Abstract
We compared the oral glucose tolerance test (I) as evaluated by six commonly used scoring methods and total glycohemoglobin assay (II) with respect to their value in the diagnosis of diabetes mellitus. Depending on the evaluation method used for I, 16.7 to 64.3% of those subjects diagnosed as diabetic or borderline by this test were judged to be normal by II. The best agreement was between II and the Unger evaluation method. High-density-lipoprotein cholesterol, which showed an inverse correlation with II, was decreased in subjects judged to be diabetic by the Unger method. We conclude that the utilization of II measurement as a screening method for diabetes mellitus is consistent with a conservative approach to the diagnosis of diabetes.