Reducing the need for carbohydrate counting in type 1 diabetes using closed‐loop automated insulin delivery (artificial pancreas) and empagliflozin: A randomized, controlled, non‐inferiority, crossover pilot trial

Abstract

Aim We assessed whether adding empagliflozin to closed‐loop automated insulin delivery could reduce the need for carbohydrate counting in type 1 diabetes (T1D) without worsening glucose control. Materials and methods In an open‐label, crossover, non‐inferiority trial, 30 adult participants with T1D underwent outpatient automated insulin delivery interventions with three random sequences of prandial insulin strategy days: carbohydrate counting, simple meal announcement (no carbohydrate counting), and no meal announcement. During each sequence of prandial insulin strategies, participants were randomly assigned empagliflozin (25mg/day) or not, and crossed over to the comparator. Mean glucose for carbohydrate counting without empagliflozin (control) was compared to no meal announcement with empagliflozin (in the primary non‐inferiority comparison) and simple meal announcement with empagliflozin (in the conditional primary non‐inferiority comparison). Results Participants were aged 40±15 years, had 27±15 years diabetes duration, and HbA1c 7.6±0.7% (59±8 mmol/mol). The system with no meal announcement and empagliflozin was not non‐inferior (and thus reasonably considered inferior) to the control arm (mean glucose 10.0±1.6 vs 8.5±1.5 mmol/L, non‐inferiority p=0.94), while simple meal announcement and empagliflozin was non‐inferior (8.5±1.4 mmol/L, non‐inferiority p=0.003). Use of empagliflozin on the background of automated insulin delivery with carbohydrate counting was associated with lower mean glucose, corresponding to a 14% greater time in target range. While no ketoacidosis was observed, mean fasting ketones levels were higher on empagliflozin (0.22±0.18 vs 0.13±0.11 mmol/L, p<0.001). Conclusions Empagliflozin added to automated insulin delivery has the potential to eliminate the need for carbohydrate counting, improves glycemic control in conjunction with carbohydrate counting but does not allow for the elimination of meal announcement.