Predictive Value of Autoantibody Markers in Children with Type 1 Diabetes Mellitus: A Systematic Review and Meta -analysis

Main Article Content

Nur Rochmah, Muhammad Faizi, Yuni Hisbiyah, Farahdina Farahdina, Anang Endaryanto, Soetjipto


Background: The association between autoantibodies and the risk of type 1 diabetes mellitus (T1DM) is well known. However, a quantitative overview of all associated autoantibodies and their effectiveness in diagnosing T1DM is still lacking.

Purpose: To perform a meta-analysis concerning the association between autoantibodies and the risk of T1DM

Methods: Published papers from PubMed, Embase, Cochrane, and Google Scholar were collected and analyzed using a fixed or random-effect model.

Results: Seven papers on 1266 T1DM patients and 982 controls relevant to our study were included in the analysis. Our pooled analysis found that autoantibody expression in children with T1DM was associated with age at diagnosis [mean diff: 4.35 (95% CI: 1.10-7.60) p = 0.009) and the number of autoantibodies detected [OR: 2.13 (95% CI: 1.65-2.76) p < 0.00001]. However, duration of disease [mean diff: 33.36 (95% CI: 8.78-75.50), p = 0.12] and HbA1c [mean diff: 21.63 (95% CI: 5.43-48.48), p = 0.12] did not differ significantly between the single and multiple autoantibody detection groups. We also found that the expression of anti-ZnT8, anti-glutamic acid decarboxylase (GAD), and IAA was associated with a higher risk of T1DM development in children [OR: 35.26 (23.28-53.41), p < 0.00001; OR: 25.59 (18.29-35.79), p < 0.00001; OR: 23.62 (15.79-35.36), p < 0.0000, respectively].

Conclusion: ZnT8 has a better predictive value than other single autoantibodies, but two or more autoantibodies give superior predictive power.

Article Details