Low pelvic incidence as a risk factor for vertebral recollapse after percutaneous vertebroplasty in the thoracolumbar region.

Abstract

Percutaneous vertebroplasty (PVP) is an effective procedure for treatment of osteoporotic vertebral compression fractures (OVCFs). Recollapse of the cemented vertebrae is not unusual and the thoracolumbar junction is the most common region. Nevertheless, not all patients suffering from OVCFs in this region develop recollapse after PVP. The aim of this study was to investigate possible risk factors related to recollapse of the cemented vertebrae in the thoracolumbar region. Retrospective study. A total of 161 patients undergoing PVP. Clinical outcomes were assessed using Visual Analog Scale (VAS) scores. Radiographic parameters included pelvic incidence (PI), thoracolumbar kyphosis (TLK), kyphotic angle, compression rate, reduction rate, and occurrence of intravertebral cleft (IVC). Patients were divided into the recollapse group and control group. Patient characteristics, clinical and radiographic parameters were compared between the two groups. Univariate and multivariate logistic regression were used to evaluate the potential risk factors for recollapse. The correlations between the variables were examined. A receiver operating characteristic curve of PI was constructed to discriminate between patients with and without recollapse. There were no significant differences in patient characteristics between the two groups except for bone mineral density (BMD). Occurrence rate of IVC was significantly higher in the recollapse group. VAS scores were significantly decreased after PVP. At last follow-up, they were increased in the recollapse group and maintained in the control group. PI was significantly lower in the recollapse group than in the control group. The univariate logistic regression found four possible risk factors for recollapse, including low PI, IVC, low BMD, and high TLK. Further multivariate logistic regression eliminated high TLK from them. The collinear analysis showed low tolerance and high variance inflation factor for preoperative and postoperative TLK, but not for PI, IVC and BMD. PI was a good predictor of recollapse, and the optimal cut-off value was 43°. The magnitude of preoperative and postoperative TLK was significantly correlated with the value of PI. Recollapse of the cemented vertebrae in the thoracolumbar region was related to low PI, IVC and low BMD. PI less than 43° was a good predictor of recollapse. TLK was dependent on PI and not a risk factor for recollapse. In addition to PVP, patients with low PI, IVC and low BMD may require personalized interventions such as combined internal fixation and trunk orthoses.