The aim of our study was to compare the results provided by the measurement of vertebral and femoral bone mineral density (BMD) for assessing the individual risk of osteoporosis as defined by either low BMD and/or rapid bone loss. Vertebral and femoral BMD were measured twice at a mean interval of 21 months in 85 normal, early post-menopausal women who had passed a natural menopause 6 months to 3 years previously. According to the measurement site, 36% (spine), 29% (femoral neck), 35% (Ward's triangle), and 25% (trochanter) fall in the “at risk” category, defined by a BMD value of 1 SD or more below the normal values for premenopausal women. Based on vertebral BMD, 39–48% of the women at risk had a normal femoral BMD. On the other hand, 24–37% of the women classified at risk based on femoral BMD maintained a low risk at the vertebral level. The annual rate of bone loss was significantly greater for the Ward's triangle (-2.7±3.8%) and femoral neck (-2.1±2.5%) than for the spine (-1.5±2.1%) and trochanter (-1.5±3.4%). There was a significant relationship between the rate of loss measured at the spine and femoral levels (r=0.34–0.58). Among the 21 women with a rapid vertebral bone loss, 48–67% had a low bone loss at the femoral level and vice versa. The ratio between mean rate of loss and the precision of the measurement sites was greater for the spine (1.6) compared with the femur (1.1–0.71). Our results indicate that vertebral and femoral BMD measurements produce discordant results in assessing the individual risk for osteoporosis.