Our objective was to evaluate the efficacy of recombinant human growth hormone (GH) on bone mineral density (BMD) in persons age 50 and older, with normal pituitary function, with or at risk for developing osteoporosis. We systematically reviewed randomized clinical trials (RCTs), searching six databases, and conducted meta-analyses to examine GH effects on BMD of the lumbar spine and femoral neck. Data for fracture incidence, bone metabolism biomarkers, and adverse events were also extracted and analysed. Thirteen RCTs met the eligibility criteria. Pooled effect sizes suggested no significant GH effect on BMD. Pooled effect sizes were largest, but nonsignificant, when compared to placebo. GH had a significant effect on several bone metabolism biomarkers. A significantly higher rate of adverse events was observed in the GH groups. Meta-analysis of RCTs suggests that GH treatment for persons with or at risk for developing osteoporosis results in very small, nonsignificant increases in BMD.