[Precision and measurement errors in bone mineral density by dual energy x-ray absorptiometry in mass health...

Decreasing mortality and the low birth rates has resulted in a high rate of increase in the elderly population in Japan. Increases in osteoporosis and its complication (fractures) are inevitable among the Japanese elderly. The Ministry of Health and Welfare of Japan and other municipalities are in the process of developing standards for noninvasive and quantitative assessment of bone minerals in mass health examinations in order to prevent osteoporosis fracture in elderly females. Dual energy x-ray absorptiometry (DXA) has undergone development to significantly improve scan speed, image resolution and precision with relatively low exposure to x-rays, and is now considered one of the most suitable instruments to measure bone mineral density (BMD) in both clinical and mass-screening for osteoporosis in population-based mass health examinations in Japan. The purpose of the present study using DXA fixed in a mobile van are: 1) to compare the precision of anteroposterior (AP) and lateral (LT) projections of the lumbar region, and three areas of the proximal femur (neck: NK, trochanter: TR, Ward's triangle: WD); 2) to analyze the physical factors influencing the precision; 3) to study the magnitude of intra- and inter-observer measurement errors and of different sets of regions of interest (ROI) in LT projection; and 4) to examine which measurement items should be selected in mass health examination in order to minimize measurement errors particularly for longitudinal observational study. Results are as follow: 1) Precision estimated by the reproducibility of measurements in the spine phantom bone were relatively good in both short and long terms. However, in female volunteers, precision of LT-BMD was remarkably inferior to that for AP-BMD and the three BMD of the proximal femoral region. Among the physical characteristics, body mass index and menopausal state were significantly associated to the precision of LT-BMD. 2) Regarding the correlations of BMD in the same region, all correlation coefficients except LT-BMD (r = 0.65) exceeded 0.97. Intra-regional BMD showed that correlations when including LT-BMD were significantly lower than the other correlations with LT-BMD excluded. 3) Concerning the magnitude of observational errors in both intra- and inter-observer measurements, LT-BMD showed the largest measurement error, largely due to lack of consensus in the ROI setting for the vertebral body and to the instability of positioning when lateral scanning was carried out.(ABSTRACT TRUNCATED AT 400 WORDS)