We have reported 1-yr results of a double blind, placebo-controlled trial of recombinant human insulin-like growth factor I (rhIGF-I) replacement in 16 children from the Ecuadorian GH receptor-deficient (GHRD) population. This report extends observations of rhIGF-I efficacy at two dosage levels [120 micrograms/kg BW twice daily (n = 15) and 80 micrograms/kg twice daily (n = 7)] over 2 yr, compares biochemical responses [serum IGF-I and IGF-binding protein-3 (IGFBP-3)] and their relationship to growth effects, and compares treatment effects of rhIGF-I in GHRD to rhGH in idiopathic GH deficiency (GHD). There were no baseline differences between the low and high dose groups for growth velocity (GV), bone age (BA), SD score for height, or percent mean body weight for height (MBWH). Over 2 yr of rhIGF-I treatment, there were no differences in GV or in changes in height SD score, height age (HA), or BA between the two groups; a subgroup of six subjects at the higher dose followed for a third year continued at the second year GV. The higher dose resulted in a greater change in percent MBWH. GV in yr 1 and 2 for the entire group and in yr 3 for a subgroup were greater for GH-treated GHD (n = 11). The GHD group showed a greater change in SD score for height and HA, but did not differ from the rhIGF-I-treated GHRD group in the change in BA (delta BA) or delta HA/delta BA over 2 yr. There was a greater change in percent MBWH in GHRD. There were no differences between dosage groups for serum IGF-I levels at baseline or the near-normal trough levels 12 h after rhIGF-I injection; these individual levels correlated with HA gain in yr 1 and 2. IGFBP-3 levels were markedly low, with no changes of significance with treatment. Comparable growth responses to the two dosage levels and the biochemical changes indicate a plateau effect at or below 80 micrograms/kg BW twice daily. The growth response and favorable trough levels of IGF-I despite the overall lack of increase in circulating IGFBP-3 levels suggest an alternative mechanism for sustaining IGF-I levels and avoiding rapid clearance of rhIGF-I. The greater increase in MBWH with treatment of GHRD than with treatment of GHD may reflect comparable effects on lean body mass without the lipolytic effects of GH in the GHRD subjects. The difference in growth response between rhIGF-I-treated GHRD and rhGH-treated GHD groups is consistent with the hypothesis that 20% or more of GH-influenced growth is due to the direct effects of GH on bone. Nonetheless, the comparable delta HA/delta BA suggests similar long term effects of replacement therapy in the two conditions.