Recompression in new levels after percutaneous vertebroplasty and kyphoplasty compared with conservative treatment

Since 1987, percutaneous vertebroplasty (PVP) and kyphoplasty (PKP) with polymethylmethacrylate (PMMA) augmentation have been highly advocated as treatment techniques for osteoporotic vertebrate compression fractures (OVCFs) [1]. Although the overall satisfaction rate ranges between 70 and 90 %, and complication rate associated is reported to be very low, some authors have indicated that PVP/PKP increases the risk for subsequent vertebral fractures [2‐5].

Multiple covariate analysis, such as patient characteristics and procedural techniques, has been used to identify risk factors [6‐7]. But it is still unclear as to what the exact relevant element is and whether such fractures are procedure-related or part of the natural course of osteoporosis. There also have been no reports to date of a large series of new vertebral fractures that developed within adjacent or nonadjacent levels comparing PVP/PKP and conservative treatment. In our original study, we prospectively investigate the incidence of new symptomatic OVCFs in patients treated by PVP/PKP or conservative therapy and discuss the possible causative mechanism of refracture.

Our study was performed at the Peking University First Hospital Orthopedics Department and approved by Institutional Review Board. Herein we report our first hand data statistics regarding China mainland.

No major adverse events were observed during procedure or follow-up period. 87.6 % patients of PVP/PKP group returned to their normal lives 1 week after operation and 59.2 % patients of conservative group exercise to their regular level 2 months later. In total, 31 of 290 (10.7 %) patients returned with 42 new OVCFs (Table 1). 14 individuals (5 males, 9 females) were initially treated by surgical technique (PVP 9, PKP 5) and 17 ones (7 males, 10 females) were from conservative group. 25 patients were subjected to repeated PVP/PKP and the other six patients were successfully managed by conservative care. Of these newly documented fractures, 14 (34.2 %) involved adjacent segments, whereas 25 (59.5 %) occurred at distant levels, and another three (7.3 %) were the same.

A total of 1.5 million new fractures, nearly half of which are vertebral (700,000), are reported in the US each year, outnumbering fractures of the hip and ankle combined. Patients with osteoporotic vertebral compression fractures often suffer from acute episodes of pain, increased morbidity, prolonged hospitalization as well as long-term disability, as the natural healing history of OVCFs is often accompanied by 1–3 months bedrest. Also, the risk of pain and disability increases progressively with the number and severity of vertebral deformities. They are not typically associated with neurological deficit, since the fractures do not extend into the spinal canal nor compress the neural elements. Generally, acute pain resolves in 4 weeks to 8 months. In the years before PVP/PKP invention, conservative management of OVCFs consists of bed rest, analgesic medications, bracing, and physical therapy. However, some patients do not respond to these therapies in either terms of pain relief or progression of deformity. Long-term staying in bed can also bring out a series of complications, such as pneumonia, anorexia, indigestion, malnutrition, even stroke. An ideal therapy should both address fracture-related pain and avoid the associated complications in a minimally invasive fashion. Since 1990s, PVP/PKP had become standard procedures for OVCFs regarding effective pain reduction and relatively low incidence of complications. But recently, many clinical investigations have reported an increased fracture risk of adjacent levels. Hulme et al. [9] reviewed 69 clinical series evaluating both kyphoplasty and vertebroplasty: 766 patients after PKP had 115 new fractures and 66 % were located at adjacent levels. Mazzantini et al. [10] found a 27.8 % cumulative incidence of new vertebral fractures in a cohort of 115 patients treated with PVP after a follow-up of mean 39 months; 68 % occurred next to pretreated segments. Tseng et al. [11] investigated 852 patients (1,131 vertebrae),141 patients (16.6 %) after PVP experienced new OVCFs, 58.8–63.8 % of which were adjacent fractures. Similarly, Trout et al. [12] also reported a significant increase (41.4 %) in the incidence of fractures in the vicinity of cemented vertebrae on evaluation of 432 patients.

Unfortunately, much of the current literature reports their studies comparing original fractures and new fractures mainly in operative group, there are only several limited data with respect to recompression incidence of PVP/PKP versus conservative treatment. Rikke et al. [13] reported that the relative risk for all new fractures in the PVP group compared with the conservative group was 2.9 after 3 months and 1.3 after 12 months. However, Klazen et al. [14] observed 202 patients; there was no difference between PVP and conservative therapy.

