Eighty osteoporotic, postmenopausal women, 50-70 years of age, with ongoing estrogen therapy (HRT), were randomized to recombinant human growth hormone (GH), 1.0 U or 2.5 U/day, subcutaneous, versus placebo. This study was double-blinded and lasted for 18 months. The placebo group then stopped the injections, but both GH groups continued for a total of 3 years with GH and followed for 5 years. Calcium (750 mg) and vitamin D (400 U) were given to all patients. Bone mineral density and bone mineral content were measured with DXA. At 18 months, when the double-blind phase was terminated, total body bone mineral content was highest in the GH 2.5 U group (p = 0.04 vs. placebo). At 3 years, when GH was discontinued, total body and femoral neck bone mineral content had increased in both GH-treated groups (NS between groups). At 4-year follow-up, total body and lumbar spine bone mineral content increased 5% and 14%, respectively, for GH 2.5 U (p = 0.01 and p = 0.0006 vs. placebo). Femoral neck bone mineral density increased 5% and bone mineral content 13% for GH 2.5 U (p = 0.01 vs. GH 1.0 U). At 5-year follow-up, no differences in bone mineral density or bone mineral content were seen between groups. Bone markers showed increased turnover. Three fractures occurred in the GH 1.0 U group. No subjects dropped out. Side effects were rare. In conclusion, bone mineral content increased to 14% with GH treatment on top of HRT and calcium/vitamin D in postmenopausal women with osteoporosis. There seems to be a delayed, extended, and dose-dependent effect of GH on bone. Thus, GH could be used as an anabolic agent in osteoporosis.