BACKGROUND: Intensive basal-bolus insulin therapy has been shown to improve glycaemic control and reduce the risk of long-term complications that are associated with type 1 diabetes mellitus. Insulin degludec is a new, ultra-longacting basal insulin. We therefore compared the efficacy and safety of insulin degludec and insulin glargine, both administered once daily with mealtime insulin aspart, in basal-bolus therapy for type 1 diabetes. METHODS: In an open-label, treat-to-target, non-inferiority trial, undertaken at 79 sites (hospitals and centres) in six countries, adults (aged ≥18 years) with type 1 diabetes (glycated haemoglobin [HbA(1c)] ≤10% [86 mmol/mol]), who had been treated with basal-bolus insulin for at least 1 year, were randomly assigned in a 3:1 ratio, with a computer-generated blocked allocation sequence, to insulin degludec or insulin glargine without stratification by use of a central interactive response system. The primary outcome was non-inferiority of degludec to glargine, assessed as a reduction in HbA(1c) after 52 weeks, with the intention-to-treat analysis. This trial is registered with ClinicalTrials.gov, number NCT00982228. FINDINGS: Of 629 participants, 472 were randomly assigned to insulin degludec and 157 to insulin glargine; all were analysed in their respective treatment groups. At 1 year, HbA(1c) had fallen by 0*40% points (SE 0*03) and 0*39% points (0*07), respectively, with insulin degludec and insulin glargine (estimated treatment difference -0*01% points [95% CI -0*14 to 0*11]; p<0*0001 for non-inferiority testing) and 188 (40%) and 67 (43%) participants achieved a target HbA(1c) of less than 7% (<53 mmol/mol). Rates of overall confirmed hypoglycaemia (plasma glucose <3*1 mmol/L or severe) were similar in the insulin degludec and insulin glargine groups (42*54 vs 40*18 episodes per patient-year of exposure; estimated rate ratio [degludec to glargine] 1*07 [0*89 to 1*28]; p=0*48). The rate of nocturnal confirmed hypoglycaemia was 25% lower with degludec than with glargine (4*41 vs 5*86 episodes per patient-year of exposure; 0*75 [0*59 to 0*96]; p=0*021). Overall serious adverse event rates (14 vs 16 events per 100 patient-years of exposure) were similar for the insulin degludec and insulin glargine groups. INTERPRETATION: Insulin degludec might be a useful basal insulin for patients with type 1 diabetes because it provides effective glycaemic control while lowering the risk of nocturnal hypoglycaemia, which is a major limitation of insulin therapy. FUNDING: Novo Nordisk.