BACKGROUND: Islet transplantation is indicated for patients with type 1 diabetes with severe hypoglycaemia or after kidney transplantation. We did a randomised trial to assess the efficacy and safety of islet transplantation compared with insulin therapy in these patients. METHODS: In this multicentre, open-label, randomised controlled trial, we randomly assigned (1:1) patients with type 1 diabetes at 15 university hospitals to receive immediate islet transplantation or intensive insulin therapy (followed by delayed islet transplantation). Eligible patients were aged 18-65 years and had severe hypoglycaemia or hypoglycaemia unawareness, or kidney grafts with poor glycaemic control. We used computer-generated randomisation, stratified by centre and type of patient. Islet recipients were scheduled to receive 11 000 islet equivalents per kg bodyweight in one to three infusions. The primary outcome was proportion of patients with a modified beta-score (in which an overall score of 0 was not allocated when stimulated C-peptide was negative) of 6 or higher at 6 months after first islet infusion in the immediate transplantation group or 6 months after randomisation in the insulin group. The primary analysis included all patients who received the allocated intervention; safety was assessed in all patients who received islet infusions. This trial is registered with ClinicalTrials.gov, number NCT01148680, and is completed. FINDINGS: Between July 8, 2010, and July 29, 2013, 50 patients were randomly assigned to immediate islet transplantation (n=26) or insulin treatment (n=24), of whom three (one in the immediate islet transplantation group and two in the insulin therapy group) did not receive the allocated intervention. Median follow-up was 184 days (IQR 181-186) in the immediate transplantation group and 185 days (172-201) in the insulin therapy group. At 6 months, 16 (64% [95% CI 43-82]) of 25 patients in the immediate islet transplantation group had a modified beta-score of 6 or higher versus none (0% [0-15]) of the 22 patients in the insulin group (p<0.0001). At 12 months after first infusion, bleeding complications had occurred in four (7% [2-18]) of 55 infusions, and a decrease in median glomerular filtration rate from 90.5 mL/min (IQR 76.6-94.0) to 71.8 mL/min (59.0-89.0) was observed in islet recipients who had not previously received a kidney graft and from 63.0 mL/min (55.0-71.0) to 57.0 mL/min (45.5-65.1) in islet recipients who had previously received a kidney graft. INTERPRETATION: For the indications assessed in this study, islet transplantation effectively improves metabolic outcomes. Although studies with longer-term follow-up are needed, islet transplantation seems to be a valid option for patients with severe, unstable type 1 diabetes who are not responding to intensive medical treatments. However, immunosuppression can affect kidney function, necessitating careful selection of patients. FUNDING: Programme Hospitalier de Recherche Clinique grant from the French Government.