In order to test the impact of a given risk profile on the incidence of osteoporosis which could justify BMD measurement, and that of a low risk profile which could render it unnecessary, BMD was measured in 217 women under 72 in whom menopause had occurred at least 6 years previously and who corresponded to one of the two following profiles: high risk (A, n = 102) = BMI < 27 kg/m2, with no estrogen replacement treatment, and with at least one of the following risk factors: BMI < 20, early menopause, positive family history, no dairy products associated with tobacco consumption (> 10 cigarettes/day for > 20 years and/or alcohol consumption of > 0.5 l wine/day during > 10 years, corticotherapy of > 6 months, rickets, anorexia nervosa. Low risk (B, n = 115) = absence of characteristics of group A, BMI > 27 kg/m2 with (B+, n = 24) or without estrogen therapy (B-, n = 91). BMD was measured by DXA in 4 centers using Lunar or Hologic equipment. Results were expressed in % of the mean of the respective young adult control groups. As expected, BMD was significantly different in these two subgroups of the population. Osteoporosis was diagnosed (BMD < 75% = < -2.5 SD, according to WHO) in 72% of group A, and in 17% (B+) and 19% (B-) respectively of group B. There was no difference between the various risk factors in group A concerning their impact on BMD, but concerning incidence, low BMI and early menopause were the most frequent. The high risk profile of group A seems to justify densitometry, since it leads to the diagnosis of osteoporosis in over 70%. However, the protective profile of group B does not exclude osteoporosis (risk still 20%); only in severe obesity (BMI > 33) does it drop to 1%.