Despite improvements in the clinical management of pregnant women with diabetes, pregnancy outcomes compare unfavourably to those of people without diabetes. The link between maternal glycaemia as a measure of metabolic control and fetal size has been the subject of previous studies. Evidence is conflicting due to the small numbers of cases and/or a mixture of different types of diabetes in most of these studies. This population‐based cohort study was carried out on 274 women with type 1 diabetes from 10 maternity units in the Mersey area. We showed that in early (10±2 weeks gestation) and late pregnancy (≥34 weeks gestation), maternal HbA1c is not associated significantly with macrosomic or large for gestational age infants. Good glycaemic control measured by HbA1c in pregnancy, although of proven benefit in reducing perinatal mortality and morbidity, is therefore unable to prevent these problems Copyright © 2004 John Wiley & Sons, Ltd.