Osteoporosis International

, Volume 22, Issue 12, pp 2915–2934

Balloon kyphoplasty and vertebroplasty in the management of vertebral compression fractures

Authors

  • S. Boonen
    • Division of Gerontology and Geriatrics and Center for Musculoskeletal Research, Department of Experimental MedicineLeuven University
  • D. A. Wahl
    • International Osteoporosis Foundation
  • L. Nauroy
    • International Osteoporosis Foundation
  • M. L. Brandi
    • Department of Internal MedicineUniversity of Florence
  • M. L. Bouxsein
    • Department of Orthopedic SurgeryHarvard Medical School
  • J. Goldhahn
    • AO Clinical Priority Program Fracture Fixation in Osteoporotic BoneETH: Institute for Biomechanics of ETH Zurich
  • E. M. Lewiecki
    • New Mexico Clinical Research & Osteoporosis Center
  • G. P. Lyritis
    • Laboratory for the Research of the Musculoskeletal SystemUniversity of Athens, KAT Hospital
  • D. Marsh
    • Institute of Orthopaedics and Musculoskeletal ScienceUniversity College London, Royal National Orthopaedic Hospital
  • K. Obrant
    • Department of OrthopaedicsSkåne University Hospital
  • S. Silverman
    • Department of Medicine and RheumatologyCedars-Sinai Medical Center
  • E. Siris
    • Toni Stabile Osteoporosis CenterColumbia University Medical Center
    • Department of OrthopedicsSkåne University Hospital Malmö, Lund University
  • for the CSA Fracture Working Group of the International Osteoporosis Foundation
Review

DOI: 10.1007/s00198-011-1639-5

Cite this article as:
Boonen, S., Wahl, D.A., Nauroy, L. et al. Osteoporos Int (2011) 22: 2915. doi:10.1007/s00198-011-1639-5

Abstract

Vertebral compression fractures (VCFs) are the most prevalent fractures in osteoporotic patients. The classical conservative management of these fractures is through rest, pain medication, bracing and muscle relaxants. The aim of this paper is to review prospective controlled studies comparing the efficacy and safety of minimally invasive techniques for vertebral augmentation, vertebroplasty (VP) and balloon kyphoplasty (BKP), versus non-surgical management (NSM). The Fracture Working Group of the International Osteoporosis Foundation conducted a literature search and developed a review paper on VP and BKP. The results presented for the direct management of osteoporotic VCFs focused on clinical outcomes of these three different procedures, including reduction in pain, improvement of function and mobility, vertebral height restoration and decrease in spinal curvature (kyphosis). Overall, VP and BKP are generally safe procedures that provide quicker pain relief, mobility recovery and in some cases vertebral height restoration than conventional conservative medical treatment, at least in the short term. However, the long-term benefits and safety in terms of risk of subsequent vertebral fractures have not been clearly demonstrated and further prospective randomized studies are needed with standards for reporting. Referral physicians should be aware of VP/BKP and their potential to reduce the health impairment of patients with VCFs. However, VP and BKP are not substitutes for appropriate evaluation and treatment of osteoporosis to reduce the risk of future fractures.

Keywords

IOFKyphoplastyOsteoporosisPain managementVertebral fracturesVertebroplasty

Introduction

Vertebral compression fractures (VCFs) are a common cause of back pain and disability. Osteoporosis is the underlying disease in most of these fractures and the remainder are a consequence of metastatic or primary cancer in the spine, high energy trauma or, more rarely, haemangioma [1, 2]. In 2005, it was estimated that there were over 2 million incident osteoporotic fractures in the US costing nearly $17 billion, with approximately 550,000 vertebral fractures, 300,000 hip fractures, 400,000 wrist fractures, 130,000 pelvic fractures and 680,000 other fractures [3]. In Europe, new cases of osteoporotic vertebral fractures are estimated at around 1.4 million each year [2]. These are the most common osteoporotic fractures of which up to 70% do not come to clinical attention [4]. These fractures of the weakened vertebral body typically result from low-energy loading during normal daily activities. Most clinically diagnosed fractures (over 80%) are detected during investigation of back pain whilst the remaining are old fractures detected incidentally during work-up [5]. Pain can have variable intensities and may radically affect a patient’s quality of life and physical function [6]. The intensity of pain depends on the site of the fracture, the number of fractures and their severity [7]. However, approximately half of patients with VCFs have no pain and therefore do not receive appropriate evaluation and treatment to reduce the risk of future fractures [5].

Vertebral fractures are predictive of future vertebral fractures and other osteoporotic fractures [810]. In women with pre-existing vertebral fractures, the risk of subsequent fracture is approximately four times that of women without a fracture history, and this risk increases with the number of prior vertebral fractures [11]. The risk of a subsequent fracture is highest (up to five times higher) in the first year after the original event [12, 13].

Effective pharmacological therapies with antiresorptive and anabolic agents for the prevention of subsequent fractures should be implemented in patients with osteoporotic VCFs [14]. There is general consensus that patients receiving such medication must have an adequate calcium and vitamin D supplementation as well, to maximize benefits of treatment, improve musculoskeletal functions and prevent future falls and fractures [1517]. Insufficiencies of both calcium and vitamin D appear to be widespread in adults, particularly older and institutionalized adults [18]. Current recommendations regarding vitamin D supplementation for older adults are 20 to 25 μg/day (800 to 1,000 IU/day) complemented with a dose of 1,000–1,200 mg/day of elemental calcium [15, 19]. Good adherence to calcium and vitamin D supplementation is essential as the benefits are lost after therapy withdrawal [20].

VCFs can lead to severe vertebral deformity or hyperkyphosis. The associated spinal deformity is linked with significant long-term consequences including decreased pulmonary function [21, 22], early satiety [23], gastric distress [23], impaired gait [23], subsequent vertebral fractures [24] and mortality risk [25]. In addition to back pain and back dysfunction [2628], VCFs cause substantial deficits in quality of life and contribute significantly to patient morbidity and mortality [2936]. Indeed, these fractures are associated with a 16% reduction in an expected 5-year survival [37].

Additionally, it has been estimated that a significant proportion of the cost of treating a fracture is not related to the fracture itself or its sequelae, but rather to co-morbidities resulting in significantly higher overall healthcare expenditures [3840]. In the US, the direct cost of medical management of osteoporotic VCFs was evaluated to just over $1 billion for approximately 550,000 incident vertebral fractures in 2005 [3]. By 2015, the 22% estimated increase in fracture incidence will raise the direct annual costs by 20% [3].

The standard treatment of painful VCFs has been conservative non-surgical management (NSM), consisting of bed rest, analgesics and bracing. Previous studies have found that medical management of symptomatic VCFs often fails to improve pain and mobility particularly in cases of chronic pain related to kyphotic deformity [27, 31, 41]. Furthermore, long-term use of narcotic analgesics and anti-inflammatory drugs are poorly tolerated by elderly patients. Narcotics may result in sedation and impaired balance that increase risk of falling and fractures. Prolonged bed rest may cause bone loss that exacerbates the underlying disease state [42] and can lead to rapid deconditioning, pulmonary compromise and increased mortality [43, 44].

Unfortunately, even the highest standard of NSM cannot prevent kyphotic deformity or permit the restoration of spinal alignment. Surgical treatment for unstable VCFs uses metal implants to stabilize the fracture [45]. However, poor bone quality and the presence of co-morbidities in older osteoporotic patients have limited the use of invasive surgical procedures due to the elevated risk for complications. Therefore, most surgeons only consider invasive surgical stabilization with metal implants when a neurologic deficit is present [4648].

Pain relief is the principal objective when treating VCFs in elderly populations. Vertebroplasty (VP) and balloon kyphoplasty (BKP) are minimally invasive approaches for vertebral augmentation available for the management of symptomatic VCFs [49]. Vertebroplasty was first performed to treat vertebral haemangioma over 20 years ago [50]. The insertion of a bone cement (polymethylmethacrylate-based, PMMA) or a bone substitute (calcium phosphate-based, CaP) into the vertebra is thought to stabilize the fracture, relieve pain, and in some cases of recent mobile fractures, restore lost vertebral height.

