Introduction

Vertebral compression fractures (VCFs) can result from a variety of conditions, which include but are not limited to osteoporosis, primary or metastatic spine neoplasms, and some benign bone tumors such as vertebral hemangiomas. Especially with the aging population, osteoporosis remains the most common cause of VCFs.¹ Direct care costs of osteoporotic VCFs have been estimated between $12.2 and $17.9 billion a year. VCFs are associated with significant morbidity and mortality, including impairment of daily activities and psychosocial performance. The kyphotic deformities caused by VCFs are associated with pulmonary dysfunction, constipation, and imbalance. Compared with age-matched control subjects, patients with VCFs have a higher mortality rate, increasing with the numbers of fractures as well as the duration of follow-up.^1–4

Vertebroplasty and kyphoplasty are minimally invasive techniques used to treat painful VCF. Vertebroplasty is the percutaneous injection of a vertebral body (VB) with bone cement, generally polymethylmethacrylate (PMMA). PMMA has been used in orthopedics since the late 1960s.⁵ Percutaneous vertebroplasty (PV) was first reported by a French group in 1987 for the treatment of painful hemangiomas. This case was a “stunner” of a C2 hemangioma, according to the authors. It was painful and got better after cement injection.^6,7 Since then, the indications for PV have expanded to include osteoporotic compression fractures, traumatic compression fractures, and painful vertebral metastasis.^8,9 Kyphoplasty is a modification of PV. It involves the percutaneous placement of balloons (called “tamps”) into the VB with an inflation and deflation sequence to create a cavity before the cement injection. Percutaneous kyphoplasty (PK) may restore some of the VB height and reduce the kyphotic angulation of the compression fracture before PMMA injection.¹⁰

Ideal candidates for PV or PK have activity-related axial pain corresponding to the level of a recent compression fracture. This pain lessens or abates completely with recumbency, sitting still, or both. A complete neurological examination and recent radiographic imaging are mandatory to rule out spinal cord compromise or retropulsed bony fragments in the canal. Magnetic resonance imaging (MRI) shows an increased T2-weighted signal caused by bone edema at the level with a recent fracture. Bone scan has also been used to target the most recent fracture(s) in patients with multiple fractures because uptake of radiotracer has been associated with a higher rate of excellent pain relief compared with PV without correlation with scintigraphy.^11,12 Spinal cord compression on MRI (in the absence of neurological findings) is a relative contraindication. There may be cases in which the procedure is still indicated in a patient with a “tight canal” such as this, but the margin of error is very small, so if a small amount of PMMA extrudes, neurological deficits may ensue. If a posterior cortical fracture is suspected on MRI, a computed tomography scan will reveal the bony architecture more precisely.

Procedural Overview and Complication Avoidance

Standards for the safe practice of these techniques have been published by the Society of Interventional Radiology (SIR) in 2003 and recently updated by the Cardiovascular and Interventional Radiological Society of Europe; highlights of this document can be found in Box 19-1.^13,14

Box 19-1 Summary of Guidelines for Percutaneous Vertebroplasty and Percutaneous Kyphoplasty According to the Society of Interventional Radiology and Cardiovascular and Interventional Radiological Society of Europe

Indications

Painful osteoporotic VCF refractory to 3 weeks of analgesic therapy

Painful vertebrae caused by benign or malignant primary or secondary bone tumors

Painful VCF with osteonecrosis (Kummell disease)

Reinforcement of vertebral body before surgical procedure

Chronic traumatic VCF with nonunion

Absolute Contraindications

Asymptomatic VCF

Patient improving on medical therapy

Active infection

Prophylaxis in a patient with osteoporosis

Uncorrectable coagulopathy

Myelopathy caused by retropulsion of bone or canal compromise

Allergy to PMMA or opacification agent

Relative Contraindications

Radicular pain

VCF >70% height loss

Severe spinal stenosis or asymptomatic retropulsion

Tumor extension into canal or epidural space

Lack of surgical backup

PMMA, polymethylmethacrylate; VCF, vertebral compression fracture.

