To evaluate the effect of alendronate combined with hormone replacement therapy (HRT) on postmenopausal osteoporotic Chinese women living in Taiwan, we treated 151 women (age range, 47-70 years; mean, 61 years) with conjugated equine estrogen (0.625 mg), medroxyprogesterone 5 mg, and elemental calcium 500 mg daily with either alendronate 10 mg (n = 79) or placebo (n = 72), and measured their bone mineral density (BMD) at the lumbar spine and hip every 6 months for 3 years. Urine N-telopeptide of type I collagen corrected by concentration of urine creatinine (NTx/Cr) and serum osteocalcin (OC) concentration was also measured at weeks 2, 4, and every 3 months from month 3 for 2 years. Significantly higher percentage increases in BMD at the lumbar spine (P < .0001, 2-way analysis of variance) throughout the 36-month treatment period were found in the alendronate plus HRT group than in the HRT-only group. However, there was no difference in BMD at the femoral neck and trochanter between these 2 groups. Treatment with alendronate plus HRT resulted in a 10.1% increase at the L-spine BMD and a 7.7% increase at the trochanter BMD at the end of the 3-year study period (P < .01, compared with baseline at both sites). A significant decline in urine NTx/Cr was observed at week 4 in the alendronate plus HRT group, whereas in the HRT-only group, a significant decline in urine NTx/Cr occurred at month 9. By the end of 24 months, urine NTx/Cr decreased by 49.7% in the alendronate plus HRT group (P = .001 compared with a 20.4% increase in the HRT group). A significant decline in serum OC level occurred at month 3 in the alendronate plus HRT group, whereas a similar decline was observed at month 6 in the HRT-only group. By the end of 24 months, serum OC decreased by 52.2% in the alendronate plus HRT group (P < .001 compared with a 1.5% increase in the HRT-only group). Subjects treated with alendronate plus HRT had a significantly greater percentage decrease in urine NTx/Cr (P = .0001) and serum OC (P = .0007) than subjects treated with HRT only throughout the 24-month treatment period by 2-way analysis of variance comparison. There was no difference in upper gastrointestinal or drug-related side effects between groups. In conclusion, our data suggest that the use of alendronate combined with HRT for 3 years was well tolerated and it significantly increased BMD at the L-spine and hip in postmenopausal Chinese women with osteoporosis. This regimen is safe and can be used in subjects who have no satisfactory response to a single agent or who have very low BMD with multiple risks. However, this study does not indicate whether HRT plus alendronate has any greater effect on BMD than alendronate alone.