The general population risk of developing type 1 diabetes mellitus (DM), 1/300, is magnified 15-20 fold in first-degree relatives of affected individuals. Because a combination of immunologic, metabolic, and genetic markers can be used to predict the disease, multicenter prevention trials in the US (DPT-1) and Europe (ENDIT) were initiated in relatives. In the DPT-1 over 80,000 relatives under 45 years of age will be screened for ICA and then'staged'to assess risk. High-risk subjects (> 50% over 5 yr) are randomized either to 4 days intravenous insulin infusion annually followed by bid low doses of subcutaneous ultralente insulin, or to close observation. To date (September 2000), 331/340 (97%) high-risk subjects have been enrolled with the intention of detecting a 35% decrease in disease over 5 years (80% power). 280/490 (57%) of intermediate risk subjects (25-50% over 5 yr) have been randomized to oral insulin or placebo. A 50% treatment difference is sought. Anticipated enrolment for the high-risk arm will be completed by year 2001, and by 2003 for the oral arm. The ENDIT study will prospectively address whether nicotinamide will reduce the rate of progression to DM in relatives. 40,000 first-degree relatives (5-40 yr) have been screened with 552 subjects (ICA titers> 20 JDF U) randomized to nicotinamide or