Biological variation of glycated haemoglobin in a paediatric population and its application to calculation of significant change between results

Abstract

To determine precisely the probability that a change between two glycated haemoglobin A1c (HbA1c) results is significant and that clinical actions may be required, the biological variation of HbA1c must be known. However, it has not been evaluated in a paediatric population. We therefore determined the long-term biological variation of HbA1c in a paediatric population and used it to generate a probability curve for significant changes between two consecutive HbA1c measurements. A group of 24 boys and 14 girls with cystic fibrosis (CF) but without diabetes or impaired glucose tolerance has been selected. HbA1c has been measured at least five times over five consecutive years for all subjects. We have used the Fraser and Harris method to calculate within-subject biological variation (CV(I)), which allowed the determination of the probability that a change is significant between results. As within-subject variances are equivalent for girls and boys (P > 0.1), both genders were merged for biological variation analysis. The CV(I) calculated for HbA1c was 4.8% and the between-subject variation (CV(G)) was 12.8%. Then, a probability curve based on the CV(I) found was generated and showed that a change of 14% between two consecutive HbA1c results corresponding to a probability of 95% was significant. We determined for the first time the biological variation of HbA1c in a paediatric population, which is higher than the ones found for adult populations. The probability curves generated from these data could be invaluable tools for clinicians to balance HbA1c results with other clinical parameters.