Are bilateral decubitus views necessary in assessing for vertebral compression fractures using DXA vertebral fracture assessment?

A Summary The purpose of this study is to assess the differences in VFA diagnostic accuracy when using bilateral decubitus views and whether diagnostic accuracy is affected by scoliosis. Our findings show that the current practice of performing only one side is valid; however, bilateral views can improve specificity in scoliosis. Introduction The diagnostic accuracy of vertebral fracture assessment (VFA) can be influenced by poor patient position and scoliosis. This study aims to assess the differences in VFA diagnostic accuracy for right and left lateral decubitus views and the effect of scoliosis. Methods One hundred fourteen postmenopausal women received right and left lateral thoracolumbar spine dual-energy VFA and radiography. Cobb angles were measured from the posteroanterior absorptiometry image, and lumbar spine radiography was the standard reference for vertebral fracture and also provides the levels investigated. McNemar's test was used to compare accuracy between the two decubitus position and Fisher's exact test was used for patients with and without scoliosis. Results Forty-two vertebral fractures (VFs) were identified. There was no significant difference in sensitivity (p = 0.125) or specificity (p = 0.866) between the left lateral decubitus (64.3, 97.2%) and right lateral decubitus (76.2, 91.1%), respectively, views. Scoliotic patients had a significantly worse specificity (92.7 vs 98.1%, p = 0.003) than patients without scoliosis; however, a combination of both decubitus positions significantly improved specificity (p < 0.001). Conclusion Right and left side lateral decubitus views have excellent agreement with radiography and similar diagnostic accuracy in the detection of VFs. Thus, the current practice of performing only one side is valid. With scoliosis, bilateral decubitus views can improve the specificity of detecting VF; however, this would increase radiation dose.