Balloon Kyphoplasty and Vertebroplasty for Vertebral Compression Fractures

Transcription

1 Balloon Kyphoplasty and Vertebroplasty for Vertebral Compression Fractures A Comparative Systematic Review of Efficacy and Safety Rod S. Taylor, PhD,* Rebecca J. Taylor, MSc, and Peter Fritzell, MD, PhD SPIE Volume 31, umber 23, pp , Lippincott Williams & Wilkins, Inc. Study Design. Systematic review and meta-regression. Objectives. To compare the efficacy and safety of balloon kyphoplasty and vertebroplasty for the treatment of vertebral compression fractures, and to examine the prognostic factors that predict outcome. Summary of Background Data. A previous systematic review of vertebroplasty by Levine et al in 2000 identified seven case series studies and no controlled studies. Methods. A number of electronic databases were searched through March 1, Citation searches of included studies were undertaken and contact was made with experts in the field. o language restrictions were applied. All controlled and uncontrolled studies were included with the exception of case reports. Prognostic factors responsible for pain relief and cement leakage were examined using meta-regression. Results. The following studies were included: balloon kyphoplasty (three nonrandomized comparative studies against conventional medical therapy and 13 case series), vertebroplasty (one nonrandomized comparative study against conventional medical care and 57 cases series), balloon kyphoplasty versus vertebroplasty (one nonrandomized comparative study). The majority of studies were undertaken in older women with osteoporotic vertebral compression fractures with long-term pain that was refractory to medical treatment. At this time, there is no good quality direct comparative evidence of balloon kyphoplasty versus vertebroplasty. From indirect comparison of case series evidence, the procedures appear to provide similar gains in pain relief while for balloon kyphoplasty there is better documentation of gains in patient functionality and quality of life. The level of cement leakage and number of reported adverse events (pulmonary emboli and neurologic injury) in balloon kyphoplasty From the *Department of Public Health & Epidemiology, University of Birmingham, United Kingdom; Health Economics Facility, University of Birmingham, United Kingdom; and Department of Orthopaedic Surgery, Falun Hospital, Sweden. Acknowledgment date: July 28, First revision date: October 22, Second revision date: January 13, Acceptance date: January 18, The device(s)/drug(s) is/are FDA-approved or approved by corresponding national agency for this indication. Corporate/Industry funds were received in support of this work. One or more of the author(s) has/have received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this manuscript: e.g., honoraria, gifts, consultancies. Supported through unrestricted funding by Kyphon Inc. The planning, conduct, and conclusions of this report are made independently from the company. Address correspondence and reprint requests to Rod S. Taylor, PhD, Department of Public Health & Epidemiology, University of Birmingham, Edgbaston, Birmingham, B15 2TT, United Kingdom; r.s.taylor@bham.ac.uk was significantly lower than for vertebroplasty. These findings were confirmed by meta-regression analysis. Conclusions. There is Level III evidence to support balloon kyphoplasty and vertebroplasty as effective therapies in the management of patients with symptomatic osteoporotic vertebral compression fractures refractory to conventional medical therapy. Although there was a good ratio of benefit to harm for both procedures, balloon kyphoplasty appears to offer the better adverse event profile. These conclusions need to be updated on the basis of the findings of ongoing randomized controlled trials. Key words: spinal fracture, osteoporosis, systematic review, meta-analysis. Spine 2006;31: Vertebral compression fractures (VCFs) constitute a major health care problem in western countries, not only because of the high incidence of these lesions but also due to their direct and indirect negative consequences for patient health-related quality of life and the costs to the health care system. 1 3 Regardless of their etiology, the mainstay of management for symptomatic VCFs is medical therapy that may include analgesics, bed rest, external fixation, and rehabilitation. 4 5 However, such treatments are only partially effective in addressing symptoms, and about one third of patients have been reported to suffer from persistent pain and progressive functional limitation and loss of mobility. 6 In addition, anti-inflammatory drugs and certain types of analgesics are poorly tolerated by elderly patients, while bed rest leads to further demineralization and may predispose to future fractures. Whether painful or not, medical management does not prevent kyphotic deformity. 5 The only surgical treatment for VCFs, refractory to medical therapy, that address deformity has been decompression and stabilization of the fractured vertebrae with different kinds of metal implants. 7 However, because of the poor quality of osteoporotic bone, surgical fixation often fails. 8 Also, because of the risks of open surgery in these often elderly patients, these procedures have generally been limited to cases where there is concurrent spinal instability, or neurologic deficit. 8,9 Balloon kyphoplasty and vertebroplasty are two minimally invasive surgery approaches developed for the management of symptomatic VCFs. 10,11 In balloon kyphoplasty, injection of polymethylmethacrylate (PMMA) follows insertion of a tamp (balloon) into the vertebral body using either a transpedicular or extrapedicular route in order to compress the cancellous bone 2747

