Percutaneous Vertebroplasty For Vertebral Fractures Caused By Osteoporosis And Malignancy

Back pain and related problems have substantial public health impact within the Medicare population each year, including discomfort, loss of mobility, and serious morbidity. CMS has identified this important issue for long-term examination. Disease conditions interest include disc degeneration, spinal stenosis, spondylolisthesis, and vertebral compression fractures.

Vertebral body compression fractures of the spine are among the most common fracture types in patients who have osteoporosis. Compression fractures of the spine are also caused by osteolytic destruction, secondary to malignancy. Pain and loss of function are the most common symptoms.

Percutaneous vertebroplasty is a minimally invasive treatment that inserts bone cement into the compressed and fractured vertebrae to provide mechanical stabilization. Kyphoplasty is a variation of vertebroplasty. Kyphoplasty uses an inflatable balloon to expand the compressed vertebral body, in an attempt to restore its natural height before injecting a cement-like substance into the vertebral cavity.

As an initial step, CMS will review the scientific evidence on the effectiveness of surgical management of vertebral compression fractures, particularly vertebroplasty and kyphoplasty procedures, in the Medicare population. We will refer this issue to the Medicare Coverage Advisory Committee (MCAC). Committee members will be given presentations on the current literature and also receive public comments to aid in the discussions and recommendations regarding the quality of the evidence on verterbroplasty and kyphoplasty procedures.

We expect to hold additional future meetings of the Medicare Coverage Advisory Committee regarding other back pain treatments and will post future notices in the Federal Register.

DRAFT

PERCUTANEOUS VERTEBROPLASTY FOR VERTEBRAL FRACTURES CAUSED BY OSTEOPOROSIS AND MALIGNANCY

ASSESSMENT OBJECTIVE

This Assessment evaluates the available evidence to determine whether percutaneous vertebroplasty (PVP) is demonstrated to be an effective treatment. Percutaneous vertebroplasty is a minimally invasive treatment involving percutaneous needle injection of bone cement into a diseased vertebral body. The primary uses reported in the literature include treatment of: 1) osteoporotic vertebral compression fracture, and 2) vertebral fractures caused by osteolytic destruction secondary to malignancy. Beneficial effects of interest include relief of associated symptoms (e.g., pain) as well as improvement in ability to function (e.g., mobility and activities of daily living). Adverse effects would include complications associated with percutaneous vertebroplasty.

BACKGROUND

Osteoporotic Vertebral Compression Fracture

Vertebral fractures are among the most common fractures in patients who have osteoporosis (Ross 1997). It is estimated that up to 50% of women and approximately 25% of men will have a vertebral fracture at some point in their lives. Multiple vertebral fractures may be expected in about half of these cases (Ross 1997). While the incidence of vertebral fracture is high, a minority of vertebral fractures (about one-third) actually reaches clinical diagnosis (Ross 1997).

Acute vertebral fracture often present with pain, although clinically silent fractures may account for one-half of all radiographically visible vertebral fractures (Ross 1997). Pain management in the acute setting is not standardized. Common management strategies include bed rest, activity modification, and local and/or systemic analgesics. Calcitonin has also been suggested to reduce pain in the acute setting. Postural bracing of the spine is another option, but this has not been reported recently in the literature (Lin and Lane 2002).

Chronic pain, usually following multiple vertebral fractures, does not tend to respond to the management strategies described for acute pain. The source of chronic pain after vertebral compression fracture is not thought to be the vertebrae itself but is believed to be related to strain on muscles and ligaments secondary to kyphosis. This type of pain frequently is not improved with analgesics and may be better addressed through exercise (Ross 1997).

Vertebral Metastasis/Multiple Myeloma Lesions

Metastatic disease involving the spine generally involves the vertebral bodies with pain being the most frequent complaint (Healey 1997). Pain may be caused by any number of factors such as intraosseous tumor, vertebral fracture with associated segment instability, or extraosseous tumor producing spinal or nerve root compression. Such compression may also cause neurological dysfunction.

Prognosis in patients with vertebral metastasis is variable and relates to a number of factors such as the patient’s underlying functional status and primary tumor as well as the anatomic location of the metastasis. Restored ambulation has been associated with increased survival (Healey 1997).

Palliative treatment options include radiation therapy, chemotherapy, and/or surgical resection with fixation for stabilization. The pain associated with metastasis alone is usually quite responsive to radiation therapy. The pain due to fractures caused by osteolytic destruction due to the metastases are often more difficult to manage. External bracing devices may also be used, but long-term use may be difficult for patients. Nonsurgical intervention may be preferable in patients with limited expected survival and poor functional status.

While radiation and chemotherapy are frequently effective in reducing tumor burden and associated symptoms, pain relief may be delayed days to weeks depending on tumor response and the presence of fracture. Furthermore, these approaches rely on bone remodeling to regain vertebral body strength, which may necessitate supportive bracing to minimize the risk of vertebral collapse during healing.

