Proinsulin Immunoreactivity in Recent-Onset IDDM: The significance of insulin antibodies and insulin autoantibodies

Abstract

OBJECTIVE To study the natural history of fasting proinsulin immunoreactivity (PIM) during the first 30 months of IDDM and its relationship to fasting C-peptide and insulin antibodies. RESEARCH DESIGN AND METHODS An incidence cohort of 204 consecutive newly diagnosed IDDM patients were followed prospectively, having blood drawn for measurements at diagnosis and at 1, 3, 6, 9, 12, 18, 24, and 30 months. A sensitive enzyme-linked immunosorbent assay was used for the determination of PIM. RESULTS All patients had detectable fasting PIM in plasma at diagnosis, with a median value and interquartile range of 3.5 pmol/l (2.2–6.2). The median PIM level increased during the first months of IDDM to reach a peak at 9–12 months (9.9–10.3 pmol/l). PIM then declined gradually to 5.6 pmol/l (1.9–13.5) at 30 months without reaching baseline. PIM at each time point was widely scattered in a skewed log-normal distribution without signs of bimodality. After the onset of insulin treatment, median insulin antibody level increased and declined in a similar pattern. Both PIM and antibody level were significantly higher in children and adolescents compared with adults. However, stepwise multiple regression analysis showed that age was only of minor importance for the PIM variation during the study period. Insulin antibody level and fasting C-peptide were the major determinants at 3–30 months, accounting for ∼40% of the variation (R2). Blood glucose was of minor importance, and insulin dose, HbA1c, and BMI were of no importance. The correlation between fasting PIM and fasting C-peptide improved (R2 doubled) if the insulin antibody level was accounted for. Further, the slope of the correlation curve between PIM and C-peptide increased threefold when antibody binding was > 4%. At diagnosis, insulin autoantibodies could be detected in 19% of the patients. Their presence predicted higher proinsulin at 1–3 months, a higher insulin dose the 1st year, and higher levels of insulin antibodies later in the study. CONCLUSIONS Circulating insulin antibodies may affect the level of PIM in IDDM, probably by adding a pool of IgG-bound PIM thereby increasing half-life and plasma concentration. This may explain why C-peptide and PIM levels do not change in concert during the 1st years of IDDM. Unlike C-peptide, PIM can not therefore quantitate β-cell secretion unless the presence of insulin antibodies is ruled out.