2 2748 Spine Volume 31 umber creating a cavity and, if possible, realignment of the endplate of the vertebral body. After the removal of the bone tamp, the PMMA fixes and stabilizes the fracture. Vertebroplasty involves a percutaneous injection of PMMA into the fractured vertebral body, generally through a transpedicular route. Although a number of reviews of literature and commentaries have promoted the use of balloon kyphoplasty and vertebroplasty in patients with symptomatic VCFs from osteoporotic or neoplastic etiology, their failure to use systematic review methods limits their conclusions fundamentally. 6,11 13 The one systematic review to date by Levine et al reported 7 case series studies examining vertebroplasty for VCF but failed to identify any controlled clinical studies. 14 Furthermore, that review completed its literature searches in We are unaware of a published systematic review for balloon kyphoplasty. The aims of this report were threefold: to systematically review the efficacy and safety of balloon kyphoplasty in VCF patients; to update the previous systematic review of the efficacy and safety of vertebroplasty in VCF patients; and, finally, to examine the prognostic factors predicting outcome after the two procedures. Methods The present review was conducted and reported in accord with the recommendations Cochrane Skeletomuscular Collaborative Group (2003) system. 15 Literature Search. A number of bibliographic databases were searched, including MEDLIE (Ovid), MEDLIE (R) Inprocess citations, EMBASE (Ovid); Cochrane Library; registers of ongoing research (i.e., ational Research Register, Meta Register of Controlled Trials, MRC Clinical Trials Register, and ClinicalTrials.gov); and Internet sites of regulating authorities (Food and Drug Administration and European Medicines Evaluation Agency). Searches were conducted from 1983 onwards until first March Search terms were selected in order to maximize both the search sensitivity and specificity. Index and text words representing the device/procedure names were combined with terms for vertebral compression fractures. Hand searching of the reference lists of included studies and reviews was undertaken. Experts in the field (see Acknowledgments), Kyphon Inc., and bone cement manufacturers were contacted to identify any studies that may have been missed, were ongoing, or were unpublished. There were no language restrictions and foreign language papers were translated. Inclusion Criteria and Data Extraction. Studies were included in this review if they met the following criteria: Study Design. Experimental studies (i.e., randomized and nonrandomized trials), observational studies (i.e., cohort studies, case-control studies, or cross sectional studies), and uncontrolled studies (i.e., case series). Population. Patients with VCFs of osteoporotic or neoplastic (i.e., myeloma, metastasis, or osteolysis) etiology. Intervention. Balloon kyphoplasty or vertebroplasty used as a single therapy, or in combination with other therapies. Studies including patients undergoing repeat interventions (balloon kyphoplasty or vertebroplasty) were excluded. Comparator. Any surgical or medical therapy. Outcomes. Studies were required to include information on at least one of the following: efficacy, pain relief, functional capacity, and health-related quality of life; safety, cement leakage, incident (adjacent and nonadjacent) fractures, complications. Case reports were excluded, as were case series published only as abstracts. Study selection was carried out independently by two reviewers (R.J.T., R.S.T.) using a standardized proforma. There was a good level of agreement between reviewers (weighted Cohen s kappa: 0.85, 95% confidence interval, ). A single reviewer (R.J.T.) conducted the data extraction using a standardized proforma that was checked by a second reviewer (R.S.T.). The following information categories were extracted: 1) details of the study population and baseline characteristics of the intervention (and, where appropriate, control groups); 2) details of the intervention and comparator, such as procedural details, bone filler used, and length of follow-up; 3) details of completion rates across the groups, reasons for withdrawal, and loss to follow up; and 4) details of individual outcome results. Results were extracted, where possible as raw numbers, plus any summary measures with standard deviations, confidence intervals and P values. Where there was missing information, study authors were contacted and requested to provide these missing items. Disagreements between the two reviewers were resolved by discussion. Detail of the quality assessment process is provided in the Appendix (available online through ArticlePlus). Data Analysis. In order to obtain a global measure of effect, where possible, the results of individual studies were combined. 17 ot all studies reported on all outcomes of interest. Separate meta-analyses were undertaken for each procedure and outcome. Imputation methods were used to estimate outcome variances where not reported. Dichotomous and continuous outcomes were summarized as proportions, rates or rate ratios (relative risks), and mean differences or standardized mean differences, respectively, and combined using a random effects model. 