31 Vertebro/Kyphoplasty for Osteoporotic Fractures

Justin E. Costello and Troy A. Hutchins

Abstract

In total, 700,000 known osteoporotic vertebral compression fractures (VCFs) occur annually in the United States. The actual incidence is likely higher as only an estimated third of all fractures are clinically diagnosed. ¹ Osteoporotic VCFs can lead to significant disability and morbidity and are associated with increased risk of comorbidity-related mortality. ² Percutaneous vertebral augmentation (PVA) techniques have been used for 30 years as adjuncts to medical management for pain palliation in VCFs. However, controversy exists in the literature regarding appropriate patient selection and clinical benefit compared to medical management. ³ A working knowledge of the current evidence is necessary for the neurointerventionalist to appropriately select patients who are most likely to benefit from PVA and to provide up-to-date information to referring clinicians and patients during consultation.

Key words

kyphoplasty – vertebroplasty – percutaneous vertebral augmentation – vertebral compression fractures – osteoporosis

31 Vertebro/Kyphoplasty for Osteoporotic Fractures

31.1 Goals

Review the literature regarding our basic understanding of percutaneous vertebral augmentation (PVA) for osteoporotic vertebral compression fractures (VCFs), emphasizing indications, contraindications, and potential complications.
Understand appropriate patient selection and effects of timing of intervention relative to fracture age when treating osteoporotic VCFs with PVA.
Critically analyze the literature regarding efficacy of pain palliation with PVA compared to medical management for osteoporotic VCFs.
Review advantages and disadvantages of vertebroplasty (VP) versus kyphoplasty (KP) for treating osteoporotic VCFs.
Review the current literature regarding mortality risk for patients with osteoporotic VCFs treated with PVA compared to those managed medically.

31.2 Case Example

31.2.1 History of Present Illness

A 77-year-old Caucasian female presents with 5 weeks unchanged lower back pain, which began suddenly after bending over in the garden. Initial magnetic resonance imaging (MRI) demonstrates a Til VCF with associated marrow edema (Fig. 31.1). Her pain is rated as an 8 out of 10 on visual analog scale (VAS), despite several weeks of conservative treatment, including bracing and oral pain medications. Her activities of daily living are severely limited by pain, and she spends most of her days in bed or in a chair. She denies neurologic deficit, bowel/bladder incontinence, or fevers.

Past medical history: Osteoporosis (T score: -2.7), no history of cancer, steroid use, unexplained weight loss, active infection, bleeding disorder, or anticoagulant therapy.

Past surgical history: Hysterectomy for uterine leiomyomas.

Family history: Noncontributory.

Social history: No IV drug use, nonsmoker.

Review of systems: As per the above.

Physical examination: Antalgic gait, tender to palpation lower back Tl 1 level, neurologically intact.

Imaging studies: See Fig. 31.1.

In Fig. 31.1(a-c) Sagittal T2, short tau inversion recovery (STIR), and Tl-weighted MR images of the thoracic spine demonstrate a Til compression fracture. There is 75% vertebral body height loss, with diffuse edema signal on STIR images (Fig. 31.1b), consistent with acute fracture. Fracture line is also evident on Tl-weighted imaging (Fig. 31.1c). There is minimal retropulsed fragmentation, with intact posterior longitudinal ligament (Fig. 31.1a).

31.2.2 Treatment Plan

After review of potential benefits, risks, and other treatment options, the patient elected to undergo treatment of the Til compression with PVA. The Tl 1 vertebral body was accessed via bilateral transpedicular approach (Fig. 31.2). Following manual drilling (Fig. 31.2), KP balloons containing saline and contrast are inflated to create small cavities within the vertebral body (Fig. 31.3a). Cement is then injected slowly into the vertebral body and monitored in real time. A small amount of cement leakage is seen within the T10-T11 disc space (Fig. 31.3b), an incidental finding that does not increase risk of adjacent-level vertebral fracture.

31.2.3 Follow-up

The patient did well with decreased use of pain medications and back brace at 2 weeks and VAS pain score reduced to 2/10. She also showed increased mobility and at 4 weeks was able to return to gardening, with no residual pain or dysfunction. Follow-up radiographs demonstrated maintained vertebral body height with no new fracture or deformity.

