For the first time the incidence of insulin autoantibodies and islet cell antibodies were evaluated in a prospective study from birth. Consecutive neonates (168) from mothers with Type 1 (insulin-dependent) diabetes mellitus (n = 113) and gestational diabetes (n = 55) were included at birth. To date, follow-up sera were obtained from 90 of 168 mother-child-pairs 9 months postpartum and from 39 of 168, 2 years postpartum. At birth, there was a strong correlation between the presence of antibodies in the cord blood of neonates and in maternal circulation [Type 1 diabetic mothers: 20% islet cell antibodies > or = 20 JDF-U (detection threshold of our islet cell antibody assay), 74% insulin antibodies > 49 nU/ml (upper limit of normal range in sera of healthy control subjects aged 0.5 to 46 years); neonates: 21% islet cell antibodies > or = 20 JDF-U, 76% insulin antibodies > 49 nU/ml; gestational diabetic mothers: 11% islet cell antibodies > or = 20 JDF-U, 18% insulin antibodies > 49 nU/ml; neonates: 13% islet cell antibodies > or = 20 JDF-U, 55% insulin antibodies > 49 nU/ml]. This supports transplacental passage of insulin antibodies and islet cell antibodies from diabetic mothers to their offspring. During follow-up, the majority of children lost antibody-positivity after birth. A few offspring, however, exhibited or developed antibodies consistently, whereby insulin autoantibodies preceded islet cell antibodies in each case (antibody-positivity: 9 months: 0% islet cell antibody positive, 3.3% insulin autoantibody positive; 2 years: 2.6% islet cell antibody positive, 7.7% insulin autoantibody positive).(ABSTRACT TRUNCATED AT 250 WORDS)