Tubeless automated insulin delivery reduces HbA1c more than multiple daily injections in type 1 diabetes

This multicenter, international randomized controlled trial evaluated the efficacy of a tubeless automated insulin delivery (AID) system compared to standard multiple daily injections (MDI) in individuals with type 1 diabetes. The study was conducted across 19 hospitals in the United Kingdom, Belgium, and France. Participants included 188 children and adults aged 4 to 70 years with type 1 diabetes who were already using both multiple daily injections and continuous glucose monitoring. Eligible participants had HbA1c levels between 7.5% and 11% (58-97 mmol/mol) at baseline. Participants were randomly assigned in a ratio favoring the intervention group, with 125 allocated to the AID group and 63 to the control MDI group. The study followed participants for 13 weeks.

The intervention group used a tubeless automated insulin delivery system, which integrates a continuous glucose monitor with an insulin pump that automatically adjusts basal insulin delivery. The comparator group continued their usual care with multiple daily injections. Both groups continued using continuous glucose monitoring throughout the study. Specific dosing protocols and insulin types were not reported in the provided data.

The primary outcome was the adjusted between-group difference in HbA1c at 13 weeks, assessed for superiority. The AID group showed a reduction in HbA1c from a baseline of 8.1% (SD 0.7) to 7.2% (SD 0.6) at 13 weeks. The control MDI group showed minimal change, from a baseline of 8.1% (SD 0.6) to 8.0% (SD 0.7) at 13 weeks. The adjusted mean difference in HbA1c reduction favored the AID group by -0.8% (95% CI -1.0 to -0.6; p<0.0001). Absolute numbers for the primary outcome were not reported.

No specific secondary outcomes were detailed in the provided evidence. The main results focused exclusively on the primary HbA1c outcome, with no reported data on time in range, hypoglycemic events beyond the absence of severe episodes, quality of life measures, or other glycemic variability metrics.

Regarding safety and tolerability, adverse events were more frequently reported in the AID group. There were 39 adverse events among 28 participants in the AID group, compared to 3 adverse events among 3 participants in the control group. The nature of these adverse events was not specified. Two serious adverse events occurred in the AID group: one case of Kawasaki disease and one case of acute coronary syndrome. Both were judged by investigators to be unrelated to the study device or procedure. The number of serious adverse events in the control group was not reported. Importantly, no episodes of severe hypoglycemia or diabetic ketoacidosis occurred in either group during the 13-week study. Data on discontinuations and specific tolerability metrics were not reported.

These results align with the broader evidence base demonstrating that automated insulin delivery systems generally improve glycemic control compared to conventional therapy in type 1 diabetes. The magnitude of HbA1c reduction (0.8%) is clinically meaningful and consistent with effects seen in other trials of similar duration for different AID systems. The absence of severe hypoglycemia or diabetic ketoacidosis in this short-term study is reassuring and consistent with the safety profile established by prior landmark studies, which have generally shown AID systems to be safe and effective.

Key methodological limitations must be considered. The study had an open-label design, which introduces potential for performance and detection bias, as participants and clinicians were aware of treatment assignment. The 13-week follow-up period is short for a chronic condition like type 1 diabetes, limiting understanding of long-term efficacy, safety, and adherence. The study was funded by Insulet Corporation, the manufacturer of the device, which represents a potential conflict of interest. The random assignment ratio (approximately 2:1) and the lack of reported data on several important outcomes, including participant discontinuations and the specific nature of most adverse events, are additional limitations.

The clinical implications of this study are that, in the short term, a tubeless AID system can provide superior glycemic control compared to multiple daily injections in children and adults with type 1 diabetes who have suboptimal HbA1c levels. For clinicians, this adds to the options for advanced diabetes technology that may benefit appropriate patients. However, decisions should consider the open-label nature of the evidence, the need for patient training and comfort with technology, the higher rate of reported adverse events with AID, and the device costs. This evidence supports considering AID as a tool for improving glycemic control in the studied population.

Several important questions remain unanswered. The long-term durability of the glycemic benefit beyond 13 weeks is unknown. The impact on patient-reported outcomes like quality of life, treatment satisfaction, and burden was not reported. The study does not clarify which patient subgroups (e.g., by age, baseline HbA1c, or prior insulin regimen) might benefit most or least from this intervention. The reasons for the higher number of adverse events in the AID group require clarification. Furthermore, comparative effectiveness against other AID systems (both tubed and tubeless) and the cost-effectiveness of this approach are important areas for future research.