Authors
Ruining Yang, Xiyi Lu, Yixian Liang, Runzhi Liu, Jianxin Rui, Lixun Xu, Nan Chen
Published in
Diabetes, obesity & metabolism. Jul 13, 2026. Epub Jul 13, 2026.
Abstract
To compare insulin delivery modalities for glycaemic control, body weight and insulin dose in adults with insulin-treated Type 2 diabetes.
We searched eight databases and two trial registries until 27 April 2026 for parallel-group randomised trials comparing pre-specified insulin delivery modalities. The primary network meta-analysis excluded the clinically heterogeneous residual Other node. Outcomes were changes in HbA1c, time in range (TIR), time above range (TAR), time below range (TBR), body weight and insulin dose; certainty was assessed with CINeMA.
Thirty-seven reports describing 36 independent trials with 4157 participants were included. Median follow-up was 12 weeks. In the primary no-Other analysis, hybrid closed-loop (HCL) reduced HbA1c versus multiple daily injections (MDI; MD -0.79%, 95% CI -1.33 to -0.25), increased TIR (MD +16.00 percentage points, 95% CI +8.77 to +23.22), reduced TAR (MD -13.48 percentage points, 95% CI -23.05 to -3.91) and reduced insulin dose (MD -15.69 U/day, 95% CI -31.33 to -0.06). Continuous subcutaneous insulin infusion also reduced HbA1c (MD -0.47%, 95% CI -0.74 to -0.20) and insulin dose. Fully closed-loop yielded larger TIR/TAR estimates, but from one short trial only and should be considered hypothesis-generating. HCL increased body weight (MD +1.58 kg, 95% CI +0.75 to +2.40). Most trials had risk-of-bias concerns and low-risk sensitivity analysis was infeasible.
HCL showed the most consistent glycaemic benefit, but body-weight and sparse safety evidence warrant caution. FCL findings require larger, longer confirmation.
PROSPERO registration: CRD420261381231.
PMID:
42443590
Bibliographic data and abstract were imported from PubMed on 14 Jul 2026.
Read full publication at:
Please sign in
to see all details.
Advertisement
Stats
- Recommendations n/a n/a positive of 0 vote(s)
- Views 4
- Comments 0