Dapagliflozin in patients with type 1 diabetes: A post hoc analysis of the effect of insulin dose adjustments on 24-hour continuously monitored mean glucose and fasting β-hydroxybutyrate levels in a phase IIa pilot study

Aims: To investigate the effects of total daily insulin dose (TDD) reductions on 24-hour continuously monitored mean glucose and fasting beta-hydroxybutyrate (a marker for diabetic ketosis/ketoacidosis [DKA]) levels, using patient-level data from a 14-day, pilot study of dapagliflozin in type 1 diabetes (T1DM). Methods: A post hoc exploratory correlation analysis was performed to determine the relationship between change in TDD and (1) 24-hour mean glucose, assessed by continuous glucose monitoring, and (2) fasting beta-hydroxybutyrate, in 70 patients with T1DM receiving insulin and dapagliflozin (1, 2.5, 5 or 10 mg) or placebo. The pharmacodynamic effect of dapagliflozin was estimated as a virtual insulin dose using 24-hour urinary glucose excretion values and a recognized insulin-to-carbohydrate counting technique. Results: Trends for correlations were observed between change in TDD and 24-hour glucose (day 7: r = -0.264, P = .056) and beta-hydroxybutyrate (day 7: r = -0.187, P = .133; day 14: r = -0.274, P = .047). The pharmacodynamic effect of dapagliflozin 5 or 10 mg was estimated as equivalent to similar to 20% of baseline TDD. Higher mean and maximum beta-hydroxybutyrate levels were observed on days 7 and 14 in patients with a TDD reduction > 20% vs <= 20%. Conclusions: Over 14 days, decreasing the insulin dose diminished the glucose-lowering effect of dapagliflozin-insulin combination therapy and increased levels of beta-hydroxybutyrate. While insulin dose adjustments should always be individualized, these analyses suggest that, as a general rule, TDD reduction in dapagliflozin-treated patients with T1DM should not exceed 20%, to ensure glycaemic control does not deteriorate and to mitigate the potential for an increased risk of DKA.