Objective:To identify risk factors for cement leakage in percutaneous vertebroplasty (PVP) for Kümmell disease.Methods:A total of 309 patients (351 levels) with Kümmell disease who underwent PVP between November 2015 and June 2019 were retrospectively reviewed. Age, gender, time of symptom onset, staging of Kümmell disease, fracture site(thoracic, lumbar), cortical disruption, type of fracture (wedge, biconcave, crush), fracture severity (mild, moderate, severe), intrusion of posterior wall, basivertebral foramen, puncture approach (unilateral, bilateral), cement distribution pattern (lumped, spongy), cement volume, cement leakage (yes, no) and cement leakage type were recorded. Cement leakage was classified into three types: through the basivertebral vein, through the cortical defect, and through the segmental vein. The data was analyzed by univariate and multivariate analysis to determine related factors of cement leakage in general and each type.Results:The rate of overall leakage was 65.8% (231/351). The leakage rate of basivertebral vein type, cortical defect type, and segmental vein type was 21.4% (75/351), 37.6% (132/351) and 22.8% (80/351), respectively. Multivariate analysis showed that three significant factors related to leakage in general were cortical disruption, basivertebral foramen and cement distribution pattern. Significant factors related to basivertebral vein type leakage were basivertebral foramen and cement distribution pattern. Significant factors related to cortical defect type leakage were cortical disruption and cement distribution pattern. Significant factors related to segmental vein type leakage were basivertebral foramen, cement distribution pattern, cement volume and fracture site.Conclusion:Risk factors of cement leakage in PVP for Kümmell disease include cortical disruption, basivertebral foramen and cement distribution pattern.

Percutaneous vertebral augmentation, including percutaneous vertebroplasty and percutaneous kyphoplasty, has been considered as an effective and safe option in treating osteoporotic vertebral compression fractures. The fractured vertebrae were strengthened by the bone cement injected, thus reducing the symptoms related to fracture. Bone cement injected intraoperatively can be divided into extraspinal (leakage) and intraspinal part, depending on its final location. The former may lead to pulmonary embolism, spinal cord or nerve injury, or some other sequelae; the latter may closely relate to the clinical outcome, radiological outcomes, surgical complications and biomechanical properties. To date, there were a large number of studies on term of the distribution type of bone cement. However, the classification criteria varied and there was lack of literature review on this issue. According to the literature reviewed, the distribution type of bone cement was a critical parameter in percutaneous vertebral augmentation; most classification systems were based on the postoperative X-ray, some based on the postoperative CT, and only a few based on postoperative MRI; in different classification systems, criteria on bone cement morphology tends to be consistent, however, criteria on bone cement range tends to be inconsistent, consistency, similarity and controversy all exited among conclusions between various studies on the morphology and range of bone cement; any single classification system can not describe the distribution of bone cement thoroughly. In this study, classification systems were reviewed, clinical significance and biomechanical conclusions of different classification systems were documented, and the reliability and limitations of classification systems were summarized, hence providing an insight for further research on classifications of the bone cement distribution.

Objective:To explore the differences in clinical and injured vertebra radiological parameters between patients with non-traumatic osteoporotic vertebral compression fracture (OVCF) and those with traumatic OVCF after percutaneous vertebroplasty (PVP).Methods:A retrospective study was conducted to analyze the 369 OVCF patients (with 458 vertebrae injured) who had been treated by PVP at Department of Orthopaedics, Beijing Luhe Hospital between October 2015 and March 2017. There were 292 females and 77 males with a mean age of 73 (60, 79) years. Based on the absence or presence of a trauma history, the patients were stratified into a non-traumatic group (127 patients with 160 vertebrae injured) and a traumatic group (242 patients with 298 vertebrae injured). Clinical parameters [age, gender, body mass index, symptomatic duration, and number of injured vertebrae, visual analogue scale (VAS), Oswestry disability index (ODI), duration of follow-up, and rate of new OVCFs] and injured vertebra radiological parameters (position of injured vertebra, fracture type, compression severity, fracture range, cortical defect, intravertebral cleft, spinal canal compromise, basivertebral foramen, morphology of bone cement, range of bone cement, cement leakage, cement volume, rate of vertebral height restoration, recollapse of cemented vertebrae) were recorded perioperatively. All the clinical and radiological parameters were compared between the 2 groups.Results:Compared with the traumatic group, the non-traumatic group had an older age [75 (71, 83) years versus 71 (65, 76) years], more females (85.0% versus 76.0%), a longer symptomatic duration [10.0 (7.0, 15.0) d versus 6.5 (2.0, 12.0) d], a lower preoperative VAS pain score [7 (6, 8) points versus 7 (7, 8) points], a lower VAS pain score at postoperative day 1 [2 (2, 3) points versus 2 (2, 3) points], a lower preoperative ODI [66% (63%, 72%) versus 70% (65%, 73 %)], a lower ODI at postoperative day 1 [32% (30%, 34%) versus 32% (31%, 34%)], a higher rate of new OVCFs during follow-up (34.6% versus 12.8%), a lower rate of thoracolumbar lesions (51.9% versus 70.1%), more deformed fractures (mostly amphicoelous type), a lower rate of cortical defects in the anterior wall (20.0% versus 31.5%), a higher rate of trabecular pattern of cement (83.1% versus 71.8%), a higher rate of type-B cement leakage (50.6% versus 31.9%), a lower rate of type-C cement leakage (5.6% versus 12.8%), a lower rate of recollapse of cemented vertebrae (43.8% versus 55.4%). All the comparisons above were statistically significant ( P<0.05). There were no significant differences between the 2 groups in the other clinical or radiological parameters ( P>0.05). Conclusions:There are statistically significant differences in a significant number of clinical and injured vertebra radiological parameters between patients with non-traumatic OVCF and those with traumatic OVCF after PVP. It is noteworthy that non-traumatic OVCFs are one specific subgroup of OVCFs.