Objective: The aim of this study was to assess the validity of fasting plasma glucose (FPG) and/or glycated haemoglobin (HbA(1c)) as screening tests for the early detection of diabetes in high-risk subjects.
Methods: A total of 392 subjects (149 male and 243 female) with risk factors for diabetes were included. All subjects underwent a 75-g oral glucose tolerance test and HbA(1c) measurement. Receiver operating characteristic curve analysis was used to examine the sensitivity and specificity of FPG and HbA(1c) for detecting diabetes, which was defined as a FPG > or = 7.0 mmol/l or a post-challenge 2-h plasma glucose > or = 11.1 mmol/l.
Results: The prevalence of newly diagnosed diabetes was 22.4% (n = 88). The current guideline of FPG > or = 7.0 mmol/l for diabetes screening detected only 55.7% of diabetic subjects. The optimal cut-off points of HbA(1c) and FPG for the diagnosis of diabetes were 6.1% (sensitivity 81.8%, specificity 84.9%) and 6.1 mmol/l (sensitivity 85.2%, specificity 88.5%), respectively. The screening model using FPG > or = 6.1 mmol/l and/or HbA(1c) > or = 6.1% had sensitivities of 71.6-95.5% and specificities of 77.6-95.7% for detecting undiagnosed diabetes.
Conclusions: The current American Diabetes Association diagnostic criteria, based only on FPG, are relatively insensitive in the detection of diabetes in high-risk subjects. The simultaneous measurement of FPG and HbA(1c) might be a more sensitive screening tool for identifying high-risk individuals with diabetes at an early stage.