The aim of the study was to evaluate the immunogenicity of different commercial recombinant-growth hormone preparations. The presence of antibodies to growth hormone was tested in 210 growth hormone-deficient children at 6-month intervals during treatment for 6-66 months. The patients were treated with three preparations (groups A, B and C of 70 cases each) having the authentic growth hormone sequence. Groups A and B received hormone synthesized by the recombinant DNA technique in E. coli, while the group C preparation was produced in a mammalian cell line. The preparations showed poor immunogenicity and antibodies were found as follows: 1.4% in patients of group A (1 case: binding capacity 0.2mg/l and Ka 3.5 107 1 M-1), 2.8% in patients of group B (2 cases; case 1 binding capacity 0.7mg/l and Ka 1.5 1071M-1; case 2 binding capacity 0.04mg/1 and Ka of 1.8 108 and 6.5 106 1 M-1), and 8.5% in group C (6 cases; binding capacity from 0.4 to less than 0.02 mg/ 1, Ka from 1.6 107 to 3.8 108 1 M-1). Only two patients of group C presented the antibodies in two subsequent examinations; in the other patients the positivity was found once. In all patients positive samples were found at intervals of 6-24 months after the start of therapy. In all antibody-positive patients growth velocity presented no decrease at the time of antibody detection and was never different to that of negative patients. We conclude that the three commercial preparations examined showed poor immunogenicity without clinical relevance.