A systematic review and network meta-analysis of interventions to preserve insulin-secreting beta cell function in people newly diagnosed with type 1 diabetes: results from randomised controlled trials of immunomodulatory therapies.

Background: Type 1 diabetes is characterised by the immune-mediated destruction of pancreatic beta cells. We aimed to determine the effectiveness of immunotherapies for preserving residual beta cell function in newly diagnosed (stage 3) type 1 diabetes.

Methods: Searches were carried out in MEDLINE, Embase, Cochrane CENTRAL and trial registries until 31st Jul 2024. RCTs of immunotherapies to preserve beta cells in newly diagnosed type 1 diabetes were included. Data were extracted using a bespoke, piloted extraction sheet. Risk of bias was assessed using Cochrane Risk of Bias Tool 1. A random effects network meta-analysis was undertaken in R. The primary outcome was C-peptide. Interventions were analysed by class.

Results: Sixty trials were included (4597 patients, 32 intervention classes). Forty-one trials of 42 interventions were eligible for network meta-analysis. Eleven interventions demonstrated statistically significantly higher levels of C-peptide than placebo at 12 months, mesenchymal stem cells (autologous and Wharton's jelly-derived cells), azathioprine, interferon-alpha (5000 IU), autologous dendritic cells, anti-TNF golimumab, low-dose ATG, 3 mg 1-course anti-CD3 teplizumab, baricitinib, cyclosporin and 9/11 mg 2-course anti-CD3 teplizumab but with substantial heterogeneity present (I² = 66%). Azathioprine ranked highest (median ranking 3rd); however, rankings demonstrated relatively wide confidence intervals and thus uncertainty in exact rank order of near adjacent therapies. Risk of bias assessment identified poor reporting, particularly in older trials, but few studies demonstrated high risk overall.

Conclusions: Eleven of 42 interventions demonstrated statistically significantly higher C-peptide levels than placebo at 12 months in the network meta-analysis. These results have identified the 11 most promising therapies trialled and help to direct future head-to-head clinical trials to support approvals for interventions to treat those newly diagnosed with type 1 diabetes. However, data for some interventions originated from small studies (mesenchymal stem cell therapies, azathioprine, autologous dendritic cells) and findings should be considered as hypothesis generating and interpreted with caution due to evidence heterogeneity.

Systematic review registration: The protocol for the systematic review was registered on PROSPERO, the international database of prospectively registered systematic reviews (registration: CRD42018107904).