Aims/hypothesis We determined the impact of insulin pump therapy on long-term glycaemic control, BMI, rate of severe hypoglycaemia and diabetic ketoacidosis (DKA) in children. Methods Patients on pump therapy at a single paediatric tertiary hospital were matched to patients treated by injections on the basis of age, duration of diabetes and HbA1c at the time of pump start. HbA1c, anthropometric data, episodes of severe hypoglycaemia and rates of hospitalisation for DKA were collected prospectively. Results A total of 345 patients on pump therapy were matched to controls on injections. The mean age, duration of diabetes at pump start and length of follow-up were 11.4 (± 3.5), 4.1 (± 3.0) and 3.5 (± 2.5) years, respectively. The mean HbA1c reduction in the pump cohort was 0.6% (6.6 mmol/mol). This improved HbA1c remained significant throughout the 7 years of follow-up. Pump therapy reduced severe hypoglycaemia from 14.7 to 7.2 events per 100 patient-years (p < 0.001). In contrast, severe hypoglycaemia increased in the non-pump cohort over the same period from 6.8 to 10.2 events per 100 patient-years. The rate of hospitalisation for DKA was lower in the pump cohort (2.3 vs 4.7 per 100 patient-years, p = 0.003) over the 1,160 patient-years of follow-up. Conclusions/interpretation This is the longest and largest study of insulin pump use in children and demonstrates that pump therapy provides a sustained improvement in glycaemic control, and reductions of severe hypoglycaemia and hospitalisation for DKA compared with a matched cohort using injections.