Review of criteria for diagnosis of diabetes mellitus based on results of follow-up study.

Comparison of the 75-g oral glucose tolerance test (OGTT) method was made with the conventional 50-g method in a fixed population followed for 1--17 yr. The possibility of using the results in establishing the diagnosis of diabetes mellitus was also considered. The following results were obtained: (1) The 75-g method showed significantly higher 2-h and 3-h postchallenge plasma glucose (PG) and immunoreactive insulin values. (2) The two methods showed good correlation in PG values at various time periods, and there was no difference between the two at 1/2 h and 1 h. The 2-h standard value of 200 mg/dl used to diagnose diabetes with the 75-g method was equivalent to 180 mg/dl by the 50-g method. The upper limits of normals were 140 mg/dl and 120 mg/dl, respectively, for the two tests. (3) Those subjects diagnosed as diabetic on the basis of fasting plasma glucose (FPG) values of 140--149 mg/dl only had a high rate of reverting to normal over time. The frequency of "nondiabetic" plasma glucose values after glucose loading steadily decreases as FPG increased, with separation into two asymptotic lines at 150 mg/dl level. Thus, the logical value for diagnosis of diabetes when based only on FPG value is considered to be 150 mg/dl. (4) A smaller number of individuals had 2-h PG values that satisfied criteria for diagnosis of diabetes mellitus, but 1/2-h and 1-h values that were less than 200 mg/dl. Nevertheless, follow-up of these subjects showed a high development rate of diabetes. Thus, the 1/2-h and 1-h values are not considered necessary to establish diagnosis of diabetes mellitus.