This randomised (1: 1), multinational, open-labelled, parallel group trial compared insulin detemir (IDet) with neutral protamine Hagedorn (NPH) insulin, in combination with mealtime insulin aspart, over 1 yr in subjects aged 2-16 yr with type 1 diabetes mellitus. Of 348 randomised subjects, 82 (23.6%) were 2-5 yr (IDet: 42, NPH: 40). This article is a descriptive subgroup analysis of these young children. Baseline characteristics (IDet vs. NPH) were similar: mean age, 4.3 vs. 4.5 yr; diabetes duration, 2.2 vs. 2.1 yr; males, 42.9 vs. 52.5%. Mean haemoglobin A1c (HbA1c) was similar between groups at baseline (8.2 vs. 8.1%), and changed little over 1 yr (8.1 vs. 8.3%). Fasting plasma glucose (FPG) was similar at baseline (8.44 vs. 8.56 mmol/L) and decreased during the study (-1.0 vs. -0.45 mmol/L). A lower rate of hypoglycaemia was observed with IDet compared with NPH (24-h; 50.6 vs. 78.3 episodes per patient-year; nocturnal hypoglycaemia, 8.0 vs. 17.4 episodes per patient-year). No severe hypoglycaemic episodes occurred with IDet, while 3 subjects reported 6 episodes with NPH. Change in weight standard deviation score standardised by age and gender was -0.17 with IDet and +0.03 with NPH. A slightly lower proportion of subjects in this age group reported adverse events with IDet than with NPH (69.0 vs. 77.5%). Serious adverse events were few (5 with IDet, 7 with NPH). In conclusion, long-term treatment with IDet in children aged 2-5 yr suggested similar glycaemic control, greater reduction in FPG, lower rates of hypoglycaemia, no inappropriate weight gain, and fewer adverse events compared with NPH.