Switching from Flash Glucose Monitoring to Continuous Glucose Monitoring on Hypoglycemia in Adults with Type 1 Diabetes at High Hypoglycemia Risk: The Extension Phase of the I HART CGM Study

摘要

Background: The I HART CGM study showed that real-time continuous glucose monitoring (RT-CGM) has greater beneficial impact on hypoglycemia than intermittent flash glucose monitoring (flash) in adults with type 1 diabetes (T1D) at high risk. The impact of continuing RT-CGM or switching from flash to RT-CGM for another 8 weeks was then evaluated. Methods: Prospective randomized parallel group study with an extension phase. After a 2-week run-in with blinded CGM, participants were randomized to either RT-CGM or flash for 8 weeks. All participants were then given the option to continue with RT-CGM for another 8 weeks. Glycemic outcomes at 8 weeks are compared with the 16-week endpoint. Results: Forty adults with T1D on intensified multiple daily insulin injections and with impaired awareness of hypoglycemia or a recent episode of severe hypoglycemia were included (40% female, median [IQR] age 49.5 [37.5-63.5] years, diabetes duration 30.0 [21.0-36.5] years, HbA1c 56 [48-63] mmol/mol, and Gold Score 5 [4-5]), of whom 36 completed the final 16-week extension. There was a significant reduction in percentage time in hypoglycemia (3.0mmol/L) in the group switching from flash to RT-CGM (from 5.0 [3.7-8.6]% to 0.8 [0.4-1.9]%, P=0.0001), whereas no change was observed in the RT-CGM group continuing with the additional 8 weeks of RT-CGM (1.3 [0.4-2.8] vs. 1.3 [0.8-2.5], P=0.82). Time in target (3.9-10mmol/L) increased in the flash group after switching to RT-CGM (60.0 [54.5-67.8] vs. 67.4 [56.3-72.4], P=0.02) and remained the same in the RT-CGM group that continued with RT-CGM (65.9 [54.1-74.8] vs. 64.9 [49.2-73.9], P=0.64). Conclusions: Our data suggest that switching from flash to RT-CGM has a significant beneficial impact on hypoglycemia outcomes and that continued use of RT-CGM maintains hypoglycemia risk benefit in this high-risk population.