Aims/hypothesis The aim of this work was to investigate whether tetraspanin 7 autoantibodies (TSPAN7A) are valuable in predicting poor beta cell function in individuals with latent autoimmune diabetes in adults (LADA). Methods The cross-sectional study involved participants with LADA (n = 173), type 1 diabetes (n = 158), type 2 diabetes (n = 204) and healthy control participants (n = 170). The longitudinal study involved 53 participants with LADA, with a 3-year follow-up. In both cohorts, TSPAN7A in the sera were measured by a luciferase immunoprecipitation system assay, and physical and clinical characteristics were recorded. Results The prevalence of TSPAN7A in LADA, type 1 diabetes, type 2 diabetes and healthy control participants was 21.4% (37/ 173), 26% (41/158), 0.5% (1/204) and 1.2% (2/170), respectively. Importantly, measurement of TSPAN7A significantly increased the number of individuals with LADA found to be positive for multiple antibodies (32.4% vs 22%; p < 0.001). Further logistic regression analysis demonstrated that positivity for TSPAN7A (OR 2.87, p = 0.034), disease duration (OR 1.83, p = 0.019) and GAD antibody titre (OR 2.67, p = 0.009) were risk factors for beta cell function in LADA, while BMI (OR 0.34, p = 0.001) was a protective factor. In the prospective study in individuals with LADA, the median annual decrease in rates of fasting C-peptide and 2 h postprandial C-peptide in individuals who were positive for TSPAN7A was significantly higher when compared with the decrease in those who were negative for TSPAN7A (34.6% vs 7.9%, p = 0.043 and 33.2% vs 11%, p = 0.041, respectively). Conclusions/interpretation TSPAN7A are valid islet autoantibodies for use in East Asian populations with autoimmune diabetes and can discriminate individuals with LADA who have lower beta cell function after disease progression.