BACKGROUND: To investigate whether the combination of HbA1c and fasting plasma glucose (FPG) can be used for the diagnosis of diabetes and impaired glucose tolerance (IGT) in people at high risk of cardiovascular disease (CVD). METHODS: A cross-sectional study was performed on 2907 people at high risk of cardiovascular events but without a previous diagnosis of diabetes. Optimal cut-off points and the diagnostic potential of FPG, HbA1c, and their combination were determined. RESULTS: The sensitivity of the usually applied FPG cut-off point of 7.0 mmol/L to diagnose diabetes mellitus was low (59.0%). Receiver operating characteristic (ROC) curve analysis indicated that the optimal cut-off points for the diagnosis of diabetes using FPG and HbA1c were 6.4 mmol/L (sensitivity 75.7%; specificity 77.5%; likelihood ratio 3.37) and 5.9% (41 mmol/mol; sensitivity 68.7%; specificity 67.1%; likelihood ratio 2.09), respectively. To diagnose IGT, the optimal cut-off points for FPG and HbA1c were 6.1 mmol/L (sensitivity 57.1%; specificity 57.9%) and 5.7% (39 mmol/mol; sensitivity 63.8%; specificity 60.3%), respectively. For diabetes, combining cut-off points for FPG and HbA1c identified four categories with likelihood ratios ranging from 5.59 to 0.21, and post-test probabilities between 69.3% and 7.8%. For IGT, likelihood ratios varied between 2.05 and 0.56, whereas post-test probabilities ranged from 84.0% to 58.8%. CONCLUSIONS: Using FPG alone results in the underdiagnosis of glucometabolic abnormalities in people at high risk of CVD. Using an algorithm with both HbA1c and FPG improves the detection of diabetes, but not IGT, and could be easily implemented in patient care.
BACKGROUND: To investigate whether the combination of HbA1c and fasting plasma glucose (FPG) can be used for the diagnosis of diabetes and impaired glucose tolerance (IGT) in people at high risk of cardiovascular disease (CVD). METHODS: A cross-sectional study was performed on 2907 people at high risk of cardiovascular events but without a previous diagnosis of diabetes. Optimal cut-off points and the diagnostic potential of FPG, HbA1c, and their combination were determined. RESULTS: The sensitivity of the usually applied FPG cut-off point of 7.0 mmol/L to diagnose diabetes mellitus was low (59.0%). Receiver operating characteristic (ROC) curve analysis indicated that the optimal cut-off points for the diagnosis of diabetes using FPG and HbA1c were 6.4 mmol/L (sensitivity 75.7%; specificity 77.5%; likelihood ratio 3.37) and 5.9% (41 mmol/mol; sensitivity 68.7%; specificity 67.1%; likelihood ratio 2.09), respectively. To diagnose IGT, the optimal cut-off points for FPG and HbA1c were 6.1 mmol/L (sensitivity 57.1%; specificity 57.9%) and 5.7% (39 mmol/mol; sensitivity 63.8%; specificity 60.3%), respectively. For diabetes, combining cut-off points for FPG and HbA1c identified four categories with likelihood ratios ranging from 5.59 to 0.21, and post-test probabilities between 69.3% and 7.8%. For IGT, likelihood ratios varied between 2.05 and 0.56, whereas post-test probabilities ranged from 84.0% to 58.8%. CONCLUSIONS: Using FPG alone results in the underdiagnosis of glucometabolic abnormalities in people at high risk of CVD. Using an algorithm with both HbA1c and FPG improves the detection of diabetes, but not IGT, and could be easily implemented in patient care.