Context: Growth hormone (GH) is known to be diabetogenic, but the risk of diabetes in individuals treated with GH in childhood has been little evaluated, and conflicting results have been obtained. Objective: To investigate the prevalence of diabetes and gestational diabetes in a population-based cohort of patients treated with GH for short stature in childhood in France. Design, setting, and participants: Participants were a population-based cohort of 5100 children with idiopathic isolated GH deficiency, idiopathic short stature, or short stature in children born short for gestational age who started GH treatment between 1985 and 1996. Data on the delivery of diabetes drugs in 2009 and 2010 were obtained from the French national health insurance database. Cases in patients and controls were identified from diabetes drugs deliveries. Main outcome measure: The prevalence of diabetes was calculated and compared with that in the general population, determined on the basis of data from the same source, with the same definition. Results: At a mean age of 30 years, no difference in the prevalence of treated diabetes (oral drugs or insulin) was found between subjects treated with GH and the general population in France, regardless of sex. Similarly, the risk of insulin-treated gestational diabetes was similar in patients and in the reference population. Conclusions: No difference in the risk of diabetes was found between GH-treated patients and the reference population. These results are reassuring, but further studies with a longer follow-up are required to evaluate the risk of diabetes with age in these patients.