Computed tomography imaging characteristics help to differentiate pyogenic spondylitis from brucellar spondylitis.

Purpose: Both pyogenic spondylitis (PS) and brucellar spondylitis (BS) can cause deformities and permanent neurologic deficits without prompt diagnosis and treatment. However, differential diagnosis is challenging. The aim of this study was to compare the computed tomography (CT) imaging features of PS with those of BS.

Methods: Thirty-two patients with PS and 44 with BS were enrolled in the study. CT images were obtained in all cases. Data on bone destruction and formation, vertebral wall destruction, and osteosclerotic changes were collected and compared using the Chi-square test or t test. A P value < 0.01 was considered statistically significant. Positive predictive values (PPV) for detecting PS or BS were reported.

Results: Involvement of the lumbar vertebrae and multiple spinal levels was more common in the BS group than in the PS group. Bone destruction was significantly greater in the PS group than in the BS group (30.8 vs 18.0%; t = 3.920, P = 0.000), with more extensive destruction of the vertebral body (35.8 vs 12.5%, χ = 12.672, P = 0.002, PPV = 63.16%). In the BS group, there was more osteosclerosis around erosions (70.5 vs 43.3%, χ = 11.59, P = 0.001, PPV = 67.74%) and fan-shaped osteosclerosis (27.3 vs 19.4%, χ = 18.556, P = 0.006, PPV = 64.86%), more bone formation around the vertebra (77.2 vs 34.3%, χ = 33.608, P = 0.000, PPV = 76.83%), more bone formation under the anterior longitudinal ligament (63.6 vs 19.4%, χ = 30.133, P = 0.000, PPV = 76.09%), more longer anterior bone formation (3.55 vs 0.78 mm, t = 3.997, P = 0.000), and more anterior and closed-bone formation with local erosion (42.0 vs 9.0%, χ = 74.243, P = 0.000, PPV = 74.36%).

Conclusions: CT images have unique advantages of revealing the morphology of erosions, osteosclerosis, and bone formation around the vertebra and help to differentiate PS from BS. These slides can be retrieved under Electronic Supplementary Material.