Patients with chronic thyroiditis were selected for study according to specific selection criteria, and comparisons of an enzyme-linked immunosorbent assay (ELISA) method to a hemagglutination method were performed. The ELISA assays contained antithyroglobulin (TG) antibody and antithyroid-peroxidase (PO) antibody, the latter using recombinant human TPO. Four groups were established from 437 patients with chronic thyroiditis; 316 with both antibodies (TG + PO group), 75 with TG antibody alone (TG group), 10 with PO antibody alone (PO group), and 36 with no such antibodies (N group). The hemagglutination assay resulted in much different antibody profiles: PO group based on ELISA occurred in only 2%, whereas a group with microsome antibody alone based on hemagglutination assay occurred in 39%. Subsequently, studies of disease characteristics including age, gender, familial predisposition, thyroid function, and morphology of the gland were performed among the four groups of patients based on ELISA profiles but not on hemagglutination results. The subgroups of TG + PO and TG were derived from randomly selected patients (40 out of a much larger number of patients in each). The results showed that the PO group had smaller thyroid volume (25+/-16 mL, mean +/- SD) with normal echogenicity, and 50% prevalence of hypothyroidism. The TG group had larger thyroid volume (57+/-42 mL) with frequent association of small nodular formation (53%) but less frequent hypothyroidism (23%). In summary, compared to the hemagglutination method, the ELISA was noted to have both a higher sensitivity and specificity for detection of chronic thyroiditis. Correlation of ELISA profiles with ultrasonography of the thyroid gland detected subtle differences in subgroups that may account for differences in thyroid gland morphology and prevalence of hypothyroidism: the PO group has normal echogenicity but high incidence of impaired thyroid function, whereas in the TG group small nodules were associated with a lesser incidence of hypothyroidism.