 1.0% (57 \ub1 11 mmol/mol), were enrolled. Paired isCGM, plasma (PG) and capillary (CG) glucose values (total of 136) were collected during an interval exercise (45 min at 55% VO2max load with 20 s sprints at 80% VO2max every 10 min). Paired isCGM and CG (total of 832) were collected during free-living condition. Results: During exercise, isCGM absolute relative difference (ARDs) means/medians were 12.5/9.4% versus PG and 15.4/10.8% versus CG. During rest, ARDs means/medians were 16.6/12.0%. The Consensus Error Grid analysis showed 98.4% of readings during exercise and 97.24% during rest in zones A + B. Percentage of readings meeting the ISO criteria for CG levels <5.55 mmol/L was 62.5% during exercise, 53.4% during rest; for CG levels 655.55 mmol/L was 64.0% during exercise, 60.4% during rest. Conclusions: isCGM demonstrated similar clinical safety and performance during exercise and in everyday life; further studies are needed to confirm its accuracy during exercise