17 Meta-regression was used to examine the influence of a number of factors on both the level of pain relief and cement leakage outcomes reported across studies. The factors were defined a priori and included: procedure (balloon kyphoplasty or vertebroplasty); average duration of fracture or pain; sample size; study bias (number of biases present); study publication date; indication (i.e., osteoporotic or neoplastic VCF); continent of study (i.e., United States or Europe or other); and duration of follow up (in months). Data are expressed as means and 95% confidence intervals or medians and ranges. All analyses were performed using Stata Software (Stata 8, StataCorp LP, TX). Results Identification and Selection of Studies A total of 487 citations were obtained from searches of the various electronic bibliographies. An additional 37 papers were obtained from either the citation list of included studies or contact with experts. For balloon kyphoplasty, four nonrandomized comparative studies

3 Balloon Kyphoplasty and Vertebroplasty Taylor et al 2749 Excluded on basis of abstract n=387 [not kypoplasty or vertebroplasty, inappropriate indication/outcome, reviews] Abstracts identified from bibliographic d/bases n=487 Full papers retrieved for detailed assessment n=100 Additional papers identified from reference lists & contact with experts/compariies n=37 Excluded studies n=24 case reports n=34 other reasons Figure 1. Summary of study selection and exclusion process. Balloon kyphoplasty Studies included 4 non-randomized comparative studies, 13 case series Vertebroplasty Studies Included 2 non-randomized comparative studies, 57 case series [E1 E4] and 13 case series [E5 E17] met the inclusion criteria, and for vertebroplasty and two nonrandomized comparative studies [E1, E18] and 57 cases series [E19 E78] did so. The study selection process and reasons for exclusions are summarized in Figure 1. Citations for included studies are provided in Appendix A, available for viewing online through ArticlePlus. Comparative Studies Study Characteristics. One prospective study and two retrospective studies compared balloon kyphoplasty and conventional medical care. One prospective study compared vertebroplasty with conventional medical care and one prospective study compared the two procedures (Table 1). Included patients were predominantly women over 65 years who had experienced a symptomatic osteoporotic VCF, the remainder having neoplastic lesions. Although not always stated, the reported duration of pain or fracture age indicated that patients were generally refractory to medical treatment. In the study of Kasperk et al [E2], patients had experienced pain for at least 12 months. In contrast, in the study of Komp et al, patients experienced pain for an average of 34 days [E3]. The quality of the comparative was variable (Table 2, Appendix). Efficacy Outcomes. Compared with medical therapy alone, balloon kyphoplasty significantly improved patients level of pain (P 0.03) and functionality (P 0.006) (Table 2). Although no difference in pain relief compared with medical therapy (P 0.97) was reported in the one vertebroplasty study, an improvement in functional capacity was observed (P 0.03) [E18]. The study of Kasperk et al [E2] reported statistically significant improvements in both vertebral height and kyphotic angle (both P ) following balloon kyphoplasty. In the one study that compared balloon kyphoplasty and vertebroplasty, both procedures appeared to provide a similar level of pain relief after surgery (P 0.68), although there was insufficient patients with follow up in order to comment on 1-year outcome [E1]. Although not reported for vertebroplasty, vertebral height (P ) and kyphotic angle (P 0.001) significantly improved following balloon kyphoplasty. Case Series Study Characteristics. A total of 641 patients (1,070 vertebral fractures) and 3,029 patients (4,861 vertebral fractures) were included in 13 and 57 cases series of balloon kyphoplasty and vertebroplasty, respectively (Table 3). Study sample size varied considerably across the studies, as did study duration, with up to 55 months of follow up available following vertebroplasty. Most patients ( 80%) had experienced painful VCF, as the result of primary or secondary osteoporosis. Only a small proportion of studies were specifically conducted in patients with neoplastic lesions. The duration of the fractures or duration of pain was infrequently reported, although where details were available, the mean across balloon kyphoplasty and vertebroplasty studies of 6.0 and 7.4 months, respectively, indicates that patients were likely to be refractory to conventional medical therapy. The quality of the case series studies was generally poor (Table 4, Appendix). Efficacy Outcomes. There was a significant reduction (P ) in the pooled level of pain following both balloon kyphoplasty and vertebroplasty (Table 5). Following balloon kyphoplasty, statistically significant (P 0.02) improvements in functional capacity (Oswestry Disability Score, Index of Back Function and physical activity levels) were reported [E5, E8, E11]. Only one vertebroplasty study reported to use a validated outcome tool (Oswestry Disability Index) and failed to report numerical outcomes [E77].

