Growth hormone replacement in adults and bone mineral density: a systematic review and meta-analysis

BACKGROUND The effect of GH replacement on bone mineral density (BMD) in adults with GH deficiency (GHD) is uncertain. We carried out a systematic review of randomized trials that compared GH to no active treatment, with BMD as an outcome. METHODS We searched electronic databases to identify articles, abstracts and conference proceedings to March 2002. We also checked reference lists in included studies and expert reviews. Two reviewers independently extracted the data on study design and change in BMD. The results of individual trials were combined by fixed effects model meta-analysis using weighted mean difference (WMD) of change in BMD at the lumbar spine (our primary outcome) and other sites. FINDINGS Eighteen trials that included 700 patients met the inclusion criteria. Maximum follow-up was for 12 weeks (1 trial), 6 months (14 trials), 12 months (1 trial), 18 months (1 trial) and 24 months (1 trial). Reporting quality of both study design and results was poor. Ten trials (458 subjects) were included in the meta-analysis. We excluded those eight trials from which sufficient data could not be extracted. We found a mean change in BMD, at the lumbar spine with GH treatment, of 0.01 g/cm(2) after 6 and 12 months, 0.02 g/cm(2) after 18 months and 0.03 g/cm(2) after 24 months. Statistical significance at the 0.05 level was just achieved at 6 and 12 months but was significant at 18 and 24 months. These changes are small and may be influenced by bias. CONCLUSION There is evidence of a small effect of GH replacement on bone mineral density in adults with GH deficiency. The clinical importance of this is uncertain.