Serious complications though rare, can be potentially devastating. Peri-operative complications can be separated into two categories: procedural and cement leakage. Reported procedural complications include infection, fractures of the transverse process, pedicle, sternum and ribs and respiratory distress caused by the anesthetic [51]. Cement leakage of bone cement out of the vertebral body is a common complication that usually has no clinical consequences. However, in some cases, cement leakage into the epidural has resulted in paralysis due to spinal cord compression or radiculopathy following compromise of a neural foramen [51, 52].

Although introducing a stiff material into a weakened vertebra may stabilise the fracture, its effect on the overall spine must be monitored, as a retrospective study of a population-based cohort showed an increased risk of new adjacent and non-adjacent vertebral fractures after augmentation [53]. Procedural parameters that may affect that risk are the number of vertebrae treated, the quantity of cement injected and the degree of cement leakage [13, 51, 54].

The Fracture Working Group of the International Osteoporosis Foundation conducted a review of the medical literature on the efficacy and safety of vertebral augmentation by VP and BKP compared to NSM. The scientific evidence was limited to prospective controlled studies of the English literature in patients with osteoporotic VCFs with other types of studies referenced when appropriate to provide background information. Efficacy was measured by the potential to reduce back pain, to improve quality of life and physical function, increase vertebral body height and improve kyphotic angle. Safety was assessed from the reported surgical complications, namely symptomatic cement leakage to incidence of new vertebral fractures. Because the two procedures are separate techniques, these will be evaluated separately.

Vertebroplasty

This procedure aims to relieve pain, stabilize the fracture and prevent further loss of vertebral body height or progression of kyphotic deformity by the percutaneous insertion of a needle through the pedicles into the vertebral body and the injection of bone cement (PMMA or CaP) into the cancellous bone [55, 56]. The cement follows the path of least resistance and the procedure is monitored directly by fluoroscopy. Partial height restoration may also be possible in appropriately selected patients with recent vertebral fractures [57]. Eight prospective controlled studies comparing VP and NSM have evaluated the efficacy and safety profile of this surgical procedure (Table 1).
Table 1

Health outcomes after vertebroplasty and non-surgical management—change from baseline and between-group difference at various follow-up periods

References

Study design

Time to operation and conditions

Radiographic findings

Health outcomes (pain, physical functioning and quality of life)

Cement leakageb

New vertebral fractures

Other complications

VAS (Change from baseline)

ODI, Barthel index or RDQa (Change from baseline)

Qualeffo-41, SF-36, EQ-5D

[55]

Prospective non-randomized VP n = 30 NSM n = 30

<4 weeks after injury

Mean recovery of kyphosis

 

VP (%)

NSM (%)

p

N/A

N/A

20% spinal canal; 6.7% vertebral disc without neurological abnormality

N/A

1 temporary respiratory insufficiency

VP +8.4% NSM −21% p = 0.0001

6 months

−91.2

−65.6

0.05

12 months

−91.5

−73.6

0.05

[59]

Prospective non-randomized VP n = 55 NSM n = 24

Acute severe vertebral fractures

N/A

 

Barthel index

N/A

N/A

N/A

N/A

 

 

VP (%)

NSM (%)

p

 

VP (%)

NSM (%)

p

24 h

−53

−5

0.001

24 h

+29

0

0.001

6 weeks

−72

−61

0.15

6 weeks

+36

+31

0.29

6–12 months

−77

−71

0.57

6–12 months

+36

+39

0.22

[58]

Prospective non-randomized VP n = 88 NSM n = 38

1–2 weeks after the fractures

N/A

 

Barthel index

N/A

N/A

At 2 years

2 transverse process fractures, 1 hemorrhage into the psoas muscle. No difference in mortality risk between groups

 

All VP: 21/88 (23.9%) NSM: 9/38 (23.7%) HR = 1.13 p = 0.76

 

VP (%)

NSM (%)

p

 

VP (%)

NSM (%)

p

Adjacent VP: 9/88 (10.2%) NSM: 4/38 (10.5%) p = 0.52

24 h

−60

−5

0.001

24 h

+29

0

0.001

6 weeks

−75

−65

0.002

6 weeks

+36

+29

0.29

6–12 months

−85

−80

0.57

6–12 months

+36

+36

0.22

24 months

−90

−85

0.85

24 months

+36

+36

0.36

[60]

Prospective non-randomized VP n = 101 NSM n = 27

Patients with painful OP VCF (loss of height <70%, <12 months) refractory to medical therapy for at least 6 weeks

N/A

VP group significantly higher VAS score at baseline (p < 0.001)

ODI

SF-36

59.6% overall, 41% peridural plexus

At 1 year

1 transitory paraparesis from cement leakage, 2 rib fractures, 5 transitory radicular neuritis

VP group significantly higher disability score at baseline (p < 0.001)

Improvement from baseline in all scores up to 12 months. At 3 months, VP scored better than NSM in bodily pain (p = 0.003) and general health (p = 0.079). No statistical differences at 6 and 12 months

40.9% overall with different cement

All VP: 31/101 (30.7%) NSM: 3/27 (11.1%) p < 0.01

 

VP (%)

NSM (%)

p

 

VP (%)

NSM (%)

p

Adjacent VP: 12/101 (11.9%) NSM: 2/27 (7.4%) p N/A

Post-op

−55

−0

N/A

3 months

−47

−20

0.001

3 months

−60

−27

<0.001

6 months

−47

−47

N/A

6 months

−65

−40

0.033

12 months

−50

−57

N/A

12 months

−67

−53

n.s.

[61]

Multi-center prospective randomized VP n = 18 NSM n = 16

Painful OP VCF (≥15% height loss) refractory to medical therapy for at least 6 weeks and <6 months

N/A

 

RDQ

Qualeffo-41 (overall)

N/A

N/A

1 cortical chip fracture of the pedicle

 

 

 

VP (%)

NSM (%)

p

 

VP (%)

NSM (%)

p

 

VP (%)

NSM (%)

p

24 h

−34

−6.6

N/A

2 weeks

−17

0

N/A

2 weeks

−11.6

0

N/A

2 weeks

−31

−16

N/A

[63, 64]

Prospective randomized VP n = 25 NSM n = 24

<2 weeks (acute) and 2–8 weeks (subacute) OP VCFs

N/A

NSM group significantly higher VAS score at baseline (p = 0.02)

Barthel index

lSF-36 PCS

All asymptomatic

At 3 months

No complications occurred

 

VP (%)

NSM (%)

p

 

 

All VP: 3/25 (12%) NSM: 1/24 (4.1%)

<24 h

−74

N/A

N/A

 

VP (%)

NSM (%)

p

 

VP (%)

NSM (%)

p

Adjacent VP: 1/25 (4%) NSM: 0/24 (0%)

3 months

−76

−70.5

0.32

3 months

+10.7

+6.5

0.07

3 months

−7.3

−12.3

0.12

At 12 months

12 months

−73.3

−67

0.29

12 months

+11.9

+8.8

0.02

12 months

−12.5

−8.7

0.63

All VP: 4/25 (16%) NSM: 3/24 (12.5%)

Tandem (balance) test

SF-36 MCS

Adjacent VP: 1/25 (4%) NSM: 0/24 (0%)

 

 

3 months

p = 0.62

 

VP (%)

NSM (%)

p

At 3 months RR = 2.9 [CI 0.3–25.7]

12 months

p = 0.29

3 months

−1.6

−6.8

0.51

At 12 months RR = 1.3 [CI N/A]

12 months

−2

−1.2

0.93

EQ-5D NSM group significantly has higher score at baseline p = 0.05)

