BACKGROUND: Results of studies in animals and human beings suggest that type 1 diabetes is preventable. Nicotinamide prevents autoimmune diabetes in animal models, possibly through inhibition of the DNA repair enzyme poly-ADP-ribose polymerase and prevention of beta-cell NAD depletion. We aimed to assess whether high dose nicotinamide prevents or delays clinical onset of diabetes in people with a first-degree family history of type 1 diabetes. METHOD: We did a randomised double-blind placebo-controlled trial of nicotinamide in 552 relatives with confirmed islet cell antibody (ICA) levels of 20 Juvenile Diabetes Federation (JDF) units or more, and a non-diabetic oral glucose tolerance test. Participants were recruited from 18 European countries, Canada, and the USA, and were randomly allocated oral modified release nicotinamide (1.2 g/m2) or placebo for 5 years. Random allocation was done with a pseudorandom number generator and we used size balanced blocks of four and stratified by age and national group. Primary outcome was development of diabetes, as defined by WHO criteria. Analysis was done on an intention-to-treat basis. FINDINGS: There was no difference in the development of diabetes between the treatment groups. Of 159 participants who developed diabetes in the course of the trial, 82 were taking nicotinamide and 77 were on placebo. The unadjusted hazard ratio for development of diabetes was 1.07 (95% CI 0.78-1.45; p=0.69), and the hazard ratio adjusted for age-at-entry, baseline glucose tolerance, and number of islet autoantibodies detected was 1.01 (0.73-1.38; p=0.97). Of 168 (30.4%) participants who withdrew from the trial, 83 were on placebo. The number of serious adverse events did not differ between treatment groups. Nicotinamide treatment did not affect growth in children or first-phase insulin secretion. INTERPRETATION: Large-scale controlled trials of interventions designed to prevent the onset of type 1 diabetes are feasible, but nicotinamide was ineffective at the dose we used.