31.3 Case Summary

What are the key indications and contraindications that should be assessed in this patient prior to performing PVA??
1. Preprocedural imaging assessment:
  
  In the setting of an osteoporotic VCF, MRI is the most important imaging modality for preprocedural assessment. ⁴ MRI should demonstrate an acute vertebral body compression deformity with marrow edema on STIR images (Fig. 31.1b). MRI may also delineate unstable VCFs with significant retropulsed osseous fragmentation or posterior longitudinal ligament injury. Careful review of prior imaging immediately prior to PVA is important for planning of vertebral body access and for potential identification of new VCFs during the procedure.
2. Pain must be attributed to the vertebral fracture: Appropriate patient selection is essential as lower back pain is very common, with multiple potential etiologies. ⁵ Pain should be new, focal, and localized to the treatment vertebral level on physical examination. ⁴ ^, ⁶ Pain will typically worsen with weight-bearing without radicular symptoms. ⁴ ^, ⁶ Pain levels should be severe enough to impair activities of daily living and can be quantified with questionnaires or the VAS. ⁴ It is also recommended that patients have failed conservative management, including use of analgesics and back bracing. Pain severity score, impact on activities of daily living, clinical examination, and failed conservative management should be documented prior to treatment. c) Absolute contraindications:
3. Absolute contraindications for PVA include irreversible coagulopathy, allergy to cement or contrast agent, and active infection, including local cellulitis, discitis-osteomy-elitis, epidural abscess, or systemic infection. ⁷ PVA should be postponed in patients with a fever or suspected sepsis.
4. Relative contraindications:
  
  Advanced vertebral collapse is a relative contraindication to PVA. ⁷ Usually at least 25% vertebral height is required to safely perform PVA. Unstable vertebral fractures with significant retropulsed fragmentation and fractures causing neurologic symptoms should generally not be treated with PVA. ⁷ Patients who are unable to lie prone are usually not candidates for PVA.

What are the most common complications of vertebral augmentation in the setting of osteoporotic VCF?
1. Cement leakage:
  
  Cement leakage is the most common complication, reported in up to 88% of patients with PVA, ⁸ though the actual rate of cement leak is likely much lower. More importantly, the majority of cement leaks are asymptomatic and of no clinical significance, occurring within the paravertebral soft tissues or disc space (Fig. 31.3b). There has been a controversy regarding disc cement leaks and higher risk of adjacent-level compression fractures ⁹ ^, ¹⁰ ^, ¹¹ ^, ¹² ; however, recent randomized controlled studies, the VERTOS II ¹³ and VERTOS IV ¹⁴ ^, ¹⁵ trials, found no increased risk of adjacent-level vertebral fracture following PVA. The authors of the VERTOS IV study concluded that adjacent-level VCFs following PVA are the result of osteoporotic disease and not related to vertebral cementation.
  
  Uncommonly, cement leaks can occur in the epidural space, neural foramina, paravertebral veins, or vertebral venous plexus, where there is potential for clinically relevant complications. The overall rate of symptomatic neurologic compromise following PVA for osteoporotic compression fracture is rare and reported at< 1%. ⁷ To minimize this risk, bone cement filling must be performed slowly and carefully monitored with real-time fluoroscopy.
  
  Cement pulmonary embolism is common following PVA and reported to occur in 3.5 to 23% of osteoporotic compression fractures. ¹⁶ The majority of pulmonary emboli related to PVA involve subsegmental pulmonary arteries and are asymptomatic. ¹⁶
2. Other complications:
  
  Other potential complications are reported to occur at <1%, including infection, significant hemorrhage, allergic reaction, fracture, symptomatic hemothorax or pneumothorax, and death. ⁷ During PVA, intravenous antibiotic prophylaxis is recommended for immunocompromised patients; however, there is no consensus for antibiotic administration in immunocompetent patients. ¹⁷

Is PVA effective for treatment of osteoporotic VCF? Are there additional PVA treatment considerations?
1. PVA effectiveness:
  
  For treatment of osteoporotic VCFs, PVA has been shown to result in effective (and often immediate) pain relief through multiple prior studies. ¹³ ^, ¹⁸ ^, ¹⁹ ^, ²⁰ ^, ²¹ The most recent randomized controlled studies conferring PVA effectiveness for osteoporotic VCFs include the fracture reduction evaluation (FREE), ¹⁹ VP versus conservative treatment in acute osteoporotic compression fractures (VERTOS II), ¹³ and safety and efficacy of VP for acute painful osteoporotic fractures (VAPOUR) ²⁰ trials.
  