 

 

VP

NSM

p

3 months

×2.0

×6.5

0.04

12 months

×1.9

×6.9

0.19

[65]

Multi-center prospective randomized single-blind VP n = 38 NSM (sham) n = 40

VCFs <12 months duration and unhealed

N/A

 

RDQ

Qualeffo-41

In 37% of patients, minimal and asymptomatic

At 1 month

2 rib fractures in the VP group, 4 in the placebo group within 6 months, 1 osteomyelitis in VP group

 

 

All VP: 2/38 (5.2%) sham: 3/40 (7.5%)

 

VP (%)

NSM (%)

p

 

VP (%)

NSM (%)

p

 

VP (%)

NSM (%)

p

At 6 months

1 week

−20.3

−29.6

N/A

1 week

−10.4

−23.1

N/A

1 week

+0.9

−6

N/A

All VP: 3/38 (7.9%) sham 4/40 (10%) p N/A

1 month

−31.1

−23

N/A

1 month

−25.4

−17.9

N/A

1 month

−4.9

−4

N/A

3 months

−35.1

−26.8

N/A

3 months

−21.4

−30.6

N/A

3 months

−10.5

−10.2

N/A

6 months

−32.4

−28.4

N/A

6 months

−23.7

−21.4

N/A

6 months

−11.2

−10.2

N/A

 

[66]

Multi-center prospective randomized single-blind VP n = 68 NSM (sham) n = 63

<12 month-old VCFs

N/A

Back pain intensity score

RDQ

SF-36 PCS

N/A

N/A

Injury to the thecal sac in 1 patient with VP

 

 

 

 

VP (%)

NSM (%)

p

 

VP (%)

NSM (%)

p

 

VP (%)

NSM (%)

p

72 h

−39.1

−45.8

0.37

72 h

−21.7

−28.6

0.30

1 month

+17.4

+13.4

0.45

2 weeks

−37.7

−37.5

0.32

2 weeks

−25.3

−29.7

0.35

SF-36 MCS

1 month

−43.5

−36.1

0.29

1 month

−27.7

−25.7

0.49

 

  

VP (%)

NSM (%)

p

1 month

+4.7

+9.9

0.83

EQ-5D

 

 

VP (%)

NSM (%)

p

1 month

+22.8

+18.5

0.13

Statistical significance (p value) is as reported in studies and reflects the improvement in health outcome after VP compared to NSM

Op osteoporotic, VCF vertebral compression fracture, RR relative risk, HR hazard ratio of new vertebral fracture after VP compared to NSM, CI confidence interval, N/A information not available, ODI Oswestry disability index, RDQ Roland–Morris disability questionnaire, SF-36 short form 36 questionnaire, MCS mental component standardised, PCS physical component standardised, VAS visual analog scale, Qualeffo-41 quality of life questionnaire of the European Foundation for Osteoporosis, EQ-5D EuroQol Group 5-Dimension Self-Report Questionnaire score, VP vertebroplasty, NSM versus non-surgical management

aThese include also in part dimensions related to quality of life

bCement leakage per vertebral level treated, if not otherwise stated

Prospective non-randomized NSM-controlled studies

Nakano and colleagues followed 30 patients with acute VCFs treated with VP (CaP-based cement) and compared pain by Visual Analogue Scale (VAS) scores and vertebral body deformity to a NSM matched case-control group [55]. Mean improvement rates in VAS score at follow-up were significantly greater in the VP group compared to control with a 91.6% versus 73.6% improvement from baseline values at 12 months. The mean duration for analgesics need was significantly reduced (8.3 days versus 62.2 days, p = 0.0005). The mean reduction of kyphosis was 8.4% after VP compared to a reported 21% further collapse of the vertebral body after NSM. Six cement leakages were reported in the spinal canal (20%) and two in the intervertebral disc (6.7%) without resulting in neurological abnormalities.

The significant immediate effects of VP compared to NSM on pain reduction (VAS score) and improved physical mobility (Barthel Index) were confirmed in two other studies of patients with acute VCFs [58, 59]. However, although the benefits of treatment were sustained over 2 years, the between-group differences were not significant anymore for pain at 6–12 months and for physical mobility at 6 weeks [58]. The mean number of hospital bed days for patients treated by VP (10.4 days) was on average less than for patients treated conservatively (17.5 days). A similar rate of new adjacent and non-adjacent VCFs was reported within 2 years of 10% and 24%, respectively with no significant difference in the 2-year mortality risk in the two study arms.

In a prospective study, 101 patients refractory to conservative management of VCFs for over 6 weeks (non-acute fractures of less than 12 months) were treated with VP and compared to 27 patients who refused surgical intervention [60]. Pain evaluated by VAS score decreased significantly post-operatively at discharge in the VP group compared to control and was significantly different at 3 and 6 months but not at 1 year compared to NSM. Physical impairment, measured by Oswestry Disability Index (ODI), decreased over the year in both groups with better functional outcomes for VP-treated patients at 3 months compared to NSM. However, at 6 months and 1 year after the operation, the mean values of the conservative treatment group were significantly better than those of the VP group. A significant improvement in bodily pain and a trend towards better general health measured by short form 36 questionnaire (SF-36) was evident in the VP-treated patients compared to NSM at 3 months, but not statistically different at 6 and 12 months post-treatment. Cement leakage occurred in 41% and 60% of cases, depending on the bone cement used. Although the majority were asymptomatic, one leakage of the PMMA cement required surgical decompression for transitory paraparesis. Within a year, significantly more patients treated with VP had new vertebral fractures (p < 0.01) with a trend towards additional adjacent fractures. It must be noted that the control group was much smaller compared to the VP-treated group and that more than half of the patients in both groups had prevalent adjacent vertebral fractures at study entry.

Prospective randomized NSM-controlled studies

VERTOS, the first prospective randomized controlled study comparing VP to NSM showed that vertebroplasty reduced pain (VAS), improved physical function and quality of life [Roland–Morris Disability Questionnaire (RDQ) and Qualeffo-41] for 2 weeks compared to baseline and to NSM [61]. However, the small number of enrolled patients (n = 44) and the high cross-over rate from non-surgical care to VP after 2 weeks precluded longer term comparisons. A new protocol has been designed to assure longer term follow-up (VERTOS II) [62].

In a randomized clinical trial, Rousing and colleagues followed patients for 1 year with acute or sub-acute osteoporotic fractures treated with either VP or conservative medical management [63, 64]. Patient groups were small (total of 50 patients) with incomplete pain assessment and unmatched pain levels at baseline. At 12–24 h post-procedure, the VP-treated group had a significant immediate decrease in pain VAS score (p = 0.00). The corresponding change in VAS score in the NSM group was not reported. However, after 3 and 12 months, both groups showed a reduction in pain with no significant difference between the groups. Other health outcomes did not significantly change compared to baseline or between groups at 3 and 12 months. Only the Barthel Index score, a measure of physical function, was significantly higher at 1 year after VP compared to NSM. The surgical procedure did produce extravertebral cement leakage but all events were asymptomatic. Although very few new vertebral fractures occurred in a small patient cohort, a higher risk for future vertebral fracture was reported in the VP group (RR = 2.9 at 3 months, RR = 1.3 at 12 months for all new VCFs).

Prospective randomized sham-controlled studies

Two recent multi-centre, randomized, single-blind trials have evaluated VP compared to a simulated sham procedure (control) [65, 66]. The procedure in the control group mainly differed by the non-insertion of the bone cement but included the local anaesthesia at the vertebral site, the perception of PMMA use and the simulation of needle insertion into the fractured vertebra.

