The aim of the study was to determine the appropriate dose of subcutaneous insulin to induce "beta-cell rest" without any hypoglycemic risk, as the first stop in the investigation of its potential effect in preventing or delaying clinical diabetes mellitus onset in high-risk subjects. Four subjects at high risk for type I diabetes mellitus (first-degree relatives, islet cell antibodies (ICA)-positive, and with diminished first-phase insulin secretion) were compared with four healthy individuals. After hospitalization, urinary C-peptide excretion (UCP) and 24-hour serum profiles for glucose were measured before and after administration of NPH insulin 0.1, 0.2, and 0.3 U x kg body weight per day subcutaneously in a single dose on 4 consecutive days. After insulin 0.1 U x kg body weight, a significant inhibition of endogenous insulin secretion was observed in high-risk subjects, but not in control subjects. There was no further inhibition when a higher insulin dose (0.2 and 0.3) was administered. A sustained beta-cell rest was obtained after 3, 6, and 12 months of treatment with 0.1 U x kg body weight per day as outpatient therapy in high-risk subjects. With this dose, no subject developed hypoglycemia (plasma glucose <50 mg/dL), whereas this adverse effect was detected after 0.2 and 0.3 U x kg body weight in both groups. In conclusion, our results indicate that administration of NPH insulin 0.1 U x kg bodyweight per day induces beta-cell rest without the undesirable effect of hypoglycemic episodes. This is a preliminary study to investigate the potential beneficial effect of insulin in preventing or delaying type I diabetes mellitus in subjects at high risk for the disease.