Insufficient Augmentation of Bone Cement Causes Recompression of Augmented Vertebra after Balloon Kyphoplasty.

INTRODUCTION: Balloon kyphoplasty (BKP) is one of the most frequently used clinical methods to relieve pain caused by osteoporotic vertebral compression fracture (OVCF); it can effectively improve the body height of the vertebra. However, recompression of the augmented vertebra (RAV) is often observed after BKP. This study aimed to report factors that are associated with RAV in terms of cement augmentation.
METHODS: A total of 78 patients (women, 60; men, 18) were included in this study. RAV was defined as anterior vertebral height loss (VHL), between immediate postoperation and 3 or 6 months after BKP, of more than 5.0 mm. Cement augmentation ratio (CAR) was calculated as the ratio of the maximal height of polymethylmethacrylate (PMMA) to the maximal distance between both end plates. Age, gender, fracture age, CAR, presence of medication for osteoporosis, intervertebral cleft (IVC), and cement leakage were evaluated using Fisher's exact test or Mann-Whitney U test to compare between RAV and non-RAV groups. Aforementioned variables were also analyzed using multiple logistic regression test. A P<0.05 was considered statistically significant.
RESULTS: The incidence rates for RAV at 3 and 6 months were 35.9% (28/78) and 38.5% (30/78), respectively. Age (80.1 vs 74.7) was significantly higher in the RAV group, whereas CAR (69.4% vs 77.6%) was lower in the non-RAV group. A multivariate regression analysis revealed that age (odds ratio (OR)=1.12, P=0.001) and CAR (OR=0.91, P=0.001) were independently associated with RAV.
CONCLUSIONS: RAV was observed in 38.5% of patients in this study. Older age and low CAR were independently associated with RAV. To prevent RAV, especially in the elderly, augmented PMMA should come in contact with both end plates.

Introduction

Osteoporotic vertebral compression fracture (OVCF) is one of the most common complications of osteoporosis in the elderly population[1]) and has a significant negative impact on the quality of life of patients. In Japan, the percentage of elderly individuals (age ≥65 years) in the population reached 25% in 2013; it is expected to exceed 30% by 2025[2]). In 2008, the incidence of painful OVCF was estimated to be 880,000, and approximately 40% of patients with painful OVCF are hospitalized owing to pain severity[3],[4]).

Because pain from acute fractures usually lasts 4-8 weeks[5]), conservative medical therapy is employed in most OVCF cases. However, traditional conservative treatments for elderly patients with OVCF are associated with high mortality[6]). In addition, progressive vertebral collapse or insufficient union following conservative treatment can trigger continuous back pain or delayed neurologic complications[7-9]).

Percutaneous vertebroplasty (PVP) is one of the most frequently used clinical methods in relieving pain caused by OVCF[10-13]). In addition, balloon kyphoplasty (BKP), which was derived from PVP, can effectively improve the body height of the vertebra and is well documented in surgical literature[14-18]). Vertebral augmentation leads to an instantaneous recovery of stability and strength; thus, it prevents continuous micromotion and further collapse.

However, during patients' follow-up, recompression of the augmented vertebra (RAV) with significant loss of vertebral height and aggravation of local kyphotic deformity was observed after BKP. Although a few studies discuss about recompression after successful BKP[19-22]), the risk factor related to recompression and the clinical significance of recompression remain vague. In our cases after BKP, recompression mainly occurred at the polymethylmethacrylate (PMMA)-unsupported portion of the vertebral body. Hence, we hypothesized that these residual unfilled spaces within the vertebrae may lead to recompression during the follow-up periods if the vertebrae are inadequately filled. The present study aimed to report factors that correlated with RAV after BKP in terms of cement augmentation.

