ABSTRACT
BACKGROUND: In 1997, the Expert Committee of the American Diabetes Association (ADA) revised the diabetes mellitus (DM) diagnostic criteria to facilitate the diagnosis of DM on the basis of fasting plasma glucose (PG). The major purpose of the study is to evaluate if oral glucose tolerance test (OGTT) is still needed after the revision of criteria. METHODS: From September 1994 to March 1995, 247 ambulatory subjects referred by clinicians for 75-g OGTT were recruited for the study. Fasting and 2-h PG and serum insulin concentrations were determined. Additional fasting blood samples were collected for the measurement of HbA1c. We used the receiver operating characteristic (ROC) curve to locate a cut-point of fasting PG corresponding to 2-h PG of 200 mg/dl. RESULTS: The optimal fasting PG level depicted by ROC curve was more than 105 mg/dl, giving 80.6% sensitivity and 85.6% specificity for the diagnosis of DM in clinical setting. If fasting PG > or = 126 mg/dl was employed, the specificity was 98.3% but the sensitivity went down to 42.6%. High rates of glucose intolerance would remain undiagnosed in subjects with fasting PG less than 126 mg/dl (41.6% of them being IGT and 38.6% DM), if the OGTT was exempted from clinical practice. HbA1c more than 6.2% could be a clue to recognize undiagnosed DM, but was unable to separate impaired glucose tolerance (IGT) from non-DM. CONCLUSIONS: In the population studied, undiagnosed glucose intolerance can still be encountered in a large number of subjects with fasting PG less than 126 mg/dl. OGTT is still indispensable in clinical setting for the diagnosis of DM and IGT after the revision of diagnostic criteria.