Percutaneous Vertebroplasty

Percutaneous vertebroplasty (PVP) is a minimally invasive treatment involving percutaneous needle injection of bone cement into a diseased vertebral body. The procedure was first reported by investigators from France in 1987 as a treatment for complicated vertebral body hemangioma (Galibert et al. 1987). Since that time, the PVP technique has been further investigated, both in the United States and in Europe, as a treatment option to provide mechanical support and symptomatic relief in other conditions involving osteolytic destruction of the spine. Osteoporosis, vertebral metastasis and vertebral involvement of multiple myeloma are the most commonly reported uses of PVP in the literature. Other reported uses for PVP include treatment for symptomatic or aggressive vertebral hemangiomas, Langerhans cell histiocytosis (also known as eosinophilic granuloma) or vertebral lymphoma (Cardon et al. 1994; Martin et al. 1999).

It has been proposed that PVP provides an analgesic effect through mechanical stabilization of a fractured or otherwise weakened vertebral body (Deramond et al. 1998). Thermal damage to intraosseous nerve fibers is another possible mechanism of effect, since the cement used in the procedure, polymethyl methacrylate (PMMA), undergoes an exothermic (heat-releasing) reaction when it hardens. One study performed PVP injections in cadaveric vertebrae and measured resultant in vitro temperature changes (Deramond 1999). The results indicate a possible role of thermal necrosis as an explanation for the pain relief observed after PVP. However, differences between in vivo conditions and the cadaveric study conditions preclude definitive conclusions. Chemical or vascular effects have also been discussed (Cotten and Duquesnoy 1997), but at present, the mechanism of analgesia is not well understood.

Description of PVP Technique. For the purpose of this Assessment, vertebroplasty is considered to be a percutaneous procedure and will be referred to as percutaneous vertebroplasty (PVP). The vast majority of published reports describe using a percutaneous approach. However, one published report from Switzerland (Wenger et al. 1999) describes vertebroplasty with an open surgical field in order to remove polymethyl methacrylate (PMMA) leakage immediately.

During PVP, PMMA is injected into a diseased vertebral body. PMMA is made radiopaque by the addition of barium sulfate powder and tantalum powder. The injection is performed by introducing a needle (usually 10–15 gauge, depending on the spinal level) through a transpedicular or paravertebral approach into the vertebral body. Either fluoroscopic or computed tomographic (CT) guidance is used to guide needle placement. Some investigators perform a venogram through the needle to delineate tip placement and venous outflow in an effort to avoid venous leakage of PMMA. The usefulness and necessity of pre-procedural venography has been debated in the literature (Wong and Mathis 2002; Do 2002; McGraw et al. 2002; Vasconcelos et al. 2002; Gaughen et al. 2002). PMMA is then injected slowly through the needle and monitored using fluoroscopic or CT imaging. If filling of the vertebral body is insufficient on a unilateral injection, then a second injection may be made using the contralateral approach.

PVP requires some degree of anesthesia; however, practices vary across reported series. Some investigators routinely used general anesthesia early in their experience, while more recent series use conscious sedation anesthesia techniques. The patient must lie prone for the entire procedure, which may add to the discomfort of the procedure, and some patients may require general anesthesia if they are unable to tolerate lying prone for several hours.

Patient Screening, Evaluation and Selection. Evaluation of the patient is necessary to establish the fractured vertebrae as the source of the pain. This includes a clinical history and physical examination to establish significant focal back pain, point tenderness, and limited mobility. Many imaging tests can identify fractured vertebrae including plain spine X-rays. Other tests commonly used include nuclear medicine bone scanning, MRI, CT, and fluoroscopy. Some of these tests may identify fractures that cannot be visualized on plain X-ray, others can differentiate between acute and old fractures, and others might indicate other reasons for back pain. Although both acute and old fractures are treated with vertebroplasty, differentiating between them might be necessary when the patient has multiple vertebral fractures. Certain anatomic configurations of the vertebral fracture are contraindications to performing vertebroplasty, such as vertebral body collapse of more than 90%. The age of fracture and specific MRI findings have been found by some investigators to be associated with outcomes, but it is uncertain whether these factors are used to select or exclude patients from consideration. In a study of the related procedure kyphoplasty, Kasperk et al. (2005) documented the eligibility of patients with severe pain and osteoporosis who were actually candidates for kyphoplasty. Out of 211 patients evaluated, 97 (43%) were considered eligible for the procedure.

Complications of PVP. Adverse effects of PVP include localized bleeding, infection, and/or resultant pain or neurological symptoms following leakage of injected material. Injected material may also leak into the systemic venous system with the potential for pulmonary embolism. In addition, complications from patient positioning and anesthesia for the procedure may include rib fracture or systemic infection.

PMMA leakage has been reported into the venous plexus, inferior vena cava, peridural space in the spinal canal, neural foramina, intravertebral disk space, and paravertebral soft tissue. The majority of PMMA leakages are asymptomatic; however, infrequently leakages have necessitated steroid therapy and/or surgical removal of extruded material. The question of whether pretreatment venography influences the safety and effectiveness of PVP has been debated recently in the literature (Wong and Mathis 2002; Do 2002; McGraw et al. 2002; Vasconcelos et al. 2002; Gaughen et al. 2002).

