Abstract
To determine whether vertical laminar fracture (VLF) can distinguish between AO type A3 and A4 fractures.
In a retrospective review of 111 consecutive acute thoracolumbar burst fractures, 5 reviewers independently analyzed computed tomography scans to classify fractures into A3 or A4 and to identify VLF. The following computed tomography parameters were measured: spinal canal stenosis >50%, anterior vertebral height ratio <50%, load sharing score >6, and local kyphosis >20°. We calculated the diagnostic performance of VLF in detecting A4 fracture. We compared the proportion of fractures with positive bony parameters, neurological deficit, dural tears, and surgical treatment between A3, A4 with VLF, and A4 without VLF.
VLF was present in 62/75 (83%) A4 fractures and 2/36 (5.5%) A3 fractures (P < 0.0001). VLF yielded a high specificity of 94% (95% confidence interval 81%–99%) and moderately high sensitivity of 83% (95% CI 72%–91%) in detecting A4 fractures. A significantly higher proportion of A4 fractures with VLF had neurological deficit (24% vs. 0, P = 0.05), spinal canal stenosis >50% (25% vs. 0, P = 0.04), and anterior vertebral height ratio <50% (24% vs. 0, P = 0.05) than A4 fractures with no VLF. Interrater and intrarater κ values for VLF and AO standard criterion were excellent (>0.85).
We found VLF to be highly specific, sensitive, and reliable in detecting A4 fractures. A higher proportion of A4 fractures with VLF had radiographic parameters and neurological deficit than A4 fractures with no VLF. VLF could be used as a severity modifier to further discriminate A3 and A4 fractures regarding severity and potentially guide treatment decision making.