We re-assess the fasting plasma
glucose (FPG) and 2-h post-load
glucose (2HPG) in the diagnosis of both
prediabetes and
type 2 diabetes mellitus by developing a gold standard based on beta-cell function. The gold standard was developed in a cohort of 2,152 adult subjects without severe renal or
liver dysfunction that also had a 2-h oral
glucose tolerance test (OGTT) during the Third National Health and Nutrition Examination Survey. Beta-cell function was computed based on a composite of insulin secretion (determined based on the
insulin and
glucose responses to the OGTT) and the homeostasis model
insulin resistance index. The X-tile program was used to generate the most appropriate categories of minor, moderate and severe impairment of beta-cell function based on the best discrimination of ln(insulin secretion). Subjects with a moderate or severe impairment in beta-cell function were used to define
prediabetes and diabetes, respectively, and was the standard against which the FPG and 2HPG were evaluated. It is shown that the current definitions of diabetes by the FPG and 2HPG mirror those derived using impairment of beta-cell function as the gold standard. It is also shown that lowering the cutoff for the FPG does little to improve its use in the screening for
prediabetes. A major finding is that the current 2HPG cutoff is inadequate and thus in need of revision to >6.7 mmol/L (>120 mg/dL) from 7.8 mmol/L (140 mg/dL) for the lower cutoff. The use of a model of beta-cell function and impairment of insulin secretion has thus put the utility of the FPG and 2HPG into perspective: We recommend that performing an OGTT be considered pivotal for accurate identification of subjects with impaired beta-cell function (and thus
prediabetes) and a revision of the OGTT lower cutoff be considered based on the results of this study.