A systematic literature review and meta-analysis of real-world evidence on commercially available automated insulin delivery systems in people with type 1 diabetes.

Abstract

Aims: Automated insulin delivery (AID) is part of the standard of care for type 1 diabetes (T1D), but real-world evidence (RWE) comparing AID systems remains limited. A systematic review and meta-analysis was conducted for outcomes across commercially available AID systems in real-world settings.

Materials and methods: PubMed and Embase were searched to March 2025 for RWE studies of commercial AID systems in ambulatory people with T1D. Eligible studies had n ≥ 250 with ≥10 days of continuous glucose monitoring data. Main outcomes were time in range (TIR) and glycated haemoglobin (HbA1c). Random-effects meta-analyses with AID system as moderator were performed, including scenario analyses by age and optimal device settings.

Results: Thirty-six records covering 34 studies with 635 463 users were included. Studies evaluated MiniMed™ 780G (n = 16), Control-IQ™ (n = 9), MiniMed™ 670G (n = 6), and other systems (≤3 studies each). Meta-analysis demonstrated frequently statistically significant between-system differences in TIR, ranging from 60.1% (95% confidence interval [CI] 54.0 to 65.9) for Omnipod® 5 to 73.9% (95% CI 72.3 to 75.5) for MiniMed 780G. For HbA1c, estimates ranged from 6.5% (95% CI 5.4 to 7.7) for Loop to 7.0% (95% CI 6.7 to 7.4) for MiniMed 780G, although differences were not significant. Significant residual heterogeneity was observed. Age was an important effect modifier, with younger users experiencing less favourable outcomes. Optimal device settings improved glycaemic outcomes.

Conclusions: RWE demonstrated differences in glycaemic outcomes, including TIR, between commercially available AID systems. User age influenced outcomes for all systems. Results must be interpreted cautiously given challenges and potential biases with RWE.