Detection of diabetes and prediabetes using glycosylated hemoglobin in Chinese adults living in Shanghai: A prospective analysis.

BACKGROUND

This study aimed to evaluate the discriminative abilities of glycosylated hemoglobin (HbA1c) and to examine the optimal HbA1c cutoff values for diabetes and prediabetes in Chinese adults.

METHODS

Data of a population-based cohort of Chinese adults aged ≥40 years living in Jiading District in Shanghai were used. At baseline, 9389 and 7241 participants were included to identify the optimal HbA1c cutoff values for diabetes and prediabetes, respectively using the 1999 World Health Organization criteria as reference. In addition, the follow-up data on incident diabetes of 4538 participants were used to determine the HbA1c cutoff value for prediabetes using the development of diabetes as reference. The discriminative abilities of HbA1c were evaluated using receiver operating characteristic (ROC) curves, and the optimal cutoff values were determined by Youden's index.

RESULTS

The areas under the ROC curves were 0.849 for diabetes, 0.614 for prediabetes using baseline data, and 0.648 for prediabetes using follow-up data. An HbA1c cutoff value of 6.0% had the largest Youden's index to diagnose diabetes with a sensitivity of 70.2% and a specificity of 87.4%. An HbA1c cutoff value of 5.6% was indicated for prediabetes using both baseline and follow-up data. However, the sensitivity and specificity were both low (55.4% and 61.1% using an oral glucose tolerance test as reference, 64.6% and 57.1% using incident diabetes as reference).

CONCLUSIONS

An HbA1c value ≥6.0% could be used to detect diabetes in Chinese adults aged ≥40 years. However, although an HbA1c value of 5.6% to 5.9% was indicated in this study, the overall discrimination of HbA1c for prediabetes was poor.