A systematic review of HbA1c variables used in the study of diabetic complications

SummaryAimThe predictive power of different HbA1c variables is of importance in prognosis of diabetic complications, clinical guidelines, health-economical analyses and in the design of clinical trials of antidiabetic agents. The aim was to review the literature with regard to the HbA1c variables used, and determine how the predictive power of these relates to diabetic complications.MethodWe reviewed 97 full-text articles on HbA1c and diabetic complications.ResultsThe most commonly used HbA1c variables were: the baseline value, the mean and the “updated” mean HbA1c. Other variables used were the logarithm of the updated mean, the standard deviation, the slope (annual average change), the initial decline (change during the first year), the final value, and the change in HbA1c between baseline and the fourth year.The updated mean, logarithm of the updated mean and mean HbA1c were found to have greater predictive power than baseline HbA1c. The slope, final value, S.D., initial decline and change of HbA1c did not add any further information. The predictive power of the mean or updated mean HbA1c became stronger with longer study lengths. There was a persistent effect over several years between HbA1c values and diabetic complications. Measurements of HbA1c varied from a single value to measurements each month.ConclusionsThe use of a mean or the updated mean HbA1c is recommended in the study design. HbA1c values several years old also influence the prognosis of diabetic complications. The possibility of finding a true effect of an antidiabetic agent increases with longer study length.