A newly developed immunoprecipitation assay, with 125I-labeled highly purified human placental insulin-like growth factor I (IGF-I) receptor, was used to search for IGF-I-receptor antibodies in human sera. Eleven of 141 patient sera tested (7.8%) immunoprecipitated labeled IGF-I receptor. Immunoprecipitation was comparable with sera and IgG prepared from these sera. Seven of the 11 sera (3 of 31 with rheumatic disorders, 3 of 48 with non-insulin-dependent diabetes, and 1 of 52 with insulin-dependent diabetes) failed to inhibit IGF-I binding to human placental membranes and thus contained non-binding-inhibitory IGF-I-receptor antibodies. Their pathophysiological function remained uncertain. The remaining 4 sera (2 of 3 with type B severe insulin resistance, 1 of 7 with polycystic ovary syndrome (PCO), and 1 of 31 with rheumatic disorders) inhibited IGF-I binding. Plasma IGF-I concentrations were elevated (663 and 802 ng/ml, respectively) in 2 patients (1 with PCO and another with systemic lupus erythematosus) with binding-inhibitory IGF-I-receptor antibodies, suggesting IGF-I resistance that was probably mediated by the IGF-I-receptor antibodies. In conclusion, we identified two species of human IGF-I-receptor antibodies. Sera from 7 of 141 patients tested contained IgG autoantibodies that bound to the IGF-I receptor at a locus different from the IGF-I binding site and did not inhibit IGF-I binding. Sera from 4 of 141 patients contained antibodies that bound to the IGF-I receptor at or near the IGF-I binding site, inhibited IGF-I binding, and probably caused IGF-I resistance.