Hadjipavlou et al ¹ provided a thorough review of the existing literature for these procedures. The reported success rates for these procedures is consistently above 80% (defined as patient-reported good to excellent pain response) with risk for certain complications. These complications include a transient increase in pain, infection, leakage of cement, and secondary vertebral compression fractures. With kyphoplasty, there have been a small number of reports of balloon rupture, but the failed balloons were withdrawn without incident. Cement leakage is the most common cause of pulmonary or neurological complications. The comparison of cement leakage risk between kyphoplasty and vertebroplasty remains controversial. Some reports suggest a clinically insignificant difference in risk, whereas others suggest that kyphoplasty is associated with less leakage.

Both vertebroplasty and kyphoplasty may increase the risk of subsequent vertebral fractures, particularly at the adjacent level. Some have hypothesized that this may be due to changes in load distribution across the endplate that occur when disc is pressure lost after endplate fracture. Fracture of the endplate increases the volume for the nucleus pulposus and reduces its ability to hydrostatically resist compressive load. In flexion, the reduced load on the nucleus causes greater load on the annulus and the anterior cortex of the vertebral body adjacent to the fractured endplate. This mechanism is being investigated by Patwardhan et al.² Some have speculated that the addition of cement to the vertebral body increases the stiffness of the vertebral body and that this may play a role in subsequent fractures. The effect of cement on the treated and adjacent levels is still being studied, but it appears that this effect is small compared to bone mineral density.³

Advocates of kyphoplasty believe that the procedure actually reduces the rate of adjacent level fractures, compared to vertebroplasty, because it more effectively reduces the fracture. However, no randomized studies have been conducted to compare the two techniques, and the natural history of adjacent level fractures has been difficult to quantify. Frankel and Vandergrift ⁴ reviewed the results of 2,000 patients enrolled in a trial evaluating bisphosphonate in patients with vertebral compression fractures. The authors noted that the rates of new fractures were 7.9% and 15% in patients treated with bisphosphonate and placebo, respectively. Moreover, in the bisphosphonate group, only 3.4% of new vertebral compression fractures were at the adjacent level, compared to 7.1% in the placebo group.⁴ Frankel and vandergrift review of the literature found the rates of subsequent adjacent level fractures following kyphoplasty to be 13% compared to 10% with vertebroplasty, suggesting that cement implantation with both of these techniques increased the risk of subsequent fractures compared to natural history.⁴

Frankel et al ⁵ compared outcomes in a series of 17 patients (20 fractures) undergoing kyphoplasty and 19 patients (26 fractures) undergoing vertebroplasty. The authors reported an average of 4.65 ml and 3.78 ml of cement per vertebral body with kyphoplasty and vertebroplasty, respectively. There were five adjacent level fractures in three kyphoplasty patients (3/17 [18%]) and none in the vertebroplasty group.⁵ Fribourg et al⁶ published a retrospective review of 38 patients (47 fractures) treated with kyphoplasty. Patients received between 1.5 and 6.0 ml cement per vertebral body. The authors reported that 10 patients (26%) had a subsequent fracture during the follow-up period (average 8 months), and 8 of those patients had a subsequent fracture within 2 months. The 8 patients with early “new fractures” all had a fracture at the adjacent level and the 2 patients with later “new fractures” all had fractures that were not adjacent to the index fracture.⁶

A larger study was performed by Harrop et al ⁶ in which 115 patients were treated with kyphoplasty. All patients had at least 3 months follow-up. In this group, 26 patients (22.6%) developed 34 new compression fractures. The authors then classified patients as having primary osteoporosis (80 patients) or secondary steroid-induced osteoporosis (35 patients) and calculated at the rate of subsequent fractures in each group. They reported that the incidence of postkyphoplasty compression factures in primary osteoporosis patients was 11%, and the incidence in the steroid-induced osteoporosis group was 49% (p < .00001).⁷ There was no mention of the use of bisphosphonates in these patients.