Materials and Methods

This study was approved by institutional review board. To rule out bias in the selection of patients and extravertebral effects, the inclusion criteria included the following: (1) single-level OVCF diagnosed by plain radiography and magnetic resonance imaging, (2) patients suffering severe fracture pain with failure of adequate conservative treatment, (3) treatment with single-level BKP via bilateral portals, (4) follow-up period of at least 6 months, and (5) no additional history of trauma after surgery. The exclusion criteria were as follows: (1) non-OVCF, including fractures related to malignancy or infection; (2) patients with hyperparathyroidism, hyperthyroidism, pituitary adenoma, hypercortisolism, or other bone metabolic disease; (3) life-threatening systemic disease such as malignancy, systemic infection, or serious disorder of a vital organ; and (4) patients whom we could not collect complete medical record or radiographic data.

Overall, this study retrospectively reviewed 122 patients receiving BKP from January 2011 to October 2014 in our single institution. After excluding 44 patients, data from 78 eligible patients (women, 60; men, 18) were analyzed. The average age was 76.8 years (range, 59-89 years). The average duration between the time of injury and surgery (fracture age) was 60.7 days. Vertebral bodies treated by BKP were distributed between the T6 and L5 levels, which were most prevalent in the thoracolumbar junction (Fig. 1).

Figure 1.

Vertebral body distribution treated by BKP.

The BKP was performed using specialized instruments (KyphonⓇ, Medtronic, Minneapolis, MN, USA) and PMMA cement via bilateral portals under general anesthesia according to routine procedures. The use of a tailor-made soft corset was prescribed for all patients after surgery for 2-3 months. Osteoporotic medications, including bisphosphonates, vitamin D, teriparatide, or raloxifene, were postoperatively used.

Each augmented vertebra was sequentially measured for its height of anterior vertebra (HA). HA was defined as the distance between the upper and lower end plates at the anterior vertebra. Data were obtained from lateral radiographs and analyzed at four different time points: preoperative, immediate postoperative, 3 months after the procedure, and 6 months after the procedure. A lateral radiograph was taken in prone position at immediate postoperation, in standing position at preoperation, and 3 and 6 months after the procedure. The vertebral height loss (VHL), which indicated RAV, was then calculated as the difference of HA between immediate postoperation and 3 or 6 months after the procedure. RAV was defined as VHL of more than 5.0 mm (Fig. 2). Moreover, as an indicator of cement augmentation, we defined a cement augmentation ratio (CAR) that was calculated as the ratio of the maximal height of PMMA to the maximal distance between the upper and lower end plates in immediate postoperative radiograph (Fig. 3). The presence of intervertebral cleft (IVC), cement leakage, and reoperation was also checked after the BKP procedure. An IVC was defined as an abnormal linear or cystic lucent region on preoperative lateral radiographs or an intravertebral fluid-/gas-filled transverse zone on preoperative CT scans. Cement leakage was evaluated using immediate postoperative front and lateral radiographs. To investigate the risk factor of RAV, clinical data, such as age, gender, fracture age, presence of medication for osteoporosis, presence of IVC, CAR and presence of cement leakage, were evaluated using chi-square test or Fisher's exact test to compare between the two patient groups (RAV and non-RAV). Descriptive statistics were expressed as mean±standard deviation for continuous variables and as percentages for categorical variables. The mean age, fracture age and CAR between the two groups were analyzed using Mann-Whitney U test. Regardless of their statistical significance, the abovementioned variables were included in the multivariate analysis. Odds ratios (ORs) for RAV and their 95% confidence intervals were calculated using multiple logistic regression test and backward selection. A P<0.05 was considered statistically significant. The SPSS 23.0 statistical software (SPSS, Inc., Chicago, IL) was used for the analysis.

Figure 2.

The measurement of HA, calculation of VHL, and definition of RAV and non-RAV.

Height of anterior vertebra (HA): A, A’.

Vertebral height loss (VHL): A-A’.

Recompression of augmented vertebra (RAV): VHL≧5mm.

Non-RAV: VHL<5mm.

Figure 3.

The calculation of CAR.

B: Max distance between the upper and lower endplates.

C: Max height of PMMA.

Cement augmentation ratio (CAR) =C/B×100.

