Comparison of unipedicular and bipedicular balloon kyphoplasty for the treatment of osteoporotic vertebral compression fractures

Osteoporosis is the most common metabolic bone disease in the United States, characterised by decreased bone mass, alteration of bone microarchitecture and increased risk of fragility fracture.¹ The most common fragility fractures associated with osteoporosis are vertebral compression fractures, affecting 25% of postmenopausal women and accounting for over 700 000 fractures annually in the United States.^2-4 The presence of vertebral compression fractures can lead to chronic and disabling symptoms.^1-5 In 2005 it was estimated that the annual cost of fractures related to osteoporosis in the United States was $16.9 billion, with the expectation that this figure would rise to $25.3 billion by 2025.⁵ The aim of surgical treatment is to address the associated morbidity by providing pain relief and early stabilisation of the fractures.

Balloon kyphoplasty achieves reduction of the vertebral fracture using a balloon tamp inserted into the vertebral body by a transpedicular approach, followed by fixation of the fracture fragments with polymethylmethacrylate (PMMA) bone cement.^6,7 It has been shown in randomised studies that balloon kyphoplasty provides significant pain relief and improved function.^8,9 It has also been shown to be cost-effective compared with non-surgical management¹⁰ and to restore vertebral body height.¹¹ The American Academy of Orthopaedic Surgeons (AAOS) has recommended that kyphoplasty should be considered for patients with osteoporotic vertebral compression fractures.¹² The standard technique for kyphoplasty involves a bipedicular approach using two balloon tamps,⁶ but recently a unipedicular approach has been advocated, reducing the operating time and risks, and increasing the cost-effectiveness of the procedure.¹³ The purpose of this study was to determine whether a uni- or bipedicular technique was superior in terms of pain relief, improvement of function and correction of deformity, based on radiological measures. Additionally, we evaluated the operating time, the amount of cement injected and the incidence of cement extravasation associated with each technique.

Patients and Methods

Patients were identified and recruited for this prospective randomised trial from the metabolic bone disease service at our institution. The trial had ethical approval and all patients gave informed consent.

A power analysis was conducted to detect a five-unit difference in the Oswestry Disability Index (ODI)¹⁴ between uni- and bipedicular kyphoplasty, with the assumption of a ten-unit standard deviation (sd) for each group. It was calculated that a minimum of 64 patients per group would be needed to achieve power of 80%. Inclusion criteria involved male and female patients aged > 50 years, with an acute vertebral compression fracture causing pain and functional limitations in their daily activities. This included patients with ≥ 40% collapse of the vertebral body, on lateral plain radiographs, or a score of ≥ 4 (of 10) on a visual analogue scale (VAS) for pain. All patients were treated conservatively for four to six weeks before enrolment in the study. Acuteness of the fracture was assessed on pre-operative MRI in which there was evidence of oedema in the fracture as seen on short-tau inversion recovery sequences, indicating a recent fracture.¹⁵ Exclusion criteria included patients with primary or secondary bone tumours, pre-existing chronic pain or functional disability unrelated to a vertebral fracture, a fracture that was secondary to trauma or those in whom both pedicles of the fractured vertebrae could not be visualised on intra-operative fluoroscopy.

Between July 2006 and February 2009, a total of 44 patients were enrolled into the study (Fig. 1). There were 23 patients in the unipedicular group and 21 in the bipedicular group, with a total of 56 vertebral compression fractures between the T5 and L5 vertebrae, which were treated by balloon kyphoplasty. There were no differences in age, gender, American Society of Anesthesiologists (ASA) score,¹⁶ fracture location or number of fractures, which were treated per patient between the groups (Table I).