The Frankel ^4,⁵ and Harrop⁷ papers demonstrate that both bisphosphonates and steroids have a significant effect on bone quality and should be considered in any analysis of adjacent-level fractures following vertebroplasty or kyphoplasty. Each of these papers reported subsequent fractures in 18% to 26% of patients, with the best case being 11% in patients specifically with primary osteoporosis. Because the natural history of adjacent level fractures is likely between 8% and 15%, standard vertebroplasty and balloon kyphoplasty may actually increase the risk of subsequent compression fractures, despite the success in reducing a patient’s pain from the index fracture.⁴^–⁷

Cadaver testing has been used to test the hypothesis that kyphoplasty is more efficacious for deformity correction than standard vertebroplasty. Belkoff et al ⁸ experimentally created compression fractures in 16 osteoporotic vertebral bodies and treated them with either balloon kyphoplasty or vertebroplasty. The vertebral bodies were compressed to 25% of their initial height; however, there was an initial elastic recovery of about 15%. The authors measured the change in height with the application of cement and then subjected those vertebral bodies to compressive failure. The authors reported that 97% of the height loss was regained with kyphoplasty, whereas only 30% of height loss was regained with vertebroplasty.⁸ These results may not reflect the in vivo situation, because muscle forces and body weight will resist height restoration, as measured clinically by Voggenreiter.⁸ The vertebral bodies in both groups were found to be stronger after the application of cement. However, those treated with kyphoplasty were found to return to their initial stiffness, whereas those treated with vertebroplasty did not.⁹ Kim et al¹⁰ performed a similar cadaver evaluation with the addition of cyclic loading to determine how vertebral fracture correction was maintained over time. They reported that balloon kyphoplasty was able to restore vertebral height, but there was significant loss of height over 100,000 cycles of compressive load. Vertebroplasty was better able to maintain height under dynamic loading. Ultimately, after the cyclic testing regimen, the vertebrae treated with kyphoplasty had less height than those treated with vertebroplasty.¹⁰ In contrast to Belkoff,⁸ the vertebral bodies treated with vertebroplasty were more stiff than with kyphoplasty.

Cadaver studies of isolated vertebral bodies cannot capture the interaction between vertebral bodies or in vivo loads. Clinical data are necessary to realistically measure the ability to achieve reduction of a vertebral compression fracture. Pradham et al ¹¹ evaluated a series of 65 consecutive patients treated with kyphoplasty between 1 to 3 levels. Kyphoplasty reduced the local kyphotic deformity by an average of 7.3 degrees (63% of preoperative kyphosis), but this did not translate into a similar correction of overall sagittal alignment. Angular correction decreased to 2.4 degrees when measured from the level above to the level below. Similarly, the reductions decreased to 1.5 and 1.0 degree at spans of 2 and 3 levels above and below the index level, respectively. The authors concluded that it was unrealistic to expect a 1- or 2 -level kyphoplasty to significantly improve sagittal alignment after vertebral compression fracture.¹¹

The StaXx FX Structural Kyphoplasty System (Spine Wave, Inc., Shelton, Conn.) was introduced to allow the physician to reduce the vertebral fracture and to correct the kyphotic deformity with a system of progressively stacked wavers made from PEEK (Figure 41-1). The permanent implant system allows controlled vertical expansion in situ and eliminates the intraoperative height loss that may occur after deflation of a balloon. Pradhan et al¹¹ remarked that using balloons to reduce the fracture is not ideal, because the balloon and subsequently inserted cement follow a path of least resistance, resulting in localized stresses on the endplate. These localized stresses compromise the endplate’s ability to maintain an improvement in the spine’s overall sagittal alignment. The geometry of the StaXx system includes a wide, flat surface to support the endplate, which encourages hydrostatic compression of the nucleus and normalization of load across the disc. This endplate support may enable the system to reduce the number of adjacent level fractures following treatment of an initial compression fracture. Tactile feedback and manual wafer implantation provide greater physician control and directed axial expansion to reduce the fractured endplate.