This national study investigated how Chinese medical institutions detect islet autoantibodies and which reagents they use. Between October 15, 2024 and March 15, 2025, investigators distributed questionnaires to core network member units of the National Clinical Research Center for Endocrine and Metabolic Diseases, member units of the Chinese Diabetes Society and Chinese Society of Endocrinology, China Type 1 Diabetes Alliance members, and other relevant institutions. Of 175 questionnaires issued, 127 valid responses were returned, covering all 31 provinces, municipalities, and autonomous regions.
Respondents reported their routine detection platforms and the provenance of reagents. The dominant platform was chemiluminescence immunoassay (CLIA), reported by 76.6% (95 of 124) of institutions that answered that item. Use of domestically produced reagents was high: domestic reagents accounted for 91.7% (121 of 132) of reagents in use among reporting sites.
To assess real-world assay performance, the study selected 10 representative institutions to participate in a blinded proficiency evaluation. Each participating laboratory received 150 coded, blinded standardized samples: 50 samples from patients with newly diagnosed type 1 diabetes and 100 from healthy donors. Participating laboratories ran their routine tests and returned results to the national center for centralized analysis.
The reference or “gold standard” comparator methods were radioligand assay (RLA) or electrochemiluminescence (ECL). The national center calculated sensitivity, specificity, accuracy, concordance with the gold standard, and Spearman correlation coefficients comparing laboratory results to gold standard values.
The 127 valid questionnaires spanned all provincial-level regions in China and represent responses from institutions within the national clinical network and professional societies. The survey gathered data on the detection methods in routine use and whether reagents were domestic or imported. Specific institutional types or further demographic breakdowns of respondents were not detailed beyond geographic coverage in the source.
Survey responses indicated that CLIA is the mainstream detection platform for islet autoantibodies in China at the time of the study. The high proportion of domestic reagents (91.7%) suggests substantial localization of reagent supply for these assays. The source did not report brand-level or manufacturer-specific performance details in the abstract.
Results from the 10 participating laboratories showed variable performance across different islet autoantibodies. Median values reported (with ranges) were:
GADA (glutamic acid decarboxylase autoantibody): sensitivity 76.0% (38.0%–90.0%); specificity 100.0% (98.0%–100.0%); accuracy 92.0% (78.0%–96.7%).
IA-2A (protein tyrosine phosphatase-2 autoantibody): sensitivity 36.0% (8.0%–66.0%); specificity 100.0% (99.0%–100.0%); accuracy 78.7% (69.3%–88.7%).
IAA (insulin autoantibody): sensitivity 24.0% (2.0%–50.0%); specificity 98.0% (95.0%–100.0%); accuracy 72.7% (66.0%–83.3%).
ZnT8A (zinc transporter 8 autoantibody): sensitivity 46.0% (0%–54.0%); specificity 99.0% (98.0%–100.0%); accuracy 80.7% (66.0%–84.7%).
Overall, specificities were high across assays, while sensitivities—particularly for IA-2A and IAA—were substantially lower and more variable. GADA exhibited the highest median sensitivity and accuracy among the four antibodies evaluated.
The study reported median concordance rates between participating laboratories’ results and the gold standard (RLA or ECL), as well as Spearman correlation coefficients (ρ) measuring quantitative agreement:
GADA: median concordance 95.7% (82.7%–100.0%); median ρ = 0.857 (range 0.780–0.905).
IA-2A: median concordance 88.3% (80.7%–96.7%); median ρ = 0.829 (range 0.507–0.871).
IAA: median concordance 87.3% (82.7%–92.0%); median ρ = 0.781 (range 0.601–0.811).
ZnT8A: median concordance 94.4% (81.3%–96.0%); median ρ = 0.805 (range 0.448–0.887).
These results indicate relatively strong concordance and correlation for GADA and ZnT8A on median, with broader variability for IA-2A and IAA across participating sites.
The study concludes that, in China at the time of the survey, CLIA is the mainstream method for detecting islet autoantibodies and that reagents are largely localized. Performance for GADA detection was described as relatively satisfactory based on median sensitivity, specificity, accuracy, concordance, and correlation with gold standard methods. By contrast, the detection performance for IA-2A, ZnT8A, and especially IAA remains suboptimal and requires further improvement.
The abstract does not report specific remedial actions, quality-improvement interventions, or manufacturer-level performance data. Those details, if present, would be found in the full text but were not reported in the abstract source used for this summary.