Fasting plasma glucose in screening for NIDDM in the U.S. and Israel

OBJECTIVE--To demonstrate the inadequacy of fasting plasma glucose for screening for NIDDM, even among groups at high risk for diabetes. RESEARCH DESIGN AND METHODS--Representative samples of adults 40-69 years of age in the U.S. (n = 2,035) and Israel (n = 2,316) were selected. Fasting plasma glucose (FPG) was measured and a 2-h oral glucose tolerance test (OGTT) was administered. Subjects with undiagnosed NIDDM were identified using internationally accepted diagnostic criteria (FPG > or = 7.8 mM or 2-h plasma glucose > or = 11.1 mM). RESULTS--Only 31-38% of subjects with undiagnosed NIDDM had fasting hyperglycemia (> or = 7.8 mM), and 36% in the U.S. and 19% in Israel had normoglycemia (< 6.1 mM). Postchallenge glucose, diagnostic of diabetes, was associated with all fasting values, including values < 5.0 mM. Based on sensitivity, specificity, and positive predictive value, no FPG level provided a satisfactory cutoff point to use in screening for undiagnosed NIDDM. Sensitivity at each FPG cutoff point varied little among groups classified by age, sex, race, blood pressure status, or body mass index (BMI) levels > 23, but sensitivity was lower among those with BMI levels < 23. CONCLUSIONS--In the clinical setting, FPG is commonly used in screening for NIDDM. However, fasting values < or = 7.8 mM are highly insensitive for detecting NIDDM. Lower FPG cutoff points tha achieve acceptable sensitivity are accompanied by inadequately low specificity, require a high percentage of patients to be retested, and result in a low yield of diabetes among those screened. Clinicians and researchers who seek detection of undiagnosed NIDDM should use the OGTT, because FPG lacks adequate sensitivity and specificity for this purpose.