An open-label, multi-centre, randomised controlled trial to assess the efficacy, safety and utility of automated insulin delivery (AID) in people with type 2 diabetes and sub-optimal glycaemia undergoing haemodialysis

Research question: Is it possible to improve glucose levels, patient safety, treatment satisfaction and quality of life using a commercially available automated Insulin Delivery system (AID) in adults with type 2 diabetes treated with insulin and undergoing regular haemodialysis treatment for end-stage kidney disease (ESKD)? Background: Diabetes is the leading cause of ESKD in the UK and disproportionately affects people of minority ethnic backgrounds. People with diabetes on haemodialysis have fluctuating glucose levels, which are challenging to manage. By automating insulin delivery to respond to glucose levels, AID may improve clinical care and patient well-being. However, the benefits and limitations of AID have not been established in this high-risk group. Aims and Objectives: This trial evaluates the efficacy, safety and acceptance of the commercially available Omnipod 5 AID system (Insulet Corporation, USA) compared to usual care with continuous glucose monitoring (CGM) in people with type 2 diabetes on insulin undergoing regular haemodialysis. Methods: This will be an open-label, multi-centre, randomised, parallel-design superiority study comparing AID with usual care with CGM. AID includes automated algorithm-driven insulin delivery by an insulin pump. Usual care includes usual insulin injections combined with CGM. We will recruit 84 participants. We based the sample size on observing an expected difference 'time-in-range' between groups of 10 percentage points and SD of 14 percentage points. Allowing for a drop-out rate of up to 10% pre-randomisation, 76 participants will be randomised, with the aim of 64 completed participants (15% post-randomisation dropout). Participants will provide written informed consent. The study will consist of a 4-week baseline run-in and a 12-week randomised treatment period. Participants will have 9 study visits. The research team will provide training and support participants in both arms with optimal AID, or CGM-supported insulin therapy throughout the study. The primary efficacy endpoint is glucose time-in-range target of 3.9 to 10 mmol/l from CGM. Safety will be assessed using the time spent in hypoglycaemia (CGM glucose levels <3.9 and 3.0 mmol/l), severe and non-severe hypoglycaemia episodes, and other adverse events. Treatment satisfaction and quality of life will be explored using semi-structured interviews and questionnaires. The primary endpoint, percent time-in-range (%) will be analysed using Analysis of covariance methods (ANCOVA), adjusting for baseline time-in-range and stratification factors. The study will be supported by the Diabetes Trials Unit, University of Oxford, which will be responsible in conjunction with the study team for ethics and regulatory submission, trial oversight and management, database creation, statistical analysis and study monitoring. Timelines for delivery: We anticipate ethics and regulatory submission by April 2025, first-participant-first-visit by July 2025, end of recruitment by Dec 2025, last-participant-last-visit by April 2026, data and statistical analysis by July 26, and submission of results by Sep 2026. Anticipated Impact and Dissemination: This trial will provide much-needed evidence for using AID in this high-risk population, and may lead to adoption of this treatment in the NHS and beyond. We will register the trial, track and measure outputs, outcomes and impacts, and disseminate results through peer-reviewed scientific journals & conference presentations.