Assessment of bone density at hip, spine, radius and calcaneus can predict fracture risk. This paper examines whether women with low bone mineral density (BMD) at the hip and spine can be identified by radial or calcaneal BMD assessment, thus enabling pre-selection of such women for further investigation. BMD in the lumbar spine (LS), femoral neck (FN), trochanter (FT) and Ward's area (FW) was measured by dual energy X-ray absorptiometry (DEXA). These measurements were compared with total (Qtot), trabecular (Qtrab), subcortical (Qscort) and cortical (Qcort) BMD of the ultradistal radius measured by peripheral quantitative computed tomography (pQCT), and ultrasound attenuation (BUA) and velocity (VOS) at the os calcis. Measurements were performed on 216 perimenopausal women aged between 45 and 55 years who attended a randomized osteoporosis screening programme. Correlations for pQCT and ultrasound with DEXA measurements were, at best, moderate (r = 0.05-0.53), being poorest for Qcort and VOS. Similar correlations were found between ultrasound and pQCT measurements (r = 0.05-0.31). None of the pQCT or ultrasound measurements predicted low DEXA measurements. 50-56% of women in the lowest quartile (QU4) of Qtot, Qtrab, Qscort and BUA were also in QU4 of LS; 45-57% were in QU4 of FT and FW, but only 22-33% were in QU4 of FN. To detect all women with osteopenia at FN or LS using pQCT or ultrasound, almost the entire population would have to be screened. In conclusion, pQCT and os calcis ultrasound measurements cannot successfully predict hip and spine osteopenia and could not be used to pre-select women for DEXA hip and spine assessment.