A study was made on the association among 2-h plasma glucose (PG) in oral glucose tolerance test (OGTT), fasting plasma glucose (FPG) using correlation and regression equation. Subjects were 13174 OGTT examinees tested between 1980 and 1998. Blood glucose was determined by the glucose oxidase method and glycated hemoglobin (HbA1c) by the HPLC method. As for correlation between 2-h PG and FPG, regression equation of the <60 year group was y=57.1+0.336x (r=0.866, P<0.0001) and that of the >==60 year group was y=61.5+0.286x (r=0.814, P<0. 0001). FPG was calculated at 124.3 in the <60 year group and 118.7 mg/dl in the >==60 year group for 2-h PG of 200 mg/dl, 2-h PG were calculated at 199.5 and 210.7 mg/dl for FPG of 126 mg/dl, respectively. In the <60 year group, FPG were calculated at 121.7 and 124.4 mg/dl and 2-h PG at 193.2 and 199.3 mg/dl for HbA1c of 6.0 and 6.1%, respectively. As for associations between HbA1c and FPG or 2-h PG being high correlation, it is possible to estimate a prevalence of DM in a group using HbA1c>==6.1%. High correlations were demonstrated among all the three measures; FPG, 2-h PG, HbA1c. If 2-h PG is used in diagnosing diabetes mellitus, an FPG of 126 mg/dl proposed by ADA and World Health Organization (WHO) as a diagnostic level of FPG is an acceptable value for the Japanese.