Automated insulin delivery in pregnant women with type 1 diabetes: a systematic review and meta-analysis

Aim The outcomes of automated insulin delivery (AID) systems in pregnant women with type 1 diabetes (T1D) have not been systematically evaluated. This study aims to evaluate the efficacy and safety of AID in pregnancy. Material and methods Literature searches were conducted until July 5, 2024, on Embase, PubMed, Cochrane Library, and ClinicalTrials.gov website. We included clinical trials and observational studies evaluating AID systems in T1D pregnant individuals. Time in the target range (TIR, 3.5-7.8 mmol/L) was the primary outcome. Secondary outcomes included time below range (TBR, < 3.5 mmol/L), time above range (TAR, > 7.8 mmol/L), and maternal and neonatal outcomes. Results Eighteen studies (550 participants) were included. Compared with standard care, AID did not improve 24-h TIR (mean differences [MD] 3.56%, 95% CI - 0.60 to 7.72). However, the overnight TIR increased by 10.05% (95% CI 6.57 to 13.53). The association between AID and decreased TBR (MD - 0.90%, 95% CI - 1.60 to - 0.20) was found, but not with deceased TAR. Only 7 of the 17 studies achieved the goal of a 24-h TIR above 70%. Additionally, the maternal and neonatal outcomes were comparable between AID and standard care, and AID might reduce maternal weight gain (MD - 2.54 kg, 95% CI - 3.96 to - 1.11). Conclusions AID did not exhibit favourable TIR when compared to standard care. However, AID could increase overnight TIR and decrease TBR. Available evidence indicates that employing AID to meet the target of a 24-h TIR above 70% remains challenging.