In the first of these trials, 78 patients with one or two painful osteoporotic fractures were randomized to undergo VP or a simulated sham procedure [65]. The primary outcome was overall pain score at 3 months, which decreased in both groups significantly compared to baseline. Pain reduction was sustained in both groups for 6 months. Similar improvements were seen in both groups with respect to physical function, quality of life and perceived improvement in pain, even after adjustment for baseline levels of previous vertebral fractures and duration of symptoms. Minimal asymptomatic leakage was found in 37% of treated vertebrae. Within 6 months, new vertebral fractures were not frequent and totalled three and four in the VP-treated and control group, respectively which limited the possible interpretation of an increased risk of future fractures.

In the second single-blind trial, 131 patients were randomly assigned to VP or a simulated sham procedure [66]. The primary endpoints of the study were scores in the modified RDQ and perceived back pain intensity after 1 month. Both procedures had an immediate and sustained improvement up to 1 month after the intervention, although not statistically different between the two arms (p = 0.19 and p = 0.49, respectively). The improvements of other measures of pain, physical function and quality of life [EuroQol Group 5-Dimension Self-Report Questionnaire score (EQ-5D), SF-36 mental component standardised (MCS) and physical component standardised (PCS)] did not also differ between groups at 1 month. Unfortunately, cross-over of patients in this study precluded longer term randomized comparisons between groups.

Discussion

Better evidence on the efficacy and safety of VP has recently become available through prospective controlled studies. Unfortunately, these still have a relatively low number of enrolled patients and have not always been powered to detect a meaningful change of health outcome. The baseline characteristics (age, time to procedure, prevalence of vertebral fractures) of patients treated with VP vary substantially across studies, and the comparator arm is seldom match-controlled. Furthermore, pain at inclusion was acute, sub-acute or chronic, and the degree of compression and number of treated vertebrae were dissimilar, which limits comparative analyses. Health outcomes were also measured differently at varying follow-up points and adverse events not systematically described. Further prospective randomized control studies with better standards for reporting data are needed in order to increase the understanding of the efficacy, safety and clinical importance of this procedure.

Published meta-analyses of past clinical case studies have shown a considerable post-operative pain reduction after VP in patients with acute or non-acute VCFs [51, 54, 6769]. However, limited studies exist that have compared pain relief with conservative management as control, and only two compared VP with a sham procedure (Table 1). Post-operatively and up to 6 weeks, pain relief after VP seems overall significantly higher than the NSM. The follow-up point at which the difference becomes insignificant varies between studies at 3 [64], 6 months [58, 60] or 1 year [55]. The decrease in pain may be due to a placebo effect, strong for invasive procedures, and the non-blinding of the treatment.

Indeed, as indicated by the two randomized sham-controlled studies, no differences were found over time in pain relief between VP and a sham procedure [65, 66]. These unexpected findings question the pain reduction potential of fixing a fractured vertebral body with bone cement and suggest that the potential benefits of the procedure may not outweigh the risks [70]. However, the controversial reports have caused much debate amongst clinicians and researchers, with great effort extended toward better understanding the design and conduct of the studies. Limitations of the studies included a selection bias of the patients, low number of study subjects (with one change in health outcome in one patient influencing the significance of results), lower amount of injected PMMA compared to previous studies, non-confirmation by imaging of vertebral fractures at baseline in patients with known fractures of less than 1 year, change in protocol (cross-overs) and a lack of standardisation of the medical treatment [7175]. The interpretation of the data is even more difficult due to the absence of an NSM group and the fact that these studies did not create a true sham procedure, but used local anaesthetic infiltration of the posterior longitudinal ligament, which is an established treatment for back pain. In comparing the values of pain reduction between the reported studies in Table 1, VP had a less pronounced effect on pain in these two studies compared to other prospective studies, which may in part explain this non-significant difference.

The efficacy of VP over NSM was assessed using complementary measures of health outcomes that quantified pain, quality of life and physical function of patients at baseline and post-operatively. Once again, there are contrasting results. Most health outcomes generally improve in both groups with time, but the increased effect of VP compared to NSM varies across studies. For example, the Barthel Index score significantly increased more after VP compared to NSM after 12 months in one study [64], but only after 1 day in another [58, 59]. Whilst the ODI score was significantly better after VP at 3 months, it was significantly worse after 6 and 12 months compared to NSM [60]. A more recent study showed no differences in the EQ-5D and SF36 MCS and PCS scores between the VP and NSM [64], which was also confirmed by a study using a sham control [66].

Only one controlled study evaluated the impact of VP on the restoration of height in acutely compressed vertebral bodies, noting a slight augmentation in height compared to a significant decrease for the conservative treatment group [55]. These were fresh acute vertebral fractures. Although some clearer evidence exist as to the short-term potential benefits of VP over NSM, it is important to consider the risks involved in injecting a foreign material into the vertebral body and performing surgery.

The most commonly reported complication after VP is cement leakage, which in the studies described in Tables 1 and 4, varied widely from 11% to 60%. The large majority were asymptomatic and only two neurological complications linked to the cement were noted [60, 76]. This is within the range of previous published meta-analyses on VP, averaging 19.7% [54], 40% [77], 41% [51] and 43% [78] of cement leakage per vertebral level treated. Symptomatic leaks were also found to be low, for only 1.1% [78], 1.6% [54], 2.6% [51] and 3% [77] of cases per vertebral level treated. The large range of cement leakage rates may be explained by differences between the studies in the surgical procedure (cement viscosity, volume and injection pressure), the methods of reporting and finally the severity of the vertebral fracture treated.

Few other non-cement-based medical and procedural complications were noted, with calculated mean rates of 2.4% [79], 3.8% [78] and 4.8% [77] in previous meta-analyses. Of the complications noted, most are transitory and possible to treat rapidly, which makes VP a relatively safe procedure for patients with VCFs.

The long-term safety of VP can be also evaluated by the number of new vertebral fractures. Indeed, the injection of a much stiffer material in a fragile osteoporotic spine may induce large stresses, which could in time result in adjacent or non-adjacent VCFs [80]. Out of the four prospective studies in Table 1 that have documented long-term incidence of new vertebral fractures, three studies reported a non-significant difference in the rate of new adjacent fractures between patients that underwent VP and those treated conservatively or with a sham procedure [58, 63, 65]. It remains uncertain as to whether stabilising spinal fractures with a cement elevates the risk of adjacent or non-adjacent vertebral fractures, or whether it is just the natural course of the patients’ future years with underlying osteoporosis.

Cost-effectiveness of vertebroplasty

Studies investigating the cost-effectiveness of VP are scarce. In a retrospective study of 140 patients comparing VP and NSM, the cost-effectiveness of VP was measured as the average cost per patient per reduction of one point in the VAS, ambulation or activity of daily living scale [81]. The results showed that after 1 week, VP was significantly more cost-effective for all measured outcomes compared to NSM (p < 0.05).

Summary

All interventions need a careful evaluation of the benefits and risks involved. Evidence has demonstrated the potential short-term benefits of VP over NSM, in terms of pain and physical function, but two sham-controlled studies have questioned these conclusions. Further studies are needed, with standardised and systematic reporting of health outcomes and complications, to elucidate the best solutions for the management of VCFs.

