How valid is fasting plasma glucose as a parameter of glycemic control in non-insulin-using patients with type 2 diabetes?

OBJECTIVE: To assess the value of fasting blood glucose as a parameter for glycemic control in type 2 diabetic patients not using insulin. RESEARCH DESIGN AND METHODS: In 1,020 type 2 diabetic patients treated with diet or oral hypoglycemic agents (OHAs), measurements of fasting plasma glucose (FPG) and HbA1c were taken. In 617 patients, the measurement could be repeated after 3 months. Cross-sectional correlation coefficients were calculated for the association between HbA1c and FPG. Receiver-operating characteristic (ROC)-curve analyses were applied to examine the performance of FPG as a diagnostic test for HbA1c. Longitudinally, the change in FPG was compared with the change in HbA1c, with both correlation measures and ROC curve analyses. RESULTS: Correlation coefficients between HbA1c and FPG and between FPG change and HbA1c change were 0.77 and 0.65, respectively. ROC curve analysis showed that HbA1c is difficult to predict from FPG values: 66% of the patients with good HbA1c (< 7.0%) were identified as such by FPG values < 7.8 mmol/l. As a test for HbA1c change, FPG change performed moderately: the highest combined values of sensitivity and specificity (87.7 and 57%, respectively) were reached at a cutoff point of zero in the range of FPG change values. CONCLUSIONS: FPG and HbA1c values that do not correspond are not rare in type 2 diabetic patients on diet or OHA treatment. HbA1c is difficult to predict from FPG values, and even more difficult is the prediction of HbA1c changes from FPG changes.