Reliability Analyses of Radiographic Measures of Vertebral Body Height Loss in Thoracolumbar Burst Fractures

Abstract

In thoracolumbar (TL) burst fractures, vertebral body height loss (VBHL) indicates the degree of instability and constitutes one of the decision criteria for surgical treatment. However, the relative reliability and variability of different measurement techniques for VBHL are unknown. We compared the reliability of different methods used to assess VBHL. A total of 144 patients with TL burst fractures were included, and lateral radiographs were taken twice at an interval of 2 weeks, which were examined by 3 observers. The measurement methods used included the anterior/posterior vertebral body height compression ratio (APCR), anterior height compression percentage (AHCP), and anterior/posterior vertebral body height compression ratio percentage. To compare the accuracy of measurements according to vertebral degeneration, subjects were divided into 2 groups based on the median age of 50 years. In intraobserver comparisons, the APCR method showed a higher inter- and intraclass correlation coefficient (ICC) (>0.714) compared with the other methods. In interobserver comparisons, the ICC of the APCR (>0.793) was excellent. In intraobserver comparisons of the aged >50-years group, only the APCR method showed an excellent ICC (>0.753), whereas the AHCP method showed a fair to good ICC, and the anterior/posterior vertebral body height compression ratio percentage method had the lowest ICC. In interobserver comparisons of the aged >50-years group, the APCR and AHCP methods showed excellent ICCs. In the aged ≤50-years group, all 3 methods showed similar fair to good ICC values. Based on comparative reliability analyses, we recommend the APCR method as the first-line technique and the AHCP as an alternative technique for measuring VBHL in TL burst fractures.