Before the procedure, the patient should be taken off all anticoagulants, and the coagulation profile should be normal. The platelet count should be at least 50,000 per microliter at the time of the procedure, although no data exist to clearly define a “cut-off” platelet count. In the cancer population, we are asked, “How high do we want the number,” and there have been cases reported of hemorrhage requiring open surgery. We have had patients with prolonged “oozing” after the procedure with marginal coagulation status. There may be instances when the risk-to-benefit ratio favors performing these procedures in patients with lower platelet counts. Active infection and sepsis are contraindications. We recommend waiting at least 2 weeks after treatment of an infection to minimize the risk of infection.

Informed consent should include lack of pain relief, osteomyelitis, fracture of the vertebra or pedicle, extravasation of cement into the spinal canal or neural foramen, paralysis or nerve root damage, and venous embolism. Rare but fatal anaphylaxis to PMMA has been reported in four cases. Also, the rare but potential need for open surgery should be discussed with the patient.

Vertebroplasty and kyphoplasty require the clinician to be trained in spinal anatomy, fluoroscopic imaging, and the use of these techniques to perform interventional procedures. The procedure should be performed in a sterile operating room suite that allows fluoroscopic imaging of the thoracolumbar spine. Biplanar or C-arm fluoroscopy of good quality is mandatory for maximal procedural safety. A radiolucent table is mandatory, as is appropriate padding for prone slightly flexed positioning. Other procedural materials needed include local anesthetic solution (the authors use a 50:50 mixture of 1% lidocaine with 0.25% bupivacaine), PMMA material, and barium or other radio-opacification material, although most cements now come opacified by the manufacturer. Eleven- or 13-gauge bone biopsy needles with connection tubing and cement injection syringes are needed. Numerous commercial “kits” are available.

General anesthesia or monitored anesthesia care (MAC) can be used. If MAC is used, the surgeon must use generous amounts of local anesthetic, especially injected onto the periosteum, where much nociception occurs. Some patients experience discomfort with advancement of the trocars across the posterior cortical margin with balloon inflation (in the case of kyphoplasty) and with PMMA injection. The anesthesiologist must be prepared to “deepen” the MAC during these phases of the procedure. Patient selection is important with consideration to the anesthesia choice. Very anxious or nervous patients may have a better experience with a general anesthetic. Careful consideration must be given to padding the pressure points of this fragile group of patients. After uni- or bipedicular VB access has been obtained, some clinicians proceed directly with injection of PMMA, but others prefer to do venography before cement injection. In theory, venography provides anatomical knowledge of large venous channels’ proximity to the trocar. This information is used to more carefully inject the PMMA. For example, if a small amount of contrast injection reveals a direct spread into a venous channel, the operator may move the trocar before injection or carefully inject relatively solidified PMMA to embolize the large vein before injecting more PMMA into the VB. The literature reveals variable efficacy of the use of venography.^15,16 Most groups no longer use venography.

PMMA injection into the VB is undertaken after careful imaging confirming location of the trocar or trocars into the anteromedial portion of the VB. The PMMA should be opacified and beginning to harden to the consistency of toothpaste before injection. Injection can be done by small syringes filled with PMMA or one of several commercially available kits. The injection must be done under live lateral or biplanar fluoroscopic guidance. If PMMA begins to go into a blood vessel or toward the posterior cortical margin, it must be halted immediately. The authors halt cement injection when it spreads to the posterior third of the VB. To minimize PMMA leakage, several groups recommend the use of high-viscosity cement and relatively small volume injection (Fig. 19-1).^17–19

Fig. 19-1 Anteroposterior and later radiographs of correct balloon placement and of postkyphoplasty results. A, Lateral fluoroscopic view of inflated kyphon balloon tamps intraoperatively. B, Anterior-posterior fluoroscopic view of inflated kyphon balloon tamps intraoperatively. C, Post kyphoplasty lateral radiograph. D, Post kyphoplasty anterior-posterior radiograph.

