This study tested whether moderate resistance training would improve femoral bone mineral density (BMD) in long-term users of hormone therapy with low BMD. The study was a 2-year randomized, controlled, trial (RCT) of moderate resistance training of either the lower extremity or the upper extremity. Eighty-five women participated in a 6-month observation period. The setting was center-based and home-based training. The participants were 189 women aged 59-78 years, with total femur T-scores from -0.8 to -2.8 and on hormone therapy (HT) for a minimum of 2 years (mean 11.8 years); 153 completed the trial. Lower extremity training used weight belts (mean 7.8 kg) in step-ups and chair rises; upper extremity training used elastic bands and dumbbells. Measurements were BMD and body composition [dual-energy X-ray absorptiometry (DXA)], bone turnover markers. Total femoral BMD showed a downward trend during the observation period: 0.35%+/-0.18% (P=0.14). The response to training was similar in the upper and lower groups in the primary outcomes. At 2 years, total femoral BMD increased 1.5% (95% CI 0.8%-2.2%) in the lower group and 1.8% (95% CI 1.1%-2.5%) in the upper group. Trochanter BMD increased 2.4% (95% CI 1.3%-3.5%) in the lower group and 2.5% (95% CI 1.4%-3.6%) in the upper group (for both analyses time effect P<0.001). At 1 year, a bone resorption marker (C-telopeptide) decreased 9% (P=0.04). Bone formation markers, bone-specific alkaline phosphatase, decreased 5% (P<0.001), and N-terminal type I procollagen peptide decreased 7% (P=0.01). Body composition (percent lean and percent body fat) was maintained in both groups. We concluded that long-term moderate resistance training reversed bone loss, decreased bone turnover, increased femur BMD, and maintained body composition. The similarity of response in upper and lower groups supports a systemic response rather than a site-specific response to moderate resistance training.