The AOSpine thoracolumbar spine injury classification system: A reliability and agreement study

AIM: Recently, AOSpine trauma knowledge forum proposed the AOSpine thoracolumbar injury classification (AOSTLIC) system and suggested that it was reliable. However, reliability data from additional institutions for the AOSTLIC system are not available. This study was to examine the reliability of the AOSTLIC system in patients with thoracolumbar (TL) fractures. MATERIALS AND METHODS: Between August 2009 and June 2012, 56 patients with 74 levels traumatic TL spinal injuries were recruited. Two classifiers, consisting of two spine surgeons, assessed clinical and imaging data. Initially, one surgeon reviewed the data in order to classify and calculate injury severity score according to the AOSTSIC system. This process was repeated on a 5-week interval by another surgeon. Then we analyzed data for intra-observer and inter-observer reliability using the kappa statistic (k). Finally, validity was assessed using the known-groups comparison. RESULTS: The mean age of patients was 59.5 ± 11.5 years. The κ values for the AOSTSIC system for intra-observer and inter-observer reliability ranged from 0.83 to 0.89, indicating nearly perfect agreement agreements. Known-groups analysis showed satisfactory results. The AOSTSIC system discriminated well between sub-groups of patients who differed in Oswestry disability index. CONCLUSION: The findings showed that the morphologic classification in AOSTSIC system appears to be reliable and reproducible classification.