4 2750 Spine Volume 31 umber Table 1. Summary of Characteristics and Quality of onrandomized Comparative Studies [E1 E4] Diamond et al (2003) Fourney et al (2003) Weisskopf et al (2003) Komp et al (2004) Kasperk et al (2004) Country Australia United States Germany Germany Germany Design Prospective cohort Retrospective Retrospective Prospective cohort Prospective cohort Population indication; mean age (SD); sex (% male) Osteoporotic VCFs and osteolytic; 67 yr; 33% Osteoporotic VCFs; 70 yr; 12.5% Osteoporotic VCFs; 69 yr; 18% Osteoporotic vertebral compression fractures; 76 yr; 30% Vertebral fractures in cancer patients (myeloma and other malignancies); 64 yr; 56% Age of fracture (mo) 0.25 to 1.5 mo ot reported ot reported Mean 1.2 mo At least 12 mo Duration of pain (mo) (range) ot reported Median 3.2 mo (0.25 to 26 mo) ot reported ot reported At least 12 mo of uncontrolled pain Intervention no. of patients (no. of vertebral fractures) Balloon kyphoplasty (PMMA) 15 (32) Balloon kyphoplasty (PMMA) 22 (37) Balloon kyphoplasty (PMMA) 19 (R) Comparator no. of patients (no. of vertebral fractures) Relevant outcomes Vertebroplasty (PMMA) conventional medical care 55 (71) Conservative medical care 24 Pain, disability, drug utilization, complications Vertebroplasty 34 (65) Pain, vertebral height, kyphotic angle, drug utilization, complications Standard medical care 20 (35) Pain, duration of hospital stay Conventional medical care 17 (R) Pain, vertebral height, disability, satisfaction, complications Balloon kyphoplasty plus conventional medical care (PMMA and Calcibon) 40 (72) Standard medical care 20 (R) Pain, vertebral height, kyphotic angle, disability, analgesia consumption, GP visits, complications Follow-up 6 to 12 mo 4.5 mo Postoperative* 6 mo 6 mo Bias Selection ot present Present Present Can t tell ot present Performance ot present ot present Can t tell ot present ot present Assessment Can t tell ot present Can t tell ot present Can t tell Attrition Can t tell Present Can t tell Can t tell ot present R not reported. *At hospital discharge. Personal communication with study authors; can t tell indicates insufficient detail reported to make a judgment. Of the 6 balloon kyphoplasty studies that assessed health-related quality of life, 4 used the Short-Form 36 (SF-36) [E5, E7, E8, E12], with substantial (P ) overall improvements in 6 of the 8 SF-36 domains. Five vertebroplasty studies reported quality of life, four using a variety of validated health-related quality of life measures, i.e., ottingham Health Profile [E30]; SF-36 [E58, E79]; and Tokuhashi Score [E20]. Of these, 3 reported Table 2. Summary of Efficacy Outcomes Across Comparative Studies Balloon Kyphoplasty ( 3) Vertebroplasty ( 1) Outcome o. of Studies With Reported Outcomes Between-Group Difference mean (95% CI) *; P o. of Studies With Reported Outcomes Between-Group Difference mean (95% CI) *; P Pre to post VAS pain (0 10 mm) Analgesic medication at follow-up (relative risk) Pre to post functional capacity (SD units) Health-related quality of life Pre to post vertebral height (%) Kyphotic angle at follow-up ( ) Satisfaction at follow-up (relative risk) *All based on random effects meta-analysis ( 7.0 to 0.3); ( 2.6 to 2.6); 1.00 ot applicable (0.5 to 1.5); 0.70 ot applicable (0.4 to 1.5); 0.57 ot applicable (8.0 to 1.4); (1.0 to 0.1); 0.03 ot applicable ot reported ot reported (19.5 to 21.0); ot applicable ot reported ( 4.2 to 3.1); ot applicable ot reported (0.6 to 11.6); 0.18 ot applicable ot reported

5 Balloon Kyphoplasty and Vertebroplasty Taylor et al 2751 Table 3. Summary of Characteristics of Case Series Balloon Kyphoplasty ( 13) Vertebroplasty ( 57) Characteristic Median (range); Frequency of Studies/Patients Median (range); Frequency of Studies/Patients Age (yr) 70.5 (58 to 77); (42 to 79); 51 Sex (% male) 28.5 (19 to 70); (0 to 72); 51 Duration of pain or age 5.0 (2.1 to 11); (0.25 to 10.6); 11 of fracture* (mo) Year of publication 2003 (2001 to 2003); (1991 to 2004); 57 Year of data collection 2001 (1999 to 2002); (1991 to 2002); 33 Sample size Patients 34 (7 to 155); 7 37 (2 to 187); 57 Vertebral fractures 66 (8 to 264); 8 48 (2 to 363); 54 Duration of follow-up (mo) 7.4 (0 to 24); (0 to 65); 55 VCF indications Osteoporosis 557 (87%) 2399 (83%) Multiple myeloma 77 (12%) 16 ( 1%) Metastatic lesions 7 (1%) 211 (7%) Metastatic lesion or multiple A 229 (7%) myeloma Hemangiomas 0 38 (1%) Other 0 9 ( 1%) Continent of data collection orth America 9 (69%) 24 (41%) Europe 4 (31%) 20 (34%) Australasia 11 (21%) South America 1 (2%) on-english language publication 3 (23%) 5 (9%) Study setting Single center 12 (77%) 57 (100%) Multicenter 1 (23%) 0 Case mix All osteoporotic 7 (54%) 29 (51%) All neoplastic 2 (15%) 5 (9%) Both osteoporotic and neoplastic 4 (31%) 23 (40%) Outcomes reported Pain relief 10 (77%) 48 (83%) Functional capacity 3 (23%) 13 (22%) Quality of life 6 (46%) 5 (9%) Vertebral height 7 (54%) 3 (5%) Kyphotic angle 3 (23%) 3 (5%) Cement leakage 10 (77%) 37 (65%) ew fractures 4 (30%) 12 (20%) Other complications 6 (46%) 31 (54%) A not applicable. *Studies