David J. Torgerson, PhD; Sally E. M. Bell-Syer, MSc

JAMA. 2001;285:2891-2897.

Context Hormone replacement therapy (HRT) is widely considered to reduce fractures, but this belief is based on observational data; evidence from randomized trials is lacking.

Objective To conduct a systematic review of all randomized trials of HRT that have reported or collected nonvertebral fracture data but that may not have focused on fracture prevention.

Data Sources The MEDLINE, EMBASE, Science Citation Index, and Cochrane Controlled Trials Register databases were searched from 1997 through 2000 and a search was conducted of all recent systematic reviews to identify older studies. Authors were contacted to establish whether fracture data had been collected but not reported. Researchers in the field and pharmaceutical companies also were contacted to try to identify unpublished studies.

Study Selection Trials were included in which participants had been randomized to at least 12 months of therapy and data on nonvertebral fractures at any other site and due to any cause were available. Of 70 initially identified studies, 22 were included in the analysis.

Data Extraction Both investigators extracted data independently and appraised trial quality according to the Jadad scale, which assesses the methods of randomization, concealment allocation, and reporting of withdrawals and dropouts. Disagreements were resolved by discussion.

Data Synthesis There was an overall 27% reduction in nonvertebral fractures in a pooled analysis (reduction favoring HRT in relative risk [RR], 0.73; 95% confidence interval [CI], 0.56-0.94; P = .02). This effect was greater among women randomized to HRT who had a mean age younger than 60 years (RR, 0.67; 95% CI, 0.46-0.98; P = .03). Among women with a mean age of 60 years or older, there was a reduced effect (RR, 0.88; 95% CI, 0.71-1.08; P = .22). For hip and wrist fractures alone, the effectiveness of HRT appeared more marked (RR, 0.60; 95% CI, 0.40-0.91; P = .02), particularly for women younger than 60 years (RR, 0.45; 95% CI, 0.26-0.79; P = .005).

Conclusions Our meta-analysis of randomized controlled trials of HRT noted a statistically significant reduction in nonvertebral fractures. However, this effect may be attenuated in older women.

Author Affiliations: Department of Health Studies and Centre for Health Economics, University of York, Heslington, England.