In the current study, we found a similar incidence of recompression comparing PVP/PKP (8.28 %) with conservative therapy (14.0 %) on accordance of patient number, which is coincident with Klazen’s conclusion. From the standpoint of vertebrae, adjacent recompression occurred more frequently than distant levels and decreased gradually with the accretion of intervals in the PKP/PVP group. By contrast, no higher incidence of recompression in adjacent levels was observed in the conservative group. Prospective and retrospective studies reported an incidence of new fractures of 12.5–36.8 % after PVP/PKP [10, 15, 16]; our data seem to be favorable and could be the consequence of long-term compliance. In addition, we only registered symptomatic fractures, whereas in most of higher rate reports recompression was defined by imaging follow-up. Moreover, in our experience, we intended to only fill the cavity left by the balloon or the fracture fissure rather than augment the entire vertebral body. Most patients followed our guidelines for anti-osteoporosis therapy, which may also interpret the difference in refracture ratio. Tanigawa et al. [16] reported that 56 new OVCFs occurred in 28 (36.8 %) of 76 patients after PVP, and 38 (67.8 %) of these fractures occurred in adjacent levels, more than half occurred within 3 months.

The exact mechanism for recompression is still unclear; several authors had published their clinical and biomechanical investigations. Most of them indicate that load transfer mechanism after augmentation is remarkably changed with increased stresses and strains in the levels of vicinity. Michael et al. [17] found the depression of the fractured endplate altered the pressure profile of the damaged disk resulting in increased compressive loading of the anterior wall of adjacent vertebra that predisposed it to wedge fracture. Rohlmann et al. [18] also suggest that a wedge-shaped fracture increases the flexion bending moment due to the upper body weight and thus a higher muscle force in the erector spinae is required to balance the spine, which results in a higher spinal load and a higher intradiscal pressure. The erector spinae is a long muscle and thus its force affects intradiscal pressure not only at adjacent levels but also the whole region. Our findings agreed with this contention for the location of refracture body all residing in the vicinity of eight levels. Viewed from this point, kyphoplasty may be an ideal alternative method in preventing further collapse compared with vertebroplasty for its rectification of kyphotic deformity. In this study, we observed that refracture incidence in PKP group is lower than that in PVP group and the conservative group (6.33 versus 10 %, 14.05 %, p = 0.426, 0.088), but not significantly. Most clinical and basic findings [19‐21] asserted that the rate of developing new fractures was not strongly influenced by the volume of cement injected, as in the present study the average amount of initial bone cement is similar between single time fracture unit and refracture unit in the PVP/PKP group (3.95 ± 1.43 versus 3.26 ± 1.54 mL, p = 0.22).

In our current report, the rate of refracture concerning adjacent bodies is truly significantly higher than distant ones in PVP/PKP group (2.5 versus 0.46 %, p = 0.000), but the difference is not conspicuous in the conservative group (1.91 versus 1.13 %, p = 0.471). Nevertheless, the unitary potentiality in the PVP/PKP group is inversely lower than in the conservative group (0.71 % versus 1.32 %, p = 0.042) from the standpoint of vertebrae. From these data we can extrapolate that both previous fracture and cement injection procedure increase the risk of new adjacent fractures, but cement injection procedure increases the risk more significantly. However, the odds is relatively low and has no significant differences compared with conservative treatment (2.5 versus 1.91 %, p = 0.636). Therefore, patients who experienced PVP/PKP may have a relatively low risk of developing new fractures, not to mention its immediate pain relief, early mobilization, and other short-term benefits.

The timing of postoperative OVCFs in our study also seems to be more favorable than that in prior studies, which have shown that between 43 and 67 % of subsequent incident fractures occur within 3 months. In the PVP/PKP group, 71.42 % of the second fractures take place within the first year, while only 47.05 % of the second fractures in the conservative group do in the same period. Mean interval was much shorter in PVP/PKP group (9.72 ± 17.8 months) than in conservative group (22.48 ± 7.99 months, p = 0.017). Both Tseng et al. [11] and Trout et al. [12] asserted that adjacent fractures emerged sooner than distant levels following cement injection. The period from a fresh compression fracture to the first adjacent compression fracture was 78.2 ± 97.1 days and to the first distant compression fracture was 231.9 ± 224.6 days according to Tseng et al. While in our clinical follow-up a similar result was observed in the PVP/PKP group, the data are 8.95 ± 7.34 and 10.75 ± 8.68 months, respectively (p < 0.05). For conservative treatment, the new adjacent segments occurred even more delayed than nonadjacent segments although no statistical discrepancy has been reached (28 ± 3.46 versus 18.8 ± 14.3 months, p > 0.05). So we can hypothesize that cement injection may precipitate the refracture, particularly to adjacent segments.