  The FREE (published in 2009) and VERTOS II (published in 2010) trials prospectively compared PVA to medical management for treatment of acute osteoporotic VCFs. Both studies concluded superior pain relief at 1 month for patients treated with PVA versus medical management. Pain relief was also durable for 1 year or greater in both PVA treatment groups. The main criticism of both studies was lack of blinding and potential for placebo effect in the PVA treatment group.
  
  The VAPOUR trial (published in 2016) prospectively compared PVA to a PVA sham procedure for treatment of acute osteoporotic VCFs. The study concluded superior pain relief for PVA versus the placebo procedure. Patients were followed to 6 months after the procedure and the primary end point of the study was pain below 4 out of 10 at 14 days postintervention.
2. PVA timing:
  
  For treatment of osteoporotic VCF, PVA has been shown to be most effective for recent vertebral fractures. ¹³ ^, ¹⁸ ^, ¹⁹ ^, ²⁰ ^, ²¹ The VERTOS II ¹³ and VAPOUR ²⁰ trials only included patients with vertebral fractures < 6 weeks old. The FREE ¹⁹ trial showed benefit for PVA treatment of subacute vertebral fractures < 3 months old. A recent single-arm prospective study, the EVOLVE ²¹ trial (published in 2019) also demonstrated pain benefit for PVA treatment of osteoporotic VCFs up to 4 months old.
  
  Currently, there is no randomized controlled data to support effective pain palliation from PVA for chronic VCFs (>4 months old), and medical management should be considered as the first-line therapy in these patients. An upcoming study, the VERTOS V trial, ²² which is not yet published, is currently investigating the efficacy of VP for patients with persistent pain and chronic unhealed osteoporotic VCFs.
3. Vertebroplasty (VP) versus kyphoplasty (KP):
  
  Both VP and KP involve similar percutaneous vertebral access using trocar systems; however, in KP, balloon-assisted cavities are created prior to cement injection. Clinical outcomes are similar for patients treated with VP or KP. A recent meta-analysis and systematic review of the literature ²³ found no difference in pain control for osteoporotic VCFs treated with VP or KP.
  
  However, because of the balloon-assisted cavities, cement control can be easier during KP and studies have found that the rate of cement leak is significantly lower for KP compared to VP. ²³ As stated previously, most cement leaks are of no clinical significance, with most occurring in the paravertebral or disc spaces. Thus, the benefit of decreased cement leaks with KP is likely marginal. Studies have also found significantly decreased kyphotic wedge angles and increased vertebral body height restoration in patients treated with KP versus VP. ²³ However, as with cement leaks, the clinical benefit of height restoration has not been well validated and it may also be clinically insignificant. More studies are needed to further investigate KP vertebral height restoration, particularly in the setting of advanced VCFs (> 50% vertebral height loss).
4. Cement injection volume:
  
  There has been controversy regarding the effect of PVA injection volume and its relationship to treatment outcomes. Two studies published in 2012 (prospective follow-up study, 196 painful osteoporotic compression fractures) ²⁴ and 2016 (retrospective review, 108 patients) ²⁵ have suggested better pain relief with higher VP cement injection volumes, albeit at the detriment of higher rate of cement leaks. An additional retrospective study published in 2006 (158 patients) ²⁶ concluded no clinical benefit for higher VP cement injection volumes and emphasized maximum safe filling of vertebral bodies.

Is PVA more effective than conservative medical management for treatment of acute osteoporotic VCFs?
1. 2009 NEJM sham procedure studies:
  
  In 2009, two randomized controlled multicenter studies were published in the New England Journal of Medicine ²⁷ ^, ²⁸ comparing VP to a sham placebo procedure for treatment of osteoporotic VCFs. Although one study reported a trend for better pain relief in the VP versus placebo group (64% pain controlled in VP group at 1 month vs. 48% in placebo group), ²⁷ both the studies concluded that there was no statistically significant difference in short-term pain relief between the two groups. At the time, these studies incited much debate in the medical community as to the efficacy and appropriateness of PVA for treatment of osteoporotic VCFs.
  
