Aims/hypothesis. The aim of the trial was to compare the efficacy and tolerability of two types of basal-bolus therapy, using either the soluble long-acting basal insulin analogue, insulin detemir, in combination with the rapid-acting analogue, insulin aspart, or NPH insulin in combination with mealtime regular human insulin. Methods. In this 18-week, 1:1 randomised, open-labelled, parallel trial, 595 patients with Type 1 diabetes mellitus received insulin detemir or NPH insulin in the morning and at bedtime in combination with mealtime insulin aspart or regular human insulin respectively. Results. Glycaemic control with insulin detemir/insulin aspart was improved in comparison with NPH insulin/regular human insulin (HbA 1 c: 7.88% vs 8.11%; mean difference: −0.22% point [95% CI: −0.34 to −0.10]; p<0.001). Self-measured 8-point plasma glucose profiles differed between the groups (p<0.001), with lower postprandial plasma glucose levels in the insulin detemir/insulin aspart group. Within-person day-to-day variation in plasma glucose was lower with insulin detemir/insulin aspart than with NPH insulin/regular human insulin (SD: 2.88 vs 3.12 mmol/l; p<0.001). Risk of overall and nocturnal hypoglycaemia (23.00-06.00 hours) was, respectively, 21% (p=0.036) and 55% (p<0.001) lower in the insulin detemir/insulin aspart group than in the NPH insulin/regular human insulin group. Body weight (adjusted for baseline and change in HbA 1 c) was 1 kg lower with insulin detemir/insulin aspart than with NPH insulin/regular human insulin (p<0.001). Conclusions/interpretation. Basal-bolus therapy using insulin detemir/insulin aspart offers a better balance of control and tolerability than with NPH insulin/regular human insulin. The low variability and more physiological action profiles generated with these insulin analogues resulted in improved glycaemic control with lower risk of hypoglycaemia and no concomitant body weight increase.