This is a retrospective study of 15 postmenopausal or amenorrheic women aged 34–78 years who had taken prednisone for 6–108 months and were followed for 1 year while continuing to take doses of 5–15 mg/day. A total of 8 patients were treated with 0.6256 mg Premarin daily for 25 days and 5 mg/day of medroxyprogesterone on days 15–25 (ERT, group 2); 7 were followed without ERT (group 1). A group of 17 women, matched for age, were randomly selected from our computerized data base to serve as a control group (group 3), and 10 women of similar age who were taking ERT only (group 4) were selected to compare the response to ERT to that of group 2. Bone density (BD) was measured in the lumbar spine baseline and at 1 year using dual-photon or dual-energy x-ray absorptiometry. Spine density did not change significantly during the year of observation in group 1. Although BD decreased in 5 of 7 patients, the change was not significant (-0.034 ± 0.018 g/cm2, p = 0.10). In group 2 BD increased significantly, with 7 of 8 patients showing an increase (0.037 ± 0.011 g/cm2, p = 0.008). BD did not change significantly in the control group (0.013 ± 0.008 g/cm2, p = 0.16). Loss of bone from the spine was significantly greater in group 1 than in controls (p = 0.02), but changes in group 2 were similar to those in the control group (p = 0.66). Changes in BD in group 2 were similar to those in group 4; that is, the response to ERT of patients taking prednisone was comparable to that of women taking ERT only. The change in spine density correlated with the cumulative dose of prednisone taken during the year of observation in the group not taking ERT (r = 0.759, p = 0.05) but not in the group taking ERT (r = 0.134, p = 0.73). This suggests that ERT protects against bone loss for at least 1 year throughout the dosage range studied. We concluded that ERT prevents vertebral bone loss for at least 1 year in the majority of women taking prednisone chronically in doses of 5–15 mg/day.