Several published case reports document significant adverse events following PVP. Case reports of major neurological complications have been reported following PMMA leakage into the surrounding neurological structures (Harrington 2001; Wenger and Markwalder 2002; Ratliff et al. 2001; Wilkes et al. 1994) though some of these complications resolved after surgical removal of PMMA. Harrington (2001) reported a 66-year-old patient who developed persistent acquired thoracolumbar spinal and foraminal stenosis, with cramping thigh pain and progressive numbness and weakness of both legs after walking only a short distance. Paradoxical cerebral artery embolization of cement was reported during an open surgical vertebroplasty with multiple pulmonary emboli of PMMA that caused pulmonary hypertension and a right-to-left shunt through a patent foramen ovale (Scroop et al. 2002). Another case of PMMA pulmonary embolus was reported by Padovani et al. (1999). Transient arterial hypotension occurred following PMMA injection in a case reported by Vasconcelos et al. (2001), although no adverse sequelae occurred in that case.

An occurrence that has been noted but is uncertain as to whether it represents a complication of vertebroplasty is subsequent fractures occurring in adjacent vertebrae. Vertebrae treated with cement are stiffer than fractured vertebrae, and this may transmit increased force to adjacent vertebrae (Fribourg et al. 2004). The occurrence of subsequent fractures has been recorded in some of the case series included in this assessment. Case series studies of Uppin et al. (2003) Grados et al. (2000) reported subsequent re-fracture rates of 12 to 52% over varying periods of time. The study by Fribourg et al. estimates a rate of symptomatic recurrent fractures in an untreated population to be 5% per year, and that increases beyond that may be due to the additional stiffness caused by vertebroplasty or kyphoplasty. However, patients undergoing these procedures are at increased risk due to their underlying osteoporosis. Thus evidence is suggestive, but not definitive, that vertebroplasty increases the risk of subsequent fracture.

Comparative Clinical Studies

Evaluation of vertebroplasty has been complicated by a lack of randomized, controlled trials. Dr. David Kallmes, a neuroradiologist and Assistant Professor at the Mayo Clinic in Rochester, received NIH funding for a randomized, sham-controlled study comparing PVP and sham procedure with crossover at 4 weeks in 294 subjects. The primary outcome will be functional status on the Roland scale (2-point improvement). Secondary endpoints will include pain assessment, SF-36, and cost-effectiveness analysis. This trial has been ongoing since September 2003 and is currently recruiting patients.

Two other small, randomized, controlled comparison studies have been reported in abstract form and these will be summarized in the Assessment's "Review of Evidence" (Kallmes et al. 2002; Do et al. 2002).

FDA Status

Vertebroplasty is a surgical procedure and is not subject to U.S. Food and Drug Administration (FDA) approval.

Polymethyl methacrylate (PMMA) bone cement was available as a drug product prior to enactment of the FDA’s device regulation and was at first considered what the FDA terms a “transitional device.” It was transitioned to a class III device requiring premarketing applications. Several orthopedic companies have received approval of their bone cement products for purposes other than vertebroplasty since 1976. On October 1999, PMMA was reclassified from class III to class II, which requires future 510(k) submissions to meet “special controls” instead of “general controls” to assure safety and effectiveness. FDA issued a guidance document on July 17, 2002 (http://www.fda.gov/cdrh/ode/guidance/668.pdf) that outlines the types of special controls required and describes the following recommended labeling information:

Kyphon, Inc. has received 510(k) clearance for their bone cement product to be used for vertebroplasty and kyphoplasty as of April 1, 2004. Continuing concern about other cement and bone void filling products led to an FDA Public Health Web Notification that notes the types of complications that can occur with these products, and offers advice to physicians regarding use of such products. FDA requires hospitals and facilities to report deaths and serious injuries associated with the use of such medical devices (U.S. Food and Drug Administration 2004). Use of cement products not receiving FDA clearance specifically for vertebroplasty represents an off-label use.

METHODS

Search Methods

Studies of vertebroplasty were identified through a computerized online search of the MEDLINE® (via PubMed) database through April 15, 2005, using the various textwords including: "vertebroplast*"; "cementoplast*"; and "methylmethacrylate" combined with ("vertebral" OR "spinal"). To identify more recent studies, the MEDLINE® search was supplemented by searches of Current Contents, by manual searches of the most recent issues of the pertinent journals, and by reading the reference lists in the most recently published papers.

Study Selection

Studies were included in the Assessment "Review of Evidence" if the study had these characteristics:

• Full-length article published in the English language
• Population consists of patients with vertebral fractures due to osteoporosis or malignancy
• Patient population is a consecutive series of patients, or near-consecutive series ( <90%)
• Treatment uses percutaneous vertebroplasty with PMMA
• Reports on relevant clinical outcomes of pain, functional status, or quality of life
• Pre- and post-procedure values for outcomes are reported, as quantitative or categorical measures
• Sample size is <20 patients for studies on osteoporosis; <10 for malignancy

Abstracts were not systematically included in the study selection process, and results of case series reported in abstract form only are not included in the Assessment. However, status updates and abstract reports for comparative clinical trials were sought and summarized in this Assessment.