Results

The incidence rates for RAV at 3 and 6 months after the BKP procedure were 35.9% (28/78) and 38.5%, respectively. Table 1 presents baseline characteristics of RAV and non-RAV groups and results of the univariable analysis. There were no statistical differences in gender, presence of medication for osteoporosis, IVC, and cement leakage. Age in the RAV group was significantly higher (age, 80.1 years vs 74.7 years, P<0.01), while CAR and fracture age were lower compared with the non-RAV group (CAR, 69.4% vs 77.6%, P<0.01; fracture age, 49.3 days vs 67.8 days, P<0.01). A multivariate regression analysis revealed that age (OR=1.12, P=0.001), and CAR (OR=0.91, P=0.001) were independently associated with RAV (Table 2). Two reoperations were performed in the RAV group. In both cases, posterior decompression and instrumented fusion were performed for paraplegia associated with RAV (Fig. 4).

Table 1.

Univariate Analysis of Clinical Data Between RAV and Non-RAV Groups.

Clinical data	RAV group (n=30)	Non-RAV group (n=48)	P-value
Age (years), mean±SD	80.1±5.1	74.7±6.8	<0.01*
Gender, n (%)			1.00
Male	7 (23.3)	11 (23.0)
Female	23 (76.7)	37 (77.1)
Fracture age (days), mean±SD	49.3±97.7	67.8±64.1	<0.01*
Presence of medication for osteoporosis, n (%)			0.50
Yes	25 (83.3)	43 (89.6)
No	5 (16.7)	5 (10.4)
Presence of intervertebral cleft, n (%)			0.574
Yes	25 (83.3)	36 (75.0)
No	5 (16.7)	12 (25.0)
CAR (%), mean±SD	69.4±12.1	77.6±7.4	<0.01*
Leakage of bone cement, n (%)			1.00
Yes	5 (16.7)	7 (14.6)
No	25 (83.3)	41 (85.4)

RAV, recompression of augmented vertebra; CAR, cement augmentation ratio; *, P<0.05

Table 2.

Multivariate Logistic Regression Analysis of Factors Correlated with RAV.

Variables	B	SE	Wald	P	ORs (95% CIs)
Age	0.18	0.06	10.4	0.001	1.12 (1.07–1.34)
CAR	−0.99	0.03	10.1	0.001	0.91 (0.85–0.96)

B, regression coefficient; SE, standard error; ORs, odds ratios; CAR, cement augmentation ratio; CI, confidence intervals

Figure 4.

Two cases in which reoperations were performed due to paraplegia after BKP. (a–b) A 76-year-old female. (a) Sagittal CT image showing spinal cord compression due to recompression of the augmented L1 vertebra (white arrow). The superior adjacent vertebra was also fractured due to the augmented PMMA mass (yellow arrow). (b) Postoperative lateral radiograph showing posterior decompression and instrumented fusion from T11 to L3. (c–d) A 86-year-old female. (c) Sagittal CT image showing spinal cord compression due to recompression of the augmented T12 vertebra (white arrow). (d) Postoperative lateral radiograph showing posterior decompression and instrumented fusion from T10 to L2.

Discussion

Although BKP can effectively restore the collapsed vertebral body height, RAV could also occur after the procedure. Li et al.[22]) revealed that the recompression rate was higher in patients who underwent BKP than those receiving PVP treatment. In this study, RAV after BKP was observed in 38.5% of the patients who underwent BKP in our single institution. Kim et al.[19]) and Niu et al.[21]) reported that the incidence rate for RAV was 14.0% and 12.5%, respectively. Although the definition of RAV in their studies was stricter than ours, the incident rate for RAV in our study (38.5%) was higher than that in their study. The high incidence rate in our study may be derived from differences in the intervention times. The mean fracture age in the study by Kim et al.[19]) was 16.6 days, and almost all vertebrae (76.9%) in the study by Niu et al.[21]) were treated within 42 days after the onset of symptoms. In contrast, the mean fracture age in our study was 60.7 days (Table 3). In cases of older fractured vertebra with a sclerotic bone environment, the augmented PMMA mass volume is prone to be restricted by the sclerotic bone area, and the remaining non-augmented cancellous bone area consequently results in RAV after the procedure. Takahashi et al.[23]) suggested that delayed surgery was related to the progression of vertebral recompression after BKP.

Table 3.