Under general anaesthesia patients were placed prone on a radiolucent Jackson table and two image intensifiers were positioned to provide simultaneous anteroposterior (AP) and lateral orthogonal views of the fractured vertebral body. All received a single dose of first-generation cephalosporin intravenously immediately before surgery. A 1 cm incision was made lateral to the pedicle of the affected vertebra in those in the unipedicular group. A Jamshidi Bone Biopsy Needle (Cardinal Health, Dublin, Ohio) was introduced into the pedicle and advanced into the centre of the vertebral body at an angle of 30° to 45° relative to the AP axis. A lateral starting point was used for both thoracic and lumbar vertebrae in order reach the midline of the vertebral body while preventing a medial broach through the pedicle.¹⁷ A guide wire was then placed through the Jamshidi needle and the needle removed. A series of dilating cannulae were then advanced over the guide wire until a working cannula was in place. A bone biopsy was taken at this point for histology to confirm a diagnosis of osteoporosis and exclude malignancy. A 15 mm or 20 mm bone tamp (Kyphon Inc., Sunnydale, California) was then introduced into the vertebral body via the cannula and inflated until the balloon was in contact with the subchondral plate, lateral vertebral body wall or anterior cortex of the vertebral body. The balloon was then deflated and removed. Subsequently, cement was injected into the cavity and allowed to harden. Injection was stopped when there was complete filling of the cavity created by the balloon tamp, or if there was a risk of breaching the borders delineated by the tamp. The cannula was then removed and the incision closed.

For patients in the bipedicular group, the same surgical steps were performed through both pedicles. However, the tips of both cannulae were advanced laterally to the midline on the AP view and the cement was introduced simultaneously through both cannulae. All patients remained in hospital post-operatively for at least 24 hours for observation, or until fit for discharge.

The self-reported questionnaires used in this study included the ODI,¹⁴ a VAS for back pain (11-point scale from 0 to 10),¹⁸ and the Roland Morris Disability Questionnaire (RDQ).^19,20 These questionnaires were given to all patients before surgery and at three months and twelve months post-operatively.

Baseline pre-operative radiological data were obtained from lateral radiographs of the spine, and the change in the kyphotic angle (Fig. 2) and the percentage of height of the vertebral body that was restored at the anterior and middle aspect of the fractured vertebrae on the post-operative radiographs was measured. Measurements of the height were made of the uninjured as well as the injured vertebrae (Fig. 3).²¹ Two of the investigators (BJR, BPG) who were blinded to the treatment groups undertook the radiological measurements. The incidence of cement leakage from the vertebral body on post-operative CT scans was recorded, as well as the complications of the procedure. The volume of cement and the length of hospital stay were also recorded.

Statistical analysis

Overall summary statistics were calculated in terms of means and sd for continuous variables and frequencies and percentages for categorical variables. Group differences among continuous variables were evaluated using independent samples t-tests and for discrete variables were evaluated using the chi-squared or Fisher’s exact test. Statistical significance was set at alpha equal to 0.05. All analyses were done using SPSS v14.0 (SPSS Inc., Chicago, Illinois).

Results

There were no differences between the uni- and bipedicular kyphoplasty groups in terms of the pre-operative ODI (p = 0.88), VAS (p = 0.95) and RDQ measurements (p = 0.79). At three months after surgery, both groups had significant improvements in all the self-reported outcome measures with no significant differences between the groups (ODI, p = 0.85; VAS, p = 0.67; RDQ, p = 0.17). The improvement, which was maintained for all self-reported measures in both groups at twelve months post-operatively, was not statistically significant except for the RDQ score in the bipedicular group that showed improvement from three months (10.6 points) to 12 months (5.9 points) (p = 0.008). Similar to the results at three months, there were no significant differences between the groups in all self-reported measures at 12 months post-operatively (ODI, p = 0.90; VAS, p = 0.87; RDQ, p = 0.36) (Fig. 4).

Pre- and post-operative measurements of height were available for 39 levels (70%); 14 patients (18 compression fractures) in the unipedicular group and 17 patients (21 compression fractures) in the bipedicular group. The mean restoration of anterior height was 20.5% (sd 15.0) in the unipedicular group, and 14.8% (sd 10.9) in the bipedicular group (p = 0.17). There were no differences between the groups in the mean restoration of middle height, which was 17.8% (sd 12.8) in the unipedicular group and 13.7% (sd 8.9) in the bipedicular group (p = 0.24). The mean pre-operative kyphotic angle was 26.5° (sd 10.3) and 24.0° (sd 9.9) in the uni- and bipedicular groups, respectively. The mean reduction of the kyphotic angle was also similar in both groups (unipedicular: 4.8 0° (sd 4.0); bipedicular: 4.7° (sd 4.6); p = 0.93). No differences were found between groups (Fig. 5).

