ANVCFV Score System: Assessment for Probability of New Vertebral Compression Fractures after Percutaneous Vertebroplasty in Patients with Vertebral Compression Fractures

Abstract

Background: Percutaneous vertebroplasty (PVP) is widely used for the treatment of painful vertebral compression fractures (VCFs). However, new VCFs occur frequently after PVP. Objectives: We aim to establish an objective risk score system to assess the possibility of new vertebral fractures in patients with VCFs undergoing PVP. Study Design: This study was a retrospective study, and it was approved by the Institutional Review Board of our 2 institutions. Setting: This study consists of patients from 2 large academic centers. Methods: Patients with VCFs who underwent their first PVP and met the inclusion criteria between January 2007 and December 2013 at Hospital A (training cohort) and Hospital B (validation cohort) were included. In the training cohort, the independent risk factors for new VCFs after PVP were identified by multivariate stepwise backward Cox regression analysis from the risk factors selected by univariate analysis and Harrell’s C-statistics and used to develop the score system (assessment for new VCFs after PVP [ANVCFV]) to predict the probability of new VCFs. Results: In total, 397 patients (training cohort: n = 241; validation cohort: n = 156) were included in this study. In the training cohort, the ANVCFV score was developed based on 5 independent risk factors for the new VCFs after PVP, including lower computed tomography (CT) values, pre-existing old VCFs, intradiscal cement leakage, more than one vertebra treated, and superior or inferior marginal cement distribution in the vertebra. The patients were divided into 2 groups by the ANVCFV score of -1.5 to 8.5 vs. > 8.5 points in the probability of new VCFs (median fracture-free time: 1846 vs. 732 days; P < 0.001) in the training cohort. The accuracy of this score system was 77.4% for the training cohort and 85.3% for the validation cohort. Limitations: The main limitations of this study are that it is a retrospective study and that there is a significant difference of the treated vertebrae of PVP per session between the 2 cohorts. Conclusion: Patients who underwent their first PVP with an ANVCFV score > 8.5 points may exhibit an increased chance of suffering from new VCFs. Key words: Vertebral compression fracture, percutaneous vertebroplasty, newly developed, risk factors, risk score system, Cox regression model, accuracy, validation