Analysis of Fasting Plasma Glucose Values for Optimal Detection of Abnormal Responses on the Oral Glucose Tolerance Test in at-Risk Study Subjects

Abstract

To identify the fasting plasma glucose (FPG) value with the best performance for detecting an abnormal response on the oral glucose tolerance test (OGTT) in patients at risk for having type 2 diabetes. All patients who underwent a 2-hour OGTT during an 18-month period were included in this study. Pretest and posttest odds, likelihood ratios, and receiver operating characteristic curves were used to identify the FPG value most strongly associated with an abnormal result on the OGTT (either diabetes or impaired glucose tolerance [IGT]). Of the 1,371 patients who underwent an OGTT during the designated study period, 1,239 fulfilled the inclusion criteria. The prevalence of IGT was 25.34% (314 patients). Diabetes was diagnosed in 141 patients (11.38%). IGT was more commonly found in the FPG strata below 115 mg/dL; above this value, diabetes was more frequently diagnosed. In general, the percentage of cases of IGT increased progressively throughout the "normal" FPG range. The prevalence varied from 11.4% (in patients with FPG values <80 mg/dL) to 32% (in those with FPG levels from 95 to 99.9 mg/dL). FPG values between 95 and 99.9 mg/dL had a likelihood ratio of 2.1 for detecting an abnormal OGTT response, of 1.8 for detecting diabetes, and of 1.66 for detecting IGT. The odds ratio for detecting either IGT or diabetes was increased 2-fold by performing an OGTT. The FPG threshold with the best ability for detecting an abnormal response on the OGTT was 95 mg/dL (sensitivity of 0.72 and specificity of 0.65). In patients at risk for type 2 diabetes, the FPG cut point (95 mg/dL) most useful for detecting an abnormal OGTT response is included in the normal range of the FPG.