Outcomes and Potential Complications

Both PK and PV have a very high acceptance and use rate. There is substantial improvement in pain and significant improvement in function after treatment by either of these techniques. Kyphoplasty improves height of the fractured vertebra and improves kyphosis by more than 50% if performed within 3 months from the onset of the fracture (onset of pain). There is some height improvement, although not as marked, along with significant clinical improvement, if the procedure is performed after 3 months.²⁰

In multiple studies, PV has been shown to provide substantial pain relief or improved mobility in 75% to 92.4% of patients with osteoporotic VCFs and in 50% to 86% of patients with pathologic VCFs secondary to neoplasm.^1,8,13

Two studies published in 2009 called the efficacy of PV into question. Although both were randomized controlled trials (RCTs), several problems exist within the methodology of each study. Kallmes et al²¹ published the larger of the two studies, in which during a 4-year period, 131 patients with VCF from 11 centers were randomized to either PV or a “sham” procedure, which was a medial branch block. Pain scores decreased in both the control and PV groups with no statistically significant difference at 1 month (3.9 ± 2.9 for the PV group vs 4.6 ± 3.0 for the control group; P = .19). The authors did note a trend toward a statistically significant difference in clinically meaningful improvement in pain (64% in the PV group vs. 48% in the control group at 1 month; P = .06). Unfortunately, this study has several flaws. Crossover was significant, with 43% of the control group crossing over to PV but only 12% of the PV crossing over to the control procedure. Patients with back pain and fractures up to 1 year old were included, but chronic back pains with an acute VCF were not specifically excluded. Also, the study had a small sample size, necessitating the liberalization of the inclusion criteria and later decreasing the target enrollment because of difficulty in patient recruitment.²¹

Another study published at the same time was performed by Buchbinder et al.²² Seventy-eight patients were recruited over a 4-year period at five institutions and randomized to PV or medial branch block. Patients were not allowed to cross over if pain control was inadequate. The PV group had a decrease in pain from 7.4 ± 2.1 preoperatively to 2.6 ± 2.9 at 3 months, and the control group had a decrease from 7.1 ± 2.3 to 1.9 ± 3.3 over the same time period. In addition, no difference was found in disability or quality of life questionnaires between the two groups. The small sample size makes interpretation of the results somewhat difficult, as does the use of a medial branch block for the control group. In addition, the authors do note they also had difficulty recruiting for their study, and approximately 30% of eligible participants declined to participate.²²

The superiority of one technique over the other remains controversial and is discussed in the following text. As experience grows with these techniques, various groups are pushing the envelope on indications for the procedure, with the possibility of seeing more complications in this patient group. There are some preliminary data and case series on efficacy in patients with radicular pain, traumatic burst fractures, severe VCF or vertebral plana, cervical spine pathology, and intraoperative PMMA augmentation of pedicle screw fixation spinal stabilization.^23–26 In our center, we have pushed “relative” contraindications in cancer patients without increased morbidity even in those with very advanced cancers.^27,28

Complications

Complications are rare but can be serious. The exact incidence is unknown. Most case series report asymptomatic PMMA extrusion rates of around 10% to 15%.^10,37 SIR divides complications for these techniques into two categories: minor and major. Minor complications are those considered to require no therapy and having no consequence, such as PMMA extrusion into the disc. Major complications are those requiring therapy, including an unplanned increase in the level of care needed or ongoing permanent sequelae (i.e., PMMA into the spinal canal with neurological deficit). SIR noted published complication rates for major complications to be less than 1%, except in those with neoplastic involvement of the vertebrae, in whom the reported level of major complications is less than 5%.¹³

