Estimated average glucose has been used to transform HbA1c into a glucose measure that might better inform patients of their glycaemic control. The data set used to obtain the estimated average glucose equation was derived in adults with Type 1 and Type 2 diabetes, along with normal healthy control subjects, and requires testing in children.


This was a cross-sectional study of 234 children and young people (106 male) with Type 1 diabetes aged 4.0–23.5 years who underwent continuous glucose monitoring over a 5-day period along with a measure of HbA1c. Regression analysis was used to determine estimated average glucose and agreement was assessed with the average glucose estimated from the Nathan equation: Nathan average glucose equation = 1.59 (HbA1c%) – 2.59.


Mean HbA1c was 76 mmol/mol (25.1) [9.1 (2.3)%] and mean continuous glucose monitoring tissue glucose was 10.4 (2.6) mmol/l. The relationship between continuous glucose monitoring tissue glucose and HbA1c was described by the paediatric equation: paediatric estimated average glucose = 0.49 (HbA1c%) + 5.95 (r = 0.45; P < 0.001). The mean paediatric estimated average glucose was 10.4 (1.1) mmol/l compared with that from the Nathan average glucose equation of 11.9 (3.7) mmol/l (P < 0.001). Overall, the paediatric estimated average glucose was 2.7 mmol/l lower than the Nathan estimated average glucose, with a 95% limit of agreement of ± 0.5 mmol/l. The agreement was very close with HbA1c values below 80 mmol/mol (9.5%).


These data suggest that the Nathan estimated average glucose could be used in children and young people with Type 1 diabetes. Caution should still be exercised in the estimates derived for average glucose as the data set is skewed in both Nathan and paediatric average glucose estimates in opposite directions because of the differences in average HbA1c.