Balloon kyphoplasty

BKP is a minimally invasive procedure aiming to relieve pain, stabilize the fracture, restore lost vertebral body height, prevent progression and correct kyphotic deformity [82]. Cannulae placed percutaneously into the vertebral body permit the insertion of two inflatable bone tamps, called “balloons” [83]. During inflation, the balloons compact the surrounding trabecular bone and create an enclosed cavity [84]. After removal of the balloons, the cavity is filled with a bone cement (PMMA) or a bone substitute (CaP) [85, 86] under low manual pressure [56]. As with VP, the procedure is monitored directly under fluoroscopy. In the English literature reviewed, only two prospective controlled studies evaluated and compared the efficacy and safety profile of BKP versus NSM (Table 2).
Table 2

Health outcomes after balloon kyphoplasty and non-surgical management—change from baseline (%) and between-group difference at various follow-up periods

References

Study design

Time to operation and conditions

Radiographic findings (Change from baseline)

Health outcomes (pain, physical functioning and quality of life)

Cement leakage**

New vertebral fractures

Other comments

VAS (Change from baseline)

EVOS-Q, SF-36 PCS, RDQ, EQ-5D (Change from baseline)

[88]a

Prospective, non-randomized controlled BKP n = 40, NSM n = 20

Chronic painful VCFs >12 months

 

 

 

7/72 (9.7%)a, c, all symptomatic

At 6 monthsa

Fewer back pain-related doctor visits in the BKP group (p = 0.015)

Midline vertebral body height*

 

EVOS-Q

All BKP: 5/40 (12.5%) NSM: 8/20 (40%) p = 0.227

 

BKP (%)

NSM (%)

p

 

BKP (%)

NSM (%)

p

 

BKP (%)

NSM (%)

p

[86]b

Post-opb, c

+15.4

N/A

N/A

Post-opc

−24.3

N/A

N/A

Post-opb, c

+23.7

N/A

N/A

At 12 monthsb

3 monthsa

+14.1

−2.6

0.0001

12 monthsc

−24.7

−1.1

0.008

3 monthsa

+20.3

+13.3

0.205

All BKP: 7/40 (17.5%) NSM: 10/20 (50%) p = 0.0084

Less opioids intake in BKP than control group

6 monthsb

+12.1

−8.6

<0.0001

36 monthsc

−26.8

−3.6

0.023

6 monthsb

+24.2

+10

0.027

At 36 monthsc

 

No burst fracture or vertebra plana

12 monthsb, c

+12.6

−8.4

<0.0001

12 monthsb, c

+24.2

+11.3

0.105

All BKP: 15/40 (37.5%) NSM: 14/20 (70%) p = 0.0148

36 monthsc

+9.3

−15.9

<0.0001

36 monthsc

+25.1

+9.5

0.082

[87]c

Kyphosis angle

 

BKP (°)

NSM (°)

p

Post-opc

−1.1

N/A

N/A

3 monthsa

−0.1

+1.3

0.038

6 monthsa

−0.4

+4.9

<0.001

12 monthsc

−0.8

+2.4

0.0001

36 monthsc

−0.9

+3.1

<0.0001

[89]

Multi-center prospective randomized

Sub-acute VCFs <3 months

N/A

 

 

51/188 (27%) all asymptomatic

At 12 months new or worsening radiographic vertebral fractures

Less opioids intake in BKP than control group

BKP n = 149, NSM n = 151

At least 1 oedema at MRI or at least 15% loss of height

 

BKP (%)

NSM (%)

p

SF-36 PCS

1 week

−46.5

−14.5

<0.0001

 

BKP (%)

NSM (%)

p

1 month

−51

−21.5

N/A

1 month

+28.3

+6.4

<0.0001

3 months

−57

−35

N/A

3 months

+32.2

+20.7

0.0008

Patients in the BKP group had about 60 fewer days of restricted activity during the year than the controls had

6 months

−60.5

−37.5

N/A

6 months

+39.4

+27

0.0064

12 months

−58.5

−45.5

0.0034

12 months

+38.1

+30.6

0.208

p < 0.0001

RDQ

 

BKP (%)

NSM (%)

p

1 month

−36

−11

<0.0001

3 months

−46

−23.5

N/A

BKP: 38/115 (33%) NSM: 24/95 (25%) p = 0.220

BKP-related adverse event, 1 soft tissue haematoma and 1 postoperative urinary tract infection

6 months

−50

−33

N/A

12 months

−49

−33

0.0012

EQ-5D

 

BKP

NSM

p

1 month

×3.3

×2.1

0.0003

3 months

×3.6

×2.85

N/A

6 months

×3.85

×2.9

N/A

12 months

×3.8

×2.95

0.0252

Statistical significance (p value) is as reported in studies and reflects the improvement in health outcome after BKP compared to NSM

*Vertebral height change compared to the intact posterior height of the fractured vertebrae or of the closest non-deformed vertebral body

**Cement leakage per vertebral level treated

pThe average treatment effect significance during follow-up

Op osteoporotic, N/A information not available, VAS visual analog scale, EVOS-Q European vertebral osteoporosis study questionnaire, EQ-5D EuroQol group 5-dimension self-report questionnaire score, RDQ Roland–Morris disability questionnaire, SF-36 PCS short form 36 questionnaire physical component standardised, BKP balloon kyphoplasty, NSM non-surgical management

Prospective non-randomized NSM-controlled studies

The first non-randomized controlled study was from Kasperk and colleagues, who prospectively followed patients for 6, 12 and 36 months that underwent either BKP (n = 40) or NSM (n = 20) [8688]. At baseline, patients were well-matched across groups in age, gender and bone mineral density and similar in baseline scores for pain and physical function measures. Midline vertebral body height increased by 15.4% post-operatively, and although this partially restored height declined with time, it was still significantly higher after 3 years compared to baseline (p < 0.0001) [87]. This translated into an improvement in kyphosis angle, which reduced by 1° over 3 years. Compared to NSM where midline vertebral body height and kyphosis angle worsened, BKP had a significant and positive effect at all follow-up points (Table 2). The sensation of pain decreased and stabilized over time after BKP, which contrasted with little change in pain in the group treated conservatively. This difference was significant up to 3 years. The mobility assessment [European vertebral osteoporosis study questionnaire (EVOS-Q)] of patients after BKP improved significantly compared to baseline and differed significantly compared to NSM only at 6 months. The study showed a trend at 3, 12 and 36 months towards a better physical function after BKP but did not reach significance compared to NSM. Traces of cement leakage after BKP occurred in 9.7% of treated vertebrae (7/72) [88] or 16.4% of cases (12/73) [86] depending on the study report, but were all asymptomatic. The occurrence of new vertebral fractures was higher after NSM compared to BKP at 6 months and reached significance at 1 and 3 years follow-up. Finally, patients treated with BKP required less pain medications and significantly less pain-related doctor visits compared to patients treated conservatively.

Prospective randomized NSM-controlled studies

The Fracture Reduction Evaluation Study enrolled 300 patients with 5 to 6-weeks old VCFs, randomized to either BKP (n = 149) or NSM (n = 151) [89]. This is the largest prospective controlled study performed so far comparing minimally invasive surgery in VCFs and NSM. Groups were similar at baseline with the exception of multiple prevalent vertebral fractures, which were more common in the BKP-treated group. The proportions of patients on bisphosphonate treatment were similar in both groups upon study entry (33% in BKP group compared to 32% in control). The primary outcome was the difference in change from baseline to 1 month in SF-36 PCS score between the BKP and NSM groups, which was found to be significantly better in BKP-treated patients (+28.3% versus +6.4%; p < 0.0001). This short-term benefit continued until 6 months only and lost its significance at 12 months. Secondary outcomes were back pain and other measures of quality of life. Although back pain scores (VAS) decreased in both groups compared to baseline, BKP-treated patients reported significantly lower back pain after 1 week and until 1 year follow-up (overall treatment significance p < 0.0001). As a result, significantly fewer BKP patients required opioid medications during the first 6 months. At 1 month, patients’ physical function and quality of life were significantly improved (RDQ and EQ-5D scores) after BKP compared to NSM (p < 0.0001 and p = 0.0003, respectively) and this marked difference was maintained up to 3 years (p = 0.0012 and p = 0.0252, respectively). In addition, patients in the NSM group had an estimated 60 more days of limited activity compared to BKP-treated patients during the year. Cement leakage occurred in 27% of treated vertebrae and all cases were asymptomatic. At 12 months, 33% of BKP-treated patients and 25% in the NSM group had new vertebral fractures or worsening of non-treated other radiographic vertebral fractures (p = 0.22). Two-year outcomes confirmed these findings [90].

