To evaluate HbA1c as a diagnostic tool in prediabetes-impaired fasting glucose (IFG) and impaired glucose tolerance (IGT), and newly detected diabetes (NDD), defined by plasma glucose and OGTT. 2,231 subjects, of mean age 50.3 ± 13.9 years and mean BMI 29.5 ± 6.2 kg/m(2), underwent an OGTT. HbA1c performance was assessed using the area under the receiver operating characteristics curve (AUC-ROC). HbA1c was significantly higher in all groups with altered glucose tolerance-5.72 ± 0.61% in IFG, 5.84 ± 0.63% in IGT, and 7.5 ± 1.69% in NDD when compared to normal glucose tolerance-5.23 ± 0.65% (P < 0.0001). HbA1c of both prediabetic groups was significantly lower in comparison with NDD (P < 0.0001); in IGT being significantly higher than in IFG (P = 0.02). ROC analysis demonstrated good performance of HbA1c for diagnosing diabetes-AUC-ROC 0.958 (95% CI: 0.946-0.970), as well as prediabetes-AUC-ROC 0.729 (95% CI: 0.702-0.755). The optimal cut-off level of HbA1c for diagnosing diabetes was 6.1% (sensitivity 86%, specificity 92%) and for undiagnosed prediabetes-5.5% (sensitivity 71%, specificity 64%). HbA(1c) appears to be a useful, convenient, and reliable tool for identifying subjects with prediabetes and diabetes and should be considered in the development of diagnostic strategies.