PMMA can flow out of the VB posteriorly into the spinal canal or neural foramina or anteriorly into the paraspinous veins with systemic consequences. There are case reports of nerve root and spinal cord compression from extravertebral PMMA.^38,39 There have been several reports of minimally symptomatic pulmonary emboli, one case of cardiovascular collapse requiring pulmonary embolectomy, one lethal pulmonary embolus, and one case of paradoxical cerebral arterial PMMA emboli.^40–43 The literature suggests that there may be less PMMA leak with PK versus PV.⁴⁴

Infectious complications, although rare, have been reported. There have been several reports of osteomyelitis requiring corpectomy.⁴⁵ Meticulous attention to sterile technique is warranted, including preoperative intravenous antibiotic administration. Most have abandoned the older technique of adding tobramycin powder to the PMMA because of uncertainty over both the efficacy and potential impact on the PMMA properties.

Complications have been reported with both procedures (vertebroplasty and kyphoplasty); a review of U.S. Food and Drug Administration safety data revealed 58 reported complications from 1999 through 2003 of approximately 200,000 procedures performed. These were approximately evenly divided among PV and PK, with more cases of pedicle fracture and spinal cord compression in PK than PV.⁴⁶ This voluntary reporting system is almost certainly flawed and most likely underreports the overall incidence of complications. As further studies are performed, a more complete risk-to-benefit ratio can be defined.

Vertebroplasty versus Kyphoplasty

Both PV and PK have been shown to be effective to reduce the pain associated with VCFs (Box 19-2). These procedures have low complication rates and tend to have islands of supporters (i.e., clinicians who always favor one over the other regardless of clinical circumstances). One study has been published in the past few years addressing PV versus PK. Liu et al⁴⁷ enrolled 100 patients with confirmed osteoporotic fractures at the thoracolumbar junction and randomized them to PV or PK. There was no difference in preoperative visual analog scale (VAS) pain score, VB height, or kyphotic wedge angle. Postoperative VAS pain scores (2.6 ± 0.6 for both PK and PV) were significantly different from preoperative pain scores (8.0 ± 0.8 for PK; 7.9 ± 0.7 for PV) but were not statistically different between PV and PK at any time in the postoperative period up to 6 months. Although no difference in pain relief was noted, there was a statistically significant increase in both VB height and reduction of the kyphotic wedge angle in the PK compared with PV. In PK, the VB height increased from 1.13 ± 0.22 cm to 2.04 ± 0.41 cm, and the kyphotic wedge angle decreased from 17. 0 degrees ± 7.3 degrees to 9.0 degrees ± 5.7 degrees. This effect was less pronounced in PV, in which the VB height increased from 1.01 ± 0.22 cm to 1.32 ± 0.26 cm and the wedge angle decreased from 15.5 degrees ± 4.2 degrees to 12.2 degrees ± 3.6 degrees.⁴⁷ Currently, the authors favor the use of PK in circumstances in which the patient has a collapse of more than 20% of vertebral height and a fracture age of less than 3 months in hopes of restoring vertebral height and maintaining functional anatomy when possible. Deramond et al⁴⁸ have stated that PK may be preferable over PV in patients with severe or multiple wedge deformity that has developed in the past 3 weeks.

Box 19-2 Vertebroplasty versus Kyphoplasty

Vertebroplasty

Less expensive

Faster for the operator and patient

Indicated to treat even older fractures

Kyphoplasty

More anatomical correction of spinal deformity than vertebroplasty

Greater height restoration in recent fractures (<3 months old)

Indicated for patients with extensive kyphosis caused by multiple recent VCFs

Less PMMA extravasation

PMMA, polymethylmethacrylate; VCF, vertebral compression fracture.

Advances in Technology

Coblation

Patients with spine tumors have an increased risk of perioperative complications, and in the setting of limited life expectancy, this can lead to an unacceptable quality of life.⁴⁹ Using less invasive techniques to treat VCF and other spinal lesions can potentially help avoid this problem. As described in the introduction, many patients with spinal tumors have VCF. Although PV and PK are highly beneficial in this patient population, additional therapies have been developed that pain physicians can use for treatment of the tumor and to potentially improve the efficacy of PV or PK.