Discussion

Although numerous retrospective and clinical case studies of BKP report good efficacy and safety of BKP [51, 77, 78, 91, 92]. To date, only a limited number of prospective controlled studies exist in osteoporotic VCF patients. The study from Wardlaw and colleagues constitutes level I evidence of treatment efficacy, with the largest number of patients enrolled [89]. Both prospective controlled studies provide informative results of BKP over NSM, but are skewed due to the non-blindness of the surgery and the strong placebo effect with invasive surgery.

Different methods for reporting data were chosen, which limits the comparison between studies. Patient inclusion criteria in both studies also differed in age, prevalent fractures and type of fractures. It is therefore difficult to characterize a group of patients with osteoporotic VCFs that would benefit most from BKP. Retrospectively however, it was found that treatment with BKP of acute fractures or chronic VCFs resulted in similar pain relief and improved physical functioning [93]. Vertebral height restoration and kyphosis reduction occurred however to a larger extent in acute fractures [93].

In Table 2, only one study reported radiographic findings and showed a direct increase in vertebral height and a reduction in kyphosis after BKP over NSM, in VCFs older than 12 months [88]. This augmentation was independent of the type of bone cement used [94]. An earlier biomechanical study suggested that the initial gains in vertebral body height after BKP are lost after repetitive loading [95], but in vivo, this change appears to have lasted up to 3 years [87].

Prospective controlled studies of BKP show that back pain is significantly reduced from baseline in patients with sub-acute VCFs up to 1 year and chronic VCFs up to 3 years. In both studies, however, BKP had a significantly better pain outcome compared to NSM at all time points. Physical functioning and quality of life (EVOS-Q) improved to a larger extent after BKP compared to NSM in chronic VCFs but was significantly different only up to 6 months [86, 87]. For sub-acute fractures, only the SF-36 PCS lost significance at 6 months, but for the other health outcomes (RDQ and EQ-5D), the positive treatment effect lasted up to 1 year.

It is possible that the improvement of vertebral morphometry after BKP contributes partially to the reduction in back pain and improvement of physical functioning, but such a correlation has not been demonstrated in clinical studies [76, 87]. This implies that the observed pain relief and improved health outcomes are a consequence of the vertebral fracture fixation or due to a placebo effect. A sham-controlled study would be helpful for further evaluation.

Cement leakage rates in the tabulated studies varied between 9.7% and 27% of vertebral fractures treated with BKP but were without clinical consequences [88, 89]. Previous published meta-analyses of clinical case studies reported cement leakage average rates of 8% to 10.5% [51, 77, 78, 91, 92], with complication rates between 0.4% and 2.2% of cases [51, 78, 91]. Neurological complications due to cement leakage into the epidural space are reported in 0.03% of BKP cases and pulmonary embolism secondary to cement flow into venous channels occur in 0.01% of cases [51]. Other non-cement-related medical complications were reported in 1.6% of cases after BKP [78].

The effect of BKP on the risk of having subsequent vertebral fractures was reported in both studies with contrasting results. BKP either reduced significantly the rate of new vertebral fractures within 1 and 3 years [86, 87] or slightly increased the rate within 1 year [89]. Patient characteristics in both studies differed at baseline for age and prevalence of vertebral fractures, suggesting that the risk for future osteoporotic VCFs varied as well. Thus, more comparative studies are needed to determine whether BKP increases the risk of future vertebral fractures over and above the risk of patients treated conservatively.

Cost-effectiveness

Taylor and colleagues calculated the incremental cost-effectiveness ratio using quality-adjusted life years (QALY) for BKP to be €7,016/QALY for Italy and €11,428/QALY for Spain [96]. At a willingness-to-pay per QALY of €45,000, the probability that BKP is cost-effective compared to conservative NSM is 83% for Italy and 78% for Spain.

In another study, Ström et al. used a Markov model to evaluate the cost-effectiveness of BKP in osteoporotic patients hospitalised with painful VCFs to NSM in a UK setting [97]. Table 3 highlights the cost per QALY gained of £8,840 when comparing BKP and NSM. The authors concluded that with a willingness to pay in the UK ranging from $20,000 to 30,000 per QALY, BKP is a cost-effective treatment in this population setting. The differences in subsequent new fracture rate and mortality risk between the two procedures were however not taken into account.
Table 3

Cost-effectiveness of balloon kyphoplasty versus non-surgical procedure in 70-year-old patients with painful osteoporotic VCFs in the UK. Adapted from Strom et al. 2009 [97]

 

BKP

NSM

Difference

Procedure cost (£)

4,090

0

4,090

Inpatient and outpatient care cost (£)

6,330

8,926

−2,596

Total cost (£)

10,420

8,926

1,494

Life years

9.390

9.390

0.00

QALYs

3.842

3.673

0.169

Cost per QALY gained (£)

  

8,840

BKP balloon kyphoplasty, NSM non-surgical management, QALY quality-adjusted life years

Summary

Current results from prospective controlled studies including a large randomized clinical trial demonstrated the acute and sustained benefits of BKP on pain relief, improved quality of life and physical functioning compared to conservative NSM. However, single-blinded or sham-controlled randomised studies are necessary to confirm these findings. Evidence is available demonstrating vertebral body height restoration and kyphosis angle improvement with BKP, though a clear clinical benefit to the patient has not been demonstrated. Overall and with the current evidence available, BKP seems to be an efficient, relatively safe and cost-effective treatment option for painful osteoporotic VCFs. Further studies are however recommended to increase our knowledge on this minimally invasive procedure.

Comparisons of vertebroplasty and balloon kyphoplasty

Table 4 shows the direct comparative studies of VP and BKP in the management of osteoporotic VCFs. These prospective studies, of which one was randomized, assessed simultaneously the efficacy and safety of both procedures.
Table 4

Prospective studies and clinical trials comparing vertebroplasty and balloon kyphoplasty

References

Study design

Time to operation and conditions

Radiographic findings (Change from baseline)

Health outcomes (pain, physical functioning and quality of life)

Cement leakageb

New vertebral fractures

Other comments

VAS (Change from baseline)

ODI (Change from baseline)

[98]

Prospective non-randomized study VP n = 10 BKP n = 11

Painful VCFs <6 months old

N/A

 

 

VP: Overall, 5/18 (27.8%) disc or paravertebral space, all asymptomatic

N/A

In both groups, no neurological compromise or cardiovascular events associated with PMMA in both groups.

 

 

 

BKP (%)

VP (%)

p

 

BKP (%)

VP (%)

p

BKP: 0%

48 h

−89.1

−92.2

N/A

6 months

−68.6

−66.4

N/A

 

1 months

−87.9

−80.4

N/A

 

3 months

−92.9

−88.0

N/A

 

6 months

−91.5

−93.4

N/A

 

[99]

Prospective non-randomized study VP n = 23 BKP n = 28

Non-recent OP VCFs; median duration of pain 8 weeks before surgery

 

 

 

VP: Overall 8/29 (27.6%), 13.8% disc space, 6.9% epidural space, 6.9% segmental vessels

At 4 months

Vertebral body heighta

 

 

 

Adjacent BKP: 6/28 (21.4%) VP: 1/23 (4.3%)

 
 

BKP (%)

VP (%)

p

 

BKP (%)

VP (%)

p

 

BKP (%)

VP (%)

p

BKP: Overall 8/35 (22.8%), 22.8% disc space, 0% epidural space or segmental vessels

 

Post-op

+5.8

0.0

N/A

Post-op

−52.7

−61.5

N/A

4 months

−37.7

−6.1

N/A

 

p N/A

Kyphosis angle

4 months

−56.8

−26.9

N/A

12 months

−31.1

−4.1

N/A

No neurological deficit recorded in both procedures

 
 

BKP (%)

VP (%)

p

12 months

−63.5

−26.9

N/A

24 months

−8.2

+6.1

N/A

 

Post-op

−6

0.0

N/A

24 months

−73

−41

N/A

 

[100]

Prospective non-randomized study VP n = 30 BKP n = 30

Fresh single segment VCFs <1 month old without neurological deficits

 

 

N/A

VP: Overall, 10/30 (33.3%)

At 13 months

One rupture balloon during BKP procedure, without any consequence.

