Glycated hemoglobin (HbA(1c)) has a key role for diagnosing diabetes and monitoring glycemic state. As recently reviewed, available data on HbA(1c) biological variation show marked heterogeneity. Here we experimentally revaluated these data using a well designed protocol. We took five EDTA whole blood specimens from 18 apparently healthy subjects on the same day, every two weeks for two months. Samples were stored at -80°C until analysis and assayed in duplicate in a single run by Roche Tina-quant® Gen.2 immunoassay. Data were analyzed by the ANOVA. To assess the assay traceability to the IFCC reference method, we preliminarily carried out a correlation experiment. The bias (mean±SD) of the Roche immunoassay was 0.3%±0.7%, confirming the traceability of the employed assay. No difference was found in HbA(1c) values between men and women. Within- and between-subject CV were 2.5% and 7.1%, respectively. Derived desirable analytical goals for imprecision, bias, and total error resulted 1.3%, 1.9%, and 3.9%, respectively. HbA(1c) had marked individuality, limiting the use of population-based reference limits for test interpretation. The estimated critical difference was ~10%. For the first time we defined biological variation and derived indices for the clinical application of HbA(1c) measurements using an accurately designed protocol and an assay standardized according to the IFCC. Copyright © 2011 Elsevier B.V. All rights reserved.