Recent studies have linked autoimmunity to nervous tissue structures and diabetic autonomic neuropathy, but data on the early stage of IDDM and on the natural history of this association are not available. For this reason, we investigated autonomic nervous function, and the presence of autoantibodies to sympathetic and parasympathetic nervous structures, to glutamic acid decarboxylase (GAD) and tyrosine phosphatase (IA-2/ICA512) in 85 adolescents with insulin-dependent diabetes mellitus (IDDM) (mean age 14.7+/-1.6 yr, mean duration of diabetes 6.8+/-3.5 yr), and 45 age and sex-matched healthy subjects. Nervous tissues autoantibodies were detected using an indirect immunofluorescent complement-fixation technique, with monkey adrenal gland, rabbit cervical ganglia and vagus nerve as substrates. GAD and IA-2/ICA512 autoantibodies were detected by radioimmunoprecipitation assay. Seven patients (8%) had anti-vagus nerve autoantibodies, 7 other patients (8%) had anti-cervical ganglia autoantibodies, while all controls were negative (P < 0.05). Anti-adrenal medulla antibodies were detected in 16 patients (19%) and in 2 control subjects (P<0.02). None of the patients had autonomic symptoms. When patients were divided according to the presence or absence of autoantibodies, values of the cardiovascular tests (deep breathing, 30:15 ratio, Valsalva ratio) were similar in the two groups and similar to those in healthy subjects. However, when considered together, patients positive for one or more autoantibody showed a trend for lower values of deep breathing test and 30:15 ratio test, compared with healthy control subjects, which failed to reach conventional significance values (P=0.17 and P=0.07, respectively). No correlation was found between cardiovascular parameters and metabolic control or diabetes duration. There was no association between autoimmunity to nervous tissue structures and presence of GAD and IA-2/ICA512 Ab, and no correlation between these two autoantibodies and values of cardiovascular tests. Our data indicate that autonomic dysfunction is not a characteristic of young diabetic patients, but that autoantibodies against autonomic nervous structures are present during the first 1 to 15 yr of diabetes. GAD and tyrosine phosphatase appear to be excluded as target autoantigens within autonomic structures. Follow-up studies are required to evaluate future autonomic dysfunction and symptoms in these patients, and to establish whether the subtle autonomic dysfunction detected and/or the nervous tissue autoantibodies, are predictive of the development of this complication.