A Comparative Analysis of International Classification Systems to Predict the Risk of Collapse in Single-Level Osteoporotic Vertebral Fractures

Abstract

Introduction: Various classifications for osteoporotic vertebral fractures (OVFs) have been introduced to enhance patient care and facilitate clinical communication. However, there is limited evidence of their effectiveness in predicting vertebral collapse, and very few studies have compared this association across different classification systems. This study aims to investigate the association between OVF categories, according to the most widely used classification systems, and vertebral collapse.

Patients and methods: A retrospective single-center study was conducted involving patients diagnosed with acute OVFs at the emergency department of a tertiary-level academic hospital with a minimum follow-up of 6 months. Vertebral fractures were independently classified by two radiologists according to several classification systems, including those proposed by Genant, Sugita, the German Society for Orthopedics and Trauma (DGOU), and the AO Spine. Associations between vertebral collapse and OVF classification systems were analyzed using bivariate and logistic regression analyses.

Results: This study included 208 patients (82.7% females; mean age of 72.6 ± 9.2 years). The median follow-up time was 15 months, with L1 being the most common fracture site (47.6%). The most frequent OVF types observed, according to Genant's morphological, Genant's quantitative, Sugita 's, DGOU's, and AO Spine's classifications, were biconcave (50%), grade 0.5 (47.6%), bow-shaped (61.5%), OF2 (74%), and A1 (61.5%), respectively. All classifications, except for Genant's quantitative system, were significantly associated with vertebral collapse in bivariate analyses. Logistic regression analyses showed a significant association (p = 0.002) between the AO Spine classification and vertebral collapse, with 85.7% of A4 fractures developing collapse on follow-up.

Conclusions: The AO Spine classification showed the highest predictive capacity for vertebral collapse. Specifically, A4 fracture types showed a very high risk of vertebral collapse, confirming the need for non-conservative management of these fractures. Further multicentric and prospective studies are warranted to confirm these findings.

Keywords: AO spine; classification; collapse; fracture; imaging; osteoporosis; prognosis; radiology; spine; vertebra.

Figure 1

Illustrative examples of osteoporotic vertebral fractures in our sample based on the AO Spine classification system. (A) A1-type fracture of L1 (white arrow). Note the subtle loss of vertebral height and the trabecular fracture line below the superior endplate, as well as the absence of posterior wall displacement. (B) A2-type fracture of L3 (white arrow) with a typical pincer-like morphology, involving the central part of the vertebral body. (C) A3-type fracture of L1 (white arrow) involving the superior endplate, with mild displacement of the posterosuperior wall to the spinal canal. (D) A4-type fracture of L2, with involvement of the superior and inferior endplates (white arrows) and displacement of the posterosuperior wall to the spinal canal (orange arrow).

Figure 2

The incidence and rate of collapse of osteoporotic vertebral fractures, according to the AO Spine classification system in our sample.

Figure 3

Receiver operating characteristic (ROC) curve analysis for the AO Spine classification in our sample. The blue line represents the ROC curve based on the logistic regression results, where the AO Spine classification was the only significant predictive variable. The gray diagonal line corresponds to the reference of a random classification (line of no discrimination).