Background: The objective of this work is to document performance of automated insulin delivery (AID) during real-life use in type 2 diabetes (T2D).
Methods: A retrospective analysis was performed of continuous glucose monitoring and insulin delivery data from 796 individuals with T2D, who transitioned from 1-month predictive low-glucose suspend (PLGS) use to 3-month AID use, in real-life settings. Primary outcome was change of time in range (TIR = 70-180 mg/dL) from PLGS to AID. Secondary outcomes included time above/below range (TAR/TBR) and total daily insulin (TDI).
Results: Compared with PLGS, AID increased TIR on average from 63.2% to 72.6%, decreased TAR from 36.2% to 26.8%, and increased TDI from 70.2 to 76.3 U (all P < .001), without significant change to TBR. Glycemic improvements were more pronounced in those with worse glycemic control during PLGS use (P < .001).
Conclusions: Real-life use of AID led to a rapid and sustained improvement of glycemic control in individuals with T2D.
Keywords: automated insulin delivery; continuous glucose monitoring; insulin pump; real life; time in range; type 2 diabetes.