Plasma-mediated radiofrequency ablation (coblation) is currently used in procedures such as tonsillectomy, disc decompression, and tendon and cartilage debridement in which dissolution of the tissue is necessary. A plasma field is created by passing radiofrequency waves through a conducting medium such as saline, creating a precisely focused plasma field composed of ionized particles. This field, formed immediately adjacent to the tip of the device, causes dissolution of soft tissue at the relatively low temperatures of 40° to 70° C.⁵⁰ Coblation has the potential to increase safety and efficacy in a number of ways. Such a device allows for cavity creation in the VB with little mechanical disturbance of tissue outside of the plasma field. Minimizing disturbance of the rest of the VB could allow for vertebral augmentation in cases of retropulsion of tumor or cortical wall that would otherwise be a contraindication to PV or PK.⁵¹ Dissolution of the surrounding soft tissue also allows for destruction of spinal primary or metastatic tumors and increases the efficacy of chemotherapy or radiation therapy by cytoreduction. This is accomplished by removing and replacing the coblation device at different positions of the clock, typically at the 12, 2, 4, 6, 8, and 10 o’clock positions, thus forming a void in the VB larger than the device itself (Fig. 19-2). In conventional PK for treatment of tumor-related VCF, the tumor is pressed against the healthy bone, forming a rim of tissue between the cement and bone. If the practitioner were to choose to perform PK in addition to coblation, removal of the tumor can allow for better interdigitation of the cement in the VB.⁵² In addition, the relatively low temperatures of the plasma field causes less heat transfer to the VB and assists with hemostasis in the VB.⁵²

Fig. 19-2 Lateral fluoroscopic images showing creation of void in the vertebral body with a coblation device. A, Lateral intraoperative fluoroscopic image of tumor coblation. B, Lateral intraoperative fluoroscopic image of tumor coblation in same case showing intra-vertebral body sweeping motion to ablate tumor.

Conclusion

PV and PK are minimally invasive techniques used to treat painful VCFs. There is a growing body of evidence, albeit of limited quality (predominately open case series), indicating that these procedures are efficacious in alleviating the pain associated with VCF. The results of the procedures in these reports are uniformly good. However, there are a growing number of case reports of serious complications.

Recent reviews and editorials have called for a more critical evaluation of these procedures. Watts et al⁵⁷ reviewed the literature and concluded that controlled multicenter trials are needed to determine the short- and long-term safety. Garfin et al²⁰ concluded that there is a 95% improvement in pain and significant improvement in function after these procedures. They emphasized that the procedure is technically demanding with the potential for significant complications. They recommended further efficacy and safety studies. Jarvik and Deyo⁵⁸ called for RCTs or some type of control cohort to compare long-term outcomes carefully. Einhorn⁵⁹ calls for careful monitoring of outcomes and minimal training standards. Birkmeyer⁶⁰ calls for randomized clinical trials, citing insufficient evidence from case series to prove safety, efficacy, and cost-effectiveness. The prospective RCTs described above seem to confuse rather than clarify our understanding of the efficacy of PV. Ideally, a prospective RCT with a larger sample size and appropriate control group can be performed in the future.

It will be difficult to conduct the RCTs needed to compare short- and long-term outcomes of PV or PK versus more conservative therapies. These procedures have gained such widespread popularity that patients would undoubtedly resist being randomized to the conservative treatment group; recent attempts at RCTs for PV have been plagued by low enrollment. Other studies need to be done to compare PV and PK in various disease states in a randomized fashion. Early studies are underway to evaluate biologic materials for spinal injection rather than acrylic (PMMA). Despite the need for more research, PV and PK have shown great promise in the treatment of painful VCFs caused by a variety of different pathologic states. With careful patient selection, adequate training, and attention to details during the procedure, serious complications are rare.