reported either age of fracture or duration of pain. In some series, these outcomes were not always reported in a usable way for this review. Some studies included both osteoporotic and malignant VCFs but did not report the numbers for each separately. significant (P 0.05) improvement in quality of life following vertebroplasty while one study reported no change [E20]. Given the differences in outcomes, pooling was not possible. Across the 3 before and after studies, there was evidence of a significant increase in vertebral height (P ) and reduction in kyphotic angle (P ) following balloon kyphoplasty [E8, E11, E16]. Four vertebroplasty [E47, E49, E60, E70] studies reported a significant increase in vertebral height (P ) and 3 of these studies [E49, E60, E70] reported on a reduction in kyphotic angle. Three of these 4 vertebroplasty studies analyzed the subset of fractures with vacuums or clefts resulting from bony necrosis due to pseudarthrosis or avascular necrosis of the vertebral body. Changes in vertebral body height in vertebral bodies without clefts were minimal. Safety Outcomes. Safety outcomes were pooled across both comparative studies and case series (Table 6). The rates of adverse events (pulmonary embolism, neurologic complications, and perioperative mortality) were low for both procedures although poorly reported across studies. A significantly higher rate of cement leakages was reported for vertebroplasty than balloon kyphoplasty (P ). o leaks were reported to be symptomatic with balloon kyphoplasty, while some 3% of leaks with vertebroplasty were reported to be symptomatic. For example, the study of Fourney et al [E1], comparing balloon kyphoplasty and vertebroplasty, reported no cement leakages in balloon kyphoplasty patients, while 6 of 65 (9.2%) patients treated with vertebroplasty were reported asymptomatic extravasations. Across all studies, the

6 2752 Spine Volume 31 umber Table 4. Summary of Quality of Case Series Studies Balloon Kyphoplasty ( 13) no. (%) Vertebroplasty ( 57) no. (%) Bias Present Bias ot Present Can t Tell* Bias Present Bias ot Present Can t Tell* Selection bias 0 (0) 12 (92) 1 (8) 0 (0) 54 (95) 3 (5) Performance bias 1 (8) 1 (8) 11 (84) 1 (2) 9 (16) 47 (82) Assessment bias Retrospective or not 0 (0) 11 (85) 2 (15) 5 (9) 51 (89) 1 (2) before/after study ot blinded or independent 0 (0) 2 (15) 11 (85) 5 (9) 5 (9) 47 (82) assessment Validated/objective 0 (0) 9 (69) 4 (31) 5 (9) 13 (23) 39 (68) outcomes not used Attrition bias 3 (24) 5 (38) 5 (38) 6 (11) 12 (21) 39 (68) *Insufficient detail reported to make a judgment. pooled reported incidence of new vertebral fractures, both total and adjacent, was somewhat higher for balloon kyphoplasty than vertebroplasty, although their 95% confidence intervals overlapped. Exploration of A substantial level of heterogeneity was observed in both the level of pain relief ( 2, 644; df, 27; P ) and cement leaks (Q, 1408; df, 31; P ) across studies. The results of the exploration of this heterogeneity are shown in Table 7. Mean preprocedure VAS pain was significant predictor of the change in VAS (P ). However, in the multivariate analysis, no factors were found to be statistically significant predictive of the level of pain relief. In contrast, for the outcome of the risk of cement leakage, the choice of procedure was found to be a highly significant predictor of cement leakage (P ), the level of cement leakage being higher for vertebroplasty than balloon kyphoplasty. Discussion Summary of Findings To the author s knowledge, this is the first published comparative systematic review of balloon kyphoplasty and vertebroplasty for the treatment of vertebral compression fractures. Compared with a previous systematic review of vertebroplasty in 2000, 14 we have identified a number of additional studies. For balloon kyphoplasty, a total of four observational nonrandomized comparative studies and 13 case series studies were identified, and for vertebroplasty two comparative studies and 57 case series studies. All these comparative studies compared the procedures with conventional Table 5. Summary of Case Series Efficacy Outcomes Balloon Kyphoplasty ( 13) Vertebroplasty ( 57) Mean (95% CI)*; P Mean (95% CI)*; P Change in pain VAS ( 4.9 to 3.1); ( 6.5 to 5.0); (0 10 mm scale) Change in functional (0.2 to 2.0); ot reported capacity (SD units) Change in quality of life SF-36 Physical functioning 4 24 ( 33 to 16); (not reported) ot applicable Role physical 3 36 ( 49 to 18); ot reported Bodily pain 4 32 ( 41 to 23); (not reported) ot applicable General health 3 2 ( 6 to 9); ot reported Vitality 4 13 ( 18 to 8); (not reported) ot applicable Social functioning 4 32 ( 43 to 21); (not reported) ot applicable Role emotional 3 20 ( 52 to 13); ot reported Mental health 4 14 ( 18 to 11); ot reported Satisfaction at follow-up 2 98 (90 to 100); (81 to 95); (% of patients) Change in vertebral 3 14 (8 to 20); (8 to 10); ot applicable height (% original height) Change in kyphotic angle (reduction in ) 4 7 ( 5to 10); ( 3to 6); *All based on random effects meta-analysis.

