Study design: Retrospective study.
Objective: To identify fracture morphology-dependent risk factors for the occurrence of cement leakage during percutaneous vertebroplasty.
Summary of background data: Extravertebral cement leakage is the most frequently reported complication and represents a major risk after vertebroplasty and balloon kyphoplasty. Despite the frequent occurrence of cement leakage and potentially dangerous complications and sequelae, there are few reports that examine its predictors.
Methods: Retrospective review of 194 consecutive patients who underwent percutaneous vertebroplasty for painful osteoporotic or malignant vertebral fracture. The influences of several parameters that might affect the occurrence of cement leakage were assessed using univariate and multivariate analyses. Cement leakage was assessed using computed tomography scanning and classified into 4 different types with different potential sequelae: through the basivertebral vein, through the segmental vein, through a cortical defect, and intradiscal leakage.
Results: Leakage of cement was detected in 209 of the 272 treated vertebrae (76.83%). The most common types of leakage detected were through the basivertebral vein (43.38%) and the segmental vein (42.27%). None of the evaluated variables showed a statistically significant effect. Only the absence of Kummell disease (P = 0.063) and a lower severity grade of collapse (P = 0.068) approached statistical significance. For basivertebral vein leakages, the location at the thoracolumbar level and the absence of Kummell disease (P < 0.05) were strong predictive factors for cement leakages. For segmental vein leakages, the odds decreased as the severity grade of collapse increased (P = 0.008).
Conclusion: Each different vertebral fracture pattern has its own risk factors for cement leakage. For certain types of leakage, Kummell avascular necrosis as a protective factor and fracture severity grade with paradoxical effect have each demonstrated influence in the occurrence of cement leakage.
Level of evidence: 4.