FORMULATION OF THE ASSESSMENT

Patient Indications

1. Patients who have painful vertebral body compression fracture(s) associated with osteoporosis. This indication does not include patients with evidence of spinal cord compression or compromise. Vertebroplasty has been performed on patients with both acute and chronic symptoms due to fractures. Clinical history and appropriate use of imaging tests, which may include MRI, should be used to exclude other causes of back pain and to locate the vertebral bone that is causing the pain. The fracture must be anatomically suited for the procedure.
2. Patients who have painful vertebral fractures associated with osteolytic destruction from malignant disease (e.g., bone metastasis). This indication does not include patients with evidence of spinal cord compression or compromise.

Technologies to Be Compared

Patient Indication #1. For these patients, the usual comparator is continued medical management. The initial treatment for osteoporotic vertebral body compression fractures includes conservative measures such as bed rest, use of an immobilization/bracing device, and analgesic medication, sometimes including narcotic analgesics. Second-line treatment alternatives in this setting are not well established for the specific indication of vertebral fractures and may include exercise, continued conservative treatment, or other methods of pain relief.

Patient Indication #2. These patients seek palliation from pain. Because these patients have limited remaining life span and may have poor health due to the underlying malignant disease, treatment choices are complex. Osteolytic vertebral body destruction may result in pain and/or instability with the potential for neurological compromise from tumor growth and vertebral collapse. Patients may simply be medically managed. Destructive lesions due to malignancy may be treated with local radiation therapy to shrink the tumor and relieve pain. The effectiveness of radiation therapy in reducing pain may be delayed several days to weeks and some tumors may not be radiation sensitive. In addition, prior radiation therapy in that same spinal location may contraindicate further radiation therapy, due to dose limitations. For patients with local spinal instability from extensive destruction, surgical stabilization may be an option; however, patients with extensive malignancy may not be considered suitable candidates for aggressive surgical intervention.

Health Outcomes

Patient Indication #1. The primary health outcomes of interest include pain and ability to function particularly with regard to activities of daily living. Beneficial effects of treatment would include reduction in pain and increased ability to function, which is primarily achieved through decreased pain and increased mobility.

Potential harmful effects include complications associated with the procedure or the associated anesthesia.

Patient Indication #2. The primary health outcomes of interest include pain and ability to function, similar to Patient Indication #1. However, fast response is important for this indication, and durability of relief may be less important given the patients’ limited remaining life span.

Specific Assessment Questions

REVIEW OF EVIDENCE

Overview

The available published evidence describing the outcomes of vertebroplasty consists mostly of uncontrolled studies including case reports and small case series. The case series are mostly retrospective. Beyond the requirements that patients have severe pain attributable to vertebral fractures, it is often difficult to determine additional criteria for patient selection for the procedure. There appears to be no universal criteria used that identify a patient that is suitable or not suitable for vertebroplasty.

There is one nonrandomized study with concurrent controls treated without PVP. Evidence from 2 small, randomized studies evaluating patients with osteoporotic compression fractures has only been reported in abstract form at the 2002 American Society of Neuroradiology meeting in Vancouver, Canada.

In the uncontrolled studies, reported outcomes most commonly included measures of pain, degree of analgesic use, and procedure-related adverse events. Some studies report functional status or activities of daily living. Many studies are limited with regard to reporting detail of outcome measures and results with significant amounts of missing follow-up data in some studies. Most studies are retrospective in nature, and the length of follow-up is relatively short in many studies.

Comparison of reported outcomes across case series studies may not be reliable because known and unknown differences in patient populations are likely to have a confounding influence on observed outcomes. Differences in study methods, methods of evaluating and selecting patients, concurrent treatments, outcome measures used, and differences in reporting may complicate comparisons.

What are the effects of vertebroplasty on health outcomes?

Patient Indication #1. Patients who have symptomatic vertebral body compression fracture(s) associated with osteoporosis

All studies enrolled patients with severe pain, but they varied with respect to the duration of the pain prior to the procedure. Specific criteria regarding patient characteristics, physical examination, or imaging results used to either select or exclude patients are generally not described. All studies used some form of visual analogue scale to report pain outcomes, but varied with respect to the methodology of reporting other outcomes. Characteristics of the included published studies are shown in Table 1. There are 11 case series studies evaluating a total of 907 patients and one nonrandomized comparative trial with 79 patients (55 of whom received PVP).

The studies are generally consistent in that all show statistically significant decreases in pain from an initial starting value between 8–9 on the VAS (or similar score proportionate to the highest possible score) to 2–4, typically within 1 day of receiving the procedure (Table 2). Such pain relief appears to be lasting in the limited number of studies that reported long term outcomes. In the 4 studies that assessed outcomes at about 1 year follow-up, the pain scores were generally in the range of the value achieved shortly after the procedure. However, in the

Table 1. Vertebroplasty for Osteoporotic Vertebral Fractures – Study Characteristics

Abbreviations: See Appendix Table

Table 2. Outcomes of Studies of Vertebroplasty for Osteoporotic Vertebral Fractures [PDF, 257KB]

studies by Zoarski et al. (2002) and Grados et al. (2000) several patients of the original cohort were not available for long term follow-up.

