International criteria for the diagnosis of diabetes and impaired glucose tolerance.

International agreement on classification and criteria for the diagnosis of diabetes is highly desirable. Two systems promulgated in 1979-80 and widely used today are those of the NIH National Diabetes Data Group (NDDG) and of the World Health Organization (WHO). Although these systems are similar in many ways, certain discrepancies between them result in different classifications of oral glucose tolerance test (OGTT) results and different estimates of prevalence of the various glucose tolerance groups. Analysis of 3704 OGTTs performed during a survey of a national probability sample of U.S. residents without known diabetes and aged 20-74 yr shows that the two systems agreed in classification of 87.7% of OGTTs. For the remainder, the NDDG and WHO classifications differed, primarily because nondiagnostic OGTTs occur in the NDDG system but not in the WHO system. The differences resulted in the prevalence of impaired glucose tolerance (IGT) using WHO criteria (11.6%) being more than twice that using NDDG criteria (4.9%), although prevalence rates of diabetic OGTTs were similar in both systems (3.6%, 3.4%). The WHO system represents a simpler, inclusive classification scheme, and there is insufficient evidence from longitudinal studies of prognostic differences that would justify the more complicated NDDG diagnostic criteria. In situations where multiple venipunctures or retesting are not possible, the venous plasma glucose concentration at 2 h after 75 g glucose appears to be the most appropriate single value to use to designate whether a person has diabetes, IGT, or neither. The use of this value alone placed 97% of diabetic subjects and all other subjects, in the same class they were in when the full WHO criteria were used.