7 Balloon Kyphoplasty and Vertebroplasty Taylor et al 2753 Table 6. Summary of Safety Outcomes Balloon Kyphoplasty ( 13) Vertebroplasty ( 57) o. (%) of Events Rate mean (95% CI) o. (%) of Events Rate mean (95% CI) Cement leakages* Overall 13 90/1,111 (8%) 0.08 (0.05 to 0.11) /1,551 (40%) 0.46 (0.31 to 0.61) Symptomatic 13 0/1,094 (0%) 0 A 7 8/275 (3%) 0.04 (0 to 0.08) ew vertebral fractures Overall 4 56/276 (20.3%) 0.29 (0.08 to 0.49) /626 (10.2%) 0.14 (0.08 to 0.20) Adjacent 4 40/276 (14.5%) 0.26 (0.06 to 0.46) /626 (7.2%) 0.08 (0.03 to 0.12) Adverse events Pulmonary embolism 3 1/291 (0.3%) 0.01 (0 to 0.03) /803 (1.8%) 0.11 (0.03 to 0.19) Spinal cord 2 0/195 (0%) 0 A 14 3/631 (0.5%) 0.05 (0 to 0.40) compression erve root pain/ 5 1/322 (0.3%) 0.02 (0 to 0.05) /1,100 (2.5%) 0.16 (0.01 to 0.27) radiculopathy Mortality Overall 9 34/462 (7.4%) 0.06 (0.04 to 1.00) /1,487 (6.4%) 0.06 (0.03 to 0.08) Perioperative 7 1/406 (0.2%) (0 to 0.04) /680 (2.1%) 0.03 (0.01 to 0.05) A not applicable. *o. of events per vertebrae. o. of events per patient. Defined as perioperative or within 4 weeks of procedure. o. of events/fracture/year or no. of events/patient/year. Random effects meta-analysis. medical management. In addition, one study compared balloon kyphoplasty with vertebroplasty. The quality of both comparative studies and case series studies varied considerably. Over 80% of patients were women of 60 years or over with osteoporotic VCFs. Although not consistently reported across all studies, the majority of vertebral compression fractures would have been refractory to conventional medical treatment (mean duration of pain or mean fracture age across studies, 4 7 months). As the one comparative study of balloon kyphoplasty and vertebroplasty to date was prone to substantial selection bias and confounding, it is not possible at this time to directly compare the two procedures [E1]. However, in this study, patients with more severe VCFs were allocated to balloon kyphoplasty that possibly could have influenced the comparison. Assessment of the relative efficacy and safety of balloon kyphoplasty and vertebroplasty is therefore dependent on an indirect comparison of case series evidence. Case series studies provide evidence of statistically significant improvements in pain relief, functional capacity, and health-related quality of life, up to 2 years, following balloon kyphoplasty. A similar level of pain relief has been reported in vertebroplasty case series that provide up to 5 years of follow up. Few vertebroplasty case series studies have assessed either patient functional capacity or quality of life. Improvements in vertebral height and kyphotic angle were consistently reported with balloon kyphoplasty and in a small number of vertebroplasty case series. However, these results should be interpreted with some caution as studies have used nonvalidated radiographic measurement methods. Furthermore, the vertebroplasty studies were generally conducted in a subgroup of patients with fractures that were readily reducible due to presence of clefts. The mechanism by which the procedures promotes pain relief and enhances patient s health related quality of life remains unclear. For example, a significant association between the magnitude of improvement in vertebral height and kyphotic angle and pain relief could not be determined in the study of Kasperk et al [E2]. A substantially higher level of cement leakage has been reported with vertebroplasty (40%) than balloon kyphoplasty (8%). A number of these vertebroplasty leaks were symptomatic and resulted in pulmonary embolism, nerve root pain, or radiculopathy. It has been suggested that this difference is reflective of the creation of the cavity and the use of more viscous cement in balloon kyphoplasty. 18 Furthermore, we found a comparably low incidence of subsequent VCF fractures with both procedures. The pooling of the results of the two comparative studies who reported incident vertebral fractures showed a significant reduction with balloon kyphoplasty compared with conventional medical care (12/59 vs. 17/37, P 0.022) [E2, E3]. These safety findings are in contrast to two recently published studies suggesting a greater level of complications with balloon kyphoplasty than vertebroplasty 19 and an increased incidence of VCFs following balloon kyphoplasty relative to the natural history for untreated fractures. 20 However, not only do the findings of these two studies have to be seen in the context of the totality of evidence, they have both received considerable subsequent criticism on the basis of their poor methodology. 21,22

