Measurement of forearm bone mineral density: comparison of precision of five different instruments

Measurement of bone mineral density (BMD) is used for clinical estimation of fracture risk in osteoporosis. The precision of the method is important for the evaluation of true and clinical relevant changes in BMD in patients with osteoporosis. We measured BMD of the forearm in 14 young, healthy probands (10 males, 4 females), aged 24. 6 +/- 1.5 years with five different instruments using dual-energy X-ray absorptiometry (DXA), single-photon absorptiometry (SPA), and peripheral quantitative computed tomography (pQCT). Precision was expressed as the percentage coefficient of variation (CV%). In addition, the standardized CV% (sCV%) and the root mean square standard deviation (rmsSD%) was calculated for long-term precision. CV% ranged from 1.04 (SPA, distal BMD) to 2.75% (pQCT, trabecular BMD) for short-term precision and from 1.49 (DXA, QDR 1000, 1/3-distal BMD) to 4.33% (SPA, ultradistal) for long-term precision, respectively. The results for the rmsSD% were higher but correlated well with the CV%. A change that exceeds 2 radical2 CV% has been considered as being significant. On this basis, 24.0 +/- 5.1% (mean +/- SEM) of the participants in our study would be expected to have a significant change in BMD without any correlation to the time-delay between the two measurements. Measurements of BMD were done at two locations with all five instruments: ultradistal and middistal BMD using DXA and SPA and total and trabecular BMD using pQCT, respectively. Coefficients of correlation for "between-instrumental" correlation were greater than 0.5 for almost all instruments. Distal and ultradistal BMD measured by SPA and trabecular and total BMD measured by pQCT correlated better with ultradistal BMD measured by DXA. Correspondingly, "within-instrumental" correlation was better for pQCT and SPA than for DXA. The coefficients of correlation between the different DXA methods were greater than 0.95 when corresponding locations were compared. We conclude that the clinical value of monitoring bone loss by measurement of forearm BMD is compromised by the low precision which was seen for DXA methods as well as for SPA and even pQCT in young healthy controls.