Analysis of intradiscal cement leakage during percutaneous vertebroplasty: Multivariate study of risk factors emphasizing preoperative MR findings

Objective: Previous reports have shown that intradiscal cement leakage during percutaneous vertebroplasty (PVP) is related to several risk factors. The purpose of this study was to evaluate preoperative MRI scans for such risk factors. Methods: The study retrospectively analyzed 136 patients (aged 43-93 years; 234 vertebral bodies) with osteoporotic compression fractures. All patients underwent both MRI and PVP. There were 28 men (20.59%) and 108 women (79.41%). Age, gender, bone mineral density (BMD) score, endplate cortical disruption, abnormal T2-weighted hyperintensity in adjacent discs, presence of Kümmell's disease, linear body fracture with extension to endplate, level of treated vertebral body and injected cement volume were considered risk factors for intradiscal cement leakage. Results: Of the 234 vertebral bodies, 55 bodies from 42 patients with no endplate cortical disruption showed no adjacent intradiscal cement leakage. Of 179 bodies from 95 patients with endplate cortical disruption, 54 (30.17%) showed intradiscal cement leakage. Of the other possible risk factors, abnormal T2 hyperintensity in adjacent discs was significantly related to intradiscal cement leakage (P= 0.016). The other possible factors (age, gender, BMD score, Kümmell's disease, linear body fracture extending to the endplate, level of treated vertebral body and injected cement volume) were not related to intradiscal cement leakage. Conclusion: There was no adjacent intradiscal cement leakage without endplate cortical disruption. Abnormal T2 hyperintensity in adjacent discs may be related to intradiscal cement leakage, but only in the presence of endplate cortical disruption. Also, not having Kümmell's disease did not prevent intradiscal cement leakage. Thus, given these circumstances, careful cement injection is needed to reduce intradiscal cement leakage.