Measurement of bone mineral density at the spine and proximal femur by volumetric quantitative computed tomography and dual-energy x-ray absorptiometry in elderly women with and without vertebral fractures

Abstract

The goal of this study was to determine the effect of vertebral fracture status on trabecular bone mineral density (BMD) measurements obtained in the proximal femur and spine by helical volumetric quantitative computed tomography (vQCT). The study population consisted of 71 Italian women (average age 73 ± 6) years. This group included 26 subjects with radiographically confirmed atraumatic vertebral fractures and 45 controls. The subjects received helical CT scans of the L1 and L2 vertebral bodies and the hip. The three-dimensional CT images were processed using specialized image analysis algorithms to extract measurements of trabecular, cortical, and integral BMD in the spine and hip. To compare the vQCT results with the most widely used clinical BMD measurement, dual X-ray absorptiometry (DXA) scans of the anteroposterior (AP) spine and proximal femur were also obtained. The difference between the subjects with vertebral fractures and the age-matched controls was computed for each BMD measure. All BMD measurements showed statistically significant differences, which ranged from 7% to 22% between subjects with fractures and controls. Although, given our small sample size, we could not detect statistically significant differences in discriminatory power between BMD techniques, integral BMD of the spine measured by vQCT and DXA tended to show stronger associations with fracture status (0.001 < p < 0.004). Measurements by QCT and DXA at the hip were also associated with vertebral fracture status, although the association of DXA BMD with fracture status was explained largely by differences in body weight between subjects with vertebral fractures and controls.

Introduction

For assessment of bone mineral density (BMD) in the elderly population, dual X-ray absorptiometry (DXA) measurements in the proximal femur are often preferred to anteroposterior (AP) measurements of the lumbar spine. This preference is based on studies showing that hip BMD is a better predictor of hip fracture than AP spinal BMD,2 and that inclusion of hypertrophic and degenerative changes into the measured region artificially elevates the BMD values in the elderly spine, reducing the sensitivity for detection of vertebral osteoporosis.13, 16, 17

Although the comparative efficacy of spinal and femoral DXA measurements for estimation of hip and spine fracture risk has been thoroughly addressed in the literature,2, 3, 18 comparisons between hip and spine measurements are lacking for volumetric quantitative computed tomography (vQCT), because protocols to assess the proximal femur, have only recently been developed.7, 11 Such information is of increasing interest because the widespread clinical dissemination of helical CT systems has resulted in higher availability of scanner time and has broadened the potential market for vQCT BMD measurements.

In this study, we compared vQCT BMD measurements of the spine and proximal femur in a cohort of elderly Italian women who were referred for routine clinical bone mineral assessment by DXA. We determined trabecular, integral, and cortical BMD values at the two skeletal sites and examined differences in the vQCT and DXA measures between subjects with one or more atraumatic vertebral fractures and subjects without vertebral fractures.