Kyphosis angle

 

BKP: Overall, 2/30 (6.7%) p = 0.021 all asymptomatic

Adjacent VP: 1/30 (3.3%) BKP: 0/30 (0%)

 

BKP (%)

VP (%)

p

 

BKP (%)

VP (%)

p

N/A

Post-op

−47.2

−18.4

<0.001

Post-op

−61

−63.8

>0.05

13 months

−43.2

−8.8

<0.001

13 months

−68.3

−66.3

>0.05

[101]

Prospective randomized study VP n = 50 BKP n = 50

Acute VCFs averaging 2 weeks old

 

 

N/A

N/A

At 6 months

BKP required significantly longer operation times and larger amounts of PMMA.

Vertebral body height

Adjacent BKP: 2/50 (4%) VP: 0/50 (0%)

 

BKP (%)

VP (%)

p

 

BKP (%)

VP (%)

p

p N/A

Post-op

+80.5

+30.7

N/A

Post-op

−67.5

−70.9

N/A

Kyphosis angle

6 month

−67.5

−67.1

N/A

 

BKP (%)

VP (%)

p

Overall significance between group of p > 0.05

Post-op

−47.0

−21.3

N/A

p<0.001 overall between groups

[76]

Prospective non-randomized study VP n = 154 BKP n = 36

Painful VCFs >1 month but <3 months

Vertebral body collapse

 

 

Overall, 14.6% observed cement leakage

At 9 month

Significantly longer operation times and larger amounts of PMMA with BKP.

BKP for VCFs equal or higher than 30%

Anterior

BKP (%)

VP (%)

 

 

VP: 18/156 (11.5%) BKP: 6/43 (14%) p < 0.05 asymptomatic except for 1 neurological complication in VP

All VP: 4/154 (2.6%) BKP: 0/43 (0%)

The days spent in hospital after procedure were similar.

VP for VCFs less than 30% and >3 months

Post-op

−17.9

0

 

BKP (%)

VP (%)

p

 

BKP (%)

VP (%)

p

 

Adjacent VP: 2/154 (1.3%) BKP: 0%

No significant correlation was found between VAS or ODI scores and radiographic measurements.

3 months

−15.4

−4.8

1 month

−54.8

−57.1

N/A

1 month

−55.0

−56.0

N/A

p N/A

6 months

−15.4

−4.8

3 months

−62.5

−61.9

N/A

3 months

−73.3

−75.7

N/A

24 months

−12.8

0

6 months

−67.5

−64.3

N/A

6 months

−85.3

−83.7

N/A

Midline

BKP (%)

VP (%)

24 months

−76.2

−76.2

N/A

24 months

−90.2

−87.2

N/A

Post-op

−18.9

+5.3

3 months

−18.9

+5.3

6 months

−16.2

0

24 months

−16.2

0

Posterior

BKP (%)

VP (%)

Post-op

−16.7

+11.1

3 months

−16.7

+11.1

6 months

−8.3

0

24 months

−8.3

0

p for all values N/A

Statistical significance (p value) is as reported in studies and reflects the improvement in health outcome after BKP compared to NSM

aHeight of the vertebral body is the distance between the connections of the frontal and dorsal edge of the vertebral body

bCement leakage per vertebral level treated, if not otherwise stated

OP osteoporotic, N/A information not available, VAS visual analog scale, ODI Oswestry disability index, BKP balloon kyphoplasty, NSM non-surgical management, PMMA polymethylmethacrylate, VP vertebroplasty, VCFs vertebral compression fractures

Prospective non-randomized comparative study

De Negri and colleagues published a prospective non-randomized study on a small sample size (21 patients) [98]. Up to 6-month follow-up, the pain score (VAS) and physical functioning (ODI) had improved significantly from baseline in both BKP- and VP-treated group, and the authors reported no between-group differences. Cement leakages occurred after VP only, in 5 of the 18 vertebral bodies, without evidence of clinical symptoms. New vertebral fractures, adjacent or non-adjacent were not reported.

Grohs and colleagues published a prospective non-randomized comparison of VP and BKP in 51 elderly patients [99]. In this study, only BKP managed to restore partially vertebral body height and reduce the measured kyphosis angle compared to baseline measurements. A large pain reduction measured by VAS is observed in both groups post-operatively compared to baseline and was sustained up to 2 years in patients treated with BKP. In VP-treated patients, pain was reduced up to 2 years also but to a lesser extent. Post-operative back function measured by ODI was statistically better (p < 0.05) only in the BKP-treated patients compared to baseline levels over 1 year follow-up but was not significantly different after 2 years. Between-group differences were not assessed. Cement was present after BKP in the disc space of eight treated vertebral bodies and after VP, it was present in the epidural space and segmental vessels of four treated vertebral bodies. No clinical consequences were noticed in either group. A higher number of new adjacent level fractures (6/28) were reported within 4 months after BKP compared to VP (1/23).

A similarly sized study prospectively followed 60 patients assigned to either BKP or VP following recent VCFs for 13 months [100]. In both groups, the kyphosis angle was improved compared to pre-surgery but significantly more after BKP (p < 0.001). Pain score (VAS) largely decreased post-operatively and was sustained over time up to 13 months in both groups, without being significantly different between groups (p > 0.05). Cement leakages were asymptomatic in both groups but occurred much more after VP (10 of 30 vertebrae treated) than after BKP (2 out of 30 vertebrae treated). Only one adjacent fracture was reported after VP within 13 months of follow-up.

The largest comparative study to date enrolled 314 patients with VCFs, selected to either have NSM, BKP or VP depending on their time since fracture [76]. When VCFs were acute, a conservative treatment was initiated for 1 month. After failure of the conservative treatment to reduce pain or when fractures were older, patients were stratified to receive VP or BKP depending on the age of the fracture and the amount of vertebral collapse. Patients were treated with BKP when the vertebral compression reached 30% or more and with a time from fracture between 1 and 3 months, whilst VP was indicated for less severe VCFs or fractures older than 3 months. All health outcomes (VAS and ODI) were significantly improved post-operatively and over a period of 2 years after VP and BKP without large differences between the two groups. Vertebral body collapse was partially restored post-operatively after BKP at the anterior, midline and posterior section of the vertebra, which sustained although to a lesser extent over 2 years. VP however did not prevent further collapse of the midline and posterior height of the vertebra. No significant correlations were found between VAS, ODI and radiographic measurements, and patients treated with both procedures stayed in hospital for similar amount of days after surgery. With regards to complications, the overall cement leakage rate was 14.6% for treated vertebra (11.5% after VP and 20.9% after BKP). All were asymptomatic except for one neurological complication after VP. Out of 154 patients treated with VP, four patients had new vertebral fractures within 9 months follow-up, of which two were adjacent. Patients treated with BKP did not report any new vertebral fractures. Finally, the authors noted significantly longer operation times and larger amounts of PMMA used during BKP.

Prospective randomized comparative study

In a larger randomized study, Liu and colleagues followed 100 patients with acute T12/L1 VCFs for 6 months [101]. The mean duration between injury and surgery was 17.0 ± 7.7 days for BKP and 15.8 ± 6.7 days for VP. A post-operative increase in vertebral body height and a reduction of kyphotic angle were observed in both techniques; however, results for these parameters were significantly greater after BKP compared to VP (p < 0.001). Both treatments provided a fast, large and sustained pain relief but without any significant difference between treatment groups from the pre-operative, through to the post-operative period and the final follow-up at 6 months. Two patients in the BKP-treated group had an adjacent segment fracture and none after VP, but prevalent fractures at baseline were not indicated. The authors reported finally that BKP required significantly longer operation times and larger amounts of PMMA compared to VP.

