Measurement of fracture kyphosis with the Oxford Cobbometer: intra- and interobserver reliabilities and comparison with other techniques.

Abstract

Statistical analysis of 3 techniques for measuring thoracolumbar kyphosis secondary to fracture. To determine the reliability of using an Oxford Cobbometer and assess the most reliable measurement technique. The reproducibility of Cobb angles for the assessment of saggital plane deformity on spine radiographs has been shown to have significant variability in both intra- and interobserver error. Twenty-four lateral spine radiographs of patients with thoracic and lumbar vertebral fractures were measured on 2 separate occasions, in random order, by 4 blinded observers using the same Oxford Cobbometer and ruler. Method 2, the angle from the inferior endplate of the vertebra above the fractured vertebra to the superior endplate of the vertebra below the fractured vertebra, had the greatest intraobserver and interobserver reliabilities (rho = 0.856-0.976 and rho = 0.95, respectively). The other 2 methods had lower reliabilities; however, all 3 methods were well above the statistically acceptable threshold of >0.8, and the intraobserver reliabilities with each observer was 99% overall. These reliabilities supersede results reported previously using the conventional Cobb technique. The absolute mean difference between readings and 95% limit of agreement also improves on previous data, 2 degrees and +/- 5.8 degrees , respectively. Highest intraclass correlation coefficients were obtained using method 2. Using the Oxford Cobbometer to measure fracture kyphosis has higher reliability than the standard Cobb angle technique. It is easy and quick to use in a clinical setting.