Bone mass measured by photon absorptiometry: comparison of forearm, heel, and spine.

We assessed bone mass quantification at different skeletal sites by single and dual photon absorptiometric (SPA and DPA) methods. Improved DPA measurement of spinal bone mineral density in young healthy subjects showed a short- and long-term precision of 1.2% and 1.6%. Compared to the conventional DPA method the imprecision was reduced by more than 50%. The appendicular measurements were more precise (0.5-1.2%). We present the intercorrelations and predictive errors between peripheral measurements and improved spinal and total body bone measurements in early postmenopausal women (n = 144) recruited in 1988. To compare the improved system with the conventional methods, we retrieved data on age-matched early postmenopausal women (n = 151) recruited in 1983. In the 1988 population all peripheral methods had similar predictive errors in estimation of spinal bone mineral density (SEE = 11-13%) and total body bone mineral density (SEE = 4-5%). Measurement of trabecular bone in the heel and distal forearm did not improve the validity of predicting spinal bone mass. In a cadaver study (n = 11) the predictive error in estimation of the spinal ash weight from forearm measurements was of the same magnitude (15%) as that in estimation of the spinal BMC from the forearm (16%). We conclude that the predictive error in estimation of spinal bone mass from peripheral bone measurements is more likely to be caused by intra-skeletal variation in bone mass than by precision errors. However, this does not reflect the inability of peripheral bone mass measurements to predict fracture risk.