Predictive ability of fasting plasma glucose for a diabetic 2-h postload glucose value in oral glucose tolerance test: spectrum effect.

摘要

The performance of diagnostic tests may vary according to patient characteristics. The aim of this study is to find out the factors, if any, that may affect the performance of fasting plasma glucose (FPG) to predict a diabetic 2-h postload glucose level (> or =200 mg/dl) in oral glucose tolerance test (OGTT). One hundred ninety-six patients with known risk factors for diabetes mellitus to whom OGTT was applied were included. Factors that may have an effect on the performance of FPG in prediction of a diabetic value in OGTT were determined by using logistic regression and likelihood ratios (LRs). The cutoff of FPG predicting a 2-h postload glucose of > or =200 mg/dl was calculated by receiver operating characteristic curve as 110 mg/dl (sensitivity, 76.7%; specificity, 75.9%). Waist-to-hip ratio (WHR) and body mass index (BMI) influenced sensitivity, whereas age, family history, and presence of hyperlipidemia affected specificity of FPG. Significant factors for positive LR were age and hyperlipidemia, whereas sex, smoking, hyperlipidemia, physical inactivity, WHR, and BMI influenced negative LR. Fasting plasma glucose performance as a diagnostic test can be affected by many factors that are clearly stated as risk factors for diabetes mellitus. These data emphasize how the interpretation of a diagnostic test varies as the patient characteristics vary; the criteria that we confidently rely on may not be that reliable, changing between just two different patients.