8 2754 Spine Volume 31 umber Table 7. Predictors of VAS Pain Improvement and Cement Leakage Univariate* Potential Limitations of the Present Study Multivariate* Pain relief Pre-mean VAS (n 28) Procedure (balloon kyphoplasty vs vertebroplasty) Patient indication (osteoporosis vs neoplastic) (n 27) Average duration of pain or fracture duration (mo) (n 12) Continent of data collection (orth America vs. Europe vs. Australasia vs. South America) Study setting (single vs. multicenter) Publication year Average duration of follow-up (mo) (n 27) Study sample size Study quality (no. of biases present) Cement leakage Procedure (balloon kyphoplasty vs vertebroplasty) (n 31) Patient indication (osteoporosis vs neoplastic) (n 30) Average duration of pain or fracture duration (mo) (n 12) Continent of data collection (orth America vs. Europe vs. Australasia vs. South America) (n 31) Study setting (single vs. multicenter) Publication year (n 31) Average duration of follow-up (mo) Study sample size (n 31) Study quality (no. of biases present) (n 31) *Adjusted for baseline VAS. o. of studies in univariate analysis. Twelve studies in multivariate model. Twenty-seven studies in multivariate model. Quality of Studies Reviewed. The principle limitation in the interpretation of the findings of this review was the absence of randomized controlled trials. Because of the nonrandomized allocation of patients to groups, such studies are prone to substantial selection bias and confounding. 23 Although case series studies are relatively low in the hierarchy of evidence, well-conducted and adequately reported studies can provide useful data on real world effectiveness and the safety of the procedure. 24 Indirect Comparisons. Given the poor quality of the one study that has compared balloon kyphoplasty and vertebroplasty [E1], comparing the efficacy and safety of the two procedures is dependent on the comparison of case series, i.e., indirect comparisons. Indirect comparisons require careful interpretation as outcome differences reported between balloon kyphoplasty and vertebroplasty case series may simply reflect systematic differences between the two groups of studies such as a difference in the case-mix of patients undergoing the two procedures. 25 evertheless, we found little difference in the patient case-mix between the balloon kyphoplasty and vertebroplasty case series, which is supportive of their comparison. Publication Bias. Inevitably, any review can be subject to publication bias, i.e., studies with positive results are more likely to be reported and published, while side effects and adverse events are more likely to be underreported. For example, we observed in this review that a higher proportion of balloon kyphoplasty case series studies (85%) reported cement leakage outcomes, than were reported by vertebroplasty cases studies (69%). We incorporated a number of procedures in order to minimize publication bias: authors of included studies and experts in the field were contacted to ask if they knew of published and unpublished studies we may have missed by our bibliographic searches; we attempted to identify, and exclude, duplicate publications (i.e., studies that included the same patients); and finally, we sought out all results in papers, rather than limiting to only those highlighted by the study authors.. There was evidence of significant statistical heterogeneity across a number of outcomes. Such statistical heterogeneity is reflective of differences in outcomes between studies beyond chance. Potential explanations for such heterogeneity include differences between studies in: their methodologic quality; the characteristics of included patients; the nature of the intervention; and the method of outcome assessment. Using meta-regression methods, we were able to explore the heterogeneity in the level of pain relief and cement leakage rate. Ecologic Fallacy. The method of meta-regression used to explore heterogeneity is this study was based on study level data (e.g., the mean age of patients, the proportion of males, the mean duration of fracture pain). A potential risk of this approach is identifying relationships that may not reflect those at patient level, so-called ecological fallacy. 26 Implications for Further Research A number of randomized controlled trials of balloon kyphoplasty and vertebroplasty are currently under way. The results of these trials should improve the quality of the evidence for both procedures and thereby clarify their place in the management of patients with symptomatic vertebral compression fractures. Future trial evidence needs to address those patients who could gain most from the procedures, taking into account important parameters such as time from fracture, multiple versus single fractures, and degree of kyphotic angle, and also provide data on cost-effectiveness.