Other outcomes also generally showed improvement after vertebroplasty. Two studies showed significant decreases in analgesic use, and 4 studies showed improvements in either physical function or disability scale scores. One study showed an improvement in a mental functioning score.

In terms of adverse outcomes (Table 3), leakage of the cement outside the vertebral body is a common occurrence, occurring between 19% and 72% in 8 studies that reported its occurrence. The study by Zoarski et al. (2002) only counted “undesirable” cement extravasation as a complication, which apparently did not occur. Of the leaks that occur, a small proportion causes symptoms. The one patient that developed a symptomatic leak in the study by Chen et al. (2004) required surgical decompression; no other studies reported that surgery was necessary. One patient in the study by Cyteval et al. (1999) developed persistent crural pain due to cement leakage. Across all studies, 2 patients had pulmonary emboli. Bleeding and rib fractures are other complications that were uncommonly reported.

In sum, the case series show a consistent and durable (beyond 1 year) improvement in pain scores and other functional scores when compared to baseline. The major limitation of case series evidence is that there is no control group; thus, placebo effects, natural history, and regression to the mean may account for some or all of the apparent benefits of treatment. The studies with comparison groups, reported later in this report, demonstrate the possibility of these other effects.

Patient Indication #2. Patients who have symptomatic vertebral body lesion(s) associated with osteolytic destruction (e.g., bone metastasis). This indication does not include patients with evidence of spinal cord compromise.

Three studies evaluating a total of 70 patients were found that met criteria for minimum sample size and quality of outcome reporting. Although this Assessment reviews fewer studies than the prior TEC Assessment, the quality of the reporting of outcomes is higher in these studies, and the results appear to be qualitatively no different than the prior TEC Assessment. In the prior TEC Assessment, results were typically reported as percent of patients having complete or substantial improvement in pain, and ranged from 60–80%.

Descriptive characteristics of the studies evaluating vertebroplasty for patients with malignant vertebral lesions are shown in Table 4. All patients had severe pain unresponsive to conservative management. All studies evaluated pain relief using VAS both before and after the procedure. However, evaluating the duration of benefit in these patients is problematic because of their very short remaining life span and due to other treatments being performed such as radiation that may

Table 3. Complication Rates of Vertebroplasty for Osteoporotic Fractures

¹Adverse event rates based on %pts with complication
* More stringent definition of leakage defined in this study, none reported

Key:
Leak all cement leaks, with or without any clinical symptoms
Bleed postoperative, clinically significant bleeding
Leak-Sx cement leaks causing clinical symptoms
PE pulmonary embolus
CHF congestive heart failure
MI myocardial infarction
Rib Fx rib fracture

Other Abbreviations: See Appendix Table

Table 4. Vertebroplasty for Vertebral Fractures Due to Malignancy – Study Characteristics

Abbreviations: See Appendix Table

alleviate pain. In all of the studies, there were substantial losses to follow-up due to death as early as 1 month after the procedure.

The change in pain scores was consistent across the 3 studies, showing that mean VAS pain scores went from 7–10 at baseline to 0–3 after the procedure, all changes from baseline statistically significant across all studies (Table 5). Regarding other outcomes, Alvarez et al. (2003) showed that the proportion of fully ambulatory patients improved from 38% to 76%, but the study by Fourney et al. (2003) showed no statistically significant improvement in ambulatory status. The study by Chow et al. (2004) reported that changes in analgesic usage were not statistically significant, and changes in nausea and depression in the Edmonton Symptom Assessment Scale were statistically significant, but specific quantitative results are not reported.

The adverse effects reported in these studies (Table 6) revealed a rate of leakage of cement ranging from 9% to “most,” with a small proportion of the patients with cement leakage having symptoms due to the leak.

How do the outcomes of vertebroplasty compare with outcomes of alternative treatments?

There was one nonrandomized comparison study of vertebroplasty for treatment of osteoporotic fractures. Diamond et al. (2003) enrolled 79 consecutive patients with acute vertebral fractures. All patients were offered vertebroplasty, and those who declined were followed as a comparison group. The 2 groups had balanced baseline characteristics. At 24 hours, the group undergoing PVP (n=55) had much improved pain compared to the control group (n=24). This comparison may not represent equal times between groups, because the 24-hour outcomes in the PVP group were measured at 24 hours after the procedure, but the 24-hour outcomes in the control group were measured 24 hours after enrollment into the study. However, at 6 weeks and between 6 and 12 months there were no differences between groups in pain score. The control group had an identical mean pain score to the PVP group at the end of follow-up. Similar findings were shown for the Barthel index of physical functioning. At long-term follow-up, there was still slightly higher functioning in the group undergoing PVP, but no difference in the percent improvement from baseline between groups. The authors interpret these findings as demonstrating that PVP produced faster resolution of symptoms than conservative management.