Discussion

Four prospective non-randomized studies and one prospective randomized trial compared directly the VP and BKP for treatment of osteoporotic VCFs. All studies reported immediate and sustained pain reductions over time for both surgical treatments in both acute and non-acute fractures. Statistical differences between groups for pain reduction were not found [100, 101]. Other health outcomes were only measured in three studies with the ODI, a condition-specific outcome measure used in the management of spinal disorders. In non-acute VCFs, ODI improved after BKP in all studies, whilst it only improved in two out of three studies after VP (Table 4). The ODI response or other health outcome changes to the treatment of acute VCFs was however not evaluated in these comparative studies. Furthermore, it is not possible from these studies to determine whether patients would benefit from either of the surgical procedures compared to NSM.

When evaluating direct comparative studies only (Table 4), partial vertebral height restoration and improved kyphotic angles can be observed in both procedures, so long as the vertebral fractures are acute. It is thought that after a period of 6 to 8 weeks, most skeletal fractures heal and stiffen, which would limit the possibility for vertebral height changes and patient manipulation [51, 52]. Nonetheless, these changes of vertebral shape after the procedure were significantly more marked after BKP compared to VP [76, 100, 101], which can be explained most likely by the action of the inflatable tamps pushing against the vertebral walls during the BKP procedure. However, no correlation was found between the vertebral shape correction and a clinical benefit to the patient in terms of pain or disability [76].

Cement leakage and new adjacent or non-adjacent vertebral fractures are the main complications reported in the reviewed comparative studies (Table 4). Most cement leakages were found to be asymptomatic and generally more pronounced after VP compared to BKP. However, it appears more important to evaluate the change in risk of vertebral fractures, as these affect patient’s quality of life and mortality risk [102, 103]. It was not possible to determine from the studies in Table 4 whether the rate of new vertebral fractures is increased after either of the surgical treatments. Only one study suggested a larger rate of new vertebral fractures after BKP compared VP [99]. In general, the literature suggests that new fractures of adjacent vertebrae occur after VP and BKP at rates approximately equivalent to that of those with prevalent osteoporotic vertebral fractures [51, 89].

To determine whether the rates of subsequent VCF are higher among subjects undergoing vertebral augmentation compared to those treated non-surgically would require a concurrently controlled study in which risk factors for fracture are evenly distributed across the treatment groups. It is important to identify the baseline levels of prevalent fractures in order to correctly analyze the effect of surgical procedures on incidental new vertebral fractures. It seems also that the quality characteristics of the bone affect the clinical outcome and the occurrence of the complications of the procedure. It should be noted that there are certain categories of patients, such as those with corticosteroid treatment [104], renal disease or solid organ transplant recipients [105] that appear to have increased incidence of the subsequent vertebral fractures.

In addition, operators’ skills, cement types and surgical apparatus/technique most likely affect the clinical and safety outcomes of VP and BKP. New types of cements are being investigated, altering their stiffness and viscosity to reduce cement leakage risks, load transfer properties and ease of injection [106]. Alternative cements can also mimic closer the mechanical properties of bone when set [85]. Other surgical technologies are emerging for vertebral augmentation addressing these specific complications [107], including vertebral body stenting [108] and radiofrequency kyphoplasty [109].

Recommendations from the IOF CSA fracture working group

Need for standardized reporting in clinical studies

The heterogeneity of the reviewed studies precludes general conclusions. Strict research reporting criteria are required to ensure comparability of studies. A recently published consensus document on research reporting standards from interventional radiology colleagues provides guidelines on mandatory data about study populations, treatment description, outcome assessments and analysis [110]. They also provide definitions for minor and major complications that are crucial for the reporting of results. Studies including these required parameters will form a much better base for decision making and should be encouraged.

Improved clinical management of vertebral fracture patients

In most health care systems, patients with VCFs are poorly identified and treated. Vertebral fractures are typically diagnosed and treated by practitioners representing a variety of medical specialties. In order to improve the management of these patients and ensure the rapid implementation of the most appropriate treatment, it is recommended to organize multidisciplinary clinics. These clinics should combine surgical and non-surgical management expertise in order to have the best global appreciation of the patient and her/his fracture status, and facilitate the decision for NSM or vertebral augmentation. Multidisciplinary consultation would also ensure that osteoporosis treatment is adequate and calcium and vitamin D supplementations initiated when needed, to treat the underlying disease and prevent subsequent fractures.

VCF is the most common type of fragility fracture, but VCFs vary widely in anatomic location, degree of deformity, degree of back pain, loss of quality of life, and, ultimately, individual outcomes and clinical consequences. There is evidence to suggest that one of the most significant factors predicting the severity and longevity of symptoms is the severity of fracture deformation [7]; the presence or absence of a previous fracture is also correlated with important clinical outcomes [111]. However, there is currently no comprehensive predictive tool based on the fracture’s character (spine localization, type of fracture, grade of fracture, previous fractures, etc.) to identify, at baseline, the best therapeutic approach for a given VCF to minimize its impact on patients’ future disability, quality of life and activities of daily living.

Finally, the mortality risk for the different treatment options of VCFs (conservative treatment, VP and BKP) after 1 year has been compared in a retrospective analysis of the US Medicare database (n = 40,354 procedures) [112]. Surgically treated patients had a 43% reduced risk of mortality compared to non-surgically treated patients (VP and BKP, adjusted OR = 0.57, p < 0.0001). However, the risk for mortality was 13% lower in BKP-treated patients compared to VP-treated patients (adjusted OR = 0.87, p < 0.0001).

Conclusion

VCFs are the most prevalent type of osteoporotic fractures, although most do not come to clinical attention. There is evidence that they are under-reported, and when they are reported, appropriate intervention is often not initiated. These fractures are associated with back pain, kyphosis disability and increased risk of death. VCFs are linked with an increased risk for subsequent vertebral and other osteoporotic fractures.

Randomized comparative studies comparing non-pharmacological treatment options for patients with VCFs have recently become available—these data are invaluable for medical decision making. This review assessed clinical outcomes of three different treatments for VCFs (NSM, VP and BKP) with endpoints that included reduction in pain, improvement of mobility, vertebral body height restoration, kyphosis amelioration and post-operative complications. In short-term studies, these procedures have not shown to reduce or increase the risk of subsequent vertebral fracture, although long-term effects remain uncertain. An important limitation in this review is the absence of standardization of trial design among the different studies. Other large randomized controlled trials with long-term follow-up are underway and will help to provide new data for physicians to consider vertebral augmentation.

Properly reporting and treating patients with VCFs remain the key if we are to reduce the burden of these debilitating fractures. Options such as VP and BKP are available for stabilizing the fracture and restoring lost vertebral body height, with varying but non-negligible clinical implication to the patient in terms of acute pain relief and improved health. It is important that referring physicians are aware of these techniques and their outcomes.

Patients with VCFs should be assessed for factors contributing to skeletal fragility and managed appropriately to reduce the risk of future fractures. Non-pharmacological treatment of VCFs per se does not reduce the risk of future osteoporotic fractures. As a result, BKP and VP should not be considered as substitutes for the medical management of osteoporosis. Calcium and vitamin D supplementations at adequate dosage should be implemented, along with osteoporosis medication, and patients followed-up regularly for fracture risk assessment.

Acknowledgements

The Fracture Working Group of the IOF Committee of Scientific Advisors was supported by the International Osteoporosis Foundation. None of the authors were compensated for their contribution. Dr. Boonen is senior clinical investigator of the Fund for Scientific Research, Flanders, Belgium (F.W.O.-Vlaanderen) and holder of the Leuven University Chair in Gerontology and Geriatrics.

Conflicts of interest

Steven Boonen has received research funding and consulting fees from Medtronic—but no compensation in relation to this manuscript. All other authors have no competing interests to declare.

Copyright information

© International Osteoporosis Foundation and National Osteoporosis Foundation 2011