9 Balloon Kyphoplasty and Vertebroplasty Taylor et al 2755 Conclusion There is Level III evidence to support balloon kyphoplasty and vertebroplasty as effective therapies in the management of patients with symptomatic osteoporotic vertebral compression fractures refractory to conventional medical therapy. Although there was a good ratio of benefit to harm for both procedures, balloon kyphoplasty appears to have a better short-term adverse event profile. These conclusions require confirmation by randomized controlled trials. Key Points A systematic review and meta-regression analysis of all comparative and noncomparative evidence of the use of balloon kyphoplasty and vertebroplasty in patients with vertebral compression fractures was undertaken. The review identified 3 nonrandomized comparative studies and 13 case series for balloon kyphoplasty and 1 nonrandomized comparative study against conventional medical care and 57 cases series for vertebroplasty. One nonrandomized study directly comparing balloon kyphoplasty and vertebroplasty was also found. Randomized controlled trial evidence is required to confirm the effectiveness of balloon kyphoplasty and vertebroplasty. Appendix available online through Article Plus. Acknowledgments The authors thank Mrs. Sue Bayliss (Department of Public Health & Epidemiology, University of Birmingham) for her assistance in conducting the electronic literature searches, and Dr. Stephan Becker, Dr. Edward Benzel, Professor Yves Boutsen, Professor Olaf Johnell, Professor Christian Kasperk, Dr. Reginald Knight, and Professor Jean-Denis Laredo for their peer review comments on the original report on which the current report is based. References 1. Old J, Calvert M. Vertebral compression fractures in the elderly. Am Fam Physician 2004;69: Hall SE, Criddle RA, Comito TL, et al. A case-control study of quality of life and functional impairment in women with long-standing vertebral osteoporotic fractures. Osteoporos Int 1999;9: Report on Osteoporosis in the European Commission. European Commission, DG employment, industrial relations and Social Affairs, Accessed ovember 1, Rapado A. General management of vertebral fractures. Bone 1996;18(suppl 3): Reginster J, Minne HW, Sorensen OH, et al. Randomized controlled trial of risedronate on vertebral fractures in women with established postmenopausal osteoporosis. Osteoporos Int 2000;11: Philips FM. Minimally invasive treatments of osteoporotic vertebral compression fractures. Spine 2003;28(suppl): Hu SS. Internal fixation in osteoporotic spine Spine 1997;22(suppl): Dickman C, Fessler RG, MacMillan M, et al. Transpedicular screw-rod fixation of the lumbar spine: operative technique and outcome in 104 cases. J eurosurg 1992;77: Essens S, Sacs BL, Drezyin V. Complications associated with the technique of pedicle screw fixation: a selected survey of ABC members. Spine 1993;18: Phillips FM, Wetzel FT, Lieberman I, et al. An in vivo comparison of the potential for extravertebral cement leak after vertebroplasty and kyphoplasty. Spine 2002;27: Liebermann I, Reinhardt MK. Vertebroplasty and kyphoplasty for osteolytic vertebral collapse. Clin Orthop 2003;415(suppl): Truumees E. The roles of vertebroplasty and kyphoplasty as parts of a treatment strategy. Curr Opin Orthop 2002;13: Gross KA Vertebroplasty: a new therapeutic option. Orthop urs 2002;21: Levine SA, Perin LA, Hayes D, et al. An evidence-based evaluation of percutaneous vertebroplasty. Manag Care 2000;9: Cochrane Collaborative Review Groups Musculoskeletal Group. Methods used in reviews. articles/muskel/frame.html. Accessed February 10, Clarke M, Oxman AD, editors. Cochrane Reviewers Handbook [updated October 2001]. In: The Cochrane Library, Issue 4, Oxford: Update Software. Updated quarterly. 17. Egger M, Smith GD, Phillips A. Meta-analysis: principles and procedures. BMJ 1997;315: Heini PF, Orler R. Kyphoplasty for treatment of osteoporotic vertebral fractures. Eur Spine J 2004;13: ussbaum DA, Gailloud P, Murphy K. A review of complications associated with vertebroplasty and kyphoplasty as reported to the Food and Drug Administration medical device related website. J Vasc Interv Radiol 2004; 15: Fribourg D, Yang C, Sra P, et al. Incidence of subsequent vertebral fractures after kyphoplasty. Spine 2004;29: Lieberman IH, Phillips FM, Togawa D, et al. Vertebral augmentation and the limits of interpreting complications reported in the Food and Drug Administration Manufacturer and User Facility Device Experience Database. J Vasc Interv Radiol 2004;15: Harrop JS, Pipa B, Reinhart MK, et al. Primary and secondary osteoporosis incidence of subsequent vertebral compression fractures after kyphoplasty. Spine 2004;29: HS Centre for Reviews & Dissemination. Report 4: Undertaking Systematic Reviews of Research on Effectiveness: CRD s Guidance for Carrying Out or Commissioning Reviews, 2nd ed. York, UK: HS, Taylor RS, Van Buyten JP, Buchser E. Spinal cord stimulation for chronic back and leg pain and failed back surgery syndrome: a systematic review and analysis of prognostic factors. Spine 2005;30: Song F, Altman DG, Glenny AM, et al. Validity of indirect comparison for estimating efficacy of competing interventions: empirical evidence from published meta-analyses. BMJ 2003;326: Lau J, Ioannidis JP, Schmid CH. Quantitative synthesis in systematic reviews. Ann Intern Med 1997;127:820 6.