  Both studies had limitations, primarily patient selection with inclusion of chronic osteoporotic VCFs. Both study designs randomized treatment for patients with VCFs <1 year in age. In the study by Kallmes et al, ²⁷ 36% of VCFs were > 6 months old and edema on MRI was not a requirement for study participants. In the study by Buchbinder et al, ²⁸ 25% of VCFs were > 3 months old and physical examination was not required for study participants.
2. VAPOUR study:
  
  The VAPOUR trial, ²⁰ published in 2016 by Clark et al, also randomized patients to either VP or a placebo sham procedure. The major difference in study design versus the 2009 NEJM studies was patient selection. The VAPOUR trial only included patients (120) with fracture ages of < 6 weeks and nearly 80% of study participants had fracture ages of < 3 weeks. Baseline VAS and numeric-rated score for pretreatment pain assessment were also greater, 7 out of 10 or more in all study participants. The study found statistically significant pain relief for the VP treatment group versus the placebo group at both 2 weeks and 6 months follow-up. Also, many of the patients in the study were admitted inpatients (57%) and the study found a reduced mean hospital stay of 5.5 days for the VP treatment group.
3. VERTOS IV study:
  
  VERTOS IV, published in 2018 by Firanescu et al, ¹⁴ was the latest randomized controlled study to compare VP to a placebo sham procedure. The study included 180 patients with VCF, pain up to 9 weeks, marrow edema on MRI, and VAS pain scores of 5 or greater. Pain scores were subsequently followed at day 1, week 1, and months 1, 3, 6, and 12 following VP or placebo procedure. The study concluded no significance in pain relief between the VP treatment and placebo groups at all follow-up time points.
4. Where do we stand now?
  
  The study designs, with inclusion of chronic VCFs, calls into question the validity of the 2009 NEJM vertebroplasty sham studies. However, two subsequent randomized controlled studies with similar placebo sham procedures, VAPOUR and VERTOS IV, included only acute or subacute fractures, with conflicting results regarding VP pain control. Thus, it is important to understand the key differences in these two studies.
  
  The placebo sham procedure in the VAPOUR study included conscious sedation, anesthetic injection into the subcutaneous soft tissues only, and did not use a poly-methyl methacrylate (PMMA) kit to produce a cement smell. For the VERTOS IV sham procedure, there was infiltration of both the subcutaneous soft tissues and pedicle periosteum, without conscious sedation, and the PMMA kit emitted a smell during the procedure. Greater than 80% of placebo patients believed they had received cement during the VERTOS IV study, versus 46% in VAPOUR, leading authors of VERTOS IV to raise the possibility of increased placebo effect in the VERTOS IV sham control group. VERTOS IV has also suggested a potential therapeutic benefit from pedicle periosteal infiltration of anesthetic.
  
  Patient selection was also different between the two studies. The VAPOUR study included only fracture ages of < 6 weeks and most (80%) were < 3 weeks old. VERTOS IV included fracture ages of up to 9 weeks, with a median fracture age of 6.1 weeks for VP patients. Preprocedure pain score criteria were higher for VAPOUR (at least 7 out of 10 pain score) versus VERTOS IV (at least 5 out of 10 pain score). VAPOUR also had a large proportion of hospital inpatients (57%) as part of the study, whereas VERTOS IV included only outpatients.
  
  These differences in study designs highlight the importance of patient selection when performing VP for osteoporotic VCFs. VCFs treated with VP should be < 6 weeks in age and preprocedure pain score should be severe (at least 7 out of 10 VAS), with a failed course of conservative therapy. An additional, similarly designed randomized trial, the VP with a sham procedure for painful acute osteoporotic vertebral fractures (VOPE) trial, ²⁹ has yet to be published. Comparison of VOPE to VAPOUR and VERTOS IV will be useful, when available.

Does PVA confer a mortality benefit in patients with acute osteoporotic VCF?

Increased mortality risk, estimated at 2 to 42% for 12 months, for patients following an osteoporotic VCF can be substantial. Multiple large retrospective studies have concluded a significant long-term mortality benefit for patients with osteoporotic VCFs treated by PVA.

In 2010, a retrospective database study ³⁰ of 5,766 non-neo-plastic VCFs found a significantly lower in-hospital mortality rate for patients treated with KP compared to conservative management (0.3 vs. 1.6%). Improved measures of hospital discharge were also identified with a higher rate of discharge to home for KP patients compared to conservative management (38.4 vs. 21%). Similar mortality benefit was demonstrated by three subsequent large retrospective studies. ³¹ ^, ³² ^, ³³ In a Taiwanese study, ³³ the risk of respiratory failure was also significantly lower for the VP group, suggesting that pulmonary impairment may contribute to increased mortality in patients treated with conservative management.