The 26 patients enrolled in this prospective study with pain-causing osteoporotic VCFs (6 men, 20 women), were seen between April 2005 and May 2007. The study protocol was approved by Changzhou No. 2 Hospital's Ethics Committee and informed consent was obtained from all patients before the procedure. The mean patient age was 65.5 years (range: 58 - 90 years). A total of 35 procedures were performed in 26 patients. The fractures occurred between the T8 and L5 vertebrae: T12 - 12 cases; L1 - 11 cases; T11 - 4 cases; L2 - 2 cases; T8-10 and L3-L5 - 1 case, respectively. Of the patients treated with the Sky Bone Expander, 20 were treated for a single VCF, whereas six were treated for multiple VCFs (2-4 vertebrae). Magnetic resonance imaging (MRI) was used to confirm the presence of an acute fracture associated with pain, as well as burst-type injuries with slightly retropulsed bone. The signs of a fracture on the MRI included a higher intensity signal in T2-weighted images and a weaker intensity signal in T1-weighted images. The diagnostic criteria for painful VCFs included obvious back pain and image examination at one or more levels, and low bone mineral density (BMD) with a T-score less than -2.5 (determined by standard dual-energy X-ray absorptiometry) (Explorer QDR, Hologic Inc. MA). Contraindications for the application of this procedure included fractures that did not cause pain, fractures with a big retropulsed bone, and patients afflicted with other diseases which rendered them unable to tolerate the operation. The initial number of VCF patients was 31; however, three patients with vertebral fractures accompanied with spinal cord compression and two patients with severe cardiovascular or pulmonary diseases were excluded due to the inability to tolerate a prone position.

The Sky Bone Expander is a device made of a highly rigid polymeric material. In the contracted configuration, it is 4 or 5 mm in diameter and can be inserted via a pedicle cannula into the vertebral body. In turn, its fully expanded shape is a trapezoid or cubic configuration with a maximal diameter of 14 or 16 mm. This shape elevates the endplates restoring the vertebral body height (Fig. 1). The procedures were carried out as described previously (2-4). In brief, all procedures were performed in the operation room with the patient positioned in the prone position with the spine extended. The skin and subcutaneous tissues were locally anesthetized with 1% lidocaine. The proper placement of the Sky Bone Expander involved identifying the transpedicular entry point into the vertebral body using a guide pin and with the aid of fluoroscopy. A hand-mounted drill bit was then used to ream a corridor for insertion of the device. With the guide of a lateral radiograph, the radiopaque polymer tube was ideally situated under the collapsed endplate (Fig. 2A). The creation of the cavity for injection of bone cement was achieved by radial expansion (Fig. 2B), followed by the contraction and removal of the SKY Bone Expander. Furthermore, with the aid of continuous fluoroscopic guidance, bone cement was injected into the created cavity through the implant core (Figs. 2C, D). Cement filling was stopped when the cement mantle reached two thirds of the way back to the posterior vertebral body cortex or when cement extrusion was observed on the lateral fluoroscopic images. The same procedure was repeated on the opposite side of the vertebra if necessary. A single procedure can treat 1-4 vertebral levels, with an average bone cement injection volume of 3.5 mL (range, 2.5-5.0 mL) per vertebrate level.

Postoperative radiographs were examined for the presence of cement leakage by looking at vertebral position and the possible correlation with any postoperative clinical symptoms.

All radiographs performed in this study were independently evaluated at one week after surgery by two unbiased researchers who were not involved in the care of these patients. The parameters measured included the Cobb angles across the fractured level, the preoperative and postoperative anterior height (AH), midline height (MH), and posterior height (PH) of the fractured vertebra on a lateral radiograph in the standing position. To determine the Cobb angle, measurements were taken from the superior endplate of the vertebra from one level above the treated vertebra to the inferior endplate of the vertebral body at one level below the treated vertebra (5). The vertebral height of the fractured vertebra was defined as the distance between corresponding points on the superior and inferior endplates at the posterior, middle, or anterior locations (Fig. 3).

The postoperative follow-up time ranged from 3 to 6 months, with an average of 4.5 months. Patients completed the questionnaire of the Visual Analog Scale (VAS) and ambulation status, as well as undergoing a follow-up comprehensive neurologic examination at 1 day, 1 week, 1 month, and 3 months after the procedure. Moreover, patients also rated back pain on a VAS, in which "1" indicated no pain, and "10" indicated severe pain. Lastly, patient ambulation status included full ambulation, assisted ambulation, and no ambulation. The SPSS 11.5 software package (Chicago, IL) was used as a data analysus tool. Paired t-tests were used to compare the pre-operation and post-operation data.

All of the 26 studied patients tolerated the bone expansion procedure well. The average operative time was 42.4 ± 15.5 min per vertebra. As well, all patients also experienced some degree of pain relief and improvements in mobility within the first 24 hours after surgery. The mean patient VAS score significantly improved from 7.7 ± 1.8 before surgery to 3.1 ± 2.0 at 1-day (p < 0.01), 2.9 ± 1.7 at 7-days (p < 0.01), 2.6 ± 1.5 at 1-month (p < 0.01), and 2.9 ± 1.3 at 3-months (p < 0.01) after surgery, respectively (Fig. 4).

The patients' mobility significantly improved after the operation (p < 0.01). Further, the proportion of fully ambulatory patients increased from 12% before surgery to 69% at 1-day and 92% at 3-months after surgery, respectively (p < 0.01) (Fig. 5).

The radiologic height measurement of all the 35 vertebrae treated demonstrated that the Cobb angle improved from an mean angle of 18.5 ± 8.2 degrees before surgery to 9.2 ± 4.0 degrees after surgery (p < 0.01). The anterior vertebral height significantly increased from 18.4 ± 5.1 mm before surgery to 20.5 ± 5.3 mm after surgery (p < 0.01). The midline height significantly increased from 15.5 ± 5.2 mm before surgery to 18.9 ± 4.0 mm after surgery (p < 0.01). The posterior heights remained unchanged (29.9 ± 3.5 mm and 30.2 ± 3.2 mm before and after surgery, respectively) (p > 0.05) (Fig. 6).

In all cases, a cavity was successfully created in the treated vertebral body. Minor cement extravasations were observed in 4 out of 35 procedures, but none of the leakages were associated with any clinical consequences. These leakages included one in the epidural space, one in the disc space, and two in the paraspinal tissues or paravertebral veins (Fig. 7).