AIMS: We investigated pathways linking offspring birth weight to maternal diabetes risk in later life by taking into account a range of prospective early-life and adult maternal factors. METHODS: In a national birth cohort study, we examined the relationship between offspring birth weight and maternal glycated haemoglobin (HbA1c) at age 53 years in 581 mothers who had a first birth between age 19 and 25 years, and had data on potential confounders or mediators. RESULTS: Mean age at first birth was 21.5 years. After adjustment for maternal body mass index (BMI), mean percentage change in maternal HbA1c per kilogram increase in offspring birth weight was -1.8%[95% confidence interval (CI) -3.5, -0.1; P = 0.03]. This relationship was mostly accounted for by gestational age that was inversely related to maternal HbA1c (-0.9%; 95% CI -1.5, -0.4; P = 0.001). Other risk factors for high HbA1c were smoking and high BMI at 53 years. There was a significant interaction between offspring birth weight and maternal childhood social class (P = 0.01). Mothers from a manual background with higher birth weight offspring had lower HbA1c (BMI adjusted: -3.1%; 95% CI -5.0, -1.1); this was not observed for mothers from a non-manual background (BMI adjusted: 1.9%; 95% CI -1.3, 5.0). CONCLUSIONS: Short gestational age and low offspring birth weight may be part of a pathway linking impaired early maternal growth to diabetes risk in later life. A second possible pathway linking higher offspring birth weight to later maternal glucose status was also identified. These potential pathways require further investigation in cohorts with a wider maternal age range so that the early targeting of public health initiatives can be assessed.