No full-text randomized, controlled trial publications were identified comparing outcomes of vertebroplasty with an alternative treatment. However, 2 abstracts presented at the 2002 American Society of Neuroradiology meeting in Vancouver, Canada report randomized controlled trials in patients with osteoporotic compression fractures. These small studies compared PVP with conventional medical management in one study (Do et al. 2002) and with a sham procedure in the other study (Kallmes et al. 2002).

Table 5. Outcomes of Studies of Vertebroplasty for Vertebral Fractures Due to Malignancy

¹ p-values for pre- post- (within-group) comparisons unless otherwise specified

Table 6. Complication rates of vertebroplasty for vertebral fractures due to malignancy

Abbreviations: See Appendix Table

Do et al. (2002) randomized 31 patients with acute painful vertebral body compression fractures to receive either PVP (n=17) or continued medical therapy (n=14). Results on pain (0–10 scale), activity (6-point scale), and analgesic use (0–5 scale) were measured after 6 weeks. Medical therapy patients were permitted to cross over to PVP after completing follow-up. Reported results were:

This study compares outcomes between PVP and continued medical management and finds significant improvements in pain, activity, and analgesic use after PVP. Contrary to the study by Diamond et al. (2003), the control group did not experience improvement over the same time period. However, this trial design does not take into account possible placebo effects. Outcomes such as pain relief and associated improvement in ability to function may respond to placebos.

Kallmes et al. (2002) conducted a small pilot study in 5 patients with subacute vertebral body compression fractures (i.e., less than 2 months in duration). This sham-controlled, randomized trial permitted blinded crossover after 14 days if a patient failed to respond to initial treatment. Three patients were initially assigned to receive the sham procedure and 2 were assigned to receive PVP, but one sham patient suffered a new compression fracture 48 hours after the sham treatment and was excluded from the analysis. Thus, results are available in 4 subjects.

Both subjects initially treated with the sham procedure had minimal relief after treatment and crossed over to PVP, but results after PVP were similar with minimal pain relief observed. Both subjects who initially underwent PVP had minimal relief in symptoms and crossed over to receive the sham procedure. One of these patients reported complete pain relief after the sham procedure. All 5 patients were asked to guess which procedure they had received in order to assess how well the blinding had worked and all guessed that they had received the sham procedure on the first treatment session.

This study provides anecdotal support for the concern that nonspecific placebo effects may affect the results of PVP. Therefore, controlled, comparative studies are important to establish the clinical effectiveness of either vertebroplasty or kyphoplasty due to the potential influence of nonspecific treatment effects (placebo effect) as well as the influence of heterogeneous patient populations and variable natural history of symptomatic vertebral body lesions. The results of the observational study by Diamond et al. (2003) showing immediate short-term improvement within 24 hours are also consistent with a placebo effect of a surgical intervention.

In sum, the evidence from comparative trials of vertebroplasty is not sufficient to determine the efficacy of the procedure compared to continued medical management. All the cited studies enrolled patients with painful acute fractures. The studies by Diamond et al. (2003) and Do et al. (2002) show findings consistent with the case series in that pain relief was achieved in patients receiving PVP over the short term and persisted over the long term. However, the results of the control groups differ, in that the control group in Diamond et al. (2003) showed improvement by 6 weeks while the control group in the study by Do et al. (2002) did not show improvement. Improvement over a 6 week period of time is expected for most patients with acute fractures. The value of improvement at a 24-hour period of time after receiving PVP is uncertain without knowing the trajectory of pain improvement among untreated patients. Finally, the results of a very small pilot trial of patients who received sham procedures raise the issue of placebo effects

SUMMARY OF APPLICATION OF THE TECHNOLOGY EVALUATION CRITERIA

Following is a summary of the available evidence on percutaneous vertebroplasty for vertebral fractures from osteoporosis or malignancy according to the Blue Cross and Blue Shield Association Technology Evaluation Center (TEC) criteria.

1. The technology must have final approval from the appropriate governmental regulatory bodies.

Vertebroplasty is a surgical procedure and, as such, is not subject to U.S. Food and Drug Administration (FDA) approval. Kyphon, Inc. has received 510(k) marketing clearance for a bone cement product to be used in vertebroplasty and kyphoplasty. Other bone cements and bone void filling products used represent an off-label use of such products.

2. The scientific evidence must permit conclusions concerning the effect of the technology on health outcomes.

The available evidence is not sufficient to permit conclusions of the effect of percutaneous vertebroplasty (PVP) on health outcomes.

The published evidence describing the outcomes of vertebroplasty consists mostly of uncontrolled studies. These uncontrolled studies were mostly retrospective and enrolled heterogeneous patient populations. Such studies cannot eliminate placebo and natural history effects as explanations for the apparent effectiveness of PVP. Two studies raise the issue of such effects. In a nonrandomized study, patients undergoing PVP had immediate pain relief from the procedure. However, at 6 weeks of follow-up and at 6–12 months’ follow-up, there was no difference between the group undergoing PVP and another group of patients that had not undergone PVP. In another pilot study reported only in abstract form, patients did not respond to PVP, but did respond to a sham procedure. These studies raise concern that nonspecific placebo effects may be important in determining results following PVP.