Total operating time was significantly less in the unipedicular group, as was the total amount of cement used (Table II). There were no surgical complications in either group. There were nine fractured vertebrae with asymptomatic cement leakage, two in the unipedicular group and seven in the bipedicular group with significant difference between groups (p = 0.09). The length of hospital stay was not significantly different in the two groups (1.0 days (1 to 3) vs 1.1 days (1 to 4) for uni- and bipedicular groups, respectively; p = 0.17).

Discussion

Osteoporotic vertebral compression fractures can lead to decreased quality of life²² and increased mortality.²³ In patients who fail conservative management, good results have been reported following balloon kyphoplasty,^{8,9,13,17,24,25} Despite previous data supporting the use of a bipedicular technique, a unipedicular technique may be preferable as a result of the advantages of reduced operating time, radiation exposure and costs.^17,26 In our prospective, randomised study we found no differences in pain and function following unipedicular and bipedicular kyphoplasty at three and 12 months post-operatively and there was no difference between the groups in the correction of the spinal deformity as measured radiologically.

In a retrospective study of 45 patients, Song et al²⁷ showed that patients who underwent unipedicular kyphoplasty (n = 15) had greater improvement in VAS scores at three days when compared with those undergoing bipedicular kyphoplasty (n = 30), suggesting that immediate post-operative pain scores were decreased with a unipedicular approach. It has also recently been shown that multi-level uni- and bipedicular kyphoplasty can provide significant and comparable clinical improvement using VAS and SF-36 at two weeks and two years post-operatively.²⁶ Although immediate post-operative assessment was not undertaken in our study, we found that patients in both the uni- and bipedicular groups had significant improvement in all outcomes (ODI, VAS, and RDQ) at three and 12 months post-operatively when compared with the baseline levels.

We found that both techniques provided similar restoration of vertebral height and reduction of kyphotic deformity. There were no differences between the groups in the anterior or middle vertebral heights post-operatively. Chen et al²⁸ reported that the restoration of vertebral height was greater in the bipedicular group than in the unipedicular group. Many factors could affect the differences found in vertebral heights post-operatively, including the amount of cement used and the formation of an adequate cavity with the balloon tamp. Although the use of balloon kyphoplasty has been reported to show significant increases in vertebral height and reduction of kyphotic deformity,^7,29-31 we found no difference in the vertebral height post-operatively in the two groups despite more cement being used in the bipedicular group. We found that the operating time was significantly reduced using the unipedicular technique, and while cost analysis was not the aim of our study, it has been previously estimated that if 5% of all vertebral compression fractures in the United States were treated by unipedicular kyphoplasty, instead of bipedicular kyphoplasty, the savings would be > $32 million per year, at a saving of $886 per level.¹³ Thus, our results support the concept that the unipedicular technique is a faster, less expensive option that still provides a comparable correction of spinal deformity to the bipedicular technique.

The main limitations of our study are that we have only examined the balloon kyphoplasty technique. Other techniques of kyphoplasty might yield similar results. The fact that 12 patients were lost to follow-up at 12 months potentially affects the validity of our study. The elderly patient population in our study may have contributed to a high rate of loss to follow-up. It has been suggested that many kyphoplasty patients will not return for follow-up unless they develop a complication or further symptoms.¹⁷ Longer follow-up might lead to a detection of differences between the groups. Also, our pilot study remains underpowered as patient enrolment was limited.

In conclusion we would encourage the use of a unipedicular approach as the preferred surgical technique for treatment of osteoporotic vertebral compression fractures.

No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

This article was primary edited by S. Hughes and first-proof edited by J. Scott.