Predictive Value of GAD Antibody for Diabetes in Normal Chinese Adults: A Retrospective Cohort Study in China

PURPOSE: To investigate the prevalence of GAD antibody (GADA) in the general adult population and to evaluate its predictive value for diabetes in China. PATIENTS AND METHODS: We searched the PUMCH-HM database and identified 36,731 adult subjects with GADA test results from 2012 to 2015. We then established a retrospective cohort of 4835 nondiabetic subjects at baseline with complete annual health evaluation records through 2019. The median follow-up time was 4.8 (3.0–7.3) years. RESULTS: The overall prevalence of GADA was 0.53% and was higher in diabetic subjects (1.25%) than in nondiabetic subjects (0.47%). We found a decrease in baseline body mass index (BMI) from the GADA- to GADA(high) subgroups among baseline diabetic and prediabetic patients and also those who developed diabetes later in the cohort study. A total of 136 subjects (2.8%) developed diabetes after a median follow-up of 3.5 years. For GADA(+) participants, BMI was not associated with the risk for diabetes. In the Cox regression model, the GADA(low) and GADA(high) exhibited 2.63-fold and 4.16-fold increased risk for diabetes, respectively. This increased risk for diabetes by GADA-positivity is only found in male adults (HR 4.55, 95% CI 2.25–9.23). CONCLUSION: GADA has a low prevalence in China but is associated with a 2.63–4.16-fold increased risk for diabetes.