Comparison of Incidence Rates for RAV, Definitions of RAV, and Fracture Ages Between Previous Studies and Our Study.

Study	Incidence rate for RAV	Definition of RAV	Fracture age
Kim et al. 2010[19)] (n=80)	12.5% (n=10)	A decrease in HA of more than 1 mm between immediate postoperation and final follow-up	16.6 days on average
Niu et al. 2015[21)] (n=131)	14.0% (n=17)	A decrease in anterior, middle, or posterior vertebral height of more than 4 mm between 1 day after surgery and final follow-up	<42 days (n=93/76.9%)>42 days (n=38/23.1%)
Our study (n=78)	35.9% (n=28) (at 3 months after BKP) 38.5% (n=30) (at 6 months after BKP)	A decrease in HA of more than 5 mm between immediate postoperation and 3 or 6 months after surgery	60.7 days on average

RAV, recompression of augmented vertebra; BKP, balloon kyphoplasty; HA, height of the anterior vertebra

This study was the first to analyze CAR as an index of cement augmentation, which identified low CAR as a risk factor for RAV. Previous studies have reported that inadequate augmentation with residual unfilled spaces within the vertebrae was an important factor related to recompression. Kim et al.[19]) found that there were non-augmented bony areas between PMMA and the two end plates, which resulted in greater height loss in BKP. Li et al.[20]) also revealed that the distance between PMMA and the end plates was an important risk factor for RAV after BKP. Kim et al.[24]) conducted a biomechanical study using osteoporotic cadaveric fractured vertebral bodies to investigate the behavior of fractured osteoporotic vertebral bodies treated with either PVP or BKP under repetitive loading conditions. In their study, they found that cyclic loading compressed the osteoporotic cancellous bone between PMMA and the end plates and resulted in VHL. Consistently, this study found that low CAR with residual cancellous bone within the augmented vertebra contributed to RAV after BKP.

This is the first report that identified older age as risk factor for RAV. In adjacent vertebral fractures after vertebral kyphoplasty, a retrospective study analyzed with multivariate analysis by Takehara et al. reported that older age contributed to recompression of adjacent vertebra after BKP, and elderly people often showed bone fragility due to osteoporosis[12]). In addition, in the augmented vertebra, we speculated that the vertebral bone fragility due to aging would make the residual cancellous bone between the cement and end plates easily collapse. Further, we should pay attention to cement augmentation in elderly people to avoid postoperative RAV.

In an ideal cement augmentation, the inserted PMMA should come in contact with both the upper and lower end plates, inducing better cement augmentation with higher CAR. In cases with sclerotic bone environments, using a curette to scrape and score sclerotic cancellous bone that may hinder vertebral body height restoration, or early intervention before the formation of sclerotic bone, would be favorable. Furthermore, in accordance with fracture type, an inflatable bone tamp should be inserted into an appropriate position where the inflated balloon comes in contact with both the upper and lower end plates. These attempts for “end plate to end plate cement augmentation” are important for the prevention of RAV.

The present study has the following limitations: (1) The number of patients who developed recompression was small, and further studies with larger sample sizes are needed to ensure the generalizability of our conclusions. (2) BMD, which may influence the prognosis of the augmented vertebrae, could not be evaluated among all patients. (3) Cement augmentation was evaluated in only the S-I direction using lateral radiograph as CAR. Further study using postoperative CT scan is needed to perform a three-dimensional analysis including the L-R and A-P directions for accurate evaluation of cement augmentation. Despite these limitations, this study presents the informative data about RAV that are useful in clinical practice, especially for spine surgeon who performs BKP.

Conclusion

In this study, RAV was observed in 38.5% of patients who underwent BKP in our single institution. Older age and low CAR were independently associated with RAV. For the prevention of RAV, especially in cases of elderly people, augmented PMMA should come in contact with both the upper and lower end plates.

Conflicts of Interest: The authors declare that there are no relevant conflicts of interest.

Sources of Funding: None

Author Contributions: Motoya Kobayashi, MD, wrote and prepared the manuscript, and all of the authors participated in the study design.

Ethical Approval: Unnecessary