Objective: Little is known about using glycated haemoglobin A₁c (HbA₁c) to diagnose diabetes in Chinese subjects over 50 years old. This study aims to evaluate HbA₁c in diagnosing diabetes and identify the optimal threshold to be used in Chinese community subjects aged over 50 years.
Methods: A community-based cross-sectional survey was conducted from October 2010 to January 2011 in Shipai community of Guangzhou, Guangdong, China. A total of 1494 subjects (72·8%) aged over 50 years were investigated. Fasting plasma glucose (FPG₁st ) and HbA₁c were assayed in each participant. Diabetic candidates with FPG₁st ≥ 5·6 mmol/l or HbA₁c ≥ 39 mmol/mol (5·7%) were informed to undergo a 75-g oral glucose tolerance test (OGTT). Diagnosis of diabetes was made by 1999 World Health Organization criteria. Sensitivity and specificity of HbA1 c for diagnosing diabetes were calculated by receiver operating characteristics (ROC) curve.
Results: Among 1494 subjects, 161 subjects (10·8%) with previously diagnosed diabetes and 21 with missing data were excluded. Among the remaining 1312 subjects (87·8%), 861 subjects (65·6%) with either FPG1st ≥ 5·6 mmol/l or HbA₁c ≥ 39 mmol/mol (5·7%) were invited to perform OGTT. Finally, 453 subjects (52·6%) performed OGTT (FPG₂nd and 2-h plasma glucose were measured) and 54 subjects (11·9%) were identified as being diabetes. The area under ROC curve was 0·916 (0·887-0·940) for HbA₁c and 0·972 (0·953-0·985) for FPG₂nd in diagnosing diabetes (P = 0·045). An HbA₁c threshold of 48 mmol/mol (6·5%) yielded the highest combination of sensitivity (75·9%) and specificity (95·5%) for diagnosing diabetes.
Conclusion: An HbA1 c threshold of 48 mmol/mol (6·5%) was highly specific and had a good sensitivity for diagnosing diabetes among Chinese subjects aged over 50 years with FPG ≥ 5·6 mmol/l or HbA₁c ≥ 39 mmol/mol (5·7%). This threshold may be suitable for diagnosing diabetes in Chinese subjects over 50 years old.
© 2013 John Wiley & Sons Ltd.