Pediatr. Res. 15: 62-65 (1981) Twenty-one children were studied for at least 2 years after the diagnosis of diabetes. Blood was obtained at the time of diagnosis for the determination of ICA and bistocompatability status. ICA status was determined 12 months after diagnosis, and again at 24 months if ICA had been detected at 12 months. Residual /hell function was assessed serially by measuring 24-hr urinary Cpeptide. The 24-hr C peptide excretion was expressed relative to 24-hr creatinine excretion. Tbis ratio is termed urinary C peptide excretion (UCP). Islet cell antibodies were measured by indirect immunofluorescence on human 0-group pancreas. Fluoresceinlabeled anti-Csc was used to determine whether the islet cell antibodies were able to fix complement. A standard NIH microlymphocytotoxicity test was used to determine bistocompatability status at the A, B, and C loci.Throughout the entire study period, the UCP of each diabetic was always below the mean of the group of nondiabetic children.For all diabetic children, the UCP decreased during the 24 months after diagnosis. Most children's (14 of 21) UCP bad fallen below 0.20 during the 12 months after diagnosis. Twelve of 21 diabetic children were ICA positive at diagnosis. Five had unaltered titers of ICA 12 months later and four had no ICA detectable at 12 months, whereas for three the titer had fallen by at least two dilutions. The magnitude of the UCP 12 months after diagnosis correlated significantly with age at the time of diagnosis (r = 0.557; P < 0.05). Similarly, there was also a significant correlation between the UCP 24 months after diagnosis and age at diagnosis Ir = 0.446. P < 0.05). There was no sidficant difference between ksulin do&, expressed as units/kg/d& at 12 and 24 months after diarmosis. Tbere was no simificant relations hi^ between ICA status and insulin dose at 120r 24 months. herd was no discernible relationship between the evolution of ICA or UCP with bistocompatability antigens at the A, B, or C loci.