Except for these comparative studies, the remaining published literature on outcomes of vertebroplasty consists of case series studies. For the indication of osteoporosis, 11 case studies meeting selection criteria were reviewed that evaluated outcomes of 907 patients. Results were generally consistent in showing significant decreases in pain from an initial preoperative level of 8 to 9 on a visual analogue scale and decreasing to 2 to 4 within one day of the procedure. For the indication of osteolytic destruction due to metastasis, 3 studies evaluating a total of 70 patients were reviewed. Results were generally similar to the studies for osteoporosis, in that mean VAS pain scores went from 7–10 at baseline to 0–3 after the procedure.

Because of the results of the comparative studies suggesting placebo or natural history effects, case series studies are insufficient to make conclusions about the effect of vertebroplasty on health outcomes. Rigorous controlled studies would help determine the efficacy of vertebroplasty.

3. The technology must improve the net health outcome.

4. The technology must be as beneficial as any established alternatives.

The available evidence does not permit conclusions regarding the effect of percutaneous vertebroplasty on health outcomes or compared with alternatives.

5. The improvement must be attainable outside the investigational settings.

It has not yet been demonstrated whether percutaneous vertebroplasty improves health outcomes in the investigational setting. Therefore, it cannot be demonstrated whether improvement is attainable outside the investigational settings.

For the above reasons, percutaneous vertebroplasty for vertebral fractures from osteoporosis or malignancy does not meet the TEC criteria.

REFERENCES

Alvarez L, Perez-Higueras A, Quinones D et al. (2003). Vertebroplasty in the treatment of vertebral tumors: postprocedural outcome and quality of life. Eur Spine J, 12:356-60.

Alvarez L, Perez-Higueras A, Granizo JJ et al. (2005). Predictors of outcomes of percutaneous vertebroplasty for osteoporotic vertebral fractures. Spine, 30:87-92.

Cardon T, Hachulla E, Flipo RM et al. (1994). Percutaneous vertebroplasty with acrylic cement in the treatment of a Langerhans cell vertebral histiocytosis. Clin Rheumatol, 13(3):518-21.

Chen JF, Lee ST, Lui TN et al. (2002). Percutaneous vertebroplasty for the treatment of osteoporotic vertebral compression fractures: a preliminary report. Chang Gung Med J, 25:306-14.

Chen LH, Niu CC, Yu SW et al. (2004). Minimally invasive treatment of osteoporotic vertebral compression fracture. Chang Gung Med J, 27:261-7.

Chow E, Holden L, Danjoux C et al. (2004). Successful salvage using percutaneous vertebroplasty in cancer patients with painful spinal metastases or osteoporotic compression fractures. Radiother Oncol, 70:265-7.

Cotten A, Duquesnoy B. (1997). Vertebroplasty: Current data and future potential. Rev Rhum [Engl Ed], 64(11), 645-9.

Cyteval C, Baron Sarrabere MP, Roux JO et al. (1999). Acute osteoporotic vertebral collapse: open study on percutaneous injection of acrylic surgical cement in 20 patients. AJR Am J Roentgenol, 173(6):1685-90.

Deramond H, Depriester, Galibert P et al. (1998). Percutaneous vertebroplasty with polymethylmethacrylate - techniques, indications, and results. Radiol Clin North Am, 36(3):533-46.

Deramond H, Wright NT, Belkoff SM. (1999). Temperature elevation caused by bone cement polymerization during vertebroplasty. Bone, 25(2) suppl:17S-21S.

Diamond TH, Champion B, Clark WA. (2003). Management of acute osteoporotic vertebral fractures: a nonrandomized trial comparing percutaneous vertebroplasty with conservative therapy. Am J Med, 114:257-65.

Do HM. (2002). Intraosseous venography during percutaneous vertebroplasty: is it needed? Am J Neuroradiol, 23:508.

Do HM, Marcellus ML, Weir RU et al. (2002). Percutaneous vertebroplasty versus medical therapy for treatment of acute vertebral body compression fractures: a prospective randomized study. American Society of Neuroradiology Meeting, Vancouver, Canada, April.

Fourney DR, Schomer DF, Nader R et al. (2003). Percutaneous vertebroplasty and kyphoplasty for painful vertebral body fractures in cancer patients. J Neurosurg Spine, 98:21-30.

Fribourg D, Tang C, Sra P et al. (2004). Incidence of subsequent vertebral fracture after kyphoplasty. Spine, 29:2270-6.

Galibert P, Deramond H, Rosat P et al. (1987). Preliminary note on the treatment of vertebral angioma by percutaneous acrylic vertebroplasty. Neurochirurgie, 33:166-8.

Gaughen JR, Jensen ME, Schweickert PA et al. (2002). Relevance of antecedent venography in percutaneous vertebroplasty for the treatment of osteoporotic compression fractures. Am J Neuroradiol, 23:594-600.

