OBJECTIVE To determine A1C cut points for glucose intolerance in Asian Indians. RESEARCH DESIGN AND METHODS A total of 2,188 participants without known diabetes were randomly selected from the Chennai Urban Rural Epidemiology Study. All had fasting plasma glucose (FPG) and 2-h postload plasma glucose measurements after a 75-g load and were classified as having impaired fasting glucose (IFG) (American Diabetes Association [ADA] criteria, FPG > or =5.5 and <7 mmol/l, and World Health Organization [WHO] criteria, FPG > or =6.1 and <7 mmol/l), impaired glucose tolerance (IGT) (2-h postload plasma glucose > or =7.8 and <11.1 mmol/l), or diabetes (FPG > or =7 mmol/l and/or 2-h postload plasma glucose > or =11.1 mmol/l). A1C was measured using the Bio-Rad Variant machine. Based on receiver operating characteristic curves, optimum sensitivity and specificity were derived for defining A1C cut points for diabetes, IGT, and IFG. RESULTS Mean +/- SD values of A1C among subjects with normal glucose tolerance, IGT, and diabetes were 5.5 +/- 0.4, 5.9 +/- 0.6, and 8.3 +/- 2.0%, respectively (P(trend) < 0.001) with considerable overlap. To identify diabetes based on 2-h postload plasma glucose, the A1C cut point of 6.1% had an area under the curve (AUC) of 0.941 with 88.0% sensitivity and 87.9% specificity. When diabetes was defined as FPG > or =7.0 mmol/l, the A1C cut point was 6.4% (AUC = 0.966, sensitivity 93.3%, and specificity 92.3%). For IGT, AUC = 0.708; for IFG, AUC = 0.632 (WHO criteria) and 0.708 (ADA criteria), and the A1C cut point was 5.6%. CONCLUSIONS In Asian Indians, A1C cut points of 6.1 and 6.4% defined diabetes by 2-h postload plasma glucose or FPG criteria, respectively. A value of 5.6% optimally identified IGT or IFG but was <70% accurate.