Objectives: Early diagnosis of diabetes is crucially important in reduction of the complications. Although HbA1c is an accurate marker for the prediction of complications, less information is available about its accuracy in diagnosis of diabetes. In this study, the association between HbA1c and FBS was assessed through a cross-sectional population based study.
Methods: A random sample of population in Kerman city was selected. The total number was 604 people. Their HbA1c and fasting blood sugar (FBS) were tested. The association between HbA1c and FBS and also their sensitivity, specificity and predictive values in detection of abnormal values of each other were determined. Results: The association of HbA1c with FBS was relatively strong particularly in diabetic subjects. Generally, FBS was a more accurate predictor for HbA1c compared with HbA1c as a predictor of FBS. Although the optimum cutoff point of HbA1c was >6.15%, its precision was comparable with the conventional cutoff point of >6%. Conclusions: In conclusion, FBS sounds more reliable to separate diabetic from non-diabetic subjects than HbA1c. In case of being interested in using HbA1C in screening, the conventional cutoff points of 6% is an acceptable threshold for discrimination of diabetics from non-diabetics.