Grados F, Depriester C, Cayrolle G et al. (2000). Long-term observations of vertebral osteoporotic fractures treated by percutaneous vertebroplasty. Rheumatology (Oxford), 39:1410-4.

Harrington KD. (2001). Major neurological complications following percutaneous vertebroplasty with polymethylmethacrylate. A case report. J Bone Joint Surg, 83-A(7):1070-3.

Healey JH. (1997). Metastatic cancer to the bone. In: Cancer: Principles and Practice of Oncology. 5th ed. DeVita VT et al. eds. Lippincott-Raven Co.; Philadelphia: 2577-2586.

Kallmes DF, Jensen ME, Marx WF et al. (2002). A pilot study for a sham-controlled, randomized, prospective, crossover trial of percutaneous vertebroplasty. American Society of Neuroradiology Meeting, Vancouver, Canada, April.

Kasperk C, Hillmeier J, Noldge G et al. (2005). Treatment of painful vertebral fractures by kyphoplasty in patients with primary osteoporosis: a prospective nonrandomized controlled study. J Bone Miner Res, 20:604-12.

Kaufmann TJ, Jensen ME, Schweickert PA et al. (2001). Age of fracture and clinical outcomes of percutaneous vertebroplasty. AJNR, 22:1860-3.

Kobayashi K, Shimoyama K, Nakamura K et al. (2005). Percutaneous vertebroplasty immediately relieves pain of osteoporotic vertebral compression fractures and prevents prolonged immobilization of patients. Eur Radiol, 15:360-7.

Lin JT, Lane JM. (2002). Nonmedical management of osteoporosis. Curr Opin Rheum, 14:441-6.

Martin JB, Jean B, Sugiu K et al. (1999). Vertebroplasty: clinical experience and follow-up results. Bone, 25(2 suppl):11S-15S.

McGraw JK, Heatwole EV, Strnad BT et al. (2002). Predictive value of intraosseous venography before percutaneous vertebroplasty. J Vasc Interv Radiol, 13:149-53.

McGraw JK, Lippert JA, Minkus KD et al. (2002a). Prospective evaluation of pain relief in 100 patients undergoing percutaneous vertebroplasty: results and follow-up. J Vasc Interv Radiol, 13:883-6.

McKiernan F, Faciszewski T, Jensen R. (2004). Quality of life following vertebroplasty. J Bone Joint Surg Am, 86-A:2600-6.

Padovani B, Kasriel O, Brunner P et al. (1999). Pulmonary embolism caused by acrylic cement: a rare complication of percutaneous vertebroplasty. Am J Neuroradiol, 20:375-7.

Ratcliff J, Nguyen T, Heiss J. (2001). Root and spinal cord compression from methylmethacrylate vertebroplasty. Spine, 26(13):E300-2.

Ross PD. (1997). Clinical consequences of vertebral fractures. Am J Med, 103(2A):30S-43S.

Scroop R, Eskridge J, Britz GW. (2002). Paradoxical cerebral arterial embolization of cement during intraoperative vertebroplasty: case report. Am J Neuroradiol, 23:868-70.

U.S. Food and Drug Administration. (2004). FDA Public Health Web Notification: Complications related to the use of bone cement and bone void fillers in treating compression fractures of the spine. Available online at www.fda.gov/cdrh/safety/bonecement.html

Uppin AA, Hirsch JA, Centenera LV et al. (2003). Occurrence of new vertebral body fracture after percutaneous vertebroplasty in patients with osteoporosis. Radiology, 226:119-24.

Vasconcelos C, Gailloud P, Martin JP et al. (2001). Transient arterial hypotension induced by polymethylmethacrylate injection during percutaneous vertebroplasty. J Vasc Interv Radiol, August, p. 1001-2.

Wenger M, Markwalder TM et al. (1999). Surgically controlled, transpedicular methyl methacrylate vertebroplasty with fluoroscopic guidance. Acta Neurochir (Wein), 141(6):625-31.

Wenger M, Markwalder TM. (2002). Cement leakage and the need for prophylactic fenestration of the spinal canal during vertebroplasty. J Bone Joint Surg, 84-A(4):689.

Wilkes RA, MacKinnon JG, Thomas WG. (1994). Neurological deterioration after cement injection into a vertebral body. J Bone Joint Surg, 76-B:155.

Winking M, Stahl JP, Oertel M et al. (2004). Treatment of pain from osteoporotic vertebral collapse by percutaneous PMMA vertebroplasty. Acta Neurochir (Wien), 146:469-76.

Wong W, Mathis J. (2002). Is intraosseous venography a significant safety measure in performance of vertebroplasty? J Vasc Interv Radiol, 13:137-8.

Zoarski GH, Snow P, Olan WJ et al. (2002). Percutaneous vertebroplasty for osteoporotic compression fractures: quantitative prospective evaluation of long-term outcomes. J Vasc Interv Radiol, 13:139-148.

APPENDIX

Table of Abbreviations