Prognostic value of haemoglobin A1c and fasting plasma glucose for incident diabetes and implications for screening

Abstract

The aim of this analysis is to compare screening strategies with haemoglobin A(1c) (HbA(1c)), fasting plasma glucose (FPG) or combined measures in the identification of individuals at high risk for diabetes. Applying American Diabetes Association thresholds for FPG and HbA(1c) screening, 6,803 subjects free of diabetes were classified as non-diabetic, pre-diabetic and possibly diabetic by FPG (<100, 100-125 and >125 mg/dl) and HbA(1c) (<5.7, 5.7-6.4 and >6.4%). Hazard ratios, sensitivity and specificity were estimated for individuals with pre-diabetes with respect to incident diabetes in the following 5 years. Areas under the receiver operating characteristic curves (AUC) were estimated for levels of FPG ≤ 125 mg/dl and HbA(1c) ≤ 6.4% in diabetes prediction. Although FPG and HbA(1c) screenings poorly agreed in classifying individuals as pre-diabetic, hazard ratios [95% confidence interval] for incident diabetes were similarly increased in univariate models in the two pre-diabetic groups: FPG 100-125 mg/dl, 4.72 [3.69; 6.05]; HbA(1c) 5.7-6.4%, 3.97 [3.05; 5.23]. HbA(1c) and FPG had comparable AUCs (FPG, 0.732; HbA(1c), 0.725) and consequently similar 5-year sensitivities and specificities for their pre-diabetes definitions (when the lower cut-off for HbA(1c)-defined pre-diabetes was increased to a level between 5.8 and 5.9%). Combining HbA(1c) and FPG increased the AUC to 0.778, and a further increase to 0.817 was seen with additional inclusion of conventional risk factors. FPG and HbA(1c) have comparable (yet insufficient) abilities in identifying individuals at high risk for diabetes. Effectiveness of a diabetes screening program could be improved by a risk score including FPG and HbA(1c).