No same level refracture was observed following PVP/PKP, while in conservative group, the refracture rate was relatively high (2.05 %) and occurred sooner than that in any other unit (7.33 ± 3.29 months). Thus usage of operation procedure may be advantageous in holding back further refracture of the same segment. However, even cemented vertebrae refracture occasionally, Lin et al. [6] reported an incidence of 63 % in their retrospective clinical analysis, and concluded that significant anterior vertebral height restoration increases the risk of subsequent fracture after vertebroplasty. This conclusion challenges some theories we have discussed above; one problem is how to guarantee enough kyphotic angle rectification on the one hand while avoiding greater anterior vertebral height restoration on the other.

Johannes et al. [8] demonstrated a remarkable propensity of refractures within three levels above or below preexisting fractures. From our data, we got a similar outcome; 83.3 % of new OVCFs in PVP/PKP group and 56.2 % in conservative group support this interpretation. The locations in thoracolumbar segments where original fractures distributed in are prone to refracture both after PMMA injection and conservative therapy. New fractures tend to occur more frequently in thoracic vertebrae in both groups but no conspicuous discrepancy. The issue of multitime fractures (>2) after PVP has been raised and analyzed to some extent by Tseng et al. [11]. In our study these data were so limited and, therefore, could not be compared.

Another risk factor of new fractures of adjacent vertebral bodies is cement leakage into the disk. Pitton et al. [22] studied 191 patients in whom 385 vertebroplasties were performed. The overall rate of cement leaks was 55.6 %, including all leaks detectable by CT. Ten of 30 adjacent fractures occurred in the presence of preexisting intradiscal cement leaks and six belonged to 11 sandwich fractures of 29 sandwich situations after mean 1.5 months. However, in three cases (1.77 %) cement leakage occurred during operation in our institution and none developed refracture. In the meantime, no sandwich fracture was observed in a total of 27 sandwich situations in both groups. Pitton et al. identified that the sandwich situation changed the contextual environment because of the stiffening and the loss of vertical elasticity of the two proximate vertebrae but was not associated with an increased secondary fracture rate. Our data support their viewpoint, yet no evidence is available until now. Further researches are needed to focus on this specific constellation. In addition, older age, gender, fracture times, operation modality (PVP or PKP), and initial number of OVCFs were also documented, but they were not the influencing factors in developing new symptomatic OVCFs in this study.

We selected a calculating method similar to that of Johannes et al. [8], in which adjacent and nonadjacent segments were divided, because the total number of adjacent vertebrae generated by first time fracture is relatively small compared with that of distant levels. As a result, only simply computing the incidence of adjacent refractures by dividing the sum number of new OVCFs cannot reflect the condition practically. However, if count the refracture incidence by simply calculating the patient number, the above two therapeutics could bring out similar results.

Some issues remain unanswered even in the current study. We discussed the effects of PVP/PKP and conservative therapy only clinically, not biomechanically, and long-term effects are still unknown. Furthermore, whether resort to cement injection or let it alone in terms of sandwich situation is not known. Therefore, further research using larger patient populations and in vitro experiment may help to resolve the unclear results concerning this study. Moreover, the results of this study need to be confirmed by additional high-quality RCT clinical studies.

From most data reported in the literatures, the risk of experiencing new OVCFs increased after PVP/PKP. However, in this prospective study, compared with conservative therapy, patients who had experienced PVP/PKP were not associated with an increased risk of recompression in new levels with a mean 49.4-month follow-up period. But one point that new OVCFs occurred sooner after cement injection than that after conservative treatment must be taken into consideration. The presence of sandwich situation is not associated with an increased secondary fracture rate and no interrelated factor has been confirmed to be influential ingredient for subsequent fractures. Existing fractures are strong independent predictors of the risk of future vertebral fractures so we tend to believe that it is more of a natural process other than induced by PVP/PKP.