High prevalence of spine–femur bone mineral density discordance and comparison of vertebral fracture risk assessment using femoral neck and lumbar spine bone density in Korean patients

The aim of this study was to evaluate the prevalence of spine–femur discordance, and to compare the effectiveness of femoral neck (FN) and lumbar spine (LS) bone mineral density (BMD) for estimation of the risk of vertebral fractures. Women who were evaluated with dual energy X-ray absorptiometry between January 2001 and December 2005 were enrolled in this study. Vertebral fracture risk was calculated using initial FN and LS BMD. The follow-up vertebral X-rays from all subjects were reviewed, and the calculated estimated risk using the Fracture Risk Assessment Tool (FRAX^®) was compared with the actual prevalence of vertebral fractures during the follow-up period. Among a total of 443 women with a mean age of 58.5 years, 130 women (29.3 %) demonstrated femur–spine discordance (i.e., a difference between FN and LS BMD of >1 SD). Most subjects having discordance showed lower LS BMD (73.1 %) compared to FN BMD. During the mean 7-year follow-up period, 12 (2.7 %) vertebral fractures occurred. In cases with high estimated fracture risk (>20 % for estimated fracture risk), using LS BMD significantly reflected the actual vertebral fracture in total subjects [odds ratio (OR) 19.29, 95 % confidence interval (CI) 4.21–88.46], in subjects with spine–femur discordance (OR 16.00, 95 % CI 1.91–134.16), and in subjects with spine–femur discordance having lower LS BMD (OR 20.67, 95 % CI 1.63–262.71). In comparison, the estimated risk using FN BMD did not reflect the actual occurrence of vertebral fractures. In conclusion, a significant number of Korean subjects exhibited spine–femur discordance, and LS BMD might be more appropriate for estimation of vertebral fracture risk.