Accuracy of spinal curvature assessed by a computer-assisted device and anthropometric indicators in discriminating vertebral fractures among individuals with back pain

Abstract

This study examined the accuracy of thoracic and lumbar kyphotic angles as well as anthropometric indicators for discriminating patients with vertebral fracture among Japanese women >50 years old with back pain. Along with region-specific kyphotic angles and anthropometric indicators, the combination of thoracic and lumbar kyphotic angles offered the highest accuracy.

Introduction: Vertebral fractures have been associated with thoracic kyphosis. However, reports on lumbar kyphotic changes in association with vertebral fracture are scarce. This study investigated the accuracy of thoracic kyphotic angle (TKA) and lumbar kyphotic angle (LKA) measurements as well as anthropometric indicators (wall-occiput distance (WOD) and rib-pelvis distance (RPD)) in discriminating patients with vertebral fracture.

Methods: Lateral radiographs of the spine were obtained in 70 postmenopausal Japanese women who visited an orthopedic clinic with low back pain (mean age, 76.2 ± 9.0 years). Radiographic vertebral fracture was diagnosed using quantitative measurement according to Japanese criteria. Osteoarthritis (OA) was defined as Kellgren-Lawrence (KL) grade 3 or higher. TKA and LKA were measured using SpinalMouse®. WOD and RPD were also measured.

Results: At least one vertebral fracture was present in 49 subjects (70 %). Women with vertebral fractures showed significant increases in LKA, TKA + LKA, and WOD and decreases in RPD. Logistic regression analysis showed significant association between TKA + LKA and vertebral fracture independent of the presence of OA. Receiver operating characteristic analysis revealed that TKA was useful for discriminating thoracic fractures (area under the curve (AUC), 0.730) and LKA was useful for lumbar fractures (AUC, 0.691). The combination of TKA + LKA offered the highest accuracy for detecting thoracic, lumbar, and any vertebral fractures, with AUCs of 0.779, 0.728, and 0.783, respectively. WOD and RPD showed low-to-moderate accuracies for thoracic, lumbar, and any vertebral fractures.

Conclusions: Assessment of spinal kyphosis by SpinalMouse® as well as anthropometric indicators proved useful in discriminating subjects with vertebral fractures. These convenient and radiation-free methods could contribute to early diagnosis of vertebral fractures and subsequent appropriate treatment, thus preventing additional osteoporotic fractures.