Background: Our study evaluated the utility of vertebral attenuation derived from low-dose chest computed tomography (LDCT) compared to dual-energy x-ray absorptiometry (DXA) for detecting bone fragility. Methods: Total of 232 subjects (78 men and 154 women) aged above 50 years who underwent LDCT and DXA within 30 days were evaluated. LDCT-measured bone attenuation in Hounsfield Units (HU) of 4 vertebrae (T4, T7. T10 and L1) using volumetric methods were evaluated for the correlation with DXA-measured bone mineral density (BMD), and the diagnosis of compression fracture, osteoporosis or low BMD (osteoporosis or osteopenia) using DXA measurements as the reference standard. Results: Average attenuation of 4 vertebrae showed strong correlation with DXA-measured BMD of the L-spine (r=0.726, p<0.05). At receiver-operating characteristic (ROC) analyses, area under curve (AUC) across LDCT-measured thresholds of the average attenuation to distinguish compression fracture was 0.827, and a threshold of 129.5 HU yielded 90.9% sensitivity and 64.4% specificity. Similarly, high AUCs and good diagnostic performance were shown for detecting osteoporosis and low BMD. Among 44 subjects with compression fractures, average bone attenuation showed strong negative correlation with both worst fracture grade (r=-0.525, p<0.05) and cumulative fracture grade score (r=-0.633, p<0.05). Conclusion: LDCT-measured bone attenuation by volumetric methods showed good correlation with BMD measured by DXA, and better diagnostic performance for for prevalent compression fractures compared with DXA.