Simulation-Based Evaluation of Dose Titration Algorithms for U-500R Insulin by Pump in Subjects with Type 2 Diabetes

The recent EValuating U-500R Infusion Versus Injection in Type 2 Diabetes Mellitus (VIVID) study demonstrated greater glucose lowering effect but higher rates of nocturnal hypoglycaemia with U-500R by continuous subcutaneous infusion (CSII) compared to multiple daily injection (MDI) in people with type 2 diabetes who require high doses of insulin. ¹ The VIVID study used a defined insulin titration algorithm, but it is of interest to know how alternative algorithms might impact the glycaemic outcome or hypoglycaemia rate. In this study, we adopted a mechanistic modeling ² and simulation approach to evaluate alternative parameters in the U-500R insulin titration algorithm for the CSII arm of the VIVID study. Our goal was to determine which parameter(s) in the algorithm could mitigate the nocturnal hypoglycaemia without sacrificing efficacy observed in the trial.

The simulations were conducted in a virtual population to mirror aspects of the VIVID study, with U-500R as the only insulin and metformin as the only oral medication. The parameters of the algorithm evaluated include total daily insulin dose (TDD), basal infusion rate, bolus dose, duration of insulin action (DIA), and glucose targets; the parameters used in the CSII and MDI arms of the VIVID study served as controls. Key outcomes were glycated haemoglobin (HbA1c), fasting plasma glucose, and hypoglycaemia.

In the simulation, all parameters evaluated in the U-500R CSII dose titration algorithms performed similarly or better in lowering HbA1c compared to the MDI control algorithm. Simulation results showed that reducing TDD by 20% to 30% when initiating U-500R by CSII substantially reduced the hypoglycaemia rate while maintaining glycaemic outcomes like the VIVID study (Figure 1). A reduction in overnight basal rate by 20% slowed the decrease in fasting plasma glucose (FPG), but was offset by an increase in the bolus insulin dose and the TDD. Reduction of only the evening bolus dose, without a basal rate reduction, resulted in a similar lowering of HbA1c with up to 23% lower hypoglycaemia rate compared to VIVID study. Increasing the DIA settings of 8, 10, and 12 hours compared to the DIA of 6 hours in VIVID study, resulted in less than 10% reduction in hypoglycaemia risk with CSII compared to the VIVID protocol with minimum effect on HbA1c outcome.

Figure 1.

Model simulated effects of total daily dose (TDD) reduction in the CSII algorithm. Considering the starting dose in the VIVID study as the control (0% reduction), TDD was reduced from 10% to 40% in the model. The MDI group was included for comparison. (A) Change from baseline in fasting plasma glucose. (B) Change from baseline in HbA1c. (C) Change from baseline in TDD. (D) Cumulative hypoglycaemia events per patient. (E) Nocturnal hypoglycaemia event rate (ratio of CSII to MDI).

To summarize, in this simulation study, reducing TDD when switching subjects to CSII had the greatest effect in reducing hypoglycaemia while maintaining the beneficial lowering of HbA1c. A 20% to 30% reduction in TDD at the time of initiation of U-500R as CSII was found to be the most optimal, which is not dissimilar to the recommendation of a 25% dose reduction when transition to CSII from MDI for U-100 insulins. ³ The benefit of switching from traditional basal bolus therapy to U-500R in this patient population has been previously demonstrated. ⁴ The VIVID study showed the additional benefit of U-500R by CSII ¹ and importance of tailoring titration to individual needs. Close monitoring of glucose responses and adjusting insulin doses accordingly are important to optimize the individual benefit/risk evaluation.