Discrimination of spinal fracture with various bone mineral measurements

Abstract

For several different bone mineral measurements and various skeletal sites, we compared capability to discriminate between women in various age decades with and without spinal fracture, and attempted to identify the most effective cutoff level in discrimination of spinal fracture. The subjects were 88 women aged 50–59 years (including 32 with fracture), 95 women aged 60–69 years (including 54 with fracture), and 34 women aged 70–79 years (including 18 with fracture). Spinal trabecular and cortical bone mineral density (BMD) were measured using quantitative computed tomography (CT), and spinal, radial (ultra-distal, 10% distal and 33% distal), and calcaneal BMD were measured by dual X-ray absorptiometry. These BMD values were obtained in each subject on the same day. Three statistical techniques—Student’s t-test, the logistic regression analysis, and the receiver operating characteristics (ROC) analysis—were applied and accuracy was calculated using the various cutoff values. The capability to discriminate between women with and those without fracture using these BMD values was different among the three age groups. In women aged 50–59 and 60–69 years, all measurements showed good capabilities for discriminating women with fracture. In women aged 70–79 years, these measurements showed lower capability than in those aged 50–59 and 60–69 years, but among them, the calcaneal and ultradistal radial BMD showed relatively good capability. The 10% and 33% distal radial BMD values were not useful in the detection of the high risk women with fracture. The cutoff BMD values for discrimination of women with fracture varied according to the sites and methods of measurement. For each specific age group, the most suitable measurement methods and the appropriate skeletal sites should be considered, and the effective cutoff values to discriminate those with fracture may differ according to the measurement methods, the skeletal sites examined, and age.