Background

We sought to identify the optimal cut-off of glycated hemoglobin (HbA1c) for defining diabetes and to assess the agreements of fasting plasma glucose (FPG), fasting serum glucose (FSG), and HbA1c in defining diabetes among rural older adults in China.

Methods

This population-based cross-sectional study included 3547 participants (age ≥61 years, 57.8% women) from the Multidomain Interventions to Delay Dementia and Disability in Rural China from 2018–2019; of these, 3122 had no previously diagnosed diabetes. We identified the optimal cut-off of HbA1c against FPG ≥7.0 mmol/L for defining diabetes by using receiver operating characteristic curve and Youden index. The agreements of FPG, FSG, and HbA1c in defining diabetes were assessed using kappa statistics.

Results

Among participants without previously diagnosed diabetes (n = 3122), the optimal HbA1c cut-off for defining diabetes was 6.5% (48 mmol/mol), with the sensitivity of 88.9%, specificity of 93.7%, and Youden index of 0.825. The correlation coefficients were 0.845 between FPG and FSG, 0.574 between FPG and HbA1c, and 0.529 between FSG and HbA1c in the total sample (n = 3547). The kappa statistic for defining diabetes was 0.962 between FSG and FPG, and 0.812 between HbA1c and FPG.

Conclusions

The optimal cut-off of HbA1c for diagnosing diabetes against FPG >7.0 mmol/L is ≥6.5% in Chinese rural-dwelling older adults. The agreement in defining diabetes using FPG, FSG, and HbA1c is nearly perfect. These results have relevant implications for diabetes research and clinical practice among older adults in China.