Testing methods used to predict disease progression in children with early-stage type 1 diabetes: A systematic review and meta-analysis

Aims: Current guidance on how best to monitor children and young people (CYP) with early-stage type 1 diabetes is evidenced mainly by expert consensus. This systematic review and meta-analysis aims to evaluate the current evidence for tests used to predict disease progression. 

Methods: Data were sourced from PubMed, Cochrane Central, Ovid Embase and Scopus. The association (hazard ratio [HR]) between test positivity and progression to stage 3 type 1 diabetes in CYP aged ≤18 years with ≥2 islet autoantibodies was examined. Data were pooled using random effects models, and the Hartung–Knapp–Sidik–Jonkman (HKSJ) method was used to adjust confidence intervals to account for greater uncertainty. The risk of bias was evaluated using the QUADAS-2 tool (CRD42023393960). 

Results: In this study, 12,923 studies were identified and 285 underwent full-text review. Thirty-four studies (n = 6866 CYP, median age 11.8 years [IQR, 6.6–13.8]) were included. Overall, 2080 (30%) CYP progressed to stage 3 type 1 diabetes over a median follow-up of 5 years (IQR 2–5). The pooled HR for tests that predicted progression were: 1.40 (95% CI 1.07–1.84) for fasting glucose (OGTT), 3.19 (1.75–5.82) for 2-h glucose (OGTT), 6.43 (1.21–34.18) for the M120 above the median value, 3.12 (2.19–4.43) per 1-unit increase in Index 60 and 1.40 (1.17–1.68) per 1.1 mmol/mol increase in HbA1c (C-statistics 0.7–0.8). Evidence for other tests, including CGM, was uncertain. 

Conclusions: The OGTT, its related tests (M120, Index60) and HbA1c predict progression to stage 3 in CYP with early-stage type 1 diabetes. Other tests, including CGM, need more evidence to support their use as predictive tests in this context.