The measurement of glycated haemoglobin (HbA(1c)) is a practical and more sensitive tool than fasting plasma glucose (FPG) in screening type 2 diabetes in current practice. Its use has been limited so far by the variability of the analytical methods. The standardization process is going on, and many laboratories are currently using valid methods. Our study is consistent with the results of other groups who recommended this measurement to identify undiagnosed diabetic patients, that are about 25% to 30% in the French population. The demonstration was provided through a survey including a screening step by both HbA(1c) and G0, and a second exam with a 2 hr OGTT in a sample of positive screenees according to at least one criterion (HbA(1c) >=6% or G0 >=1.26 g/L), as well as in a sample of negative screenees. We showed that nine confirmed diabetic subjects out of ten had HbA(1c) >=6% at the screening step, while only a half had G0 >=1.26 g/L. Conversely, 22% of the positive screenees according to HbA(1c) were not confirmed as diabetic by the OGTT, including however more than half with abnormal glucose values. A chart for practical use is proposed to define patients at risk, the process of screening, and the patient follow-up according to the results of the tests.