ABSTRACT Aims Studies evaluating automated insulin delivery (AID) in type 2 diabetes are limited in number and often conducted in small cohorts. We aimed to summarize efficacy and safety data through a systematic review and meta‐analysis. Materials and Methods We searched MEDLINE, PubMed, Web of Science, and CENTRAL databases and performed hand searching until July 1, 2025. We included randomized controlled trials enrolling individuals with type 2 diabetes, evaluating AID against other glucose‐lowering treatments, and reporting 24‐h time spent in the 70–180 mg/dL glucose range (TIR 70–180 mg/dL) as an outcome. Risk of bias was assessed through the RoB2 tool. Results A total of nine studies, accounting for 714 adults with type 2 diabetes, were included. AID was associated with improved TIR 70–180 mg/dL (mean difference MD 18.43%, 95% confidence interval CI: 12.40 to 24.46; low certainty), time above range (TAR; low certainty), TAR > 250 mg/dL, mean glucose, standard deviation (moderate certainty), and HbA1c (moderate certainty) compared to controls, without differences in time below range (TBR; moderate certainty), TBR < 54 mg/dL, and coefficient of variation. Despite similar total daily insulin doses, AID led to a modest gain in body weight (MD 1.58 kg; 95% CI: 0.75 to 2.40) compared to controls. Severe adverse events, including severe hypoglycemia, were rare. No episodes of diabetic ketoacidosis were reported. Conclusions In individuals with type 2 diabetes, AID is associated with short‐term improvements in glycemic control, although the certainty of evidence is low to moderate. Trail Registration: PROSPERO 2025 CRD420251083874 .
Molfetta et al. (2026) studied this question.